This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
sta. N
0.1 0.1
• y^%
"A \
bioturbated/bioirrigated sediments
10
100 -2 j - l x
Sediment Carbon Oxidation Rate (mmol m" d" ) Figure 7 The maximum DOC concentration in the upper ~20-30 cm of sediment versus the depthintegrated sediment carbon oxidation rate. Open symbols represent bioturbated/bioirrigated sediments while closed symbols represent more strict anoxic sediments. The two lines "through" these data sets are not meant to imply any functional relationships, but are simply presented here to show the different general trends in the data sets. Data sources: Chesapeake Bay sites M3 and S3—Burdige and Homstead (1994), Burdige and Zheng (1998), and Burdige (2001); California Borderlands and central California margin sites—Berelson et al. (1996) and Burdige et al (1999, unpubHshed data); midAtlantic-shelf/slope break (site WC4)—Burdige et al. (1996, 2000, unpublished data); Cape Lookout Bight, NC (CLB)—Martens et al. (1992) and Alperin et al. (1994); Skan Bay (SB), Alaska—Alperin et al. (1992); station N (see Fig. 2)—Bauer et al. (1995).
anoxic sediments, consistent with the discussion in the previous section regarding the role of macrofaunal processes in affecting pore-water DOC concentrations. At least two possibilities may explain these observation in anoxic sediments. The first is that a balance occurs at depth between DOC production (from sediment POM) and DOC consumption (Alperin et al, 1994; Burdige and Gardner, 1998). This explanation is implicitly incorporated into the ANS model since in Eq. [A-2] the concentration of pLMW DOC at depth (which is essentially [DOC]oo) equals the parameter Qs (= aR^/kp). Based on the model in Fig. 3 Qs is the steady-state concentration of pLMW DOC at depth that is achieved when consumption and production are balanced. If a and kp are roughly constant in these sediments (see Section II.D for further details), then Q^ is proportional to Roo and the observations in Fig. 7 are consistent with this explanation if /?oois positively correlated with Cox (which does not appear to be an unreasonable assumption). Interestingly, Alperin et al. (1999) used a very different modeling approach to examine sediment DOC
626
David]. Burdige
cycling and obtained the following proportionality: [DOCloc oc CoxZ*,
[3]
where z*is the ^-folding depth for sediment organic matter remineralization. Assuming that z* is roughly constant in all of these sediments this equation is also consistent with the observations in Fig. 7. A second explanation for asymptotic DOC concentrations with depth is that DOC production rates go to zero with depth and biotic or abiotic changes in the composition of the refractory pore water DOC pool (^pLMW DOC) continually decreases its reactivity. This would eventually lead to a situation in which pore water DOC found at depth is effectively nonreactive on early diagenetic time scales and is therefore selectively preserved. In this case, one might think of this DOC at depth much like one thinks of "inert" inorganic nutrient end products such as phosphate, ammonium, or X!C^2, which also show similar exponential-like profiles in anoxic sediments (e.g., Bemer, 1980). This analogy would predict that greater amounts of DOC accumulate with depth in sediment pore waters as rates of sediment carbon oxidation increase (e.g., Krom and Westrich, 1981), as is seen in Fig. 7. Additional insights into this possible explanation will be discussed in the next section. Finally, recent studies have shown that sorption of DOC to sediment particles can affect pore-water DOC concentrations (Hedges and Keil, 1995; Henrichs, 1995), and that pore water DOC concentrations may be "buffered" by reversibly sorbed DOC in equilibrium with the pore waters (Thimsen and Keil, 1998). While it is not inmiediately apparent how these processes could explain the results in Fig. 7 they could possibly affect DOC concentrations at depth depending on the intrinsic reactivity of pore water DOC and that which is adsorbed to sediment particles (Lee, 1994; Mayer, 1994b; Henrichs, 1995), the relative sizes of the pore water and sorbed DOC pools (Thimsen and Keil, 1998), and the extent to which sorption sites on particles deposited in a marine sediment are "saturated" by DOC-particle interactions in the water colunm. However, Alperin et al. (1999) recently concluded that the buffering of pore-water DOC concentrations by reversible sorption is not an important controlling factor in explaining pore-water DOC concentrations in these North Carolina continental slope sediments (see Section VI for further details).
D. PORE WATER DOC PROFILES IN DEEP SEDIMENT CORES In contrast to the numerous DOC profiles that have been collected in "shallow" surficial sediments (see Section I.A), far fewer studies have examined porewater DOC concentrations over larger depth and time scales (Starikova, 1970; Nissenbaum et al, 1972; MichaeHs et al, 1982; Chen et al, 1993; Alperin et al.
627
Sediment Pore Waters DOC (mM) 0
1
2
DOC (mM)
DOC (mM) 3
2
4
4
6
8
12
1
^
1 • x
1 L
•
I fit to entire data set
sta. M3, Chsapeake Bay (CH XVII, 8/96)
sta. C, North Carolina continental slope
\*
• •
!
•
fit to upper 115m
1 | •
ODP site 1082, Walvis Basin
Figure 8 Pore-water DOC concentrations versus depth from three contrasting anoxic marine sediments fit to Eq. [4]. The resulting best fit parameters are Hsted in Table II, along with the references to the sources of the data used in these calculations. Note the factor of 4 difference in concentration scales and factor of > 1000 difference in depth scales as one moves from the estuarine Chesapeake Bay sediments to the deep sediment (ODP) cores collected in Walvis Basin. The different symbols for the Chesapeake Bay plot (left) represent replicate cores collected on this date at this site. At site C on the North Carolina continental slope (middle) the data were fit to Eq. [4] starting at a sediment depth of 25 cm based on the observation that the upper portion of these sediments are extensively bioturbated (AlpQTinetaL, 1999). Therefore the term yx inEq. [4] was replaced here with y(x-25) and CQ (at 25 cm) was used as an adjustable fitting parameter. The resulting best value of Co was 1.42 mM versus the measured value of 1.50 mM. Finally, for the Walvis Basin sediments (right) both the entire data set and the upper 115 m of sediment were both fit to Eq. [4]. Although there are factor of ~2 differences in each of the resulting fitting parameters (see Table II), both sets of results are consistent with the general trends discussed in the text regarding the comparison of the fitting parameters from all three sediments.
1999). Interestingly, however, when such profiles are compared with results from shallow sediments (Fig. 8), one observes a general similarity in the exponentiallike shape of the profiles, in spite of significant differences in both the depth and concentration scales for the profiles. Examining these observations in the context of the proportionality between [DOCJooand Cox in Eq. [3] leads to the conclusion that the increasing depth scale over which organic matter remineralization occurs in these sediments (e.g., z*) likely plays a major role in explaining these observations. This point is further reinforced by the fact that at least between site M3 in Chesapeake Bay and site C on the North Carolina continental slope Cox values decrease (~20 vs ~5 mmol m~^ day~^) as [DOC]oo values increase (Fig. 8); Burdige and Zheng, 1998; Alperin et al, 1999). The results from these cores can be used to further examine suggestions posed in the previous section regarding the controls on pore-water DOC concentrations with depth. To do this I initially attempted to "fit" the data from these different
628
David J. Burdige
cores to Eq. [A-2]. However, taking this approach led to the observation that the fit of these data to this equation was insensitive to several of the model fitting parameters (e.g., R^, A, and A:H). In part this occurs because all but one of the exponential terms in Eq. [A-2] decay rapidly with depth, and therefore their actual values have a minimal impact on the overall fit of the data to this equation (see Burdige and Martens, 1990, for a discussion of a similar problem encountered in fitting dissolved free amino acid pore water profiles to an analogous equation). I am currently working to more tightly constrain the parameters in the ANS model with other measured quantities that define sediment carbon remineralization, and therefore limit the number of adjustable, fitting parameters in the model (e.g., see similar discussions in Alperin et al, 1999). Given these observations, Eq. [A-2] can be rewritten by simply removing these terms, yielding [DOC] = (Co - 25)^"^" + 25,
[4]
and since the parameters a and R^o cannot be separated from one another in the equation for Qs (Eq. [A-8]), Eq. [4] actually has only two fitting parameters (A:p [in Eq. [A-9] for y], and the combined parameter aRoo), assuming Co can be fixed with the bottom water value. Equation [4] is very similar to an analogous equation derived by Alperin et al. (1999), although here by explicitly defining the kinetics of sediment DOC production and consumption I am also able to estimate the rate constant for total DOC consumption, which is also essentially the rate constant for the consumption of pLMW DOC. The results of fitting the data from these three sites to Eq. [4] are shown in Fig. 8, where it can be seen that this modified equation does a reasonably good job of fitting data from these very different marine sediments. The resulting rate parameters from these fitting efforts are listed in Table II. As a first observation, I note that the k^ value for Chesapeake Bay sediments is ^ 2 to more than 3000 times smaller than analogous first-order rate constants for the decomposition of monomeric dissolved organic compounds such as acetate or individual amino acids (see Henrichs, 1993, for a summary of these results). This observation is consistent with previous discussions that the bulk of the porewater DOC pool (^pLMW DOC) represents relatively refractory material that turns over much more slowly than components of the mLMW DOC pool. At the same time, the Chesapeake Bay rate constant is of the same order of magnitude as those determined in degradation studies in Alaska coastal sediments of synthetic glutamic acid and alanine melanoidins (0.2-0.7 and <0.09 yr~\ respectively; Henrichs and Doyle, 1986). Since melanoidins have been discussed in the literature as models for refractory marine humic materials (Nissenbaum et al, 1972; Krom and Sholkovitz, 1977; Hedges, 1988) this observation is not surprising. Perhaps more interesting, however, is the six-order-of-magnitude decrease in k^ values as one moves from surficial Chesapeake Bay sediments to deep Walvis
629
Sediment Pore Waters Table II
Best Fit Rate Parameters for DOC Cycling in Contrasting Marine Sediments^ Depth interval offit
POC range {%f
kp (year-^)
aRpo (mM/year)
0-25 cm
4-8 to 3 ^
6.4 ± 2 . 1
1.3 ± 0 . 4
North Carolina continental slope (site C)^
25-225 cm
3 to 2-2.5
1.7 ± 0.7 x 10"^
1.1 ± 0.3 x 10"^
Southwest African Margin (Walvis Basin, site 1082/
1.5-115 m 1.5-370 m
3-6 to 3 ^ 3-6 to 1-3
5.2 ± 1.9 x 10"^ 14 ± 5 x lO'^
4.6 ± 1.2 x 10"^ 9.3 ± 2.9 x 10"^
Site Chesapeake Bay (site M3)^
^See Fig. 8 for the best-fit profiles. ^Surface POC value and value at depth. '^Core collected 8/96. Data from Burdige and Zheng (1998). "DOC data and other model input parameters from Alperin et al. (1999). ^DOC data from Burdige et al. (unpublished data), other model input parameters from Wefer etal.(199S).
Basin sediments. To further examine the significance of this observation I first note that there is a similar large decrease in the parameter a/^oo in these sediments. This decrease is almost certainly due to a decrease in the overall reactivity of the source POM since sulfate reduction dominates organic matter remineralization at all three sites, the amounts of sediment POM at all three sites are comparable (Table II), and it is unlikely that the parameter a undergoes a multiple-order-of-magnitude decrease in its size. This decrease in POM reactivity is not surprising in the context of models for sediment POM remineralization such as the multi-G model or power model (Westrich and Bemer, 1984; Middelburg, 1989), since remineralization of sediment POM continually fractionates this material by preferentially using the most reactive material that is available and continually decreasing the overall reactivity of the remaining material (also see Burdige, 1991a). At first glance then, the observations in Table II might be interpreted as implying that the remineralization of increasingly refractory sediment POM leads to the net production of increasingly refractory pore-water DOC. While this is certainly a possibility, I suggest that these results can equally be interpreted using the scheme shown in Fig. 9. In this explanation I have assumed that within a given time window the bulk of the pLMW DOC produced is ultimately remineralized (i.e., production ^ consumption as discussed above), although a small amount also undergoes internal transformations that lead to a decrease in its reactivity. In part for ease of explanation, I show this as a series of three pLMW DOC pools of decreasing reactivity that are produced from one another (recognizing at the same time that this model and other DOC and POC models that attempt to "quantize"
630
David J. Burdige Pathways ofpLMW-DOC remineralization
^
I
• • • • • • • • • I 'TifmeH^ifOH'S^HHHiHHJI "Time Window 1" • • • • • I Time WinOow r
P O C WB^
HMW-DOCI
"1
pLMW-DOC, m ^
'•"'kp'.'
pLMW-DOC, — ^
I
pLMW-DOCj
\~\^
\"K.
mLMW-DOC
• p Increasing Time, Increasing [pLMW-DOCj] — and Decreasing pLMW-DOMj reactivity (l^i)
Figure 9 A reexamination of the model shown in Fig. 3 in which the reactivity of the pLMW DOM pool is viewed in terms of different subcomponents with decreasing reactivities. In any given time window the majority of the pLMW DOM can be thought of as accumulating in only one of these pools, with pLMW DOM loss then occurring predominantly from remineralization of matrial in this pool. As a result, the k^ value for this pool defines the apparent rate constant for the entire pLMW DOM pool in this time window. At the same time, it is proposed here that during remineralization a small amount of the pLMW DOM in a given pool undergoes internal transformations that decrease its reactivity. Therefore, the overall reactivity of the pLMW DOM pool decreases with time.
or compartmentalize the continuum of these materials and their reactivities are almost certainly oversimplifications). In this formalism then, we would think of pLMW DOC reminerahzation in Chesapeake Bay sediments occurring almost exclusively from this first pool, since reactions transforming pLMW DOCi to pLMW DOC2 would be too slow to be observed to any appreciable extent within the time scale of biogeochemical processes in these estuarine sediments. At the other extreme, in the deep Walvis Basin sediments the remineralization of the more "reactive" pLMW DOCi and pLMW DOC2 would be fast in comparison to the overall time scale of diagenetic processes in these sediments. Therefore, pLMW DOC would accumulate here in the pLMW DOC3 pool. Furthermore, as discussed above (also see Eq. [3]) the significantly longer time scales over which processes in Walvis Basin sediment occur would also lead to the accumulation of higher concentrations of this more refractory DOC. In some sense this observation is analogous to those made by other workers about sediment POC dynamics, namely that the POC degradation rates one observes are strongly dependent on the observational time scales (e.g., Emerson and Hedges, 1988; Middelburg, 1989). In summary, this explanation supports the general observation that a balance between production and consumption roughly controls DOC concentrations at depth. However, it also suggests that a second but equally important aspect of sediment DOC cycling is that with time there are also changes in the pLMW DOC pool (e.g., in its composition and/or structure) that lead to an overall decrease in its
Sediment Pore Waters
631
bulk reactivity. This latter process allows this material to continually accumulate with depth on long (geological) time scales, and may play a role in overall sediment carbon preservation.
III. DISSOLVED ORGANIC NITROGEN (DON) Much of what has been said about DOC in sediment pore waters applies equally well to pore-water DON, in that DON is also a heterogeneous class of organic compounds that range from well-defined biochemicals such as urea or amino acids, to larger dissolved proteins and peptides to more complex (and poorly characterized) N-containing humic and fulvic acids (e.g., Walsh, 1989; Antia et al, 1991; Bronk, Chapter 5). Nitrogen found in DOM can have several types of functionality, although in the water column much of this functionality appears to be in the amide form (-NH2; McCarthy et al, 1997). This issue has not yet been examined in sediment pore waters. In contrast, a much higher proportion of heterocyclic nitrogen functionality (i.e., nitrogen in aromatic rings such as five-member pyrrole-like or six-member pyridine-like structures) has been observed in sediment organic matter (Patience et al, 1992). The occurrence of this material may be due to either selective utilization of amino nitrogen compounds (i.e., selective preservation of heterocyclic nitrogen compounds) or the occurrence of in situ rearrangement (i.e., condensation) reactions that produce new heterocyclic "compounds" from amino-nitrogen compounds. Patience et al. (1992) suggest that the latter suggestion is more likely, based on mass balance considerations, although this issue remains unresolved. The fact that recent sediments contain levels of heterocyclic nitrogen that are comparable to amino nitrogen (Patience et al, 1992) implies that if such rearrangement reactions do occur then they may begin during the early diagenesis of sedimentary organic matter. Based on arguments presented here, this could occur through DON intermediates, although this problem has not been extensively studied (e.g., see discussions in Rubinsztain et al, 1984). The simultaneous determination of DOC and DON in pore waters allows one to examine the C/N ratio of pore-water DOM (= C/NpDOM). gaining further insights into the composition and reactivity of pore-water DOM. Pore-water profiles of DON and C/NpDOMfrom selected marine sediments are shown in Fig. 10, and Table III contains a more detailed sunmiary of C/NpooMvalues and depth trends from a wide range of marine sediments. At least four general conclusions can be made from these observations. Although not directly shown here, the first of these is that when C/NpDOM values in estuarine Chesapeake Bay sediments are compared with the C/N ratio of DOM benthic fluxes, we observe that DOM accumulating in sediment pore waters is depleted in nitrogen compared to that which escapes the sediments as a benthic
DON (|iM) 0
0 I—r-
20
C/N,pDOM
40
60
5
10
15
20
15
20
10
15
20
10
15
20
sta. S3, Chesapeake Bay
s
10
a. 20
0
40
80
.
^
^
120
5
10
• i •l ^
^0
m
sta. M3, Chesapeake Bay
0
u
10
Cu (U
60
5
V>
>r )L Jr ^\ \ v>
^
Q
40
^•^
/^—V
sc>
20
r
20
San Clemente Basin 0
20
40
60
5
Patton Escarpment
Sediment Pore Waters
633
flux (Burdige and Zheng, 1998; Burdige, 2001). This observation will be discussed in further detail in section V.C. A second observation is that in sediments where there is a significant input of terrestrial organic matter, C/NpDOM values tend to be higher (e.g., site N3 in the Chesapeake Bay and ODP core 1075 from the Southwest African Margin). Such observations are consistent with the fact that terrestrially derived organic matter is generally depleted in nitrogen as compared to marine organic matter (C/N values of ~20-80 vs - 6 - 8 ; e.g., Emerson and Hedges, 1988; Meyers, 1997). A third observation is that in oxic or mixed redox sediments C/NpDOM values tend to be relatively low as compared to more strict anoxic sediments. This can be seen at site S3 in the Chesapeake Bay, in the Patton Escarpment sediments, and in oxic, pelagic sediments in the southwest Pacific. In past work we suggested that this may occur as a result of benthic macrofaunal processes that produce specific low C/N ratio organic compounds. Urea (C/N = 0.5) is one such compound (Lomstein et al, 1989; Burdige and Zheng, 1998), although at site S3 in Chesapeake Bay our recent results suggest that urea is not a significant component of the DOM pool in these sediment pore waters (Burdige, 2001). In addition to macrofaunal sources of low C/N ratio pore-water DOM compounds, another possibility is bacterial sources. Bacteria have C/N ratios that range from ~ 3 to 5 (Saunders et al, 1983; Fenchel et al, 1998) and grazing of bacteria by higher organisms (Kemp, 1990; Lee, 1992) could lead to the production of DOM with a low C/N ratio in either oxic or mixed redox sediments. The absence of such bacterial grazing in anoxic sediments would then preclude the production of low C/N ratio bacterial DOM. Much of this low C/N ratio material is likely amino acids (see discussions in Burdige, 2001), although evidence for this in pore water amino acid data from strictly anoxic versus bioturbated/bioirrigated sediments is equivocal at best (see Section IV.B for further details). The final observation based on the results in Table III is that in most of the continental margin sediments examined to date, C/NpDOMvalues increase with depth (also see the San Clemente Basin profiles in Fig. 10). Since there is evidence for the occurrence of terrestrially derived organic matter in these sediments (Steinberg
Figure 10 Pore-water DON concentrations (left) and C/NpooM (right), both versus depth in marine sediments. Symbols on the upper x-axes represent concentrations in bottom-water samples obtained by hydrocasts. (First row) Cores collected at site S3 in the southern Chespapeake Bay in 10/96 ( • ) and 8/97 ( • ) . Data from Burdige and Zheng (1998) and Burdige (2001). (Second row) Cores collected at site M3 in the mesohaline Chespapeake Bay in 7/95 ( • ) and 10/95 ( • ) . Data from Burdige and Zheng (1998). (Third row) Replicate cores collected in San Clemente Basin in 3/94 (Burdige et al, unpublished data). (Fourth row) Replicate cores collected at the Patton Escarpments in 3/94 (Burdige et al, unpublished data). See Burdige et al (1999), McManus et al (1997), and Shaw et al (1990) for general information on the geochemistry of the San Clemente Basin and Patton Escarpment sediments. Note that DOC data from these Chesapeake Bay site M3 cores are shown in Fig. 1, and those from the site S3 cores are shown in Fig. 2.
Table III C/NpDOM Values in Marine Sediments
Site
Maximum sediment depth
Coastal, estuarine sediments Chesapeake Bay 30 cm site MS*^
C/NpDOM
C/NpDOM
(range)
(depth variations)
-12-18
Depth variations less important than seasonal variations (that are out of phase with temperature and sediment remineralization rates) No coherent depth trends
Chesapeake Bay site SS"
30 cm
-8-14
Chesapeake Bay site N^"
30 cm
-10-30
Increase with depth
5 cm
-10-20
20-21 (upper 1 cm), constant below this depth ( = - 1 0 - 1 1 )
Danish coastal sediments^
Continental margin and deep sea sediments Mid-Atlantic shelf/ 30 cm -7-17 slope break (400-750 m water depths)^
Santa Monica Basin (California Borderlands; 900 m water depth)^ San Clemente Basin (California Borderlands; ~2000 m water depth)'^
Increase with depth in most cores
30 cm
-7-18
Increase with depth (7-12 near sediment surface, 7-18 at depth)
30 cm
-7-17
Increase with depth (7-10 near sediment surface, —17 at depth)
Sediment characteristics Anoxic, nonbioturbated; sulfate reduction and methanogenesis dominate organic matter remineralization
Mixed redox conditions; sediments are bioturbated and bioirrigated Sediment organic matter largely terrestrially derived; some bioturbation in upper —5 cm of sediment Shallow water depth (4 m); upper —10 cm of sediment bioturbated Suboxic sediments with minimal bioturbation in the upper 20-30 cm of sediments; nitrate goes to zero in the upper 1-4 cm of sediment; linear sulfate gradients in the upper 25 cm of sediment Anoxic, sulfidic sediments with no bioturbation or bioirrigation
Suboxic sediments; pore water O2 depleted in the upper 1 cm of sediment, nitrate depleted by —2-5 cm sediment depth (Continues)
635
Sediment Pore Waters Table III (Continued)
Site
Maximum sediment depth
C/NpDOM
C/NpDOM
(range)
(depth variations)
Patton Escarpment (eastern North Pacific; ~3700 m water depth)'^
30 cm
-5-15
Minimum value observed—1-2 cm below the sediment surface; increase with depth below
Hatteras continental rise (northwest Atlantic; ~4200 m water depth)^
30 cm
-6-11
No obvious depth trends
-2-7
No obvious depth trends
Southwest Pacific pelagic sediments (~2800-5400 m water depths)^
30-50 cm
"Deep" (ODP) sediment cores (southwest African Margin)^ Lower Congo Basin 60 m 20 ± 7 No consistent depth (core 1075; trends ~3000m water depth)
Walvis Basin (core 1082;-1300 m water depth)
360 m
11.3 ± 3 . 2 No consistent depth trends
Sediment characteristics Pore water O2 depleted at a sediment depth of —2.5 cm, nitrate depleted by —4-10 cm sediment depth; some sediment bioturbation and bioirrigation Suboxic sediments; nitrate depleted by —8 cm sediment depth after an initial increase in the upper 1-2 cm of sediment Oxic sediments; nitrate increases with depth in an exponential-like fashion
Anoxic sediments; sulfate goes to zero by —30 m sediment depth; marine and terrestrial organic matter sources Anoxic sediments; sulfate goes to zero by —20 m sediment depth; predominant marine organic matter sources
^ Data from Burdige and Zheng (1998) and Burdige (2001). ^Data from Lomsteki et al. (1998). '^Data from Burdige and Gardner (1998) and Burdige et al (2000, unpublished data). ^DOC and DON data from Burdige et al. (1999, unpublished data). Other data from Shaw et al. (1990), Berelson et al. (1996), and W. Berelson (pers. commun). ^Data from Heggie et al. (1987). /Data from Suess et al. (1980). ^DOC and DON data from Burdige et al. (unpublished data). Other data from Wefer et al. (1998).
636
David J. Burdige
et al, 1987; Eganhouse and Venkatesan, 1993; Harvey, 1994) one explanation of these C/NpDOM values is that the increased remineralization of terrestrially derived POM becomes increasingly important with depth, leading to the production of DOM with an increasing C/N ratio. Implicit in this assumption (from the standpoint of any of the sediment organic matter remineralization models discussed above) is that N-depleted, terrestrially derived organic matter deposited in these sediments is less reactive than marine-derived POM (e.g., see discussions in Burdige, 1991a). At the same time, decomposition studies with Chesapeake Bay sediments suggest that there is terrestrially derived POM that undergoes remineralization with depth (Burdige, 1991a), yet similar trends in C/NpooMvalues are not observed in these estuarine sediments (Burdige and Zheng, 1998). The reasons for these differences are not well understood, although as discussed in Burdige and Zheng (1998), they suggest that there may not be a tight coupling between the C/N ratio of the sediment organic matter undergoing remineralization and that of its DOM intermediates (or its refractory "end products," e.g., pLMW DOM). Kristensen and Blackburn (1987) observed that the C/N ratio of sediment organic matter undergoing remineralization is not always a good indicator of its reactivity, a fact that could also explain these differences.
IV. DOM COMPOSITIONAL DATA Much of the biologically produced organic matter can (at least operationally) be categorized as carbohydrates, proteins (amino acids), and lipids. In the water column, however, such an approach accounts for less than 15% of the measured DOC (e.g., Williams and Druffel, 1988; Bauer et ai, 1992; Benner, Chapter 3). Much of the DOC in pore waters too is uncharacterized at the molecular or even compound-class level (e.g., total carbohydrates, lipids, etc.; see Burdige, 2001, for further details). Recently, Lomstein et al. (1998) were able to quantify ^40% of the DON in Danish coastal sediment pore waters as dissolved free and combined amino acids, although the details about what these combined amino acids actually represent is still uncertain (see Section IV.B for further details). In 1981, Krom and Westrich observed that "few attempts have been made to identify individual biochemically important compounds in marine pore waters." Here, much has clearly changed in the past 20 years. Looking at these studies characterizing pore-water DOM in the context of the PWSR model, I note that most efforts have focused on examining concentrations and cycling of compounds that fall into the mLMW DOM category (see the discussion below and also see Henrichs, 1993, for a sunmiary). While some work has been carried out examining the dynamics of the HMW DOM pool (Mayer, 1989; Mayer and Rice, 1992; Amosti etai, 1994;Boschker^r«/., 1995; Pantoja and Lee, 1999; Amosti, 2000), less work has been carried out examining its chemical composition.
Sediment Pore Waters
637
A. VOLATILE FATTY ACIDS ( V F A S ) Interest in the study of VFAs in marine sediments stems from the observation that acetate and other short-chain VFAs are important in situ substrates for both sulfate reducing and methanogenic bacteria in anoxic sediments (S0rensen et ah, 1981; Sansone and Martens, 1982; King et al, 1983; Christensen, 1984; Parkes et al, 1989; Wellsbury and Parkes, 1995). Rephrasing this in the context of the PWSR model, this suggests that VFAs represent major components of the mLMW DOC pool through which much of the carbon flow ultimately occurs in anoxic sediments (e.g., see Fig. 3). In most anoxic sediments acetate concentrations range from <1 to up to ~100 /xM (Ansbaek and Blackburn, 1980; Barcelona, 1980; Christensen and Blackburn, 1982; Parkes and Taylor, 1983; Shaw et al, 1984; Crill and Martens, 1986;Novelli^ra/., 1988;Michelson^^a/., 1989; Shaw and Mcintosh, 1990; King, 1991; Mines et al, 1994; Wellsbury and Parkes, 1995; Albert and Martens, 1997). These concentrations increase with depth (in the upper 20-30 cm of sediment) in most of these sediments. In many of these studies total DOC concentrations were generally not determined along with acetate, although where both measurements were made (or where DOC values can be obtained from other published studies) I estimate that acetate accounts for ~ 5 % (at most) of the total pore-water DOC and often times less than 1% of the DOC. Concentrations of other VFAs such as formate or propionate have generally not been determined as frequently as acetate, although based on the available literature, concentrations of these other VFAs can be comparable to those of acetate (e.g., Barcelona, 1980; Albert and Martens, 1997). Many of the studies discussed above have also concluded that much of the acetate and other VFAs that can be chemically measured may not be biologically available (see Henrichs, 1993, for a sunmiary of these studies). This observation further reinforces the notion that material in the mLMW DOM pool represents a small fraction of the total sediment pore-water DOM pool whose concentration is held at relatively low levels due to rapid bacterial turnover. There are two interesting sedimentary environments where pore water acetate concentrations do not follow these general trends. The first is in the anoxic sediments of Cape Lookout Bight, NC, in the transition zone between where sulfate reduction and methanogenesis predominate (Sansone and Martens, 1982). Here it is observed that during the warm summer months a distinct subsurface maximum in acetate develops at the sediment depths where sulfate concentrations go to zero. At these depths acetate concentrations are as high as ^ 2 mM and account for -^25% of the total DOC (Crill and Martens, 1986; Alperin et al, 1994; Albert and Martens, 1997). Pooling of acetate in this sediment transition zone appears to occur as seasonal temperatures rise, rates of sulfate reduction increase, and the zero sulfate horizon in the sediments moves upward toward the sediment surface. As a result of these changes, acetate consumption by sulfate-reducing bacteria
638
David]. Burdige
apparently decreases at the base of the sulfate reduction zone, and a lag in the development of a methanogenic population in this portion of the sediments that can consume this acetate leads to this transient subsurface acetate maximum (Alperin et al, 1994). The extent to which this phenomena occurs in other organic-rich, anoxic sediments has not been determined. A second sedimentary environment where significantly elevated acetate concentrations have been observed are deeply buried marine sediments (the so-called deep marine biosphere; Parkes et ai, 1994). In sediments on the Blake Ridge in the Atlantic Ocean, temperatures increase from deep-sea values (~2-3°C) to ~20-30°C within several hundred meters of sediment burial, and this apparently stimulates microbial activity in the sediments and also leads to a significant increase in the reactivity of the sediment organic matter found at depth. These depth changes results in pore-water acetate concentrations increasing with sediment depth below ^300 m, such that by 700 m acetate concentrations are as high as 15 mM (Wellsbury et al, 1997). The significance of this process in terms of the overall dynamics of the deep marine biosphere will require further study.
B. AMINO ACIDS Amino acids represent significant amounts of the carbon and nitrogen that are remineralized in marine sediments (Henrichs and Farrington, 1987; Burdige and Martens, 1988; Cowie and Hedges, 1992b), and pore water dissolved amino acids are likely important intermediates in this process (see references cited below). Amino acids in pore waters have been examined in a wide range of marine sediments including salt marsh, estuarine, and coastal sediments (Gardner and Hanson, 1979; Henrichs and Farrington, 1979; J0rgensen et ai, 1981; Caughey, 1982; Henrichs and Farrington, 1987; Burdige and Martens, 1990; Colombo etal, 1998; Landen and Hall, 1998; Lomstein et ai, 1998) and continental margin and deepsea sediments (Henrichs and Farrington, 1980; Henrichs et ai, 1984; Haberstroh and Karl, 1989; Landen and Hall, 2000). Several of these studies have also examined amino acid adsorption to marine sediments as well as dissolved amino acid turnover rates in sediments (also see Rosenfeld, 1979; Christensen and Blackburn, 1980; Sugai and Henrichs, 1992; Henrichs and Sugai, 1993). In most sediments, concentrations of total dissolved free amino acids (TDFAAs) decrease with sediment depth, with surface pore water concentrations generally ranging from ^20 to 200 /xM and concentrations below 10-20 cm being less than 5-10 /xM. In the few studies where total DOC and DON have been examined along with amino acids, TDFAAs represent 1-13% of the DON and <4% of the DOC (Henrichs and Farrington, 1987; Lomstein etai, 1998; Landen and Hall, 2000). The predominant amino acids in the DFAA pool include glutamic acid, alanine, glycine.
Sediment Pore Waters
639
aspartic acid, and, in some sediments, the nonprotein amino-acid ^-aminoglutaric acid ()S-aga), an isomer of glutamic acid (see the next paragraph for further details). For the protein amino acids the DFAA pool is generally enriched in glutamic acid relative to the sediment hydrolyzable amino acid pool. Other nonprotein amino acids in addition to )S-aga can also be enriched in pore waters relative to the sediments (e.g., )S-alanine). These (and other) compositional differences between pore water and sediment amino acids are likely related to both biological and physical (e.g., adsorption) processes that affect pore water amino acids as they undergo remineralization (see discussions in Henrichs and Farrington, 1987; Burdige and Martens, 1990). The nonprotein amino acid ^-aga was first observed in pore waters by Henrichs and Farrington (1979). In anoxic sediments they noted that the mole percentage of )S-aga generally increases with depth (upper ^30 cm), often becoming the predominant DFAA (>~40 mol %; also see Henrichs and Farrington, 1980,1987; Henrichs et al., 1984; Burdige and Martens, 1990). In contrast, ^S-aga appears to be a much less important component of the pore water DFAA pool in sediments that are described as either "oxic" or, more likely, are bioturbated and/or bioirrigated (Henrichs and Farrington, 1980; Caughey, 1982; Landen and Hall, 1998, 2000). The source(s) of ^^-aga in anoxic sediment pore waters is not well characterized (Henrichs and Cuhel, 1985; Burdige, 1989), although its relative accumulation with depth in these sediments may occur because it is more refractory than other amino acids (Henrichs and Farrington, 1979). The observed differences in relative concentrations of j6-aga in anoxic versus oxic/mixed redox sediments suggests that the proposed refractory nature of )S-aga could also be a function of sediment redox conditions, similar to that which has been observed for other refractory components of pore water DOM (Burdige, 2001). Along the same lines, the observations in Section III suggest that there could be broader compositional differences in the dissolved amino acid pools in anoxic versus oxic/mixed redox sediments that might help explain the low C/NpDOM values in these latter sediments. Focusing on the amino acid glycine (C/N = 2), my examination of the pore water amino acid data in the references cited above suggests that glycine may be preferentially enriched in the pore-water amino-acid pool of mixed redox sediments versus more strict anoxic sediments. Unfortunately, however, given the way some of these data are presented in the literature, this conclusion should be viewed as tentative. Furthermore, glycine is an abundant amino acid in many benthic invertebrates (Awapara, 1962; Henrichs, 1980), and the possibility exists that these elevated glycine levels might simply result from release of this amino acid by benthic organisms during core collection and/or pore water processing (J0rgensen et al, 1981; Burdige and Martens, 1990). Nevertheless, these glycine results appear to be consistent with the C/NpDOM values discussed in Section III and would therefore appear to warrant further examination (also see discussions in Burdige, 2001).
640
David J. Burdige
Finally, another pool of dissolved amino acids that has been given little study is that of dissolved combined amino acids (DCAAs). These represent amino acids that are either found in dissolved peptides/proteins or incorporated into abiotic (humic?) compounds. In the few studies of DCAAs in marine sediment pore waters (Caughey, 1982; Colombo et al, 1998; Lomstein et al, 1998; Pantoja and Lee, 1999); Burdige, unpub. data) concentrations of total DCAAs appear to be ^l.5-A times that of total DFAAs. Given the small data set on pore-water DCAAs it is difficult to determine whether the DCAA pool compositionally looks more like DFAAs or hydrolyzable sediment amino acids. Such information could be important in determining the extent to which the DCAA pool represents dissolved peptides or proteins (i.e., "reactive" high-molecular-weight intermediates of sediment organic matter remineralization) or abiotic condensation products of, e.g., melanoidin-type reactions. In the former case, the DCAA pool might be expected to be more similar to that of sediment amino acids, while in the latter case DCAAs might be expected to look more like DFAAs. At the same time nonprotein amino acids such as )S-alanine and y-aminobutyric acid are often times enriched in the free amino acid pool of many sediment pore waters, and become increasingly important in the sediment organic matter pool as this material becomes increasing diagenetically altered (Whelan, 1977; Cowie and Hedges, 1994; Cowie etal, 1995; Hedges etal, 1999). Thus, information on a possible linkage between the occurrence of these amino acids in different pore water pools (DFAAs vs DCAAs) and their preservation in sediments could be important in understanding carbon preservation in sediments in general.
C. CARBOHYDRATES Carbohydrates represent a significant component of the pools of living and nonliving organic matter in the marine environment. In marine sediments, particulate carbohydrates (PCHOs) account for 10-20% of the total sediment POC, and a similar percentage of the sediment POC undergoing remineralization (see recent discussions in Burdige et al, 2000). Despite the apparent biological lability of carbohydrates, some fraction of the chemically recognizable carbohydrates deposited in marine sediments escape remineralization and are "preserved" on time scales of decades to >1 million years (Cowie and Hedges, 1992a; Martens et al, 1992; Whelan and Emeis, 1992; Cowie et al, 1995; Burdige et al, 2000). Interest in the study of pore-water carbohydrates therefore stems from a desire to better understand their role in carbohydrate remineralization and carbohydrate (and perhaps total carbon) preservation. Total dissolved carbohydrates (DCHOs) have been determined in a limited number of coastal and continental margin sediments, with concentrations that
Sediment Pore Waters
641
generally range from ~10 to 400 /xM C (Lyons et al, 1979; Amosti and Holmer, 1999; Burdige et al, 2000). In most cases DCHO concentrations increase with sediment depth (upper '^20-30 cm) and represent ^10-40% of the total DOC. Higher percentages have been observed just below the sediment-water interface in Cape Lookout Bight, North Carolina, sediments (Amosti and Holmer, 1999) and in some Bermuda carbonate sediments (Lyons et al, 1979). Relative DCHO concentrations generally decrease with sediment depth (again upper ^^20-30 cm), although the magnitude of these decreases vary among the few sites that have been examined to date (see discussions in Burdige et al, 2000, for further details). Concentrations of dissolved carbohydrates in sediment pore waters are uncoupled from particulate (sediment) carbohydrate concentrations, and DCHO concentrations appear to be more strongly controlled by sediment remineralization processes (Burdige et al, 2000). Evidence-to-date also suggests that DCHOs may be preferentially found in the high-molecular-weight (HMW) pore-water DOC pool and therefore likely represent some of the initial HMW intermediates produced and consumed during sediment POC remineralization (Amosti and Holmer, 1999; Burdige ^r a/., 2000). Burdige etal (2000) determined individual aldoses (monomeric neutral sugars) in selected pore-water samples from Chesapeake Bay and mid-Atlantic shelf/slope break sediments. We observed that ~30 to 50% of the DCHOs in these pore waters could be identified as individual aldoses, and that total aldose yields (total individual aldose concentrations as a percentage of DOC) were higher in the continental margin sediment pore waters {^9%) than they were in the estuarine sediment pore waters (<5%). Dissolved glucose was the predominant aldose (28%), with the following order of abundance for the remaining aldoses: xylose (16%); fucose, rhamnose, and galactose (each average value between 12 and 14%); mannose (10%); arabinose (7%). These aldose distributions are different than those observed in the water column (e.g., Skoog and Benner, 1998), consistent with the suggestion that pore-water DCHO concentrations are controlled by sediment remineralization processes.
V. THE ROLE OF BENTHIC DOM FLUXES IN THE OCEAN CARBON AND NITROGEN CYCLES A. BENTHIC DOC FLUXES The fact that concentrations of both DOC and DON in sediment pore waters are elevated over bottom waters implies that sediments could be a potential source of DOM to overlying waters. Prior to the early 1990s the possible occurrence of these processes was discussed in the literature (Emerson and Dymond, 1984; Heggie et al, 1987; Bender et al, 1989), and in these and other works (Williams
642
David J. Burdige
and Druffel, 1987; Mopper et al, 1991; Hedges, 1992) it was suggested that sediments might represent a significant source of DOC to the deep ocean. In these works it was also proposed that benthic DOC fluxes might provide an explanation for the apparent discrepancy between the "old" (^^6000 ybp) ^^C age of deepwater DOC, the average oceanic mixing time (^^1000 years), and other chemical properties of deep-water DOC (i.e., 5^^C and lignin concentrations) that suggest that this material is primarily of marine origin. It was not until recent years that direct measurements of these fluxes were undertaken using either core incubation techniques or in situ benthic landers or chambers (Hall et ai, 1990; Burdige et al, 1992, 1999; Burdige and Homstead, 1994;Hulth^ra/., 1997; Burdige and Zheng, 1998; Alperin^ra/., 1999; Holcombe et ai, 2001). To date, benthic DOC fluxes have been determined in this manner in a number of estuarine, coastal and continental margin sediments, with values that range from ^0.1 to 3 mmol m~^ day"^ Recently, we compiled these results and observed that there was a positive, but nonhnear, relationship between benthic DOC fluxes (BDF) and depth-integrated sediment carbon oxidation rates (Cox) expressed as BDF = 0.36Cox^-^^
[5]
(BDF and Cox both expressed here as nmiol m"^ day~^; Burdige et al, 1999). The relationship between BDF and Cox impUes that the ratio of BDF to Cox decreases with increasing Cox (also see Table IV). With these results we were able to estimate the global significance of benthic DOC fluxes in oceanic and sediment carbon cycUng (Table IV). This calculation suggests that the integrated DOC flux from coastal and continental margin sediments (0-2000 m water depth) is ^^190 Tg C year"^ This quantity is larger than (though of roughly similar magnitude) a similar estimate by Alperin et al. (1999), who used a subset of the data we used and also assumed that benthic DOC fluxes were a constant fraction (1.6 ih 0.2%) of sediment carbon oxidation rates. Their approach yields an integrated DOC flux of ^^40 Tg C year~^ from this same portion of all marine sediments. The implications of our results are several-fold. First, it can be seen that DOC fluxes from coastal and margin sediments are generally less than ^10% of sediment carbon oxidation rates. Thus, sediments appear to be quite efficient in retaining (oxidizing) DOM produced during remineralization processes, consistent with past discussions here. Similar trends also appear to be the case for sediment DON cycling (see next section) and both observations imply that net sediment DOM production is small in comparison to gross sediment DOM production. This occurs because under steady-state conditions the former (net DOM production) is balanced by, or equals, the benthic DOM flux (e.g., see general discussions in Bemer, 1980), and the latter (gross DOM production) is approximately equal to gross DOM consumption or inorganic nutrient production (which results in
643
Sediment Pore Waters Table IV Benthic DOC Fluxes from Coastal and Continental Margin Sediments"
Sediment regime^
Benthic DOC flux Sediment carbon oxidation (Cox) Integraiea^ Average^ Integrated*^ Average*^ (Tg C/year) (mmol/m'^/day) (mmol/m^/day) Tg C/year %ofCox^
"Coastal" Sediments (0-200 m; 9%)
1630 (52%)
14.7
0.91
88 ± 3 0
5.4
"Margin" Sediments (200-2,000 m; 7%)
940 (30%)
6.6
0.65
89 ± 2 5
9.5
177 ± 56
6.9
Coastal plus Margin Sediments (0-2000 m)
2570 (82%)
Note. Note that Tg C/year = 10^^ g C/year. ^ Adapted from Burdige et al (1999). ^The percentage of all marine sediments found in each sediment regime is listed in parentheses. ^ Taken from Midddelburg et al (1997), who estimated globally integrated rates of sediment processes in these regimes using an extensive database of published rates of sediment biogeochemical processes. Listed in parentheses in this column is the integrated Cox in each region as a percentage of the integrated Cox for all marine sediments (= 3130 Tg C/year). ^Obtained by dividing the integrated Cox in each region by the sediment surface area in the region. ^Obtained using the average Cox in each region and Eq. [5] in the text. ^Obtained by multiplying the average benthic DOC flux in each region by the sediment surface area in the region. ^The average benthic DOC flux as a percentage of the average Cox in each sediment regime.
inorganic nutrient benthic fluxes). These trends are consistent with discussions in Section LA regarding carbon and nitrogen flow through DOM intermediates during sediment POM remineralization and the role of DOM as an intermediate in the remineralization process. A second implication of the results in Table IV is that the integrated benthic DOC flux reported here is comparable to estimates of the organic carbon burial rate in all marine sediments (~160 Tg C year"^; Hedges and Keil, 1995) and the riverine DOC input (200 Tg C year"^ Meybeck, 1982). Thus, as has been noted previously (Burdige et al, 1992; Burdige and Homstead, 1994) marine sediments may be an important net source of DOC to the oceans. However, the actual impact these fluxes have on the oceanic carbon cycle ultimately depends on the extent to which sediment-derived DOM is reactive or refractory in the water column. This point will be discussed in greater detail in Section V.C. Finally, Burdige et al. (1999) noted that there might be additional significance to the fact that integrated benthic DOC fluxes and sediment carbon burial rates
644
David J. Burdige
are of comparable magnitude. Since pore-water DOC cycling may play a role in sediment carbon preservation (see Section VI), these observations suggest a linkage between benthic DOCfluxesand sediment carbon preservation "mediated" by pore water DOC concentrations and cycling. Changes in the factors controlling benthic DOC fluxes (and/or the cycling of DOC in sediments) may then affect or control sediment carbon preservation (also see similar discussions in Hedges and Keil, 1995; Henrichs, 1995). The exact details of the relationship between benthic DOC fluxes and sediment carbon preservation await further studies.
B. BENTHIC DON FLUXES Interest in benthic DON fluxes and their role in the marine nitrogen cycle is similar to that discussed above for benthic DOC fluxes and the marine carbon cycle. However, because nitrogen can be a limiting nutrient in marine ecosystems (Carpenter and Capone, 1983), and because marine phytoplankton can use DON as their nitrogen source (Jackson and Williams, 1985; Antia et al, 1991; Bronk, Chapter 5), there has additional interest in understanding the role of sediments as a source of DON to the water column. Several studies have examined benthic DON fluxes from coastal and estuarine sediments (Hartwig, 1976; Nixon et ai, 1976; Boynton et ai, 1980; Nixon, 1981; Nixon and Pilson, 1983; Hopkinson, 1987; Teague et al, 1988; Doflar et aL, 1991; Enoksson, 1993; Cowan and Boynton, 1996; Burdige and Zheng, 1998), and two studies have also examined these fluxes from high-latitude continental margin sediments (Blackburn et ai, 1996; LandenHillmeyr, 1998). Benthic DON fluxes show quite a tremendous range both in absolute magnitude and direction (into and out of the sediments). However, at estuarine or coastal sites where repeated (or seasonal) studies have been carried out, mean or annual averages generally suggest that these fluxes are small, usually out of the sediments, and only a small percentage of the benthic DIN flux (see Burdige and Zheng, 1998, for more details; DIN, dissolved inorganic nitrogen, i.e., ammonium plus nitrate+nitrite). For example, in Chesapeake Bay sediments benthic DON fluxes range from ^^0.08 to 0.2 mmol m~^ day~^ in the anoxic sediments at site M3 and essentially zero to 0.4 mmol m"-^ day~^ in the bioturbated and bioirrigated sediments at site S3 (Burdige and Zheng, 1998; Burdige, 2001). However, at both sites benthic DON fluxes were only ^3-4% of benthic DIN fluxes. Interest in benthic DON fluxes also stems from a desire to better understand the relative importance of these fluxes in sediment nitrogen budgets relative to sediment denitrification rates (Nixon, 1981; Boynton and Kemp, 1985; Bender et ai, 1989; Kemp et ai, 1990; Devol and Christensen, 1993; Blackburn et al, 1996; Landen-Hillmeyr, 1998). This problem is of some significance in continental
Sediment Pore Waters
645
margin sediments since here denitrification is generally thought to be an important (and perhaps the major) sink for combined nitrogen in the entire marine environment (see, e.g., Codispoti, 1995, and references therein). However, until recently, sediment denitrification rates were generally determined indirectly and used to balance sediment nitrogen budgets by implicitly assuming that benthic DON fluxes were of minor significance. At site M3 in Chesapeake Bay we were able to examine this problem in a slightly different fashion, using data on sediment nitrogen deposition and burial, DIN and DON benthic fluxes, and several independent estimates of sediment denitrification rates to develop a sediment nitrogen budget (Burdige and Zheng, 1998). This approach led us to conclude that here benthic DON fluxes are less than ^20% of sediment denitrification, based on an integrated annual average benthic DON flux of 0.18 ± 0.07 mmol m"-^ day~^ and estimates of sediment denitrification that range from 0.8 to 2.4 mmol m"^ day"^ (Burdige and Zheng, 1998). In contrast, in Arctic sediments (water depths 170 to 2600 m) Blackburn et al. (1996) observed benthic DON fluxes (0.93 mmol m~^ day"^) that were substantially larger than either benthic DIN fluxes (0.1 mmol m"-^ day~^) or sediment denitrification rates (0.03 nmiol m~^ day~^). However, these workers also suggested that these large DON fluxes may have been a temporary phenomena associated with the recent sedimentation of fresh detrital material. At the same time, in sediments of the Skagerrak (a continental margin region of the North Sea), Landen-Hillmeyr (1998) reported benthic DON fluxes that range from ^0.3 to 1.4 mmol m'^day"^ and were again substantially larger than benthic DIN fluxes. The ratio of benthic DON fluxes to denitrification rates ranged from 0.3 to 17.5 at the five sites they studied, although without the high value of 17.5, this ratio was ^^1.5. In all but one of these sites denitrification rates were determined indirectly from benthic DIN fluxes and integrated sediment ammonium production rates, although at the one site where denitrification rates were directly determined, the ratio of benthic DON fluxes to denitrification rates was 1.2. The reasons why these very different results were obtained in anoxic Chesapeake Bay sediments versus these high-latitude continental margin sediments is unclear at the present time. Whether this is due to differences in sediment setting (estuarine versus continental margin), geography (temperate versus high latitude sites), or sediment redox conditions (anoxic versus more mixed redox) will require further study. Interestingly, results from irrigated Washington state continental margin sediments suggest that benthic DON fluxes are not an important component of nitrogen cycling in these sediments (Devol and Christensen, 1993). These workers were able to essentially balance sediment nitrogen budgets at these sites using only inorganic nitrogen benthic fluxes and measured denitrification rates (determined with benthic lander N2 measurements). However, given the fact that benthic DON fluxes do appear to be significant relative to sediment
646
David J. Burdige
denitrification rates in at least some continental margin sediments suggests that this problem should be further examined to more conclusively resolve the reasons for these differences. C. THE EXTENT TO WHICH BENTHIC D O M
FLUXES AFFECT
THE COMPOSITION AND REACTIVITY OF DEEP-WATER
DOM
Interest in DOM fluxes from marine sediments stems in part from a recognition of the need to better understand the sources and sinks of DOM in the oceans (Hedges, 1992; Carlson, Chapter 4; Cauwet, Chapter 12; Hansell, Chapter 15; Hedges, Chapter 1; also see Section V. A). Although benthic DOC fluxes appear to be of similar magnitude as other DOC inputs such as riverine inputs (Burdige et al, 1999) the impact of these fluxes on water-column DOC concentrations and properties depends on the reactivity of sediment-derived DOC in the water colunm. If this material is sufficiently refractory, then these fluxes could represent an important source of DOC to deep waters and might help explain some of the observations about deep-water DOC discussed in Section V. A. Conversely, if sediment-derived DOC is reactive in the water column (i.e., it undergoes remineralization on time scales shorter than deep-water residence times) then it will have minimal impacts on deep-water DOC properties (Alperin et aL, 1999). Several lines of evidence from contrasting marine sediment suggest that not all of the sediment-derived DOM (i.e., from benthic fluxes) is refractory and that it does have the potential to be reactive in the water column. In estuarine Chesapeake Bay sediments a comparison of measured benthic DOM fluxes versus calculated, diffusive DOM fluxes suggests that there is enhanced production of N-rich DOM at or near the sediment-water interface (Burdige, 2001). This is consistent with the observation that DOM accumulating in these sediment pore waters is carbon-rich (C/NpDOM > 10) relative to the more N-rich DOM that escapes the sediments as a benthic flux (which has a C/N ratio of ~4-6 at site M3 and ^^2-4 at site S3; Burdige and Zheng, 1998). Similar trends in DOM elemental ratios have been observed in other sediments (Blackburn et aL, 1996; Landen-Hillmeyr, 1998), and were explained as being due to diffusional loss of low C/N ratio DOM produced during the initial hydrolysis of fresh (i.e., low C/N ratio) detrital organic matter near the sediment surface. This explanation is consistent with our Chesapeake Bay observations discussed above and discussions in Burdige and Gardner (1998), regarding the spatial separation in sediments between processes that produce the initial high-molecular-weight intermediates of sediment POM remineralization, and processes responsible for the production of refractory DOM in sediment pore waters (i.e., pLMW DOM). This spatial separation of HMW DOM and pLMW DOM net remineraUzation rates can also be inferred from the model results in Fig. 4. Based on these results the
Sediment Pore Waters
647
uncoupling between DOM that escapes as a benthic flux and that which accumulates in pore waters can be explained if one assumes that the C/N ratio of humic-like, pLMW DOM is greater (i.e., more carbon-rich) than that of this reactive HMW DOM, which appears to be a reasonable assumption (see Burdige, 2001, for further details). Results in Fig. 4 also show that in spite of the fact that model-derived concentrations of HMW DOC were significantly lower than those of pLMW DOC, their pore water gradients near the sediment surface were similar in magnitude. This observation is further quantified in Table I where it can be seen that calculated HMW DOC benthic fluxes are --50-80% of the total benthic DOC flux. Thus, in both anoxic and mixed redox sediments, model results suggest that fluxes of both refractory and reactive DOM can be similar in magnitude. In another approach to this problem, Bauer et al. (1995) examined the stable carbon and radiocarbon content of pore water DOM from the upper ~5 cm of two contrasting eastern North Pacific sediments (in Santa Monica Basin and at a 4100 m water depth site on the continental rise at the base of the Monterey deep-sea fan). At both sites the 8^^C values of the pore-water DOC had values consistent with a predominant marine source (—21 to —22 %o) and the pore-water DOC was greatly enriched in ^"^C as compared to bottom-water DOC (approx. —150 to —250 %oin the pore waters versus approx. —500 %o in the water column). While pore-water DOC profiles at both site predict diffusive, benthic DOC fluxes that are near-equal to the sediment carbon oxidation rate, this sediment DOC source does not appear to have a significant impact on bottom-water DOC radiocarbon values. To explain this latter observation, Bauer et al. (1995) suggested that either ^"^C-enriched pore water DOC is not released from the sediments in significant quantities (i.e., they have overestimated the benthic DOC flux in their simple benthic flux calculation) or that this radiocarbon-enriched, sediment-derived DOC "does not persist in the water colunm" (i.e., it is sufficiently reactive that it is rapidly remineralized). Consistent with the first suggestion is the fact that benthic DOC fluxes at these sites that I estimate using reported sediment carbon oxidation rates and Eq. [5] are seven to eight times smaller than those calculated by Bauer et al. (1995). At the same time though, the model results discussed earlier in this section (and shown in Fig. 4) suggest that benthic DOC fluxes may contain both refractory and reactive DOM components, which are likely to have different radiocarbon ages (e.g., see Eglinton et a/., 1997, for a discussion of analogous studies of radiocarbon ages of biomarker compounds in sediment POC). This would imply that the radiocarbon values determined by Bauer et al. (1995) could result from refractory (radiocarbon-depleted) and reactive (radiocarbon-enriched) components. This reactive sediment-derived DOC would likely undergo remineralization in the bottom waters, thus further minimizing the importance of sediment pore waters as a source or refractory DOC to the deep ocean.
648
David J. Burdige
While the discussion here shows that benthic DOC fluxes may not be as important a source of refractory DOC to the water colunm as once suggested, recent results suggest another way that sediment pore-water DOC dynamics could play a role in explaining the ^"^C age of deep-water DOC. In recent studies, Guo and Santschi (2000) observed that simple desorption of colloidal (>1 kDa) organic matter from continental margin sediments yields DOC that is substantially older than the bulk sediment organic matter (^3000 years vs 700 years, respectively). Desorption of this material from sediments in continental margin benthic nepheloid layers, coupled with transport of this material to the deep ocean (e.g., Bauer and Druffel, 1998), could then play an important role in explaining the age of deepocean DOC. Assuming that this old, desorbed DOC is in some kind of reversible equilibrium with the sediment pore waters while in the sediments, the relatively refractory nature of the bulk pore water DOC pool (^pLMW DOC) would then aid in the aging process of this sorbed organic matter while it is resides in the sediments.
VI. THE ROLE OF PORE-WATER DOM IN SEDIMENT CARBON PRESERVATION As was discussed at the beginning of this chapter, two common models for sediment carbon preservation, the geopolymerization model (Nissenbaum et ai, 1972; Tissot and Welte, 1978; Krom and Westrich, 1981) and the mesopore protection model (Mayer, 1994a,b; Hedges et ai, 1999), suggest that DOM in sediment pore waters may play an important role in the preservation process. A third model for carbon preservation, the selective preservation of refractory biomacromolecules (Hatcher and Spiker, 1988; de Leeuw and Largeau, 1993) is generally thought to not involve DOM intermediates. In its simplest sense, the geopolymerization model involves condensation reactions in which low-molecular-weight dissolved organic compounds (such as amino acids or simple sugars; i.e., mLMW DOM compounds in the PWSR model) react to form higher-molecular-weight dissolved humic substances. The continued condensation of these dissolved humics is thought to eventually lead to the formation of particulate material such as humin or kerogen, and to the preservation of this organic matter in sediments. Although there have been numerous studies of these condensation reactions in the lab, there is little direct evidence for their occurrence in nature (Hedges, 1988; Henrichs, 1992). Furthermore, based on our molecularweight data (Burdige and Gardner, 1998), we noted that if these reactions do occur on early diagenetic time scales their products still have relatively low molecular weights (i.e., less than 3 kDa). In the mesopore protection model, DOM sorption is proposed to occur in small mesopores on mineral surfaces, where the sorbed DOM is physically protected
Sediment Pore Waters
649
from attack by microbial enzymes (Mayer, 1994a,b, 1999; Hedges and Keil, 1995). The mesopore protection model, however, is not mutually exclusive of the occurrence of geopolymerization reactions, since rates of abiotic condensation reactions may be accelerated in mesopore sites, by either steric- or concentration-related phenomena, thus further enhancing the preservation of sorbed DOM (Mayer, 1994b; Collins et al, 1995; Hedges and Keil, 1995). This latter point is of some interest, since under most natural conditions aqueous phase (i.e., pore water) geopolymerization reactions are thought to be quite slow (Hedges, 1988; Alperin et al, 1994). Since mesopore size places some constraint on the upper limit of the size/molecular weight of DOM molecules that can be taken up in mesopores (Mayer, 1994b), the relatively small size (<3 kDa) of most pore-water DOM would appear to aid in its possible interaction with these adsorption sites. Similarly, the refractory nature of the bulk pore-water DOM pool may also enhance its preservation, if these sorption processes are reversible to any significant extent (see discussions in (Lee, 1994; Mayer, 1994b; Henrichs, 1995). At the same time, however, if this process does play a role in controlling carbon preservation, then the nature of the sorption process and the composition of the DOM molecules involved in sorption (and ultimately carbon preservation) are likely more complicated than that which can be described by bulk DOC concentrations and simple sorption processes and coefficients (Henrichs, 1995; Mayer, 1995; Alperin et al, 1999). Finally, these observations can be used to briefly examine questions related to the relationship between anoxia and sediment carbon preservation (compare, e.g., discussions in Pedersen and Calvert, 1990; Canfield, 1994, and references therein; also see Aller, 1994; Hedges and Keil, 1995; Hedges etal, 1999). In general, field observations, experimental studies, and model calculations (Burdige, 2001) show that pore-water DOC concentrations are generally higher under anoxic conditions (also see sections II.A and II.B). Burdige (2001) also shows that changes in sediment redox conditions alter the pathways of sediment DOM remineralization, leading to the enhanced relative accumulation of refractory DOC (i.e., pLMW DOC) under anoxic conditions. Such elevated DOC concentrations in anoxic sediments could therefore enhance sediment carbon preservation in both the "conventional" geopolymerization model and the surface area adsorption/mesopore protection model (the caveats/concerns at the end of the last paragraph notwithstanding). Furthermore, given the importance of DOC as an intermediate in sediment POM remineralization these observations may help explain how oxygen exposure time apparently enhances sediment POM remineralization (Hedges and Keil, 1995; Hartnett et al, 1998; Hedges et al, 1999). Overall, then, these arguments suggest that some amount of enhanced carbon preservation may indeed occur in anoxic sediments, in spite of the fact that anoxia and sediment carbon preservation may not necessarily be causally linked.
650
David J. Burdige
VII. CONCLUSIONS AND SUGGESTIONS FOR FUTURE RESEARCH Throughout this chapter I have attempted to summarize and synthesize the existing data on DOM in marine sediment pore waters and present explanations that (at least to me) appear to best explain the observations. I have also attempted to avoid repeatedly reminding the reader that "more work is clearly needed to examine these problems " although it should be apparent that this is indeed the case in most of these instances. The fact that much of the DOM in sediment pore waters appears to be refractory and is not well characterized to date (at least in terms of known biochemicals) suggests that "structural" approaches, similar to those used in water-colunm DOM studies (e.g., Aluwihare et al, 1997; McCarthy et ai, 1997; Boon et al, 1998), may prove useful here as well. Further studies linking DOM composition and reactivity will also be required to examine many of the problems discussed here. It is interesting to note that many of these questions could be addressed in carefully conducted sediment incubation experiments (S0rensen et ai, 1981; Burdige, 1991a,b). The work discussed here has focused on dissolved organic carbon and nitrogen, both in terms of bulk compositions and depth distributions and in specific compositional studies of the DOM pool. To date, little work has focused on dissolved organic phosphorus or sulfur in sediment pore waters, although it seems highly likely that such studies could be important in understanding the sedimentary and oceanic cycles of these elements as well. Such studies might also be of relevance in better understanding sediment carbon preservation depending on how (or if) natural vulcanization reactions affect carbon preservation (Tegelaar et al., 1989; de Leeuw and Largeau, 1993). As discussed throughout this chapter, DOM compounds are important intermediates in sediment carbon remineralization and perhaps sediment carbon preservation as well. Therefore, differences in DOM composition and reactivity due to differences in sediment redox conditions might play a role in how sediment redox conditions affect overall sediment carbon preservation in these different sediment types. Again, however, future studies will be needed to examine how (or even if) this occurs.
ACKNOWLEDGMENTS Preparation of this chapter was supported by grants from the Office of Naval Research (Harbor Processes Program and Coastal Benthic Optical Properties Program). My work on pore-water DOM has also been supported by grants from the National Science Foundation. I thank Will Berelson,
651
Sediment Pore Waters
Yo Chin, Marc Alperin, Juli Homstead, Mark Hines, Jack Middelburg, Annelie Skoog, Bob Haddad, and Paul Kepkay for discussions over the years that helped me clarify many of the thoughts and ideas that now appear here. Special thanks go to Chris Martens for first getting me involved in studies of porewater DOM sometime in the past century. I also thank Rick Keil and an anonymous reviewer for their comments on an earlier version of this chapter. Finally, musical inspiration while preparing this chapter came from two great FM stations that now broadcast over the Internet, KPIG (Freedom, CA) and WFUV (Bronx, NY).
APPENDIX: A DESCRIPTION OF THE DOM ADVECTION/DIFFUSION/REACTION MODEL ANOXIC, NONBIOTURBATED SEDIMENTS ( A N S MODEL) Equations [1] and [2] are solved using the boundary conditions H = H° and P = P° at X = 0 and dH/dx = 0 and dP/dx = 0 as x -^ cx), yielding H = We-^"" + G2(l - e-^"") + Qxie-^"" - e'^"")
[A-1]
where Qi =
—
T
k}{ — Xo) — DgA^
Qi = Roo/ku
[A-3] [A-4]
2A
kp — Q)X — A ^
^ =
Qs = aRoo/kp
[A-8]
-co + y/(o^ + 4Dskp m
f^-'^
652
David J. Burdige
BioTURBATED
AND/OR BIOIRRIGATED SEDIMENTS
(BBS
MODEL)
The equations developed here are general equations for sediment systems in which bioturbation and bioirrigation can both occur. In these equations it is assumed that there is a surficial bioturbated zone in which bioturbation is modeled as a random, diffusion-like process with a bioturbation coefficient D^ (Boudreau, 1997). This then leads to Dj = Ds + DB in the "mixed" layer and Dj = D^ below the mixed layer. The bioturbation coeffiecient DB can either be a constant value (leading to Dj being a step function across the mixed layer boundary) or any other function of depth. Bioirrigation is modeled as a nonlocal advective process of the form a(C - C")
[A-10]
(e.g., Boudreau, 1997), where a is the bioirrigation exchange coefficient and C° is the bottom-water concentration. Again, a can be either a constant in the "irrigated" zone or a function of depth. Given these formulations for these biological transport processes, Eqs. [1] and [2] can be rewritten as. — ( • T - ^ I - co-^ - a(H - H') -h ARe-^^ + Roc - k^H = 0 dx \ dx I dx
[A-11]
or 72
dx^
dx dx
dx
[A-12]
and dP d_ / „ dP\ DT— - CO—- - a(P - P°) + akuH - Ux)P dx J dx dx \
=0
[A-13]
or Dj^
dx"^
+^ ^ - c o ^ - a { P - P ' ) dx dx dx
+ aknH - k^(x)P = 0,
[A-14]
where kp is now written as kp(x) to indicate that its value is likely higher in the upper portions of the sediments that are bioturbated and/or bioirrigated (Burdige, 2001). The boundary conditions for these equations are the same as those for the equations described above for the ANS model. To keep the depth-dependence of DT (i.e., DB), a, and k^ as general as possible, Eqs. [A-12] and [A-14] have been solved by numerical techniques. This involves using the Crank-Nicolson implicit centered-differencing scheme to
653
Sediment Pore Waters
approximate the first and second derivatives of P and H (Crank, 1975; Dhakar and Burdige, 1996; Boudreau, 1997). Here the sediment column is divided into n depth intervals of equal size AJC and after numerically approximating the concentration derivatives, one is left with n equation of the form -at C,_i + bi d + Ci G+i = di
[A-15]
for both P and H. In these equations C/ is the concentration of either P ox H at depth interval /* Ax, and at, bf, c,, and df are based on the values of the other parameters in Eqs. [A-12] and [A-14]. Depending on the depth-dependence of DB, the first derivative of Dj can be either determined analytically at each depth interval, or numerically approximated as described above. In solving these equations the first and last equation in each set of equations (i.e., for either P or H) is rewritten based on the boundary equations (see Dhakar and Burdige, 1996, for further details). Each set of equations can then be expressed in the form of a tridiagonal matrix and solved by Gaussian elimination (Boudreau, 1997). Although the equations for P and H are coupled, the coupling is such that the set of equations for H can first be independently solved for H versus depth. This solution is then used to solve for P versus depth. This systems of equations can be solved in an Excel spread sheet (a copy of which is available from the author). The validity of this numerical scheme was verified several ways. In the first case, bioturbation and bioirrigation were "turned off' and numerical solutions were compared with analytical solutions to the equations developed above for anoxic, nonbioturbated sediments. In the second case, analytical solutions to a simple two-layer bioturbation model (AUer, 1982) were compared with numerical solutions to the same equations. In all of these comparisons, exact agreement between the two solutions was observed. Finally, it was observed that varying the value of Ax by a factor of two did not affect the solution of model calculations that included a range of bioturbation and bioirrigation coefficients. This insensitivity of numerical solutions to the grid size is a further indication of a stable numerical solution (see discussions in Alperin, 1989).
REFERENCES Albert, D. B., and Martens, C. S. (1997). Determination of low-molecular weight organic acid concentrations in seawater and pore-water samples via HPLC. Mar. Chem. 56,27-37. AUer, R. C. (1982). The effects of macrobenthos on chemical properties of marine sediments and overlying water. In "Animal-Sediment Relations: The Biogenic Alteration of Sediments" (R L. McCall and M. J. S. Tevesz, Eds.), pp. 53-102. Plenum Press, New York. AUer, R. C. (1994). Bioturbation and remineralization of sedimentary organic matter: Effects of redox osciUation. Chem. Geol 114, 331-345.
654
David J. Burdige
Alperin, M. J. (1989). "The Carbon Cycle in an Anoxic Marine Sediment: Concentrations, Rates, Isotope Ratios and Diagenetic Models." Ph.D. dissertation, Univ. of Alaska. Alperin, M. J., Albert, D. B., and Martens, C. S. (1994). Seasonal variations in production and consumption rates of dissolved organic carbon in an organic-rich coastal sediment. Geochim. Cosmochim. Acta 58,4909^929. Alperin, M. J., Martens, C. S., Albert, D. B., Suayah, I. B., Benninger, L. K., Blair, N. E., and Jahnke, R. A. (1999). Benthic fluxes and porewater concentration profiles of dissolved organic carbon in sediments from the North Carolina continental slope. Geochim. Cosmochim. Acta 63, 427448. Alperin, M. J., Reeburgh, W. S., and Devol, A. H. (1992). Organic carbon remineralization and preservation in sediments of Skan Bay, Alaska. In "Productivity, Accumulation, and Preservation of Organic Matter in Recent and Ancient Sediments" (J. K. Whelan and J. W. Farrington, Eds.), pp. 99-122. Columbia Univ. Press, New York. Aluwihare, L. I., Repeta, D. J., and Chen, R. F. (1997). A major biopolymeric component to dissolved organic carbon in surface seawater. Nature 387,166-169. Amon, R. M. W., and Benner, R. (1994). Rapid cycling of high-molecular-weight dissolved organic matter in the ocean. Nature 369,549-552. Amon, R. M. W., and Benner, R. (1996). Bacterial utihzation of different size classes of dissolved organic matter. Limnol. Oceanogr. 41,41-51. Ansbaek, J., and Blackburn, T. H. (1980). A method for the analysis of acetate turnover in a coastal marine sediment. Microb. Ecol. 5, 253-264. Antia, N. J., Harrison, P. J., and Oliveira, L. (1991). The role of dissolved organic nitrogen in phytoplankton nutrition, cell biology and ecology. Phycology 30,1-89. Amosti, C. (2000). Substrate specificity in polysaccharide hydrolysis: Contrasts between bottom water and sediments. Limnol. Oceanogr 45,1112-119. Amosti, C , and Holmer, M. (1999). Carbohydrate dynamics and contributions to the carbon budget of an organic-rich coastal sediments. Geochim. Cosmochim. Acta 63, 353^03. Amosti, C , Repeta, D. J., and Blough, N. V. (1994). Bacterial degradation of polysaccharides in anoxic marine sediments. Geochim. Cosmochim. Acta 58,2639-2652. Awapara, J. (1962). Free amino acids in invertebrates: A comparative study of their distribution and metabolism. In "Amino Acid Pools" (T. Holden, Ed.), pp. 158-175. Elsevier, Amsterdam. Barcelona, M. J. (1980). Dissolved organic carbon and volatile fatty acids in marine sediment pore waters. Geochim. Cosmochim. Acta 44,1977-1984. Bauer, J. E., and Dmffel, E. R. M. (1998). Ocean margins as a significant source of organic matter to the deep open ocean. Nature 392,482^85. Bauer, J. E., Reimers, C. E., Dmffel, E. R. M., and WiUiams, P M. (1995). Isotopic constraints on carbon exchange between deep ocean sediments and sea water. Nature 373, 686-689. Bauer, J. E., Wilhams, P. M., and Dmffel, E. R. M. (1992). ^"^C activity of dissolved organic carbon fractions in the north-central Pacific and Sargasso Sea. Nature 357, 667-690. Bender, M., Jahnke, R., Weiss, R., Martin, W., Heggie, D. T., Orchardo, J., and Sowers, T. (1989). Organic carbon oxidation and benthic nitrogen and silica dynamics in San Clemente Basin, a continental borderland site. Geochim. Cosmochim. Acta 53, 685-697. Benner, R. H. (2002). Chemical composition and reactivity. In "Biogeochemistry of Marine Dissolved Organic Matter" (D. A. Hansell and C. A. Carlson, Eds.), pp. 59-90. Academic Press, San Diego. Benner, R., Pakulski, J. D., McCarthy, M., Hedges, J., and Hatcher, R G. (1992). Bulk chemical characteristics of dissolved organic matter in the ocean. Science 255,1561-1564. Berelson, W. M., McManus, J., Kilgore, T., Coale, K., Johnson, K. S., Burdige, D., and Pilskaln, C. (1996). Biogenic matter diagenesis on the sea floor: A comparison between two continental margin transects. J. Mar Res. 54, 731-762.
Sediment Pore Waters
655
Bemer, R. A. (1980). "Early Diagenesis: A Theoretical Approach." Princeton University Press, Princeton, NJ. Blackburn, T. H., Hall, R O. J., Hulth, S., and Landen, A. (1996). Organic-N loss by efflux and burial associated with a low efflux of inorganic N and with nitrate assimilation in Arctic sediments (Svalbard, Norway). Mar. Ecol Prog. Sen 141, 283-293. Boon, J. J., Klap, V., and Eglinton, T. (1998). Molecular characterization of microgram amounts of colloidal dissolved organic matter (UDOM) in ocean water samples by direct temperature resolved anmionia chemical ionization mass spectrometry. Org. Geochem. 29,1051-1062. Boschker, H. T. S., Bertilsson, S. A., Dekkers, E. M. J., and Cappenberg, T. E. (1995). An inhibitorbased method to measure initial decomposition of naturally occurring polysaccharides in sediments. Appl. Environ. Microbiol. 61,2186-2192. Boudreau, B. P. (1997). "Diagenetic Models and Their Implementation." Springer-Verlag, Amsterdam. Boynton, W. R., and Kemp, W. M. (1985). Nutrient regeneration and oxygen consumption by sediments along an estuarine salinity gradient. Mar. Ecol. Prog. Sen 23,45-55. Boynton, W. R., Kemp, W. M., and Osborne, C. G. (1980). Nutrient fluxes across the sediment-water interface in the turbid zone of a coastal plain estuary. In "Estuarine Perspectives" (V. S. Kennedy, Ed.), pp. 93-109. Academic Press, San Diego. Bronk, D. A. (2002). Dynamics of DON. In "Biogeochemistry of Marine Dissolved Organic Matter" (D. A. Hansen and C. A. Carlson, Eds.), pp. 153-247. Academic Press, San Diego. Burdige, D. J. (1989). The effects of sediment slurrying on microbial processes, and the role of amino acids as substrates for sulfate reduction in anoxic marine sediments. Biogeochemistry 8,1-23. Burdige, D. J. (1991a). The kinetics of organic matter mineralization in anoxic marine sediments. J. Mar Res. 49,121-161. Burdige, D. J. (1991b). Microbial processes affecting alanine and glutamic acid in anoxic marine sediments. FEMS Microbiol. Ecology 85, 211-231. Burdige, D. J. (2001). Dissolved organic matter in Chesapeake Bay sediment pore waters. Org. Geochem. 32,487-505. Burdige, D. J., Alperin, M. J., Homstead, J., and Martens, C. S. (1992). The role of benthic fluxes of dissolved organic carbon in oceanic and sedimentary carbon cycling. Geophys. Res. Lett. 19, 1851-1854. Burdige, D. J., Berelson, W. M., Coale, K. H., McManus, J., and Johnson, K. S. (1999). Fluxes of dissolved organic carbon from California continental margin sediments. Geochim. Cosmochim. Ac^fl 63,1507-1515. Burdige, D. J., Chen, W., and Kline, S. W. (2002). The cycling of fluorescent dissolved organic matter in contrasting marine sediments, ms in prep, for Mar Chem. Burdige, D. J., and Gardner, K. G. (1998). Molecular weight distribution of dissolved organic carbon in marine sediment pore waters. Mar Chem. 62,45-64. Burdige, D. J., Gardner, K. G., and Skoog, A. (2000). Dissolved and particulate carbohydrates in contrasting marine sediments. Geochim. Cosmochim. Acta 64,1029-1041. Burdige, D. J., Gardner, K. G., and Zheng, S. (1996). The molecular weight distribution of dissolved organic matter (DOM) in marine sediment pore waters. EOS 76 (3), OSl IK (abstr.). Burdige, D. J., and Homstead, J. (1994). Fluxes of dissolved organic carbon from Chesapeake Bay sediments. Geochim. Cosmochim. Acta 58, 3407-3424. Burdige, D. J., and Martens, C. S. (1988). Biogeochemical cycling in an organic-rich marine basin 10. The role of amino acids in sedimentary carbon and nitrogen cycling. Geochim. Cosmochim. Acta 52,1571-1584. Burdige, D. J., and Martens, C. S. (1990). Biogeochemical cycling in an organic-rich marine basin. 11. The sedimentary cycling of dissolved free amino acids. Geochim. Cosmochim. Acta 54, 30333052.
656
David J. Burdige
Burdige, D. J., and Zheng, S. (1998). The biogeochemical cycHng of dissolved organic nitrogen in estuarine sediments. Limnol. Oceanogr. 43,1796-1813. Cai, W.-J., Reimers, C. E., and Shaw, T. J. (1995). Microelectrode studies of organic carbon degradation and calcite dissolution at a California continental rise site. Geochim. Cosmochim. Acta 59,497-511. Canfield, D. E. (1994). Factors influencing organic matter preservation in marine sediments. Chem. Geo/. 114,315-329. Carlson, C. A. (2002). Production and removal processes. In "Biogeochemistry of Marine Dissolved Organic Matter" (D. A. Hansell and C. A. Carlson, Eds.), pp. 91-151. Academic Press, San Diego. Capone, D. G., and Kiene, R. P. (1988). Comparison of microbial dynamics in marine and freshwater sediments: Contrasts in anaerobic carbon catabohsms. Limnol Oceanogr. 33,725-749. Carpenter, E. J., and Capone, D. G. (1983). "Nitrogen in the Marine Environment." Academic Press. Caughey, M. E. (1982). "A Study of the Dissolved Organic Matter in the Pore Waters of Carbonate-Rich Sediment Cores from Florida Bay." MS thesis, Univ. of Texas at Dallas. Cauwet, G. (2002). DOM in the coastal zone. In "Biogeochemistry of Marine Dissolved Organic Matter" (D.A. Hansell and C.A. Carlson, Eds.), pp. 579-609. Academic Press, San Diego. Chen, R. R, Bada, J. L., and Suzuki, Y. (1993). The relationship between dissolved organic carbon (DOC) and fluorescence in anoxic marine porewaters: Implications for estimating benthic DOC fluxes. Geochim. Cosmochim. Acta 57,2149-2153. Chin, Y.-P, and Gschwend, P. M. (1991). The abundance, distribution, and configuration of porewater organic colloids in recent sediments. Geochim. Cosmochim. Acta 55,1309-1317. Christensen, D. (1984). Determination of substrates oxidized by sulfate reduction in intact cores of marine sediments. Limnol. Oceanogr. 29, 189-192. Christensen, D., and Blackburn, T. H. (1980). Turnover of tracer (^"^C, ^H labelled) alanine in inshore marine sediments. Mar. Biol. 58,97-103. Christensen, D., and Blackburn, T. H. (1982). Turnover of ^"^C-labelled acetate in marine sediments.
Mar Biol.
l\,\\3-n9.
Codispoti, L. A. (1995). Is the ocean losing nitrate? Nature 376, 724. Collins, M. J., Bishop, A. N., and Farrimond, P. (1995). Sorption by mineral surfaces: Rebirth of the classical condensation pathway for kerogen formation? Geochim. Cosmochim. Acta 59,23872391. Colombo, J. C , Silverberg, N., and Gearing, J. N. (1998). Amino acid biogeochemistry in the Laurentian Trough: vertical fluxes and individual reactivity during early diagenesis. Org. Geochem. 29,933-945. Cowan, J. L. W., and Boynton, W. R. (1996). Sediment-water oxygen and nutrient exchanges across the longitudinal axis of Chesapeake Bay: Seasonal patterns, controlling factors and ecological significance. Estuaries 19, 562-580. Cowie, G. L., and Hedges, J. I. (1992a). The role of anoxia in organic matter preservation in coastal sediments: Relative stabihties of the major biochemicals under oxic and anoxic depositional conditions. Org. Geochem. 19, 229-234. Cowie, G. L., and Hedges, J. I. (1992b). Sources and reactivities of amino acids in a coastal marine environment. Limnol. Oceanogr 37,703-724. Cowie, G. L., and Hedges, J. I. (1994). Biochemical indicators of diagenetic alteration in natural organic matter mixtures. Nature 369, 304-307. Cowie, G. L., Hedges, J. I., Prahl, F. G., and deLange, G. J. (1995). Elemental and biochemical changes across an oxidation front in a relict turbidite: An oxygen effect. Geochim. Cosmochim. Acta 59, 33^6. Crank, J. (1975). "The Mathematics of Diffusion," 2nd ed. Clarendon Press. Crill, P. M., and Martens, C. S. (1986). Methane production from bicarbonate and acetate in an anoxic marine sediment. Geochim. Cosmochim. Acta 50, 2089-2097.
Sediment Pore Waters
657
de Leeuw, J. W., and Largeau, C. (1993). A review of macromolecular organic compounds that comprise living organisms and their role in kerogen, coal, and petroleum formation. In "Organic Geochemistry" (M. H. Engel and S. A. Macko, Eds.), pp. 23-72. Plenum Press. Devol, A. H., and Ahmed, S. I. (1981). Are high rates of sulfate reduction associated with anaerobic oxidation of methane? Nature 291,407^08. Devol, A. H., and Christensen, J. P. (1993). Benthic fluxes and nitrogen cycling in sediments of the continental margin of the eastern North Pacific. J. Mar. Res. 51, 345-372. Dhakar, S. P., and Burdige, D. J. (1996). A coupled, non-linear, steady state model for early diagenetic processes in pelagic sediments. Am. J. Sci. 296, 296-330. Dollar, S. J., Smith, S. V., Vink, S. M., Obrebski, S., and HoUibaugh, J. T. (1991). Annual cycle of benthic nutrient fluxes in Tomales Bay, California, and contribution of the benthos to total carbon metabohsm. Mar. Ecol. Prog. Sen 79,115-125. Eganhouse, R. P., and Venkatesan, I. (1993). Chemical oceanography and geochemistry. In "Ecology of the Southern Cahfomia Bight: A Synthesis and Interpretation" (M. D. Dailey, D. J. Reish, and J. W. Anderson, Eds.), chap. 3, pp. 71-189. Univ. of California Press. Eglinton, T. I., Benitez-Nelson, B. C , Pearson, A., McNichol, A. R, Bauer, J. E., and Druffel, E. R. M. (1997). Variability in radiocarbon ages of individual organic compunds from marine sediments. Science 277,796-799. Elderfield, H. (1981). Metal-organic associations in interstitial waters of Narragansett Bay sediments. Am./.5d. 281,1184-1196. Emerson, S., and Dymond, J. (1984). Benthic organic carbon cycles: Toward a balance of fluxes from particles settling and pore-water gradients. In "Global Ocean Flux Study, Proceedings of a Workshop," pp. 285-304. National Academic Press, Washington, DC. Emerson, S., and Hedges, J. I. (1988). Processes controlling the organic carbon content of open ocean sediments. Paleocean. 3, 621-634. Enoksson, V. (1993). Nutrient recyling by coastal sediments: Effects of added algal material. Mar Ecol. Prog. Sen 92, 245-254. Fenchel, T., King, G. M., and Blackburn, T. H. (1998). "Bacterial Biogeochemistry: The Ecophysiology of Mineral Cycling." Academic Press, San Diego. Gardner, W S., and Hanson, R. B. (1979). Dissolved free amino acids in interstitial waters of Georgia salt marsh soils. Estuaries 2, 113-118. Guo, L., and Santschi, P. H. (1997). Composition and cycling of colloids in marine environments. Rev. Geophys. 3 5 , 1 7 ^ 0 . Guo, L., and Santschi, P. H. (2000). Sedimentary sources of old high molecular weight dissolved organic carbon from the ocean margin nepheloid layer. Geochim. Cosmochim. Acta 64, 651-600. Guo, L., Santschi, P. H., Cifuentes, L. A., Trumbore, S. E., and Southon, J. (1996). Cycling of high-molecular-weight dissolved organic matter in the Middle Atlantic Bight as revealed by carbon isotopic (^^C and ^^C) signatures. Limnol. Oceanogr. 41,1242-1252. Haberstroh, P. R., and Karl, D. M. (1989). Dissolved free amino acids in hydrothermal vent habitiats of the Guaymas Basin. Geochim. Cosmochim. Acta 53, 2937-2945. Hall, R O. J., Anderson, L. G., Holby, O., KoUberg, S., and Samuelsson, M.-O. (1990). Chemical fluxes and mass balance in a marine fish farm. I. Carbon. Mar Ecol. Prog. Ser 61, 61-73. Hammond, D, E., McManus, J., Berelson, W. M., Kilgore, T. E., and Pope, R. H. (1996). Early diagenesis of organic material in equatorial Pacific sediments: Stoichiometry and kinetics. Deep-Sea Res. 43,1365-1412. Hansen, D. A. (2002). DOC in the global ocean carbon cycle. In "Biogeochemistry of Marine Dissolved Organic Matter" (D. A. Hansell and C. A. Carlson, Eds.), pp. 685-715. Academic Press, San Diego. Hartnett, H. E., Keil, R. G., Hedges, J. I., and Devol, A. H. (1998). Influence of oxygen exposure time on organic carbon preservation in continental margin sediments. Nature 391,572-574.
658
David J. Burdige
Hartwig, E. O. (1976). The impact of nitrogen and phosphorus release from a siliceous sediment on the overlying water. In "Estuarine Processes" (M. Wiley, Ed.), pp. 103-117. Academic Press, New York. Harvey, H. R. (1994). Fatty acids and sterols as source markers of organic matter in sediments of the North Carolina continental slope. Deep-Sea Res. 41,783-796. Hatcher, P. G., and Spiker, E. C. (1988). Selective degradation of plant biomolecules. In "Humic Substances and Their Role in the Environment" (F. C. Frimmel and R. C. Christman, Eds.), pp. 59-74. J. Wiley, New York. Hedges, J. I. (1992). Global biogeochemical cycles: Progress and problems. Mar. Chem. 39,67-93. Hedges, J. I. (1988). Polymerization of humic substances in natural environments. In "Humic Substances and Their Role in the Environment" (F. C. Frimmel and R. C. Christman, Eds.), pp. 45-58. Wiley, New York. Hedges, J. I. (2002). Why dissolved organics matter. In "Biogeochemistry of Marine Dissolved Organic Matter" (D. A. Hansell and C. A. Carlson, Eds.), pp. 1-33. Academic Press, San Diego. Hedges, J. I., Hu, F S., Devol, A. H., Hartnett, H. E., Tsamakis, E., and Keil, R. G. (1999). Sedimentary organic matter preservation: A test for selective degradation under oxic conditions. Am. J. Sci. 299, 529-555. Hedges, J. I., and Keil, R. G. (1995). Sedimentary organic matter preservation: An assessment and speculative synthesis. Mar. Chem. 49, 81-115. Heggie, D., Maris, C , Hudson, A., Dymond, J., Beach, R., and CuUen, J. (1987). Organic carbon oxidation and preservation in NW Atlantic continental margin sediments. In "Geology and Geochemistry of Abyssal Plains" (P. P. E. Weaver and J. Thomson, Eds.), pp. 215-236. Blackwell Sci., Oxford. Henrichs, S. M, (1980). "Biogeochemistry of Dissolved Free Amino Acids in Marine Sediments." Ph.D. thesis, MITAVHOI. Henrichs, S. M. (1992). Early diagenesis of organic matter in marine sediments: Progress and perplexity. Mar Chem. 3% 119-149. Henrichs, S. M. (1993). Early diagenesis of organic matter: the dynamics (rates) of cycHng of organic compounds. In "Organic Geochemistry" (M. Engel and S. Macko, Eds.), pp. 101-117. Plenum Press. Henrichs, S. M. (1995). Sedimentary organic matter preservation: An assessment and speculative synthesis—A conunent. Mar. Chem. 49, 127-136. Henrichs, S. M., and Cuhel, R. (1985). Occurrence of 6-aminoglutaric acid in marine bacteria. Appl. Environ. Microbiol. 50, 543-545. Henrichs, S. M., and Doyle, A. P. (1986). Decomposition of ^^C-labelled organic substrates in marine sediments. Limnol. Oceanogr. 31, 765-778. Henrichs, S. M., and Farrington, J. W (1979). Amino acids in interstitial waters of marine sediments. Nature 21% 319-322. Henrichs, S. M., and Farrington, J. W. (1980). Amino acids in interstitial waters of marine sediments: A comparison of results from varied sedimentary environments. In "Advances in Organic Geochemistry" (A. G. Douglass and J. R. Maxwell, Eds.), pp. 435^143. Pergamon, New York. Henrichs, S. M., and Farrington, J. W (1987). Early diagenesis of amino acids and organic matter in two coastal marine sediments. Geochim. Cosmochim. Acta 51,1-15. Henrichs, S. M., Farrington, J. W., and Lee, C. (1984). Peru upweUing region sediments near 15°S. 2. Dissolved free and total hydrolyzable amino acids. Limnol. Oceanogr 29, 20-34. Henrichs, S. M., and Sugai, S. F (1993). Adsorption of amino acids and glucose by sediments of Ressurrection Bay, Alaska, USA: Functional group effects. Geochim. Cosmochim. Acta SI, 823-835.
Sediment Pore Waters
659
Hines, M. E., Banta, G. T., Giblin, A. E., Hobbie, J. E., and Tu, J. B. (1994). Acetate concentrations and oxidation in salt marsh sediments. Limnol. Oceanogr. 39,140-148. Holcombe, B. L., Keil, R., and Devol, A. H. (2001). Determination of pore-water dissolved organic carbon fluxes from Mexican margin sediments. Limnol. Oceanogr. 46,298-308. Holmer, M., and Nielsen, S. L. (1997). Sediment sulfer dynamics related to biomass-density patterns in Zostera marina (ellgrass) beds. Mar Ecol. Prog. Ser 146, 163-171. Hopkinson, C. S., Jr. (1987). Nutrient regeneration in shallow-water sediments of the estuarine plume region of the nearshore Georgia Bight, USA. Mar Biol. 94,127-142. Howes, B. L., Dacey, J. W. H., and Wakeham, S. G. (1985). Effects of sampling techniques on measurements of pore-water constituents in salt marsh sediments. Limnol. Oceanogr 30, 221227. Hulth, S., Tengberg, A., Landen, A., and Hall, P. O. J. (1997). Minerahzation and burial of organic carbon in sediments of the southern Wedell Sea (Antarctica). Deep-Sea Res. 44, 955-981. Jackson, G. A., and Williams, P. M. (1985). Importance of dissolved organic nitrogen and phosphorus in biological nutrient cycling. Deep-Sea Res. 32, 223-235. Jahnke, R. A. (1990). Early diagenesis and recycling of biogenic debris at the seafloor, Santa Monica Basin, California. /. Mar Res. 48,413-436. J0rgensen, N. O. G., Lindroth, P., and Mopper, K. H. (1981). Extraction and distribution of free amino acids and ammonium in sediment interstitial waters from the Limfjord, Denmark. Oceanol. Acta 4,465-474. Kemp, P F. (1990). The fate of benthic bacterial production. Rev. Aq. Sci. 2, 109-124. Kemp, W. M., Sampou, P., Mayer, M., Henricksen, K., andBoyton, W. R. (1990). Ammonium recycling versus denitrification in Chesapeake Bay sediments. Limnol. Oceanogr 35,1545-1563. King, G. M. (1991). Measurement of acetate concentrations in marine pore waters by using an enzymatic approach. App/. Environ. Microbiol. 57, 3476-3481. King, G. M., Klug, M. J., and Lovely, D. R. (1983). Metabolism of acetate, methanol, and methylated amines in intertidal sediments of Lowes Cove, Maine. Appl. Environ. Microbiol. 45, 18481853. Koepfler, E. T., Benner, R., and Montagna, P. A. (1993). Variability of dissolved organic carbon in sediments of a seagrass bed and and unvegetated area within an estuary in southern Texas. Estuaries 16,391-404. Kristensen, E., and Blackburn, T. H. (1987). The fate of organic carbon and nitrogen in experimental marine systems: Influence of bioturbation and anoxia. J. Mar Res. 45, 231-257. Krom, M. D., and Sholkovitz, E. R. (1977). Nature and reactions of dissolved organic matter in the interstitial waters of marine sediments. Geochim. Cosmochim. Acta 41,1565-1573. Krom, M. D., and Westrich, J. T. (1981). Dissolved organic matter in the pore waters of recent marine sediments: A review. In "Biogeochemie de la matiere organique a I'interface eau-sediment marin," pp. 103-111. CoUoques Intemationaux du C.N.R.S. Laanbroek, H. J., and Veldkamp, H. (1982). Microbial interactions in sediment communities. Phil. Trans. R. Soc. London B 297, 533-550. Lahajnar, N., Rixen, T, Schaefer, P., and Ittekkot, V. (2000). Porewater dissolved organic carbon (DOC) fluxes to the deep-sea:Arabian Sea vs. northeast Atltantic Ocean. EOS 81 (48), OS 1 IB-17. [Abstract]. Landen, A., and Hall, P. O. J. (1998). Seasonal variation of dissolved and adsorbed amino acids and anmionium in a near-shore marine sediment. Mar Ecol. Prog. Ser 170,67-84. Landen, A., and Hall, P. O. J. (2000). Benthic fluxes and pore water distributions of dissolved free amino acids in the open Skagerrak. Mar Chem. 71,53-68. Landen-Hillmeyr, A. (1998). "Nitrogen Cycling in Continental Margin Sediments With Emphasis on Dissolved Organic Nitrogen and Amino Acids." Ph.D. dissertation, Goteborg University.
660
David J. Burdige
Lee, C. (1992). Controls on organic carbon preservation: The use of stratified water bodies to compare intrinsic rates of decompositin in oxic and anoxic systems. Geochim. Cosmochim. Acta 56, 3233-3335. Lee, C. (1994). Kitty litter for carbon control. Nature 370,503-504. Lomstein, B. Aa., Blackburn, T. H., and Henricksen, K. (1989). Aspects of nitrogen and carbon cycling inthe northern Bering Shelf sediments. I. The significance of urea turnover in the mineraUzation of NH4+. Mar. Ecol Prog. Ser SI, 237-247. Lomstein, B. Aa., Jensen, A.-G. U., Hansen, J. W., Andreasen, J. B., Hansen, L. S., Bemtsen, J., and Kuzendorf, H. (1998). Budgets of sediment nitrogen and carbon cycling in the shallow water of Knebel Vig, Denmark. Aquat. Microb. Ecol 14,69-80. Lyons, W. B., Gaudette, H. E., and Hewitt, A. D. (1979). Dissolved organic matter in pore waters of carbonate sediments from Bermuda. Geochim. Cosmochim. Acta 43,433^37. Martens, C. S., Haddad, R. I., and Chanton, J. P. (1992). Organic matter accumulation, remineralization and burial in an anoxic marine sediment. In "Productivity, Accumulation, and Preservation of Organic Matter in Recent and Ancient Sediments" (J. K. Whelan and J. W. Farrington, Eds.), pp. 82-98. Columbia Univ. Press, New York. Martin, W. R., and McCorkle, D. C. (1993). Dissolved organic carbon concentrations in marine pore waters determined by high-temperature oxidation. Limnol. Oceanogr. 38,1464-1479. Mayer, L. M. (1989). Extracellular proteolytic enzyme activity in sediments of an intertidal mudflat. Limnol Oceanogr. 34, 973-981. Mayer, L. M. (1994a). Relationships between mineral surfaces and organic carbon concentrations in soils and sediments. Chem. Geol 114, 347-363. Mayer, L. M. (1994b). Surface area control of organic carbon accumulation in continental margin sediments. Geochim. Cosmochim. Acta 58, 1271-1284. Mayer, L. M. (1995). Sedimentary organic matter preservation: An assessment and speculative synthesis—A comment. Mar. Chem. 49, 123-126. Mayer, L. M. (1999). Extent of coverage of mineral surfaces by organic matter in marine sediments. Geochim. Cosmochim. Acta 63, 207-215. Mayer, L. M., and Rice, D. L. (1992). Early diagenesis of proteins: A seasonal study. Limnol Oceanogr. 37, 280-295. McCarthy, M., Pratum, T., Hedges, J., and Benner, R. (1997). Chemical composition of dissolved organic nitrogen in the ocean. Nature 390, 150-154. McManus, J., Berelson, W. M., Coale, K., Johnson, K., and Kilgore, T. E. (1997). Phosphorus regeneration in continental margin sediments. Geochim. Cosmochim. Acta 61, 2891-2907. Meybeck, M. (1982). Carbon, nitrogen and phosphorus transport by worlds rivers. Am. J. Scl 282, 401-450. Meyers, P. A. (1997). Organic geochemical proxies of paleoceanographic, paleoUmnologic, and paleoclimatic processes. Org. Geochem. 11, 213-250. Michaehs, W., Mycke, B., Vogt, J., Schuetze, G., and Degens, E. T. (1982). Organic geochemistry of interstitial waters. In "Initial Reports of DSDP 64" (J. Blakeslee et. al, Eds.), pp. 933-937. US Govt. Printing Office, Washington, DC. Michelson, A. R., Jacobsen, M. E., Scranton, M. I., and Mackin, J. E. (1989). Modeling the distribution of acetate in anoxic estuarine sediments. Limnol Oceanogr 34, lAl-151. Middelburg, J. J. (1989). A simple rate model for organic matter decomposition in marine sediments. Geochim. Cosmochim. Acta 53, 1577-1581. Middelburg, J. J., Soetaert, K., and Herman, P. M. J. (1997). Empirical relationships for use in global diagenetic models. Deep-Sea Res. 44, 327-344. Mopper, K., Zhou, X., Keiber, R. J., Keiber, D. J., Sikorski, R., and Jones, R. D. (1991). Photochemical degradation of dissolved organic carbon and its impact on the oceanic carbon cycle. Nature 353, 60-62.
Sediment Pore Waters
661
Nie, Y., Suayah, I. B., Benninger, L. K., and Alperin, M. J. (2001). Modeling detailed sedimentary ^^^Pb and fallout ^39,240 p^ profile, to allow episodic events: An application in Chesapeake Bay. Limnol. Oceanogr. 46,1425-1437. Nissenbaum, A., Baedecker, M. J., and Kaplan, I. R. (1972). Studies on dissolved organic matter from interstitial waters of a reducing marine fjord. In "Advances in Organic Geochemistry 1971" (H. R. von Gaerter and H. Wehner, Eds.), pp. 427-440. Pergamon Press, New York. Nixon, S. W. (1981). Remineralization and nutrient cycling in coastal marine ecosystems. In "Estuaries and Nutrients" (B. J. Neilson and L. E. Cronin, Eds.), pp. 111-138. Humana Press, clifton. Nixon, S. W., Oviatt, C. A., and Hale, S. S. (1976). Nitrogen regeneration and the metaboHsm of coastal marine bottom conmiunities. In "The Role of Terrestrial and Aquatic Organisms in Decomposition Processes" (J. Anderson and A. MacFayden, Eds.), pp. 269-283. Blackwell, Sci. Oxford. Nixon, S. W., and Pilson, M. E. Q. (1983). Nitrogen in estuaries and coastal marine ecosystems. In "Nitrogen in the Marine Environment" (E. J. Carpenter and D. G. Capone, Eds.), pp. 565-648. Academic Press, New York. NoveUi, R C , Michelson, A. R., Scranton, M. I., Banta, G. T., Hobbie, J. E., and Howarth, R. W. (1988). Hydrogen and acetate cycling in two sulfate-reducing sediments: Buzzards Bay and Town Cove, Mass. Geochim. Cosmochim. Acta 52, 2477-2486. Orem W. H., and Gaudette H. E. (1984). Organic matter in anoxic marine pore waters: oxidation effects. Org. Geochem. 5,175-181. Otto, S., and Balzer, W. (1998). Release of dissolved organic carbon (DOC) from sediments of the N. W. European Continental Margin (Goban Spur) and its significance for benthic carbon cycling. Prog. Oceanogr. 42, 127-144. Pantoja, S., and Lee, C. (1999). Peptide decomposition by extracellular hydrolysis in coastal seawater and salt marsh sediment. Mar. Chem. 63,273-291. Parkes, R. J., Cragg, B. A., and Bale, S. J. (1994). Deep bacterial biosphere in Pacific Ocean sediments. Nature 371,410-413. Parkes, R. J., Gibson, G. R., Mueller-Harvey, I., Buckingham, W. J., and Herbert, R. A. (1989). Determination of the substrates for sulfate-reducing bacteria within marine and estuarine sediments with different rates of sulfate reduction. J. Gen. Microbiol. 135,175-187. Parkes, R. J., and Taylor, J. (1983). Analysis of volatile fatty acids by ion-exclusion chromatography, with special reference to marine pore water. Mar. Biol. 77,113-118. Patience, R. L., Baxby, M., Bartle, K. D., Perry, D. L., Reiss, G. W., and Rowland, S. J. (1992). The functionality of organic nitrogen in some recent sediments from the Peru upwelling region. Org. Geochem. 18, 161-169. Pedersen, T. P., and Calvert, S. E. (1990). Anoxia vs. productivity: What controls the formation of organic-carbon-rich sediments and sedimentary rocks? AAPG Bull. 74,454-466. Rosenfeld, J. K. (1979). Amino acid diagenesis and adsorption in nearshore anoxic marine sediments. Limnol. Oceanogr. 24,1014-1021. Rubinsztain, Y, loselis. P., Ikan, R., and Aizenshtat, Z. (1984). Investigations on the structural units of melanoidins. Org. Geochem. 6,791-804. Sansone, F. J., and Martens, C. S. (1982). Volatile fatty acid cycling in organic-rich marine sediments. Geochim. Cosmochim. Acta 45,101-121. Santschi, P. H., Guo, L., Baskaran, M., Trumbore, S., Southon, J., Bianchi, T. S., Honeyman, B., and Cifuentes, L. (1995). Isotopic evidence for the contemporary origin of high-molecular-weight organic matter in oceanic environments. Geochim. Cosmochim. Acta 59,625-631. Saunders, E. W., Roberts, C. E., Lidstrom, M. E., Pearsall, N. N., and Nester, M. T. (1983). "Microbiology," 3rd ed. Saunders College, Philadelphia. Shaw, D. G., Alperin, M. J., Reeburgh, W. S., and Mcintosh, D. J. (1984). Biogeochemistry of acetate in anoxic sediments of Skan Bay, Alaska. Geochim. Cosmochim. Acta 48, 1819-1825.
662
David J. Burdige
Shaw, D. G., and Mcintosh, D. J. (1990). Acetate in recent anoxic sediments: direct and indirect measuraients of concentration and turnover rates. Estuarine Coastal Shelf Sci. 31,775-788. Shaw, T. J., Gieskes, J. M., and Jahnke, R. A. (1990). Early diagenesis in differing depositional environments: The response of transition metals in pore waters. Geochim. Cosmochim. Acta 54, 1233-1246. Sholkovitz, E. R., and Mann, D. R. (1984). The pore water chemistry of ^^^'^"^^Pu and ^^^Cs in sediments of Buzzards Bay, Massachusetts. Geochim. Cosmochim. Acta 48, 1107-1114. Skoog, A., and Benner, R. (1998). Aldoses in various size fractions of marine organic matter: Implications ofr carbon cycling. Limnol. Oceanogr. 42, 1803-1813. Skoog, A., Hall, R O. J., Hukth, S., Paxeus, N., and Rutgers van der Loef, M. (1996). Early diagenetic production and sediment-water exchange of fluorescent dissolved organic matter in the coastal environment. Geochim. Cosmochim. Acta 60, 3619-3629. Skrabal, S. A., Donat, J. R., and Burdige, D. J. (1997). Fluxes of copper-complexing ligands from estuarine sediments. Limnol. Oceanogr. 42,992-996. Skrabal, S. A., Donat, J. R., and Burdige, D. J. (2000). pore water distributions of dissolved copper and copper-complexing ligands in esmarine and coastal marine sediments. Geochim. Cosmochim. Ac/fl 64,1843-1857. Smith, C. R., Pope, R. H., DeMaster, D. J., and Magaard, L. (1993). Age-dependent mixing in deep-sea sediments. Geochim. Cosmochim. Acta 57, 1473-1488. S0rensen, J., Christensen, D., and J0rgensen, B. B. (1981). Volatile fatty acids and hydrogen as substrates for sulfate-reducing bacteria in anaerobic marine sediments. Appl. Environ. Microbiol. 42, 5-11. Starikova, N. D. (1970). Vertical distribution patterns of dissolved organic carbon in sea water and interstitial solutions. Oceanology 10,796-807. Steinberg, S. M., Venkatesan, M. I., and Kaplan, I. R. (1987). Organic geochemistry of sediments from the coninental margin off southern New England, U.S.A. I. Amino acids, carbohydrates and Hgnin. Mar. Chem. 21, 249-265. Suess, E., MuUer, P. J., and Reimers, C. E. (1980). A closer look at nitrification in pelagic sediments. Geochem. J. 14, 129-137. Sugai, S. P., and Henrichs, S. M. (1992). Rates of amino acid uptake and remineralization in Resurrection Bay (Alaska) sediments. Mar. Ecol. Prog. Sen 88, 129-141. Teague, K. G., Madden, C. J., and Day, J. J., Jr. (1988). Sediment-water oxygen and nutrient fluxes in a river-dominated estuary. Estuaries 11, 1-9. Tegelaar, E. W., Leeuw, J. W. de, Derenne, S., and Largeau, C. (1989). A reappraisal of kerogen formation. Geochim. Cosmochim. Acta 53, 3103-3106. Thimsen, C. A., and Keil, R. G. (1998). Potential interactions between sedimentary dissolved organic matter and mineral surfaces. Mar Chem. 62,65-76. Thurman, E. M. (1985). "Organic Geochemistry of Natural Waters." Martinus Nijhoff/Junk, The Hagve. Tissot, B., and Welte, D. H. (1978). "Petroleum Occurrence and Formation." Springer-Verlag, Amsterdam. Walsh, T. W. (1989). Total dissolved nitrogen in seawater: A new high-temperature combustion method and a comparison with photo-oxidation. Mar Chem. 26, 295-311. Wefer, G., Berger, W H., Richter, €., etal. (1998). "Proceedings of the Ocean Drilling Program, Initial Reports," Vol. 175. College Station, TX. Wellsbury, P, Goodman, K., Barth, T, Cragg, B. A., Barnes, S. P, and Parkes, R. J. (1997). Deep marine biosphere fuelled by increasing organic matter availability during burial and heating. Nature 388, 573-576. Wellsbury, P., and Parkes, R. J. (1995). Acetate bioavailability and turnover in an estuarine sediment. FEMS Microbiol. Ecol. 17, 85-94, Westrich, J. T, and Bemer, R. A. (1984). The role of sedimentary organic matter in bacterial sulfate reduction: The G model tested. Limnol. Oceanogr 29, 236-249.
Sediment Pore Waters
663
Whelan, J. K. (1977). Amino acids in a surface sediment core of the Atlantic abyssal plain. Geochim. Cosmochim. Acta 41, 803-810. Whelan, J. K., and Emeis, K.-C. (1992). Sedimentation and preservation of amino compounds and carbohydrates in marine sediments. In "Productivity, Accumulation, and Preservation of Organic Matter in Recent and Ancient Sediments" (J. K. Whelan and J. W Farrington, Eds.), pp. 176-200. Columbia Univ. Press, New York. WilHams, P., and Druffel, E. (1988). Dissolved organic matter in the ocean: Comments on a controversy. Oceanogr. Mag. 1,14-17. Williams, P M., and Druffel, E. R. M. (1987). Radiocarbon in dissolved organic matter in the central North Pacific Ocean. Nature 330,246-248.
This Page Intentionally Left Blank
Chapter 14
DOC in the Arctic Ocean Leif G. Anderson Analytical and Marine Chemistry Goteborg University Goteborg, Sweden I. Introduction A. Water Masses and Circulation II. Sources of DOC to the Arctic Ocean A. River Runoff Sources B. Seawater Sources C. Biological Sources within the Arctic Ocean III. Composition and Distribution of DOC within the Arctic Ocean
A. Lignin Oxidation Products and Stable Carbon Isotopes B. C/N Molar Ratios C. Distribution IV. Summary of Sources and Sinks References
I. INTRODUCTION The objective of this chapter is to summarize the present knowledge on the input of terrigenous dissolved organic matter (DOM) to the Arctic Ocean, the input of marine DOM in water masses flowing into the Arctic from the Pacific and Atlantic oceans, and the distribution of terrigenous and marine-origin DOM within the Arctic Ocean. The Arctic Ocean is, together with the Greenland, Iceland, and Labrador seas, a major area of deep-water formation in the Northern Hemisphere (Anderson et al, 1999). As this deep water contributes to the global thermohaline circulation and thus adds to the deep waters of all global oceans, it is of global interest to evaluate the vertical flux of DOM within the Arctic Ocean (and the other deep-water formation sites). In order to address the above aspects of DOM it is essential to consider the water mass formation and circulation within the area. Biogeochemistry of Marine Dissolved Organic Matter Copyright 2002, Elsevier Science (USA). All rights reserved.
665
666
LeifG. Anderson
Several rivers draining large areas of Siberia and North America flow into the Arctic Ocean. The four dominating are the Ob, Yenisey, and Lena from Siberia and the Mackenzie from North America. In addition, much of the runoff from the Yukon River reaches the Arctic Ocean after entering the Bering Sea and flowing north through the Bering Strait. Consequently, the Arctic Ocean receives much freshwater, about 10% of the global runoff (Aagaard and Carmack, 1989), while constituting only about 1% of the global ocean volume (Menard and Smith, 1966). Theserivers,with the major fractions supplied by Siberian rivers, add large amounts of terrigenous DOM to the Arctic Basins (e.g. Gordeev et ai, 1996). A significant fraction of this DOM is dissolved organic carbon (DOC), which is the focus of this review. The large terrestrial component of DOM distinguishes the Arctic Ocean from the Southern Ocean. A. WATER MASSES AND CIRCULATION
Ocean water from the Pacific enters the Arctic Ocean through the Bering Strait and from the Atlantic, through the eastern Fram Strait and the Barents Sea (Fig. 1). Fresh water, in the form of runoff and sea ice meltwater mixes with ocean water in the upper Arctic Ocean and exits the Arctic Basin mostly through the Canadian
5 Russia
Figure 1 Map with geographic information and schematic circulation of surface water (gray arrows) and intermediate water (black arrows). The straight arrows indicate the mouths of the rivers included in Table I.
DOC in the Arctic Ocean
667
Archipelago and western Fram Strait (Jones et al, 1998). Some of the upper waters are entrained and transported to deeper regions. Intermediate-depth water largely follows the topography, resulting in several large loops within the deep central Arctic Ocean (Rudels et al, 1994). Deep water both enters and exits the Arctic Ocean through Fram Strait over a sill at a depth of about 2200 m. The inflowing Pacific water is relatively fresh and contributes significantly only to the upper water masses. However, a small amount of high-salinity water formed during sea ice production penetrates to the deepest parts of the Canadian Basin (Jone5 et al, 1995). The Adantic water, on the other hand, has a saHnity of close to 35 when entering the Fram Strait and the Barents Sea, but the temperature is high enough (around 4°C) for this water to stay at the surface. However, during the transit over the Barents Sea, heat is lost to the atmosphere and, together with brine release from sea-ice production, the density increases to form waters that penetrate to intermediate depths of the Arctic Ocean (e.g., Swift et al, 1983; Schauer et al, 1997). Most of this high-density water enters the Arctic Ocean through the St. Anna Trough, though some water of Atlantic origin passes through the Kara Sea into the Laptev Sea, with a fraction continuing into the East Siberian Sea before meeting water of Pacific origin (Jone5" et al, 1998). The Atlantic water that flows through Fram Strait meets sea ice and the upper part of this warm water melts the sea ice, forming an approximately 100-m-thick surface water layer of low salinity (5 ~34.2) and with temperatures close to the freezing point (Rudels et al, 1996). This constitutes the formation of the lower halocline water (LHW) and prevents deep-water formation within the central Arctic Ocean. In addition, the LHW hampers the penetration of heat from the Adantic Layer water to the overlying sea-ice cover. The Pacific water, and to some extent the Atlantic water, transports significant amounts of nutrients into the Arctic Ocean shelf seas. The high nutrient supply and hydrographic conditions stabilizing the water column result in primary production rates that are high even in a global perspective. In the Bering-Chukchi Sea region, new production has been estimated to be 288 g C m~^ year"^ (Hansell et al, 1993). New productivity in the Barents Sea is also considered high, being stifl higher within the marginal ice zone (Sakshaug and Skjoldal, 1989). Because of the patchy productivity it is difficult to estimate a mean productivity rate, but the vertical carbon flux at 75 m, as simulated by a 3-D model, generally varied between 10 and 40 g C m~^ year"\ depending on forcing conditions (Slagstad and Wassmann, 1996).
11. SOURCES OF DOC TO THE ARCTIC OCEAN The highest concentrations of DOC in source waters to the Arctic Ocean are found in river runoff, with a mean of more than 500 JJM (e.g., Gordeev et al, 1996; Lobbes et al, 2000). This concentration is about an order of magnitude higher than
668
LeifG. Anderson
in the inflowing Atlantic water, but the volume flux of the latter is about 50 times larger than that of the continental runoff. Nevertheless, there is a clear signature of terrigenous DOC in the surface water over the central Arctic Ocean (Opsahl et al, 1999). Many of the investigations of organic carbon in the Arctic Ocean have reported data on unfiltered samples and are therefore total organic carbon (TOC) concentrations. However, often the waters of the Arctic Ocean are very low in particles, which makes the difference between DOC and TOC very little. This is not the case for samples collected during high primary productivity or in turbulent coastal waters.
A. RIVER RUNOFF SOURCES Numerous rivers enter the Arctic Ocean. They drain enormous areas (total drainage basin area > 10 x 10^ km^) with variable vegetation and soil conditions. Consequently, DOC concentrations vary significantly between rivers (Table I). There are also significant seasonal differences in river TOC concentration, as reported for the Lena River by Cauwet and Sidorov (1996). The maximum concentration (980 ^M) was found during the maximum water discharge in early summer, followed by a lower concentration (700 /xM) in the summer and autumn, and the lowest concentration (310 /xM) during winter. The mean annual, discharge-weighted, concentration was 830 jiM. It should be noted that several investigations were performed after maximum water discharge in summer, and not always is the date of sampling given in the literature. To get average discharge weighted concentrations, DOC concentrations were multiplied by each river discharge and divided by the total annual discharge (Table II). Another uncertainty is that around one-third of the discharge to the Arctic Ocean takes place through smaller rivers and creeks that are not included in Table I. An alternative approach for estimating an average discharge weighted concentration for the rivers entering a given area is to sample the estuary and the surrounding sea and make a DOC versus salinity plot (Fig. 2). Assuming that DOC behaves conservatively, the intercept at 5 = 0 corresponds to a discharge weighted mean of the rivers entering the area investigated. This estimate includes the seasonal variability, as the residence time of the runoff on the Eurasian shelves has been estimated to ^ 3 years (Schlosser et al, 1994). This approach is not suitable for the Beaufort Sea area, where the Mackenzie River discharges, as the residence time of the surface water on the shelf in summer is short (Macdonald et al, 1989). The Mackenzie River dominates the discharge from North America into the Arctic Ocean and it is thus more straightforward to evaluate the DOC concentration in the runoff from this continent, than from the Eurasian. The regression lines of Figs. 2C and 2D (from the Laptev Sea region) are in excellent agreement with an intercept of 579 and 580 /xM. The data of Fig. 2B fall
669
DOC in the Arctic Ocean Table I Reported Concentrations of DOC and TOC in Arctic Rivers No.
River
1 2 3 4 5 6 7 8 9 10 11 B.S.
Pechora Ob Pyr Yenisey Katanga Olenek Lena Yana Indigirka Kolyma Mackenzie Yukon
DOC {^iM)
111 850 538 to 558 232 to 264 404 387 375 to 863 357 to 733
TOC (^lM)
References
Shelf seas
1083 592 to 733 558 617 525 600 792 to 842 558 to 611 642 to 754 389 to 675 642 to 1050 476 to 833
d d,g d d,e,g d d,e a, c, d, e, g c, d, e, g c,d,g c,d,g b,dj,g
Barents Kara Kara Kara Laptev Laptev Laptev Laptev East Siberian East Siberian Beaufort Bering
b,f,g
Note. The Yukon river enters the Bering Sea (B.S.), outside the range of Figure 1, but most of its w^ater enters the Arctic Ocean through Bering Strait. "Cauvet and Sidorov (1996). ^Degens et al. (1991). ^Fitznar (1999). '^Gordeev et al (1996). ^Lobbesetfl/. (2000). ^Pocklington (1987). ^Telang et al (1991).
Table II DOC Flux from Major Rivers into the Arctic Ocean (Lobbes et al, 2000) River Mezen
Ob Yenisey delta Olenek Lena delta Yana Indigirka Kolyma Mackenzie Total
Discharge (km"^ year~^)
Drainage basin area (km^ 10^)
DOC flux D O C (jLtM)
(lO^gCyear-i)
56
1006
2,990 2,440
1,805
735 711 850 538 232 404 387 640
248 3,690 4,860 323 3,380 85 241 458 1,917
10,994
636^^
15,200
21 419 569 32 524 31 50 98 249
2,430
1,993
198 244 305 526
"The mean concentration is computed as (DOC flux)/(discharge) x (12).
CD
o o
8 8 CO
(£>
o o
8
in
o o
"
o o
CO
in (t.1 lOUJrl) o o a
<£>
(^.1 lOLurl) o o a
cvj
w CM
s+ S
<0
'
9
II
1
u •
o /o 9Pi
1
'
o/ CD 1-•—r^ '
1 -
1
J-
1
1
\
CM
'
QM
1
L
'
oi
r
1 1
1
CO
8 8 8
'
O" {O
'
°.^ c^ •
1
(t-l lOLurl) o o a
ON
>
o
6, -
G
15 W
2
ON
OH
„
'/J
0\
I-
P. S
DOC in the Arctic Ocean
671
around the same regression line, but without data in the salinity interval 1 to 28 it is not meaningful to make a linear regression calculation. The data West of 115°E in Fig. 2A also falls along the same regression line, but with too few low-salinity data to make a linear regression calculation. Hence, it is not possible to draw any conclusion with regard to the DOC concentration in the runoff entering the Kara Sea relative to that entering the Laptev Sea. A striking feature in Fig. 2A is the large scatter in data east of IIS'^E and the low DOC concentrations in the lowsalinity waters. Three possible explanations for the scatter can be considered. First, these samples were not filtered prior to analysis, which might both result in higher concentration and larger scatter. Second, these samples were analyzed on board the ship, using a Shimadzu TOC 5000, which is sensitive to vibrations, and thus also likely to contributes to scatter. Finally, the low-salinity samples were mainly collected in the East Siberian Sea, where sea-ice melt contributed significantly to the freshwater (Olsson and Anderson, 1997). Since sea ice can have highly variable TOC concentrations depending on biological activity in and under the ice, this can add to the scatter. The consequence of all these uncertainties is that it is difficult to evaluate any mean runoff DOC (or TOC) concentration for the East Siberian Sea, using the available data. Guay et al. (1999) used a UV fluorometer on the SCICEX-97 cruise to record a continuous in situ record of fluorescence at excitation 320 nm and emission 420 nm along the submarine track, at a depth of about 50 m. The fluorescence measurements can be used to estimate the concentrations of the humic-rich terrestrial component of DOM. The measured fluorescence detector response (V) gave a linear correlation to the TOC concentrations (TOC = 94.8 x V, r^ = 0.84, n = 186), the latter measured by high-temperature combustion on samples collected roughly every hour during the cruise. Also particulate organic carbon was determined on some samples and that accounted for less than 4.2% of the measured TOC, leading the authors to conclude that DOC ^ TOC. The DOC and salinity data collected along the continental slope of the Makarov and Amundsen Basins, show a linear correlation of DOC = -18.5 x 5 + 705 /xM (r^ = 0.76, n = 4914). This is interpreted as data from a region with high DOC Eurasian runoff mixing with waters of Atlantic origin with low DOC. The resulting runoff concentration of 705 /xM is higher than that found from measurements in the Laptev Sea. However, there might be temporal variability in the runoff DOC concentration and also the data reported by Guay et al. (1999) covered a salinity range of about 32 to 34, making the computed intercept at 5' = 0 somewhat uncertain. These data help identify regions with runoff from the shelf seas to the deep central Arctic Ocean.
B.
SEAWATER SOURCES
The reported DOC concentrations in the inflowing water from the Atlantic vary between 52 and 75 JJLM (Wheeler et al, 1997; B0rsheim and Mycklestad, 1997;
672
LeifG. Anderson
Opsahl et al, 1999; Fransson et al, 2001). Some of this variation might be a result of analytical errors, but in addition different water masses were sampled, and they were not sampled at the same time. Inflowing water from the Atlantic includes surface water flowing through Fram Strait and the Barents Sea, as well as deeper water flowing through Fram Strait. Deeper Atlantic water is modified between the Greenland-Scotland Ridge and the Arctic Ocean. The Norwegian Sea deep water, flowing north through Fram Strait, is a mixture of Eurasian Basin deep water and Greenland Sea deep water (Swift et ai, 1983). Some of the Eurasian Basin deep water that exits through Fram Strait flows around the Greenland Sea and mixes with the Greenland Sea deep water before it reenters the deep Arctic Ocean. It is difficult to assign a specific DOC concentration to the water flowing in through Bering Strait as the inflow consists of several different water masses and considerable modifications take place during the transit through the Bering Sea. The Arctic Ocean Section expedition in 1994 showed a TOG concentration span of 50 to 110 /JLM in the waters with Pacific-derived characteristics (Wheeler et al, 1997). When the particulate carbon fraction was subtracted the DOC concentrations had a span of 20 to 100 /xM (their Fig. 8). The mean concentration computed for the samples at the slope stations equals ~70 ± 1 5 /xM. Using a Lagrangian model, Walsh et al. (1997) computed a monthly depth average DOC concentration of between 67 and 134 fiM at positions over the 80-m isobath of the northwestern Chukchi Sea. The annual average of the monthly values is ^^90 /iM, which is in fairly good agreement with the Wheeler et al. (1997) mean DOC measurements of the slope stations north of the Chukchi Sea.
C. BIOLOGICAL SOURCES WITHIN THE ARCTIC OCEAN
An additional source of DOM in the Arctic Ocean is through biological processes within the Arctic Ocean and its shelves. Primary productivity over the continental shelves is substantial and results in a significant seasonal production of marine DOM. This seasonal signal can be observed in surface waters. The outflow from the Barents Sea into the central Arctic Ocean through the St. Anna Trough (containing insignificant fraction of river runoff) showed elevated surface DOC concentrations relative to waters below 150 m (Fransson et al, 2001). At depths shallower than 150 m, the nutrient distribution indicated that primary production occurred in the surface water during the transit over the Barents Sea. Subtracting the deep-water DOC concentration (average of 52 /xM) from surface DOC values estimates the labile, i.e., freshly produced, part of the DOC (cf. Hansell and Carlson, 1998). The labile DOC amounts to 1.4 mol C m"^ integrated over the top 150 m (Fransson et al., 2001), which will be exported to the central Arctic Ocean. The Chukchi Sea also has a high biological productivity. In analogy with the above estimate of exported DOC, the difference between the highest (134 /xM)
DOC in the Arctic Ocean
673
and lowest (67 /xM) monthly depth average, DOC concentrations at the continental break of the northwestern Chukchi Sea (Walsh et al, 1997) should reflect the marine DOC exported into the central Arctic Ocean from this area. The productivity of the central Arctic Ocean is small compared to the shelf seas. Based on the computed deficit of phosphate in surface waters, Anderson et al. (2000) estimated an average export production of 0.04 mol m~^ year~^ over the central Arctic Ocean. In contrast an in situ DOC production of over 0.5 mol m~^ year~^ (assuming a 120-day productive season) was computed for the central Arctic Ocean in 1994 (Wheeler et al, 1997). If this DOC production is distributed over the top 50 m, it will result in a concentration increase of 10 /xM. In order to sustain such an annual DOC production over the residence time of the Arctic Ocean surface water, 5-10 years, an unrealistic DOC concentration would follow, indicating extensive recycling. Hence, it is essential to consider the seasonal production and degradation of marine DOC in the Arctic Ocean. Data are unfortunately not available to make such an evaluation, not even in the shelf seas. From extensive measurements, DOC in sea ice has been attributed to ice algae production (e.g.. Smith et al, 1997). In shelf seas receiving much runoff, the sea ice produced will include some terrigenous DOC. In the spring (beginning of April to end of May) when the sea ice algae develops, Smith et al. (1997) found a good correlation between chlorophyll a and DOC in the bottom ice of Resolute Passage in the Canadian Archipelago. DOC concentrations were much higher in ice than in underlying water, especially in ice covered with only a thin snow layer. The highest DOC concentration (>3000 /xM) was measured in the bulk of the bottom ice on May 14 (Smith et al, 1997). However, the volume with this high concentration is small and thus the integrated contribution of DOC from ice to the underlying water mass is small. Measurements of DOC release rates by ice algae were performed by Gosselin et al (1997) along the track of the Arctic Ocean section in 1994. The release rate varied from less than 25 /xmol m~^ day~^ to 1600 lb 1500 /xmol m~^ day~^ with the highest rates in the Chukchi Sea. Thomas et al (1995) collected three ice cores of more than 2 m length in the Fram Strait. In two of these the DOC concentration was mostly below 100 /xM all through the core. In the third the concentration was close to 100 /xM in the top ^1.8 m, and increased to a maximum of ~700 /xM some 10 cm from the bottom. This increase was explained by a combination of DOM excretion by biota and decomposition of organisms (Thomas et al, 1995). The mean bulk concentration of DOC in sea ice from the central Arctic Ocean is 316 ± 99 /xM (Melnikov, 1997). If 1 m of ice melts annually, the concentration in the top 50 m (typical winter surface mixed layer (Rudels et al, 1996)) would increase by just over 6 /xM. A further source of DOC from biological processes is release from the sediment surface caused by decomposition of particulate organic material. Hulth et al (1996) measured DOC concentrations in the range of 500 to 8000 /xM in pore
674
LeifG. Anderson
water in the Svalbard area. The lowest concentrations were found at stations east of Svalbard, where also a significant inverse linear correlation (r^ = 0.849) of DOC concentrations with a sediment reactivity index (defined as sediment oxygen consumption rate normalized to the organic content) was found. This suggests a coupling between reactivity of organic matter in sediment and DOC lability in pore water. In a study of the eastern Eurasian Basin and adjacent shelves (Hulthe and Hall, 1997), DOC fluxes out of the sediment were evaluated to be in the range from close to zero to 3.6 mmol m"-^ day~^. The highest fluxes were found on the shelves and the lowest in the deep basins and on the slopes. A positive correlation of the DOC and dissolved inorganic carbon fluxes was observed, with DOC constituting up to 50% of the total benthic carbon flux at stations with the highest total benthic carbon fluxes. This indicates that the fraction of DOC that is oxidized to inorganic carbon is decreasing with increasing decomposition rates.
III. COMPOSITION AND DISTRIBUTION OF DOC WITHIN THE ARCTIC OCEAN Before the transport of DOC to and from of the Arctic Ocean is discussed, the quality of the terrigenous DOM has to be considered. Does it flow with the water as a biogeochemically stable solute or is it available to diagenetic alteration or photochemical decomposition? Several investigations have studied the composition of DOM in rivers entering the Arctic Ocean (e.g., Gordeev et al, 1996; Cauwet and Sidorov, 1996; Lara et al, 1998; Lobbes et ai, 2000) as well as in the Arctic Ocean itself (e.g., Wheeler ^r a/., l991;OpsahletaL, 1999; Kattner^r^/., 1999). One general conclusion is the stability of terrigenous DOC in the surface waters of the Arctic Ocean. Except for the Unear mixing line of runoff and seawater in a DOC vs salinity plot, the fairly constant composition of the DOM in all of the Arctic Ocean supports this conclusion.
A. LiGNiN OXIDATION PRODUCTS AND STABLE CARBON ISOTOPES The most useful quantitative tracers of terrestrial organic matter are lignin oxidation products, which have been determined in runoff to the Arctic Ocean (Opsahl et ai, 1999; Lobbes et al, 2000) and in the surface waters of the Arctic Ocean (Opsahl et al, 1999; Kattner et al, 1999). Kattner et al (1999) determined lignin in the "humic" fraction of DOM and used this as a tracer for terrigenous influence, with the result that the riverine-derived freshwater contribution to the Laptev Sea is 8 to 30%. Combining this proportion with DOC concentrations in the Lena River and Laptev Sea indicates that about 60% of the DOC in the surface layer of the Laptev Sea and adjacent Eurasian Basin would be of terrigenous
675
DOC in the Arctic Ocean
origin. In contrast, terrigenous dissolved organic nitrogen (DON) only accounted for 20 to 30% of the total DON (Kattner et al, 1999). However, as stressed by the authors, the distribution of DON is generally more influenced by biological processes, making this last estimate more uncertain. The fraction of terrigenous DOM in surface waters of the central Arctic Ocean was estimated from the carbon-normalized yields of lignin oxidation products (Ae) and (5^^C in ultrafiltered dissolved organic matter (UDOM) (Opsahl et al, 1999), resulting in 5-22% and 16-33%, respectively. The UDOM represents the highmolecular-weight fraction of DOM (>1 kDa), which is about 20-30% of total DOM. In Fig. 3 the mean values (ibvariability) of samples from the Kara Sea (low (5^^C), the polar surface water (medium 5^^C), and deep Fram Strait and Greenland Sea (high <5^^C) (Opsahl et al, 1999) are plotted versus A6. The polar surface water (32.04<5'<34.49) samples in this investigation were collected from submarines at depths of 38 to 165 m, within the SICEX program. Hence, these data do not include low-salinity surface waters. The relative contribution of terrigenous DOM in the polar surface water was computed from the mixing line of Fig. 3 to 15 ± 6%, where the error represents the extreme variability of the data. This computation is based on the same hypothesis as the estimate by Opsahl et al (1999), that the deep Fram Strait and Greenland Sea data represents marine-derived organic matter and the Kara Sea data represents terrigenous-derived organic matter. If A6 and 8^^C are conservative, the data would fall along a straight line, and the estimates of Opsahl et al (1999) and that from Fig. 3 would be equal.
-20 , Deep Fram Strait -21 -Pt.. and Greenland Sea -22 «
^-23. O S -24
P -25.
:
*
•
Polar Surface Water Kara Sea
-26-1 -27 H -28 0.0
—I—
0.2
—I—
0.4
—I—
0.6
— 1 —
0.8
1.0
AeCmg/IOOmgOC) Figure 3 The mean values (± standard deviation) of carbon normalized yields of lignin oxidation products (A6) versus stable carbon isotopic composition (5^^C) of DOM samples from the Kara Sea, the polar surface water and deep Fram Strait and Greenland Sea. The open circle indicates value from one sea ice sample. All data from Opsahl et al. (1999).
676
Leif G.Anderson
Syringyl and vanillyl phenols are two of the oxidation products from lignin. The ratio of syringyl and vanillyl (SA^) has been shown to be an indicator of oxidative changes. Investigations suggest that the SA^ ratio is reduced by diagenetic alterations in the Atlantic and Pacific oceans (Opsahl and Benner, 1997). Likewise, photochemical degradation can selectively alter SA^ of terrigenous DOM oceans (Opsahl and Benner, 1998). Lobbes et aL (2000) showed that the S/V ratio also is a biomarker to distinguish between DOM originating from angiosperm plants (high SA^ ratios) and gymnosperms (low SA^ ratios) in Russian rivers entering the Arctic Ocean. Opsahl et al. (1999) determined the SA/^ ratio in UDOM for different regions of the Arctic Ocean, showing ratios not too different in the Kara Sea (0.3-0.5; n = 9) and central Arctic Ocean (0.12-0.31; n= 13) samples. Consequently, the relatively constant S A'^ ratio within the Arctic Ocean indicates a limited alteration of terrigenous DOM in this region.
B. C/N MOLAR RATIOS The molar ratio of C/N is the most studied property of DOM and can be used as a tracer for the origin of DOM. The C/N ratio is generally high in terrigenous DOM and low in marine DOM. An average C/N ratio of 20.5 ± 2.6 was reported for seven Siberian rivers (Gordeev et al, 1996, recalculated by Wheeler et al, 1997). Cauwet and Sidorov (1996) found a similar value (22) for the Lena river, while significantly higher ratios (30 to 58) were reported by Lara et al. (1998) for different locations along the Lena River. Lobbes et al. (2000) reported data from several rivers, where the mean C/N ratio for Yenisey, Olenek, Lena, Yana, and Indigirka was 47 ib 10. The variabiHty in the reported ratios is mainly a result of variable DON concentrations. The high C/N ratios of DOM in the runoff are characteristic of riverine fulvic acids (Thurman, 1985). The C/N ratio of marine DOM is dependent on biological activity in the investigated water mass. When the C/N ratio is plotted versus salinity for samples collected in the outer Laptev Sea, at the continental margin and in the eastern part of the Eurasian Basin, two regimes can be identified (Fig. 4). At salinities below 34.5 (depth <100 m), the C/N ratio increases with decreasing salinity (o in Fig. 4) and at saUnities above 34.5 (depth >100 m), the ratio varies from 12 to 30 (x in Fig. 4). The signature at 5'<34.5 is mainly a result of water of Atlantic origin mixing with runoff, supported by the intercept 49.7 at 5 = 0 of the fitted line. No trend but a large C/N span can be seen in the waters of iS'>34.5, which likely is a result of these samples being deep waters and thus have a signal affected by decay of sinking organic particulate matter. A large variability in the C/N ratio of DOM, ranging from 10 to 40 with a peak around 15, was also observed in the Fram Strait (Lara et al, 1998). The variable C/N ratio in marine dominated waters is a result of variable DON concentrations. This makes the C/N ratio less useful for
677
DOC in the Arctic Ocean
30-
C/N = -0.943*S + 49.7 F^= 0.2799
25 H
,
O Q
o o o
I 20. c3
X
o
o o
15H
X X
o^o
X
10-
I
28
I
29
30
—r32 Salinity
31
33
"~T—
34
35
Figure 4 C/N ratio in dissolved organic matter (DOC/DON) in the eastern Eurasian Basin. Open circles represent S <34.5, crosses represent S >34.5. The linear regression line is fitted to the open circles. Data are from Fitznar (1999).
quantitative computations, but it is valuable as a qualitative tracer of terrigenous DOM.
C. DISTRIBUTION Too few DOC data are available from the central Arctic Ocean surface waters to produce a map of the concentration distribution. Several processes considerably influence the distribution by producing and consuming DOC. The relative importance of these processes can be seen in a DOC versus salinity plot of the surface waters with 5<34.5 (Fig. 5). A line representing the conservative mixing of Atlantic water (S = 34.926 and DOC = 60 /xM) and runoff (S = 0 and DOC = 550 jjM) is included as reference. Data with 5'>34 (Anderson et al, 1994; Opsahl et al, 1999) are spread around the mixing line, showing that consumption and production of DOC balance. At saHnities around 33, the Opsahl et al. (1999) data are below the mixing line, while the Anderson et al. (1994) data are above the mixing line. The latter is likely caused by biological activity as shown by Wheeler et al. (1997). Their data from the Arctic Ocean section 1994 show a similar trend (see Fig. 5 in Wheeler et al, 1997), but with fewer data above the mixing line.
LeifG. Anderson
678 160
Salinity Figure 5 DOC versus salinity for the samples from the central Arctic Ocean with 5 <34.5. The open circles are data from the Oden 91 cruise (Anderson et al, 1994), while the square illustrates the range of data from Opsahl et al. (1999). A line representing the conservative mixing of Atlantic water {S = 34.926 and DOC = 60 /LIM) and river runoff (5 = 0 and DOC = 555 /AM) is included as a reference.
However, sea-ice meltwater also lowers the salinity, and the DOC concentration (316 ± 99 /xM; Melnikov, 1997) is lower in sea-ice meltwater than in the runoff. Nevertheless, the DOC concentration in the surface waters of the central Arctic Ocean is negatively correlated with salinity to a large degree. The DOC concentrations in Arctic Ocean deep waters are lower than in the inflowing Atlantic water (Opsahl et ai, 1999; Bussmann and Kattner, 2000). Opsahl et al. (1999) found 61 /xM in the inflowing Atlantic water of Fram Strait and 65 /xM in that recirculating in Fram Strait, while Bussmann and Kattner (2000) found 59 /xM (n = 37) in the Atlantic layer of the central Arctic Ocean. The mean deep-water concentration was 50 /xM in the Nansen Basin (n = 53), 54 /xM in the Amundsen Basin (n = 67), and 56 /xM in the Makarov Basin (Bussmann and Kattner, 2000). These values agree well with the observations in the outflowing deep waters from the Canadian and Eurasian Basins, 53 and 49 /xM, respectively (Opsahl et al, 1999). The difference in DOC concentration between the inflowing Atlantic water and the Arctic Ocean deep waters is in the order of 10 /xM, which is not much above the analytical range of accuracy. However, it is realistic to expect a higher DOC concentration in the Atlantic water relative to the Arctic Ocean deep waters as the former has been exposed to a larger flux of particulate organic matter from above. Furthermore, the Arctic Ocean deep waters have a long residence time (>100 years), with a limited flux of particulate organic matter from above,
DOC in the Arctic Ocean
679
resulting in a DOC decomposition rate that could be larger than the production rate by decay of particulate organic matter. The lower DOC concentrations in the Arctic Ocean deep waters do not exclude an export of terrigenous DOM to the deep waters, as this is a function of sources and sinks. However, both the low concentration of lignin oxidation products and the predominance of a marine 8^^C signature indicate that terrigenous DOM is a minor contribution to the DOC of the deep waters of the Arctic Ocean (Opsahl et al, 1999).
IV. SUMMARY OF SOURCES AND SINKS A budget of the fluxes to and from the Arctic Ocean is given in Table III, based on measured concentrations of DOC and reported volume fluxes of the different waters. This budget does not distinguish between terrigenous and marine DOC. Generally the terrigenous DOC is high in the surface waters and low in the deep waters (Opsahl et al, 1999; Fitznar, 1999). It should be noted that the DOC budget of Table III is around 15% lower than that reported by Anderson et al (1998), a result of much new high-quality data being collected during the past few years, as referred to in Section III. The uncertainties given in Table III are based on the variability in reported DOC concentrations for the different water masses. No considerations of uncertainties in volume fluxes are included. Fortunately, the largest uncertainties are in the Atlantic and deep-water volumefluxesand these waters have a fairly constant DOC concentration. Consequently, an error in the volume influx has to be compensated by a comparable error in the volume outflux and hence have a small impact on the net DOC flux out of the Arctic Ocean. Adding the in- and outfluxes of Table III gives —5 ± 9x 10^^ g C year~\ indicating that the Arctic Ocean is neither a sink nor a source of DOC considering the uncertainty in the estimate. The in situ production of marine DOC within the central Arctic Ocean has been estimated to 6.1 g C m"-^ year~^ and the in situ respiration to 8.8 g C m~^ year~^ (Wheeler et al, 1997). Combining these numbers with the area of the deep central Arctic Ocean (5.8 x 10^^ m^) gives a total m^to production of 35 x lO^^gCyear"^ and a total in situ respiration of 51 x 10^^ g C year~^ These numbers are based on one summer investigation in a limited area and the uncertainties must be significant when applying them to a whole year and the whole central Arctic Ocean. Nevertheless, it is interesting to note that the in situ respiration of DOC exceeds that of in situ production of marine DOC, while the latter is of the same order as the added terrigenous DOC (35 x 10^^ g C year~^ relative to 23 x 10^^ g C year~^). These results indicate that the in situ respiration of DOC in the central Arctic Ocean will quantitatively consume all marine DOC produced in the central Arctic Ocean and some of that added by river runoff. This estimate does not include the DOC produced by the biota on the shelves.
680
LeifG. Anderson Table III A Budget of the DOC Fluxes to and from the Arctic Ocean Water mass
Volume flux (Sv)
DOC (AtM)
Atlantic water Deep water Pacific water Runoff
2.5 0.58 0.83 0.11
58 ±5^^ 53 ± 5 ^ 71 ± 20^ 555 ± 50^
Total in
4.02
Organic carbon transport, (1012• g C year-i)
In
Out Sea ice From EB:
From CB:
Total out Net outflow
- Surface mixed layer - Halocline - Atlantic layer - Deep water - Surface mixed layer - Halocline - Atlantic layer - Deep water
0.11 0.165 0.25 0.9 0.42 0.362 0.54 0.698 0.575 4.02
55 12 22 23
±5 ±1 ±6 ±2
112 ± 8 316 ±50^ 82 ± 15/ 70 ± 6 / 58 ± 4 / 51 ± 5 ^ 100 ± 10/
75 ± nf 53d=4/ 55 lb 5^
13 ± 2 5±1 1± 1 20 ± 1 8±1 14 ± 1 15 ± 2 14 ± 1 12 ± 1 107 ± 4 -5
±9
Note. The volume fluxes of the water masses are from Anderson et al (1998), while the DOC concentrations are means of literature values. The river runoff includes all continental freshwater input, and outflows are from the Eurasian Basin (EB) and Canadian Basin (CB), respectively. As discussed in the text, errors in the organic carbon transport figures do not include errors in the volume fluxes. ^Mean of Wheeler et al (1997), Opsahl et al (1999), and (Bussmann and Kattner, 2000). ^Data in the Greenland Sea at 1800 m (Opsahl et al, 1999). ^Mean of Walsh et al (1997), Wheeler et al (1997), and Guay et al (1999). ^The mean of the regression lines at S = 0 of Figures 2B-D. ^Melnikov (1997). /Wheeler effl/. (1997). ^Mean of Opsahl et al (1999) and Bussmann and Kattner (2000).
Even if the Arctic Ocean itself is neither a sink nor a source of DOC there is a significant export of DOC to the North Atlantic. This flux (29 x 10^^ g C year"^) is a combined result of inflow from the Pacific Ocean (22 x 10^^ g C year"^), river runoff (23 x 10^^ g C year~^), and the difference between in situ production (35 X 10^^ g C year"^) and respiration (51 x 10^^ g C year"^) within the Arctic Ocean. The above arguments together with the balanced budget support the idea that terrigenous DOC is relatively stable within the Arctic Ocean.
DOC in the Arctic Ocean
681
The finding that the outflowing deep water DOC concentrations are lower than (or very similar to) the inflowing water concentrations indicate (/) that little terrigenous DOM is exported to deep layers (as was also concluded by Opsahl et al, (1999) on the basis of lignin analysis) and (//) that little net export of marine DOM occurs to deep layers. The latter statement is supported by the arguments above that most marine DOC produced in the central Arctic Ocean is respired in the surface layers. The fact that little terrigenous DOC is exported to the deep waters of the Arctic Ocean through dense plumes originating on the shelves, where they are initiated by brine drainage from sea ice production, is an important finding as it put constraints on the global DOC budget. With regard to the total carbon budget in and out of the Arctic Ocean, the DOC fluxes are about 5% of the total carbon fluxes, calculated as the sum of inorganic and organic carbon. However, while the dissolved inorganic carbon concentration largely has a positive correlation with salinity, the DOC concentration has a negative one. Consequently, in situ production and respiration of DOC plays a relatively more important role for the carbon cycle in the low-salinity surface waters, relative to deeper layers, and it is the surface water that is in contact with the atmosphere linking the marine carbon cycle to climate.
ACKNOWLEDGMENTS Gerhard Kattner, Annelie Skoog, and Robert Benner gave valuable comments that helped improve the present manuscript. Financial support from the Swedish Research Council is greatly acknowledged.
REFERENCES Aagaard, K., and Carmack, E. C. (1989). The role of sea ice and other fresh water in the Arctic circulation. /. Geophys. Res. 94,14,485-14,498. Anderson, L. G., Bjork, G., Holby, O., Kattner, G., Koltermann, R K., Jones, E. P., Liljeblad, B., Lindegren, R., Rudels, B., and Swift, J. H. (1994). Water masses and circulation in the Eurasian Basin: Results from the Oden 91 North Pole Expedition. J. Geophys. Res. 99, 3273-3283. Anderson, L. G., Jones, E. P., and Rudels, B. (1999). Ventilation of the Arctic Ocean estimated by a plume entrainment model constrained by CFCs. /. Geophys. Res. 104,13,423-13,429. Anderson, L. G., Jones, E. P., and Swift, J. H. (2000). Export production in the central Arctic Ocean as evaluated from phosphate deficit. Submitted for publication. Anderson, L. G., Olsson, K., and Chierici, M. (1998). A carbon budget for the Arctic Ocean. Global Biogeochem. Cycles. 12,455^65. B0rsheim, K. Y., and Myklestad, S. M. (1997). Dynamics of DOC in the Norwegian Sea inferred from monthly profiles collected during 3 years at 66°N, 2°E. Deep-Sea Res. 44, 593-601. Bussmann, I., and Kattner, G. (2000). Distribution of dissolved organic carbon in the central Arctic Ocean: The influence of physical and biological properties. /. Mar. Sys. 27,209-219.
682
Leif G. Anderson
Cauwet, G., and Sidorov, I. (1996). The biogeochemistry of Lena River: Organic carbon and nutrients distribution. Mar. Chem. 53, 211-227. Degens, E. T., Kempe, S., and Richey, J. E. (1991). Summary: Biogeochemistry of major world rivers. In "Biogeochemistry of Major World Rivers" (E. T. Degens, S. Kempe, and J. E Richey, Ed.), pp. 323-347. Wiley, New York. Fitznar, H. R (1999). D-Amino acids as tracers for biogeochemical processes in the river-shelf-oceansystem of the Arctic. Ber. Polarforsch. 334, [in German]. Fransson, A., Chierici, M., Anderson, L. G., Bussman, I., Kattner, G., Jones, E. P., and Swift, J. H. (2001). The importance of shelf processes for the modification of chemical constituents in the waters of the eastern Arctic Ocean. Com. Shelf Res. 21, 225-242. Gosselin, M., Levasseur, M., Wheeler, P. A., Homer, R. A., and Booth, B. C. (1997). New measurements of phytoplankton and ice algal production in the Arctic Ocean, Deep-Sea Res. II44, 1623-1644. Gordeev, V. V., Martin, J. M., Sidorov, I. S., and Sidorova, M. V. (1996). A reassessment of the Eurasian river input of water, sediment, major elements, and nutrients to the Arctic Ocean. /. Am. Sci. 296, 664-691. Guay, C. K., Klinghammer, G. P, Falkner, K. K., Benner, R., Coble, P G., Whitledge, T. E., Black, B., Bussel, F. J., and Wagner, T. A. (1999). High-resolution measurements of dissolved organic carbon in the Arctic Ocean by in situ fiber-optic spectrometer. Geophys. Res. Lett. 26,1007-1010. Hansen, D. A., and Carlson, C. A. (1998). Net community production of dissolved organic carbon. Global Biogeochem. Cycles. 12,443^53. Hansen, D. A., Whitledge, T. E., and Goering, J. J. (1993). Patterns of nitrate utilization and new production over the Bering-Chukchi shelf. Cont. Shelf Res. 13,601-628. Hulth, S., Hall, P O. J., Blackburn, T. H., and Landen, A. (1996). Arctic sediments (Svalbard): Pore water and solid phase distributions of C, N, P and Si. Polar Biol. 16,447^62. Hulthe, G., and Hall, P. (1997). Benthic carbon fluxes—DOC versus ECO2 in shelf, slope and deep-sea environments, and relation to oxygen fluxes. Rep. Polar Res. 22^, 115-116. Jones, E. P., Anderson, L. G., and Swift, J. H. (1998). Distribution of Atlantic and Pacific waters in the upper Arctic Ocean: Implications for circulation. Geophys. Res. Lett. 25,765-768. Jones, E. P., Rudels, B., and Anderson, L. G. (1995). Deep waters of the Arctic Ocean: Origin and circulation. Deep-Sea Res. 42,131-160. Kattner, G., Lobbes, J. M., Fitznar, H. P, Engbrodt, R., Nothig, E.-M., and Lara, R. J. (1999). Tracing dissolved organic substances and nutrients from the Lena River through Laptev Sea (Arctic). Mar Chem. 65, 25-39. Lara, R. J., Rachold, V., Kattner, G., Hubberten, H. W, Guggenberger, G., Skoog, A., and Thomas, D. N. (1998). Dissolved organic matter and nutrients in the Lena River, Siberian Arctic: Characteristics and distribution. Mar Chem. 59, 301-309. Lobbes, J. M., Fitznar, H. P., and Kattner, G. (2000). Biogeochemical characteristics of dissolved and particulate organic matter in Russian rivers entering the Arctic Ocean. Geochim. Cosmochim. Acta. 64, 2973-2983. Macdonald, R. W, Carmack, E. C , McLaughlin, F. A., Iseki, K., Macdonald, D. M., and O'Brien, M. C. (1989). Composition and modification of water masses in the Mackenzie shelf estuary. J. Geophys. Res. 94,18,057-18,070. Melnikov, I. A. (1997). "The Arctic Ice Ecosystem." Gordon and Breach Science Pubhsher, The Netherlands. Menard, H. W., and Smith, S. M. (1966). Hypsometry of ocean basin provinces. /. Geophys. Res. 71, 4305^325. Olsson, K., and Anderson, L. G. (1997). Input and biogeochemical transformation of dissolved carbon in the Siberian shelf seas. Cont. Shelf Res. 17, 819-833. Opsahl, S., and Benner, R. (1997). Distribution and cycling of terrigenous dissolved organic matter in the ocean. Nature. 386,480-482.
DOC in the Arctic Ocean
683
Opsahl, S., and Benner, R. (1998). Photochemical reactivity of dissolved lignin in river and ocean waters. Limnol. Oceanogr. 43,1297-1304. Opsahl, S., Benner, R., and Amon, R. M. W. (1999). Major flux of terrigenous dissolved organic matter through the Arctic Ocean. Limnol Oceanogr. 44,2017-2023. Pocklington, R., (1987). "Arctic rivers and their discharge," Vol. 64, pp 261-268. Mitt. Geol.-Palaontol. Inst. Univ., Hamburg. Rudels, B., Anderson, L. G., and Jones, E. P. (1996). Formation and evolution of the surface mixed layer and halocline of the Arctic Ocean. /. Geophys. Res. 101, 8807-8821. Rudels, B., Jones, E. P., Anderson, L. G., and Kattner (1994). On the intermediate depth waters of the Arctic Ocean. In "The Polar Oceans and Their Role in Shaping the Global Environment" (O. M. Johannessen, R. D. Muench, and J. E. Overland, Eds.), pp. 3 3 ^ 6 . American Geophysical Union, Washington, DC. Sakshaug, E., and Skjoldal, H. R. (1989). Life at the ice edge. Ambio. 18,60-67. Schauer, U., Muench, R., Rudels, B., and Timokhov, L. (1997). The impact of eastern Arctic shelf waters on the Nansen Basin intermediate layers. J. Geophys. Res. 102, 3371-3382. Schlosser, P., Bauch, D., Fairbanks, R., and Bonisch, G. (1994). Arctic river-runoff: mean residence time on the shelves and in the halocline. Deep-Sea Res. 41,1053-1068. Slagstad, D., and Wassmann, P. (1996). Climate change and carbon flux in the Barents Sea: 3-D simulations of ice-distribution, primary production and vertical export of particulate organic carbon. Mem. Nat. Inst. Polar Res. 51,119-141. Smith, R. E. H., Gosselin, M., Kudoh, S., Robineau, B., and Taguchi, S. (1997). DOC and its relationship to algae in bottom ice conununities. /. Mar. Syst. 11,71-80. Swift, J. H., Takahashi, T, and Livingstone, H. D. (1983). The contribution of the Greenland and Barents Seas to the deep water of the Arctic Ocean. /. Geophys. Res. 88, 5981-5986. Telang, S. A., Pocklington, R., Naidu, A. S., Romankevich, E. A., Gitelson, L I., and Gladyshev, M. L (1991). Carbon and Mineral Transport in Major North American, Russian Arctic, and Siberian Rivers: The St. Lawrence, the Mackenzie, the Yukon, the Arctic Alaskan Rivers, the Arctic Basin Rivers in the Soviet Union, and the Yenisey. In "Biogeochemistry of Major World Rivers" (E. T. Degens, S. Kempe, and J. E. Richey, Eds.), pp. 75-104. Wiley, New York. Thomas, D. N., Lara, R. J., Eicken, H., Kattner, G., and Skoog, A. (1995). Dissolved organic matter in Arctic multi-year sea ice during winter: Major components and relationship to ice characteristics. Polar Biol. 15,417-4S3. Thurman, E. M. (1985). Aquatic humic substances. In "Organic Geochemistry of Natural Waters," pp. 273-361. Nijhoff/Junk PubHshers, Dordrecht. Walsh, J. J., Dieterle, D. A., MuUer-Karger, F. E., Aagaard, K., Roach, A. T, Whitledge, T. E., and Stockwell, D. (1997). CO2 cycling in the coastal ocean. II. Seasonal organic loading of the Arctic Ocean from source waters in the Bering Sea. Cont. Shelf Res. 17,1-36. Wheeler, P. A., Watkins, J. M., and Hansing, R. L. (1997). Nutrients, organic carbon and organic nitrogen in the upper water column of the Arctic Ocean: Implications for the sources of dissolved organic carbon. Deep-Sea Res. II44,1571-1592.
This Page Intentionally Left Blank
Chapter 15
DOC in the Global Ocean Carbon Cycle Dennis A. Hansell^ Bermuda Biological Station for Research, Inc., St. Georges Bermuda
I. Introduction II. Distribution of DOC A. Spatial Variability at the Basin Scale B. Temporal Variability III. Net Community Production of DOC A. Evidence for Net Production of DOC B. Regional and Global Estimates for Net Production of DOC
C. Nutrient Depletion and Net Production of DOC IV. Contribution of DOC to the Biological pump A. Evidence for DOC Export B. Exportable DOC V. Research priorities VI. Summary References
L INTRODUCTION Dissolved organic carbon (DOC) makes up the second largest of the bioreactive pools of carbon in the ocean (680-700 Pg C; WilHams and Druffel, 1987; Hansell and Carlson, 1998a), second to the very large pool of dissolved inorganic carbon (38,000 Pg C). The size of the reservoir, as well as its positions as a sink for autotrophically fixed carbon and as a source of substrate to microbial heterotrophs, ^ Present address: University of Miami, Rosenstiel School of Marine and Atmospheric Science, 4600 Rickenbacker Causeway, Miami, FL 33149. Biogeochemistiy of Marine Dissolved Organic Matter Copyright 2002, Elsevier Science (USA). All rights reserved.
685
686
Dennis A. Hansell
indicates that DOC plays a central role in the ocean carbon cycle. But what is this role, how is it realized, and what are its mechanisms and controls. The fundamentals of these questions have remained unchanged over the past 40 years (Duursma, 1962) and continue to challenge the ocean carbon research community today. A considerable amount of financial and intellectual capital has been expended and significant progress has been made over the past decade. DOC concentrations in the ocean range from a deep-ocean low of 34 JJLM to surface-ocean highs of >90 jxM (Section II). Biological processes set up this vertical gradient (net production at the surface and net consumption at depth), while certain physical conditions (high vertical stability) are required to maintain the gradient (Section II.A). The bulk DOC in the ocean can be resolved into at least three fractions, each qualitatively characterized by its biological lability (see Carlson, Chapter 4). All ocean depths contain (1) the very old, biologically recalcitrant DOC (see Bauer, Chapter 8). Its distribution is thought to be fairly uniform in the ocean, largely comprising the DOC of the deep ocean. Built upon the recalcitrant DOC at intermediate and upper layer depths is (2) material of intermediate (or semi-) lability (months to years). It is this material that is produced in the surface ocean and then mixed into the main thermocline, thereby reducing the vertical concentration gradient and contributing to carbon export (Section IV). Concentrations of this fraction can be 10-30 /xM in the upper ocean, and near zero in the deep ocean. The surface ocean alone contains (3) the highly biologically labile fraction of DOC, with lifetimes of days to months and concentrations of just a few to tens of micromolar of C. This latter material is most important for supporting the microbial heterotrophic processes in the ocean (see Carlson, Chapter 4) and shows high variability seasonally. In this chapter, the role of DOC in the ocean carbon cycle is considered in its broadest temporal and spatial scales. The chapter begins with an evaluation of the spatial distribution of DOC at the regional and basin scales, in both the surface and deep ocean. In this context, the distribution of DOC relative to the distribution and timing of marine productivity is evaluated. The older data sets reporting DOC distributions are appraised here as well. The next Section evaluates temporal variability, with consideration of how DOC varies seasonally from high to low latitudes. Following the assessment of variability, the net community production of DOC is examined. The focus is on DOC that accumulates for durations with biogeochemical relevance. This Section is followed by an evaluation of the contribution of DOC to the biological pump. We examine the mechanisms and locations of DOC export, and thus develop an understanding of the controls on export. The chapter concludes with priorities for present and future research, as well as a brief synthesis of the findings reported. Note that organic carbon in the ocean is distributed between the dissolved and particulate (POC) fractions. Summed, these fractions are referred to as total organic carbon (TOC). It is common to measure TOC directly in the water column
DOC in the Global Ocean Carbon Cycle (analysis of unfiltered water), even when DOC is the pool of interest, when the POC concentrations are very low relative to the DOC concentrations. This situation is common at ocean depths well below the surface layer (at depths >200 m), as well as in some surface ocean regions where POC concentrations are normally a few micromolar. The latter conditions are found in the oligotrophic ocean and in highlatitude systems during winter. In these situations (deep water and low-POC surface water), TOC serves as a very close approximation to the DOC concentrations. A primary reason for measuring TOC in these waters, rather than measuring DOC directly, is to avoid contamination by handling (filtering, transfers, etc.) the sample. The term DOC is used in this chapter both for true DOC analyses, and when TOC was measured in deep or POC-impoverished surface waters. The term TOC is reserved for use when DOC and POC, measured separately, are sunmied.
11. DISTRIBUTION OF DOC A. SPATIAL VARIABILITY AT THE BASIN SCALE During the decades leading up to the 1990s, DOC data were relatively sparse because relatively few laboratories made the measurements. An early body of work that stands as providing some of the greatest spatial coverage of an ocean region is that by Duursma (1962). He conducted extensive DOC surveys in the northern reaches of the Gulf Stream and its offshoots south of Greenland, finding the spatial variability associated with hydrographic features that we would likely find today. Since his early work, the few additional ocean sections occupied in the decades leading to the 1990s produced data of uncertain quality (see discussion by Wangersky, 1978, and below). Consequently, our sense for the distribution of DOC in the ocean has been highly uncertain. In this discussion, findings from recently occupied sections in various ocean basins will be discussed (locations identified in Fig. 1) and some of the older sections evaluated. 1. Upper Ocean Distributions Meridional sections from the eastern and western South Pacific and the central Indian Ocean show that the highest upper ocean DOC concentrations are typically found in the low to mid latitudes (Fig. 2 [see color plate]). Concentrations decrease into colder water, whether as a horizontal gradient along the surface from low to high latitudes or vertically with increasing depth. Vertical stability provided by the main thermocline of the open ocean supports the accumulation of DOC in the surface waters. Where vertical stability is strong, DOC concentrations are relatively high; where stability is weak, DOC concentrations can remain at low levels (Fig. 2c). The cold, deep waters have the lowest concentrations, and where
687
688
Dennis A. Hansell
Figure 1 Map depicting locations in the global ocean from which data are shown in this chapter. The solid Unes (gray) represent ocean sections; the triangles are the sites of the time-series stations near Bermuda and Hawaii; the filled circles and the triangles are the sites of the deep-water DOC analyses in addition to the time series sites.
these waters ventilate at high latitudes, similarly low DOC values are present (Figs. 2b and 2c). Where low-DOC, subsurface water mixes to the surface its impact is felt in the surface DOC concentrations. Upwelling sites, both in coastal regions and along the equator, normally have relatively low DOC values at the surface where upwelling is strongest. In the central Equatorial Pacific, DOC is depressed at the surface because of upwelling (note the upward doming of the subsurface DOC contours at the equator; Figs. 2b and 2c). In the central Indian Ocean, where equatorial upwelling is weak, DOC is rather uniform from the subtropical gyre to across the equator (Fig. 2a). DOC along the equator in the Pacific shows the controls by hydrography and biology (Fig. 3 [see color plate]). The Equatorial Undercurrent, near 200 m west of the dateline, has a DOC concentration of ~55 /xM (Hansell et al, 1997b). This water is transported to the east, shoaUng to near surface in the central and eastern Equatorial Pacific, bringing with it low-DOC water. The return flow of surface water to the west undergoes an increase in DOC (to ~65 /xM) due to biological activity. The highest DOC concentrations in Fig. 3 (>70 /xM C; largely west of 165° W in the surface 100-120 m) are associated with the Western
DOC in the Global Ocean Carbon Cycle Pacific Warm Pool (Hansell et al, 1997b; Hansell and Feely, 2000). The front separating the DOC-enriched Warm Pool to the west and the recently upwelled water to the east varies with the ENSO state, being found further to the west during La Nina conditions (Dunne et al, 2000). The impact of upwelling on equatorial DOC concentrations exists at coastal upwelling sites as well. Along the coast of Oman in the Arabian Sea, strong upwelling occurs during the Southwest Monsoon. Low surface DOC concentrations are present in coastal water during such periods although productivity can be quite high (Hansell and Peltzer, 1998). UpwelUng along the northwest margin of the African continent similarly forces a shoaling of the DOC isolines (Postma and Rommets, 1979). Similarly, Doval et al (1997) reported a decrease in subsurface DOC in northwest Spain due to upwelling. Ocean margins influenced more by riverine inputs than by upwelling tend to show increases in DOC concentrations. Rivers introduce water with high DOC concentrations (see Cauwet, Chapter 12), thus raising concentrations along the coast. One example is in the Chesapeake Bay outfall, where DOC concentrations increase from 70 ^xM in the surface Sargasso Sea to >200 JJM in the Chesapeake Bay mouth (Bates and Hansell, 1999). Guo et al (1994) reported onshore DOC concentrations of 131 /xM off Galveston, Texas, and moderate concentrations of 83 /xM offshore in the Gulf of Mexico. Property/property plots of DOC and salinity show the conservative nature of riverine DOC as it mixes with oceanic water. In general, the strength and direction of concentration gradients between the surface open ocean and the coastal ocean depend on the degree of upwelling of low-DOC water from below or invasion of DOC-enriched freshwater from the continent. Comparing two zonal sections in the North Atiantic provides further evidence for the control physical properties of the water column play on DOC distributions (Fig. 4 [see color plate]). A Section at 24°N shows strongly enhanced DOC concentrations in the upper 200 m (up to 80 /xM C), reflecting the strong stratification present in the subtropical gyre. In a more northerly Section, surface DOC is lower (>60 jjM C) and the concentration contours are pushed deeper into the water column. Note, for example, the 55 /xM DOC contour at 200-300 m along 24°N, but at 200-600 m on the northern Section. This change in contour depths reflects the weaker stratification at higher latitudes, and subsequent downward mixing of the surface produced DOC. 2. Deep-Ocean Distributions Reports on the distribution and variability of DOC in the deep ocean have been conflicting. Measurements from the 1960s (discussed below) suggested strong, horizontal gradients in DOC. More recently, Druffel et al (1992) reported a modest 5 /xM concentration difference between the deep waters near Bermuda and Hawaii. Martin and Fitzwater (1992), in contrast, reported the complete absence of DOC
689
690
Dennis A. Hansell
gradients in the deep ocean. Hansell and Carlson (1998a), in an effort to narrow the uncertainty, surveyed representative sites in the deep ocean (Fig. 1). They found a 29% decrease in DOC concentration from the northern North Atlantic (48 /xM in the Greenland Sea) to the northern North Pacific (34 /JM in the Gulf of Alaska) (Fig. 5, top). The gradient reflects the export of DOC-enriched (formerly subtropical) water during North Atlantic deep water (NADW) formation (Fig. 5, bottom) and the decrease in DOC (by mineralization and mixing) along the path of deep ocean circulation away from the North Atlantic. The formation of Antarctic bottom water (AABW) does not introduce additional DOC to the deep ocean (see Section IV), so the concentrations remain low near those sites. The small increase in DOC concentrations from the Southern Ocean into the deep South Pacific and Indian oceans is enigmatic and the source unidentified (Hansell and Carlson, 1998a). Possible causes include inputs from marginal seas (Red Sea, Arabian Sea, and Bay of Bengal for the Indian Ocean), inputs due to dissolution of sinking biogenic particles, non-steady-state conditions in the deep-ocean concentration gradients of DOC, and, of course, unidentified processes. The highest deep-water DOC concentrations may be those in the deep Eurasian Basin of the Arctic Ocean (see Anderson, Chapter 14), where concentrations >50 /j^M C have been reported (Anderson et aL, 1994). The sources of this material must be terrestrial runoff and Arctic continental shelf produced DOC (Opsahl et aL, 1999; Wheeler etai, 1997). It is interesting to speculate as to the mechanisms responsible for DOC concentration decreases in the deep ocean. Certainly microbial mineralization and mixing contribute, but, based on our present knowledge, these mechanisms appear to be inadequate. The DOC concentration decreases by 14 /xM over the length of the deep limb of the "global conveyor belt," but would the marine microbes we are most familiar with today be satisified with such meager rations over the half millennium required for transport over that distance? The apparent rate of oxidation (^30 nM year"^ over ^500 years), and the amount of energy derived over these several centuries, is miniscule. Perhaps the poorly understood Archaea, now known to inhabit the deep ocean, are designed to catabolize recalcitrant DOC at such low rates. Perhaps microbes play only a secondary role, and DOC is removed primarily by coagulation and formation of sinking particles, or it is stripped from the water column by particles passing through the water colunm. The true mechanisms for DOC loss need to be resolved. 3. Relation to Productivity Given the surface DOC distributions described here (Fig. 5), of low DOC near sites of upwelling or deep mixing and high values in stratified water, a general observation can be made: upper ocean DOC concentrations are relatively high in oligotrophic waters where regenerated production dominates, and low in systems
691
DOC in the Global Ocean Carbon Cycle 50
Pacific/Indian Sector
Indian
32' 80°N
40"N
80°S
40"S
40°S
40°N
80"N
Latitude
Atlantic Sector
Pacific/Indian Sectors
av^ace NPOH^//
AABW
North
Circumpolar water
South
CDW North
Figure 5 (Top) Distribution of DOC in the deep-ocean. The x-axis is viewed in the context of the deepocean circulation, with fonnation in the North Atlantic, circulation around the Southern Ocean, and flow northward into the Indian and Pacific oceans. Station locations in Fig. 1. (Bottom) The general patterns of ocean circulation driving the deep ocean DOC signal. DOC-enriched surface water is introduced to the deep ocean in the North Atlantic. This water moves south as North Atlantic deep water (NADW), to the circumpolar waters of the Southern Hemisphere. DOC-impoverished Circumpolar deep water (CDW) flows north into the Pacific and Indian oceans. Deep return flow to the North Atlantic is via Antarctic bottom water (AABW) and to Antarctica via North Pacific (NPDW) and Indian Ocean deep waters (lODW).
692
Dennis A. Hansell
where new nutrients are introduced to the surface. Such a meridional gradient has been reported for the Equatorial Pacific (Tanoue, 1993), the South Pacific (Hansell and Waterhouse, 1997; Doval and Hansell, 2000) and the North Atlantic (Duursma, 1962; Kahler and Koeve, 2001) and is evident in Fig. 2. The regenerated vs new production nature of these systems is a reflection of the conmiunity compositions within them. The mechanisms by which conununity composition controls DOC concentrations are not understood (see Carlson, Chapter 4). DOC concentrations are also controlled by the vertical stability of the water colunm. The highest DOC concentrations in the open ocean are normally found where stratification of the water column is highest (Hansell and Waterhouse, 1997; Hansell and Feely, 2000). This finding suggests that stability facilitates the retention of DOC in the upper ocean. The lowest DOC concentrations, to the contrary, exist where DOC-depleted subsurface water is introduced to the surface, either by vertical mixing or upwelling. These high nutrient sites can experience large but brief seasonal increases in DOC concentrations, however (see below). Because of the role of ocean stratification in controlling DOC concentrations, a positive correlation between DOC concentrations and primary productivity (an oft predicted relationship) is absent in much of the oligotrophic, open ocean. Menzel and Ryther (1970) reported the absence of this correlation early and evidence for the generality will be given using data from the Sargasso Sea later in the chapter (Section II.B.2). In fact, in the highly stratified portions of the open ocean, DOC broadly correlates positively with temperature (Hansell and Waterhouse, 1997; Doval and Hansell, 2000), another sign of the importance of physical control on concentrations. At higher latitudes, however, where DOC concentrations are depressed during the winter, elevated DOC values indeed follow springtime elevation of primary productivity (Borsheim and Myklestad, 1997; Chen et al, 1996; Carlson et al, 2000). This positive relationship between primary production and DOC was reported early by Duursma (1963) and has been discussed elsewhere (Williams, 1995). In high-latitude systems, increased water colunm stability favors both phytoplankton growth and DOC accumulation in the upper ocean. The data indicate that low-latitude, highly stratified environments behave very differently than high-latitude environments in terms of the coupling between DOC dynamics and primary production. So, while their observations are in apparent conflict, both Menzel and Ryther (1970) and Duursma (1963) were correct about the relationship between DOC and productivity; but they were correct specifically for the hydrographic systems they were evaluating. 4. Historical Data With the onset of discussions surrounding the use of the high-temperature combustion (HTC) systems for DOC analysis (Sugimura and Suzuki, 1988), much attention has been paid to whether or not the earlier data are accurate and, therefore,
DOC in the Global Ocean Carbon Cycle of value (Sharp, 1997). A comparison of what we find in the ocean today with that reported in earlier decades shows some older data and findings to have serious flaws. A comparison of historical and recent data from the surface ocean cannot be easily made because of the wide natural variability in those waters (see below). The most useful comparisons between historical and recent data are made in the intermediate and deep ocean, where significant changes in concentration over a few decades (the sampling interval) are unlikely. Menzel (1964) reported DOC concentrations in the intermediate depths (400800 m) of the Arabian Sea and western Indian Ocean to range from 0.4 to 1.6 mg/L (30 to 130 /xM DOC). This wide range is not reproducible anywhere in the intermediate or deep ocean using modem techniques, nor was it evident during the US Joint Global Ocean Flux (US JGOFS) program in the Arabian Sea during 1995 (Hansen and Peltzer, 1998). Menzel and Ryther (1970) also reported a very unlikely DOC concentration doubling at all depths > 1000 m between the waters northeast and southeast of South America. Romankevich and Ljutsarev (1990), reviewing investigations conducted by the Soviet Union, reported DOC off Peru at 500-1000 m to be an unlikely 1 mg/L (^83 /xM). Soviet measurements in the deep Bay of Bengal exceeded 1 mg/L as well. These latter DOC concentrations are probably high by a factor of two. Williams et al (1980) reported DOC concentrations in the central North Pacific a few meters off bottom (5650 m) that were elevated by twofold relative to the values at 2000-5000 m. Such a strong DOC gradient, indicative of sedimentary input of DOC to the bottom layer, has not been confirmed using modem techniques and extensive near-bottom surveys. Recent data, using modem HTC techniques, should be viewed with caution as well. Dileep Kumar et al. (1990) reported a strong DOC concentration gradient from the central Arabian Sea to the westem Indian Ocean, increasing from 100 to 300 /zM at 2500 m. The high DOC concentrations and wide range reported are unlikely to be accurate representations of that system.
B. TEMPORAL VARIABILITY The temporal variability of DOC concentrations in the surface ocean has been noted since the earliest days of the measurement. Duursma (1963) reported a twofold increase in DOC concentrations in the North Sea, from winter lows of 0.8 mg/L (~66 jiM) to spring and early summer highs of 1.8 mg/L (~150 /xM). The increase in DOC concentrations started some weeks after the spring phytoplankton bloom. Holmes et al. (1967) reported large spikes in DOC concentrations, from a baseline of 1 mg/L up to 4-5 mg/L (330-415 /xM), during several red water dinoflagellate blooms off La JoUa Bay, Califomia. Here, too, the DOC peaks followed the decline of the blooms. Williams (1995) evaluated the seasonal accumulation of DOC using data from Parsons et al. (1970), Banoub and WiUiams (1973),
693
694
Dennis A. Hansell
and Duursma (1963), suggesting that the accumulation of C-rich dissolved organic matter resulted from nitrogen limitation. The possible role of nutrient depletion in the generation of DOC is discussed below. (See Carlson, Chapter 4, for a more complete listing of publications reporting temporal variability of DOC.) 1. High Latitudes Strong seasonal increases in DOC concentrations associated with phytoplankton blooms appear to be characteristic of systems that receive high input of new nutrients over the winter periods. The waters of the Ross Sea polynya, for example, undergo deep mixing over the winter such that nitrate concentrations exceed 30 fiM prior to the spring bloom (Bates et ai, 1998). DOC concentrations increase in the surface layer from winter lows of 42 /xM to summer highs of 65-70 /xM (Carlson etai, 1998). High southern latitude systems can experience large increases in DOC concentrations (tens of micromolar C), with the wintertime baseline concentration as low as the much deeper waters (Carlson et aL, 2000; Wiebinga and de Baar, 1998; Kahler et aL, 1997). The Ross Sea undergoes DOC concentration increases of 15-30 /xM where the Phaeocystis and diatom blooms are particularly strong (Carlson etal, 2000). Where the blooms are small because of various controls on plant growth (deep mixing, iron limitation, etc.), the DOC concentrations remain low (e.g., over the Ross Sea shelf break, with a gain of <5 JJLM relative to winter) (Sweeney et aL, 2000). The North Sea (Duursma, 1963), as well as other highlatitude/strong-bloom regions (e.g., the Norwegian Sea, Borsheim and Myklestad, 1997), likely behaves similarly to the Ross Sea. In Trondheimsfjord, Borsheim et aL (1999) reported a >2x increase in DOC concentrations during the summer. It is apparent, though, that the winter lows of DOC concentrations in the high northern latitudes are not as low as the local deep-water values (in contrast to the conditions found in the Southern Ocean). This finding holds true along 20°E in the North Atlantic, where Kortzinger et aL (2001) reported the winter low DOC to be 53 fiM C, well above the deep-ocean values in the region. The more physically stratified nature of the northern systems prevents full water column overturn and homogenization of the DOC each winter. 2. Mid-latitudes The more oligotrophic, mid-latitude zones of the ocean do not show the same seasonality (in either strength or direction) as the high latitudes or other nutrientrich areas. In the Sargasso Sea, where convective overturn during the winter introduces small amounts of new nutrients to the euphotic zone and phytoplankton blooms follow (Michaels and Knap, 1996), the seasonality of DOC in the surface ocean contrasts that found at high latitudes (Carlson et aL, 1998; Hansell and Carlson, 2001a). Overturn of the water column coincides with the spring bloom
DOC in the Global Ocean Carbon Cycle there because adequate light is present at these mid latitudes. The effect is to mix low DOC subsurface water upward, thereby reducing the DOC concentrations during the periods of highest primary productivity. Once stratification reasserts itself with warming of the surface ocean, and the bloom terminates, DOC concentrations rebuild to normal summer levels (Fig. 6 [see color plate]). The concentration change from the annual low to the annual high is only 3-6 /xM, a very small range compared to high-latitude systems. The lowest winter concentrations remain well above the deep-water values. While the DOC concentrations in the Sargasso Sea are lowest during the winter overturn/spring bloom period, the same cannot be said for the integrated DOC stocks. Relatively deep convective overturn maintains the low surface DOC concentrations but the bloom still supports the net production of as much as 1-1.5 mol m"^ of DOC over the upper 250 m (Fig. 7; Carlson et al, 1994; Hansen and Carlson, 2001a [see color plate]). This increase in DOC stock is as large as that seen in the much more productive Ross Sea (Carlson et al, 2000). DOC and bloom dynamics in the Arabian Sea during the NE Monsoon are similar to that in the Sargasso Sea. Convective overturn in the Arabian Sea, forced by cool dry winds off the Tibetan plateau, mixes moderate amounts of nutrient into the euphotic zone. There, too, DOC concentration changes are not large during the bloom, but the increase in DOC stock can be 1.5-2 mol C m~^ (Hansell and Peltzer, 1998). It is interesting that while the seasonal range for DOC in the western Sargasso Sea (at ~31°N) is only 3-6 luM, the seasonal range at the same latitude in the eastern North Atlantic can reach 10-20 /xM (Kortzinger et al, 2001). The western North Atlantic is generally warmer and more stratified than in the east, suggesting differing community composition and productivity between the sites. This follows from the gyre circulation patterns: the northward flow of water in the west, from the warm equatorial region to higher latitudes, lends itself to high vertical stability and highly oligotrophic conditions; the southern flow in the east, carrying cooler water from the north, would lend itself to less stability and less oligotrophic conditions. It may be that the more stable system in the west experiences less primary productivity and net DOC production than the system in the east. Physical characteristics of the systems and the biological regimes they support are centrally important in controlling DOC dynamics. 3. Low Latitudes Low-latitude systems that do not undergo winter freshening of the surface layer do not show seasonality in DOC concentrations. The waters at Station ALOHA, north of Hawaii at 23°N, represent such a system. There, variability in DOC occurs at interannual time scales, but there is no recurring trend with seasons (Fig. 6). Church et al (2001) reported a net accumulation from 1993 to 1999 of a DOM
695
696
Dennis A. Hansell
pool that was enriched in C and N, relative to P. These long-term changes may be a manifestation of the broad, ecosystem-wide shift from N to P limitation described by Karl (1999). No such shifts have been noted in the Sargasso Sea, hence the near constancy in summer time DOC highs from year to year. Note that the surface DOC concentrations at ALOHA are much higher than the highs at BATS (Fig. 6) and higher than any values found along 24°N in the North Atlantic (Fig. 4). Why this difference exists between these similarly low latitude zones of the North Atlantic and North Pacific is unknown. Community composition may be key, but an evaluation has not been conducted. 4. Deep Ocean Whether or not there is measurable temporal variability of DOC in the deep ocean remains debatable. Hansell and Carlson (2001a) did not resolve DOC variability in the deep Sargasso Sea over 6 years of time series measurements. Similarly, Hansell and Peltzer (1998) found no variability in the deep Arabian Sea over a single year, even through periods of very high sinking particle flux. Bauer et al (1998), in contrast, reported significant (8 /xM) long-term (2-year) changes in DOC in the deep eastern North Pacific. They tied these variations to natural variability ("patchiness") and exchanges with sinking POC. Why there may be variability at this site and not at the others studied needs to be resolved. 5. Short-Term Biological Events Further variabiUty in DOC concentrations can be expected to occur with biological "events." Examples are blooms of red tide organisms described by Holmes et al (1967) and of diazotrophs. Onset of enhanced nitrogen fixation rates in openocean systems can increase DOM stocks considerably. Karl et al. (1997) reported organic nitrogen concentration increases of several micromolar which should coincide with several tens of micromolar increase in DOC. A case in point is the western tropical South Pacific, where relatively high DOC is present under the zone of the atmospheric South Pacific Convergence Zone. Hansell and Feely (2000) suggested that the excess precipitation in this system increased vertical stability, thereby favoring nitrogen fixers and in turn increasing concentrations of DON and DOC. Near the continental margins, DOC concentrations will vary with the strength of DOC-enriched riverine inputs or coastal upwelling, both of which vary seasonally (Cauwet, Chapter 12; Hansell and Peltzer, 1998). High riverine input may result in high-DOC concentrations; strong upwelling reduces the DOC concentrations. Zones of equatorial upwelling similarly exhibit the lowest DOC concentrations during strong upwelling (e.g., La Nina), and the highest values during reduced upwelling (e.g.. El Nino; Peltzer and Hayward, 1996). In this way, physical stability plays a major role in controlling DOC concentrations both along the margins, in
DOC in the Global Ocean Carbon Cycle the open ocean and in equatorial upwelling systems (Carlson and Ducklow, 1995; Hansen and Waterhouse, 1997; Tanoue, 1993). 6. Summary Our present understanding of seasonal variability in DOC can be summarized here: At high latitudes, where spring blooms are intense, we expect to see large DOC concentration changes. Because the winter DOC concentrations are low in these high-latitude systems, the highest concentrations during sunmier may be no higher than the summer highs in the low-latitude gyre systems, but the concentration change between seasons may be large. However, the large increases in concentration do not necessarily translate into large accumulations of DOC stock (vertically integrated loads of DOC) because of the normally shallow euphotic zones in these highly productive systems (high concentrations but over little depth). In mid-latitude open-ocean regions, such as the Sargasso and Arabian Seas, where convective overturn introduces moderate nutrient loads, DOC concentration ranges between seasons can be relatively small, though the change in integrated stocks can be a relatively large signal (comparable to the change in DOC stock in the Ross Sea). The overturning water column mixes the DOC too deeply for a strong surface accumulation to occur, as found in blooms occurring in more stratified systems but the small concentration change over large depths results in significant increase in stock. At mid-latitude coastal sites with significant winter recharge of surface nutrients, DOC seasonality will be strong. Upwelling reduces the DOC concentrations while high riverine inputs increase them. At low-latitude sites where spring blooms are absent, no seasonality is evident; but, as with all ocean regions, interannual variability exists. The DOC that accumulates each year at mid-latitudes has a lifetime exceeding the season in which it was produced, so it can be transported elsewhere with surface currents, or be available for export during the subsequent winter overturn events. At higher latitudes, the seasonally produced DOC is seen to have a lifetime shorter than that of the season of production; thus it undergoes net consumption by microbes once primary production is reduced with the onset of Fall conditions. This material is not as available for export (see Section IV.A).
III. NET COMMUNITY PRODUCTION OF DOC DOC is produced on a daily basis as part of the primary and secondary production systems in the surface ocean. Most of the DOC released is mineralized on the time scale of hours to days. For DOC to play a role in the ocean carbon cycle beyond serving as substrate for surface ocean microbes, it must act as a reservoir for carbon on the time scales of ocean circulation. This it does, given that the
697
698
Dennis A. Hansell
ET-^ Deep and Bottom water m High density mode water • I Low density and Subtropical mode water —— Subtropical gyre circulation
Figure 8 Distribution of sites of water column overturn (from Talley, 1999), general patterns of surface circulation in the subtropical gyres, and proposed distribution of exportable DOC. Overlap in the distribution of exportable DOC (background field of white) and sites of ocean ventilation (sites colored by gray scale) favors DOC export; a lack of overlap precludes export. The waters of the Southern Ocean (slanted stripes) are without exportable DOC present, so where these waters overlap sites of ventilation, little export is expected.
production and accumulation of DOC in the surface ocean has been demonstrated (Figs. 2-8). The rates of, and controls on, the net production of DOC, topics not well understood at this time, are the focus of this section. Because so few DOC data exist, particularly from ocean systems for which accumulation has been evaluated, it is useful to normalize estimates of DOC accumulation to a more broadly available and easily measured variable. DOC accumulation as a function of net community production (NCP) has proven useful in this way (Hansell and Carlson, 1998b; Kortzinger et aL, 2001). NCP occurs when autotrophic production exceeds heterotrophic consumption, such as during a spring bloom. It is a process that largely results in the export of carbon and new nitrogen from the euphotic zone as sinking biogenic particles and in this way is analogous to new production (Dugdale and Goering, 1967). If DOC accumulates, then DOC too is a sink for NCP. NCP is estimated most directly by measuring the biological drawdown of the reactants (dissolved inorganic carbon and/or nitrate) or as the flux of the products (i.e., accumulation of DOC, suspended POC, export of sinking biogenic particles.
DOC in the Global Ocean Carbon Cycle and contributions by migrating zooplankton). The Section above on temporal variability of DOC sheds light on the net production of DOC. As a rule, oceanic regions showing seasonality of DOC concentrations are experiencing some transfer of NCP into the DOC pool.
A. EVIDENCE FOR NET PRODUCTION OF DOC Seasonal increases of DOC stocks in the Ross Sea indicate that 8-20% of NCP in the polynya system accumulates each growing season as DOC (Bates et al, 1998; Carlson et al, 2000; Hansell and Carison, 1998b; Sweeney et aU 2000). The balance of NCP is lost to the deep ocean as sinking biogenic particles, mostly Phaeocystis and diatoms. Annual rates of NCP in the Ross Sea polynya are 6-14 mol C m"^, so net DOC production of 1.2-2 mol C m"^ occurs over the growing season (Bates et al, 1998; Carlson et al, 2000; Sweeney et al, 2000). The net production of DOC in the Ross Sea is about that in the Sargasso Sea (1-2 mol C m~^; see above), but the Sargasso Sea has a much lower annual rate of net commiunity production. The rate of DOC production in the Ross Sea, normahzed to NCP, is similar to that found in the Equatorial Pacific. Estimates of net DOC production as a percentage of NCP in the central Equatorial Pacific range from 6 to 40%, with most estimates near the 20% level (Archer et al, 1997; Hansell et al, 1997a,b; Zhang and Quay, 1997). These values from the Equatorial Pacific are similar to the Equatorial Atlantic (20%; Thomas et al, 1995), but significantly lower than prior estimates in the Equatorial Pacific by Murray et al (1994), Feely et al (1995), and Peltzer and Hay ward (1996). Those latter authors estimated net DOC production closer to 75% of NCP, but those findings have been challenged (Hansell et al, 1997b; Zhang and Quay, 1997). Noji et al (1999) suggested that more than half of NCP in the Greenland Sea accumulated as DOC, high compared to findings from other nutrient-rich sites. Alvarez-Salgado et al (2001) reported that 20% of net ecosystem production accumulated as DOC in a coastal upwelling environment along the Iberian margin in the North Atlantic. This rate is very similar to that reported for the Equatorial Pacific and the Ross Sea. Net DOC production in the Ross Sea, the Equatorial Pacific and the Iberian margin takes place when the conditions are right for net autotrophy. In the Ross Sea, this occurs when vertical stability and light are available, while in the Equatorial Pacific and the coast of Spain light becomes available following upwelling. At these three sites, vertical stability is relatively strong during the periods of net production. The Sargasso Sea contrasts those systems. Light is generally available year round but nutrients are not, so a reduction in vertical stability (convective overturn of the water column and entrainment of nutrients) is required for net autotrophy. A representative year (July 1994 to July 1995) for DOC in the Sargasso
699
700
Dennis A. Hansell
Sea is useful for demonstrating net DOC production (Fig. 7). Winter overturn and mixing of the water column was both the cause of concentration reductions and the trigger for net DOC production each year following nutrient entrainment and subsequent new production (Carlson et aL, 1994; Hansell and Carlson, 2001a). The net production of DOC at the BATS site varies interannually as a function of the maximum in the winter mixed layer depth. The greater the vertical mixing (and nutrient entrainment) in the Sargasso Sea, the greater the net production of DOC (Hansell and Carlson, 2001a). In winter 1995 (Fig. 7), the DOC stock increased by 1.4 mol C m~^in response to maximum mixing depths of 260 m (note the net production of DOC in the upper 250 m of the water colunm; Fig. 7b). In subsequent years experiencing shallower maxima in mixed layer depth (<220 m), DOC stocks increased <0.7 mol C m~^ during the overturn event (Hansell and Carlson, 2001a). For the 1995 spring bloom, net DOC production was estimated to be 59-70% of the NCP (Hansell and Carlson, 1998b). This high rate does not hold for the entire year though. Much of the DOC produced during overturn and then mixed into the subsurface layers is remineralized upon restratification (note the rise and fall in integrated stocks of DOC at 100-250 m; Fig. 7b). Indeed, NCP continues throughout the summer and fall periods (Steinberg et ai, 2001), while DOC concentrations remain relatively static (no net production of DOC). As such, net DOC production reduces to ^ 8 % of NCP on an annual basis. Left unanswered is why so much of the NCP during the bloom accumulates as DOC, compared to the much lower NCP-normalized accumulation of DOC in the Ross Sea and the Equatorial Atlantic and Pacific. Community compositions and their responses to physical dynamics vary between these systems, and the answer likely lies in those variables (Carlson etai, 1998; Hansell and Carlson, 1998b). B. REGIONAL AND GLOBAL ESTIMATES FOR N E T PRODUCTION OF
DOC
As mentioned in the introduction to this Section, the value of normalizing the DOC accumulation rates to NCP is that NCP is a variable for which there is high data density. We can use the findings developed here on net DOC production, along with existing estimates for new (nitrate-based) production in various ocean regions (Chavez and Toggweiler, 1995), to estimate annual rates of DOC accumulation (Table I). The ratio of net DOC production to new production for each ocean region was assigned given what is known for the few systems (reviewed above) in which data exist. As would be expected, the regions of highest new production, such as tropical and coastal upwelling areas, contribute most to net DOC production globally. The weakest sites for the entrainment of nitrate to the
DOC in the Global Ocean Carbon Cycle
701
Table I Estimates of Annual Net DOC Production Based on Regional Estimates of New Production (Chavez and Toggweiler, 1995) Region Tropical open ocean Upwelling Turbulent Mixing Southern ocean Subarctic gyres Coastal upwelling Monsoonal Subtropical gyre Continental margins Western boundary currents Estuarine influenced shelves Total
Net DOC production (pg C year~^ )
New production (pg C year"^)
ADOCiNP
1.5 (21%) 0.7 (9.5%) 1.1 (15.5%) 0.3 (4%) 0.8(11%) 0.4 (5.5%) 0.5 (7%)
0.2 0.1 0.12 0.15 0.2 0.2 0.1
0.3 (24.6%) 0.07 (5.7%) 0.13(10.8%) 0.04(3.7%) 0.16(13.1%) 0.08 (6.6%) 0.05 (4.1%)
0.7(9.5%) 1.2(17%) 7.2
0.2 0.2
0.14(11.5%) 0.24 (19.7%) 1.2 (17% of total NP)
Note. Values in parentheses represent percentages of the global estimate. Adapted with permission from Hansell and Carlson (1998b).
surface (such as the subtropical gyres) are the weakest in DOC net production. Net DOC production, based on this analysis, is the sink for about 17% of global new (nitrate based) production each year. A recent estimate for global new production is ^1 Pg C year-i (Chavez and Toggweiler, 1995; Fung et al, 2000; Lee, 2001), so net DOC production in the global ocean could be ^ 1.2 Pg C year~^. This estimate is for nitrate-based new production, thus excluding other forms of new production such as nitrogen fixation. We do not at present know the fraction of carbon fixed by active diazotrophs accumulating as DOC.
C. NUTRIENT DEPLETION AND NET PRODUCTION OF DOC One apparent axiom of net DOC production has been that nutrient depletion drives high rates of DOC production. This perception was based on numerous batch phytoplankton culture experiments where nutrients were allowed to run to very low values. When nutrients were present, and the plants were in the exponential growth phase, DOC release was very small; when nutrients were depleted and the plants went into stationary phase, carbon fixation exceeded the N stocks available to support biomass growth, so C-rich DOC was released (e.g., Goldman et al, 1992). Because of these results, it has been assumed that nutrient depletion forces high levels of DOC production everywhere in the ocean. It would follow
702
Dennis A. Hansell
that high levels of DOC in oligotrophic systems were expected because of nutrient limitation. Such arguments can be tested for relatively oligotrophic systems using the time-series data from the Sargasso Sea (Fig. 6). As outlined above, net DOC production took place largely while rates of primary production were highest (the spring bloom period). During the summers, when nutrients were most depleted, there was no further accumulation of DOC; nor was there a major enhancement in bacterial respiration rates relative to primary productivity to indicate high release of labile DOC (Carlson et ai, 1996). The data from the Sargasso Sea indicate that nutrient depletion alone does not drive high rates of DOC production. Indeed, Parsons et al. (1970) reported DOC accumulation during the sunamer in the Strait of Georgia, but without depletion of the macronutrients (nitrate went as low as 7 /xM only). Where the culture experiments fail to reflect nature is in their inability to support reasonable changes in community composition. In nature, new assemblages of autotrophs will develop given changes in the nutrient forcing, and the stress of nutrient depletion on the sunmier population will be dissimilar to that felt by the spring bloom population. It is apparent that nutrient depletion alone is not adequate to force significant net production of DOC in the open ocean.
IV. CONTRIBUTION OF DOC TO THE BIOLOGICAL PUMP A. EVIDENCE FOR DOC EXPORT DOC plays perhaps its most important role in the biogeochemistry of the ocean carbon cycle when it contributes to the biological pump (Copin-Montegut and Avril, 1993; Carlson et ai, 1994; Ducklow et ai, 1995). The export of sinking biogenic particles has long been understood to drive respiration in the ocean interior and to help maintain the ocean's strong vertical gradient of inorganic carbon. The contribution of DOC to the biological pump has been debated for several decades, and is only now being resolved. Menzel (1970) stated that deep-ocean changes in oxygen and nutrients could not be attributed to the long-term in situ decomposition of dissolved organic matter. He acknowledged that some published data suggested that 15% of the total oxygen consumption was due to oxidation of DOC, but he thought this value to be unreaUstic. Craig (1971), in a direct counter to these findings, suggested that one-third of the oxygen consumption in recently formed North Atlantic Deep Water could be due to DOC oxidation. While he did not have faith in the accuracy of the DOC data available at the time, and he used calculations that he described as "brute force," he was satisfied that deep DOC oxidation was substantive. Ogura (1970) took a refined approach to the question. He evaluated property/property plots of DOC and apparent oxygen utilization (AOU) along
DOC in the Global Ocean Carbon Cycle isopycnal surfaces in the western North Pacific, reporting that one-third of the AOU was due to DOC consumption. He noted that the contribution was restricted to the upper ocean (<500 m); DOC oxidation was not resolvable in the deep waters. Kahler and Koeve (2001) argued that DOC is unlikely to be substantial in the export of biogenic material to the deep sea; otherwise the deep-ocean C:N ratios should deviate from Redfield stoichiometry as required by the input of C-enriched dissolved organic matter. The export of DOC in the ocean is a consequence of its accumulation in the surface ocean (Fig. 2), redistribution with the wind-driven circulation, and eventual transport to depth with overturning (thermohaline) circulation at high latitudes and subduction in the subtropical gyres. Thermohaline circulation refers to the vertical movement of water that takes place when its density increases by a significant change of temperature or salinity. The most important sites for ventilation of the intermediate and deep ocean are located in the North Atlantic (Labrador Sea Intermediate Water and North Atlantic Deep Water formation sites), the North Pacific (site of North Pacific Intermediate Water formation), the area around Drake Passage (Antarctic Intermediate Water formation), and around Antarctica (Antarctic Bottom Water formation) (Fig. 8). Water mass formation at these sites occurs primarily during winter. Subduction occurs year round in subtropical gyres due to Ekman pumping of the surface layer. The upper thermocline ventilates by this process. Because the mixed layer density and depth reach their local maxima at late winter, the water subducted into the permanent thermocline has properties that are strongly biased toward the late winter values (Stommel, 1979). As a rule, DOC export with ventilation of the ocean occurs if there is a vertical DOC concentration gradient at the onset of overturn. In such a situation, DOC accumulated at the surface undergoes net transport downward. Where vertical DOC gradients are weak or absent, there is little net downward movement of DOC with overturn and subduction. 1. Export into the Upper Pycnocline The main thermoclines of the major ocean basins ventilate following buoyancy loss (via cooling) in DOC-enriched warm water previously transported poleward in the western boundary currents. Equatorward subduction beneath the less dense surface waters results in the export of DOC along isopycnal surfaces, a process evaluated by Ogura (1970), Doval and Hansell (2000), and Abell et at. (2000). Establishing the contribution of DOC to export is perhaps best examined by normalizing the gradients in DOC concentrations to the gradients in AOU. AOU derives from the mineralization of sinking biogenic particles and subducted DOC and so reflects the total oxidation of biogenic carbon along specified surfaces. A representative isopycnal surface along 170°W in the western South Pacific is that of (7^26-26.5 (Doval and Hansell, 2000). DOC concentrations along that isopycnal decreased away from the sea surface commensurate with an increase in AOU,
703
704
Dennis A. Hansell
such that oxidation of DOC drove ^40% of the AOU. Mineralization of sinking biogenic particles formed overhead in the euphotic zone drove the balance of the AOU. In the isopycnal surfaces lying between aelS.S and 27.0 along 170°W, 21-47% of the AOU was driven by DOC oxidation (Doval and Hansell, 2000). Like Ogura's (1970) findings in the North Pacific, DOC oxidation along 170°W was restricted to the upper 500 m; at greater depths, only biologically refractory DOC remained in the water colunm so sinking particles alone drove additional AOU development. Guo et al. (1994) similarly found that 20-30% of AOU may be due to DOC mineralization in the Gulf of Mexico. Subtropical mode water formation is a strong form of subduction ventilating the upper thermocline primarily during winter at mid-latitudes. Such a process is important to ventilation of the western North Atlantic in the north Sargasso Sea. The waters of the Sargasso Sea near Bermuda are strongly stratified through much of the year. During the fall and winter months, regular storm fronts deliver cold, dry winds that cool the surface water, causing a deepening of the mixed layer by convective overturn. Overturn of the water colunm, when deep enough (~250 m), ventilates Worthington's (1976) 18° Mode Water. DOC near Bermuda is present at its highest concentrations during the long warm (summer) periods (Fig. 7). At overturn, DOC concentrations and stocks at depths >100 m increase (Fig. 7), reflecting the downward mixing of the carbon; its disappearance with time at those depths reflects net microbial mineralization and removal by horizontal advection and mixing into mode water. During the overturn periods of 1992-1998, DOC export near Bermuda ranged from 0.4 to 1.4 mol C m"-^, rates representing 15 to 41% of the total biogenic carbon mineralized annually in the upper 400 m of the water colunm (Carlson et al, 1994; Hansell and Carlson, 2001a). 2. Export into the Deep Pycnocline Ventilation of the deep pycnocline occurs with the formation of intermediate water masses. At present few data exist for evaluating DOC export during this process. An exception is DOC export with North Pacific intermediate water (NPIW) formation (Hansell et al, 2002). NPIW forms by brine rejection in the Sea of Okhotsk north of Japan. It is delivered to intermediate depths (400-1000 m and bounded by cr6i26.6-27.4) of the subtropical North Pacific as a mixture of recently ventilated, relatively fresh subpolar (Oyashio) water and older, more saline subtropical (Kuroshio) water of the same density range. New NPIW is transported eastward from the site of subduction east of Japan as a low-salinity tongue, with eventual circulation into the intermediate layers of the subpolar and subtropical gyres, thus replenishing those systems (Talley, 1997). Data collected at 1000 m (^27.4 ae), between 10 and 45°N along 1527158°W (Fig. 1), demonstrate the export of DOC-enriched subpolar water and its impact on DOC distribution in the NPIW of the subtropical gyre (Fig. 9). The samples were from deep in the
705
DOC in the Global Ocean Carbon Cycle
1400
I8OOK
a
2000
^D
44
•
M
1
•••
—
35 N
30 N
25 N
r
1
1
••
20 N
Latitude
•
•
15N
T
T
—
••
• •
1,- -
0 0 40
•
0
1
40 N
1
30 N
25 N
Latitude
—
• <\
•
•
38
1
•
b
1
15N
Figure 9 Evidence for the export of DOC with formation of new North Pacific intermediate water, (a) Salinity contours in the upper 2000 m along 152°/158°W in the North Pacific, overlain (bold) by cr^ surfaces (26.6 and 27.4) indicating the upper and lower bounds of NPIW. Filled circles indicate sample locations for DOC (b) along the section. Note the freshening of the water column to the north, where the influence of new NPIW is strongest.
706
Dennis A. Hansell
NPIW, across a salinity gradient of high subpolar (fresh) water influence in the north and high subtropical (sahne) water influence in the south (Fig. 9a). Samples comprising subtropical water (south of 25°N and salinity >34.4) had a mean DOC concentration of 38.7 ±0.7 /xM (Fig. 9b). Waters to the north, with a greater subpolar contribution (salinity <34.4), had a mean of 44.7 ±0.6 /iM, for a 15% enrichment. The evidence for DOC export with NPIW formation is clear. 3. Export to the Deep Ocean Removal of carbon from exchange with the atmosphere is greatest with DOC export associated with deep- and bottom-water formation. Evidence for the export of DOC with NADW formation is the strong meridional gradient in deep-water DOC, along the proximal path of the deep western boundary current (Fig. 5; Hansell and Carlson, 1998a). Data from 75°N in the deep Greenland Sea are presently our best representation for concentrations of DOC in the source water for NADW formation (48 /xM). NADW overrides and entrains northward-flowing Antarctic bottom water (AABW), of a largely Weddell Sea source and with a DOC concentration of 41 /xM (Fig. 5). Mixing of these two source waters, along with microbial degradation, results in a DOC concentration gradient in the Atlantic Ocean. AABW forms in the cyclonic gyres that develop south of the Antarctic Circumpolar Current system, particularly the Weddell and Ross Sea gyres. The contribution of AABW formation to DOC export appears to be very small. The best studied of the high-latitude cyclonic gyre systems with regard to DOC distributions is the Ross Sea (Carlson et al, 1998, 2000). During the 1997 austral summer season in the Ross Sea, DOC concentrations in the surface 50 m reached >20 ^M above background in certain areas. By the time of winter overturn in fall, biological mineralization of the DOC reduced the concentrations to a mean of <5 /xM above background. The vertical export of this material, if still remnant at the time of complete overturn, would contribute to only 2% of the annual export of particulate plus dissolved organic carbon in the Ross Sea (Sweeney et al, 2000). Mineralization of the DOC prior to overturn prevented it from making a major contribution to export in this system. Evaluating horizontal gradients of DOC in the deep layers between the Ross Sea and the circumpolar deep water is a second test for DOC export with AABW formation. Because DOC exported with NADW formation demonstrated a strong gradient along the path of advection, so too should a gradient be present from the sites of AABW formation to the deep circumpolar waters if export is significant. The deep Ross Sea (data from 76°S in the Pacific sector of Fig. 5) has a DOC concentration (42 /xM) that is indistinguishable from the concentration in the circumpolar water to the north at 60°S (Fig. 5). The lack of a gradient suggests the absence of DOC export.
DOC in the Global Ocean Carbon Cycle
707
B. EXPORTABLE DOC We see from the analyses presented here that the export of DOC contributes in a significant way at some, but not all, sites of water mass formation. The water masses formed in the Northern Hemisphere, as well as those formed in the low- to mid-latitudes of the Southern Hemisphere, carry with them significant DOC loads for export. NADW, formed in the region of 60-70°N of the North Atlantic, exports DOC at ^32 Tg C year"^ and contributes to 40-50% of AOU in the deep northern North Atlantic (Hansell and Carlson, 1998b; Hansell et al, 2002). NPIW, formed at 40°N in the western North Pacific, exports ?^13 Tg C year"\ driving 30% of oxygen utilization near the site of formation (Ogura, 1970; Hansell et al, 2002). Overturn at 32°N in the Sargasso Sea exports DOC adequate to drive 15-40% of AOU in the underlying waters (Hansell and Carlson, 2001a). Similarly, ventilation of the South Pacific and Indian Ocean main thermoclines delivers DOC adequate to drive 20-45% of AOU (Doval and Hansell, 2000). In contrast, DOC export does not take place with AABW formation. Export with Antarctic intermediate water formation, a poorly resolved process, is likely low as well. In order for DOC export to occur its concentration must be elevated at the surface relative to the deeper water into which mixing takes place. This excess DOC at the surface must be present at the onset of overturn. Exportable DOC is the term used here to refer to DOC present at overturn that is in excess of the DOC concentrations at the depth to which vertical mixing takes place. It is the fraction of DOC whose lifetime exceeds the season of production and is therefore transportable to sites of water mass formation by surface currents. (So as to avoid confusion, some comments about the terminology used here and in the Introduction must be made. The exportable DOC introduced here is approximately the same fraction of DOC as the "semilabile" pool given in the Introduction. In this Section, though, the DOC pool is resolved into fractions that describe their export functionality. Terms such as "semilabile" DOC highhght the functional nature of DOC relative to microbial turnover of the various fractions.) Significant amounts of exportable DOC appear to be absent at high latitudes in the Southern Ocean. Sharp fronts in a variety of properties separate the Antarctic Circumpolar Current System (ACCS) from the DOC-enriched subtropical gyres, as evidenced by the abrupt shifts in DOC concentrations south of the gyres (Fig. 2). It is likely to be in the gyres that exportable DOC is stored. Certainly it is in these waters that this fraction is resident year round (Fig. 6). The primary hydrographic front separating the subtropical and subantarctic water masses in the Southern Ocean is the Subtropical Front (also known as the Subtropical Convergence). Where subtropical gyre waters serve as precursors for water mass formation, we expect DOC export to be important. Where high-density waters of the Southern Ocean are the primary source for water mass formation (such as for AABW formation), DOC export appears to be negligible. Kahler et al (1997), Wiebinga and
708
Dennis A. Hansell
de Baar (1998), and Carlson et al (1998, 2000) noted the apparent lack of DOC export in the Southern Ocean. Kahler et al (1997) highlighted the absence of the greater-than-seasonal fraction of DOC, a fraction they described as being of intermediate stability and one that is defined here as making up the exportable fraction of DOC. The distribution of the world ocean's primary sites for water mass formation (Talley, 1999), when overlain by the proposed global distribution of exportable DOC, indicates where DOC export is likely to be of consequence (Fig. 8). Excluded from this map are the zones of subduction in the subtropical gyres, which will force a relatively shallow (<500 m) export of DOC. Hansell and Carlson (2002 b) hypothesize that the primary sites of formation for exportable DOC are the subtropical gyres and the equatorial/coastal upwelling regions. It is in the gyres that exportable DOC is present year round, that this fraction has ample residence time in the euphotic zone to accumulate, and from which surface water is exported to higher latitudes as precursor for water mass formation. Further evidence for this hypothesis is that the exportable fraction is absent at high southern latitudes missing inputs of subtropical water (i.e., the Southern Ocean). Given this, high-latitude regions replenished by subtropical gyre waters via western boundary currents, such as the northern North Pacific and northern North Atlantic, will have exportable DOC present at the time of overturn. High-density water masses formed in high northern latitudes (NADW, NPIW, and likely Labrador Sea intermediate water), but fed by subtropical water, will export DOC. In contrast, DOC export in high-density water masses formed in the high-latitudes of the Southern Hemisphere will not contribute to the total export of biogenic carbon. The presence of the strong frontal systems in the ACCS prevents the DOC-enriched subtropical water from serving as source water during convection. The absence, as well, of locally produced exportable DOC at high southern latitudes precludes the process. Formation of Antarctic intermediate water (AAIW) should support a more variable export of DOC, dependent on the ratio of highlatitude and subtropical waters serving as precursor to this widely distributed water mass. Much more work is required to quantify DOC export with AAIW formation. Hansell and Carlson (2002 b) have estimated the global, annual export of DOC. Export with intermediate, deep, and bottom water was estimated at ^9.8x10^^ mol C year~\ or ^10% of the total export (oxygen utilization) to depths >500 m. The contribution of DOC to exported C oxidized at depths <500 m is elevated relative to the deep ocean. DOC contributes 15 to 41% of oxygen utilization at the BATS site in the Sargasso Sea (Hansell and Carlson, 2001a) and 25 to 45% in the main thermoclines of the South Pacific and Indian Oceans (Doval and Hansell, 2000). The global contribution of DOC to export must lie between the ~10% value of the deep ocean and the ~30% value of the main thermocline. The contribution of DOC to global export in the open ocean, then, must fall in the range of 20 =b 10%.
DOC in the Global Ocean Carbon Cycle
709
As referenced above, 7 Pg C year" ^ is the recent estimate for global new production, so total DOC export is 1.4 ± 0.7 Pg C year~^ DOC export to depths >500 m is only 0.13 Pg C year~\ so most of the global export of DOC takes place through the many paths of shallow subduction. This rate of export is about the same as the earlier estimate of global net production of DOC (1.2 Pg C year"^), suggesting that most of the annual net production of DOC is exported. The sites of production are spatially distinct from the sites of export; surface currents provide the essential connectivity.
V. RESEARCH PRIORITIES In this chapter, a holistic picture of DOC in the ocean carbon cycle was developed. DOC was evaluated in the context of its wide-scale distribution (i.e., spatial and temporal variability), net production, and export. Evaluating DOC in this way was possible through collection of high quality data at key ocean sites. Unfortunately, our data density remains far too small to generate a higher resolution picture of the system. Many of the details await discovery and, the conceptual models and hypotheses, critical testing. The major constraint on improving data density in the past has been that most of the DOC data generated by the international community were not referenced against common materials. Data from within a laboratory could be used to quantify, for example, wide-scale spatial variability, but data could not be combined from several laboratories for a similar analysis. The analytical differences between laboratories often exceeded the natural variability assessed, thus precluding such data compilations. The solution is increasing dedicated use of standardized reference materials for DOC analysis, now available to the ocean science community through funding by the U.S. National Science Foundation. One important issue not covered in this chapter is the oceanic transport of carbon as DOC with surface currents. We evaluated DOC production and export, but we did not physically connect the processes with an evaluation of transport. What is required is a strong analysis of DOC transport with the surface-ocean, wind-driven circulation. The data now in existence are inadequate for extensive analyses of transport. Some efforts can be made in the North Atlantic where zonal sections have been completed (Fig. 4). The distribution of DOC, with highest concentrations in the low to mid-latitude gyres, suggests that the subtropical gyres and low to mid-latitude upwelling systems are important sources of DOC for export at higher latitude sites of ocean ventilation. The strong western boundary currents carry the warm, saline subtropical water to higher latitudes, where cooling drives overturn of the water column. The subtropical waters are DOC-enriched, so the boundary currents must be important for the net transport of DOC as well. Little fieldwork or data analysis exists
710
Dennis A. Hansell
to understand or quantify this process for the various ocean basins. The assessment made here that much of the net DOC production occurs in coastal and equatorial upwelling sites in turn suggests that the gyres must serve as a reservoir for that DOC produced. From these ocean systems, the DOC is transported to the high latitudes for export. This model will be evaluated over the next decade. Changes in the circulation and net DOC production patterns with ocean climate change will modify the importance of the system described. One form of DOC export not considered in this chapter is vertical diffusive mixing into the water colunm. This process is likely important on a global scale but it has not been adequately evaluated. Loh and Bauer (2000), Emerson et al (1997), Guo et al (1995), Christian et al (1997), and Vidal et al (1999) provide recent calculations for downward diffusive flux of DOC. One major uncertainty with this method is the choice of the vertical diffusion coefficient, which may be uncertain by an order of magnitude. Also, it is difficult to test the vertical diffusive flux calculations independently. Confidence in rate estimates comes with independent evaluations. Clearly, this mechanism of export needs a good deal more analysis. Modeling efforts may prove the most useful test of the reported findings. Other issues requiring high-priority effort include understanding controls on the variability in surface DOC concentrations between the various major subtropical gyres. DOC in the subtropical North Pacific, for example, may be elevated relative to the North Atlantic (Fig. 6), but we do not know why. Will the size of the marine reservoir of organic carbon vary with climate-induced changes in the unknown controls on the reservoir? Further, we do not know the factors controlUng the NCP-normaUzed net production of DOC, such that the rate is high in the Sargasso Sea and lower in the areas with higher nutrient loads. In the deep ocean, we do not understand the processes causing the variability seen in Fig. 5. What are the mechanisms for removal of DOC along the entire path, and what are the mechanisms for introducing DOC in the low-latitude southern hemisphere? We also need to identify the truly important mechanisms of DOC production. Many times in publications a "laundry list" of mechanisms is presented (phytoplankton exudation, sloppy feeding, viral lysis, particle dissolution, etc.), but in truth we do not know which are significant and when. We know just as little about the processes involved in DOC decomposition. Progress is required here. While a great deal of field work is still needed, certainly we can make more significant headway by modeling DOC dynamics, particularly when performed in the context of ocean physics. Advances in understanding processes such as DOC export will be made whole when models are employed to expand our understanding of how the system works. Meanwhile, it is important to collect high-quality data and to interpret these in the context of the biological and physical conditions present at collection.
DOC in the Global Ocean Carbon Cycle
711
VL SUMMARY In this chapter an understanding for the role of DOC in the ocean carbon cycle was developed. Net DOC production is a function of net community production in the ocean, so where NCP is highest, net DOC production is high as well (Table I). The coastal and equatorial upwelling environments are particularly important in this regard. DOC produced at mid-latitudes, given sufficient lifetime, converges by surface currents into the subtropical gyres. The gyres, as the recipients of the DOC enriched waters and as sites of strong vertical stability, have the highest year-round concentrations and stocks of DOC in the global ocean. Subduction in the gyres carries DOC into the upper pycnocHne. The western boundary currents are important in the transport of the DOC from the gyres to higher latitudes, where DOC export occurs at sites of deep-water mass ventilation. Where the western boundary currents reach high latitudes, both intermediate and deep water formation carries exported DOC. Where western boundary currents are impeded from reaching high latitudes by frontal systems, such as in the Southern Ocean, DOC export is minimal. Annually, net DOC production represents about ~20% of net community production and ~20% of export production. Surface currents connect the sites of net production and export.
ACKNOWLEDGMENTS An understanding for the role DOC plays in the ocean carbon cycle is developed only with great support from many individuals and institutions. I thank the U.S. National Science Foundation (Ocean Chemistry) and the U.S. National Oceanographic and Atmospheric Administration (Office of Global Programs) for their continued support of this work. Numerous individuals have suffered many hours over hot instrumentation to generate the thousands of DOC values required for the analyses reported. Rachel Parsons did the lion's share of the DOC analyses; I owe her a great debt of gratitude. Also important are Dr. Wenhao Chen, Paula Hansell, Tye Waterhouse, Susan Becker, Amy Ritchie, and Bemie CuUen; each spent many hours in the laboratory and at sea. Finally, I thank Craig Carlson for being a good friend and collaborator.
REFERENCES Abell, J., Emerson, S., and Renaud, R (2000). Distributions of TOP, TON, and TOC in the North Pacific subtropical gyre: Implications for nutrient supply in the surface ocean and remineralization in the upper thermocline. J. Mar. Res. 58, 203-222. Alvarez-Salgado, X. A., Gago, J., Miguez, B. M., and Perez, F. F. (2001). Net ecosystem production of dissolved organic carbon in a coastal upwelling system: The Ria de Vio, Iberian margin of the North Atlantic. Limnol. Oceanogr. 46,135-147. Anderson, L. G. (2002). The Arctic Ocean. In "Biogeochemistry of Marine Dissolved Organic Matter" (D. A. Hansell and C. A. Carlson, Eds.), pp. 665-715. Academic Press, San Diego.
712
Dennis A. Hansell
Anderson, L. G., Olsson, K., and Skoog, A. (1994). Distribution of dissolved inorganic and organic carbon in the Eurasian Basin of the Arctic Ocean, In "The Polar Oceans and Their Role in Shaping the Global Environment" (O.M. Johannessen, R. D. Muench, and J. E. Overiand, Eds.), pp. 255-262. Geophysical Monograph Series, American Geophysical Union, Washington, DC. Archer, D., Peltzer, E. T., and Kirchman, D. L. (1997). A timescale of DOC production in equatorial Pacific surface waters. Global Biogeochem. Cycles 11,435-452. Banoub, M. W., and Williams, P. J. leB (1973). Seasonal changes in the organic forms of carbon, nitrogen and phosphorus in the EngUsh Channel in 1968. J. Mar. Biol. Assoc. UK 53,695-703. Bates, N., and Hansell, D. A. (1999). Hydrographic and biogeochemical signals in the surface ocean between Chesapeake Bay and Bermuda. Mar. Chem. 67,1-16. Bates, N. R., Hansell, D. A., Carlson, C. A., and Gordon, L. I. (1998). Distribution of CO2 species, estimates of net conmiunity production and air-sea CO2 exchange in the Ross Sea polynya. J. Geophys. Res. 103,2883-2896. Bauer, J. E. (2002). Carbon isotopic composition of DOM. In "Biogeochemistry of Marine Dissolved Organic Matter" (D. A. Hansell and C. A. Carlson, Eds.), pp. 405^53. Academic Press, San Diego. Bauer, J. E., Druffel, E. R. M., WilUams, P M., Wolgast, D. M., and Griffin, S. (1998). Temporal variability in dissolved organic carbon and radiocarbon in the eastern North Pacific Ocean. J. Geophys. Res. 103, 2867-2881. Borsheim, K. Y., and Myklestad, S. M. (1997). Dynamics of DOC in the Norwegian Sea inferred from monthly profiles collected during 3 years at 66°N, 2°E. Deep-Sea Res. I 36,497-507. Borsheim, K. Y., Myklestad, S. M., and SneU, J. A. (1999). Monthly profiles of DOC, mono- and polysaccharides at two locations in the TrondheimsQord (Norway) during two years. Mar. Chem. 63,255-272. Carlson, C. A. (2002). Production and removal processes. In "Biogeochemistry of Marine Dissolved Organic Matter" (D. A. Hansell and C. A. Carlson, Eds.), pp. 91-151. Academic Press, San Diego. Carlson, C. A., and Ducklow, H.W. (1995). Dissolved organic carbon in the upper ocean of the central equatorial Pacific Ocean, 1992: Daily and finescale vertical variations. Deep-Sea Res. II42,639656. Carlson, C. A., Ducklow, H. W, Hansell, D. A., and Smith, W O., Jr. (1998). Organic carbon partitioning during spring phytoplankton blooms in the Ross Sea polynya and the Sargasso Sea. Limnol. Oceanogr. 43, 375-386. Carlson C. A., Ducklow, H. W, and Michaels, A. F. (1994). Annual flux of dissolved organic carbon from the euphotic zone in the northwestern Sargasso Sea. Nature 371,405-408. Carlson, C. A., Ducklow, H. W, and Sleeter, T. D. (1996). Stocks and dynamics of bacterioplankton in the northwestern Sargasso Sea. Deep-Sea Res. II43,491-515. Carlson, C. A., Hansell, D. A., Peltzer, E. T, and Smith, W. O., Jr. (2000). Stocks and dynamics of dissolved and particulate organic matter in the southern Ross Sea, Antarctica. Deep-Sea Res. II47, 3201-3225. Cauwet, G. (2002). DOM in the coastal zone. In "Biogeochemistry of Marine Dissolved Organic Matter" (D. A. Hansell and C. A. Carlson, Eds.), pp. 579-609. Academic Press, San Diego. Chavez, F. P., and Toggweiler, J. R. (1995). Physical estimates of global new production: The upwelling contribution. In "Upwelling in the Ocean: Modem Processes and Ancient Records" (C. P. Summerhayes et al., Eds.), pp. 313-320. Wiley, New York. Chen, R. F , Fry, B., Hopkinson, C. S., Repeta, D. J., and Peltzer, E. T. (1996). Dissolved organic carbon on Georges Bank. Cont. Shelf Res. 16,409-420. Christian, J. R., Lewis, M. R., and Karl, D. M. (1997). Verticalfluxesof carbon, nitrogen, and phosphorus in the North Pacific Subtropical Gyre near Hawaii. J. Geophys. Res. 102,15,667-15,677. Church, M. J., Ducklow, H. W., and Karl, D. M. (2002). Decade-scale secular increase in dissolved organic carbon and nitrogen inventories in the North Pacific subtropical gyre. Limnol. Oceanogr. 47,1-10.
DOC in the Global Ocean Carbon Cycle
713
Copin-Montegut, G., and Avril, B. (1993). Vertical distribution and temporal variation of dissolved organic carbon in the Northwestern Mediterranean Sea. Deep-Sea Res. 740,1963-1972. Craig, H. (1971). The deep metaboHsm: Oxygen consumption in abyssal ocean water. J. Geophys. Res. 76, 5078-5086. Dileep Kumar, M., Rajendran, A., Somasundar, K., Haake, B., Jenisch, A., Shuo, Z., Ittekkot, V., and Desai, B. N. (1990). Dynamics of dissolved organic carbon in the northwestern Indian Ocean. Mar. Chem. 31,299-316. Doval, M. D., Alvarez-Salgado, X. A., and Perez, F. F. (1997). Dissolved organic matter in a temperate embayment affected by coastal upwelling. Mar. Ecol. Prog. Ser. 157, 21-37. Doval, M., and Hansell, D. A. (2000). Organic carbon and apparent oxygen utilization in the western South Pacific and central Indian Oceans. Mar. Chem. 68, 249-264. Druffel, E. R. M., WiUiams, P. M., Bauer, J. E., and Ertel, J. R. (1992). Cycling of dissolved and particulate organic matter in the open ocean. J. Geophy. Res. 97,15639-15659. Ducklow, H. W., Carlson, C. A., Bates, N. R., Knap, A. H., and Michaels, A. F (1995). Dissolved organic carbon as a component of the biological pump in the North Atlantic Ocean. Phil. Trans. R. Soc. London B 348,161-167. Dugdale, R. C , and Goering, J. J. (1967). Uptake of new and regenerated forms of nitrogen in primary productivity. Limnol. Oceangr. 12,196-207. Dunne, J. P., Murray, J. W., Rodier, M., and Hansell, D. A. (2000). Export production in the western and central equatorial Pacific using ^^^Th calibrated sediment traps and TOC accumulation. Deep-Sea Res. I 47,901-936. Duursma, E. K. (1962). Dissolved organic carbon, nitrogen and phosphorus in the sea. Neth. J. Sea Res. 1, 1-148. Duursma, E. K. (1963). The production of dissolved organic matter in the sea, as related to the primary gross production of organic matter. Neth. J. Sea Res. 2, 85-94. Emerson, S., Quay, P., Karl, D. M., Winn, C , Tupas, L., and Landry, M. (1997) Experimental determination of the organic carbon flux from open-ocean surface waters. Nature 389, 951-954. Feely, R. A., Wanninkhof, R., Cosca, C. E., Murphy, R R, Lamb, M. F , and Steckley, M. D. (1995). CO2 distributions in the equatorial Pacific during the 1991-1992 ENSO event. Deep-Sea Res. II 42, 365-386. Fung, I. Y, Meyn, S. K., Tegen, I., Doney, S. C , John, J. G., and Bishop, J. K. B. (2000). Iron supply and demand in the upper ocean. Global Biogeochem. Cycles 14, 281-295. Goldman, J. C , Hansell, D. A., and Dennett, M. R. (1992). Chemical characteristics of two large chain-forming oceanic diatoms: Impact on water column chemistry. Mar Ecol. Progn Ser. 88, 257-270. Guo, L., Coleman, C. H., Jr., and Santschi, P. H. (1994). The distribution of colloidal and dissolved organic carbon in the Gulf of Mexico. Mar Chem. 45,105-119. Guo, L., Santschi, P. H., and Wamken, K. W. (1995). Dynamics of dissolved organic carbon (DOC) in oceanic environments. Limnol. Oceanogr 40,1392-1403. Hansell, D. A., Bates, N. R., and Carlson, C. A. (1997a). Predominantly vertical losses of carbon from the surface layer of the Equatorial Pacific Ocean. Nature 386, 59-61. Hansell, D. A., and Carlson, C. A. (1998a). Deep ocean gradients in dissolved organic carbon concentrations. Nature 395, 263-266. Hansell, D. A., and Carlson, C. A. (1998b). Net community production of dissolved organic carbon. Global Biogeochem. Cycles 12,443-453. Hansell, D. A., and Carlson, C. A. (2001a). Biogeochemistry of total organic carbon and nitrogen in the Sargasso Sea: Control by convective overturn. Deep-Sea Res. II48,1649-1667. Hansell, D. A., and Carlson, C. A. (2002 b). Dissolved organic carbon export by ocean circulation and mixing. Submitted for publication. Hansell, D. A., Carlson, C. A., Bates, N. R., and Poisson, A. (1997b). Horizontal and vertical removal
714
Dennis A. Hansell
of organic carbon in the equatorial Pacific Ocean: A mass balance assessment. Deep-Sea Res. II 44,2115-2130. Hansell, D. A., Carlson, C. A., and Suzuki, S. (2002). Dissolved organic carbon export with North Pacific Intermediate Water formation. Global Biogeochem. Cycles, in press. Hansell, D. A., and Feely, R. A. (2000). Atmospheric intertropical convergences impact surface ocean carbon and nitrogen biogeochemistry in the tropical Pacific Ocean. Geophys. Res. Lett. 27,10131016. Hansell, D. A., and Peltzer, E. T. (1998). Spatial and temporal variations of total organic carbon in the Arabian Sea. Deep-Sea Res. II45,2171-2193. Hansell, D. A., and Waterhouse, T. Y. (1997). Controls on the distribution of organic carbon and nitrogen in the eastern Pacific Ocean. Deep-Sea Res. 144, 843-857. Holmes, R. W., Williams, P M., and Eppley, R. W. (1967). Red water in La JoUa Bay, 1964-1966. Limnol. Oceanogr. 12,503-512. Kahler, P., Bjomsen, P. K., Lochte, K., and Anita, A. (1997). Dissolved organic matter and its utiUzation by bacteria during spring in the Southern Ocean. Deep-Sea Res. II 44, 341353. Kahler, P., and Koeve, W. (2001). Dissolved organic matter in the sea: Can its C:N ratio explain carbon overconsumption? Deep-Sea Res. I 48,49-62. Karl, D. M. (1999). A sea of change: Biogeochemical variability in the North Pacific subtropical gyre. Ecosystems 2,181-214. Karl, D. M., Letelier, R., Tupas, L., Dore, J., Christian, J., and Hebel, D. (1997). The role of nitrogen fixation in biogeochemical cycling in the subtropical North Pacific Ocean. Nature 388, 533-538. Kortzinger, A., Koeve, W, Kahler, P., Mintrop, L. (2001). C:N ratios in the mixed layer during the productive season in the northeast Atlantic Ocean. Deep-Sea Res. I 48,661-688. Lee, K. (2001). Global net conmiunity production estimated from the annual cycle of surface water total dissolved inorganic carbon. Limnol. Oceanogr. 46,1287-1297. Loh, A. N., and Bauer, J. E. (2000). Distribution, partitioning and fluxes of dissolved and particulate organic C, N and P in the eastern North Pacific and Southern Oceans. Deep-Sea Res. I 47, 2287-2316. Martin, J. H., and Fitzwater, S. E. (1992). Dissolved organic carbon in the Atlantic, Southern and Pacific oceans. Nature 356,699-700. Menzel, D. W. (1964). The distribution of dissolved organic carbon in the Western Indian Ocean. Deep-Sea Res. 11,757-765. Menzel, D. W. (1970). The role of in situ decomposition of organic matter on the concentration of non-conservative properties in the sea. Deep-Sea Res. 17,751-764. Menzel, D. W, and Ryther, J. H. (1970). Distribution and cycling of organic matter in the oceans. In "Organic Matter in Natural Waters" (D. W. Hood, Ed.), Institute of Marine Science Occasional Publication 1, pp. 31-54. Univ. of Alaska, Fairbanks, AK. Michaels, A. F , and Knap, A. H. (1996). Overview of the U.S. JGOFS Bermuda Adantic Time-series Study and the Hydrostation S program. Deep-Sea Res. II43,157-198. Murray, J. W., Barber, R. T., Roman, M. R., Bacon, M. P., and Feely, R. A. (1994). Physical and biological controls on carbon cycling in the Equatorial Pacific. Science 266,58-65. Noji, T. T., Rey, F , Miller, L. A., Borsheim, K. Y., and Urban-Rich, J. (1999). Fate of biogenic carbon in the upper 200 m of the central Greenland Sea. Deep-Sea Res. II46,1497-1509. Ogura, N. (1970). The relation between dissolved organic carbon and apparent oxygen utilization in the western North Pacific. Deep-Sea Res. 17,221-231. Opsahl, S., Benner, R., and Amon, R. M. W. (1999). Major flux of terrigenous dissolved organic matter through the Arctic Ocean. Limnol. Oceanogr. 44,2017-2023. Parsons, T. R., LeBrasseur, R. J., and Barraclough, W E. (1970). Levels of production in the pelagic
DOC in the Global Ocean Carbon Cycle
715
environment of the Strait of Georgia, British Columbia: A review. /. Fish. Res. Bd. Can. 27, 1251-1263. Peltzer, E. T., and Hayward, N. A. (1996). Spatial and temporal variabiHty of total organic carbon along 140°W in the equatorial Pacific Ocean in 1992. Deep-Sea Res. II43,1155-1180. Postma, H., and Rommets, J. W. (1979). Dissolved and particulate organic carbon in the north Equatorial Current of the Atlantic Ocean, Neth. J. Sea Res. 12, 85-98. Romankevich, E. A., and Ljutsarev, S. V. (1990). Dissolved organic carbon in the ocean. Mar. Chem. 30,161-178. Sharp, J. H. (1997). Marine dissolved organic matter: Are the older values correct? Mar. Chem. 56, 265-277. Steinberg, D. K., Carlson, C. A., Bates, N. R., Johnson, R. J., Michaels, A. E, and Knap, A. H. (2001). Overview of the US JGOFS Bermuda Atlantic Time-series Study (BATS): A decade-scale look at ocean biology and biogeochemistry. Deep-Sea Res. II48,1405-1447. Stommel, H. M. (1979). Determination of water mass properties of water pumped down from the Ekman layer to the geostrophic flow below. Proc. Natl. Acad. Sci. USA 76, 3051-3055. Sugimura, Y., and Suzuki, Y. (1988). A high-temperature catalytic oxidation method for the determination of non-volatile dissolved organic carbon in seawater by direct injection of liquid sample. Mar. Chem. 24,105-131. Sweeney, C , Hansell, D. A., Millero, F. J., Takahashi, T, Gordon, L. I., Carlson, C. A., Codispoti, L. A., Smith, W. O., Jr., and Marra, J. (2000). Biogeochemical regimes, net conmiunity production and carbon export in the Ross Sea, Antarctica. Deep-Sea Res. II47, 3369-3394. Talley, L. D. (1997). North Pacific Intermediate Water transports in the Mixed Water Region. /. Phys. Oceanogr. 27,1795-1803. Talley, L. D. (1999). Some aspects of ocean heat transport by the shallow, intermediate and deep overturning circulations. Geophys. Monogr. Sen 112,1-22. Tanoue, E. (1993). Distributional characteristics of DOC in the central Equatorial Pacific. /. Oceanogr. 49,625-639. Thomas, C , Cauwet, G., and Minster, J.-F. (1995). Dissolved organic carbon in the equatorial Atlantic Ocean. Mar Chem. 49,155-169. Vidal, M., Duarte, C. M., and Agusti, S. (1999) Dissolved organic nitrogen and phosphorus pools and fluxes in the central Atlantic Ocean. Limnol. Oceanogr 44,106-115. Wangersky, P. J. (1978). Production of dissolved organic matter. Mar Ecol. 4,115-220. Wheeler, P. A., Watkins, J. M., and Hansing, R. L. (1997). Nutrients, organic carbon and organic nitrogen in the upper water column of the Arctic Ocean: implications for the sources of dissolved organic carbon. Deep-Sea Res. II44,1571-1592. Wiebinga, C. J., and de Baar, H. J. W. (1998). Determination of the distribution of dissolved organic carbon in the Indian sector of the Southern Ocean. Mar Chem. 61,185-201. Williams, P. J. le B. (1995). Evidence for the seasonal accumulation of carbon-rich dissolved organic material, its scale in comparison with changes in particulate material and the consequential effect on net C/N assimilation ratios. Mar Chem. 51,17-29. Williams, P. M., Carlucci, A. E, and Olson, R. (1980). A deep profile of some biologically important properties in the central North Pacific. Oceanol. Acta 3,471-476. WiUiams, P. M., and Druffel, E. R. M. (1987). Radiocarbon in dissolved organic carbon in the central north Pacific Ocean. Nature 330,246 -248. Worthington, L. V. (1976) "On the North Atlantic Circulation," Vol. 6. The Johns Hopkins Oceanographic Studies, Baltimore, MD. Zhang, J., and Quay, P. D. (1997) The total organic carbon export rate based on ^^C and ^^C of DIC budgets in the equatorial Pacific region. Deep-Sea Res. II44,2163-2190.
This Page Intentionally Left Blank
Chapter 16
Modeling DOM Biogeochemistry James R. Christian^
Thomas R. Anderson
Universities Space Research Association NASA Goddard Space Flight Center, Code 970.2 Greenbelt, Maryland
George Deacon Division Southampton Oceanography Centre, Southampton United Kingdom
I. Introduction II. Ecosystem Modeling Studies A. Compartments and Currencies B. Modeling Production of DOM C. Modeling Utilization and Remineralization of DOM III. Modeling the Role of DOM in Ocean Biogeochemistry A. Major Classes of Ocean Models B. Early Results: Equatorial Nutrient-Trapping C. DOM Turnover Times
D. DOM and Atmospheric CO2 E. Diagnostic Models of the North Atlantic F. Basin-to-Global Scale Ecosystem Modeling G. Inverse Models of Flows within Food Webs H. Coastal and Estuarine Systems I. Small-Scale Spatial Structure IV. Discussion and Conclusions References
I. INTRODUCTION In the 1970s and 1980s it was realized that planktonic food webs were far more complex than had previously been appreciated, and that microorganisms ^ Present address: Earth System Science InterdiscipHnary Center, University of Maryland, College Park, MD 20742. Biogeochemistry of Marine Dissolved Organic Matter Copyright 2002, Elsevier Science (USA). All rights reserved.
717
718
Christian and Anderson
constitute the majority of biomass and respiration in most planktonic systems (Pomeroy, 1974; Williams, 1981; Azam et al, 1983). This new paradigm of the "microbial loop" developed contemporaneously with a heightened interest in the ocean's role in the global carbon cycle, following the realization that changes in ocean biogeochemistry may have forced changes in atmospheric CO2 during glacial periods (Broecker, 1982) and that a large "missing sink" for atmospheric CO2 existed, in which the ocean biota might play a role (Wong, 1978; Walsh etal, 1981). Although the literature of ocean biogeochemical modeling has sometimes been concerned with one or the other of these developments without explicitly linking the two (i.e., the microbial food web is highly parameterized in some ocean biogeochemical models, and some microbial food web models are concerned primarily with organismal interactions rather than biogeochemical cycles), at least some investigators drew the connection quite early on (e.g., Pace etal, 1984). These developments predated the "discovery" of dissolved organic carbon (DOC) measured by high-temperature catalytic oxidation (HTCO) but not detected by other methods, in the late 1980s (Sugimura and Suzuki, 1988). This development heightened interest in DOC in the biogeochemical modeling community and led to the first global-scale simulations of oceanic DOC. While the original results of Sugimura and Suzuki (1988) are now considered erroneous (Suzuki, 1993), they opened the door to a different view of dissolved organic matter (DOM). In particular, it is now known that there is a substantial pool of DOM with a turnover time much less than the ocean overturning time scale (~1000 years), but long enough to allow carbon and nutrients to be remineralized far from where they were incorporated into organic matter. After elevated levels of DOC and dissolved organic nitrogen (DON) were reported using HTCO methods (Suzuki et al, 1985; Sugimura and Suzuki, 1988), a number of modeling experiments were undertaken to assess the implications of these observations for the global carbon cycle (e.g., Bacastow and Maier-Reimer, 1991; Najjar et al, 1992; Paillard et al, 1993). These experiments are of considerable interest in terms of understanding the evolution of the field, although their contemporary relevance is somewhat diminished by new information that calls into question some the assumptions employed. Most importantly, these authors took the concentration estimates of Suzuki and coworkers to be substantially accurate, and assumed that a "Redfield equivalent" pool of dissolved organic phosphorus (DOP) would eventually be observed as well. Neither of these assumptions remained tenable for long. By 1993, it had become clear that the HTCO-DOC estimates of Sugimura and Suzuki (1988) were excessive, although a pool of DOC not measured by other methods does exist (Suzuki, 1993; Hedges and Lee, 1993). Furthermore, the DON estimates of Suzuki et al (1985) were shown to be erroneous, and no clear evidence for either DON or DOP not measurable by non-HTCO methods exists (Hansen, Chapter 15).
Modeling DOM Biogeochetnistry
719
The conclusions drawn in the early global modeling experiments must therefore be reconsidered in light of an HTCO-DOC pool that is (a) considerably smaller than originally assumed, although not negligible and with similar vertical distribution, and (b) considerably out of Redfield ratio, i.e., strongly enriched in C relative to N and P. More recently, it has been shown that the equatorial nutrient-trapping and anoxicity that Bacastow and Maier-Reimer (1991) and Najjar et al. (1992) attempted to remedy by adding a DOM component to their models may be largely an artifact of the coarse resolution of the ocean circulation models employed, and can be eliminated simply by increasing resolution with no modification to the biogeochemical model (Aumont et al, 1999). Matear and Holloway (1995) also showed that these artifacts could be remedied by changing circulation fields without including any DOM component in the biogeochemical model. However, the basic conclusion of Toggweiler (1989), that advection of DOM will substantially alter the distribution of nutrients, oxygen and dissolved inorganic carbon (DIC) relative to an ocean model with only sedimentation of particulate organic matter (POM), remains sound and relevant. The need for prognostic, mechanistic models of the processes that create and consume this pool is acute, and the fact that its elemental composition deviates from Redfield ratio underscores the need for a more sophisticated treatment of variable stoichiometry in biogeochemical models.
11. ECOSYSTEM MODELING STUDIES A. COMPARTMENTS AND CURRENCIES The bulk DOM pool is still largely uncharacterized (Benner, Chapter 3) and cycling of elements through DOM is poorly understood from a mechanistic perspective (Kirchman etai, 1993a; Azam, 1998). Reducing the complexity of DOM biogeochemistry to representative and quantifiable structures in models is therefore difficult, and a diversity of approaches and model structures have been utilized (Table I). Early ecosystem modeling studies examined the role of the microbial loop in recycling of nutrients and as a pathway for carbon transfer to higher trophic levels. The models of Fasham et al. (1990) and Taylor and Joint (1990) included labile DOM and heterotrophic bacteria (HBAC) as state variables, but no slowturnover DOM pools. More recently, interest has focused on the biogeochemical role of longer-lived pools of DOM and, in particular, their contributions to export from the euphotic zone. Despite its heterogeneous nature, modelers have endeavored to categorize DOM into different classes to distinguish material that turns over rapidly from that which accumulates and can potentially be exported. Highmolecular-weight organic matter requires enzymatic hydrolysis in order to provide the simple monomers that can be taken up by bacteria (Chrost, 1990), and so in principle should be utilized more slowly than monomers. The "HSB" model developed
720
Christian and Anderson Table I Model CharacteristicsPOM DOM
Reference
Type
State variables
Fashamefa/., 1990
OD
NPZDB
L
N
N
Billen and Becquevort, 1991
OD
B
L,S
C
C
Connolly and Coffin, 1995
OD
ZDB
L,S
C
C
Kawamiyaera/., 1995
ID
NPZD
S
N
N
Six and Maier-Reimer, 1996
3D
NPZD
S
P(C)
C
Anderson and Williams, 1998
OD
NPZDB
L,S
N(C)
N,C
LcwyetaL, 1998
ID
NPZDB
L,S
N
N
Anderson and Williams, 1999
ID
B
L,S,R
C
C
BissetX etai, 1999
ID
NPZDB
L,R
N(C)
N,C
Waish etai, 1999
3D
NPZDB
L,S
N(C)
C
Tianera/., 2000
ID
NPZDB
L
N(C)
c
Vallino, 2000
OD
NPZDB
US
N,C
N,C
DOM pools
Note. State variables: nutrient (N), phytoplankton (P), zooplankton (Z), detritus (D), and bacteria (B) are listed if present. Others not listed. DOM pools: labile (L, turnover rate hours to days), semilabile (S, weeks to months), refractory (R, decades and longer); terminology may differ in original texts. Currencies: parentheses indicate fixed C / N or C/P ratios.
by the Brussels group (Billen, 1990; Billen and Becquevort, 1991; Lancelot etai, 1991) exploited this principle by including two polymeric pools, with fast and slow rates of hydrolysis by bacterial ectoenzymes, which are converted to a common monomeric pool which is consumed rapidly by bacteria. However, the correlation between molecular weight and lability is surprisingly weak in natural DOM (Benner, Chapter 3). High-molecular-weight material can be highly bioreactive, while conversely the bulk of oceanic DOM comprises small molecules that cycle slowly or are relatively unavailable to microorganisms (Amon and Benner, 1994, 1996; Kepkay, 2000). The simplest distinction between different DOM pools can be made simply on the basis of turnover rates, without necessarily invoking underlying causes. The early work of Ogura (1975) indicated that DOM decomposition in coastal seawater occurred in two distinct phases with rates of 0.1-1 and 0.007 day"^ with a third fraction remaining unutilized. Experiments with DOM derived from phytoplankton in the laboratory also suggest a relatively small number of fractions, although the rate coefficients are quite variable (Pett, 1989; Chen and Wangersky, 1996). The bulk DOM pool can be usefully categorized into labile, semilabile, and
Modeling DOM Biogeochemistry
721 DOC (mmol m-^)
0
10
20
30
40
0 •
50
60
70
jQsemi-labil^/Q^
80
90
JQ^ 1
Qf/om
200-
(Se)
/MO
# 1 **
400-
j
# T3
600-
i
QefractorJ)
o
800-
1000- " " " " " ^ ^ ^ ^ " H P " ^ ^ ^
0
10
20
o
> •
p
1
30-9-91^ 40
o
1 50
60r
1
1
1
70
80
90
Figure 1 Vertical profile of DOC as simulated by the model of Anderson and Williams (1999), split into its component parts: labile, semilabile, and refractory. Data are for various profiles in the Atlantic (solid points) and Pacific oceans (open points) (see Anderson and Williams, 1999, for details).
refractory fractions (Kirchman etal, 1993a; Carlson and Ducklow, 1995; Cherrier et al, 1996) on the basis of turnover rates. Labile material is consumed rapidly, on time scales of hours to days, semilabile material degrades on seasonal time scales, while refractory material degrades very slowly and may be biologically inert. Anderson and Williams (1999) modeled vertical profiles of DOC in the ocean using a model based on these three fractions (Fig. 1). Many other models have employed two pools representing labile and semilabile compounds without necessarily assuming that this distinction is exactly analogous to, for example, monomers vs polymers (e.g., Connolly and Coffin, 1995; Anderson and Williams, 1998; Levy et al, 1998; Walsh et al, 1999; Vallino, 2000). The terminology regarding different fractions has been variously applied in the literature, e.g., the semilabile pool as defined above has been described as both labile (e.g.. Six and Maier-Reimer, 1996) and refractory (e.g., Walsh and Dieterle, 1994; Levy et al, 1998). Few models contain long-lived refractory pools (Anderson and WiUiams, 1999; Bissett etal, 1999a).
722
Christian and Anderson
Phytoplankton production is often limited by nutrient elements such as N or P, so one of these is usually employed as a model currency. Nitrogen is in some cases the only model currency (e.g., Fasham et al, 1990; Kawamiya et al., 1995; Levy et al, 1998). However, the growth of heterotrophic bacteria may be carbon- or energy-Umited (Kirchman, 1990; Carlson and Ducklow, 1996). Fasham et al. (1990) derived a relationship to balance the uptake of DON and ammonium based on assumed C/N ratios of bacteria and DOM of 5 and 8, respectively. Flows of carbon in multielement models are often calculated by assuming fixed C/N (or C/P) ratios for state variables. Ratios in zooplankton and bacteria are commonly different (lower) than those in phytoplankton and DOM. Elemental ratios in zooplankton and bacteria, and to a lesser extent phytoplankton, are relatively constant, whereas ratios in DOM are more variable, for example having highest C/N ratios during accumulation in spring (Williams, 1995). It is therefore necessary to stoichiometrically balance N cycling with DOC uptake and respiration by bacteria (e.g., Anderson, 1992; Goldman and Dennett, 2000). Two approaches have been used to overcome difficulties with variable DOC/DON: DOC can be included in models without associated DON, or DOC and DON can be included as separate state variables permitting varying C/N. DOC and DON are, of course, inextricably Hnked; although N-free DOC exists, organic compounds contain C and so there is no DON without DOC. The first approach is useful for modeling accumulation and turnover of DOC, but neglects the role of DON in recycling of nutrients. This approach does not permit nitrogen to enter slow-turnover DON pools, and may therefore result in overestimation of remineralization rates. Bacterial growth is assumed to be limited by DOC, and nitrogen requirements are taken from the inorganic pool (Fig. 2). The second approach is to have separate state variables for DOC and DON, giving rise to a dynamic C/N. Models of this type (e.g., Moloney and Field, 1991; Anderson and Williams, 1998; ValUno, 2000) permit a detailed examination of the roles of DOM in nutrient
CO.
co^ / I bacteria
I-co.
,.i-l phytoplankton
1 ^ m'- I ^ ^ 1
I
^ J__, U ^
-I I
DOC
zooplankton
iLr J - , 'I I I I I
U I
Figure 2 Ecosystem model structure which includes DOC but not DON. Solid lines, N or P flows; dashed lines, C flows.
723
Modeling DOM Biogeochemistry
cycling and accumulation and export of DOC and DON, but parameterization of the interactions between C and N is far from simple.
B. MODELING PRODUCTION OF DOM DOM is produced from a variety of sources, so ecosystem models need to be complex if production processes are to be fully addressed. Most ecosystem models contain phytoplankton, zooplankton, and detritus, thus providing the potential for sources to be adequately defined. Models which do not include a full ecosystem, but do contain HBAC and DOM as state variables, have been used to study DOM cycling. Billen and Becquevort (1991) modeled bacterial production in Antarctic coastal waters using observed phytoplankton biomass to estimate production of DOC. Anderson and WiUiams (1999) defined DOC production rate in the euphotic zone as afixedfraction of primary production and examined the fate of the organic carbon in a deep-water column. Some modeling studies that do include a full ecosystem include external as well as internal sources of DOM. Parsons and Kessler (1986) included a riverine source of DOC in an estuarine model. Walsh and Dieterle (1994) included a sedimentary source in their shallow-water model. Bissett et al (1999a) included a source from dinitrogen fixation without having an explicit population of N2-fixers. There are a large number of processes by which DOM is produced (Fig. 3), most of which are poorly understood. As a first approximation, these processes can be
photo-oxidation I
A
lysis
DOCI
semi-labileT
bacteria
labile
solubilization^^
detritus
A
JI.-*
exudation
phytoplankton
refractory
lysis \ s osolubilization ]
grazer-associated losses
pellets
zooplankton Figure 3 Idealized food web illustrating the four main DOC production terms (phytoplankton exudation, grazer-associated losses, lysis, detrital solubilization) and the two principal loss routes (bacterial uptake, photo-oxidation).
724
Christian and Anderson Table U Sources of DOM in Models That Include an Ecosystem Phytoplankton
Zooplankton
Lysis
Detritus
Fasham et ai, 1990
0.05pp
0.025Z
—
0.05D
Connolly and Coffin, 1995''
0.15pp
{0.07-0.14}g
—
{0.1-0.3}D
Kawamiya etai, 1995
0.135pp
—
—
0.5(0.03+ 0.0693T)D^
Six and MaierReimer, 1996
0.03(P-0.01)
0.06(Z-0.01)
Anderson and Williams, 1998
(0.05 + 0.34)pp^
0.23g
0.04B + 0.015P
0.05D
L&yy etai, 1998
0.05pp
{0.005-0.025}Z
Bissett et ai, 1999
0.05P
{0.33-0.46}g
—
Walsh et ai, 1999
0.04pp
0.5g
0.0075P
{0.02-0.04}?
0.4g
Reference
Tian et ai, 2000 Vallino, 2000
0.564pp
{0.008-0.075}D
0.132D d
—
49.6D^
Note. P, phytoplankton; Z, zooplankton; B, bacteria; D, detritus; T, temperature; pp, primary production; g, grazing rate. All rates are per day. Brackets indicate a range of values; parentheses have their standard meaning in mathematical expressions. "Includes only partial ecosystem (input observed pp). ^An equivalent amount is remineralized directly to DIN. '^First term (0.05) leakage (C, N), second term exudation (C only). ^Includes breakdown of large (sinking) to small (nonsinking) particles. ^Final value after optimization from a first guess of 0.1 day~^ with a range of 0-50, indicating that this parameter was essentially unconstrained by the data.
divided into four categories: phytoplankton exudation, grazer-associated losses, lysis, and detrital turnover. Different models have included different combinations of these terms, and parameterized them in different ways (Table II). 1. Phytoplankton Exudation Phytoplankton exudation may involve both passive "leakage" and active release. Leakage results from the permeability of the plasma membrane to low-molecularweight compounds (Bj0msen, 1988). Metabolic instabilities in algae may cause "extra" carbon to be actively exuded as DOC (Williams, 1990). High release rates
Modeling DOM Biogeochemistry
725
may be a common feature of oligotrophic waters, possibly a result of continued use of photosynthetic machinery after nutrient exhaustion (Norrman et al., 1995; Obemosterer and Hemdl, 1995). Exuded polymers may serve a number of functions, few if any of which are well understood (Decho, 1989; Hoagland et aly 1993). Literature estimates of percentage extracellular release (PER) of DOC have a mean of about 13% of primary production, although considerably higher estimates exist (Baines and Pace, 1991; Nagata, 2000). There is significant uncertainty about a mean value due to both methodological considerations and adequacy of data coverage; oceanic environments are underrepresented (Baines and Pace, 1991; Nagata, 2000). The balance of the data seem to suggest that passive leakage is not the dominant mechanism, i.e., that exudation is more closely related to primary production rather than to phytoplankton biomass (Baines and Pace, 1991), although some studies suggest the opposite (e.g., Fuhrman et al, 1980). Bj0msen (1988) argued for the representation of exudation as a "property tax" (some fraction of biomass per unit time) rather than an "income tax" (a fraction of photosynthesis) and estimated the rate as about 5% of carbon biomass per day. Models have variously used both approaches (Table II). Rates of production may differ for carbon and nitrogen. A range of compounds including simple sugars and amino acids may be leaked from cells, containing both C and N, whereas exudation due to an overflow of photosynthate might be expected to be dominated by nonnitrogenous compounds. Anderson and Williams (1998) distinguished in their model between leakage, which occurred in the C/N ratio of the phytoplankton, and exudation which was only carbon. The Bissett et al (1999a) exudation term supplied only DOC. Vallino (2000) optimized model parameters to fit mesocosm data and found that a very high C/N in phytoplankton exudate (43200), as well as an exuded fraction of greater than 50% of photosynthesis, achieved the best results. Anderson and Williams (1998) adjusted the exuded fraction in their model in order to simulate the seasonal accumulation of DOC at a station in the English Channel and similarly found that a high DOC exudation rate (PER of 29%) was required. A few models have treated exudation rate as being inversely related to photosynthetic or phytoplankton growth rate (e.g.. Pace et al, 1984; Bratbak and Thingstad, 1985), but this approach has not been favored in the more recent literature and does not appear to be supported by the available data (Baines and Pace, 1991). Anderson and Williams (1998) compared two formulations of exudation—a constant fraction of photosynthesis, or a fraction of the difference between nutrient-limited and nutrient-saturated growth rates (cf. Bratbak and Thingstad, 1985). They were unable to assert that either of these parameterizations was more useful than the other. 2. Grazer-Associated DOM Production There are several mechanisms of grazer-associated DOM production, including so-called "sloppy feeding" (release of dissolved compounds when prey cells are
726
Christian and Anderson
broken by mouthparts of crustacean zooplankton), direct exudation of DOM, and egestion and dissolution of fecal material. Sloppy feeding losses are negligible if prey can be ingested whole and are likely restricted to metazoa (Nagata, 2000). Rapid pellet dissolution may provide an important mechanism of DOC production (Jumars et ai, 1989). Strom et al (1997) estimated that between 16 and 37% of algal C is released as DOC during grazing by phagotrophic protozoa. The most common modeling approach is to direct a fixed fraction of grazed material to DOM. Values assigned in this manner typically range between 0.2 and 0.4 (Table II), giving rise to large fluxes of DOM via zooplankton. Another approach has been to direct a fraction of zooplankton losses (excretion, mortality) to DOM. Several models do not have any explicit term for production of DOM by zooplankton, but direct losses to "particulate" detritus which is subsequently solubilized to DOM. Including this lag may cause small differences in the dynamic behavior of models, but it is unlikely that these differences are meaningful given the overall uncertainty regarding definitions of particulate and dissolved. Most modelers will likely continue to assign losses to "DOM" or "detritus" depending on assumptions about the size of the grazers that dominate in a particular ecosystem; in open ocean ecosystems, aflowthat is predominantly directly to DOM is appropriate (Nagata, 2000). 3. Lysis Planktonic microorganisms may be lysed by a variety of agents including viruses, bacteria, and under certain circumstances their own intrinsic hydrolytic enzymes (autolysis). Lytic processes potentially explain a great deal of DOM production, especially the more refractory fractions, as cell walls and membranes are sources of important components of DOM (Tanoue et al, 1995; McCarthy et al, 1998). There have been relatively few attempts to model the details of these processes. Models of viral infection have been developed (e.g., Murray and Jackson, 1992), although these have not yet been incorporated into biogeochemical models that fully account for the fate of the DOM produced in this process, with the exception of the very comprehensive treatment by Blackburn et al (1996). These authors modeled four biochemical pools: protein, carbohydrate, nucleic acid, and phospholipid. The DOM released during viral lysis was divided about equally between protein and nucleic acid with only minor contributions from carbohydrate and phosphoUpid, which were largely retained in "ghost" cells. Thingstad (2000) modeled multiple bacterial strains with a common protozoan predator but host-specific viruses, and showed that the overall rate of viral infection and the importance of lysis in biogeochemical cycles depends strongly on the diversity of the bacterial conmiunity. Plankton models often include a nonspecific mortality term to account for phytoplankton loss processes other than grazing and sedimentation. This term may be
Modeling DOM Biogeochemistry
727
directed to POM, DOM, or inorganic nutrients. The first models to have assigned some of this loss term to DOM appear to have been Taylor and Joint (1990), who applied a first-order loss term to all of their biota groups and assigned some fraction of this to DOC, and Moloney and Field (1991), who defined a population of "senescent" cells that were then subject to lysis to DOC. More recent models that assign some of the loss to lysis to DOC include Anderson and Williams (1998) and Walsh et al, (1999). More commonly, this term is applied to particulate organic carbon (POC), which is subsequently subject to solubilization and/or remineralization.
4. Solubilization of Particles It has been observed that the rate of solubilization of particles by hydrolytic enzyme activity may be orders of magnitude greater than the rate at which the DOM produced is respired (Smith et al, 1992), implying that most particle mass enters the dissolved phase as DOM. Particle turnover is therefore often passed directly to DOM in models (e.g., Fasham et al, 1990; Anderson and Williams, 1998; Levy et al., 1998; ValHno, 2000), although other models recycle it at least in part directly to inorganic nutrients (e.g., Bissett et al, 1999a). Fluxes are modeled as first-order rate processes in many models. Modeling the underlying mechanisms of this process is in its infancy. A basic theory exists for fragmentation of aggregates by turbulence, although there are a number of ill-constrained parameters involved (Hill, 1996). The importance of turbulence in the fragmentation process is in dispute, however, because the turbulent energy required exceeds that normally found in the ocean by several orders of magnitude (Alldredge et al, 1990; Hill, 1998). Solubilization of POM to DOM is in general a biological process, for which few quantitative models exist. Vetter et al. (1998) determined the steady-state distribution of a freely released extracellular enzyme and its hydrolysate in an idealized aggregate with a single bacterium at its center. This simple model provides little quantitative information about rates of solubilization, but can be used to estimate the ratio of hydrolysate respired by attached bacteria to that lost to the environment as DOM, which was one to two orders of magnitude greater. The flux of hydrolysate to the cell depends strongly on the diffusivity of the enzyme, with smaller enzymes producing less benefit because hydrolysis occurs, on average, further from the cell. Within the range of diffusivities and partition (between solid and liquid phase) coefficients considered, the flux of hydrolysate increased linearly with increasing release of enzyme, i.e., there was no optimal rate of release. This simple model demonstrates that it is possible for bacteria in porous aggregates to survive by releasing hydrolytic enzymes and that literature estimates of "uncoupled solubilization" from field experiments are reasonable.
728
Christian and Anderson
5. Lability of DOM Produced In models with multiple DOM pools, production of DOM must be allocated between the various pools. Phytoplankton exudation is usually assumed to consist solely of labile molecules (e.g., Billen and Becquevort, 1991; Anderson and Williams, 1998; Walsh et ai, 1999). However, there is marked variability between models in how other fluxes are allocated. Billen and Becquevort (1991) assumed that DOC produced by lysis and sloppy feeding was partitioned equally between their two polymeric fractions. Walsh et al (1999) allocated 40% of DOCrelated grazing losses to the labile pool, with the remainder to semilabile. Connolly and Coffin (1995) assumed that both bacteria and phytoplankton biomass contain 15% labile and 20% semilabile C, part of which is released as DOC during zooplankton grazing. Various modeling studies have adjusted the allocation of organic fluxes to different pools in order to achieve acceptable fits to data. Anderson and Williams (1998) were able to simulate the seasonal DOC increase in the English Channel (34 /xmol L~^) by partitioning 90% of DOM produced by various processes (lysis, sloppy feeding, detrital turnover) to the semilabile pool (with the remainder labile). Lowering this fraction required unrealistically high phytoplankton exudation rates in order to generate sufficient DOC to match observations. In a mesocosm experiment, Vallino (20(X)) found the best fit to data when 67% of detrital turnover went to the semilabile pool. By contrast, in the Mediterranean Sea, Levy et al (1998) found that only 15% of DOC fluxes needed to be allocated to the semilabile pool. Processes contributing to the production of refractory DOM are poorly understood and have not been extensively modeled. Anderson and Williams (1999) allocated a small fraction of the turnover of labile and semilabile pools to the refractory pool; a value of 0.35% led to a balance between production and ultraviolet (UV) photooxidation (see Section II.C.3). Bissett etal (1999a) assumed that 4.0 % of DOC consumption was released as refractory material.
C. MODELING UTILIZATION AND REMINERALIZATION OF DOM The primary loss mechanism for DOM is uptake by heterotrophic bacteria. Measurements of bacterial production and growth efficiency show that bacterial respiration accounts for a large fraction of primary production in most oceanic ecosystems (Ducklow, 1999). Some eukaryotic microorganisms (Sherr, 1988; Marchant and Scott, 1993) and metazoa (Wright and Manahan, 1989) can take up dissolved or colloidal organic matter, but it is not known how widespread or quantitatively significant this process is, and it has not to our knowledge been explicitly incorporated into models. The other sink is abiotic photooxidation by solar ultraviolet radiation. Direct photooxidation of DOC to DIC (photomineralization) may be as great as photolysis to monomers and subsequent respiration by bacteria (Miller
729
Modeling DOM Biogeochemistry
and Zepp 1995; Miller and Moran 1997), although it is not known how quantitatively important this process is in the open ocean. Photochemical effects have been incorporated into several ecosystem models (Bissett et al, 1999a; Anderson and WilHams, 1999). 1. Turnover of Semilabile DOM Semilabile material is variously defined in different models, so it is unsurprising that parameterizations of its turnover vary. Connolly and Coffin (1995, p. 682) describe it as compounds that are "readily used, but are less optimal for bacterial growth." They assumed that semilabile material was utilized only after exhaustion of labile substrates and with a lower growth efficiency. Another definition is "molecules whose eventual assimilation by the bacteria requires ectoenzyme hydrolysis to the labile pool" (Anderson and Williams, 1998). Many models therefore employ Michaelis-Menten kinetics to describe turnover, which is usually passed to labile pools (Table III). However, estimates of the kinetic parameters are rare. Billen (1990) estimated parameter values from degradation of DOM derived from an algal culture; Walsh and Dieterle (1994) used these parameters in their model. Lamy et al (1999)fitthe HSB model directly to time-series data from experimental microcosms using nonlinear regression. Connolly et al (1992) and Connolly and Coffin (1995) derived values for a variety of coastal and freshwater environments; values from Santa Rosa Sound were applied to the English Channel by Anderson and Williams (1998). 2. Bacterial Utilization of Labile DOM The predominant model of uptake of dissolved organics by bacteria is a hyperbolic form similar to Michaelis-Menten kinetics, which has been widely used in ecosystem models (Davidson, 1996). Monod (1942) showed that the relationship of growth rate of bacteria in culture to substrate concentration has the form J W ^ Ks + S'
[1]
where S is the substrate concentration, /Xmax is the maximal growth rate, and ^ s is the substrate concentration at which /x=/Xinax/2. Monod's formulation has been widely adopted by the modeling community, although it lacks a strong theoretical basis (Button, 1998) and there is evidence that other hyperboHc functions give a better fit to observed growth rates (Bader, 1982). Monod's result was derived for cultures grown on simple monomers such as glucose, which some early models assumed to be the predominant form of substrate (e.g., Bratbak and Thingstad, 1985). As models have come to consider other forms of substrate the basic formulation has been retained, although experimental evidence of its applicability is limited.
730
Christian and Anderson Table III Models of DOM Cycling and l\imover Model
Bacterial model
Parameters
Fasham et al, 1990
FDM (see text)
v^ = 2.0 day"^ A:=0.5AtMN
Billen and Becquevort, 1991
Monod fn DOC
coQ = 0 . 3 v^ = 4.3 day~^ K=0.%ixMC
Connolly and Coffin, 1995
Monod fn DOC VB = 5.0 day-i
Semilabile turnover
Parameters
—
—
Monod fn DOC
z;s = 6 day~^ K=20S^iMC
Monod fn DOC
coc = 0.2 i^s = 10 day~^
UV
—
^=20.8A6MC
Kawamiya et al, 1995 Six and MaierReimer, 1996
—
Anderson and Williams, 1998
CN stoichiometry
Levy et al, 1998
FDM
Temperature dependent rate
0.03 day-i at 0°C
Monod fn phosphate
vs = 0.025 day-^ K=0.5fiMP
coc = 0.27 UB = 3.6 day~^ /i:=25/iMC
Monod fn DOC
Vs = 4.0 day~^ K=411 fiMC
= 2.0 day~^ K=0.5 fiMN
fixed rate
vs = 1.0 year~^
—
VB
Bissett et al, 1999
Min[C, N terms], C:Monod, N:after FDM
UB = 2.0 day-i /s:=130/xMC
Walsh et al, 1999
Monod fn DOC
a>C = 0.5 VB = 1.6day~^
Tian et al, 2000
Monod fn DOC
Vallino, 2000
CN stoichiometry
—
— Y
Monod fn DOC
Vs = 0.6 day~^ /s:=0.83/iMC
Y
= 0.5 day~^ K=l2.5fiMC
VB
Fixed rate coc = 0.804 VB = 40.0 day-^ ^ = 48.8AtMC
0.128 day-i
Note."\JW" ir\dicates models that include photolysis a n d / o r photomineralization. coc, carbon gross growrth efficiency; VQ, maximum bacterial growth rate (in some cases may be the product of maximum uptake rate and gross grow^th efficiency); vs, maximum semilabile uptake rate; K, half-saturation constant.
731
Modeling DOM Biogeochemistry
The basic formulation of Monod has been extended in ecosystem models to address the simultaneous use of organic and inorganic nitrogen (Fasham et al, 1990; Ducklow, 1994). The N-based model of Fasham et al (1990) defines a ratio of inorganic to organic nitrogen uptake for balanced growth, given by ^=
<^c^0M
;
T— - 1,
r^i
[2]
where cOx is the growth efficiency for carbon or nitrogen, and ^DOM ^^'^ ^B are C/N ratios of DOM and bacteria, respectively. This model assumes that if there is sufficient NH4'^, dissolved inorganic nitrogen (DIN) and DON are taken up in fixed ratio (^ = 0.6). If not, DIN and DON jointly limit the bacterial growth rate. This model does not include carbon and does not consider the dependence of NH4"^ excretion on substrate C/N; excretion occurs at a constant biomass-specific rate (0.05 day-i). Excretion of nitrogen by bacteria is thought to decrease at high substrate C/N ratio (Goldman et al, 1987) as N is conserved for growth and respiration costs are met using C-rich substrates. Net nitrogen excretion, £^B, can be described by the following expression (Goldman et al, 1987; Anderson, 1992; Goldman and Dennett, 2000):
E. = uJ-^-'^),
[3]
where Uc is DOC uptake. A negative E^ requires ammonium uptake to supplement DOC as a growth substrate. Net excretion occurs only at low C/N, and supplementation of DON by anmionium occurs only if there is insufficient DON to meet N demand. If sufficient C is available then no net excretion of N occurs. Depending on the C/N of DOM, either C or N is predicted to limit growth; above a threshold C/N ratio OB/COC, N is Hmiting. At high C/N there may be insufficient NH4^ to permit full utilization of DOC for growth, and DOC can accumulate (Anderson and WiUiams, 1998; see also Thingstad et al, 1997). Net regeneration of NH4"^ is predicted to occur only when nonnitrogenous C sources are scarce and N-rich DOM is being utilized. Simple relationships between regeneration of ammonium by bacteria and respiration may not occur (Ducklow, 1994). Values of co vary considerably among models (Table III), typically between 0.25 and 0.50. A recent review of bacterial growth efficiency in the open ocean indicates a mean value of 0.15 for carbon (del Giorgio and Cole, 2000). The model of Vallino et al (1996) provides a more mechanistic basis for modeling bacterial respiration and growth efficiency, but ecosystem models to date have not generally moved beyond assuming constant values. Blackburn et al (1996) used a two-component model of bacterial respiration, with a constant fraction of DOC
732
Christian and Anderson
uptake respired in addition to a constant mass-specific basal metabolism. Values of the half saturation constant for labile DOC uptake also show marked differences between models. Moreover, there is evidence that bacteria physiologically adapt to changing substrate levels, e.g., by altering the maximal uptake rate (Kirchman et al, 1993b, 1995), a phenomenon not currently considered in models. A mathematical treatment of the mechanisms underlying such adaptation is given by Button (1998). A slightly different approach to dual element modeling was taken by Bissett et al (1999a). The equations of Fasham et al (1990) were used to define rates of DON and NH4^ uptake, except that the value of ^DOM (and therefore r]) was variable, and a separate Monod function was defined for DOC. The rate of carbon uptake was then set to be the minimum of the carbon- and nitrogen-limited rates of carbon uptake. If the former, nitrogen uptake was adjusted downward to achieve balanced growth. DIN uptake was reduced first, and if DON uptake was in excess once DIN uptake was eliminated, the "excess" N was transformed into NH4'^, permitting remineralization of N when the ambient substrate pool was N rich. The model of Thingstad et al (1997, 1999) indicates that labile DOC may not be consumed rapidly by bacteria because of a "malfunctioning microbial loop." These models contain a steady-state representation of the bacteria-phytoplanktonphosphate-flagellate system, which is subject to external grazing pressure from ciliates and higher predators. Biomass of HBAC can be limited by a combination of nutrient stress and predation so that an increase in the total nutrient content (and therefore phytoplankton biomass and production) of the system can result in accumulation of labile DOC (Thingstad et al, 1997). Lags between peaks of primary and secondary (bacterial) production during blooms can be explained even if the DOM is labile (Thingstad et al, 1999). This work provided a theoretical and experimental demonstration of how a combination of predation and mineral nutrient limitation, rather than the availability of organic substrates, may control bacterial production and thus consumption of labile DOM. 3. Photochemical Effects The photochemistry of DOM has been an area of increasing interest in recent years, although most models do not contain such processes. The basic equations of aquatic photochemistry have been reviewed by Miller (1998) and Mopper and Kieber (Chapter 9) and will not be repeated here. The absorption spectrum of absorbing or "colored" DOM (CDOM) has a negative exponential shape, increasing monotonically toward shorter wavelengths. Most of the photochemical reactivity is in the ultraviolet B range (UVB, approximately 280-320 nm). A number of models exist that describe photochemical production of various compounds from DOM (e.g., Sikorski and Zika, 1993; Gnanadesikan, 1996; Preiswerk
Modeling DOM Biogeochemistry
733
and Najjar, 2000), but are not discussed here as they do not explicitly model the DOM pool itself. For a general treatment of the theory of modeling meteorological and solar forcing of photochemistry in the upper ocean see Doney etal (1995). Anderson and WilHams (1999) included refractory DOM (RDOM) as one of three fractions in their model. This RDOM was not subject to direct bacterial oxidation, but was photooxidized to labile DOM which was available to bacteria. Using a photochemical breakdown rate at the ocean surface (ofo) of 0.0015 day~^ and an attenuation coefficient for UVB (A:uv) of 0.33 m ~ \ they found that a steady state was attained with an RDOM production rate of 0.35% of labile and semilabile DOM utilization by bacteria. The depth-integrated rate of photooxidation in this one-dimensional model can be approximated as aoRo/K^ (their Eq. [7]), where Ro is the concentration of RDOM at the ocean surface. This model suggested that a 10% increase in UV radiation at the ocean surface would decrease the global ocean stockof RDOM by less than 1% over 200 years. Klepperetal (1994) estimated that increased UV-induced DOM photolysis would decrease ocean carbon storage in 2070 by more than 25% relative to their baseline simulation of no change in ocean circulation or biogeochemistry with CO2-induced climate change. This effect was the largest of the seven "feedback" terms that they quantified, but was considered the most uncertain. Few details about the model are given, so it is difficult to evaluate these conclusions. The model incorporates in at least rudimentary form the ocean's overturning circulation, which is impossible in the type of model employed by Anderson and Williams (1999), and is highly relevant to the question of global DOM photooxidation rates. Bissett et al. (1999b) coupled a complete ecosystem model to a spectral model of inherent and apparent optical properties, by specifying a certain fraction of the DOC produced within the ecosystem model as "colored" and subject to photochemical reactions. Estimating these fractions is difficult, and their values in nature are largely unknown (Nelson and Siegel, Chapter 11). Bissett et al. (1999b) had observations of spectral attenuation coefficients with which to compare the model output, although their model is very complex and difficult to constrain. Modeled values of CDOM absorption in the Sargasso Sea were maximal at 60-80 m depth in autumn, which is consistent with observations (Siegel and Michaels, 1996). Ratios of the downwelling attenuation coefficients at 412 and 487 nm, which is an approximate index of attenuation due to CDOM relative to phytoplankton (Siegel and Michaels, 1996) were in the range of 1-1.2 and were generally lower than observed values except at the surface in summer. This may imply low rates of photooxidation or it may be the modeled seasonal cycle of DOC itself that is in error, rather than the photochemical model. This model was also the first to our knowledge to consider direct photomineralization to CO2.
734
Christian and Anderson
III. MODELING THE ROLE OF DOM IN OCEAN BIOGEOCHEMISTRY A. MAJOR CLASSES OF OCEAN MODELS At basin to global scale, ocean models generally belong to one of two classes: ocean general circulation models (OGCMs) and box-diffusion models. The former represent the integration of the equations of motion, or a partially linearized approximation, on a grid whose resolution is generally on the order of 1-3° for basin- to global-scale models. The latter divide the ocean up into a number of boxes that exchange heat and chemical tracers according to specified rates of advective exchange and mixing coefficients ("eddy diffusivities"). The processes that are represented by these coefficients are also relevant to OGCMs, as the grid is too coarse to represent mixing processes explicitly, and the results are frequently highly sensitive to how these "sub-grid-scale" processes are represented (e.g., Danabasoglu et ai, 1994; McWilliams, 1996). An important innovation in box-diffusion models has been the inclusion of an "outcrop" box in the high latitudes that extends from the surface to greater depths than the surface boxes in lower latitudes (e.g., Siegenthaler and Joos, 1992). This permits rapid exchange of oxygen and carbon between the atmosphere and the low-latitude deep ocean via the high latitudes, as these fluxes occur in the ocean primarily along isopycnal surfaces. B. EARLY RESULTS: EQUATORIAL NUTRIENT-TRAPPING Experiments with both OGCMs and box-diffusion models were conducted in the wake of the "discovery" of HTCO-DOC to assess the effects of DOC production on the large-scale distribution of nutrients, oxygen, and DIG (e.g., Bacastow and Maier-Reimer, 1991; Najjar et al, 1992; Paillard et ai, 1993). It is important to note that not all of these simulations employed prognostic biological models: in many cases they simply estimated "new production" (NP, which is actually net community production) either by restoration of the model nutrient fields to climatological values as these are altered by upwelling and advection (Najjar et al, 1992) or as a function of nutrient concentration (Paillard et al, 1993) or nutrient concentration and irradiance (Bacastow and Maier-Reimer, 1991; Matear and Holloway, 1995). This NP is then redistributed downward to simulate sedimentation, usually according to the hyperbolic expression of Martin et al (1987). Early experiments with DOM assigned some fraction of NP to the DOM pool rather than to the sedimentation flux, allowing it to be mixed and advected in the same fashion as dissolved inorganic nutrients. Bacastow and Maier-Reimer (1991) assigned a decay rate for DOM of 0.02 year~^ (turnover time, r = 5 0 years). Najjar et al (1992) did not specify
Modeling DOM Biogeochemistry
735
r, but their model gives values of the same order. This results in a significant penetration of DOM into the mesopelagic and alters the vertical distribution of nutrients and oxygen relative to particle-only simulations. An important component to these experiments was the search for solutions to the equatorial "nutrient-trapping" problem. It had been found that sedimentation-only biogeochemical models produced large accumulations of (inorganic) nutrients and depletions of oxygen in the equatorial thermocline that were not consistent with observations (Toggweiler, 1989). Assigning some of the NP to DOM rather than to the sedimentation scheme (which results in remineralization directly below the point at which the particles are formed) seemed to remedy this problem, resulting in nutrient maxima that were deeper, weaker, and less restricted to the equatorial region (Toggweiler, 1989; Bacastow and Maier-Reimer, 1991; Najjar et al, 1992). These simulations employed OGCMs with horizontal resolution on the order of 3 ^ ° , which is inadequate for accurate simulation of the equatorial current system. These early results have been called into question on the grounds that artificialities in the circulation fields are as likely an explanation for the nutrienttrapping problem as the use of particle-only biogeochemical models. Matear and HoUoway (1995) used a simple POM-based biogeochemical model similar to that used by Bacastow and Maier-Reimer (1991) and an adjoint technique that allows particular model parameters or fields to vary in order to force selected fields closer to observations or other a priori constraints. By retaining the circulationfieldsof the basic model but allowing the new production rate at a given nutrient concentration and remineralization length scale (RLS) for sinking particles to vary, they found that an increased RLS and reduced NP could alleviate nutrient-trapping without an explicit DOM component, but not with reaHstic values of these terms. By retaining the base values of the biogeochemical parameters but allowing the circulation fields to vary, they found that small changes in circulation could achieve the same result and concluded that uncertainties about the modeled circulation were too large for definitive conclusions to be drawn regarding the relative roles of DOM and POM. More recent experiments by Aumont et al. (1999) have shown that increasing the resolution of the circulation model largely eliminates nutrient trapping without any change in the biogeochemical model. Note that the authors of the original studies appear to have been well aware that this might prove to be the case (Toggweiler, 1989; Najjar and Toggweiler, 1993). A more extensive discussion of the effects of advection schemes and grid resolution on biogeochemical models is given by Oschlies (2000).
C. D O M TURNOVER TIMES The decay rates employed in the early simulations of Bacastow and MaierReimer (1991) and Najjar et al. (1992) were called into question by Archer et al.
736
Christian and Anderson
(1997) and Yamanaka and Tajika (1997), using HTCO-DOC data not available at the time of the earUer experiments. Using a steady-state model of DOC production and consumption and an OGCM simulation of the tropical Pacific circulation, Archer et al (1997) defined a "grow in" time scale for semilabile DOC. This model reduced the production and remineralization rates to a single value (based on the observed differences between surface and deep concentrations), assuming that oligotrophic surface waters with the highest observed DOC concentrations are near steady-state with respect to production and consumption of DOC, while recently upwelled waters are not. This simple model generated optimal values of r between 30 and 120 days, suggesting that the bulk of semilabile DOC is not nearly as refractory as assumed in earlier studies. Yamanaka and Tajika (1997) used a slightly more complex model to estimate the ratio of DOC to POC produced ("production ratio") and the decay rate of semilabile DOC. They found that r's of 0.3-1 year, and production ratios of 1-2 were the most consistent with observed surface concentrations and penetration depths of semilabile DOC. These values increase and decrease, respectively, by about a factor of 2 if POM solubilization produces labile rather than semilabile DOM. An innovative aspect of this analysis is that while either the observed surface concentrations or penetration depths generate a range of approximately equivalent solutions, the two sets of observations provide orthogonal constraints (in terms of positive or negative correlation between turnover time and production ratio for statistically equivalent solutions). Only a narrow range of solutions are consistent with both. The authors note, however, that estimates of the penetration depth are imprecise, and that more observations in the 1(X) to 4(X) m depth range would make these solutions more robust. These estimates of the mean Ufetime of semilabile DOM are also consistent with the ID model results of Anderson and Williams (1999); these authors found the poorest fit to data in the 200 to 400 m depth range. It has also been shown that alleviation of nutrient trapping is possible with turnover times in this lower range (Anderson and Sarmiento, 1995), using the same circulation fields employed by Najjar ^rtz/. (1992).
D.
DOM
AND ATMOSPHERIC
CO2
There are several examples of global ocean box models applied to questions about atmospheric CO2. Models of ocean-atmosphere CO2 exchange on glacialinterglacial time scales tend to underestimate the glacial-interglacial difference (ACO2, ~100 ppm). Paillard et al. (1993) attempted to determine whether including DOM in the ocean model would alleviate this, but found that it actually increased the discrepancy (decreased ACO2 in the model) because it reduced the overall geochemical stratification of the model ocean, i.e., less transport of carbon to the deep ocean. Decreasing the turnover time of the DOM would bring ACO2
Modeling DOM Biogeochemistry
737
more in line with the particle-only model, while increasing it tends to further "smooth out" the glacial-interglacial cycle. The biogeochemical model resembles those of Bacastow and Maier-Reimer (1991) and Najjar et al (1992) in that DOM was assumed to have Redfield C/N/P ratios and a Ufetime of order 100 years. Keller and Goldstein (1995) used a variant of the model of Siegenthaler and Joos (1992) to assess the long-term consequences of a pulse of CO2 into the atmosphere or an injection into the oceanic thermocline. They found that at steady state (the simulations ran for 1500 years) a negUgible fraction (0.21%) of the "excess" carbon was found in the DOC pool. Sensitivity studies showed that an increased upwelling velocity caused a significant increase in the total DOC, due to increased input of nutrients to the surface layer. They assumed Redfield stoichiometry for DOM and modeled remineralization as a first-order process with a turnover time that appears to have been on the order of 100 years. Klepper et al (1994) used a box-diffusion model to analyze biogeochemical feedbacks to rising atmospheric CO2, but provide so few details that it is difficult to determine what if any effect DOM had on their results.
E. DIAGNOSTIC MODELS OF THE NORTH ATLANTIC
Schlitzer (1989) constructed a simple model of the North Atlantic Ocean and estimated values of new production, particle flux, and air-sea exchange of CO2 in each box by fitting the model to historical observations of nutrients, oxygen, DIG, and alkalinity. He then attempted to add a DOC component to the model based on the very Hmited data available at the time, by (1) using pre-HTGO values, (2) using values derived from the Sugimura and Suzuki (1988) Pacific data, with their surface concentrations in the first model layer {GQ < 25.5) and their deep concentrations in the other layers, and (3) calculating DOC from apparent oxygen utilization based on the correlation reported by Sugimura and Suzuki (1988). The result of this experiment was that with the pre-HTGO DOG concentrations, solutions could be derived that were consistent with the range of data constraints available, whereas when the results of Sugimura and Suzuki (1988) were used this was not possible. In experiment (2), the distributions of total G, N, and P were "incompatible with the circulation pattern of the model and with [North Adantic Deep Water] formation rates greater than 5 Sv" (Schlitzer, 1989, p. 12,791). The method used allows the circulation to change to give the best fit of the model to biogeochemical fields, but the changes required in this case were outside acceptable bounds. In experiment (3), the optimal solution was better but still unacceptable; new production was much less than the smallest literature values, and diapycnal mixing coefficients were negative in some places. Note that this experiment was conducted considerably before Suzuki's (1993) retraction of his early results. As a cautionary note, however, like other early modeling efforts in the
738
Christian and Anderson
"post-HTCO" era, Schlitzer (1989) appears to have assumed large pools of DON and DOP as well as DOC. Walsh et al (1992) also attempted to estimate meridional fluxes of DOM in the North Atlantic from a handful of early (and probably erroneous) HTCO-DOC estimates, using zonally integrated transport estimates for various depth strata (cf. Rintoul and Wunsch, 1991). They too found mass-balance difficulties, such as ratios of carbon and oxygen flux that were far out of Redfield ratio. The concentration estimates are generally high, but if this is attributed largely to blank-correction problems (Suzuki, 1993) the net meridional transport estimates may not be entirely fanciful. They estimated the net (southward) flux of DOC as about \~^ X 10^^ mol year~^ which would imply that DON could not balance the apparent poleward transport of DIN estimated by Rintoul and Wunsch (1991) with reasonable C/N ratios.
E BASIN-TO-GLOBAL SCALE ECOSYSTEM MODELING None of the studies discussed above employed prognostic biological models (e.g., state variables for phytoplankton and zooplankton). Prognostic ecosystem models including a DOM component have been coupled to OGCMs in the North Atlantic (Fasham etal, 1993; Sarmiento etal, 1993), the North Pacific (Kawamiya et al.y 2000), the equatorial Pacific (Toggweiler and Carson, 1995), the Arabian Sea (Ryabchenko et al, 1998), and, in only one case that we are aware of, globally (Six and Maier-Reimer, 1996). Not all of these models included an explicit population of heterotrophic bacteria; Six and Maier-Reimer (1996) and Kawamiya et al. (2000) used concentration-dependent rate equations for remineralization of DOM. In the simulations of Fasham et al (1993) and Sarmiento et al (1993), DON constituted a small fraction of total N in the North Atlantic (~0.01 /xM N, much less than particulate detritus or plankton biomass), reflecting the fact that the ecosystem model employed (Fasham et al, 1990) simulates only labile DON. Fasham et al (1993) noted that at both Bermuda (subtropical) and Ocean Weather Station "India" (OWSI, subarctic) about half of the annual supply of DON came from breakdown of detritus. Although this result is sensitive to parameter choices, it was not anticipated by the authors of the model. This source was most dominant below about 40 m depth. Convective overturning was the dominant process for removal of DON from the euphotic zone (more than an order of magnitude greater than vertical mixing or downwelling when averaged over the model grid), consistent with observations collected near Bermuda (Carlson et al, 1994; Hansell and Carlson, 2001). Convective losses were much greater at Bermuda than at OWSI, although annual mean surface concentrations were quite similar. This may reflect the lack of production in winter in the subarctic and periodic restratification in
Modeling DOM Biogeochemistry
739
winter in the subtropics, but caution is required in interpreting these results as the OGCM may overestimate the vertical exchange of nitrogen by wintertime convection (McGillicuddy et al, 1998). Kawamiya et al. (2000) simulated realistic concentrations of DON (6-8 /xM) in the tropical Pacific Ocean. Their model included a prognostic phytoplankton and zooplankton model; DON consumption was parameterized by a first-order remineralization term. Their model equations indicate a temperature-dependence of this rate, but the temperature-dependence parameter was assigned a value of zero, implying a constant specific remineralization rate of 0.01 day~^. Solubilization of particulate organic nitrogen (PON) was temperature-dependent with a base rate of 0.05 day~^ and rates of ~0.3 day~^ at the temperatures characteristic of the tropics. A simple calculation gives the steady-state concentration of DON as a function of temperature, PON concentration, and the rate of production of DON by processes other than solubilization; the modeled values of 6-8 /xM are consistent with values of these expected for tropical-subtropical surface waters. The specific rate of PON solubilization is consistent with measurements of ectoenzyme activities on particles (Smith et al, 1992), and calculations based on fluxes of particles collected in sediment traps (Christian et al, 1997) or on thorium disequilibria (Mumane, 1994). Kawamiya^f al. (2000) compared their modeled DON distributions with data collected by Libby and Wheeler (1997) between 10°S and 10°N and about 95-140°W, noting the model's reproduction of a minimum at the equator and a zonal gradient (increasing westward) north of the equator, which they attributed to greater mixed layer depths in the central tropical Pacific. Six and Maier-Reimer (1996), like Kawamiya et al. (2000), employed prognostic phytoplankton and zooplankton models, but did not model bacteria and used first-order remineraHzation of DOC, with a maximal rate of 0.025 day"^. The actual rate was a function of inorganic nutrient concentration; i.e., it was assumed that nutrient limitation of HBAC would reduce the rate of remineralization (see also Thingstad et al, 1997). They did not include solubilization of POC as a source of DOC. Like Kawamiya et al. (2000), their simulated DOC concentrations were within the range of observed values (15-40 /xM at the surface, not counting the refractory "background" fraction). The highest concentrations were in the summer hemisphere, with strong enrichments to ~60° latitude in the sunmier months and maximal concentrations around 20°. DOC dominated the poleward transport of organic matter that balances equatorward transport of mineral nutrient; these transports were maximal at 10-15°N or S. These simulations showed little seasonality of surface DOM concentrations at latitudes less than about 20°, while at higher latitudes most of the surface-layer DOM appears to be turned over on annual time scales. These results could be highly sensitive to the choice of temperature or nutrients as the factor determining the remineralization rate, as these are negatively correlated in surface waters at large space and time scales.
740
Christian and Anderson
G. INVERSE MODELS OF FLOWS WITHIN FOOD WEBS An "inverse method" is a statistical method for fitting a model to data, and can be applied to both diagnostic (e.g., Vezina and Piatt, 1988; Jackson and Eldridge, 1992) and prognostic (e.g., Matear and Holloway, 1995; Spitz etaly 1998; Fasham etaly 1999; Vallino, 20(K)) models. Fitting of a Hnear, steady-state model of aplanktonic food web was described by Vezina and Piatt (1988), and applied to data sets collected in the English Channel and the Celtic Sea. This model estimates the flows of matter and energy among the model compartments (e.g., phytoplankton, zooplankton, heterotrophic bacteria) that best fit the data in a least-squares sense, subject to specified constraints (e.g., an upper limit to the fraction of gross photosynthesis that can be exuded as DOC), but does not specify the mathematical form of relationships among these compartments. This methodology has since been applied to data collected in various regions of the ocean (e.g., Jackson and Eldridge, 1992; Donali et ai, 1999; Vezina et al, 2000), as well as freshwater (Vezina and Pace, 1994) and sea ice (Vezina et ai, 1997) ecosystems. The solutions derived by Vezina and Piatt (1988) suggest that a majority of DOC (~65%) came from heterotrophs (DON was assumed to come only from heterotrophs). A substantial fraction (11-19%) of DOC came from the heterotrophic bacteria. There were no a priori constraints placed on this flow (whereas other groups had upper limits set on the fraction of energy intake lost to DOC), so it is not surprising that in the optimal solutions this flow would have a positive value, and little can be concluded with certainty from it. The upper limit to detrital dissolution was set at 1% day"^ Jackson and Eldridge (1992), applying the method to data collected in the Southern California Bight, discarded this constraint and found that this rate was about 6% day~^ for their data set. The flux of carbon to bacteria from DOM in the solutions of Vezina and Piatt (1988) was 16.2 nmiol C m"^ day~^ in the English Channel and 17.5 mmol C m~^ day~^ in the Celtic Sea. In the English Channel this is equivalent to 0.56% of the total DOC pool each day, or 36-40% of net primary production. In the Celtic Sea, the estimatedfluxof carbon to bacteria exceeded measured bacterial production by a factor of 21, implying a very low growth efficiency. In both systems, the calculated C/N molar ratio of the DOM consumed by bacteria was 14, and uptake rates of DON and NH4"'" were similar. In the Gulf of Riga, Donali et al. (1999) found that bacterial carbon demand significantly exceeded DOC supply in spring and autunm. Because the calculation takes account only of the mean flows during the period of the field observations, such imbalances do not violate mass balance constraints, but imply that a large amount of DOC (which was not measured) was present in the water prior to the cruises. These authors attributed this DOM to phytoplankton production rather than terrestrial input. In the subarctic Pacific, Vezina and Savenkoff (1999) calculated flows for cruises in September, February, and May. In May, steady-state solutions could not be derived, i.e., changes in the mass of some compartments over the
Modeling DOM Biogeochemistry
741
course of the cruise were necessary. They calculated a net DOC accumulation on the order of 100 mg m~^ day~^ which would imply a seasonal accumulation of ~10 fjM. As with the data sets employed by Vezina and Piatt (1988), heterotrophs dominated DOC production in all seasons, with the highest autotrophic input in spring. Heterotrophs also dominated DOC production in the Gulf of St. Lawrence, which was approximately half of gross primary production during winter-spring, with protozoa accounting for the largest fraction (Vezina et al, 2000).
H. COASTAL AND ESTUARINE SYSTEMS Coastal systems are not our area of expertise, and we have not attempted to cover exhaustively work in this area, especially that which is primarily concerned with anthropogenic DOM (e.g., Yassuda etal, 2000). The principles of modeling DOM in coastal systems are essentially the same as in pelagic systems, except that one must consider allocthonous (e.g., fluvial) sources of DOM, and exchange of organic matter at the water-sediment interface. Some ecosystem models take account of the DOM flux from sediments in some fashion (e.g., Walsh and Dieterle, 1994). Decomposition of fluvial DOM has been modeled using a variety of approaches, few of which take explicit account of the biochemical composition of this material. Hopkinson et al. (1998) defined equations relating measurable quantities such as elemental composition and molecular weight to the chemical composition (e.g., aromaticity) and degree of carbon reduction of terrestrially derived DOM, which are expected to be correlated with bacterial growth rate and efficiency (Vallino etal, 1996). An important issue in oceanography is the lateral transport of autocthonously produced (plankton source) DOM, as well as inorganic nutrients and DIC, from coastal regions to the open ocean (e.g.. Pace et al, 1984; Walsh et al, 1997; Tusseau-Vuillemin et al, 1998). Walsh et al (1997) used a Lagrangian model to estimate fluxes from the coastal zone to the open ocean, concluding that there are substantial fluxes of DOM from the Bering and Chukchi seas to the Pacific and Arctic oceans. Tusseau-Vuillemin et al. (1998) found that the continental shelf of the Gulf of Lions was a source of nitrate to the open Mediterranean in winter but a sink for oceanic DIN for much of the rest of the year. The shelf sink for DIN in sunmier may imply an export of DON, and this model could in principle be used to quantify fluxes of DOM (including fluvial DOM) from the margin to the open ocean. The magnitude of this flux, however, depends on exchanges at the sediment-water interface that the authors describe as "roughly parameterized." Cifuentes and Eldridge (1998) developed a simple model of DOC dynamics along estuarine salinity gradients, which they used to identify additional allocthonous (e.g., wetlands) and autocthonous (e.g., phytoplankton) sources of DOC
742
Christian and Anderson
when the model deviated from patterns expected for the central processes of mixing, advection and decomposition of fluvial DOM. These authors provide a useful analysis of the relationship between DOC decomposition time scales and estuarine mixing and advection time scales, noting that the behavior of DOC in strongly mixed estuaries will be quite different than in those where advection dominates. Accurate identification of nonfluvial sources will likely improve with improved models of mixing and DOC decomposition.
I. SMALL-SCALE SPATIAL STRUCTURE Virtually all of the models cited thus far treat the plankton conmiunity as homogeneous in space on scales ranging from a few meters to hundreds of kilometers. The interaction between bacteria and their substrates takes place on viscous scales orders of magnitude smaller. The basic concepts for understanding how organisms function on these scales have been reviewed by Jackson (1987) and Jumars et al (1993). Significant results from these studies are that (a) oceanic turbulence does not enhance the flux of substrate to cells in the bacterial size range over that resulting from molecular diffusion alone, (b) motile bacteria "swim" along a biased random walk trajectory rather than a consistently up or down gradient path, and (c) the minimum size of phytoplankton cells that can be "found" by chemotactic bacteria is 2-5 /xm. Several investigators have addressed the question of whether bacteria can gain energetic advantage from chemotactic "clustering" around phytoplankton cells "leaking" DOM. Jackson (1989) addressed this question in relation to laminar sinking of phytoplankton cells, concluding that only in conditions where cells were large, abundant, and leaky did chemotaxis appear to confer significant energetic advantage for nonattached bacteria, with the caveat that the effects of turbulent fluid motion needed to be assessed before this question could be resolved. Bowen et al (1993) showed that there is a small but significant gain from chemotaxis under reahstic conditions of oceanic turbulence. Individual cells do not remain in a particular cell's halo for long except under the most quiescent conditions, and the fraction of cells found within enriched microzones is small, but chemotactic cells spend enough time within these microzones on average to derive a significant energetic advantage. Blackburn et al (1997) addressed the question of chemotaxis with a microbial food web model that was spatially structured (70 x 70 grid cells) but purely viscous (molecular diffusion only). Rather than exudation, they treated "events" of protozoan predation or cell lysis as the source of DOM. A chemically homogeneous and presumably labile pool of DOM was employed, whose concentration remained low (much less than plankton biomass). DOM concentration varied by more than two orders of magnitude in both time and space, although the total
Modeling DOM Biogeochemistry
743
volume simulated was less than 1 mL. This experiment provides an additional qualitative confirmation of the value of chemotaxis and shows that the aggregate behavior of the spatially structured model differs significantly from that of a homogeneous version of the same food web model. The spatial scales and diffusivities employed are at (or perhaps beyond) the limits of the viscous assumption, so the results should be taken as an illustration and not at as a realistic simulation of the "invisible world." An important aspect of the Blackburn et al. (1997) study is that they addressed temporal as well as spatial variability of DOM production, whereas Jackson (1989) and Bowen et al. (1993) treated the phytoplankton cell as a continuous point source of DOM. In all of these experiments it is assumed that autotroph or micrograzer cells from which DOM is generated are much larger than bacteria (e.g., 20 /xm). The "encounter rate" of bacteria with the enriched "microzones" scales linearly with the number of such microzones but with the third power of their diameter (Jumars et al, 1993), so results calculated for a relatively small number of large microzones can not necessarily be extrapolated to oligotrophic waters with a larger number of smaller ones.
IV. DISCUSSION AND CONCLUSIONS Over the past two decades there have been considerable advances in the methodologies of both seawater chemistry and numerical ocean modeling. The relatively small number of high-quality observations, as well as the fundamental weakness of our understanding of interactions within microbial conmiunities (Nagata, 2000) and of the physiology of heterotrophic bacteria (Kirchman, 2000), limit what can be achieved with numerical models. For example, few estimates of the kinetic parameters defining degradation of semilabile DOM are available, and it is questionable how reliably it is possible to simulate this process in models. Estimates of the decay time scale of semilabile DOM in early models (e.g., Bacastow and Maier-Reimer, 1991; Najjar et al, 1992) are much longer than those in more recent studies (e.g.. Archer et al, 1997; Yamanaka and Tajika, 1997). The rate of attenuation of DOM concentration with depth is important for biogeochemical cycling and model validation; more observations in the thermocline and mesopelagic zone (e.g., 100-500 m) would be useful. The question of what are the optimal biological structures for use in large-scale biogeochemical modeling studies is a subtle one. The one component which appears to be required is a semilabile DOM fraction, which provides a significant contribution to export flux in many areas. But what about labile and refractory fractions and indeed heterotrophic bacteria? One important argument for the inclusion of heterotrophic bacteria and the whole microbial loop is that these organisms provide the enzymes for hydrolysis of macromolecular and even monomeric
744
Christian and Anderson
DOM. Without explicit treatment of bacteria it is necessary to resort to empiricism to model DOM turnover. However, this apparent advantage must be weighted against the reliability with which bacteria can be simulated, as well as our ability to mechanistically parameterize semilabile DOM turnover. Other reasons for explicit treatment of the microbial loop in models are that it may provide a link between the microbial and metazoan food webs and that bacteria compete with phytoplankton for inorganic nutrients (e.g., Bratbak and Thingstad, 1985). The question of whether to include refractory DOM in models would appear to be a matter of time scales. Anderson and Williams (1999) examined the response of this pool to increased UV and concluded that it was so slow that it may not be necessary to include refractory material dynamically in models for examining climate change within the next 200 years. Processes by which DOM is created are represented in models in varying ways. In models that do not include explicit DOM state variables, terms representing exudation, respiration, and lysis and other forms of nongrazing mortality can have essentially identical mathematical form, although the choice of processes considered and terminology used to describe them varies. Here we have shown the wide diversity that exists in the ways models that consider DOM explicitly represent cycling of organic matter. One can speculate on various causes of this variabihty— differences between systems, varying objectives, or a lack of consensus on the importance of representing different processes in models. We suggest that the last of these causes is likely to be a significant source of model variability, highlighting the need for further process studies and improved models. Determining model sensitivities to choices of processes and parameters is a necessary first step, which requires data relative to which the models' sensitivities can be adequately assessed (e.g., relatively complete seasonal cycles of DOC concentration, direct measurement of rate processes such as exudation). Ultimately, evaluation of different model formulations' performance relative to common data sets will be required. One of the earhest marine ecosystem models to consider DOM assumed that all organisms produce DOM (Pace et ai, 1984). Since then there have been both models that assumed just one or two processes were responsible, and models that specified a variety of different processes (Table II). The recent models of Anderson and Williams (1998), Walsh et al (1999), and Tian et al (2000) have incorporated a broad spectrum of processes. The mathematical formulation of these terms remains tentative and speculative, and the sensitivity of the models with respect to the formulation of these various sources need to be assessed carefully. Analysis of carbon flows within food webs seems to confirm that heterotrophs, not autotrophs, account for the largest fraction of DOC production. This conclusion needs to be examined for biases resulting from the structure of the food web models and the data to which they are fitted, but is consistent with other results and should be taken seriously by developers of ecosystem models.
Modeling DOM Biogeochemistry
745
Differences in terminology between studies also confound model comparison. The distinction between what is dissolved and what is particulate is not necessarily clear (Sharp, 1973). From a biogeochemical modeling perspective, the distinction between dissolved material and suspended particles with neghgible sinking velocities is not necessarily meaningful, although the mechanisms and rates of utilization by bacteria may differ (Joint and Morris, 1982; Kepkay, 2000). Bissettetal. (1999a) gave three of their four fecal pellet fractions zero sinking velocity, but treated these differently than DOM and did not make them available to HBAC. Tian et al (2000) considered only a single "dissolved" organic fraction and two of "particulate" detritus, but the smaller of these did not sink. The inclusion of additional fractions may stem from a desire to include a reasonably complete representation of the range of microbial processes. Vertical transport is a key issue in biogeochemical cycling, so more complete tests of these various formulations' performance are desirable. There has been progress on the "demand" side (modeling the utilization of DOM by HBAC), for example, the models of Anderson and Williams (1998), Bissett et al (1999a), and VaUino (2000) allow for both C- and N-limited growth and net uptake or remineralization of DIN. Spitz et al (2001) found that the fit of their model (based on that of Fasham et al, 1990) to data from the Bermuda Atlantic Time-Series station was significantly improved by including these processes. Nevertheless, the assumptions underpinning these models need further validation. A major weakness of most models is that the biochemical composition of DOM is not explicitly considered. The composition of the mixture of substrates utilized in nature is not well known, and the variable energy content of different biochemicals with similar stoichiometry is not accounted for in models (VaUino et al, 1996; Weber, 2000). An important challenge for the near future will be to model the full C/N/P stoichiometry of bacterial growth. The biochemistry and biogeochemistry of N and P are quite different (Kirchman, 2000; Karl and Bjorkman, Chapter 6), and few models have addressed these differences. Furthermore, bacteria in aquatic ecosystems are a diverse group of organisms, and so care has to be exercised when using, for example, stoichiometric models. The physiological capacities of dominant bacterial groups can vary seasonally (Pinhassi and Hagstrom, 2000). Nutrient additions may stimulate particular types of bacteria, changing overall community composition (Fuchs et al, 2000). Only a few models have addressed the significance of UV radiation in DOM cycling, and those have focused on increases in lability resulting from photooxidation of refractory material. While exposing refractory DOM to UV radiation can increase its lability, there is also evidence that photochemical reactions can render labile DOM less available for bacterial consumption (Keil and Kirchman. 1994; Benner and Biddanda, 1998; Obemosterer et al, 1999). Simulating the balance between these opposing effects of UV radiation on DOM labiUty presents a significant future challenge for modelers.
746
Christian and Anderson
Among the most promising recent developments is the use of optimization techniques for statistical fitting of prognostic models to data (Spitz et al, 1998, 2001; Fasham et al, 1999; Lamy et al, 1999; Vallino, 2000). These experiments have employed physical frameworks such as a mesocosm (Vallino, 2000) and a Lagrangian, drifter-tracking cruise (Fasham et al, 1999) to minimize advective effects and make use of zero-dimensional models viable. The limited information available from these experiments results from the mismatch of model complexity and the available observational data, that is, the available analytical methodologies may not be adequate to place strong constraints on many of the terms in the models. ValUno (2000) applied a variety of parameter-estimation techniques to a model of a mesocosm experiment with four experimental treatments (control, -hDOM, +DIN, and -hDOM-hDIN). The results are somewhat disconcerting: despite the extraordinary effort expended in finding the most statistically probable solutions for the model parameters, these solutions could not be generalized from one experimental treatment to another. It is therefore unlikely that the model accurately represented the mechanisms regulating the response of the microbial community to the different experimental treatments. It is simple to find fault with particular aspects of the model or the experimental treatments; it is quite another matter to demonstrate that these artificialities and not a fundamental lack of understanding of the underlying biology are responsible for the weakness of the solutions derived. There are many areas in which biogeochemical models have not yet addressed the underlying biological mechanisms. For example, the results of Thingstad (2000) suggest that the diversity of the bacterial community may represent an important control on biogeochemical cycles. Even for a single species, a more mechanistic treatment of bacterial growth is clearly possible (Vallino et al, 1996; Button, 1998). Nonetheless, the development of modeling in the study of dissolved organic matter in the oceans has advanced rapidly over the past decade. While there are many remaining uncertainties, and many of the results cited are quite tentative, a fair amount of progress has been made. A variety of sophisticated new models, have been developed in a simplified (OD or ID) physical context, and their ability to simulate spatial and temporal variability of DOM and its effects on large-scale ocean biogeochemistry will hopefully be evaluated in the near future. There is a continuing need for improved parameterizations of the underlying processes, and coupling of these to contemporary models of the ocean circulation (cf. Doney, 1999). All of the global and most of the regional simulation experiments to date have employed fairly coarse-resolution models, and conclusions regarding biogeochemistry must be received cautiously until more realistic circulation fields are employed. Equatorial nutrient-trapping and anoxicity can be eliminated in particleonly models and may be an artifact of poor simulation of the equatorial circulation (Matear and HoUoway, 1995; Aumont etal, 1999). However, the "negative result" that DOM is not needed to eliminate these artifacts cannot be taken to imply that DOM plays a minor role in ocean biogeochemistry.
Modeling DOM Biogeochemistry
747
Inclusion of DOM in models can alter the biogeochemical stratification of the ocean and the partitioning of carbon between ocean and atmosphere, decouple the C, N, and P cycles, and change the structure of food webs and the grazing pressures on phytoplankton. Quantifying these effects is difficult due to lack of data and uncertainty about model structure. Early experiments in large-scale modeling were clustered at opposite extremes, i.e., biogeochemical models with highly parameterized biology and long DOM lifetimes, and microbial food web models with a single, labile DOM pool. Both the long lifetimes assumed by the former and the low concentrations generated by the latter are probably erroneous. Proper simulation of the role of DOM in biogeochemistry would appear to require at the least that the semilabile pool be simulated, and that its lifetime is on the order of 1 year. Some recent experiments have included a single pool with a first-order remineralization rate in the semilabile range, and generated realistic concentrations and spatial and seasonal variations (Six and Maier-Reimer, 1996; Kawamiya et al, 2000). Dependence of these remineralization rates on temperature and inorganic nutrient concentration needs to be better understood, assuming that such dependence even exists. Rigorous evaluation of competing formulations against observations, combined with development more mechanistic models of the remineralization process (e.g., Vallino et al, 1996) could potentially identify weaknesses in empirical formulations and also identify situations in which they are adequate to the task at hand. Models with both a semilabile pool and an explicit HBAC population have generated realistic depth profiles (Anderson and Williams, 1999), but need to be more extensively tested with regard to temporal variations. To understand biogeochemical fluxes on interannual to interdecadal time scales, more data in the mesopelagic zone are required, along with a better understanding of why some models fit the observations poorly in this region. Fluxes between the coastal zone and the open ocean are also poorly characterized; improved models of exchanges at the sediment-water interface and of decomposition of terrestrially derived DOM need to be explored.
ACKNOWLEDGMENTS The authors acknowledge support from the NSF Biological Oceanography program, the NASA Ocean Biogeochemistry program, and the Natural Environment Research Council, UK. Hugh Ducklow, Alain Vezina, Ray Najjar, and two anonymous reviewers made helpful comments on an earlier draft of this chapter. US-JGOFS Contribution 669.
REFERENCES Alldredge, A. L., Granata, T. C , Gotschalk, C. C , and Dickey, T. D. (1990). The physical strength of marine snow and its implications for particle disaggregation in the ocean. Limnol Oceanogr. 35, 1415-1428.
748
Christian and Anderson
Amon, R. M. W., and Benner, R. (1994). Rapid cycling of high-molecular-weight dissolved organic matter in the ocean. Nature 369, 549-552. Amon, R. M. W., and Benner, R. (1996). Bacterial utilization of different size classes of dissolved organic matter. Limnol Oceanogr. 41,41-51. Anderson, L. A., and Sarmiento, J. L. (1995). Global ocean phosphate and oxygen simulations. Global Biogeochem. Cycles 9, 621-636. Anderson, T. R. (1992). Modelling the influence of food C:N ratio, and respiration on growth and nitrogen excretion in marine zooplankton and bacteria. J. Plankton Res. 14, 1645-1671. Anderson, T. R., and Williams, R J. le B. (1998). ModeUing the seasonal cycle of dissolved organic carbon at Station Ei in the English Channel. Estuarine Coastal Shelf Sci. 46,93-109. Anderson, T. R., and Williams, P. J. le B. (1999). A one-dimensional model of dissolved organic carbon cycUng in the water column incorporating combined biological-photochemical decomposition. Global Biogeochem. Cycles 13, 337-349. Archer, D., Peltzer, E. T., and Kirchman, D. L. (1997). A timescale for dissolved organic carbon production in equatorial Pacific surface waters. Global Biogeochem. Cycles 11,435-452. Aumont, O., Orr, J. C , Monfray, P., Madec, G., and Maier-Reimer, E. (1999). Nutrient trapping in the equatorial Pacific: The ocean circulation solution. Global Biogeochem. Cycles 13, 351-369. Azam, F. (1998). Microbial control of oceanic carbon flux: The plot thickens. Science 280,694-696. Azam, P., Fenchel, T, Field, J. G., Gray, J. S., Meyer-Reil, L. A., and Thingstad, F. (1983). The ecological role of water-colunm microbes in the sea. Mar. Ecol. Prog. Ser. 10, 257-263. Bacastow, R., and Maier-Reimer, E. (1991). Dissolved organic carbon in modehng oceanic new production. Global Biogeochem. Cycles 5, 71-85. Bader, F. B. (1982). Kinetics of double-substrate limited growth. In "Microbial Population Dynamics" (M. J. Bazin, Ed.), pp. 1-32. CRC Press, Boca Raton, FL. Baines, S. B., and Pace, M. L. (1991). The production of dissolved organic matter by phytoplankton and its importance to bacteria: Patterns across marine and freshwater systems. Limnol. Oceanogr 36, 1078-1090. Benner, R., and Biddanda, B. (1998). Photochemical transformations of surface and deep marine dissolved organic matter: Effects on bacterial growth. Limnol. Oceanogr 43,1373-1378. Billen, G. (1990). Delayed development of bacterioplankton with respect to phytoplankton: A clue for understanding their trophic relationships. Arch. Hydrobiol. Beih. Ergeb. Limnol. 34, 191-201. Billen, G., and Becquevort, S. (1991). Phytoplankton-bacteria relationship in the Antarctic marine ecosystem. Polar Res. 10 (1), 245-253. Bissett, W R, Carder, K. L., Walsh, J. J., and Dieterle, D. A. (1999b). Carbon cychng in the upper waters of the Sargasso Sea. II. Numerical simulation of apparent and inherent optical properties. Deep-Sea Res. 46, 271-317. Bissett, W. R, Walsh, J. J., Dieterle, D. A., and Carder, K. L. (1999a). Carbon cycling in the upper waters of the Sargasso Sea. I. Numerical simulation of differential carbon and nitrogen fluxes. Deep-Sea Res. 46, 205-269. Bj0msen, P. K. (1988). Phytoplankton exudation of organic matter: Why do healthy cells do it? Limnol. Oceanogr 33, \5\-l54. Blackburn, N., Azam, F., and Hagstrom, A. (1997). Spatially expHcit simulations of a microbial food web. Limnol. Oceanogr 42, 613-622. Blackburn, N., Zweifel, U. L., and Hagstrom, A. (1996). Cychng of marine dissolved organic matter. II. A model analysis. Aquat. Microb. Ecol. 11,79-90. Bowen, J. D., Stolzenbach, K. D., and Chisholm, S. W. (1993). Simulating bacterial clustering around phytoplankton cells in a turbulent ocean. Limnol. Oceanogr 38, 36-51. Bratbak, G., and Thingstad, T. F. (1985). Phytoplankton-bacteria interactions: An apparent paradox? Analysis of a model system with both competition and commensalism. Mar Ecol. Prog. Ser 25, 23-30.
Modeling DOM Biogeochemistry
749
Broecker, W. S. (1982). Ocean chemistry during glacial time. Geochim. Cosmochim. Acta 46, 16891705. Button, D. K. (1998). Nutrient uptake by microorganisms according to kinetic parameters from theory as related to cytoarchitecture. Microbiol Mol Biol Rev. 62,636-645. Carlson, C. A., and Ducklow, H. W. (1995). Dissolved organic carbon in the upper ocean of the central equatorial Pacific Ocean, 1992: Daily and finescale vertical variations. Deep-Sea Res. II42, 639-656. Carlson, C. A., and Ducklow, H. W. (1996). Growth of bacterioplankton and consumption of dissolved organic matter in the Sargasso Sea. Aquat. Microb. Ecol 10,69-85. Carlson, C. A., Ducklow, H. W., and Michaels, A. F. (1994). Annual flux of dissolved organic carbon from the euphotic zone near Bermuda. Nature 371,405-408. Chen, W., and Wangersky, P. J. (1996). Rates of microbial degradation of dissolved organic carbon from phytoplankton cultures. J. Plankton Res. 18,1521-1533. Cherrier, J., Bauer, J. E., and Druffel, E. R. M. (1996). UtiUzation and turnover of labile dissolved organic matter by bacterial heterotrophs in eastern North Pacific surface waters. Mar. Ecol Prog. Sen 139, 267-279. Christian, J. R., Lewis, M. R., and Karl, D. M. (1997). Verticalfluxesof carbon, nitrogen, and phosphorus in the North Pacific Subtropical Gyre near Hawaii. J. Geophys. Res. 102, 15,667-15,677. Chrost, R. J. (1990). Microbial ectoenzymes in aquatic environments. In "Aquatic Microbial Ecology: Biochemical and Molecular Approaches" (R. J. Chrost and J. Overbeck, Eds.), pp. 47-78. SpringerVerlag, New York. Cifuentes, L. A., and Eldridge, P. M. (1998). A mass- and isotope-balance model of DOC mixing in estuaries. Limnol Oceanogr. 43,1872-1882. Connolly, J. P., and Coffin, R. B. (1995). Model of carbon cycling in planktonic food webs. /. Environ. Eng. 121, 682-690. Connolly, J. P., Coffin, R. B., and Landeck, R. E. (1992). Modeling carbon utilization by bacteria in natural water systems. In "Modelling the Metabolic and Physiologic Activities of Microorganisms" (C. J. Hurst, Ed.), pp. 249-276. Wiley, New York. Danabasoglu, G., McWilliams, J. C , and Gent, P. R. (1994). The role of mesoscale tracer transports in the global ocean circulation. Science 264,1123-1126. Davidson, K. (1996). Modelling microbial food webs. Mar Ecol Prog. Ser 145, 279-296. Decho, A. W (1989). Microbial exopolymer secretions in ocean environments: Their role(s) in food webs and marine processes. Oceanogr Mar Biol Ann. Rev. 28, 73-153. del Giorgio, P. A., and Cole, J. J. (2000). Bacterial energetics and growth efficiency. In "Microbial Ecology of the Oceans" (D. L. Kirchman, Ed.), pp. 289-325. Wiley, New York. Donah, E., OUi, K., Heiskanen, A.-S., and Andersen, T. (1999). Carbon flow patterns in the planktonic food web of the Gulf of Riga, the Baltic Sea: A reconstruction by the inverse method. /. Mar Syst. 23,251-268. Doney, S. C. (1999). Major challenges confronting marine biogeochemical modeling. Global Biogeochem. Cycles 13,705-714. Doney, S. C , Najjar, R. G., and Stewart, S. (1995). Photochemistry, mixing and diurnal cycles in the upper ocean. J. Mar Res. 53, 341-369. Ducklow, H. W. (1994). Modeling the microbial food web. Microb. Ecol 28, 303-319. Ducklow, H. W (1999). The bacterial component of the oceanic euphotic zone. FEMS Microbiol Ecol 30, 1-10. Fasham, M. J. R., Ducklow, H. W, and McKelvie, S. M. (1990). A nitrogen-based model of plankton dynamics in the oceanic mixed layer. / Mar Res. 48,591-639. Fasham, M. J. R., Sarmiento, J. L., Slater, R. D., Ducklow, H. W, and WiUiams, R. (1993). Ecosystem behaviour at Bermuda station "S" and ocean weather station "India": A general circulation model and observational analysis. Global Biogeochem. Cycles 7, 379^15.
750
Christian and Anderson
Fasham, M. J. R., Boyd, P. W., and Savidge, G. (1999). Modeling the relative contributions of autotrophs and heterotrophs to carbon flow at a Lagrangian JGOFS station in the northeast Atlantic: The importance of DOC. Limnol Oceanogr. 44, 80-94. Fuchs, B. M., Zubkov, M. V., Sahm, K., Burkhill, P. H., and Amann, R. (2000). Changes in community composition during dilution cultures of marine bacterioplankton as assessed by flow cytometric and molecular biological techniques. Environ. Microbiol. 2,191-201. Fuhrman, J. A., Ammerman, J. W., and Azam, F. (1980). Bacterioplankton in the coastal euphotic zone: Distribution, activity, and possible relationships with phytoplankton. Mar Biol. 60,201-207. Gnanadesikan, A. (1996). Modeling the diurnal cycle of carbon monoxide: Sensitivity to physics, chemistry, biology and optics. J. Geophys. Res. 101, 12,177-12,191. Goldman, J. C , Caron, D. A., and Dennet, M. R. (1987). Regulation of gross growth efficiency and anunonium regeneration in bacteria by substrate C:N ratio. Limnol. Oceanogr 32,1239-1252. Goldman, J. C , and Dennett, M. R. (2000). Growth of marine bacteria in batch and continuous culture under carbon and nitrogen limitation. Limnol. Oceanogr 45, 789-800. Hansen, D. A., and Carlson, D. A. (2001). Biogeochemistry of total organic carbon and nitrogen in the Sargasso Sea: Control by convective overturn. Deep-Sea Res. 7/48, 1649-1667. Hedges, J. I., and Lee, C , Eds. (1993). Measurement of dissolved organic carbon and nitrogen in natural waters. Mar Chem. 41, 1-290. Hill, P. S. (1996). Sectional and discrete representation of floe breakage in agitated suspensions. DeepSea Res. 43,679-702. Hin, P S. (1998). Controls on floe size in the sea. Oceanogr Mag. 11,13-18. Hoagland, K. D., Rosowski, J. R., Gretz, M. R., and Roemer, S. C. (1993). Diatom extracellular polymeric substances: function, fine structure, chemistry, and physiology. J. Phycol. 29, 537-566. Hopkinson, C. S., Buffam, I., Hobbie, J., Vallino, J., Perdue, M., Eversmeyer, B., Prahl, F , Covert, J., Hodson, R., Moran, M. A., Smith, E., Baross, J., Crump, B., Findlay, S., and Foreman, K. (1998). Terrestrial inputs of organic matter to coastal ecosystems: An intercomparison of chemical characteristics and biolabihty. Biogeochemistry 43, 211-234. Jackson, G. A. (1987). Physical and chemical properties of aquatic environments. In "Ecology of Microbial Communities" (M. Fletcher, T. R. G. Gray, and J. G. Jones, Eds.), pp. 213-233. Cambridge University Press, Cambridge. Jackson, G. A. (1989). Simulation of bacterial attraction and adhesion to falUng particles in an aquatic environment. Limnol. Oceanogr 34, 514-530. Jackson, G. A., and Eldridge, P. M. (1992). Food web analysis of a planktonic system off Southern California. Prog. Oceanogr 30, 223-251. Joint, I. R., and Morris, R. J. (1982). The role of bacteria in the turnover of organic matter in the sea. Oceanogr Mar Biol. Ann. Rev. 20, 65-118. Jumars, P A., Deming, J. W., Hill, P S., Karp-Boss, L., Yager, P L., and Dade, W. B. (1993). Physical constraints on marine osmotrophy in an optimal foraging context. Mar Microb. Food Webs 7, 121-159. Jumars, P A., Penry, D. L., Baross, J. A., Perry, M. J., and Frost, B. W. (1989). Closing the microbial loop: Dissolved carbon pathway to heterotrophic bacteria from incomplete ingestion, digestion and absorption in animals. Deep-Sea Res. 36,483^95. Kawamiya, M., Kishi, M. J., and Suginohara, N. (2000). An ecosystem model for the North Pacific embedded in a general circulation model. I. Model description and characteristics of spatial distributions of biological variables. J. Mar Syst. 25,129-157. Kawamiya, M., Kishi, M. J., Yamanaka, Y, and Suginohara, N. (1995). An ecological-physical coupled model appUed to Station Papa. J. Oceanogr 51, 635-664. Keil, R. G., and Kirchman, D. L. (1994). Abiotic transformation of labile protein to refractory protein in seawater. Mar Chem. 45, 187-196.
Modeling DOM Biogeochemistry
751
Keller, A. A., and Goldstein, R. A. (1995). Oceanic transport and storage of carbon emissions. Clim. Change 30, 367-395. Kepkay, P. E. (2000). Colloids and the ocean carbon cycle. In "The Handbook of Environmental Chemistry vol. 5, part D: Marine Chemistry" (P. Wangersky, Ed.), pp. 35-56. Springer-Verlag, Berlin. Kirchman, D. L. (1990). Limitation of bacterial growth by dissolved organic matter in the subarctic Pacific. Mar. Ecol Prog. Ser. 62,47-54. Kirchman, D. L. (2000). Uptake and regeneration of inorganic nutrients by marine heterotrophic bacteria. In "Microbial Ecology of the Oceans" (D. L. Kirchman, Ed.), pp. 261-288. Wiley, New York. Kirchman, D. L., Keil, R. G., Simon, M,, and Welschmeyer, N. A. (1993b). Biomass and production of heterotrophic bacterioplankton in the oceanic subarctic Pacific. Deep-Sea Res. 40,967-988. Kirchman, D. L., Lancelot, C , Fasham, M. J. R., Legendre, L., Radach, G., and Scott, M. (1993a). Dissolved organic matter in biogeochemical models of the ocean. In "Towards a Model of Ocean Biogeochemical Processes (G. T. Evans and M. J. R. Fasham, Eds.), pp. 209-225. Springer-Verlag, Berlin. Kirchman, D. L., Rich, J. H., and Barber, R. T. (1995). Biomass and biomass production of heterotrophic bacteria along 140°W in the equatorial Pacific: Effect of temperature on the microbial loop. Deep-Sea Res. II42,603-619. Klepper, O., de Haan, B. J., and van Huet, H. (1994). Biochemical feedbacks in the oceanic carbon cycle. Ecol. Modell. lSn6,459^69. Lamy, F., Bianchi, M., Van Wambeke, F., Sempere, R., and Talbot, V. (1999). Use of data assimilation techniques to analyze the significance of ectoproteolytic activity measurements performed with the model substrate MCA-Leu. Mar Ecol. Prog. Ser 111, 27-35. Lancelot, C , Billen, G., Veth, C , Becquevort, S., and Mathot, S. (1991). Modelling carbon cycling through phytoplankton and microbes in the Scotia-Weddell Sea area during sea ice retreat. Mar Chem. 35, 305-324. Levy, M., Memery, L., and Andre, J.-M. (1998). Simulation of primary production and export fluxes in the Northwestern Mediterranean Sea. /. Mar Res. 56,197-238. Libby, P. S., and Wheeler, P. A. (1997). Particulate and dissolved organic nitrogen in the central and eastern equatorial Pacific. Deep-Sea Res. 44, 345-361. Marchant, H. J., and Scott, F. J. (1993). Uptake of sub-micrometre particles and dissolved organic material by Antarctic choanoflagellates. Mar Ecol. Prog. Ser 92, 59-64. Martin, J. H., Knauer, G. A., Karl, D. M., and Broenkow, W W (1987). VERTEX: Carbon cycling in the northeast Pacific. Deep-Sea Res. 34,267-285. Matear, R. J., and HoUoway, G. (1995). Modeling the inorganic phosphorus cycle of the North Pacific using an adjoint data assimilation model to assess the role of dissolved organic phosphorus. Global Biogeochem. Cycles 9, 101-119. McCarthy, M. D., Hedges, J. L, and Benner, R. (1998). Major bacterial contribution to marine dissolved organic nitrogen. Science 281, 231-234. McGilhcuddy, D. J., Robinson, A. R., Siegel, D. A., Jannasch, H. W, Johnson, R., Dickey, T. D., McNeil, J., Michaels, A. R, and Knap, A. H. (1998). Influence of mesoscale eddies on new production in the Sargasso Sea. Nature 394, 263-266. McWilliams, J. C. (1996). Modeling the oceanic general circulation. Ann. Rev. Fluid Mech. 28, 215248. Miller, W L. (1998). Effects of UV radiation on aquatic humus: Photochemical principles and experimental considerations. In "Aquatic Humic Substances" (D. O. Hessen and L. J. Tranvik, Eds.), Ecological Studies 133, pp. 125-143. Springer-Verlag, Berlin. Miller, W. L., and Moran, M. A. (1997). Interaction of photochemical and microbial processes in the degradation of refractory dissolved organic matter from a coastal marine environment. Limnol. Oceanogr 42,1317-1324.
752
Christian and Anderson
Miller, W. L., and Zepp, R. G. (1995). Photochemical production of dissolved inorganic carbon from terrestrial organic matter—Significance to the oceanic organic carbon cycle. Geophys. Res. Lett. 22,417^20. Moloney, C. L., and Field, J. G. (1991). The size-based dynamics of plankton food webs. I. A simulation model of carbon and nitrogen flows. / Plankton Res. 13, 1003-1038. Monod, J. (1942). "Recherches sur la croissance des cultures bacteriennes." Hermann, Paris. Mumane, R. J. (1994). Determination of thorium and particulate matter cycling parameters at station P: A reanalysis and comparison of least squares techniques. J. Geophys. Res. 99, 3393-3405. Murray, A. G., and Jackson, G. A. (1992). Viral dynamics: A model of the effects of size, shape, motion and abundance of single-celled planktonic organisms and other particles. Mar. Ecol. Prog. 5eA: 89, 103-116. Nagata, T. (2000). Production mechanisms of dissolved organic matter. In "Microbial Ecology of the Oceans" (D. L. Kirchman, Ed.), pp. 121-152. Wiley, New York. Najjar, R. G., Sarmiento, J. L., and Toggweiler, J. R. (1992). Downward transport and fate of organic matter in the ocean: Simulations with a general circulation model. Global Biogeochem. Cycles 6, 45-76. Najjar, R. G., and Toggweiler, J. R. (1993). Reply to the comment by Jackson. Limnol. Oceanogr 38, 1331-1332. Norrman, B., Zweifel, U. L., Hopkinson, C. S., and Fry, B. (1995). Production and utiHzation of dissolved organic carbon during an experimental diatom bloom. Limnol. Oceanogr. 40, 898-907. Obemosterer, I., and Hemdl, G. J. (1995). Phytoplankton extracellular release and bacterial growth: Dependence on the inorganic N:P ratio. Mar Ecol. Prog. Ser. 116,247-257. Obemosterer, I., Reitner, B., and Hemdl, G. J. (1999). Contrasting effects of solar radiation on dissolved organic matter and its bioavailability to marine bacterioplankton. Limnol. Oceanogr. 44,1645-1654. Ogura, N. (1975). Further studies on the decomposition of dissolved organic matter in coastal seawater. Mar. 5/o/. 31, 101-111. Oschhes, A. (2000). Equatorial nutrient-trapping in biogeochemical ocean models: the role of advection numerics. Global Biogeochem. Cycles 14,655-667. Pace, M. L., Glasser, J. E., and Pomeroy, L. R. (1984). A simulation analysis of continental shelf food webs. Mar Biol. 82,47-63. Paillard, D., Ghil, M., and Le Treut, H. (1993). Dissolved organic matter and the glacial-interglacial pC02 problem. Global Biogeochem. Cycles 7, 901-914. Parsons, T. R., and Kessler, T. A. (1986). Computer model analysis of pelagic ecosystems in estuarine waters. In "The Role of Freshwater Outflow in Coastal Marine Ecosystems" (S. Skreslet, Ed.), pp. 161-181. Springer-Verlag, Berlin. Pett, R. J. (1989). Kinetics of microbial mineralization of organic carbon from detrital Skeletonema costatum cells. Mar Ecol. Prog. Ser 52, 123-128. Pinhassi, J., and Hagstrom, A. (2000). Seasonal succession in marine bacterioplankton. Aquat. Microb. Ecol. 21,245-256. Pomeroy, L. R. (1974). The ocean's food web, changing paradigm. Bioscience 24,499-504. Preiswerk, D., and Najjar, R. G. (2000). A global, open-ocean model of carbonyl sulfide and its air-sea flux. Global Biogeochem. Cycles 14, 585-598. Rintoul, S. R., and Wunsch, C. (1991). Mass, heat, oxygen and nutrient fluxes and budgets in the North Atlantic Ocean. Deep-Sea Res. (Suppl.) 38, 355-377. Ryabchenko, V. A., Gorchakov, V. A., and Fasham, M. J. R. (1998). Seasonal dynamics and biological productivity in the Arabian Sea euphotic zone as simulated by a three-dimensional ecosystem model. Global Biogeochem. Cycles 12, 501-530. Sarmiento, J. L., Slater, R. D., Fasham, M. J. R., Ducklow, H. W., Toggweiler, J. R., and Evans, G. T. (1993). A seasonal three-dimensional ecosystem model of nitrogen cycHng in the North Atlantic euphotic zone. Global Biogeochem. Cycles 7,417-450.
Modeling DOM Biogeochemistry
753
Schlitzer, R. (1989). Modeling the nutrient and carbon cycles of the North Atlantic. 2. New production, particle fluxes, CO2 gas exchange, and the role of organic nutrients. / Geophys. Res. 94, 12781-12794. Sharp, J. H. (1973). Size classes of organic carbon in seawater. Limnol Oceanogr. 18,441. Sherr, E. B. (1988). Direct use of high molecular weight polysaccharide by heterotrophic flagellates. A^flmre 335, 348-351. Siegel, D. A., and Michaels, A. F. (1996). Quantification of non-algal light attenuation in the Sargasso Sea. Deep-Sea Res. II43, 321-345. Siegenthaler, U., and Joos, F. (1992). Use of a simple model for studying oceanic tracer distributions and the global carbon cycle. Tellus 44B, 186-207. Sikorski, R. J., and Zika, R. G. (1993). Modeling mixed-layer photochemistry of H2O2: optical and chemical modeling of production. J. Geophys. Res. 98,2315-2328. Six, K. D., and Maier-Reimer, E. (1996). Effects of plankton dynamics on seasonal carbon fluxes in an ocean general circulation model. Global Biogeochem. Cycles 10,559-583. Smith, D. C , Simon, M., AUdredge, A. L., and Azam, F. (1992). Intense hydrolytic enzyme activity on marine aggregates and implications for rapid particle dissolution. Nature 359,139-142. Spitz, Y. H., Moisan, J. R., and Abbott, M. R. (2001). Configuring an ecosystem model using data from the Bermuda Atlantic Time-series (BATS). Deep-Sea Res. II A&, 1733-1768. Spitz, Y. H., Moisan, J. R., Abbott, M. R., and Richman, J. G. (1998). Data assimilation and a pelagic ecosystem model: Parameterization using time-series observations. /. Mar. Systems 16, 51-68. Strom, S. L., Benner, R., Ziegler, S., and Dagg, M. J. (1997). Planktonic grazers are a potentially important source of marine dissolved organic carbon. Limnol Oceanogr. 42,1364-1374. Sugimura, Y, and Suzuki, Y (1988). A high-temperature catalytic oxidation method for the determination of non-volatile dissolved organic carbon in seawater by direct injection of a liquid sample. Mar Chem. 24,105-131. Suzuki, Y (1993). On the measurement of DOC and DON in seawater. Mar Chem. 41, 287-288. Suzuki, Y, Sugimura, Y, and Itoh, T. (1985). A catalytic oxidation method for the determination of total dissolved nitrogen in seawater. Mar Chem. 16, 83-97. Tanoue, E., Nishiyama, D., Kamo, M., and Tsugita, A. (1995). Bacterial membranes: Possible source of a major dissolved protein in seawater. Geochim. Cosmochim. Acta 59, 2643-2648. Taylor, A. H., and Joint, I. (1990). A steady-state analysis of the 'microbial loop' in stratified systems. Mar Ecol. Prog. Ser 59,1-17. Thingstad, T. F. (2000). Elements of a theory for the mechanisms controlling abundance, diversity, and biogeochemical role of lytic bacterial viruses in aquatic systems. Limnol. Oceanogr 45,13201328. Thingstad, T. F., Hagstrom, A., and Rassoulzadegan, F. (1997). Accumulation of degradable DOC in surface waters: Is it caused by a malfunctioning microbial loop? Limnol. Oceanogr 42, 398404. Thingstad, T. R, Havskum, H., Kaas, H., Nielsen, T. G., Riemann, B., Lefevre, D., and Williams, P. J. le B. (1999). Bacteria-protist interactions and organic matter degradation under P-limited conditions: Analysis of an enclosure experiment using a simple model. Limnol. Oceanogr 44, 62-79. Tian, R. C , Vezina, A. F., Legendre, L., Ingram, R. G., Klein, B., Packard, T., Roy, S., Savenkoff, C , Silverberg, N., Therriault, J.-C, and Tremblay, J.-E. (2000). Effects of pelagic food-web interactions and nutrient remineralization on the biogeochemical cycling of carbon: A modeling approach. Deep-Sea Res. 47,637-662. Toggweiler, J. R. (1989). Is the downward dissolved organic matter (DOM) flux important in carbon transport? In "Productivity of the Ocean: Present and Past" (W. H. Berger, V. S. Smetacek, and G. Wefer, Eds.), pp. 65-83. Wiley, New York. Toggweiler, J. R., and Carson, S. (1995). What are upwelling systems contributing to the ocean's
754
Christian and Anderson
carbon and nutrient budgets. In "Upwelling in the Ocean: Modem Processes and Ancient Records" (C. P. Summerhayes, K.-C. Emeis, M. V. Angel, R. L. Smith, and B. Zeitzschel, Eds.), pp. 337-360. Wiley, Chichester. Tusseau-Vuillemin, M.-H., Mortier, L., and Herbaut, C. (1998). ModeUng nitrate fluxes in an open coastal environment (Gulf of Lions): Transport versus biogeochemical processes. J. Geophys. Res. 103,7693-7708. Vallino, J. J. (2000). Improving marine ecosystem models: Use of data assimilation and mesocosm experiments. J. Mar. Res. 58, 117-164. Vallino, J. J., Hopkinson, C. S., and Hobbie, J. E. (1996). Modehng bacterial utilization of dissolved organic matter: Optimization replaces Monod growth kinetics. Limnol. Oceanogr. 41, 15911609. Vetter, Y. A., Deming, J. W., Jumars, P. A., and Krieger-Brockett, B. B. (1998). A predictive model of bacterial foraging by means of freely released extracellular enzymes. Microb. Ecol 36,75-92. Vezina, A. R, Demers, S., Laurion, I., Sime-Ngando, T., Juniper, S. K., and Devine, L. (1997). Carbon flows through the microbial food web of first-year ice in Resolute Passage (Canadian High Arctic). J. Mar. Systems 11, 173-189. Vezina, A. F., and Pace, M. L. (1994). An inverse model analysis of planktonic food webs in experimental lakes. Can. J. Fish. Aq. Sci. 51,2034-2044. Vezina, A. R, and Piatt, T. (1988). Food web dynamics in the ocean. I. Best-estimates of flow networks using inverse methods. Mar. Ecol. Prog. Sen 42, 269-287. Vezina, A. R, and Savenkoff, C. (1999). Inverse modeUing of carbon and nitrogen flows in the pelagic food web of the northeast subarctic Pacific. Deep-Sea Res. II46,2909-2939. Vezina, A. R, Savenkoff, C , Roy, S., Klein, B., Rivkin, R., Therriauh, J.-C, and Legendre, L. (2000). Export of biogenic carbon and structure and dynamics of the pelagic food web in the Gulf of St. Lawrence. 2. Inverse analysis. Deep-Sea Res. II47, 609-635. Walsh, J. J., Carder, K. L., and Muller-Karger, R E. (1992). Meridional fluxes of dissolved organic matter in the North Atlantic Ocean. J. Geophys. Res. 97, 15625-15637. Walsh, J. J., and Dieterle, D. A. (1994). CO2 cycling in the coastal ocean. I. A numerical analysis of the southeastern Bering Sea with appUcations to the Chukchi Sea and the northern Gulf of Mexico. Prog. Oceanogr 34, 335-392. Walsh, J. J., Dieterie, D. A., Muller-Karger, R E., Aagaard, K., Roach, A. T., Whitledge, T. E., and Stockwell, D. (1997). CO2 cychng in the coastal ocean. II. Seasonal organic loading of the Arctic Ocean from source waters in the Bering Sea. Com. Shelf Res. 17, 1-36. Walsh, J. J., Dieterie, D. A., Muller-Karger, P. E., Bohrer, R., Bissett, W. P., Varela, R. J., Aparicio, R., Diaz, R., Thunnell, R., Taylor, G. T, Scranton, M. I., Panning, K. A., and Peltzer, E. T. (1999). Simulation of carbon-nitrogen cycling during spring upwelling in the Cariaco Basin. J. Geophys. Res. 104, 7807-7825. Walsh, J. J., Rowe, G. T, Iverson, R. L., and McRoy, C. P. (1981). Biological export of shelf carbon is a sink of the global CO2 cycle. Nature 291, 196-201. Weber, A. L. (2000). Sugars as the optimal biosynthetic carbon substrate of aqueous life throughout the universe. Orig. Life Evol. Biosphere 30, 33^3. WiUiams, P. J. le B. (1981). Incorporation of microheterotrophic processes into the classical paradigm of the planktonic food web. Kiel. Meeresforsch. Sonderh. 5,1-28. WiUiams, P. J. le B. (1990). The importance of losses during microbial growth: Commentary on the physiology, measurement and ecology of the release of dissolved organic material. Man Microb. FoodWebs4,\15-2(d6. WiUiams, P. J. le B. (1995). Evidence for the seasonal accumulation of carbon-rich dissolved organic material, its scale in comparison with changes in particulate material and the consequential effect on the C/N assimilation ratios. Mar Chem. 51,17-29.
Modeling DOM Biogeochemistry
755
Wong, C. S. (1978). Atmospheric input of carbon dioxide from burning wood. Science 200,197-200. Wright, S. H., and Manahan, D. T. (1989). Integumental nutrient-uptake by aquatic organisms. Ann. Rev. Physiol. 51, 585-600. Yamanaka, Y, and Tajika, E. (1997). Role of dissolved organic matter in the marine biogeochemical cycle: Studies using an ocean biogeochemical general circulation model. Global Biogeochem. Cycles 11,599-612. Yassuda, E. A., Davie, S. R., Mendelsohn, D. L., Isaji, T., and Peene, S. J. (2000). Development of a waste load allocation model for the Charleston Harbor estuary. II. Water quality. Estuarine Coastal Shelf Sci. 50,99-107.
This Page Intentionally Left Blank
Index
AABW, see Antarctic bottom water Abiotic formation of biologically recalcitrant DOM, 133-134 removal processes, 122-123 phototransformation, 122-123 sorption of DOM onto particles, 123 Absorption of light, 513-526 CDOM as principle, 513-515 magnitude of, 515 Absorption of UV light, role of DOM in, 456-457 Absorption/fluorescence, relationship to DOC, 529-531 Accumulation of DOM, 133-137 abiotic formation of biologically recalcitrant DOM, 133-134 biotic formation of recalcitrant DOM, 134-135 limitation of bacterial growth and accumulation of biodegradable DOM, 135-136 microbial community structure and DOM utiUzation, 137 Accumulation of DOM in coastal zone and export processes export of DOM from coastal zone, 598-600 mechanism of accumulation, 595-598 Adenosine 3^ 5'-cyclic monophosphate (c-AMP), 322-323 Advection/diffusion/reaction model for DOM, description of, 65\-65%app) anoxic, nonbioturbated sediments (ANS) model, 651
bioturbated and/or bioirrigated sediments (BBS) model, 652-653 Advection/diffusion/reaction model for sediment DOC cycling, 617-623 application to anoxic sediments, 618-619 application to bioturbated and bioirrigated sediments, 619-623 Air-sea exchange of important atmospheric trace gases, photo production and, 495 Air-sea surface, organic matter at, 10 Algal blooms, harmful, potential link between DON and, 210-211 Amberlite XAD resin series, 62 Amino acids, 638-640 common, in seawater, 69-70, 72, 76-77 Ammonium formation in natural waters, 473-474 Anabaena, 330 Analytical interferences in SRP and TDP estimation, 279-280 Analytical methods for total DOM pools broad community methods comparisons, 41^4 historical perspective, 37 problem, 37-39 small group methods comparisons, 39^1 Antarctic bottom water (AABW) decrease of DOC in, 690 forms in cyclonic gyres, 706 Anoxia and sediment carbon preservation, relationship between, 649 Anoxic nonbioturbated sediments (ANS) model, 651 (app)
757
758 Anoxic sediments acetate concentration range, 637-638 application to, 618-619 APase activity, 342, 343, 344 Aquatic environments, concentrations of DON in, 155 Arabian Sea, DOC and bloom dynamics of, 695 Arachaea, 137, 220 Arctic Ocean, DOC in composition and distribution of DOC within, 674-679 C/N molar ratios, 676-677 distribution, 677-679 lignin oxidations products and stable carbon isotopes, 674-676 sources of DOC to Arctic Ocean, 667-674 biological sources with Arctic Ocean, 672-674 river runoff sources, 668-671 seawater sources, 671-672 summary of sources and sinks, 679-681 water masses and circulation, 665-667 Arctic Ocean, productivity of central, 673 Arctic Ocean waters, DOC concentrations of deep, 678 Arsenic (As) in seawater, 279 Atlantic Ocean, and, 562-563 Atmospheric CO2, DOM and, 736-737 Atmospheric trace gases, important, photo production and air-sea exchange, 495 Atmospheric ventilation, 468 ATP and related nucleotides, 319-322 Autotrophic versus heterotrophic DON utilization, 208-211 Autotrophs, 208-210 B Bacteria as DON consumer, 190-191 role in metabolizing photochemical products ofCDOM,537 Bacterial carbon demand, 120-122 Bacterial contributions to DOM, research on, 82-83,113,134-135 Bacterial DOM growth and accumulation of biodegradable DOM, limitation of, 135-136 origination and transformation of, 115-116
Index Bacterial growth efficiency, 117-120 selected from literature, 119-120 changes in, 480 Bacterial lysis, 113 Bacterial utilization of labile DOM, 116-117, 729-732 of labile radiolabeled substrates, 83 Bacterioplankton, examples of DOM released from, 116 Baltic Sea, optical oceanographers working in, 548-549 Basin scale, spatial variabihty at, 687-693 Basin-to-global scale ecosystems modehng, 738-739 Benthic DOC fluxes (BDF), 642, 643-644 Benthic DOM fluxes in ocean carbon and nitrogen cycles, role of benthic DOC fluxes, 641-644 benthic DON fluxes, 644-646 extent to which benthic DOM fluxes affect composition and reactivity of deep-water DOM, 646-648 Bermuda Atlantic Times-Series Study (BATS), 553-554 Bioactive metals in seawater, chemical speciation of, 386 Bioassay procedure, 330 Biodegradability of riverine DOM, 584-588 Biological labihty of DOM, continuum of, 130-133 Biological pump, contribution of DOC to evidence for DOC export, 702-706 exportable DOC, 707-709 Biological sources of DOC within Arctic Ocean, 672-674 Biologically available P, 329-330 Biologically labile DOM, 128-129 Biologically refractory DOM, 126-128 Biologically semilable DOM, 129-130 Biomolecules, well-defined, 537 Bioreactive pools of carbon in ocean, 123-133, 685-687 Biorefractory DOM mechanisms of formation and removal of, 80-82 reservoir, 81 Biorefractory LMW DOM in ocean, 82 Biotic consumption of DOM, 116-122 eukaryotes, 122
Index prokaryotes, 116-122 bacterial carbon demand, 120-122 ^ bacterial growth efficiency, 117-120 Bidtic DON release rate, methods for estimating, 200-202 uptake, methods for estimating, 211-212 Biotic formation of recalcitrant DOM eukaryotic sources, 134 prokWyotic sources, 134-135 Biotic sources of DON in water column, 186-200 Bioturbated and/or bioirrigated sediments (BBS) model, 652-653 (app) application to, 619-623, 624 Blackwatenrivers, distribution of CDOM of, 532 Blank problem, recognition of, 41 Blue Hght absorption, fraction of, 562 Blue shifts, observation of in excitation/emission spectra of CDOM, 533 Bomb ^^C, old oceanic carbon reservoirs, 412 Broad community methods comparison current and future, 47-48 for total DOM pools, 4 1 ^ 4 Budget of DOC fluxes to and from Arctic Ocean, 680 Bulk DON, 202 Bulk marine DOM, distribution and chemical characteristics of bulk concentration and composition, 64-68 chemical composition of isolated fractions of, 71-80 molecular composition, 68-71 Bulk seawater DOM and biochemical content, 10-11
C/N molar ratios, 676-677 C:N ratio within UDOC, 439 C:N:P stoichiometry of dissolved and particulate matter pools, 292-294 Carbohydrates, 640-641 during mesocosm, distribution of, 597 Carbon, distribution of along size continuum of organic matter in seawater, 64 Carbon and sulfur cycles, photochemistry and, 569-570 Carbon dioxide (CO2), 456 Carbon isotopic composition of DOM application of ^^^C and A^^^C of DOC in ocean margins, 430-443
759 distributions of DOC in ocean margins, 433-439 exchanges of DOC between ocean's margins and interior, 441-443 sources and inputs of UDOC to ocean margins, 439^41 vertical mixing and distribution of A^^^C-DOC in open ocean, 430-433 conventions and definitions for expressing isotopic contents carbon 13, 407-^09 carbon 14, 409-413 future challenges, 445 measurements and distributions of 8^C and A^^C of DOC in oceanic systems, 415-429 distribution of in ocean margins, 423-429 distribution of in oceanic systems, 417-423 methods used for extraction, 416 methods for extracting DOC from seawater for isotopic analysis sample oxidation, 414 sample processing, 413^14 studies, 405-407 summary, 443-445 Carbon monoxide (CO), 456 Carbon 13 isotopic contents, conventions and definitions for expressing, 407^09 signatures of organic matter, 408 Carbon 14, isotopic contents, conventions and definitions for expressing, 409^13 content of carbon pool, 409 expression of, 410 Carbon 14 PP rates, daily, and seasonally produced DOC, DON, and DOP stocks and concentrations during phytoplankton blooms for marine sites, 94-97 Carbon-oxygen mass balances and absorbence changes, photochemical DOM degradation in relation to, 503-507 Carbonyl sulfide (OCS), 456,468-472 CDOM composition, assessing, 553 Cell lysis, DOM production via, 112-113 bacterial lysis, 113 viral lysis, 112 Cell-wall-associated compound classes, 332 Cellular metabolism, coordination and regulation of, 333
760 Cellular P metabolism, 262-263 CFF colloid studies, use of, 376 Characterization by ^^P NMR, 311-313 Chemical characterization of DOM, 61-63 soHd-phase extraction for DOM isolation, 61-62 tangential-flow (or cross-flow) ultrafiltration for DOM isolation, 62-63 Chemical composition and reactivity chemical characterization of DOM, 59-64 approaches for chemical characterization, 61-63 size distribution of, 63-64 distribution and chemical characteristics of bulk marine DOM bulk concentration and composition, 64-68 chemical composition of isolated fractions of, 71-80 molecular composition, 68-71 major topics on research on cycling of DOM, ongoing and future fate of terrigenous DOM in ocean, 83-84 mechanisms of formation and removal of biorefractory DOM, 80-82 Chemical composition of DOM, 552-554 Chemical composition of DON pool, 174-185 Chemical form of colloidal metals, 385-388 Chemocentric definition for colloidal matter, 371 Chesapeake Bay sediments, 628 anoxic, 645 remineralization, 630 Chlorophyll absorption, CDOM from, 556 Chromophores, absorbing, 555 Chromophoric DOM in coastal environment distribution, 532-534 fiiture areas of research, 539 interest in, 509-510 optical properties, 513-531 absorption, 513-526 fluorescence, 526-531 sources and sinks sinks, 536-539 sources, 534-536 sunmiary, 539 Chromophoric DOM in open ocean characterization of DOM chemical composition, 552-554
Index methods for quantifying, 554-556 operational definition, 549-550 optical properties, 550-552 future advances, needs for, 571-572 global CDOM distribution controls on global CDOM distribution, 563-566 global distribution of CDOM absorption from sea WiFS, 561-563 temporal variability of global CDOM distribution, 563 implications for photochemistry and photobiology, 568-571 photochemistry and carbon and sulfur cycles, 569-570 photoinhibition of photosynthesis and microbial growth, 570-571 observed DOM dynamics impact of CDOM on underwater light field, 557 local sources and sinks of CDOM, 557-561 recent research, 547-548 short history, 548-549 relationship between DOM and CDOM in open ocean, 567-568 Cloud condensation nuclei (CCN), 486 Coastal and estuarine systems, 741-742 Coastal zone, DOM in accumulation of DOM in coastal zone and export processes, 595-600 export of DOM from coastal zone, 598-600 mechanism of accumulation, 595-598 conclusions, 600-601 estuarine processes, 588-595 degradation of DOM during estuarine mixing, 590-595 physical processes, 588-590 river inputs in, 579-588 biodegradability of riverine DOM, 584-588 estimates of discharge from river, 580-584 Colloid, definition of, 64 Colloid cycHng models, 392 Colloid-associated metals in offshore waters, studies of, 383-384 Colloidal bioactive metals, biological availability of, 395-396
Index Colloidal ligands, sources of metal-complexing, 389-390 Colloidal matter isolation of for bulk analysis, 375-380 reaction rates, 390-394 Colloidal metals, chemical form of, 385-388 Community precision for TDN analysis, 47 Composition and concentration of DON pool, 154^186 Composition and distribution of DOC within, 674-679 C/N molar ratios, 676-677 distribution, 677-679 lignin oxidation products and stable carbon isotopes, 674-676 Concentration and composition of bulk marine DOM, 64-68 C, N, and P stoichiometry in DOM, 66 colormetric analyses of dissolved carbohydrates, 66-68 Concentration and composition of DON pool, 154-186 chemical composition of DON pool, 174-185 DON distributions and correlative relationships between DON and other parameters, 155-174 methods for measuring DON concentrates, 154-155 research priorities, 186 Concentration gradients of DOM between surface and deep water, 68 Consumption processes of DOM, see Production and consumption processes of DOM Contrasting marine sediments, best fit rate parameters for DOC cycling in, 629 Controls on DOC concentrations with depth in surficial sediments, 623-626 Controls on global CDOM distribution, 563-566 Cross flow filtration (CFF), 307, 375 Cross flow ultrafiltration for DOM isolation, 62-63 Cross-linked polystyrene (XAD) resins, 414 Cyclic AMP, DOP in, 322-323 D D-DNA, 183-184 DCAA chemical structure of, 175 and DFAA, 205, 206, 213, 220-221
761 Deep ocean, concentrations in, 286-292 distributions of DOC, 689-690 DOC samples, data from, 43 DOM, extent to which benthic DOM fluxes affect composition and reactivity of, 646-648 DOM samples refractory, 5-6 as static pool, 4 exportable DOC to, 706 and surface DOC, 599 temporal variability of DOC in, 696 transit times and DOC age differences, 421 Deep pycnocline, exportable DOC into, 704, 706 Deep sediment cores, pore water DOC profiles in, 626-631 Degradation of DOM during estuarine mixing, 590-595 S^^C and A^^C of DOC in ocean margins, 430-443 distributions of DOC in ocean margins, 433-^39 exchanges of DOC between ocean's margins and interior, 441-^43 sources and inputs of UDOC to ocean margins, 439^41 vertical mixing and distribution of A^^^C-DOC in open ocean, 430-433 Denges-Atkins method of seasonal phosphate concentration dynamics, 259-260 Depth variations in DOP, 281-286 Derivative high-temperature catalytic oxidation method for DOC analysis, 15 DFAA chemical structure of, 180-181 and DCAA, 205, 206, 213, 220-221 Dialysis, 202 Dimethyl sulfide (DMS), 456, 468-469 Dimethyl sulfoxide, 469 Dinitrogen (N2) fixation, 303-304 Direct measurement of DOP compounds, 314-329 ATP and related nucleotides, 319-322 cyclic AMP, 322-323 inorganic poly-P/ and pyro Fi, 327-329 lipids, 323-326 nucleic acids, 315-319 vitamins, 326-327
762 Direct utilization of DOP, 339-342 Discharge from river, estimates of, 580-584 Dispersed HA and FA compounds, 332 Dissolved inorganic carbon (DIC), removal of, 37-38 Dissolved organic carbon (DOC) advection/diffusion/reaction model for sediment DOC cycling, 617-623 analysis, improved methods for, 9 benthic carbon fluxes, 641-646 controls on DOC concentrations with depth in surficial sediments, 623-626 deep ocean profiles of versus DON, 18 examples of light intensity, nutrient limitation, temperamre, and community structure on PER of, 104-105 general observations, 614-617 in seawater, 3 large blanks intrinsic to HTCO-based instruments, 22 measurements of concentrates, 2 pore water DOC profiles in deep sediment cores, 626-631 release rates and PER for natural assemblages of marine origin, 100-103 Dissolved organic matter (DOM) global carbon cycle and, 23 importance of, 23-25 learning experiences, 25-26 measurement of DOC and DON concentrates, 2 new DON and DOC, 13-23 operational definition of, 549-550 quantifying DOM, 36 research on, 1, 2-13 pre-1970's, 2-7 inl970's,7-ll in 1980's, 11-13 UV-based, 20 Dissolved organic matter (DOM) compositional data, 636-Ml amino acids, 638-640 carbohydrates, 640-641 volatile fatty acids (VFAs), 637-638 Dissolved organic nitrogen (DON), 631-636 analysis current and future broad community methods comparison, 47-48 benthic fluxes, 644-646 concentrates deep ocean profiles versus DON, 18
Index definition of, 153-154 dynamics concentration and composition of DON pool, 154-186 chemical composition of DON pool, 174-185 DON distributions and correlative relationships between DON and other parameters, 155-174 methods for measuring DON concentrates, 154-155 research priorities, 186 historical perspective and analytical problem, 45^6 in seawater, 2-3 sinks for DON, 207-226 heterotrophic versus autotrophic DON utilization, 208-211 literature values of DON uptake in aquatic environments, 212-222 methods for estimating biotic DON uptake, 211-212 photochemical decomposition as sink for DON, 222-223 research priorities, 223-226 small group methods comparisons, 46-47 sources of, 186-213 biotic sources of DON in water column, 186-200 literature values of DON release rates in aquatic environments, 202-212 research priorities, 206-207 methods for estimating biotic DON release rate, 200-202 summary, 227-231 turnover times for DON, 226-227 values in marine sediments, 634-635 Dissolved organic phosphorus (DOP) analysis historical perspective and analytical problem, 49-50 small group methods comparison, 50 Dissolved organic phosphorus (DOP), dynamics of conclusions and prospecms, 347-348 early years of pelagic marine P-cycle research, 258-262 pelagic marine P cycle, key pools and processes, 262-266 phosphorus (P), as essential nutrient for living organisms, 250-253
763
Index pool characterization, 306-334 biologically available P, 329-330 characterization by enzymatic characterization, 309-311 characterization by ^^P NMR, 311-313 direct measurement of DOP compounds, 314-329 majority view of DOP, 330-334 molecular weight characterization of DOP pool, 308-309 by partial photochemical oxidation, 313-314 production, utilization and remineralization, 334-347 direct utilization of DOP, 339-342 DOP interactions with light and suspended minerals, 346-347 DOP production and remineralization, 335-339 enzymes as P-cycle facilitators, 342-346 sampling, incubations, storage, and analytical considerations analytical interferences in SRP and TDP estimation, 279-280 detection of Fi and P-containing compounds in seawater, 271-279 isotopic tracers, use of in P-cycle research, 267-268 sample processing, preservation and storage, 269-270 sampling, 266-267 in sea, variation in space, 280-294 C:N:P stoichiometry of dissolved and particulate matter pools, 292-294 in deep sea, concentrations in, 286-292 regional and depth variations in DOP, 281-286 in sea, variation in time, 294-306 English Channel, 295-297 North Pacific Subtropical Gyre (NPSG), 297-306 terms, definitions and concentration units, 253-258 Dissolved polysaccharides, release of, 597 Distribution of DOC spatial variabiUty at basin scale, 687-693 temporal variability, 693-697 DOC, see Dissolved organic carbon DOC/DON Workshop, 26 DOM, see Dissolved organic matter
DON, see Dissolved organic nitrogen DOP, see Dissolved organic phosphorus Dynamic light scattering (DLS), 373
Ecological significance of DOM, 92 Ecosystem model coupled to spectral model, 733 model structure, 722 sources of DOM that include, 724 Ecosystem modeling studies, 719-743 coastal and estuarine systems, 741-742 compartments and currencies, 719-723 diagnostic models of North Atlantic, 737-738 inverse models of flows within food webs, 740-741 modeling production of DOM, 723-728 modeling utilization and remineralization of DOM, 728-733 small-scale spatial structure, 742-743 Elemental analyses, limits of, 51-52 Elemental cycles, impact of photochemistry on carbon, 458, 460-467 nitrogen and phosphorus, 473 sulfur, 467-473 English Channel DOP in, 295-297 DOP samples, 259 Enzymatic characterization of P, 309-311 Enzymes as P-cycle facilitators, 342-346 schematic presentation of role of selected, 263 Equatorial nutrient-trapping, early results, 734-735 Equatorial Pacific, DOC production in, 699 Escherichia coli, 265, 340 Estuarine and coastal systems, 741-742 Estuarine processes of DOM degradation of during mixing, 590-595 physical processes, 588-590 Eukaryotes, 122, 134 Excitation and emission spectra, 527-528 Excitation-emission matrix spectra, 527-528 Exogenous radioisotopic tracers, use of, 268 Exogenous radiotracers, 336 Export of DOM from coastal zone, 598-600 Exportable DOC, 702-709 to deep ocean, 706 into deep pycnocline, 704, 706 into upper pycnocline, 703-704
764 Extracellular phytoplankton production, 98-106 release models, 99-105 model comparison, 105-106 overflow model, 99 passive diffusion model, 105
Fate colloidal organic carbon phases in ocean waters, 369 Flow-field fractionation (flow FFF), 378-380 components and theory of FFF separation, 379 techniques, 380 Flows within food webs, inverse models of, 740 Ruorescence measurements of humic substances and CDOM, 526-531 fluorescence quantum yield and fluorescence/absorption relation, 529 fluorescent-excitation and emission spectra, 527-528 relationship of absorption/fluorescence to DOC, 529-531 Fluorescent chromophore distribution, 555 Ruorescent substances dissolved in seawater, 11 Food webs idealized, 723 inverse models of flows within, 740-741 plankton, complexity of, 717-718 Fractions of DOC, characteristics of divided by biological labiHty, 125-126 Fulvic and humic substances, 181-183
Gel formation via colloid aggregation, 394 Geopolymerization model for sediment preservation, 648 Global carbon reserves, factors controlling distribution of ^^C in, 411 Global carbon cycle importance of, 23 interest in, 60 Global CDOM distribution controls on global CDOM distribution, 563-566 global distribution of CDOM absorption from sea WiFS, 561-563 temporal variability of global CDOM distribution, 563 Global distributions of DON, 155-157 Global ocean carbon cycle, DOC in
Index bioreactive pools of carbon in ocean, 685-687 contribution of DOC to biological pump evidence for DOC export, 702-706 exportable DOC, 707-709 distribution of DOC spatial variability at basin scale, 687-693 temporal variability, 693-697 net community production of, 697-702 evidence for net production of DOC, 699-700 nutrient depletion and net production of DOC, 701-702 regional and global estimates for net production of DOC, 700-701 research priorities, 709-710 sunmiary, 711 Global open ocean DOP, 280, 281 Global pattern of CDOM absorption, 562 Global scales, extrapolating photochemical rates to, 478 Global-scale meridonial CDOM increase, 565 Grazer-associated DOM production, 725-726 Grazing-induced DOM production, 106-112 biogeochemical significance, 107, 112 macrozooplankton, 106-107 macrozooplankton, 107 GulfofMexico,DON, 173 Gulf of Riga, DOC in, 740-741
H Heterotrophic bacteria, 395 Heterotrophic versus autotrophic DON utilization, 208-211 Heterotrophs, 208 High latitudes regions, chemical characteristics of DOM in, 65 temporal variability of DOC in, 694 High-temperature catalytic oxidation (HTCO), 718 HMWDOM,613 HMWDON characteristics of, 184-185 mucous exopolymers, 115 HT(C)0, use of with oxidation, 25 method, 16, 17 HTCO-based instruments, large DOC blanks intrinsic to, 22 Humic substances, 221-222 and fulvic substances, 181, 183
765
Index Humic-bound organics, photolysis of, 492 Humification, 331 Hydrogen sulfide, photochemical degradation of, 472 Hypothesized interactions regulating mixed layer concentrations, 566
Iberian margin, DOC production in, 699 Ice in sea, DOC in, 673 In-house comparison and refereed type of comparison, differences between, 43 small, 44 Indian Ocean, and %CDOM contributions, 562-563 Individual organic N compounds and bulk DON pool, published rates of, 214-219 release rates of in field, 203 Inorganic carbon formation, photochemical dissolved, 484-485 Inorganic forms of P, 253, 256-257 Inorganic N, link between DON and, 172-174 Inorganic poly-P/ and pyro Fi, DO? in, 327-329 Intercalibration effort, 48 Inverse models of flows within food webs, 740-741 Ion retardation, 201-202 Isolated fractions of chemical composition of DOM, 71-80 amino acids analyses (THAA), 74, 75 average, 73 hydrolysis, variability in susceptibility to, 76 NMR spectroscopy, 72-73 Isotope dilution, 328 Isotopes of carbon in DOC, natural, 406 Isotopic analysis, methods for extracting DOC from seawater for, 413-414 Isotopic contents, conventions and definitions for expressing carbon 13,407-409 carbon 14, 409-413 Isotopic tracers, use of in P-cycle research, 267-268
K Key pools and processes, 262-266
Labile DOM component in surface ocean, perception, 7 Lability of DOM, 123-133 biologically labile DOM, 128-129 biologically refractory DOM, 126-128 biologically semilable DOM, 129-130 continuum of biological lability, 130-133 produced, 728 Light absorption of CDOM, methods for estimating, 555-555 Light and suspended minerals, DOP interactions with, 346-347 Lignan-derived phenols, concentrations of, 84 Lignin oxidation products and stable carbon isotopes, 674-676 Lipids, DOP in, 323-326 Literature values of DON concentrations of DOC in aquatic systems, 176-178 concentrations of N associated with humic and fulvic acids, 182 concentrations of TDN and DON, 158-170 release rates in aquatic environments, 202-206 uptake in aquatic environments, 206-222 Low latitudes, temporal variabihty of DOC in, 695-696 Low-DOC subsurface water mixes, 688 Low-molecular-weight (LMW) organic compounds, 512 Lysis, 726-727
M Macro- and macrozooplankton, 191, 198 selected examples of PER DOM release by, 106-107,108-111 Magnesium sulfate-hydrochloric acid hydrolysis, 276 Magnesium-induced-coprecipitation (MAGIC) method for Fi analysis, 274 Majority view of DOP, 330-334 Malfunctioning microbial loop hypothesis, 136 Marine colloids and trace metals analytical methods, 372-380 isolation of colloidal matter for bulk analysis, 375-380
766 Marine colloids and trace metals (continued) number concentrations of marine colloidal matter, 372-374 biological availability of colloidal bioactive metals, 395-396 chemical form of colloidal metals, 385-388 definition of, 369-372 dissolved organic carbon concentrations, 367-369 fate of colloidal organic carbon phases in ocean waters, 369 metal availability, 368 metal content of marine colloidal matter, 380-385 measurement of colloid reaction rates, 390-394 particulate-based estimates of colloidal metal concentrations, 388-389 sources of metal-complexing colloidal ligands, 389-390 sunmiary, 396-397 Marine DOP concentrations, selected, 288-289 Marine food web dynamics, DOM photochemistry and, 477 Marine microorganisms, 263 Marine sediments carbon/nitrogen values in, 634-635 sediments, values in, 634-635 Mass-balance calculation, 431 Measurement of Dissolved Organic Carbon and Nitrogen in Natural Waters, workshop on, 21 Measurements and distributions ofS^^C and A^'^C of DOC in oceanic systems, 415^29 distribution of in ocean margins, 423-429 distribution of in oceanic systems, 417-423 Measuring DON concentrates, methods for, 154-155 Mechanistic studies of DOM, 476 Meridonial fluxes of DOM in North Atlantic, 738 Meridonial trend in incident solar irradiance, 565 Mesopore protection model, 648-649 Metal content of marine colloidal matter, 380-385 availability, 368 Metal/colloidal interactions, control of by metal-specific reactions, 386-387
Index Metal-complexing colloidal ligands, sources of, 389-390 Methane thiol, 472 Micro- and macrozooplankton, 191, 198 selected examples of PER DOM release by, 107,108-111 Microbial conmiunity structure and DOM utilization, 137 Microbial growth, photoinhibition of photosynthesis and, 570-571 Microbial loop, 24, 92 Microgels, 394 Mid-latitudes, temporal variability of DOC in, 694-695 Millipore HA filters, 15 Minor sulfur species, 490-491 Mississippi plume, 220 Mississippi river DON, 173 Mixed redox sediments, 633 Mixed or no effects of photochemical DOM alteration on microbial DOM utilization, 500-502 Model comparisons of extracellular phytoplankton, 105-106 Modeling DOM chemistry ecosystem modeling studies, 719-743 basin-to-global scale ecosystems modeling, 738-739 coastal and estuarine systems, 741-742 compartments and currencies, 719-723 diagnostic models of North Atlantic, 737-738 inverse models of flows within food webs, 740-741 modeling production of DOM, 723-728 modeling utilization and remineralization of DOM, 728-733 small-scale spatial structure, 742-743 modeling role of DOM in ocean biogeochemistry, 734-737 DOM and atmospheric CO2, 736-737 DOM turnover times, 735-736 early results, equatorial nutrient-trapping, 734-735 major classes of ocean models, 734 planktonic food webs, complexity of, 717-718 Molar stoichiometrics of dissolved inorganic nutrient pools, 303 Molecular composition of DOM, 68-71 Molecular highlights of 1980's, 13
Index Molecular weight characterization of DOP pool, 308-309 Molecular-level analyses of DOM in surface ocean, 70 Monod, basic formulation of, 731 Morphoclimactic zones, water exports from, 583 Multi elemental methods of DOM pools analysis, 51
N N compounds, other organic, 222 N2 fixer, 189-190 NCP, estimate of, 698-699 Negative effects of photochemical DOM alteration on microbial DOM utilization, studies showing, 498-499 Neutral sugars, measured in seawater, 69 Nitrogen and phosphorus, impact of photochemistry on, 473 Nitrogen-to-phosphorus (N:P) ratios for total dissolved matter, 301, 302, 304 NOAA-NODC global ocean search, 280-281 Nonionic macro porous sorbents, 62 Nonlinear dependence of CDOM absorption on salinity, 532-533 Nonprotein amino acid beta-aga, observation of, 639 North Adriatic Sea, investigation in, 316 North Atlantic deep water, oxygen consumption in, 702-703 diagnostic models of, 737-738 Meridonial fluxes of DOM in, 738 North Pacific Ocean, DOC concentrates in, 8 North Pacific Subtropical Gyre (NPSG), DOP in, 297-306 NPSG alternating ecosystem state hypothesis, schematic presentation of, 305 Nucleases and related enzymes, 346 Nucleic acids, DOP in, 315-319 Number concentrations of marine colloidal matter, 372-374 Nutrient depletion and net production of DOC, 701-702 Nutrient trace metals, 382
O Observed DOM dynamics impact of CDOM on underwater light field, 557 local sources and sinks of CDOM, 557-561
767 Ocean, open, see Open ocean Ocean biogeochemistry, modeling role of DOM in, 734-737 DOM and atmospheric CO2, 736-737 DOM turnover times, 735-736 early results, equatorial nutrient-trapping, 734-735 major classes of models, 734 Ocean carbon and nitrogen cycles, role of benthic DOM fluxes in, 641-648 Ocean margins distributions of DOC in, 423^29, 433^39 and interior, exchanges of DOC between, 441^43 sources and inputs of UDOC to, 439-441 Oceanic depth profiles of DOC, analysis of, 44 Oceanic systems, distribution of A^^C and 8^^C from oceanic systems, 417^23 OCGMs with horizontal resolution, 735 OCS, photochemical production of, 471 Open ocean P cycle, 251 vertical mixing and distribution of A^^C-DOC in, 430-433 Open ocean, chromophoric DOM in characterization of DOM chemical composition, 552-554 methods for quantifying, 554-556 operational definition, 549-550 optical properties, 550-552 future advances, needs for, 571-572 global CDOM distribution controls on global CDM distribution, 563-566 global distribution of CDM absorption from SeaWiFS, 561-563 temporal variability of global CDOM distribution, 563 implications for photochemistry and photobiology, 568-571 photochemistry and carbon and sulfur cycles, 569-570 photoinhibition of photosynthesis and microbial growth, 570-571 observed DOM dynamics impact of CDOM on underwater light field, 557 local sources and sinks of CDOM, 557-561 recent research, 547-548 short history, 548-549
768 Open ocean (continued) relationship between DOM and CDOM in open ocean, 567-568 Optical properties of CDOM absorption, 513-526 for different geographic areas, 516-522 fluorescence, 526-531 Optical properties of DOM, 550-552 Organic matter released as DOM due to solubilization of particles, 114-115 Organic molecules during CFF, breakthrough of, 377-378 Organic N compounds, turnover time estimates of DON and, 228-230 Organic P fractions, 253-254, 256-257 Organic solute DOC, 406 Organic sulfur compounds, 467^73 Overflow model of extracellular phytoplankton production, 99 Oxic or mixed redox sediments, 633 Oxidation, sample, 414 Oxygen consumption, photochemical dissolved, 466-467
Pacific Ocean, and %CDOM contributions, 562-563 Pacific water, inflowing, 667 Parameter S, 523 Partial photochemical oxidation, DOP pool characterization by, 313-314 Particles, solubilization of, 727 Particulate carbohydrates (PCHOs), 640 Particulate organic carbon (POC) distributions in discrete water samples, 4-5, 8-9 Particulate-based estimates of colloidal metal concentrations, 388-389 Passive diffusion model of extracellular phytoplankton production, 105 Pee Dee Belemnite (PDB) standard, 407-408 Pelagic marine P-cycle key pools and processes, 262-266 research, early years of, 258-262 Pelagic production of DOM, study of, 595 Percent extracellular release (PER), 98-99 Permanganate oxidation, 275 Permeate, changing analyte concentrations in during processing, 377 Pho regulon, 340
Index Phosphate, photochemical formation of, 493 Phosphoinositol derivatives, 334 Phospholipids, 323-324, 325 Phosphorus and nitrogen, impact of photochemistry on, 473 Phosphorus (P), as essential nutrient for living organisms, 250-253 Phosphorylation of LPS, 333 Photobleaching and photosensitizing properties of seawater, 12 Photochemical and cycling of carbon, sulfur, nitrogen and phosphorus impact of photochemistry on elemental cycles carbon, 458, 460-467 nitrogen and phosphorus, 491 sulfur, 467^73 photochemical transformation of riverine and marsh-derived DOM inputs to sea, 457-458 role of in absorption of UV Hght, 456-457 unresolved questions and future research, 476-478 DOM photochemical reactions involving trace elements, 477^78 DOM photochemistry and marine food web dynamics, 477 extrapolating photochemical rates to global scales, 478 mechanistic studies, 476 photo production and air-sea exchange of important atmospheric trace gases, 477 Photochemical decomposition decompensation of CDOM, 569 Die photo production from DOM, 485, see also Dissolved inorganic carbon as sink for DON, 222-223 Photochemical effects of DOM, 732-733 Photochemical release from DON in whole water, rates of, 224-226 Photochemistry, impact on carbon, 458, 460-467 photochemical dissolved inorganic carbon formation and oxygen consumption, 466-467 sequential photochemical-microbial DOC degradation, 459, 460-475 Photochemistry and photobiology, implications for, 568-571 photochemistry and carbon and sulfur cycles, 569-570
Index photoinhibition of photosynthesis and microbial growth, 570-571 Photodegradation products, main, 458 Photoinhibition of photosynthesis and microbial growth, 570-571 Photolysis of humic-bound organics, 492 Photon correlation spectroscopy, 373 Photoplankton, 722 Phototransformation of recalcitrant DOM, 122-123 Phytoplankton, 188-189 and colloid phase, 395 exudation, 724-725 and zooplankton models, 739 P/ and P-containing compounds in seawater, detection of, 271-279 analysis of P/, 271-274 analysis of TDP, 274-279 Planktonic food webs, complexity of, 717-718 Polymer gel size in seawater as function of time after conventional filtration, 393 Polymeric low-molecular-weight compounds (pLMWDOM),613 Polysaccharides, synthesis of, 333 Pool characterization, 306-334 biologically available P, 329-330 characterization by enzymatic characterization, 309-311 characterization by ^^P NMR, 311-313 direct measurement of DOP compounds, 314-329 majority view of DOP, 330-334 molecular weight characterization of DOP pool, 308-309 by partial photochemical oxidation, 313-314 pools and processes, key, 262-266 Pore water DOC profiles in deep sediment cores, 626-631 DOM in sediment carbon preservation, role of, 648-649 DON concentrations, 632-633 size/reactivity (PWSR) model, 613 Positive effects of photochemical DOM alteration on microbial DOM utilization, studies showing, 490-497 (app) Potentiometric versus manometric methods of measuring seawater, 23-24
769 Process of production of DOM bacterial DOM, origination and transformation of, 115-116 DOM production via cell lysis, 112-113 extracellular phytoplankton production, 98-106 grazing-induced DOM production, 106-112 solubilization of particles, 113-115 Prochlorococcus species, 299, 301 Production, utilization and remineralization of DOP, 334-347 direct utiUzation of DOP, 339-342 DOP interactions with light and suspended minerals, 346-347 DOP production and remineralization, 335-339 enzymes as P-cycle facilitators, 342-346 Production and consumption processes of DOM accumulation of DOM, 133-137 abiotic formation of biologically recalcitrant DOM, 133-134 biotic formation of recalcitrant DOM, 134-135 limitation of bacterial growth and accumulation of biodegradable DOM, 135-136 microbial community structure and DOM utilization, ecological significance of DOM, 92 lability of DOM, 123-133 biologically labile DOM, 128-129 biologically refractory DOM, 126-128 biologically semilable DOM, 129-130 continuum of biological lability, 130-133 production process, 92-116 bacterial DOM, origination and transformation of, 115-116 DOM production via cell lysis, 112-113 extracellular phytoplankton production, 98-106 grazing-induced DOM production, 106-112 solubilization of particles, 113-115 removal process of DOM, 116-123 abiotic removal processes, 122-123 biotic consumption of DOM, 116-122 Production of DOC, net community, 697-702 evidence for net production of DOC, 699-700 nutrient depletion and net production of DOC, 701-702
770 Production of DOC, net community (continued) regional and global estimates for net production of DOC, 700-701 Production of DOM, modeling, 723-728 grazer-associated DOM production, 725-726 lability of DOM produced, 728 lysis, 726-727 phytoplankton exudation, 724-725 solubilization of particles, 727 Prokaryotes, 116-122, 134-135 bacterial carbon demand, 120-122 bacterial growth efficiency, 117-120 Protein-like class, 528 Pt-catalyzed combustion column and procedure, 15 HTCO units for DOC and DON analyses, 21 Published rates of DON release rates of individual organic N compounds in field, 203 summary of, 192-197 Pycnocline, exportable DOC into deep pycnocline, 704, 706 upper Pycnocline, 703-704
Quantifying DOM, methods for, 554-556 Quantifying pool of DOM in sea and other aquatic environments analysis broad community methods comparisons, 41-^W historical perspective, 37 problem, 37-39 small group methods comparisons, 3 9 ^ 1 dissolved organic nitrogen analysis current and future broad community methods comparison, 47-48 historical perspective and analytical problem, 45-^6 small group methods comparisons, 46-47 dissolved organic phosphorus analysis historical perspective and analytical problem, 49-50 small group methods comparison, 50 interest in, 35-37 limits of elemental analyses, 51-52 methods, diagram of, 38 multielemental methods, 51 reference materials, need for continual, 52-54
Index Quantitative assay for total dissolved phosphomonoester P, 310 Quantitative constraints on organic matter cycHng, 24 Quantitative determination of D-ATP, 320 Quantitative estimation of DOP, 271 Quantitative importance of coupled biological-photochemical pathways for photoremineralization of DOM, 459 Quantitative tracers of terrestrial organic matter, use of, 674-675 Quantum yield spectra for photochemical reactions, 569 Quantum yields for OCS photo production, apparent, 471
Radiocarbon ages of DOM in deep Atlantic and Pacific oceans, 82 Radiocarbon-based equivalent age of carbon-containing materials, estimation of, 410 Rapid cycling of bioactive elements with ocean, 24 Recalcitrant DOM, biotic formation of, 134-135 Reference materials for total DOM pools, need for continual, 52-54 Refractory CDOM, 572 Refractory DOM (RDOM), 733 Regional and depth variations in DOP, 281-286 Regional and global estimates for net production of DOC, 700-701 Remineralization of DOP, 335-339 and modeling utiUzation of DOM, 728-733 bacterial utilization of labile DOM, 729-732 models of DOM cycling and turnover, 730 photochemical effects, 732-733 turnover of semilabile DOM, 729 Removal process of DOM, 116-123 abiotic removal processes, 122-123 biotic consumption of DOM, 116-122 Research priorities, 709-710 for DON, 186, 206-207, 223-226 River inputs to coastal zone, 579-588 biodegradabiUty of riverine DOM, 584-588
Index estimates of discharge from river, 580-584 River runoff sources of DOC, 668-671 Riverine DOC, 457-458 DOM, biodegradability of, 584-588 and marsh-derived DOM inputs to sea, photochemical transformation of, 457^58 Rivers in Arctic concentrations of DOC and TOC in, 669 flux of DOC from into Arctic Ocean, 669 Ross Sea, rate of DOC production of, 699
S, parameter, 523-524 value of, 524 S, slope parameter, 550-511 S, values of observed for river end members, 533 Sample processing, preservation and storage, of DOP, 269-270 Sampling DOP, 266-267 Sargasso Sea CDOM observations, 563 seasonal scale, 561 DOC concentrations in, 694-695, 700 western, 695 Sea, DOP in, variation in space, 280-294 C:N:P stoichiometry of dissolved and particulate matter pools, 292-294 in deep sea, concentrations in, 286-292 regional and depth variations in DOP, 281-286 variation in time, 294-306 English Channel, 295-297 North Pacific Subtropical Gyre (NPSG), 297-306 Sea ice, DOC in, 673 SeaWiFS, global distribution of CDOM absorption from, 561-563 Seasonal dependence of CDOM fluorescence, 530 Seasonal variations of DON, 172 Seawater bar and shield representation of dissolved matter in, 252 DOM, origin of, 331
771 methods for extracting DOC from for isotopic analysis sample oxidation, 414 sample processing, 413^14 photochemical loss of DMS in, 468^69 sources of DOC, 671-672 Seawater HMW DOM, analysis of by solid state, 77 Seawater humic substances (SHS), 10 Sediment carbon oxidation rate, 625 Sediment pore waters appendix: description of DOM advection/diffusion/reaction model, 651-653 benthic DOM fluxes in ocean carbon and nitrogen cycles, role of benthic DOC fluxes, 641-644 benthic DON fluxes, 644-646 extent to which benthic DOM fluxes affect composition and reactivity of deep-water DOM, 646-648 conclusions and suggestions for future research, 650 Dissolved organic carbon in sediment pore waters (DOC) advection/diffusion/reaction model for sediment DOC cycling, 617-623 controls on DOC concentrations with depth in surficial sediments, 623-626 general observations, 614-617 pore water DOC profiles in deep sediment cores, 626-631 DOM compositional data, 636-641 amino acids, 638-640 carbohydrates, 640-641 volatile fatty acids (VFAs), 637-638 dissolved organic nitrogen (DON), 631-636 values in marine sediments, 634-635 pore-water DOM in sediment carbon preservation, role of, 648-649 scientific background, 612-614 Semilabile DOM, turnover of, 729 Sephadex-gel-filtration method, 15 Sequential photochemical-microbial DOC degradation, 459, 460-465 evaluating, 460 Shallow sediments, see Surficial sediments Short-term biological events, 696-697
772 Siberian Shelf Seas, DOC concentration versus salinity in, 670 Sinks and sources of CDOM, local, 557-561 and sources, summary of, 679-681 of terrestrial CDOM, 536-539 Sinks for DON, 207-226 heterotrophic versus autotrophic DON utilization, 208-211 literature values of DON uptake in aquatic environments, 212-222 methods for estimating biotic DON uptake, 211-212 photochemical decomposition as sink for DON, 222-223 research priorities, 223-226 Size distribution of DOM, 63-64 Small group methods comparisons, 46-47 of total DOM pools, 3 9 ^ 1 of DOP analysis, 50 Small-scale spatial structure, 742-743 Solar irradiance, incident, 565 Solid-phase extraction for DOM isolation, 61-62 Solid-state cross-polarized magic angle spinning (CPMAS), 78 SolubiUzation of DOM particles, 113-115, 727 Soluble ligand model, 387 Sorption of DOM onto particles, 123 Sources and sinks, 534-536 ofCDOM, local, 557-561 Sources of DOC to Arctic Ocean, 667-674 biological sources within Arctic Ocean, 672-674 and links, summary of, 679-681 river runoff sources, 668-671 seawater sources, 671-672 Space, DOP variations in, 280-294 Spatial and temporal scales, examination of, 533 Spatial variability at basin scale, 687-693 deep-ocean distributions, 689-690 historical data, 692-693 relation to productivity, 690-692 upper ocean distributions, 687-689 Spectral irradiance, decrease, 556 Spectral model coupled to ecosystem model, 733 Stable carbon isotopes and lignin oxidation products, 674-676
Index Stratification in ocean, role in controlling DOC concentrations, 692 Subeuphotic zone DOP measurements, 285 Subsurface maxima in fluorescent CDOM, 558-559 Subtropical mode water formation, 704 Sugars, neutral, measured in seawater, 69, 76-77 Sulfur and carbon cycles, photochemistry and, 569-570 impact of photochemistry on, 467-473 carbonyl sulfide, 468-472 dimethyl sulfide, 468-469 dimethyl sulfoxide, 469 minor sulfur species, 472-473 Surface water and DOC transfer, 599 map with geographic information and schematic circulation of, 666 Surficial sediments, controls on DOC concentrations with depth in, 623-626 Synchronous spectra, 527-528 Syringyl and vanillyl phenols, 676
Tangential-flow or cross-flow ultrafiltration for DOM isolation, 62-63 ultrafiltration and solid-state-cross-polarized magic angle spinning (CPMAS), 312-313 TDP analysis of, 274-279 measurement of, 255 methods, comprehensive, 277-278 Temporal variability of DOC, 693-697 deep ocean, 696 high latitudes, 694 low latitudes, 695-696 mid-latitudes, 694-695 short-term biological events, 696-697 summary, 697 Temporal variability of global CDOM distribution, 563 Terrestrially dominated regions, 534-535 Terrigenous dissolved organic matter (DOM) input of to Arctic Ocean, 665-666, 675 in ocean, fate of, 83-84 mechanisms of removal of, 84
773
Index Thermoclines of major ocean basins, main, 703-704 Three-box scavenging model, 391 Three-dimensional excitation, 527-528 Three-source isotopic mixing models, 437 results, 438 Time, DOP variations in, 294-306 Time series of CDOM absorption coefficient, 564 Total dissolved carbohydrates (DCHOs), 640-641 Total dissolved free amino acids (TDFAAs), 638-639 Total dissolved nitrogen (TDN) in seawater, 13-14 Trace elements DOM photochemical reactions involving, 477-^78 atmospheric trace gases, atmospheric, 477 Trace metals and marine colloids analytical methods, 372-380 biological availability of colloidal bioactive metals, 395-396 chemical form of colloidal metals, 385-388 definition of, 369-372 dissolved organic carbon concentrations, 367-369 measurement of colloid reaction rates, 390-394 particulate-based estimates of colloidal metal concentrations, 388-389 sources of metal-complexing colloidal Ugands, 389-390 sunmiary, 396-397 Transmission electron microscope images of marine colloidal matter from depth profile, 374 Trichodesmium, 190 population, 304 Turnover times DOM, 735-736 estimates of DON and organic N compounds, 228-230 Two- and three-source mass balance models, 429, 432^33 calculations, 429, 437 Two-component steady-state inputs of margin-derived and surface-ocean-derived DOC to deep open ocean, 442
U UDOC, sources and inputs to ocean margins, 439-441 Ultrafiltration (UF), or cross-flow filtration (CFF), 414 Ultraviolet (UV) absorption to detect yellow organic substances, 3 ^ fluorometer, use of, 671 photooxidation, 255, 276-277 UV-based DOC measurements, 20 Underwater light field, impact of CDOM on, 557 Upper ocean distributions of DOC, 687-689 Upper pycnocline, exportable DOC into, 703-704 Urea, 174, 205, 219 Utilization and remineralization of DOM, modeling, 728-733
Vanillyl and syringyl phenols, 676 Ventilation of deep pycnocline, 704, 706 Vertical diffusive mixing into water column, 710 Vertical distributions of D-DNA, 317, 318 Vertical mixing and distribution of A^^C-DOC in open ocean, 430-433 Vertical profiles of DON, 157, 171-172 Vertical Transport and Exchange (VERTEX) program, 299 Viral lysis, 112 Viruses, 198-200 Vitamins, DOP in, 326-327 Volatile fatty acids (VFAs), 637-638 Volume-based concentrations of colloidal trace metals, 84
W Water, deep, see Deep water Water column overturn, distribution of sites of, 698 Water discharge and DOC, relationship between, 586-588 Water insoluble or hydrophobic DOP compounds, 341 Water masses and circulation, 665-667
Index
774 Wavelength dependence for photochemical loss ofDMS,469 WCO, see Wet chemical oxidation Wet chemical isolation, 201, 204 Wet chemical oxidation (WCO), 21 method, 37, 38-39, 45, 51 World rivers drainage area, total discharge volume and carbon fluxes in, 581-582 major role of in global water cycle, 580
X XAD ion, exchange resins, 550 resins, 414 XAD-2, 9 retaining seawater on, 62 XAD-8 resins, 62 Zooplankton and phytoplankton models, 739
IGY @ 40° N
u 1000
'nwmJ\MSPM^^9n|^^ f^l
A'^i I ^KK^'
Bf.|MM|fc^ W -9HH
^^^H
g^2000 ^3000
I -
-
^^9^1
0.25
0.1
-
5000 1
TOW
0.3
0.15
Q H^^t^SK^^ \ * \
10.35
0.2
- ri-* • . Hi?-. • w^M
4000
0.2 0.4
1
60W
1
1
SOW
i,
1
40W
•
.,
1
30W
1
1
20W
1 ^
.1.,,
10W
Figure 10 Phosphorus concentration (|xM) versus ocean depth contours of TDP (top), SRP (center) and DOP (bottom) for samples collected on a transect at 40°N latitude in the Atlantic Ocean during the International Geophysical Year (IGY). TDP was measured using the method of Harvey (1948). These data were originally presented by McGill (1963) and were obtained by using the NOAANODC World Wide Web-based online search program.
IGY @ 24.25'' S
W''^^^^
^3000 CO
u
1000
v^i
^ %^m^,..^^^KmM^^^^'i ^
i^'-
2000
"
«i^l
f
a. 3000 O Q
4000
5000 AQ\N
lA. IniJL.f 30W
i 20W
10W
/IJ wW 0
_ 10E
Figure 11 As in Figure 10, except for a transect at 24.25°S latitude.
Surface DOP 70 N
65N
60 N
55 N
SON
Figure 13 Compilation of near surface ocean DOP concentrations (|LLM P ) for samples collected in the Baltic Sea and surrounding coastal regions. These data were obtained from the NOAA-NODC using their World Wide Web-based online search program and are neither time-averaged nor contemporaneous measurements. Stations are indicated by white circles.
Surface DOP 10N
11.28
11.12
0.96
10S
10.80
15S
0.64
20 S
10.32
25 S
30 8
100W
95W
90W
85W
SOW
75W
70 W
•••o.oo
Figure 14 As in Figure 13, except for samples collected off the west coast of South America.
Temperature
(C)
DOC (/zmolkg"^)
(m) q^daa
a^dom (325nm, m"^)
-r-
0)
'S & ^
5 o r^
CIH
73
O
cd
W ) ^ ^ CO
'^ ^ •l ^ . b ^ ^ S (D §.co "^ o
0^
CH
(0
(U
S5 8.S c3
C/5
o .-s y^
X>
^ b
^< PH
i
13 o
r^
c«
o ^ T3
^
CO
1 ^ ^
"ti
SJ flS C/,5 D
0)
00 O)
05
.52 &.
g ^ '^.i ^ ao CIH
c^
^ ^
in
3 3
rin
73
o
^
D
J::^ c)
^ IL
2 a
o
^_. c^
^1 irj 5-1
O +^
fl
3 <^
^H 3
O
a
X
^3
73 V-i
w
o o 'o
^^
43
^ 8
a
-a a o .2
ffj
WD - O
S g
-60
0
60
120
Longitude E Figure 5 Global distribution of (A) absorption coefficient of CDM at 443 nm (ensemble mean estimates using data from Sep. 1997 to Jun. 2000), (B) the % contribution of CDM to total non-water light absorption based on SeaWiFS observations of ocean color spectra (Siegel et al, in press), and (C) global surface ocean DOC distribution (^mol I'l) inferred from global surveys of DOC and empirical relationships between DOC and sea surface temperature (D. Hansell, pers. commun. 2000, Siegel et al, in press). Central gyre regions where DOC accumulates are typically regions of lower CDM concentration (Fig. 5), as increased stratification leads to increased bleaching of CDM. (Panel C courtesy of D. Hansell and C. Pequignet.)
60°S
40^8
20°S
Equ
20*'N
60^8
WS
20^3
Equ
20°N
Figure 2 Meridional sections of DOC in the Southern Hemisphere. (A) the central Indian Ocean nominally along 80°E; (B) the western South Pacific along 170°W; and (C) the eastern South Pacific along 1057110°W. Note the uplifting of contours in the Equatorial Pacific and surface accumulations in the subtropical gyres. See Fig. 1 for locations of sections. Labels on the figures indicate ocean section designations by the World Ocean Circulation Experiment (WOCE). In the section from the eastem South Pacific (C), the location of the Equatorial Undercurrent and its contribution to equatorial upwelling and downwelling is indicated. The approximate position of the main thermocline is indicated by the dashed line, and the general circulation patterns are shown by the arrows [low DOC Circumpolar Deep Water (CDW) upwells to the surface; Antarctic Intermediate Water (AAIW) subducts toward the equator; convergence in the gyres leads to downwelling of DOC enriched surface waters].
130°W Longitude
Figure 3
200 H :S400^ Q.
s
60045°W
35°W Longitude
800-
70°W
60°W
50°W 40°W Longitude
Figure 4
15°W
25^
30°W
20°W
Figure 6 Time-series of DOC in the western Sargasso Sea near Bermuda (top) and the North Pacific near Hawaii (bottom). Note the differences in DOC contour scales. SeasonaHty in the Sargasso Sea is evident both in the DOC and temperature contours (shallow contour is 24°C; deeper contour is 20°C), while DOC seasonahty is absent near Hawaii. Data from the Bermuda Atiantic Time-series Study (BATS) and Hawaiian Ocean Time-series (HOT) programs, respectively.
Figure 3 Zonal section of DOC along the equator in the Pacific Ocean. The data are a compilation collected in Autumn 1992 (Peltzer and Hayward, 1996; 110°W to 140°W) and in Autumn 1994 (Hansen etal, 1997b; 150°Wto 170°E), during similar ENSO states. Note the shoaling of low DOC, Equatorial Undercurrent (EUC) to the east. The high concentrations west of the dateline are associated with the Western Pacific Warm Pool (WPWP). The dashed lines serve to delineate the various water masses described. Figure 4 Zonal sections of DOC in the North Atlantic Ocean, along 43°-49°N (top) and 24°N (bottom). Labels on the figures indicate ocean section designations by WOCE.
400 Juiy94
Oct.94
Jan.95
Apr.95
July95
DOC Stock (molesC/m^)
U o Ho O
o
Q E
8
July94
Oct.94
Jan.95
Apr.95
July95
Figure 7 Top: DOC, overlain by isotherms, during 12 months (July 1994-July 1995) in the Sargasso Sea at the BATS site. Note the seasonal overturn of the water column, the downward mixing of DOC, and the summer maximum in DOC associated with high water column stability. Bottom: DOC stocks (moles m-^) during the period. Note the increase in DOC stock during the period of overturn, both over the upper 250 m (reflecting total net DOC production) and at 100-250 m (reflecting export of DOC to those depths).