SEARCH

SEARCH BY CITATION

Keywords:

  • remineralization ratios;
  • nutrient cycles;
  • global ocean circulation;
  • latitudinal change;
  • GEOSECS;
  • WOCE

Abstract

  1. Top of page
  2. Abstract
  3. 1. Introduction to a New Three-End-Member Mixing Model
  4. 2. Steps in the Model Calculation
  5. 3. Results and Discussion
  6. 4. Overall Discussion
  7. 5. Summary and Conclusions
  8. Acknowledgments
  9. References
  10. Supporting Information

[1] A new three-end-member mixing model is introduced to obtain remineralization ratios of organic matter in the water column. Remineralization ratios (P/N/Corg/−O2) of organic matter in the deep water column change systematically from the northern Atlantic to the Southern Oceans, then to the equatorial Indian and the northern Pacific oceans, more or less along the global ocean circulation route of deep water. Average remineralization ratios of organic matter for the northern Atlantic Ocean are P/N/Corg/−O2 = 1/(16 ± 1)/(73 ± 8)/(137 ± 7), and for the Southern Oceans P/N/Corg/−O2 = 1/(15 ± 1)/(80 ± 3)/(133 ± 5). Those values are similar to the traditional Redfield ratios of P/N/Corg/−O2 = 1/16/106/138 for marine plankton, except for the low Corg/P ratio. Average remineralization ratios for the equatorial Indian Ocean are P/N/Corg/−O2 = 1/(10 ± 1)/(94 ± 5)/(130 ± 7), and for the northern Pacific Ocean P/N/Corg/−O2 = 1/(13 ± 1)/(124 ± 11)/(162 ± 11). The apparent low N/P ratio for both ocean basins suggests that organic nitrogen was converted partly into gaseous N2O and N2 by bacteria through nitrification/denitrification processes in a low-oxygen or reducing microenvironment of organic matter throughout the oxygenated water column. The actual N/P ratio of remineralized organic matter is probably around 15 ± 1. The −O2/Corg ratio of remineralized organic matter also decreases systematically along the global ocean circulation route of deep water, indicating changes in relative proportions of biomolecules such as lipids, proteins, nucleic acids, and carbohydrates. No temporal trends of remineralization ratios are detected when comparing the results obtained by GEOSECS and WOCE data sets.

1. Introduction to a New Three-End-Member Mixing Model

  1. Top of page
  2. Abstract
  3. 1. Introduction to a New Three-End-Member Mixing Model
  4. 2. Steps in the Model Calculation
  5. 3. Results and Discussion
  6. 4. Overall Discussion
  7. 5. Summary and Conclusions
  8. Acknowledgments
  9. References
  10. Supporting Information

[2] Redfield [1958] proposed an idealized molar formula for an average marine plankton, i.e., (CH2O)106(NH3)16(H3PO4). Oxidation or remineralization of one mole of this idealized plankton can be represented by the following overall net reaction: (CH2O)106(NH3)16(H3PO4) + 138 O2 [RIGHTWARDS ARROW] 106 CO2 + 16 NO3 + H2PO4 + 17 H+ + 122 H2O. Molar ratios among remineralized P, N, Corg, and consumed O2 (P/N/Corg/−O2) are called remineralization ratios or Redfield ratios. Remineralization ratios of P/N/Corg/−O2 = 1/16/106/138 for the idealized marine plankton are known as the “traditional Redfield ratios”. Remineralization ratios are useful in explaining the coupled nature of nutrients and carbon cycles in the oceans. They are also important constants needed to estimate conservative tracers such as “NO” and “PO” [Broecker, 1974; Broecker et al., 1985]; the preformed nitrate, phosphate, and total dissolved inorganic carbon (NO30, PO40 and DIC0) [Chen and Pytkowicz, 1979; Li et al., 2000]; and the extent of denitrification and nitrogen fixation [Gruber and Sarmiento, 1997; Deutsch et al., 2001] in the oceans. It has been assumed that remineralization ratios remain constant over the global oceans. The availability of high quality data of nutrients and carbon chemistry as a result of WOCE/JGOFS global ocean survey provides an opportunity to reevaluate the remineralization ratios of organic matter and challenge the validity of their uniformity in the world oceans. Possible spatial and temporal changes of remineralization ratios in the oceans are the subject of the present study.

[3] Li et al. [2000] presented a new two-end-member mixing model to estimate remineralization ratios P:N:Corg:−O2 = 1:(rp/rn):(rp/rc):rp. Here, rp, rn, and rc are respectively the molar ratios of consumed O2 to remineralized P, N, and Corg during oxidation of organic matter in the water column, i.e. rp = −O2/P, rn = −O2/N, and rc = −O2/Corg. The model essentially involves a multiple-variable linear regression (hereafter simplified as multiple regression) of hydrographic data of water samples within the linear segment of a chosen θ-S (potential temperature versus salinity) plot to obtain rp, rn, and rc values. Hydrographic variables include potential temperature (θ), dissolved oxygen (O2) along with dissolved inorganic phosphate (PO4) or dissolved nitrate (= NO3 + NO2), or DA (= DIC-Alkalinity/2). The advantage of this model over earlier ones [Takahashi et al., 1985; Broecker et al., 1985; Minster and Boulahdid, 1987; and Boulahdid and Minster, 1989] is that it requires no knowledge of the end-member values for each variable. As noted by Minster and Boulahdid [1987], the way end-member values were chosen in earlier models could be problematic. The model by Anderson and Sarmiento [1994] also requires no knowledge of end-member values. However, their assumption that ΔCinorg (carbon increment from dissolution of carbonates) divided by ΔP (phosphate increment from oxidation of organic material) is constant along any neutral surface in their model is unrealistic, especially in water masses above the carbonate saturation depth. As summarized by Morse and Mackenzie [1990], the calcite saturation depth is about 4000 ± 500 m in the Indian and Pacific Oceans and 4500 ± 500 m in the Atlantic Ocean. In a saturated or oversaturated water mass, the formation processes of carbonates are completely decoupled from the oxidation process of organic matter. Only in some microenvironments may the oxidation of organic matter enhance the dissolution of carbonates [Troy et al., 1997].

[4] Hupe and Karstensen [2000] introduced a linear inverse mixing model with multiple end-members to estimate remineralization ratios in the Arabian Sea. However, the model requires knowledge of end-member values for each variable. The choice of proper end-member characteristics is always subjective, especially for nonconservative tracers. Shaffer [1996] fitted global averaged profiles of phosphate, nitrate, and dissolved oxygen to his high-latitude-exchange/interior-diffusion-advection model to estimate remineralization ratios, and he concluded that remineralization ratios change with depth. Using a new diffusion-advection model with source or sink terms and allowing local exchange across neutral surfaces, Shaffer et al. [1999] obtained a similar conclusion but the remineralization ratios reach constant values below 1500 m. The model by Shaffer et al. [1999] also requires a prior knowledge of end-member characteristics. In contrast, Broecker et al. [1985], Peng and Broecker [1987], and Anderson and Sarmiento [1994] concluded that remineralization ratios do not change much with depth below 400 m.

[5] To use available data more effectively and to cover a greater portion of ocean regions for the evaluation of remineralization ratios, a logical extension of the two-end-member mixing model of Li et al. [2000] is to introduce the additional conservative tracer salinity (S) and construct a three-end-member mixing model. In this model, selected data need not be confined in the linear segment of the θ-S diagram as required for any two-end-member mixing model. The model also requires no prior knowledge on the end-member values for each variable.

[6] The following mass balance equations for three conservative tracers, θ, S, and (NO), hold in a water sample that resulted from three-end-member mixing:

  • equation image
  • equation image
  • equation image
  • equation image

Here θ, S, O2, and NO3 are observed values for the water sample; subscripts 1, 2, and 3 represent the three end-members; fi is the fraction of the end-member i (= 1, 2, and 3) in the sample; (NO)i is the concentration of conservative tracer “NO” (= O2 + rn · NO3; here rn is assumed constant [Broecker, 1974]) for the end-member i. First, we solve for f1, f2, and f3 from equations (1a) to (1c) in terms of observed θ and S, and end-member values θi and Si; then substitute the solutions into equation (1d). The result is a linear equation:

  • equation image
  • equation image

Note that the coefficients α0, α1, α2, and D are all functions of θi, Si, and (NO)i (i = 1, 2, and 3) for the three end-members, and are unknown but stay constant during the mixing process. One may regard end-member values θi, Si, and (NO)i as mathematical dummies, and they need not have explicit values.

[7] By multiple regression of O2, θ, S and NO3 concentration data from a properly selected hydrographic transect that contains three end-members (see the next section), one can estimate the regression coefficients α0, α1, α2, and rn in equation (2a). However, our focus is only on the value of rn. As discussed by Li et al. [2000], one can rewrite equation (2a) to make NO3 a dependent variable, i.e.,

  • equation image

[8] By multiple regression, one can also obtain independently 1/rn and thus its inverse, designated here as rn′. The rn and rn′ values are always not equal, unless the square of the multiple correlation (R2) is one. The results given in this paper are all averaged values, i.e., (rn + rn′)/2 with the estimated uncertainty of ±[(rn − rn′)/2] or ±[(Δrn)2 + (Δrn′)2]1/2/2, whichever is larger. Here, Δrn and Δrn′ are the standard errors for rn and rn′, given by any multiple regression program for personal computer (in our case, we used the Statistical Package for the Social Sciences, SPSS). When the square of the multiple correlation (R2) is 0.98 or higher for the multiple regression, the rn and rn′ values always overlap each other within the estimated standard errors.

[9] Similarly, replacing “NO” in equation (1d) with another conservative tracer “PO” (= O2 + rp · PO4 [Broecker, 1974]), and NO3 with PO4, one obtains the following:

  • equation image

where A's are similar in formula to α's given beneath equation (2a), except replacing end-member values (NO)i with (PO)i. Again, by multiple regression of O2, θ, S, and PO4 data, one can obtain rp, rp′ and (rp + rp′)/2.

[10] When O2 data are missing, the rp/rn (= N/P) ratio still can be estimated by multiple regression of NO3, θ, S, and PO4 data according to the following equation, which is obtained by eliminating O2 from equations (2a) to (2c):

  • equation image

[11] As shown by Li et al. [2000], the following relationships are valid for total dissolved inorganic carbon (DIC) and total alkalinity (Alk) in a water sample:

  • equation image
  • equation image
  • equation image

where x and y are respectively cumulative increments of DIC from dissolution of carbonate minerals and from oxidation of organic matter during the transport of a water sample from its source region to the current position; “a” = H+/Corg (molar ratio of the produced H+ to the remineralized Corg during oxidation of organic matter) = (N + P)/Corg = rc/rn + rc/rp ≈ rc/rn (since rc/rn ≫ rc/rp); DIC0, Alk0, and O20 are the initial values at the source region. If O20 is the saturated oxygen concentration at the given S and θ of the water sample in equilibrium with air, then (O20 − O2) by definition becomes AOU (apparent oxygen utilization), and DIC0 and Alk0 are the preformed DIC and alkalinity.

[12] By assuming O20 to be equal to the saturated oxygen in equilibrium with air, and eliminating x and y from equations (3a) to (3c), one obtains the following:

  • equation image

where DA = DIC – Alk/2, and DA0 (= DIC0 − Alk0/2) is the preformed DA.

[13] The new variable DA (combining capital letters of DIC and Alk) can be visualized as alkalinity-corrected DIC [Li et al., 2000]. Like any other preformed parameter, DA0 is also a conservative tracer that changes in water samples only through water mass mixing before the industrial time. Strictly speaking, DA0 is no longer a conservative tracer for the shallow part of the ocean where DIC0 changes with time due to anthropogenic CO2 inputs. However, the increment rate of DIC in the surface mixed layer is only about 0.05% per year [Winn et al., 1998]. Furthermore, the regionwide vertical mixing coefficient for the oceanic thermocline below the surface mixed layer is also finite (0.5 to 2.8 cm2/s, which corresponds to 56 to 94 m of average mixing depth per year [Li et al., 1984]). Therefore, DA0 still can be considered as near constant in the lower thermocline waters during a short period of one to five years, and it contains both the preindustrial and cumulative anthropogenic CO2 components. DA0 for any deep-water sample, where anthropogenic CO2 input is negligibly small [Sarmiento et al., 1992; Wanninkhof et al., 1999], can be expressed by

  • equation image

if the water sample is formed by three-end-member mixing.

[14] By substituting equations (1a), (1b), (1c) and (4b) into equation (4a), one obtains

  • equation image

where β0, β1, and β2 are similar in formula to α0, α1, and α2 in equation (2a), except (NO)i are replaced by (DA0)i; and β3 = (a/2 + 1)/rc, thus rc = 1/[β3 − 0.5(1/rn + 1/rp)] ≈ 1/[β3 − 0.5/rn]. As will be shown in Table 1, the range of rn is between 8 and 13, therefore the change in rn will not affect the rc value much. Again, one can estimate β3, thus rc as well as rc′ and (rc + rc′)/2, by multiple regression of DA, θ, S, and AOU data for deep waters from lower thermocline and below, using equation (4c).

Table 1a. Remineralization Ratios Based on the WOCE and GEOSECS Sections in the Atlantic Oceana
SectionLatitude IntervalStation Numberθ, °Crp (−O2/P)rn (−O2/N)rp/rn (N/P)β3rc (−O2/Corg)rp/rc (Corg/P)
  • a

    No bottom waters in the Atlantic Ocean were used in the model calculation, because the change of oxygen concentration in bottom waters can be explained solely by water mass mixing.

Redfield   1388.6160.831.30106
a2042°N–33.5°N12–333–18142 ± 28.6 ± 0.116.5 ± 0.30.66 ± 0.021.67 ± 0.0685 ± 2
 33°N–13.5°N34–633–6132 ± 28.5 ± 0.215.5 ± 0.40.58 ± 0.041.93 ± 0.1568 ± 2
   6–12145 ± 38.7 ± 0.216.7 ± 0.50.62 ± 0.021.79 ± 0.0781 ± 3
 13°N–7°N64–923–5.2140 ± 39.6 ± 0.314.6 ± 0.60.52 ± 0.022.15 ± 0.1065 ± 2
a176°N–0°175–2354.2–12125 ± 37.9 ± 0.115.8 ± 0.40.64 ± 0.011.75 ± 0.0472 ± 2
 0°–25°S88–1744–12140 ± 28.5 ± 0.116.5 ± 0.30.67 ± 0.011.65 ± 0.0385 ± 2
 25°S–40°S33–870.2–3124 ± 38.0 ± 0.115.5 ± 0.40.74 ± 0.011.48 ± 0.0384 ± 2
   3–12127 ± 18.6 ± 0.114.8 ± 0.20.66 ± 0.011.67 ± 0.0376 ± 1
 40°S–50°S5–320.4–2.6126 ± 48.5 ± 0.214.8 ± 0.60.74 ± 0.011.48 ± 0.0385 ± 3
   2.6–12126 ± 28.8 ± 0.214.3 ± 0.40.64 ± 0.011.73 ± 0.0373 ± 2
 average  133 ± 88.6 ± 0.415.5 ± 0.80.65 ± 0.061.73 ± 0.1977 ± 7
Geosecs42°N–12°N27–382–6132 ± 57.9 ± 0.316.7 ± 0.9   
   6–18127 ± 48.1 ± 0.215.7 ± 0.6   
 9°N–8°S39–491.8–3.8116 ± 18.9 ± 0.113.0 ± 0.20.70 ± 0.061.56 ± 0.1574 ± 1
   3.8–12120 ± 38.8 ± 0.113.6 ± 0.40.71 ± 0.041.54 ± 0.1078 ± 2
 8.5°S–24°S50–571.8–4121 ± 29.0 ± 0.113.4 ± 0.3   
   3.2–12133 ± 28.9 ± 0.114.9 ± 0.30.69 ± 0.021.59 ± 0.0584 ± 2
 27°S–49°S59–682.6–12128 ± 19.0 ± 0.114.2 ± 0.20.71 ± 0.021.54 ± 0.0583 ± 1
 average  126 ± 68.8 ± 0.614.5 ± 1.20.70 ± 0.011.55 ± 0.0282 ± 5
Table 1b. Remineralization Ratios Based on the WOCE and GEOSECS Sections in the Indian Oceana
SectionLatitude IntervalStation Numberθ, °Crp (−O2/P)rn (−O2/N)rp/rn (N/P)β3rc (−O2/Corg)rp/rc (Corg/P)
  • a

    Character “b” after a number in the θ column represents the potential temperature of bottom water. Without “b” means no bottom water samples were used in the model calculation. Symbols: *, θ ≥ (S − 34.74) × 3.8/0.24 + 1.8; **, θ ≥ (S − 34.72) × 4/0.12 + 2 in the θ-S diagram, i.e., samples above the Circumpolar Water.

Redfield   1388.6160.831.30106
i8N6°N–0.5°S279–3031b–9120 ± 313.0 ± 0.39.2 ± 0.30.79 ± 0.041.34 ± 0.0790 ± 3
0.5°S–18.5°S304–3430.9b–8129 ± 212.6 ± 0.410.2 ± 0.40.78 ± 0.041.36 ± 0.0795 ± 3
19°S–29.5°S344–3932–8129 ± 110.4 ± 0.112.4 ± 0.20.67 ± 0.031.62 ± 0.0880 ± 1
30°S–33°S394–4422–8127 ± 210.4 ± 0.212.2 ± 0.30.69 ± 0.071.57 ± 0.1781 ± 2
i8S30°S–51°S4–492–11133 ± 39.5 ± 0.314.0 ± 0.50.66 ± 0.031.66 ± 0.0880 ± 3
51°S–63.5°S50–84*130 ± 38.5 ± 0.215.3 ± 0.50.72 ± 0.031.52 ± 0.0785 ± 3
i9N20°N–2.5°N213–2770.8b–11131 ± 113.0 ± 0.210.1 ± 0.20.77 ± 0.021.37 ± 0.0495 ± 2
2°N–18.5°S172–2120.8b–8142 ± 113.4 ± 0.210.6 ± 0.20.76 ± 0.021.39 ± 0.04102 ± 2
19°S–31°S148–1712–10134 ± 110.5 ± 0.212.8 ± 0.20.69 ± 0.011.57 ± 0.0386 ± 1
i9S34.5°S–50°S114–1472–11152 ± 29.8 ± 0.215.5 ± 0.40.60 ± 0.051.83 ± 0.1783 ± 2
50°S–64.5°S85–113**129 ± 28.5 ± 0.115.2 ± 0.30.70 ± 0.051.57 ± 0.1282 ± 2
Geosecs15°N–10°S440–4500.7b–12129 ± 211.5 ± 0.211.2 ± 0.30.73 ± 0.021.46 ± 0.0488 ± 2
10°S–30°S436–4390.4b–12137 ± 110.2 ± 0.113.4 ± 0.20.71 ± 0.011.52 ± 0.0290 ± 1
39°S–63°S429–4340–2.6126 ± 29.0 ± 0.114.0 ± 0.3   
Table 1c. Remineralization Ratios Based on the WOCE and GEOSECS Sections in the Pacific Oceana
SectionLatitude IntervalStation Numberθ, °Crp (−O2/P)rn (−O2/N)rp/rn (N/P)β3rc (−O2/Corg)rp/rc (Corg/P)
  • a

    Character “b” in the θ column represents the potential temperature of bottom water.

  • b

    Samples above the Circumpolar Water.

Redfield   1388.6160.831.30106
p14N52°N–45°N20–341b–4.3176 ± 413.1 ± 0.413.4 ± 0.50.76 ± 0.031.39 ± 0.06127 ± 5
45°N–26.5°N35–721b–12165 ± 312.0 ± 0.213.8 ± 0.30.70 ± 0.011.53 ± 0.03108 ± 3
26°N–11.5°N73–1020.8b–6162 ± 411.6 ± 0.214.0 ± 0.40.83 ± 0.021.28 ± 0.03127 ± 4
  6–12136 ± 112.0 ± 0.111.3 ± 0.10.82 ± 0.011.29 ± 0.02105 ± 1
11°N–4°N103–1260.8b–10.5173 ± 612.7 ± 0.413.7 ± 0.60.85 ± 0.031.24 ± 0.05140 ± 7
p15N54°N–41°N2–311b–6174 ± 212.8 ± 0.113.6 ± 0.20.71 ± 0.011.50 ± 0.02116 ± 2
40°N–26.5°N32–591b–12165 ± 412.5 ± 0.213.2 ± 0.40.71 ± 0.021.50 ± 0.05110 ± 3
26°N–12.5°N66–870.8b–6152 ± 611.7 ± 0.413.0 ± 0.70.83 ± 0.031.28 ± 0.05119 ± 6
12°N–0°88–1110.65b–12170 ± 712.9 ± 0.413.2 ± 0.70.80 ± 0.031.32 ± 0.05129 ± 7
0°–12°S115–1360.65b–6165 ± 412.7 ± 0.113.0 ± 0.30.80 ± 0.021.32 ± 0.04125 ± 5
p15S9.5°S–15.5°S146–1820.6b–2173 ± 311.9 ± 0.414.5 ± 0.60.87 ± 0.051.21 ± 0.07143 ± 5
  2–12128 ± 18.9 ± 0.114.4 ± 0.20.68 ± 0.011.61 ± 0.0379 ± 1
16°S–42°S89–1450.5b–2158 ± 412.0 ± 0.313.2 ± 0.50.90 ± 0.081.17 ± 0.11135 ± 5
  2–10123 ± 39.2 ± 0.213.4 ± 0.40.62 ± 0.031.78 ± 0.1069 ± 2
43°S–53°S61–871.8–11.3148 ± 19.9 ± 0.114.9 ± 0.20.59 ± 0.011.87 ± 0.0479 ± 1
53°S–64°S36–60−1–8b136 ± 19.2 ± 0.114.8 ± 0.20.62 ± 0.011.78 ± 0.0376 ± 1
p14S53°S–67°S4–321.7–7.6136 ± 49.3 ± 0.214.6 ± 0.50.66 ± 0.021.66 ± 0.0682 ± 3
P1816°S–28.5°S82–1071.5–12123 ± 28.7 ± 0.214.1 ± 0.40.61 ± 0.021.82 ± 0.0767 ± 2
29°S–40.5°S58–811.5–6.5129 ± 210.1 ± 0.212.8 ± 0.30.63 ± 0.011.73 ± 0.0374 ± 2
54.5°S–67°S10–332–7131 ± 28.8 ± 0.314.9 ± 0.60.66 ± 0.031.67 ± 0.0979 ± 3
Geosecs50°N–40°N 1b–8176 ± 111.6 ± 0.115.2 ± 0.20.74 ± 0.021.44 ± 0.04122 ± 1
35°N–25.5°N 0.9b–12158 ± 212.1 ± 0.113.1 ± 0.20.80 ± 0.031.32 ± 0.05119 ± 2
25°N–10°N 0.8b–6152 ± 411.6 ± 0.313.1 ± 0.5   
9°N–13.5°S 0.6b–12159 ± 511.8 ± 0.213.5 ± 0.50.83 ± 0.051.27 ± 0.08125 ± 4
15°S–42°S 2–12152 ± 510.0 ± 0.215.2 ± 0.6   
46.5°S–53°S 2–12137 ± 39.9 ± 0.313.8 ± 0.5   
54°S–69°S 2–6148 ± 38.8 ± 0.316.8 ± 0.70.70 ± 0.071.56 ± 0.1795 ± 4

[15] One needs not assume O20 to be equal to the saturated oxygen in equilibrium with air, and rewrite equation (4c) to

  • equation image

Substituting O20 = f1 · (O20)1 + f2 · (O20)2 + f3 · (O20)3 into equation (4d), one obtains

  • equation image

where constants B0, B1, and B2 are bulky functions of θi0, Si0, DAi0, O20i, for three end-members and β3.

[16] By multiple regression of DA, θ, S, and O2 data using equation (4e), one can estimate β3, thus rc, as well as rc′ and (rc + rc′)/2. In practice, both equations (4c) and (4e) give the same β3 values within the estimated uncertainty. Therefore, the assumption that O20 represents the saturated oxygen in equilibrium with air is a good approximation.

[17] In short, the advantage of the present three-end-member model is that the data points selected for the model need not be confined in an isopycnal horizon [Takahashi et al., 1985; Broecker et al., 1985; Minster and Boulahdid, 1987] or neutral surface [Anderson and Sarmiento, 1994]. The data points can be expanded greatly (easily to more than one hundred to several hundreds data points for each run) to include those in a vertical cross section or in a large volume of a given regional basin without having to worry about distinguishing horizontal/vertical or isopycnal/diapycnal mixing.

2. Steps in the Model Calculation

  1. Top of page
  2. Abstract
  3. 1. Introduction to a New Three-End-Member Mixing Model
  4. 2. Steps in the Model Calculation
  5. 3. Results and Discussion
  6. 4. Overall Discussion
  7. 5. Summary and Conclusions
  8. Acknowledgments
  9. References
  10. Supporting Information

[18] To illustrate how these model calculations are conducted, we use the WOCE (World Ocean Circulation Experiment) hydrographic data from the transect p15 (roughly along the meridian 170°W) of the Pacific Ocean as an example. First, we grouped the stations in such a way that data points in a θ-S plot for a chosen latitudinal interval can be easily explained by mixing of three end-members. For example, in Figure 1a (stations 2 to 31 within latitudinal interval between 54°N and 41°N), data points with θ between 1° and 6°C (dotted horizontal lines) can be formed by mixing end-members I, II, and III. Similarly, in Figure 1b, data points with θ between 1 and 12°C can be formed by the mixing of end-members I, III, and IV, and so forth. The end-members of water masses depicted in Figures 1a to 1h are as follows. I = North Pacific Bottom Water, I* = South Pacific Bottom Water, Ic* = Antarctic Circumpolar Water, II = Subarctic Water, II* = Subantarctic Water, III = North Pacific Intermediate Water, III* = Antarctic Intermediate Water, IV = North Pacific Shallow Salinity Minimum Water, IV* = South Pacific Shallow Oxygen Minimum Water, and V* = South Pacific Subtropical Salinity Maximum Water. One end-member may appear more than once in Figures 1a to 1h, but the end-member values for θ and S may not be always the same due to mixing effect of other end-members. For example, end-member III* (Antarctic Intermediate Water) appears in Figures 1d to 1h, but its θ and S values change gradually from south to north due to mixing inputs from end-members I*, Ic*, II*, and IV*. Therefore, in each chosen latitudinal interval, end-member III* has its own local end-member characteristics. Fortunately, our model requires no knowledge on end-member values of θ, S, and other variables for any chosen latitudinal interval.

image

Figure 1. The θ-S plots of the WOCE Pacific p15 transect data from different latitudinal intervals (north to south). Roman numerals represent various end-members of water masses in the Pacific Ocean: I = North Pacific Bottom Water, I* = South Pacific Bottom Water, Ic* = Antarctic Circumpolar Water, II = Subarctic Water, II* = Subantarctic Water, III = North Pacific Intermediate Water, III* = Antarctic Intermediate Water, IV = North Pacific Shallow Salinity Minimum Water, IV* = South Pacific Shallow Oxygen Minimum Water, and V* = South Pacific Subtropical Salinity Maximum Water.

Download figure to PowerPoint

[19] The data in Figure 1e are further separated into three temperature intervals, i. e., θ = 0.6 ∼ 2.1, θ = 2.1 ∼ 12, and θ = 12 ∼ 22°C. Hydrographic data from the first two intervals are plotted in Figures 2 and 3. The obvious outliers in those plots, especially in the θ-S (Figure 2a) and PO4-NO3 plots (Figures 2d and 3d), can be eliminated before regression analysis. The PO4-NO3 plot is useful in testing the internal consistency of phosphate and nitrate data and in detecting any additional new end-member with a low N/P ratio due to localized denitrification in the water column or at the nearby water-sediment interface [Broecker and Peng, 1982; Gruber and Sarmiento, 1997]. Figure 4 exemplifies the good correlation between the observed and calculated O2 from equation (2a), and between the observed and calculated DA from equation (4c) for the data from three different temperature segments in Figure 1e. By eliminating obvious outliers in Figures 2 and 3, one can easily improve the square of the multiple correlation (R2) up to 0.98 or higher for most cases. According to Figure 1e, the break at θ = 2.1°C is not obvious. However, if one fits the data between θ = 0.6 and 12°C together, there are clear breaks in the plots of the observed O2 versus the calculated O2, and observed versus calculated DA as shown in Figures 4e and 4f. Those breaks correspond to θ around 2.1°C. Water with θ = 2.1°C represents a local end-member probably formed by mixing the North and South Pacific Bottom Waters, Antarctic Circumpolar Water, and Antarctic Intermediate Water.

image

Figure 2. Various xy plots of hydrographic data from the p15 transect in latitudinal interval between 9.5°S and 15.5°S (as shown in Figure 1e) with potential temperature (θ) range between 0.6° and 2.1°C.

Download figure to PowerPoint

image

Figure 3. Same as Figure 2, except with potential temperature (θ) range between 2.1° and 12°C.

Download figure to PowerPoint

image

Figure 4. Comparison of the model calculated versus observed O2 and DA for the p15 transect in latitudinal interval between 9.5°S and 15.5°S and at three different potential temperature ranges (θ = 0.6–2.1°C; θ = 2.1–12°C; and θ = 0.6–12°C).

Download figure to PowerPoint

[20] The WOCE station data were obtained from the website administered by the WOCE Hydrographic Program Office (http://whpo.ucsd.edu). For the present study, we selected only meridional sections a17 (1994) and a20 (1997) from the western Atlantic; i8 (1994–1995) and i9 (1994–1995) from the eastern Indian Ocean; and p14 (1993 and 1996), p15 (1994 and 1996) and p18 (1994) from the central and eastern Pacific oceans. Exact locations of those sections can be found in the same website. Those selected sections provide a complete set of data on θ, S, O2, NO3, PO4, DIC, and Alk. When all WOCE data become publicly available, especially DIC and alkalinity data, we will expand our scope of analysis. Also we present here only the results from water samples with θ below 12°C. The water samples with θ greater than 12°C often require more than three-end-member mixing and are thus beyond the present model analysis. Since θ and S are often no longer conservative in the surface oceans due to air-sea heat exchange and water evaporation/precipitation, the samples shallower than 100 m are also excluded. As will be discussed later, some bottom water samples, where the oxygen profile can be explained solely by water mass mixing (that is, the water mixing rate is much faster than the oxygen consumption rate), are also excluded from the regression analysis. Data from the northern part of the WOCE section p18 (<15°S) indicate intensive denitrification both in sediments and the water column [Gruber and Sarmiento, 1997; Deutsch et al., 2001], thus they are also excluded here but will be treated in a separate paper.

[21] Some useful cross-sectional contour diagrams for the WOCE data can be found in a website at the Alfred Wegener Institute, Germany (http://www.awi-bremerhaven.de/GEO/eWOCE/Gallery/#). The selected GEOSECS (Geochemical Ocean Section Study) data in this paper include only station numbers between 27 and 68 from the western Atlantic Ocean (in the year 1972), between 429 and 450 from the eastern Indian Ocean (1978), and between 212 and 287 from the middle Pacific Ocean (1973–1974). Those stations are close to the WOCE stations used in this paper. The GEOSECS data can be obtained from the NOAA/AOML data management website http://www.aoml.noaa.gov/ocd/oaces).

3. Results and Discussion

  1. Top of page
  2. Abstract
  3. 1. Introduction to a New Three-End-Member Mixing Model
  4. 2. Steps in the Model Calculation
  5. 3. Results and Discussion
  6. 4. Overall Discussion
  7. 5. Summary and Conclusions
  8. Acknowledgments
  9. References
  10. Supporting Information

[22] Results of multiple regression analysis from three major oceans between given latitudinal intervals and potential temperature ranges are summarized in Tables 1a, 1b, and 1c, and are plotted in Figures 5, 6, and 7 as a function of latitude. The rp, rn, and rc in those tables are all averaged values (i. e. [ri + ri′]/2). Other ratios are basically calculated from rp, rn, and rc. The rc and rp/rc values for some GEOSECS segments are not given, because DIC and alkalinity data for those segments are too noisy to provide meaningful results. Since the chosen data all can be fitted nicely to our model, the implicit assumption of our model that remineralization ratios are constant with depth within the chosen potential temperature interval is a good one.

image

Figure 5. Various remineralization ratios as a function of latitude in the Atlantic Ocean (data from Table 1a). The solid boxes are derived from WOCE data, and dashed boxes are from GEOSECS data. The traditional Redfield Ratio is drawn in each panel for reference.

Download figure to PowerPoint

image

Figure 6. Various remineralization ratios as a function of latitude in the Indian Ocean (data from Table 1b). The solid boxes are derived from WOCE data, and dashed boxes are from GEOSECS data. The traditional Redfield Ratio is drawn in each panel for reference.

Download figure to PowerPoint

image

Figure 7. Various remineralization ratios as a function of latitude in the Pacific Ocean (data from Table 1c). The solid boxes are derived from WOCE data, and dashed boxes are from GEOSECS data. The traditional Redfield Ratio is drawn in each panel for reference.

Download figure to PowerPoint

[23] Based on the general variation patterns of remineralization ratios shown in Figures 5, 6, and 7, the stations in the Atlantic Ocean are divided into a northern group (45°N–5°N) and a southern group (5°N–50°S). Similarly, the Indian Ocean stations are divided into an equatorial group (20°N–20°S) and a southern group (30°S–65°S), and the Pacific Ocean also into a northern group (55°N–10°S plus bottom water mass between 10°S and 40°S) and a southern group (10°S–70°S). The average remineralization ratios for those groups, as based only on the WOCE results in Tables 1a to 1c, are summarized in Table 2.

Table 2. Average Remineralization Ratios of Organic Matter in Three Major Oceans (Based Only on WOCE Data)a
 Latitude IntervalP/PN/P (rp/rn)Corg/P (rp/rc)−O2/P (rp)−O2/N (rn)−O2/Corg (rc)Corg/N (rn/rc)
  • a

    Values in parenthesis are obtained by assuming rn = rp/15, i.e. N/P = 15.

  • b

    Plus bottom water mass between 10°S and 40°S.

  • c

    Average of the south from three oceans.

Redfield 1161061388.61.306.6
 
Atlantic
North45°N–5°N116.1 ± 1.073 ± 8137 ± 78.5 ± 0.81.9 ± 0.24.6 ± 0.6
South5°N–50°S115.2 ± 0.779 ± 6128 ± 58.4 ± 0.31.6 ± 0.15.2 ± 0.4
 
Indian
Equator20°N–20°S110.3 ± 0.794 ± 5130 ± 712.7 ± 0.71.4 ± 0.19.2 ± 0.5
  1(15 ± 1)(92 ± 5)130 ± 7(8.7 ± 0.6)(1.4 ± 0.1)(6.1 ± 0.5)
South30°S–65°S114.8 ± 0.783 ± 2134 ± 99.1 ± 0.81.6 ± 0.15.6 ± 0.6
 
Pacific
North55°N–10°Sb113.3 ± 0.9124 ± 11162 ± 1112.2 ± 0.51.3 ± 0.19.2 ± 0.8
  1(15 ± 1)(123 ± 11)162 ± 11(10.8 ± 0.7)(1.3 ± 0.1)(8.2 ± 0.8)
South10°S–70°S114.5 ± 1.078 ± 7136 ± 109.4 ± 0.51.7 ± 0.15.4 ± 0.4
 
Other
Southern Oceansc 115 ± 180 ± 3133 ± 59.0 ± 0.41.6 ± 0.15.4 ± 0.3

3.1. Atlantic Ocean

[24] Bottom water samples with θ below 2 to 3°C in the northern and below 0.2 to 0.4°C in the southern stations are not included in the regression analysis (Table 1a). As mentioned earlier, oxygen concentration profiles of those bottom samples can be explained solely by water mass mixing (no apparent in situ oxygen consumption).

[25] According to Table 1a, differences in the remineralization ratios between low and high temperature intervals are small, thus, to a first approximation, the remineralization ratios can be considered constant with depth in the Atlantic Ocean with θ below 12°C. As shown in Figures 5a, 5b, and 5d, the −O2/P (= rp), −O2/N (= rn) and N/P (= rp/rn) ratios are near constant from north to south for both WOCE and GEOSECS station data. However, the −O2/P and N/P ratios have a slight tendency to decrease from north to south for WOCE data. There are some minor disagreements in the −O2/P and N/P values between WOCE and GEOSECS data in the latitudinal interval between 0°S and 20°S. The −O2/P and N/P ratios from GEOSECS data tend to be slightly lower (Table 1a). The −O2/Corg (= rc) ratio is near constant in the latitudinal interval between 50°S and 10°N and increases suddenly northward (Figure 5c). Since the −O2/N ratio is constant (Figure 5b), the sudden increase in the −O2/Corg ratio is mostly caused by decrease in organic carbon content of remineralized organic matter in the northern Atlantic Ocean. The −O2/Corg ratio (Figure 5c) is much higher and the Corg/P ratio (Figure 5e) much lower than the traditional Redfield ratios. This implies a low carbon content of remineralized organic matter in the Atlantic Ocean as compared to Redfield's average marine plankton. Interestingly, the Corg/P ratio is near constant from north to south (Figure 5e).

[26] The average remineralization ratios, as normalized to P, for the northern Atlantic group are P/N/Corg/−O2 = 1/(16.1 ± 1.0)/(73 ± 8)/(137 ± 7) (Table 2). Those values are in good agreement with the traditional Redfield ratios of P/N/Corg/−O2 = 1/16/106/138, except for the low Corg/P ratio. However, the uncertainties of the traditional Redfield ratios are never given. If the traditional Redfield ratios represent the average composition of living marine plankton, then those ratios can easily have a standard deviation of more than 50% [Li et al., 2000]. The model-derived remineralization ratios given here represent the average composition of remineralized nonliving organic matter (both particulate and dissolved) in the water column. The average remineralization ratios of P/N/Corg/−O2 = 1/(15.2 ± 0.7)/(79 ± 6)/(128 ± 5) for the southern Atlantic group (Table 2) are the same as those for the northern Atlantic group within the estimated uncertainties. Average remineralization ratios for the Atlantic Ocean as obtained from WOCE and GEOSECS data sets are also the same statistically (Table 1a).

[27] Using GEOSECS data, Takahashi et al. [1985] obtained remineralization ratios of P/N/Corg/−O2 = 1/(16.8 ± 1.3)/(92 ± 12)/(178 ± 11) for the Atlantic thermocline water (σθ = 27.20) between a wide latitudinal interval (38°N to 45°S). Those values agree reasonably well with the present study, except that their −O2/P value is much higher than our model-derived 133 ± 5. Broecker et al. [1985] and Anderson and Sarmiento [1994] also gave a high −O2/P ratio of 170 ± 10 for the whole ocean. The disagreement on the estimate of the −O2/P ratio primarily results from the way samples were chosen for the earlier models. For example, the θ-S diagrams for the same samples used by Takahashi et al. [1985] in their two-end-member mixing model show a distinct break in slopes (Figures 8a and 8b), indicating mixing of at least three end-members. Therefore, they should not have used all those samples, which cover a wide latitudinal range, in their two-end-member mixing model. When their northern Atlantic data alone were fitted to the two-end-member mixing model of Li et al. [2000], an −O2/P ratio of 140 ± 10 was obtained, in good agreement with the present result. Anderson and Sarmiento [1994] put together neutral surface data from a wide latitudinal interval (20°N to 50°S) for the South Atlantic Ocean. Peng and Broecker [1987] also treated together isopycnal horizon data (with θ = 1.4 ± 0.1°C; depth about 3000 m) from a wide latitudinal interval (45°N to 43°S) in the Pacific Ocean. The θ-S plot of samples used by Peng and Broecker [1987] appears to suggest a two-end-member mixing at first glance (Figure 8c), but diapycnal mixing inputs of other water masses is unavoidable over such a wide latitudinal interval. Indeed, an irregular O2 distribution pattern in the θ-O2 plot of their original data (Figure 8d) suggests mixing of at least three to four end-members. As shown in Figures 1a to 1g, the water mass with θ = 1.4 ± 0.1°C is formed not only by mixing North Pacific Bottom Water (I) and South Pacific Bottom Water (I*), but also Antarctic Circumpolar Water (I*), Antarctic Intermediate Water (III*), and North Pacific Intermediate Water (III). One may avoid the problem by choosing data from a narrower latitudinal interval, but very often the number of available data points becomes rather small for any horizontal two-end-member mixing model.

image

Figure 8. Plots of salinity (S) versus potential temperature (θ) data from (a) the Atlantic thermocline water along the σθ = 27.2 horizon [Takahashi et al., 1985], (b) the Indian thermocline water along σθ = 27.2 horizon [Takahashi et al., 1985], and (c) the deep Pacific Ocean along the σ4 = 45.83 per mil isopycnal horizon [Peng and Broecker, 1987]. (d) Plot of dissolved oxygen (O2) versus θ along the σ4 = 45.83 per mil isopycnal horizon in the deep Pacific Ocean [Peng and Broecker, 1987].

Download figure to PowerPoint

[28] An additional problem with earlier models was the way the end-member values were assigned. For example, Takahashi et al. [1985] plotted θ versus O2 data from an isopycnal horizon and extrapolated linearly the correlation line to intercept with the oxygen solubility line in equilibrium with the atmosphere (as a function of θ). The values of θ and O2 at the intercept were adapted as the end-member values. However, O2 is a nonconservative tracer; thus there is no a priori reason to believe a plot of θ versus O2 should be always linear. If it is really linear, then there is no oxidation of organic matter. Broecker et al. [1985] plotted PO4 versus O2 from an isopycnal horizon to obtain the slope of −O2/P (= rp). Their implicit assumption was that the concentrations of PO4 and O2 for the two mixing end-members are identical or similar, which is not always a good assumption. Broecker et al. [1985] also plotted (PO4 − PO40) versus (O20 − O2) (= AOU) to obtain the −O2/P ratio. However, the procedure for assigning values of preformed phosphate (PO40) is problematic, as was also the case for Takahashi et al. [1985].

[29] The average composition of sediment trap materials obtained at the Ocean Flux Program (OFP)/ Bermuda Atlantic Time Series (BATS) site is P/N/Corg = 1/(20 ± 7)/(180 ± 70) over a depth range between 500 and 3200 m [Conte et al., 2001]. Therefore, the composition of remineralized organic matter in the water column is quite different from that of sediment trap materials. Sediment trap materials must be mostly leftover materials after remineralization and may have preferentially lost P over N (N/P ratio of 20 versus 15), and N over Corg (Corg/N ratio of 9 ± 4.7 versus 4.6 ± 0.6) as compared to the remineralized organic matter. The Corg/N ratio for the total organic matter (dissolved and particulate) at BATS site at a depth interval between 1000 and 4000 m is 14.1 ± 1.8 [Hansell and Carlson, 2001]. This ratio is again much larger than those of marine plankton (Corg/N = 6.6) and remineralized organic matter (Corg/N = 4.6 ± 0.6) in the water column, again suggesting preferential loss of N.

3.2. Indian Ocean

[30] The O2 data for the bottom water samples with θ less than 2°C in the southern Indian Ocean show no apparent in situ oxygen consumption and can be explained by water mass mixing, thus are again excluded (Table 2b). Remineralization ratios given in Figures 6a to 6e for the southernmost stations of the Indian Ocean are all similar to those for the southernmost stations of the Atlantic Ocean. The −O2/P ratio is nearly constant (close to the traditional Redfield ratio; Figure 6a), but the −O2/N ratio increases steadily from the southern (close to the traditional Redfield ratio) to equatorial Indian oceans (Figure 6b). This may indicate a steady decrease of net nitrate input relative to both P input and oxygen consumption during oxidation of organic matter in the water column. The steady decrease of the N/P ratio from the south (close to the traditional Redfield ratio) to the equator (Figure 6d) is consistent with this explanation. The −O2/Corg (Figure 6c) and P/Corg (inverse of Figure 6e) ratios decrease slightly from the south to the equator, while the −O2/P ratio is nearly constant (Figure 6a). These observations suggest that Corg input increases slightly relative to both P input and oxygen consumption from the south to the equator during oxidation of organic matter in the Indian Ocean.

[31] The average remineralization ratios of P/N/Corg/−O2 = 1/(14.8 ± 0.7)/(83 ± 2)/(134 ± 9) for the southern Indian group (30°S–65°S) are statistically identical to those for the southern Atlantic group (Table 2). In contrast, average remineralization ratios for the equatorial Indian group are P/N/Corg/−O2 = 1/(10.3 ± 0.7)/(94 ± 5)/(130 ± 7). The low N/P ratio here may indicate either a truly low N/P ratio in the remineralized organic matter, or more likely the loss of nitrate during the oxidation process of organic matter throughout the water column. If the actual mineralization ratio of N/P (= rp/rn) is 15 ± 1 (or rn = rp/15 = 8.7 ± 0.6) for remineralized organic matter in the equatorial Indian Ocean, then the model derived N/P ratio of 10 ± 1 may suggest the following. One third (= (15–10)/15) of organic nitrogen converts into gaseous N2 and N2O instead of usual NO3 during oxidation process of organic matter. Therefore, the true average remineralization ratios of organic matter in the equatorial Indian Ocean could be P/N/Corg/−O2 = 1/(15 ± 1)/(92 ± 5)/(130 ± 7) (Table 2, parenthesis), which are similar to those for the southern Atlantic and Indian oceans, except slightly higher Corg/P ratio (Table 2). The fraction of organic nitrogen converted into N2 and N2O must have been fairly constant throughout the water column. If the conversion were uneven and localized, one would not expect to obtain the observed good model fit. In order to calculate the preformed nitrate (= NO3 − AOU/rn), one still needs to use the model-derived rn of 12.7 ± 0.7 to account for a partial conversion of organic nitrogen into N2 and N2O. Notice here that the change of the rn value from 12.7 to 8.7 would increase only slightly the rc value from 1.39 to 1.42, and the rp/rc ratio from 94 to 92.

[32] One may postulate that the partial conversion of organic nitrogen, as (NH3)org, into N2 and N2O is facilitated by bacteria within a low-oxygen microenvironment of organic matter during the nitrification process of organic nitrogen. For example, overall net reactions could be:

  • equation image
  • equation image
  • equation image

The traditional thinking is that all organic nitrogen is first converted into NO2 + NO3 during water column nitrification process [von Brand et al., 1937]. Then, some fraction of NO2 + NO3 is reduced back to gaseous N2O and N2 during denitrification process within a reducing microenvironment of organic matter by bacteria. For example:

  • equation image
  • equation image

The low oxygen or reducing microenvironment is a necessary condition here, because all the water samples selected for our model calculations have moderate to high oxygen concentrations. Production of N2O and N2 gaseous species during the denitrification process is well documented [Karl and Michaels, 2001, and references therein; Naqvi et al., 2000]. Production of N2O during the nitrification process has also been demonstrated [Yoshinari, 1976; Elkins et al., 1978; Goreau, 1980]. Unfortunately none of them specifically looked for N2. It will be interesting to design experiments to test the formation of N2 during the nitrification process of organic nitrogen. It has been shown recently [Strous et al., 1999] that planctomycete bacteria can produce N2 by the anammox (anaerobic ammonia oxidation) reaction:

  • equation image

The lesson is that we are still quite ignorant about many novel metabolic pathways of nitrogen by bacteria in natural environments [Amend and Shock, 2001].

[33] Peng and Broecker [1987] gave average remineralization ratios of P/N/Corg/−O2 = 1/(13.4 ± 1.0)/(135 ± 18)/(176 ± 11) for the deep Indian Ocean. Takahashi et al. [1985] gave P/N/Corg/−O2 = 1/(14.5 ± 0.5)/(125 ± 7)/(174 ± 6) for the Indian thermocline water. These values are quite different from the present study. Their Corg/P and −O2/P ratios are especially high. Possible causes for their high ratios were already discussed in the last section.

3.3. Pacific Ocean

[34] Bottom water samples with θ less than 1.5 to 2°C in the southern Pacific Ocean (latitude >40°S; Table 1c) are excluded again, because the oxygen profiles there can be explained solely by water mass mixing. The northern Pacific group (55°N to 10°S plus the bottom water mass between 10°S and 40°S) is characterized by constantly high −O2/P and −O2/N, and low N/P ratios (Figures 7a, 7b, and 7d), as compared to the traditional Redfield ratios and to those for the southern Pacific group. The implication is that oxygen consumption is high in the northern Pacific group and there is a moderate but significant conversion of organic nitrogen into N2 and N2O through nitrification/denitrification reactions, as seen for the equatorial Indian group. The constantly low −O2/Corg (Figure 7c) and P/Corg ratios (inverse of Figure 7e), as compared with the southern Pacific group, also suggest a high carbon content of remineralized organic matter in the water column of the northern Pacific group.

[35] As shown in Table 2, the average remineralization ratios of P/N/Corg/−O2 = 1/(14.5 ± 1.0)/(78 ± 7)/(136 ± 10) for the southern Pacific group are essentially the same as those for the southern Atlantic and the southern Indian oceans within the estimated uncertainties. We may classify them together as the Southern Oceans group, having average remineralization ratios of P/N/Corg/−O2 = 1/(15 ± 1)/(80 ± 3)/(133 ± 5) (Table 2). The average remineralization ratios of P/N/Corg/−O2 = 1/(13.3 ± 0.9)/(124 ± 11)/(162 ± 11) for the northern Pacific group are in good agreement with P/N/Corg/−O2 = 1/(13 ± 1)/(135 ± 18)/(170 ± 9) for the ALOHA station in the subtropical North Pacific [Li et al., 2000]. Peng and Broecker [1987] gave P/N/Corg/−O2 = 1/(14.8 ± 1)/(130 ± 16)/(199 ± 10) for the deep Pacific Ocean, which is also similar to the present estimate for the northern Pacific group, except for their high −O2/P ratio. The low N/P ratio of 13 ± 1 in the northern Pacific group may represent either a true low N/P ratio in remineralized organic matter or most likely the conversion of some organic nitrogen into N2O and N2 during the nitrification/denitrification processes, as is the case for the equatorial Indian group. The actual N/P ratio for the remineralized organic matter could also be around 15 ± 1 (or rn = rp/15 = 10.8 ± 0.7). As discussed in the Indian Ocean section, the fraction of organic nitrogen converted into N2 and N2O must have been fairly constant throughout the water column. If not, one would not expect to have a tight correlation between PO4 and NO3 concentration data, as shown in Figure 2d, and obtain the observed good model fit. To calculate the preformed nitrate one still should use the model-derived rn value of 12.2 ± 0.5 to account for the partial conversion of organic nitrogen into N2O and N2. The unusually high Corg/P ratio of 124 ± 11 (Table 2) for the northern Pacific group indicates high carbon content of organic matter as compared with other groups.

[36] The shallow sediment trap materials obtained at ALOHA station have an average composition of P/N/Corg = 1/24/187 at a depth of 100 m, and P/N/Corg = 1/29/303 at a depth of 500 m [Karl et al., 1996, 2001a]. Those ratios suggest the preferential loss of P over N over Corg during shallow remineralization as compared to both plankton and remineralized organic matter, as already discussed in the Atlantic Ocean. The suspended particles collected at depths between 500 and 1000 m have an average composition of P/N/Corg = 1/20/155 [Hebel and Karl, 2001]. Furthermore, Benner [2002] gave an average composition of P/N/Corg = 1/22/300 for dissolved organic matter (DOM) from surface oceans and of P/N/Corg = 1/25/444 for DOM from deep oceans worldwide. Therefore, the composition of remineralized organic matter in the water column is again quite different from those of sediment trap materials, suspended particles, and DOM. Additional study is needed on the genetic relationship and mass balances among remineralized organic matter and other organic species such as plankton, sediment trap materials, suspended particles and DOM in a given water column.

4. Overall Discussion

  1. Top of page
  2. Abstract
  3. 1. Introduction to a New Three-End-Member Mixing Model
  4. 2. Steps in the Model Calculation
  5. 3. Results and Discussion
  6. 4. Overall Discussion
  7. 5. Summary and Conclusions
  8. Acknowledgments
  9. References
  10. Supporting Information

[37] One may formulate the chemical composition of one mole of a model organic matter in a water column from the average remineralization ratios (P/N/Corg/−O2). The assumptions are as follows: phosphorus exists as H3PO4, nitrogen as NH3, and organic carbon as a mixture of hydrocarbon and variable water (CmHn · zH2O) in the model organic matter. Oxidation of organic matter from the northern Atlantic Ocean with P/N/Corg/−O2 = 1/16/73/137 gives the following:

  • equation image
  • equation image
  • equation image
  • equation image
  • equation image

By adding equations 6a through 6d, one obtains the oxidation reaction of the model organic matter, i.e.,

  • equation image

Similar calculations for the Southern Oceans, equatorial Indian, and northern Pacific groups are summarized in Table 3. For the equatorial Indian and northern Pacific oceans, one should also include equation 5c to take care of the partial conversion of (NH3)org into N2 (including N2O). For comparison, one may rewrite the average composition of three mixed plankton samples from the southern Pacific Ocean (C81H134O29N15P; assuming N/P ratio of 15 and normalized to P) by Hedges et al. [2002] to C81H36(H2O)25(NH3)15(H3PO4). Its oxidation gives

  • equation image

Therefore, one obtains the remineralization ratios of P/N/Corg/−O2 = 1/15/81/120 for the mixed plankton, which is similar to that for the remineralized organic matter in the Southern Oceans (Table 2) except that −O2/P ratio is lower. Anderson [1995] proposed an average formula of C106H48(H2O)38(NH3)16(H3PO4) for marine phytoplankton with P/N/Corg/−O2 = 1/16/106/150, which is similar to the traditional Redfield ratios except the −O2/P ratio is higher.

Table 3. Model Compositions of 1 Mole of Remineralized Organic Matter in Different Ocean Basinsa
Ocean BasinP/N/Corg/−O2Model Organic Matter−O2/CorgH/CH+N2
  • a

    Calculated from remineralization ratios (P/N/Corg/−O2). H/C is the ratio of hydrogen and carbon in hydrocarbon fraction of the model organic matter. H+ and N2 are the moles of produced hydrogen ion and nitrogen during oxidation of 1 mole of the model organic matter.

  • b

    The z in the model organic matter is unknown number of water molecules.

  • c

    Excluded one high value at 13°N–7°N in Table 1a.

Northern Atlantic1/16/73/137C73H128(zH2O)(NH3)16(H3PO4)b1.8c1.75170
Southern Oceans1/15/80/133C80H92(zH2O)(NH3)15(H3PO4)1.61.15160
Equatorial Indian1/15/92/130C92H57(zH2O)(NH3)15(H3PO4)1.40.62112.5
Northern Pacific1/15/123/162C123H46(zH2O)(NH3)15(H3PO4)1.30.37141
Redfield1/16/106/138C106(106H2O)(NH3)16(H3PO4)1.30170

[38] The −O2/Corg and H/C ratios of hydrocarbon fraction in the model organic matter (Table 3) decrease systematically from the northern Atlantic Ocean to the Southern Oceans, then to the equatorial Indian Ocean and the northern Pacific Ocean. The implication is that the relative proportions of biomolecules, such as lipids, proteins, nucleic acids, and carbohydrate (polysaccharides), in remineralized organic matter change systematically from ocean to ocean. According to the calculation by Laws [1991], the average molar ratio of production (or consumption) rate of O2 and photosynthesis (or respiration) rate of Corg (−O2/Corg) is about 1.6 ± 0.3 for proteins and nucleic acids, 1.4 ± 0.1 for lipids, and 1.0 ± 0.1 for polysaccharides (carbohydrate). His assumption is that organic carbon and organic nitrogen in those biomolecules are all converted from (or oxidized to) CO2 and nitrate. Hedges et al. [2002] gave an average formula of C106H14(H2O)34(NH3)28(H2S) for protein, C18H30(H2O)2 for lipid, and C6(H2O)5 for carbohydrate in mixed plankton samples. Oxidation of those biomolecules gives the −O2/Corg ratio of 1.63, 1.42, and 1, respectively. These values are in good agreement with Laws' general estimations. Therefore, the relatively high −O2/Corg ratio of 1.8 to 1.6 in the northern Atlantic Ocean and the Southern Oceans (Table 2) may indicate relatively high protein and nucleic acid contents for remineralized organic matter in the water column there. The relatively low −O2/Corg ratio of 1.3 to 1.4 in the equatorial Indian and the northern Pacific oceans may suggest relatively high carbohydrate and lipid contents of remineralized organic matter in those regions. Indeed, Hedges et al. [2002] showed that the model-derived protein content is higher, and carbohydrate and lipid are lower in the mixed plankton samples from the Southern Oceans than from the equatorial Pacific (belongs to the northern Pacific group in our classification) and the Arabian Sea.

[39] Systematic change in the biochemical compositions of both remineralized organic matter and mixed living plankton may reflect the difference in plankton populations in different oceans. For example, prokaryotic picoplankton (Prochlorococcus and Synechococcus), eukaryotic nanoplankton, coccolithophores and pennate diatoms are dominant species in the Sargasso Sea of the North Atlantic Ocean [DuRand et al., 2001]. The centric and pennate diatoms are the major phytoplankton species in the Southern Oceans [Brown and Landry, 2001]. In contrast, prokaryotic picoplankton (Prochlorococcus and Synechococcus) and eukaryotic picoplankton dominate in the equatorial Pacific [Chavez et al., 1996; S. Brown et al., Microbial absorption and backscattering coefficients from in situ and satellite data during an ENSO cold phase in the equatorial Pacific (108°), submitted to Journal of Geophysical Research, 2002] and in the Arabian Sea (except during southwest monsoon, diatoms become dominant) [Garrison et al., 2000]. Prochlorococcus is the major phytoplankton taxon in the North Pacific Subtropical Gyre [Karl et al., 2001b]. Certainly further study is needed.

[40] We have shown that there is a systematic change in the average remineralization ratios (P/N/Corg/−O2) of remineralized organic matter from the northern Atlantic Ocean [1/(16.1 ± 1)/(73 ± 8)/(137 ± 7)] to the Southern Oceans [1/(15 ± 1)/(80 ± 3)/(133 ± 5)], then to the equatorial Indian Ocean [1/(15 ± 1)/(92 ± 5)/(130 ± 7)] and the northern Pacific Ocean [1/(15 ± 1)/(123 ± 11)/(162 ± 11)], more or less along the global deep ocean circulation route. As discussed earlier, the N/P ratio of 15 ± 1 is adopted here for the equatorial Indian Ocean and the northern Pacific Ocean to correct for the effect of nitrification/denitrification on the model-derived rn. In contrast, Anderson and Sarmiento [1994] gave P/N/Corg/−O2 = 1/(16 ± 1)/(117 ± 14)/(170 ± 10) for three major oceans at depths below 400 m. Shaffer et al. [1999] also concurred with those estimates for the major three oceans (at depth below 1500 m). However, each model has its own assumptions, and thus its own weakness and limitation. For example, the assumption by Anderson and Sarmiento [1994] that ΔCinorg/ΔP is constant along a neutral surface is not justified by any observation. Furthermore, their neutral surfaces cover wide latitudinal intervals, thus their two-end-member mixing model ignored the effect of possible diapycnal mixing inputs from other water masses, as mentioned earlier. Shaffer et al. [1999] allowed some variability for initial values of end-members. However, the choice of initial values is still subjective. In our model, the preformed DA (= DIC0–Alk0/2) is not strictly conservative in the shallow waters due to input of anthropogenic CO2. No matter which model we use, it seems to be reasonable to apply the remineralization ratio of N/P = 16 ± 1 to the whole ocean for estimating the extent of nitrogen fixation and conversion of Norg into N2 and N2O [Gruber and Sarmiento, 1997; Deutsch et al., 2001].

[41] With a few exceptions, the remineralization ratios obtained using WOCE and GEOSECS data sets are the same within estimated uncertainties (Figures 5 to 7). The claim that the remineralization ratios in the deep oceans change with time [Pahlow and Riebesell, 2000] is premature as has been discussed by Zhang et al. [2000].

5. Summary and Conclusions

  1. Top of page
  2. Abstract
  3. 1. Introduction to a New Three-End-Member Mixing Model
  4. 2. Steps in the Model Calculation
  5. 3. Results and Discussion
  6. 4. Overall Discussion
  7. 5. Summary and Conclusions
  8. Acknowledgments
  9. References
  10. Supporting Information

[NaN] 

  1. A new three-end-member mixing model has been introduced and successfully applied to GEOSECS and WOCE data sets to estimate remineralization ratios of organic matter (dissolved and particulate) in the water column.
  2. Remineralization ratios (P/N/Corg/−O2) of organic matter change systematically from the northern Atlantic to the Southern Oceans, then to the equatorial Indian and northern Pacific oceans, almost along the global deep ocean circulation route.
  3. The average remineralization ratios of P/N/Corg/−O2 for the northern Atlantic and the Southern Oceans groups are the same within the estimated uncertainties. The average is P/N/Corg/−O2 = 1/(15 ± 1)/(78 ± 8)/(134 ± 10), which is similar to the traditional Redfield ratios of P/N/Corg/−O2 = 1/16/106/138, except for the low Corg/P ratio. This may imply that the average compositions of remineralized organic matter and mixed marine plankton are not much different.
  4. If the N/P ratio of 15 ± 1 (or rn = rp/15 = 8.7 ± 0.7) is adopted for the equatorial Indian Ocean, the remineralization ratios there become P/N/Corg/−O2 = 1/(15 ± 1)/(92 ± 5)/(130 ± 7), which are similar to those for the northern Atlantic and the Southern Oceans groups, except for a slightly higher Corg/P ratio. In order to calculate the preformed nitrate, one still needs to use the model-derived rn value of 12.7 ± 0.7.
  5. The equatorial Indian and the northern Pacific groups are characterized by the high model-derived rn (= −O2/N) and low N/P (= rp/rn) values (Table 2). The high rn and low N/P ratio are most likely caused by partial conversion of organic nitrogen into gaseous N2O and N2 throughout the water column during the nitrification/denitrification process within reducing microenvironments of organic matter by bacteria.
  6. The northern Pacific group (55°N–10°S plus bottom water mass between 10°S and 40°S) is further characterized by the high Corg/P ratio (124 ± 11) as compared to other groups (83 ± 10; Table 2), indicating a high carbon content of remineralized organic matter.
  7. The −O2/Corg ratio of remineralized organic matter decreases from the northern Atlantic (1.8) to the Southern Oceans (1.6), then to the equatorial Indian (1.4) and the northern Pacific (1.3) oceans. This systematic change may indicate a decreasing trend of protein plus nucleic acid and an increasing trend of carbohydrate (polysaccharide) and lipid contents in remineralized organic matter as well as mixed plankton.
  8. No obvious temporal trends are detected from the remineralization ratios obtained from the GEOSCS and WOCE data sets.

Acknowledgments

  1. Top of page
  2. Abstract
  3. 1. Introduction to a New Three-End-Member Mixing Model
  4. 2. Steps in the Model Calculation
  5. 3. Results and Discussion
  6. 4. Overall Discussion
  7. 5. Summary and Conclusions
  8. Acknowledgments
  9. References
  10. Supporting Information

[42] Discussions with Bob Bidigare, Susan Brown, Paul Falkowski, David Karl, and Edward Laws are most fruitful. Detailed comments by James Murray as a reviewer have greatly improved this paper. Editorial assistant from Diane Henderson is indispensable. This work is supported by the NOAA/OAR/OGP grant GC99-032. SOEST contribution 6038.

References

  1. Top of page
  2. Abstract
  3. 1. Introduction to a New Three-End-Member Mixing Model
  4. 2. Steps in the Model Calculation
  5. 3. Results and Discussion
  6. 4. Overall Discussion
  7. 5. Summary and Conclusions
  8. Acknowledgments
  9. References
  10. Supporting Information
  • Amend, J. P., and E. L. Shock, Energetics of overall metabolic reactions of thermaphilic and hyperthermophilic archaea and bacteria, FEMS Microbiol. Rev., 25, 175243, 2001.
  • Anderson, L. A., On the hydrogen and oxygen content of marine phytoplankton, Deep Sea Res., Part I, 42, 16751680, 1995.
  • Anderson, L. A., and J. L. Sarmiento, Redfield ratios of remineralization determined by nutrient data analysis, Global Biogeochem. Cycles, 8, 6580, 1994.
  • Benner, R., Chemical composition and reactivity, in Biogeochemistry of Marine Dissolved Organic Matter, edited by D. A. Hansell, and C. A. Carlson, pp. 5990, Academic, San Diego, Calif., 2002.
  • Boulahdid, M., and J. F. Minster, Oxygen consumption and nutrient regeneration ratios along isopycnal horizons in the Pacific Ocean, Mar. Chem., 26, 133153, 1989.
  • Broecker, W. S., NO: A conservative water-mass tracer, Earth Planet. Sci. Lett., 23, 100107, 1974.
  • Broecker, W. S., and T. H. Peng, Tracers in the Sea, Lamont-Doherty Earth Obs., Palisades, N. Y., 1982.
  • Broecker, W. S., T. Takahashi, and T. Takahashi, Sources and flow patterns of deep-ocean water as deduced from potential temperature, salinity, and initial phosphate concentration, J. Geophys. Res., 90, 69256939, 1985.
  • Brown, S. L., and M. R. Landry, Microbial community structure and biomass in surface waters during a Polar Front summer bloom along 170°W, Deep Sea Res., Part II, 48, 40394058, 2001.
  • Chavez, F. P., K. R. Buck, S. K. Service, J. Newton, and R. T. Barber, Phytoplankton variability in the central and eastern tropical Pacific, Deep Sea Res., Part II, 43, 835870, 1996.
  • Chen, C. T. A., and R. M. Pytkowicz, On the total CO2-titration alkalinity-oxygen system in the Pacific Ocean, Nature, 281, 362365, 1979.
  • Conte, M. H., N. Ralph, and E. H. Ross, Seasonal and interannual variability in deep ocean particle fluxes at the Ocean Flux Program (OFP)/Bermuda Atlantic Time Series (BATS) site in the western Sargasso Sea near Bermuda, Deep Sea Res., Part II, 48, 14711505, 2001.
  • Deutsch, C., N. Gruber, R. M. Key, J. L. Sarmiento, and A. Ganachaud, Denitrification and N2 fixation in the Pacific Ocean, Global Biogeochem. Cycles, 15, 483506, 2001.
  • DuRand, M. D., R. J. Olson, and S. W. Chisholm, Phytoplankton population dynamics at the Bermuda Atlantic Time-series station in the Sargasso Sea, Deep Sea Res., Part II, 48, 19832003, 2001.
  • Elkins, J. W., S. C. Wofsy, M. B. McElroy, C. E. Kolb, and W. A. Kaplan, An aquatic sources and sinks for nitrous oxide, Nature, 275, 602606, 1978.
  • Garrison, D. L., et al., Microbial food web structure in the Arabian Sea: A US JGOFS study, Deep Sea Res., Part II, 47, 13871422, 2000.
  • Goreau, T. J., Production of NO2 and N2O by nitrifying bacteria at reduced concentrations of oxygen, Appl. Environ. Microbiol., 40, 526532, 1980.
  • Gruber, N., and J. L. Sarmiento, Global patterns of marine nitrogen fixation and denitrification, Global Biogeochem. Cycles, 11, 235266, 1997.
  • Hansell, D., and C. A. Carson, Biogeochemistry of total organic carbon and nitrogen in the Sargasso Sea: control by convective overturn, Deep Sea Res., Part II, 48, 16491667, 2001.
  • Hebel, D., and D. M. Karl, Seasonal, interannual and decadal variations in particulate matter concentrations and composition in the subtropical North Pacific Ocean, Deep Sea Res., Part II, 48, 16691695, 2001.
  • Hedges, K. I., J. A. Baldock, Y. Gelina, C. Lee, M. L. Peterson, and S. G. Wakeham, The biochemical and elemental compositions of marine plankton: A NMR perspective, Mar. Chem., 78, 4763, 2002.
  • Hupe, A., and J. Karstensen, Redfield stoichiometry in Arabian Sea subsurface waters, Global Biogeochem. Cycles, 14, 357372, 2000.
  • Karl, D. M., and A. F. Michaels, Nutrient cycling: nitrogen cycle, in Encyclopedia of Ocean Sciences, edited by J. Steele, S. Thorpe, and K. Turekian, Academic, San Diego, Calif., 2001.
  • Karl, D. M., J. R. Christain, J. E. Dore, D. V. Hebel, R. M. Letelier, L. M. Tupas, and C. D. Winn, Seasonal and interannual variability in primary production and particle flux at Station ALOHA, Deep Sea Res., Part II, 43, 539568, 1996.
  • Karl, D. M., K. M. Bjoerkman, J. E. Dore, L. Fujieki, D. V. Hebel, T. Houlihan, R. M. Letelier, and L. M. Tupas, Ecological nitrogen-to-phosphorus stoichiometry at station ALOHA, Deep Sea Res., Part II, 48, 15291566, 2001a.
  • Karl, D. M., R. R. Bidigare, and R. M. Letelier, Long-term changes in plankton community structure and productivity in the North Pacific Subtropical Gyre: The domain shift hypothesis, Deep Sea Res., Part II, 48, 14491470, 2001b.
  • Laws, E. A., Photosynthetic quotients, new production and net community production in the open ocean, Deep Sea Res., 38, 143167, 1991.
  • Li, Y. H., T. H. Peng, W. S. Broecker, and H. G. Oestlung, The average vertical mixing coefficient for the oceanic thermocline, Tellus, Ser. B, 36, 212217, 1984.
  • Li, Y. H., D. M. Karl, C. D. Winn, F. T. Mackenzie, and K. Gans, Remineralization ratios in the subtropical north Pacific gyre, Aquat. Geochem., 6, 6586, 2000.
  • Minster, J., and M. Boulahdid, Redfield ratios along isopycnal surfaces: A complementary study, Deep Sea Res., 34, 19812003, 1987.
  • Morse, J. W., and F. T. Mackenzie, Geochemistry of Sedimentary Carbonates, 707 pp., Elsevier Sci., New York, 1990.
  • Naqvi, S. W. A., D. A. Jayakumar, P. V. Narvekar, H. Naik, V. V. S. S. Sarma, W. D'Souza, S. Joseph, and M. D. George, Increased marine production of N2O due to intensifying anoxia on the Indian continental shelf, Nature, 408, 346349, 2000.
  • Pahlow, M., and U. Riebesell, Temporal trends in deep ocean Redfield ratios, Science, 287, 831833, 2000.
  • Peng, T. H., and W. S. Broecker, C/P ratios in marine detritus, Global Biogeochem. Cycles, 1, 155161, 1987.
  • Redfield, A. C., The biological control of chemical factors in the environment, Am. Sci., 46, 205221, 1958.
  • Sarmiento, J. L., J. C. Orr, and U. Siegenthaler, A perturbation simulation of CO2 uptake in an ocean general circulation model, J. Geophys. Res., 97, 36213645, 1992.
  • Shaffer, G., Biogeochemical cycling in the global ocean, 2, New production, Redfield ratios, and remineralization in the organic pump, J. Geophys. Res., 101, 37233745, 1996.
  • Shaffer, G., J. Bendtsen, and O. Ulloa, Fractionation during remineralization of organic matter in the ocean, Deep Sea Res., Part I, 46, 185204, 1999.
  • Strous, M., J. A. Fuerst, E. H. M. Kramer, S. Logemannn, G. Muyzer, K. T. Van De-Pas-Schoon, R. Webb, and J. G. Kuenen, Missing lithotroph identified as new planctomycete, Nature, 400, 446449, 1999.
  • Takahashi, T., W. S. Broecker, and S. Langer, Redfield ratio based on chemical data from isopycnal surfaces, J. Geophys. Res., 90, 69076924, 1985.
  • Troy, P. J., Y. H. Li, and F. T. Mackenzie, Changes in surface morphology of calcite exposed to the oceanic water column, Aquat. Geochem., 3, 120, 1997.
  • von Brand, T., N. W. Rakestraw, and C. E. Renn, The experimental decomposition and regeneration of nitrogenous organic matter in sea water, Biol. Bull., 72, 165175, 1937.
  • Wanninkhof, R., S. Doney, T. H. Peng, J. L. Bullister, K. Lee, and R. A. Feely, Comparison of methods to determine the anthropogenic CO2 invasion into the Atlantic Ocean, Tellus, Ser. B, 51, 511530, 1999.
  • Winn, C. D., Y. H. Li, F. T. Mackenzie, and D. M. Karl, Rising surface ocean dissolved inorganic carbon at the Hawaii Ocean Time-series site, Mar. Chem., 60, 3347, 1998.
  • Yoshinari, T., Nitrous oxide in the sea, Mar. Chem., 4, 189202, 1976.
  • Zhang, J. Z., C. W. Mordy, L. I. Gordon, A. Ross, and H. E. Garcia, Temporal trends in deep ocean Redfield ratios, Science, 289, 1839a, 2000.

Supporting Information

  1. Top of page
  2. Abstract
  3. 1. Introduction to a New Three-End-Member Mixing Model
  4. 2. Steps in the Model Calculation
  5. 3. Results and Discussion
  6. 4. Overall Discussion
  7. 5. Summary and Conclusions
  8. Acknowledgments
  9. References
  10. Supporting Information

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.