SEARCH

SEARCH BY CITATION

Keywords:

  • calcium carbonate;
  • remote sensing;
  • MODIS;
  • coccolithophores;
  • coccoliths;
  • particulate inorganic carbon

Abstract

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Methods
  5. 3. Results
  6. 4. Discussion
  7. Acknowledgments
  8. References
  9. Supporting Information

[1] We describe a two-band algorithm for the remote quantification of the ocean's suspended calcium carbonate (also known as particulate inorganic carbon (PIC)), based on normalized water-leaving radiance at 440 and 550 nm. We tested this algorithm using ship-derived and satellite-derived results from a variety of marine environments. From this validation work we calculated the overall accuracy of the satellite-based PIC estimates, assuming different timescales and space scales for binning. After performing the validation work we applied the two-band algorithm to water-leaving radiance data from 2002, sampled by Moderate-Resolution Imaging Spectroradiometer (MODIS)/Terra (a 36-band satellite spectrometer designed to observe land, ocean, and atmosphere), and we derived seasonal, global maps of the standing stock of pelagic PIC as well as particulate organic carbon (POC). These data, along with limited observations on the turnover time of calcium carbonate coccoliths in the euphotic zone, provide some new insights into global rates of pelagic calcite production.

1. Introduction

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Methods
  5. 3. Results
  6. 4. Discussion
  7. Acknowledgments
  8. References
  9. Supporting Information

1.1. Production and Distribution of Particulate Inorganic Carbon in the Surface Ocean

[2] Particulate inorganic carbon (PIC; or CaCO3) is found throughout the coastal and open oceans of the world and it represents about 1/4 of all marine sediments [Broecker and Peng, 1982; Seibold and Berger, 1982]. CaCO3 is produced in shallow waters by either coral reefs or macrophytic algae (e.g., Halimeda), or in the plankton, by coccolithophores, foraminifera, and pteropods [Milliman, 1993]. These plankton eventually fall through the water column, and are deposited in shallow and deep sea sediments. Mineralization of their CaCO3 shells can involve five allotropic modifications, with calcite and aragonite being the most abundant [Stumm and Morgan, 1981]. Previous calculations of the inventory of total alkalinity and residence times of various water masses have demonstrated that calcification by oceanic and shelf plankton accounts for ∼2/3 of global calcification [Milliman, 1993]. The current global CaCO3 production rate is thought to be ∼0.6 Gt C as CaCO3 yr−1 with ∼0.3 Gt C produced in the deep sea [Milliman, 1993]. Two other estimates of global calcification in shelf, slope and oceanic waters are ∼0.9 Gt C yr−1 [Morse and Mackenzie, 1990] and ∼1.1 Gt C yr−1 [Wollast, 1994], almost 2 × higher than Milliman [1993]. Wollast [1994] suggested that sediment trap artifacts or overestimates of dissolution were the cause of lower global calcification estimates. Milliman et al. [1999] estimated that pelagic production of calcium carbonate was 0.7 Gt PIC yr−1. This value was closer to the range of deep sea CaCO3 dissolution estimates (0.7–0.9 Gt PIC yr−1) given by Archer [1997], but still less than the estimated global CaCO3 rain rate of ∼1 Gt PIC yr−1 [Archer and Maier-Reimer, 1994; Archer, 1997]. More recently, Feely et al. [2004] used the seasonal cycle of euphotic zone alkalinity to estimate the range of annual calcite production. Their values varied from 0.8–1.4 Gt PIC yr−1.

1.2. Marine CaCO3 Cycle: Biogeochemistry Implications

[3] The burning of fossil fuels appears to have major impact on oceanographic, ecological and physiological conditions [Sabine et al., 2004] through effects (sometimes opposing) on temperature, stratification, pH, etc. [Bates et al., 1996; Lavrentyev et al., 2001; Peng et al., 1998; Riebesell et al., 1993, 2000, 2001; Stockwell et al., 2000]. Nonetheless, it is not yet known whether the net effect of increased pCO2 will shift the ratio of export of particulate organic carbon to particulate inorganic carbon. Such a change could then affect the ability of the ocean to act as a CO2 sink. It is clear that not enough is known about the calcium carbonate standing stock in the sea to allow unambiguous prediction of its response to different forcing.

[4] Anthropogenic CO2 can be neutralized by reactions with marine calcium carbonate via the dissolution reaction

  • equation image

As the oceans become enriched in fossil fuel-derived CO2, the locations and extent of dissolution will increase as a function of changes in the CaCO3 saturation state. Feely et al. [2004] estimated that the depths of calcite and aragonite saturation have shoaled as much as several hundred meters since the beginning of the industrial revolution, due to the ventilation of anthropogenic CO2 into the sea. On longer timescales, variations in calcium carbonate cycling have been proposed to account for a significant fraction of the ∼80 ppm increase in atmospheric CO2 from glacial to interglacial times [Archer and Maier-Reimer, 1994; Boyle, 1988; Broecker and Peng, 1989]. Even on thousand year timescales, sedimentary CaCO3 is a major buffer of atmospheric CO2 [Archer et al., 1998; Broecker and Takahashi, 1977; Sundquist, 1993]. Surface production and deep dissolution of CaCO3, the ocean “carbonate pump,” has the effect of raising atmospheric pCO2 by roughly 25 ppm [Maier-Reimer, 1996; Najjar et al., 1992; Volk and Hoffert, 1985], with local surface ocean impacts as high as 50 ppm in coccolithophore blooms [Holligan et al., 1983; Robertson et al., 1994]. Thus even on short timescales, changes in the carbonate standing stocks could be significant to atmospheric CO2 budgets.

[5] Knowledge of the global distribution of calcium carbonate may help explain significant anomalies in alkalinity that can be found in the surface ocean, which ultimately result from variability in calcification and dissolution. While calcium carbonate is generally oversaturated in the surface ocean (as much as 5 times), significant negative Ca++ anomalies can be seen in the top 200 m presumably due to net calcification and subsequent sinking of the CaCO3. In slightly deeper waters (while still above the lysoclines for aragonite and calcite), there are well-documented positive anomalies in Ca++, suggesting net calcite dissolution [Milliman et al., 1999]. Taken together with sediment trap measurements, it appears that at least half of the surface-produced calcium carbonate must dissolve shallower than 1000 m to be consistent with sediment trap fluxes at that depth [Balch et al., 2000; Balch and Kilpatrick, 1996; Feely et al., 2004; Troy et al., 1997].

1.3. Case for Using Remote Sensing to Study the Calcium Carbonate Cycle

[6] Clearly, the ability to remotely map suspended CaCO3 from space would provide new insights into the global carbonate cycle. In addition to its impact on atmospheric CO2 on a variety of timescales, there are several reasons for using remote sensing to study suspended CaCO3. The first reason is the impact of CaCO3 producers on the upper ocean light field, especially during coccolithophore bloom events [Ackleson et al., 1988; Balch et al., 1989; Holligan et al., 1993, 1983; Tyrell et al., 1999]. In typical nonbloom situations, backscattering of light by suspended coccoliths routinely accounts for 10–20% of all visible backscattered light from the sea [Balch et al., 1999]. This percentage can exceed 90% in coccolithophore blooms [Balch et al., 1991] when the concentration of calcite is much higher. The second reason for using remote sensing to study the global distribution of CaCO3 is that CaCO3 producers are well known for the production of dimethylsulfide (DMS) which has links to planetary albedo [Charlson et al., 1987; Shaw, 1983]. One coccolithophore species, Emiliania huxleyi, is a significant producer of dimethylsufide [Keller et al., 1989; Malin et al., 1993; Matrai and Keller, 1993]. A third reason for remotely studying the global marine CaCO3 distribution is that our interpretation of the paleorecord in marine sediments hinges on what we know about the production and dissolution of CaCO3 in the modern ocean. Understanding extant global distributions of PIC can aid in understanding paleopatterns of PIC deposition. Lastly, CaCO3 serves as ballast for particulate organic matter, especially in the deep sea, where the relative abundance of CaCO3 in particulate material becomes dominant. Recent work [Armstrong et al., 2002; Francois et al., 2002; Iglesias-Rodrigues et al., 2001] has focused on the dominant role of ballast minerals such as calcium carbonate as a carrier of organic carbon into the ocean interior. Understanding global PIC distributions through remote sensing will eventually allow better calculation of global export fluxes of carbon.

1.4. Optical Properties and Carbon Content of Coccoliths: The Heart of the Optical Algorithm

[7] Given the importance of suspended calcium carbonate to the global carbon cycle, a remote means to detect it is essential. Unlike phytoplankton pigments, which, when present, decrease the radiance in the blue but increase it only slightly in the green, coccolithophores increase the radiance uniformly in both the blue and green [Gordon et al., 1988]. Thus their remote study requires an understanding of the actual water-leaving radiance rather than just radiance ratios as in the case of pigments [Gordon and Morel, 1983]. Furthermore, the “flattening” of the reflectance spectrum in coccolithophore blooms implies that the standard ratio pigment algorithms [Gordon and Morel, 1983] will not provide correct pigment retrievals within the blooms [Balch et al., 1989]. A previous solution to this problem was to use a coccolithophore “flag” to check for regions of high coccolithophore abundance, whereupon the associated pigment concentrations could be flagged as suspect [Brown and Yoder, 1994].

[8] A fundamentally different remote sensing approach for the remote sensing of coccolithophores is to use individual water-leaving radiance values, not ratios, to resolve the interacting effects of coccolithophore calcite and chlorophyll in seawater. Gordon et al. [1988] developed a prototype model for explaining the dependence of the water-leaving radiance on the concentration of constituents in Case 1 waters. Briefly, the normalized water-leaving radiance is related to the absorption and scattering properties of the biogenic components of the water, phytoplankton and their associated detritus. This model provides the basis for extraction of the concentration of the detached coccoliths, or the coccolith PIC concentration, based on water-leaving radiance measurements and known optical properties of calcium carbonate.

[9] The calcite-specific backscattering cross section for coccolithophores is critical to the above optical PIC algorithm. This has been previously determined to be about 1.37 m2 mol PIC−1, based on lab and field studies [Balch et al., 1999]. Obviously, this property is related to the size, shape and mass of the calcite coccoliths [Gordon and Du, 2001]. It is important to point out that even within the species E. huxleyi, there can be variability in the PIC content of coccoliths, especially when comparing results from culture and field measurements [Paasche, 2002]. Such variance will invariably affect the coccolith- and PIC-specific backscattering cross sections, and, in turn, the accuracy of any optically derived PIC estimates. This variance exists because of assumptions about the numbers of coccoliths attached to each coccolithophore as well as effects of growth conditions on coccolith production. For example, PIC content per E. huxleyi coccolith has been estimated in many studies: 0.16 pg [Paasche, 1962], 0.45–0.5 pg for field measurements from a NE Atlantic bloom [Holligan et al., 1983], 0.2 pg for measurements of a Gulf of Maine bloom (see Balch [1991], but note correction [Balch, 1991]), 0.26 pg based on both field and culture measurements [Balch et al., 1992], 0.21 pg (P. M. Holligan, unpublished calcite measurements for the NE Atlantic bloom measurements as cited by Balch et al. [1992]), 0.47–1.05 pg for measurements from a North Atlantic bloom [Balch et al., 1996a; Fernández et al., 1993].

[10] The above observations suggest that field measurements of coccolith carbon content often are greater than measurements from cultures. Results from nonbloom measurements in the Equatorial Pacific [Balch and Kilpatrick, 1996] and Arabian Sea [Balch et al., 2000] certainly support this contention, as do observations of Paasche [2002]). We surmise that PIC present in other relatively rare, undocumented, calcite particles (e.g., other rare coccolithophores, foraminifera, and pteropod species, resuspended calcite particles or even fecal pellets containing coccoliths [Keller et al., 1992]) contribute to the elevated PIC coccolith content measured in the field.

[11] More recent culture-based estimates of PIC per coccolith for E. huxleyi (based on Ca measurements from cell-free suspension) have shown values of about 0.2 pg PIC, consistent with previous culture measurements: 0.18–0.25 pg PIC for an Atlantic clone [Paasche, 1998, 1999], 0.20–0.23 pg for a clone from coastal Scandinavia [Paasche, 1999; Paasche et al., 1996], 0.27 pg for a Mediterranean clone [Paasche, 1999]. X-ray analysis of individual coccoliths showed 0.201 pg PIC per coccolith [Fagerbakke et al., 1994]. The above cell-free PIC estimates are reasonably close to the theoretical value of 0.276 pg estimated by Young and Ziveri [1999] using coccolith dimensions.

[12] These observations suggest that the overall strategy for designing a radiance-based PIC algorithm clearly should be to incorporate the light scattering properties of E. huxleyi, consistent with (1) its large numerical abundance in the global ocean (including frequent mesoscale blooms) and (2) its relatively stable PIC coccolith content and an optical backscattering cross section for E. huxleyi–sized coccoliths that is orders of magnitude greater than for larger calcite-containing particles in the sea, such as foraminifera and pteropods [Balch et al., 1996a]. This means that the bulk of calcite-related reflectance likely originates from small, E. huxleyi–sized coccoliths, not larger (and much less abundant) foraminfera and pteropods. While most historical literature supports a value of ∼0.2 pg PIC per coccolith to convert coccolith concentration to PIC, it is likely that algorithm PIC estimates based on this conversion may be underestimates of the total PIC concentration since they nonetheless neglect rare, relatively large, suspended calcite particles.

[13] In this paper, we present a formalized description of a two-band algorithm for the determination of the concentrations of detached coccoliths, suspended calcium carbonate and chlorophyll based on normalized water-leaving radiance at 440 and 550 nm. We present validation results for this optical algorithm based on shipboard measurements from a variety of environments. Next we calculate the overall accuracy of the satellite-based PIC estimates, assuming different time and space binning. We apply the two-band algorithm to water-leaving radiance data from MODIS/Terra (a 36-band satellite spectrometer designed to observe land, ocean and atmosphere) to produce global PIC maps. Last, we make some preliminary inferences about the expected global rates of calcite production based on limited observations from a few ocean regions of the turnover time of calcite coccoliths in the euphotic zone.

2. Methods

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Methods
  5. 3. Results
  6. 4. Discussion
  7. Acknowledgments
  8. References
  9. Supporting Information

[14] As described below, the two-band PIC algorithm is based on a semianalytic model of ocean color developed by Gordon et al. [1988] to explain the variation of water-leaving radiance with pigment concentration observed in radiance measurements of Clark [1981]. (Note that the pigment concentration (C) is defined as the sum of the concentrations of chlorophyll a and phaeophytin a measured fluorometrically.)

[15] Field validation measurements used to test this algorithm were made in the Straits of Florida, Arabian Sea, North Atlantic (south of Iceland), and Gulf of Maine. We also include measurements associated with an experimental addition of Cretaceous chalk to surface seawater in the NW Atlantic (dubbed “Chalk-Ex,” complete details to be published elsewhere). Most of the data, in fact, come from the Gulf of Maine, due to the proclivity of naturally occurring coccolithophore blooms in the region as well as the presence of our long-term sampling program aboard a passenger ferry, the M/S Scotia Prince [Balch, 2001; Balch et al., 2004].

[16] Shipboard validation consisted of parallel measurements of acid-labile backscattering (bb) with a multiangle light-scattering photometer [Balch et al., 2000, 2001; Balch and Drapeau, 2004; Balch et al., 1999, 1996a] and PIC determinations [Balch et al., 1996a; Balch and Kilpatrick, 1996; Fernández et al., 1993] on sea water samples. We also validated the algorithm using measurements of upwelling radiance, sky radiance and downwelling irradiance at 443 and 555 nm, made with a Satlantic SAS system mounted on the above-mentioned ferry, using protocols described by Mueller et al. [2003]. Briefly, the location of the radiance sensors on the bow of the ferry avoided any potential wake contamination in the upwelling radiance data. The upwelling (ocean) and downwelling (sky) radiance sensors were always pointed 40° from nadir and zenith, respectively, and 90°–120° from the Sun's azimuth to minimize Sun glint. A downwelling irradiance sensor was located on the uppermost deck of the ship, away from any of the ship's superstructure, in order to achieve a full view of the sky. All data were sampled at 16 Hz and only the lowest 5% of the data were used, in order to eliminate highly reflective white caps. SAS-based PIC estimates are reported only if made under clear-sky conditions.

[17] MODIS/Terra data used in the validation activities were sampled at 1 km resolution. Data for global images were used at 36 km resolution. Version 4.1 of the MODIS processing software was used to process the data. MODIS processed data are assigned a product quality level ranging from 0 (best) to 3 (worst). The quality of the MODIS level 2 products for the coccolith concentration, the PIC concentration and pigment concentration in the presence of coccoliths, depend directly on the quality of the input data (normalized water-leaving radiance at 443 nm and 551 nm). Therefore these three products have been assigned the quality level of the input data. Moreover, if the calculated PIC concentration was ≤0 or >1000 mg PIC m−3 (exceeding the highest level ever observed in a bloom) then the quality level for the coccolith/PIC products and coccolithophore pigment concentration was assigned a quality level of 3 (worst). For the results reported here, we required that the input data, as well as the resultant calcite products, all had quality flags set to zero (highest quality).

3. Results

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Methods
  5. 3. Results
  6. 4. Discussion
  7. Acknowledgments
  8. References
  9. Supporting Information

3.1. Moderate-Resolution Imaging Spectroradiometer (MODIS) Two-Band Algorithm for Coccoliths and Suspended Calcite

[18] Using the data of Clark [1981], the Gordon et al. [1988] semianalytic radiance model successfully explained the dependence of the blue-green water-leaving radiance ratio on the pigment concentration (Figure 1a). It actually fits the ratio data better over the full range of pigment concentration than the power law fits usually used in the analysis of remotely sensed ocean color data [e.g., see Gordon and Morel, 1983]. It was also moderately successful at relating the actual radiances themselves to the pigment concentration. Figures 1b and 1c compare the computed and observed relationship between [Lw(443)]N, [Lw(550)]N and pigment concentration (C) for the same data shown in Figure 1a. In the figures, the plankton-detritus scattering parameter was adjusted to provide the best fit for C < 0.3 mg m−3. The resulting value of this parameter was well within the range generally found for Case 1 waters [Gordon and Morel, 1983]. Also, the “noise” in the relationship for C > 0.3 mg m−3 appears to be due to the natural variation in the backscattering of plankton and detritus, as it is consistent with the noise observed in the scattering-chlorophyll relationship for Case 1 waters [Gordon and Morel, 1983]. It was straightforward to introduce detached coccoliths from coccolithophores into the model by simply including their contribution to the backscattering.

image

Figure 1. (a) Normalized water-leaving radiance ratio between 443 and 550 nm as a function of the pigment concentration. (b) Normalized water-leaving radiance at 443 nm as a function of the pigment concentration. (c) Normalized water-leaving radiance at 550 nm as a function of the pigment concentration. Data points are from Clark [1981], and the curve is the result of the semianalytic model [Gordon et al., 1988]. The plankton-detritus scattering parameter in the model has been adjusted to provide the best fit to the Clark [1981] data for concentrations less than 0.3 mg m−3.

Download figure to PowerPoint

[19] By direct measurement of detached coccoliths in the Gulf of Maine, Balch et al. [1991] showed that at 436 and 546 nm the backscattering coefficient, bb(λ), of the detached coccoliths was approximated by

  • equation image

where Ccc was the concentration of detached coccoliths, A (436) = 1.5 × 10−13 m2 coccolith−1, and A(546) = 1.1 × 10−13 m2 coccolith−1. On the basis of these measurements the spectral variation of bb(λ) was best described by equation (3):

  • equation image

Figures 2a–2c provide the radiance ratio, [Lw(443)]N/[Lw(550)]N as functions of C and Ccc, as derived from the radiance model. Figure 2a shows that if the blue-green ratio were applied to ocean color data without regard for the presence of coccoliths considerable error would be present in the retrieved pigment concentration. For example, if the measured ratio was 4 and Ccc was 75 × 106 coccoliths L−1, C would be approximately 0.05 mg m−3 but if coccoliths were ignored, the retrieved value of C would be closer to 0.11 mg m−3. Examination of Figures 2b and 2c shows that, while the coccolith concentration exerts a strong influence on both [Lw(443)]N, C still has a large influence on [Lw(443)]N but only a moderate influence on [Lw(550)]N. This suggests that it should be possible to separate, with reasonable accuracy, C and CCC in measurements of [Lw(443)]N and [Lw(550)]N. The two-band algorithm simply consists of inverting the semianalytic model to derive C and Ccc from measurements of [Lw(443)]N and [Lw(550)]N. This is accomplished with the aid of a lookup table that is graphically represented in Figure 3. Examination of Figure 2b suggests that the natural variation in phytoplankton backscattering for C < 10 mg m−3 corresponds to 0 to 25 × 106 coccoliths L−1. Thus given accurate values of [Lw(λ)]N there will always be an uncertainty in Ccc of at least 25 × 106 coccoliths L−1. Figure 3 suggests that the sensitivity of the radiances to Ccc falls by about a factor of 2 from low to high C. Note, Figure 3 provides a simultaneous method for deriving both Ccc (or the equivalent PIC concentration assuming the coccoliths are from E. huxleyi) and C in coccolithophore blooms; however, the sensitivity when the pigment concentration is >2 mg m−3 is poor.

image

Figure 2. Results of the semianalytic model [Gordon et al., 1988] with increasing coccolith concentration (Ccc), plotted along with data points from Clark [1981]. (a) Normalized water-leaving radiance ratio between 443 and 550 nm as a function of the pigment concentration. Curved lines are for different values of Ccc, increasing from 0 to 200 × 106 coccoliths L−1, in steps of 25 × 106 coccoliths L−1 (the higher coccolith concentrations yield the “flatter” curves). (b) Normalized water-leaving radiance at 443 nm as a function of the pigment concentration. (c) Normalized water-leaving radiance at 550 nm as a function of the pigment concentration. For both Figures 2b and 2c, curved lines are for different values of Ccc from 0 to 200 × 106 coccoliths L−1, increasing in steps of 25 × 106 coccoliths L−1, from bottom to top.

Download figure to PowerPoint

image

Figure 3. Graphic representation of the lookup table used to estimate the coccolith concentration and the pigment concentration from [Lw(443)]N and [Lw(550)]N. The lines with the steeper slopes are lines of constant pigment concentration specified in units of μg pigment L−1 (written along the top curve on the figure). The lines with more gentle slopes are lines of constant coccolith concentration. The concentration is written along the rightmost curve of constant pigment either in units of 106 coccoliths L−1 (upper numbers) or after conversion to particulate inorganic carbon (PIC) in units of μg PIC L−1 (lower numbers). The natural variability of phytoplankton-detritus backscattering approximately corresponds to the separation between adjacent Ccc isopleths.

Download figure to PowerPoint

3.2. Influence of Atmospheric Correction on the Accuracy of the Two-Band Algorithm

[20] As the two-band PIC algorithm uses absolute values of the water-leaving radiances, it is more susceptible to error in atmospheric correction than algorithms employing radiance ratios. Thus atmospheric correction could be an important source of error, over and above the inherent error in the algorithm due to natural variability in calcite particles in the sea. The atmospheric correction algorithm [Gordon and Wang, 1994; Gordon, 1997] uses near infrared (NIR) spectral bands to assess the atmospheric interference based on the observation that [Lw(λ)]N is usually negligibly small in the NIR. The algorithm has an inherent error of ±0.002 and ±0.0005 in normalized water-leaving reflectance at 443 and 550 nm, respectively. (Normalized water-leaving reflectance [ρw(λ)]N is related to [Lw(λ)]N through [ρw(λ)]N = π[Lw(λ)]N/F0(λ), where F0 is the extraterrestrial solar irradiance.) However, with increasing coccolith concentration, there will be a nonnegligible [Lw(λ)]N in the NIR, and this will lead to additional error in atmospheric correction: the NIR [Lw(λ)]N will be interpreted by the algorithm as a contribution from the atmosphere. Thus we examine the influence of atmospheric correction errors for both low (negligible [Lw(λ)]N in the NIR) and high (significant [Lw(λ)]N in the NIR) coccolith concentrations.

[21] For low coccolith concentrations and the Sun near the zenith, the errors in atmospheric correction described above correspond to uncertainties of ±0.12 and ±0.03 mW cm−2 m−1 Sr−1 in [Lw(λ)]N at 443 and 550, respectively. It should be noted that by the nature of the atmospheric correction algorithm, these errors will have the same sign, i.e., both will be positive or negative. For a coccolith concentration of 15 × 106 coccoliths L−1, the error in Ccc will be approximately ±2 × 106 coccoliths L−1 for C near 0.2 mg m−3 and approximately ±1 × 106 coccoliths L−1 for C near 1 mg m−3. Thus for low Ccc the atmospheric correction–induced error in Ccc is negligible compared to that induced by the natural variability in plankton-detritus backscattering. The error induced in C however, is quite large for C > about 1 mg m−3.

[22] To assess the influence of atmospheric correction errors at high Ccc, we need to estimate [ρw(λ)]N in the NIR. As the absorption coefficient of water is very large in the NIR, the pigment concentration is almost irrelevant in predicting the reflectance, so [ρw(λ)]Nbb(λ)/6aw(λ), where aw(λ) is the absorption coefficient of water. For Ccc = 100 × 106 coccoliths L−1, this gives [ρw(765)]N ≅ 0.00045, and [ρw(865)]N ≅ 0.00022, and doubling Ccc will simply double these values. These reflectances will be interpreted by the correction algorithm as an addition to the reflectance of aerosols ρA(λ). Typical values of ρA(λ) in a clean maritime atmosphere are ∼0.01 with little dependence on wavelength. The correction algorithm uses the aerosol reflectances at 765 and 865 nm to assess the spectral variation ρA(λ) in the NIR and then uses aerosol models to extrapolate this into the visible. Following the approximate single scattering development provided by Gordon [1997], ρA(λ) = exp[a (865 − λ)] ρA(865), where the parameter a is found by setting λ to 765 nm. Adding the NIR reflectance contribution from the coccoliths to the true ρA(λ) for a clean maritime atmosphere, i.e., 0.01, determining the apparent a, and computing the apparent ρA(443) and ρA(550), we arrive at the error in [Lw(443)]N and [Lw(443)]N provided in Table 1a. The error induced in the retrieved Ccc by the coccolith-induced NIR water-leaving reflectance can be determined by picking a point (Ccc,C) on Figure 3, adding Δ[Lw(443)]N and Δ[Lw(550)]N to the normalized water-leaving radiances at the point and determining the change in Ccc and C. The resulting change in Ccc for both low and high pigment concentration is provided in Table 1b. This suggests that the error induced by NIR coccolith reflectance is also negligible compared to the natural variability in plankton-detritus scattering.

Table 1a. Error in Retrieved [Lw(λ)]N Due to Coccolith-Induced Water-Leaving Reflectance in the Near-Infrared (NIR)
Ccc, coccoliths L−1Δ[Lw(443)]N, mW/cm2μ SrΔ[Lw(550)]N, mW/cm2μ Sr
100 × 106−0.066−0.057
200 × 106−0.155−0.115
Table 1b. Error in Retrieved Ccc Due to Coccolith-Induced Water-Leaving Reflectance in the NIR
C, mg/m3Ccc, coccoliths L−1ΔCcc, %
0.07100 × 106−4.2
1.2100 × 106−6.4
0.07200 × 106−5.3
1.4200 × 106−5.4

3.3. Validation Results

[23] When ship and satellite validation data were plotted using log axes (which highlights errors at low PIC concentration (Figure 4a)) and linear axes (which better shows the high PIC concentration results from blooms (Figure 4b)), the statistics showed an overall RMS error of 28 μgPIC L−1 for 1 kilometer pixel data on any given day. It was possible that the variance about the lines in Figures 4a and 4b, particularly at low PIC concentrations, arose from other non-PIC, nonorganic particles such as biogenic silica. Using only the satellite-derived validation data plotted (i.e., not including ship radiance data), the RMS error was reduced by about a factor of two (to ∼15 μgPIC L−1 (Figure 4c)). This may indicate that the satellite-derived radiances are more accurate than the ship-derived radiances (which might be expected, given errors in azimuthal and nadir viewing geometries due to ship roll). The causes of the above RMS error may have resulted from a mismatch in the background, noncalcite, bb value assumed in the algorithm versus that present in the field. For example, it can be seen in the 2002 Gulf of Maine bloom satellite results (designated with open circles in Figure 4c) retrievals where the satellite-derived PIC values approached zero while the ship-derived values were about 30 ug C L−1. Such “blank” bias varied for the other field data sets in Figure 4c, with data falling above and below the 1:1 regression line. Another potential source of variability between ship and satellite measurements might have been due to errors in atmospheric correction.

image

Figure 4. (a) Plot of optically derived PIC concentration versus ship-measured values (based on inductively coupled plasma atomic absorption spectrometry measurements of particulate material). The y axis includes an extra scale with the intermediate acid-labile backscattering value (bb) used to optically estimate PIC. Conversion of bb to PIC assumes 1.37 m2 mol PIC−1 = 1.14 × 10−4 m2 [mg PIC]−1. The data sets used to make this plot were 1991 Iceland coccolithophore bloom (asterisks), Arabian Sea 1995 (open triangles), Straits of Florida 1995 (solid triangles), flow cytometer analysis of sorted coccoliths (solid stars), Gulf of Maine Ferry 1998–2001 (solid dots), Chalk-Ex ship measurements November 2001 (black crosses), Moderate-Resolution Imaging Spectroradiometer (MODIS) Gulf of Maine 2000–2001 (shaded squares with white cross within), Sea-viewing Wide Field-of-view Sensor (SeaWiFS) Gulf of Maine 1998–2001 (solid inverted triangles), Gulf of Maine coccolithophore bloom ship measurements (open crosses, pound symbols), Gulf of Maine July 2002 coccolithophore bloom MODIS measurements (open circles), and Chalk-Ex MODIS measurements November 2001 (solid shaded squares). There are both satellite and ship data shown in this plot. The two bold, black, curved lines that enclose the data distribution were drawn by eye. The bold straight line is the least squares linear fit to the data, with the standard error given in parentheses [bb(0.0033) = 1.145E-4 (1.576E-6) × PIC; r2 = 0.60; n = 1783; RMS error in PIC concentration = 28.56 ug PIC L−1; P < 0.001]. (b) Same as Figure 4a, but using linear axes instead of log axes. The complete statistics for the plot are the same as in Figure 4a. (c) Same as Figures 4a and 4b, except only satellite-derived PIC data are shown. The bold straight line is the 1:1 line for PIC. The equation describing the best least squares linear fit to the data is [bb = 9.665E-5 × PIC; r2 = 0.55; n = 463; RMS error in PIC concentration = 14.9 ug PIC L−1; P < 0.001].

Download figure to PowerPoint

[24] Even with the increased accuracy associated with using just the satellite-derived PIC estimate, it is difficult using 1 km daily data to accurately derive background oceanic PIC concentrations from satellite since the average global PIC concentration is ∼2 μgPIC L−1, less than the RMS error about the line in Figure 4c.

[25] The solution to the above problem, that the background PIC concentration is less than the standard deviation of the technique used to measure it, is to bin the satellite data in space and time, the sample size could be made sufficiently large that the standard errors were then significantly reduced. This is shown in the table of standard errors for the PIC algorithm (Table 2) where, at spatial scales of 4.6 km and time binning of 8 days, the standard errors of the PIC estimates were ∼1.2 μgPIC L−1, slightly less than the average global PIC concentration. For data binned over 36 km and 30 days, the standard error was 0.08 μgPIC L−1, 1/16 of the average PIC concentration in the global ocean. Space/time binning clearly allowed more confidence in interpreting global images of surface PIC.

Table 2. Standard Error Estimates for Remote Particulate Inorganic Carbon (PIC) Measurements Binned at Different Timescales and Space Scalesa
Time Bins, daysSpatial Resolution, km
14.6336111.2
  • a

    PIC measurements are in μg PIC L−1. The 111.2 km spatial scale is equivalent to 1 degree of latitude. Bold values show binning, which provides a standard error less than globally averaged PIC of ∼2 μg PIC L−1.

114.903.2180.4140.134
85.6321.2160.1560.051
302.7200.5880.0760.024
3650.7800.1680.0220.007

4. Discussion

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Methods
  5. 3. Results
  6. 4. Discussion
  7. Acknowledgments
  8. References
  9. Supporting Information

4.1. Potential Interactions From Other Suspended Minerals

[26] A potential source of error, not considered in the PIC algorithm, was the presence of other suspended minerals in the water, such as large concentrations of diatom silica (opal). The data used to make the two-band PIC algorithm were from the field, and indeed contained diatoms and their associated opal silica. Thus some degree of “opal contamination” is already inherent in the algorithm, which likely contributed to its overall error budget (but also we implicitly compensated for the presence of this mineral since it occurs naturally in the same seawater that the original algorithm was derived from). Nonetheless, it is well known that there are regions of the ocean that can have high concentrations of diatoms (e.g., Southern Ocean [Brzezinski et al., 2001]). In the Bering Sea, some regions of high scattering, once thought to be caused by coccolithophores, have been shown to be resuspended diatom frustrules from the sea floor [Broerse et al., 2003]. Thus the issue is to define the error in the PIC algorithm when the opal:PIC ratio in surface waters was different from that used in the original algorithm development. In other words, we need to define how robust the PIC algorithm was under typical and atypical oceanic concentrations of biogenic silica.

[27] A first-order estimate of the error from suspended silica can be made quite easily. Brzezinski et al. [2001] observed concentrations of 12–16 μmol of biogenic silica per liter in intense diatom blooms in the Southern Ocean. In our laboratory, we have measured the mass-specific backscattering coefficient of suspended diatom frustules (0.624 m2 (mol Si)−1) [Broerse et al., 2003]. Thus the amount of Si observed by Brzezinski would contribute 7.5 × 10−3 to 10 × 10−3 m−1 of backscattering over and above that due to POC- or PIC-containing particles. This quantity of biogenic silica in a diatom bloom indeed could cause significant error in the PIC determination. However, more typical biogenic silica concentrations (1–3 μmol L−1) would only produce additional backscattering of 0.62 × 10−3 to 1.87 × 10−3 m−1; such quantities likely are already “built into” the current PIC algorithm. In short, the extent of error due to biogenic silica cannot be estimated until radiometric data sets become available that would allow the partitioning of variance between opal and calcium carbonate.

4.2. Recent Observations of Global Particulate Inorganic Carbon (PIC)

[28] The launch of MODIS/Terra and MODIS/Aqua, along with the above-described validation results, have allowed the first opportunities to map global PIC with definable errors (Table 2). We have assembled version 4.1 MODIS/Terra data from 2002 into images of seasonal average PIC (Figure 5). There were some striking features in these data. From October to March (including Austral spring and summer), there were large regions of the Polar Convergence Zone and subpolar front that appear to have relatively high concentrations of PIC. Note, the peak concentration shown in Figure 5 is 10 μg PIC L−1, which, by no means should be considered a discolored coccolithophore bloom, but still is significantly higher than typical values seen in the central ocean gyres. From April to September (including Northern Hemisphere spring and summer) the Bering Sea, North Atlantic and Barents Sea showed high concentrations of PIC. The Namibian upwelling region, off of West Africa, showed dramatic increases in PIC concentration between October and December (Figure 5d).

image

Figure 5. Global composite images of suspended PIC concentration calculated from MODIS/Terra data using two-band calcite algorithm. See text for other details of how the data were processed. The color scale is highlighted in Figure 5c. These data were binned into 36 km and 90 day averages, and thus the standard error will be <0.08 μgPIC L−1 (see Table 2), well below the average seawater concentration of ∼2 μgPIC L−1. (a) January–March. (b) April–June. (c) July–September. (d) October–December.

Download figure to PowerPoint

[29] Given the seasonal images in Figure 5, we integrated the total surface PIC over the euphotic zone, within 10° latitudinal bands (Figures 69; Tables 3 and 4). First, the depth of the euphotic zone was estimated using satellite-derived chlorophyll concentration; the MODIS product “chlor-a2” was used as input to the algorithm for calculation of the average Kpar over the euphotic zone to the 1% light depth [Morel, 1988, equation (5)]. We then made the first-order assumption, admittedly over simplified but still reasonable based on field PIC results in non-bloom conditions, that the concentration of PIC was vertically uniform over the euphotic zone.

image

Figure 6. Statistics for 2002 PIC estimates as a function of latitude between January and March. (a) Mean PIC integrated over the euphotic zone. (b) Total euphotic zone PIC, integrated aerially in each latitudinal band. (c) Number of pixels analyzed in each latitudinal band. See text for details.

Download figure to PowerPoint

image

Figure 7. Statistics for PIC estimates as a function of latitude between April and June 2002. All else is identical to Figure 6.

Download figure to PowerPoint

image

Figure 8. Statistics for PIC estimates as a function of latitude between July and September 2002. All else is identical to Figure 6.

Download figure to PowerPoint

image

Figure 9. Statistics for PIC estimates as a function of latitude between October and December 2002. All else is identical to Figure 6.

Download figure to PowerPoint

Table 3. Total Euphotic Zone PIC as a Function of Season in 2002, Estimated From the Surface to the 1% Light Levela
 Midpoint of 10° LatitudinalAll Latitudes
85756555453525155−5−15−25−35−45−55−65−75−85
  • a

    Values given in 10° latitude increments (in megatons of carbon). These same data are graphed in Figures 6b, 7b, 8b, and 9b. Values of 0.00 represent regions with no satellite radiance estimates available, mostly due to low Sun angle at high latitudes.

Jan–Feb0.000.000.000.030.710.951.351.541.561.511.981.601.922.742.321.300.040.0019.56
Mar–Jun0.000.140.370.951.511.311.421.701.551.331.451.431.720.520.000.000.000.0015.40
Jul–Sep0.000.610.751.601.751.111.601.971.891.581.561.471.970.850.000.000.000.0018.71
Oct–Dec0.000.000.000.000.720.911.381.611.851.932.401.922.593.082.560.550.010.0021.50
Table 4. Percentage of Total Global, Euphotic PIC Poleward of 30° Latitude or the Equator as a Function of Season in 2002
 Jan–MarApr–JunJul–SepOct–Dec
% PIC N of 30°N8.6527.8031.077.57
% PIC S of 30°S42.5514.5915.0840.88
% PIC N Hemisphere31.3958.0760.3030.10
% PIC S Hemisphere68.6141.9339.7069.90

[30] Equation (25) of Morel [1988] (POC = 90 C0.57) provided a means to estimate surface POC (mg m−3) based on the mean remotely sensed pigment concentration, C. The pigment data used to generate Morel's [1988] relationship were sampled from a variety of environments ranging from oligotrophic to eutrophic regions (n = 409), spanning over three orders of magnitude of pigment concentration. Moreover, this relationship reasonably provided a range in the carbon:chlorophyll ratio of 1000 in the most oligotrophic environments to 25 in eutrophic environments, which is the typical range of field observations [Eppley et al., 1977; Geider, 1987; Steele and Baird, 1962]. In order to integrate the POC values to the base of the euphotic zone, we assumed a constant POC concentration with depth. We caution that this only provides a first-order estimate of euphotic POC since there is well known vertical heterogeneity in phytoplankton biomass over the euphotic zone [Cullen, 1990].

[31] Seasonal global totals of euphotic PIC (within all latitudes visible to the satellite) ranged from 15.4–21.5 Mt PIC (Figures 69). Moreover, the results demonstrated two global patterns of PIC distribution over the year. The first pattern prevailed between October and March, when about 69% of the PIC was in the Southern Hemisphere. During the period between April and September, the pattern shifted hemispheres, but not symmetrically; about 59% of the PIC was in the Northern Hemisphere during this period. It is clear from the images (Figure 5) that the bands with elevated PIC in the Southern Hemisphere represented a large fraction of the total global euphotic PIC. About 40% of the global PIC was found south of 30°S between October and March compared to the April to September period in the Northern Hemisphere (in which ∼29% of the PIC was found north of 30°N). Unfortunately, as of this writing, few shipboard data exist that could be used to validate these regions of elevated PIC in the Southern Convergence Zone.

4.3. Regional Analysis of PIC Distributions

[32] To provide more quantitative regional estimates of the global PIC standing stock, we used the biogeographic provinces described by Longhurst [1998] and calculated the average surface PIC concentration and its variance within each province over a season (Table 5). Owing to the high PIC concentrations observed in the Black Sea and Persian Gulf, we added separate provinces to Longhurst's [1998] original list. The Caspian Sea is not included in our budget calculations due to its isolation from the global ocean. The Chesapeake Bay province listed by Longhurst [1998] also is not included here due to its small size. To aid in global estimates, we also generated a table of average integrated euphotic zone PIC concentration (calculated as described above) and its standard deviation within each biogeochemical province (Table 6).

Table 5. Surface PIC Aerially Integrated Within the Biogeochemical Provinces Defined by Longhurst [1998]a
Surface PICJan–Mar AverageApr–Jun AverageJul–Sep AverageOct–Dec Average
ProvinceBiomeTot. PIC, MtSurf. PIC, mg/m3SD, mg/m3PIC:POCPixelsTot. PIC, MtSurf. PIC, mg/m3SD, mg/m3PIC:POCPixelsTot. PIC, MtSurf. PIC, mg/m3SD, mg/m3PIC:POCPixelsTot. PIC, MtSurf. PIC, mg/m3SD, mg/m3PIC:POCPixels
  • a

    Provinces are grouped into their respective biome: Polar, Westerlies, Trades, and Coastal. Calculations have been made for each 3 month period of 2002. Results are given for each province for (1) total PIC (essentially, the total PIC found in the top meter of the water column, aerially integrated over the province), (2) average surface PIC (mg m−3), (3) standard deviation of the average PIC (mg m−3), (4) PIC:POC ratio (where PIC was taken from the average value in the second column of each 3 month period and POC was taken from the same column of Table 7), and (5) the total number of pixels available for the analysis. Any values of 0.00 in the first three columns of the polar biome for each 3 month period represent regions where no satellite radiance data were available, mainly due to low Sun angle. At the bottom of the table the global statistics are provided as well as statistics for each of the four biomes.

Boreal Polarpolar0.00000.000.00NA00.00312.271.340.02926060.01674.122.050.06471230.00000.000.00NA0
Atl Arcticpolar0.00000.000.000.00000.00672.200.470.03148030.01113.541.120.05849820.00000.000.00NA0
Atl Subarcticpolar0.00000.000.00NA00.00824.110.990.04732250.01256.322.400.09332080.00000.000.00NA0
N Pac Epicontinentalpolar0.00081.560.990.0294910.01213.401.350.03739520.01303.291.150.05044200.00101.831.140.032501
Antarcticpolar0.03541.850.780.048251370.00000.000.00NA00.00000.000.00NA00.02571.770.470.04518530
Austral Polarpolar0.00962.560.860.04561390.00000.000.00NA00.00000.000.00NA00.00071.610.940.039685
N Atl Drift(WWDR)westerlies0.00171.733.420.0509230.00972.731.030.04136060.00832.341.100.04336080.00040.650.210.017508
Gulf Streamwesterlies0.00111.712.600.0355540.00121.070.330.0209960.00100.840.600.0279960.00081.021.320.026686
N Atl Subtrop Gyre (W)westerlies0.00250.490.290.01739020.00410.680.320.02646220.00350.580.130.03046220.00230.410.180.0174415
Mediterranean Seawesterlies0.00281.101.010.02620740.00240.930.710.02821150.00160.610.420.02321060.00291.221.440.0361978
Black Seawesterlies0.00246.071.430.0703490.009825.044.690.2993480.005112.962.750.1813500.00236.172.040.086329
N Atl Subtrop Gyre (E)westerlies0.00230.620.500.01828910.00711.601.170.03635590.00240.550.190.02135580.00180.470.330.0163034
Pac Subarctic Gyres (E)westerlies0.00040.650.260.0175840.00521.630.740.02933070.01324.181.050.07932920.00081.510.400.033473
Pac Subarctic Gyres (W)westerlies0.00080.810.420.0209060.00411.880.790.02920970.00894.021.280.06721330.00141.320.400.028982
Kuroshio Currentwesterlies0.00301.112.310.02620840.00491.530.940.02925120.00341.050.720.03024970.00281.021.950.0272116
N Pac Polar Frontwesterlies0.00410.670.240.01751250.00921.300.570.03059960.00751.030.580.03361570.00520.780.300.0215646
N Pac Subtrop Gyrewesterlies0.00540.400.200.01995670.00660.480.200.02398370.00640.460.100.02698390.00530.400.180.0209584
Tasman Seawesterlies0.00110.650.280.02113840.00181.071.200.02313720.00231.361.740.03013700.00281.650.770.0341386
S Pac Subtrop Gyrewesterlies0.01580.570.270.032196870.01800.510.230.021253910.01940.550.270.021255710.02370.800.360.03721359
S Subtrop Convergencewesterlies0.02251.310.670.036147240.01290.940.450.022115780.01430.930.670.025130390.02821.640.620.04014790
Subantarcticwesterlies0.05131.650.690.042311570.00331.180.650.02724500.00460.800.490.02351890.06001.900.590.04531728
N Atl Trop Gyretrades0.00500.683.600.02950240.00640.813.110.03054890.00610.763.050.03255140.00510.663.400.0275360
W Trop Atltrades0.00210.530.560.01925640.00320.590.280.01735140.00360.660.350.01935350.00320.620.300.0203314
East Trop Atltrades0.00130.490.190.01517580.00280.550.220.01433510.00430.940.420.02229810.00450.970.310.0283048
S Atl Gyretrades0.01120.640.320.032127500.01040.580.210.020131330.01330.750.380.024128930.01971.120.660.04012847
Caribbeantrades0.00791.904.650.05428720.00511.233.500.03828750.00571.383.450.04328750.00731.774.530.0512871
Indian Monsoon Gyrestrades0.00840.540.340.020102820.00740.470.310.016102870.01130.720.510.021102840.01070.690.320.02310270
Indian S Subtrop Gyretrades0.00920.530.170.030123920.00950.550.220.020124050.01050.610.270.020124070.01210.700.250.03112400
N Pac Trop Gyretrades0.00600.440.260.02095840.00760.510.160.024105110.00800.540.150.026105860.00580.400.200.01910281
N Pac Eq Countercurrenttrades0.00570.640.310.01959180.00730.810.690.02359460.00650.730.300.02559290.00600.670.240.0245941
Pac Eq Divergencetrades0.00870.600.280.01994890.00830.570.230.01694920.00920.640.340.01993850.01080.750.390.0249444
W Pac Warm Pooltrades0.00590.480.230.02480990.00570.460.140.02181910.00650.520.160.02481900.00640.520.180.0258190
Archipelagic Deep Basinstrades0.00760.822.130.03062940.00680.741.520.02463210.00880.961.720.02763170.00760.821.470.0296340
NE Atl Shelvescoastal0.001819.009.120.258950.00426.522.740.0507770.00274.401.570.0507380.00016.285.190.13013
Canary Coastal (EACB)coastal0.00182.762.900.0364800.00223.262.710.0344980.00111.751.590.0324850.00152.382.420.045465
Guinea Current Coastalcoastal0.00051.351.080.0242420.00121.211.720.0206420.00111.651.090.0324290.00131.731.220.034484
Guianas Coastalcoastal0.00544.497.870.0757960.00423.416.180.0468120.00463.827.450.0617980.00604.889.100.085811
NW Atl Shelvescoastal0.00584.335.490.06711520.00542.602.980.03118770.00864.103.890.07118980.00413.063.910.0541150
Brazil Current Coastalcoastal0.00121.191.500.0317240.00151.511.900.0367230.00181.792.620.0417200.00161.552.090.037728
SW Atl Shelvescoastal0.00483.691.580.04013110.00032.301.300.0311120.00216.946.960.1112780.00675.172.360.0591310
Benguela Current Coastalcoastal0.00312.852.520.0407830.00222.041.250.0267750.00272.611.230.0407520.00444.132.110.057779
E Africa Coastalcoastal0.00451.281.810.03725650.00511.452.370.03225660.00641.823.170.04125670.00561.592.300.0422568
Red Seacoastal0.00081.822.340.0363180.00071.461.680.0353300.00061.371.710.0353280.00081.721.850.036321
Persian Gulfcoastal0.00179.887.570.0991240.00074.282.360.0501250.00063.642.370.0561260.00169.297.970.122125
NW Arabian Upwellingcoastal0.00421.231.300.01923220.00320.951.240.02122670.00712.351.840.03620320.00461.330.930.0252325
E India Coastalcoastal0.00222.434.750.0416110.00262.895.400.0506200.00374.326.470.0795890.00253.455.640.060502
W India Coastalcoastal0.00354.337.960.0555420.00303.757.850.0525400.00355.328.570.0644430.00283.506.510.050546
Australia Westcoastal0.00381.382.160.04119610.00602.193.090.04719650.00822.954.470.06419760.00451.632.140.0441966
Alaska Downwelling Coastalcoastal0.00000.920.000.01720.00183.741.570.0325620.00224.571.930.0585540.00000.000.00NA0
California Upwelling Coastalcoastal0.00281.251.680.02717470.00471.841.340.03220420.00451.671.410.03721660.00301.301.340.0311818
Centr American Coastalcoastal0.00252.143.020.0327910.00242.092.100.0327870.00252.173.880.0477770.00252.132.920.042785
Chilecoastal0.00492.001.570.03318730.00422.021.510.02915170.00381.981.350.03514380.00652.711.390.0451867
China Sea Coastalcoastal0.008012.5713.560.1594710.00202.853.390.0455140.00192.703.720.0545200.00537.819.200.130498
Sundacoastal0.00932.474.650.05924900.00641.663.410.04025280.00741.943.560.04725240.00862.244.450.0582532
East Australian Coastalcoastal0.00080.770.970.0317480.00080.800.990.0237460.00100.991.150.0277500.00090.870.870.030751
New Zealand Coastalcoastal0.00211.731.050.03711080.00112.803.630.0513410.00121.892.200.0425790.00312.601.390.0501110
Unclassifiedcoastal0.00041.032.470.0343470.00051.275.310.0343070.00071.715.930.0493270.00051.111.930.038324
Summary 0.31752.123.160.0482383070.27712.133.330.0412098600.32902.232.170.0512187500.34601.901.920.048232814
Polar 0.0461.990.880.040317670.0302.991.040.036145860.0534.321.680.067197330.0271.730.850.03819716
Westerlies 0.1171.300.970.033959110.1002.840.940.060797860.1022.150.810.057843270.1411.400.740.03699014
Trades 0.0790.691.090.026870260.0800.660.880.022915150.0940.770.930.025908960.0990.811.020.02990306
Coastal 0.0753.733.760.062232560.0662.512.730.037236660.0792.903.230.052234670.0783.103.360.06023454
Table 6. Euphotic Zone PIC Aerially Integrated Within the Biogeochemical Provinces Defined by Longhurst [1998]a
Int. PIC Over Euphotic ZoneJan–MarApr–JunJul–SepOct–Dec
ProvinceBiomeTot. PIC, Mt% TotalAvg Int. PIC, mg/m2SD, mg/m2PixelsTot. PIC, Mt% TotalAvg Int. PIC, mg/m2SD, mg/m2PixelsTot. PIC, Mt% TotalAvg Int. PIC, mg/m2SD, mg/m2PixelsTot. PIC, Mt% TotalAvg Int. PIC, mg/m2SD, mg/m2Pixels
  • a

    Provinces are grouped into their respective biome: Polar, Westerlies, Trades, and Coastal. Calculations have been made for each 3 month period of 2002. The algorithm used to estimate POC is described in the text. Results are given for each province for (1) total euphotic zone PIC (essentially, the total PIC found above the 1% light depth of the water column, aerially integrated over the province), (2) average integrated PIC (mg m−2), (3) standard deviation of the average PIC (mg m−2), and (4) the total number of pixels available for the analysis. Any values of 0.00 in the first three columns of the polar biome for each 3 month period represent regions where no satellite radiance data were available, mainly due to low Sun angle. At the bottom of the table the global statistics are provided as well as statistics for each of the four biomes.

Boreal Polarpolar0.0000.000.000.0000.1150.7582.6637.0326060.6623.54162.7770.0571230.0000.000.000.000
Atl Arcticpolar0.0000.000.000.0000.2651.7287.3617.6348030.4722.52150.1241.3549820.0000.000.000.000
Atl Subarcticpolar0.0000.000.000.0000.2771.80138.8932.0832250.4892.62246.4686.6732080.0000.000.000.000
N Pac Epicontinentalpolar0.0370.1967.9539.084910.3792.46105.8933.8439520.5152.75130.4435.8344200.0430.2077.6141.00501
Antarcticpolar1.98810.17104.0230.24251370.0000.000.000.0000.0000.000.000.0001.4936.95102.9222.1618530
Austral Polarpolar0.3841.96102.0726.2461390.0000.000.000.0000.0000.000.000.0000.0410.1991.2243.59685
N Atl Drift(WWDR)westerlies0.0850.4387.03122.799230.3932.55110.5132.0236060.3691.97103.8337.2736080.0220.1039.4111.62508
Gulf Streamwesterlies0.0480.2473.7490.515540.0590.3850.4710.509960.0670.3657.6027.549960.0430.2053.7452.55686
N Atl Subtrop Gyre (W)westerlies0.1920.9837.8218.9539020.3282.1454.9513.7746220.3601.9260.2216.4946220.2040.9535.7213.594415
Mediterranean Seawesterlies0.1520.7859.8443.5120740.1541.0059.5632.4621150.1200.6446.7622.6921060.1740.8171.8267.831978
Black Seawesterlies0.0820.42210.4147.443490.3472.26889.93150.723480.1971.05501.7986.323500.0870.40235.6362.42329
N Atl Subtrop Gyre (E)westerlies0.1460.7540.2422.3728910.3442.2477.5138.3335590.1991.0744.9411.3535580.1330.6234.8115.863034
Pac Subarctic Gyres (E)westerlies0.0230.1238.2414.575840.2381.5574.8433.3833070.6363.40200.6045.7732920.0400.1979.9319.47473
Pac Subarctic Gyres (W)westerlies0.0440.2245.6221.639060.1621.0675.2227.0620970.3802.03172.6741.6721330.0720.3367.6017.75982
Kuroshio Currentwesterlies0.1500.7755.5767.7520840.2161.4167.1724.2225120.2061.1064.4026.6424970.1560.7256.7368.922116
N Pac Polar Frontwesterlies0.2371.2139.0913.7951250.4983.2470.4524.7959960.4912.6367.7426.8361570.3191.4948.0516.055646
N Pac Subtrop Gyrewesterlies0.5302.7139.4417.7295670.6354.1345.9813.4098370.6883.6849.8312.1498390.5432.5240.3015.619584
Tasman Seawesterlies0.0750.3844.6312.4813840.0900.5853.7345.2013720.1140.6168.3467.0813700.1400.6583.2130.631386
S Pac Subtrop Gyrewesterlies1.6338.3659.3229.48196871.59110.3545.0414.53253911.6098.6045.2716.50255712.1489.9972.1930.0621359
S Subtrop Convergencewesterlies1.3476.8978.4927.00147240.7034.5851.3221.59115780.8584.5955.9328.88130391.5987.4392.6826.4014790
Subantarcticwesterlies2.95915.1495.2729.49311570.1551.0155.7231.8324500.2851.5349.6121.6251893.33415.50105.5724.7831728
N Atl Trop Gyretrades0.3441.7647.38135.1550240.4452.8956.0299.7354890.4452.3855.77101.4355140.3391.5843.81113.765360
W Trop Atltrades0.1570.8040.1722.3925640.2061.3438.4611.5335140.2301.2342.7116.1235350.2261.0544.7817.273314
East Trop Atltrades0.0990.5137.0015.3417580.1721.1233.6510.6333510.2351.2651.7718.2329810.2941.3763.2717.913048
S Atl Gyretrades1.1155.7163.9126.65127500.7985.1944.2911.43131330.8944.7850.6716.28128931.5127.0385.9935.7412847
Caribbeantrades0.3781.9491.55171.6328720.2881.8769.60122.3628750.3271.7579.01127.7428750.3591.6786.99163.412871
Indian Monsoon Gyrestrades0.6743.4543.1919.98102820.5373.5034.4515.16102870.7073.7845.3021.15102840.7833.6450.2417.1110270
Indian S Subtrop Gyretrades0.9975.1057.9917.54123920.7434.8443.2112.96124050.7494.0143.5313.61124071.0765.0062.5414.2012400
N Pac Trop Gyretrades0.5512.8240.6721.1295840.7044.5847.5912.79105110.7734.1351.8814.32105860.5472.5537.7415.6310281
N Pac Eq Countercurrenttrades0.3811.9542.6612.6759180.4442.8949.4817.4159460.4762.5453.1114.0759290.4662.1751.9513.095941
Pac Eq Divergencetrades0.5983.0641.3314.8894890.5283.4336.4510.5094920.5943.1841.4918.1493850.7623.5452.8521.389444
W Pac Warm Pooltrades0.5923.0348.1422.0080990.5443.5443.7113.7581910.6193.3149.7616.5381900.6082.8348.8914.048190
Archipelagic Deep Basinstrades0.5282.7057.5259.0062940.4342.8247.0345.6463210.5172.7756.1158.4463170.5232.4356.5048.206340
NE Atl Shelvescoastal0.0650.33675.17282.83950.1110.72174.2677.717770.0870.46141.3745.537380.0030.02252.22181.4113
Canary Coastal (EACB)coastal0.0590.3093.0979.974800.0630.4195.5253.064980.0460.2572.0645.424850.0650.30104.7385.82465
Guinea Current Coastalcoastal0.0200.1055.4931.212420.0490.3249.9753.116420.0480.2673.2046.174290.0580.2778.4842.22484
Guianas Coastalcoastal0.1880.96155.72212.667960.1390.90112.66156.048120.1640.88135.37181.307980.2171.01176.37248.78811
NW Atl Shelvescoastal0.2071.06154.34179.9011520.1801.1786.1179.1618770.3611.93171.46132.5918980.1630.76121.47123.351150
Brazil Current Coastalcoastal0.0610.3161.5749.077240.0730.4773.3467.177230.0850.4585.5984.247200.0780.3678.0172.97728
SWR Atl Shelvescoastal0.1530.78117.3235.2813110.0100.0781.0948.241120.0810.43263.26235.932780.2261.05173.5661.891310
Benguela Current Coastalcoastal0.1020.5294.6349.747830.0750.4969.8726.827750.1090.58104.9435.837520.1670.78155.7456.88779
E Africa Coastalcoastal0.2461.2669.6949.4125650.2281.4964.6568.7325660.2961.5883.7696.9325670.2981.3984.3971.312568
Red Seacoastal0.0370.1981.2385.573180.0340.2272.2668.203300.0340.1872.1967.273280.0370.1780.0567.66321
Persian Gulfcoastal0.0510.26301.90220.311240.0240.16141.7667.591250.0250.13146.3577.051260.0570.26332.59259.81125
NW Arabian Upwellingcoastal0.1690.8649.0534.8323220.1530.9945.4331.8722670.2741.4790.7647.8320320.2181.0163.2230.412325
E India Coastalcoastal0.0740.3882.03103.516110.0900.5999.25113.176200.1370.74159.13161.755890.0920.43124.38142.15502
W India Coastalcoastal0.1030.53129.03183.795420.0890.58112.26184.555400.1020.54154.71208.504430.0960.45119.37170.12546
Australia Westcoastal0.2101.0776.5379.2619610.2691.7597.78114.5319650.3822.04138.01175.7219760.2521.1791.4792.961966
Alaska Downwelling Coastalcoastal0.0000.0037.220.0020.0490.32101.3140.925620.0780.42164.6963.355540.0000.000.000.000
California Upwelling Coastalcoastal0.1200.6253.9743.7417470.1971.2877.6734.4220420.2151.1579.4137.0721660.1450.6762.1741.321818
Centr American Coastalcoastal0.0890.4676.5173.737910.0910.5978.6548.267870.1090.5895.3496.337770.1060.4991.8186.07785
Chilecoastal0.1971.0181.4236.3818730.1541.0074.5135.9015170.1660.8987.2342.0214380.2831.32117.5742.821867
China Sea Coastalcoastal0.2421.24379.29366.314710.0740.48106.22100.985140.0850.46121.68127.935200.1910.89283.81280.71498
Sundacoastal0.4332.21114.41158.4924900.3082.0080.17114.4125280.3782.0298.68133.7325240.4261.98110.85148.382532
E Australian Coastalcoastal0.0640.3361.8567.837480.0530.3451.6355.797460.0610.3359.1260.727500.0630.2961.0250.91751
New Zealand Coastalcoastal0.1040.5386.7741.0111080.0440.29111.89134.353410.0590.3290.4389.015790.1460.68121.0749.511110
Unclassified 0.0260.1360.3766.043470.0240.1660.35122.273070.0340.1880.47156.643270.0290.1369.0761.27324
Summary 19.546100.0088.40101.2123830715.372100.0085.26113.1020986018.699100.0099.4777.0321875021.501100.0088.2164.95232814
Polar 2.40812.3291.3531.85317671.0356.73103.7030.15145862.13811.43172.4558.47197331.5787.3490.5835.5819716
Westerlies 7.70339.4166.9838.63959115.91338.47118.8334.25797866.58035.19105.9732.59843279.01141.9174.4931.5799014
Trades 6.41532.8250.9644.86870265.84338.0145.3331.99915156.56735.1251.7636.34908967.49634.8657.1340.9890306
Coastal 2.99415.32134.27107.17232562.55616.6389.4977.17236663.38118.08116.9099.66234673.38715.75125.41104.6723454

[33] Tabular budgets of surface PIC concentration demonstrate some striking trends. The seasonal average PIC concentration (Table 5) for all provinces was 2 μg PIC L−1 (which varied from 1.9–2.2 μg PIC L−1). Clearly, the PIC concentrations within the Trades biome were consistently the lowest, regardless of season, averaging 0.66–0.81 μg PIC L−1. Northern polar biomes (including boreal polar, Atlantic Arctic, Atlantic Subarctic, North Pacific Epicontinental) consistently showed above average PIC concentrations, at least when they were visible to the MODIS sensor between April and September. Northern polar PIC concentrations were highest between July and September 2002. The Atlantic Subarctic province, within the polar biome, showed 2–3 × the global average PIC concentration during those months, fully consistent with previous observations of mesoscale coccolithophore blooms in these areas [Balch et al., 1992, 1996a, 1996b; Brown and Yoder, 1994; Fernández et al., 1993; Holligan et al., 1993]. Southern polar biomes, when visible to MODIS/Terra between October and March, showed PIC concentrations slightly lower than the Northern Hemisphere polar biome. There are few validation measurements from this region to confirm these MODIS estimates. Coastal provinces had the greatest PIC concentrations of all the biomes, averaging between 2.5 to 3.8 μg PIC L−1, depending on the season. As might be expected, the variance in PIC concentration was greatest over space and time in the coastal provinces.

[34] Specific provinces that had above average PIC concentrations, in at least one 3 month period, were: boreal polar, Atlantic arctic, Atlantic subarctic, North Pacific epicontinental, Black Sea, Pacific subarctic gyres (east and west), northeast Atlantic Shelves, Canary coastal, Guianas coastal, northwest Atlantic shelves, southwest Atlantic shelves, Benguela Current Coastal, Persian Gulf, east India Coastal, west India Coastal, Australia west, Alaska downwelling coastal, China Sea Coastal and New Zealand Coastal. The observations of coastal shelf coccolithophore blooms are consistent with historical remote sensing observations from a number of these areas such as the northwest Atlantic shelves [Ackleson et al., 1994; Balch, 2004; Balch et al., 1991; Keller et al., 1992; Matrai and Keller, 1993; Townsend et al., 1994], the northeast Atlantic shelves [Berge, 1962; Birkenes and Braarud, 1952; Brown and Yoder, 1994; Brussaard et al., 1996; Buitenhuis et al., 1996; Burkill et al., 2002; Garcia-Soto et al., 1995; GREPMA, 1981; Head et al., 1998; Holligan et al., 1983; Malin et al., 1993; Samtleben and Bickert, 1990; Van der Wal et al., 1995; Wal et al., 1995], Australian coastal [Blackburn and Cresswell, 1993], Alaska downwelling and North Pacific Epicontinental [Kai et al., 1999; Lavrentyev et al., 2001; Napp and Hunt, 2001; Stockwell et al., 2000; Takahashi et al., 1995; Broerse et al., 2003], Black Sea [Cokacar, 2001], the southwest Atlantic shelves [Gayoso, 1995], and Pacific subarctic gyres [Fukushima and Ishizaka, 1993; Takahashi et al., 1995].

[35] Satellite-derived PIC concentrations can be checked for consistency using independent ship measurements from the same area. Note, however, this approach only allows for a gross comparison of the PIC levels since the ship and satellite observations, while measured in the same season, were made in different years. Satellite-derived values of 0.64 ± 0.34 μg PIC L−1 were observed in the Pacific equatorial divergence province during July–September 2002 (Table 5). These were lower than the average shipboard value of 2.52 (±0.58) μg PIC L−1 described by Balch and Kilpatrick [1996] for the period between August and September 1992, between 5°N and 5°S along 140°W, top 10 m only (n = 17; note one anomalously high surface value of 15 μg PIC L−1 at 2°N was not included in this average). Satellite-derived PIC values from the NW Arabian upwelling province were estimated to be 2.35 ± 1.84 μg PIC L−1 during August–September 2002 (Table 5). During July and August 1995, the average PIC measured by Balch et al. [2000] in the Arabian Sea was 2.38 ± 2.14 μg PIC L−1 [see Balch et al., 2000, Table 1], not significantly different from the satellite-derived value. During October and November 1995, average PIC concentrations measured by Balch et al. [2000] were 1.62 ± 2.03 μg PIC L−1, also not significantly different from the satellite-derived values between October and December 2002 (1.33 ± 0.93 μg PIC L−1 (Table 5)).

[36] A different view of the PIC budget can be seen in the table of PIC concentrations integrated over the euphotic zone, and aerially over each province (Table 6). The average integrated concentrations of PIC approximately mirrored the trends in surface concentration described above (but not exclusively). Aerially integrating over the various biomes, the majority of the euphotic standing stock of calcium carbonate (70–77%) occurred in the combined Westerlies and Trades biomes. This is consistent with results suggesting that low-latitude, nonbloom coccolithophores are responsible for large export ratios of CaCO3: organic matter [Sarmiento et al., 2002]. While the polar biomes had the highest concentrations of PIC, they only contributed 7–12% of the total PIC and the coastal biomes contributed ∼15%. Note, however, these polar values are likely underestimates during the fall and winter months of each hemisphere, since no satellite-derived radiance estimates are possible due to low Sun angle. Provinces that contributed over 0.5 Mt of PIC (greater than ∼2% of the global total PIC) for at least one 3 month period, were boreal polar, Atlantic subarctic, North Pacific epicontinental, Antarctic, Austral Polar, North Atlantic drift, Black Sea, Pacific subarctic gyres (east and west), North Pacific polar front, North Pacific subtropical gyre, South Pacific subtropical gyre, south subtropical convergence, subantarctic, South Atlantic gyre, Indian Monsoon gyres, Indian subtropical gyre, North Pacific tropical gyre, Pacific Equatorial divergence, West Pacific warm pool, and Archipelagic deep basins. Not a single coastal province contributed more than 0.5 Mt PIC in any 3 month period.

[37] We summarize the POC budget as well, in order to support our estimates of the PIC:POC ratios globally. As with the PIC concentrations, the polar POC estimates are likely underestimates due to the fact that remote sensing is limited during winter and fall months of either hemisphere. Average surface POC concentrations were ∼40–52 mg m−3, with highest average global POC from April to June (mean for all biomes = 52.1 mg m−3; SD = 26.05 (Table 7)). Provinces with POC concentrations >80 mg m−3 (which is ∼1.5–2 × the mean) for at least one 3 month period of the year were Atlantic subarctic, Black Sea, northeast Atlantic shelves, northwest Atlantic shelves, southwest Atlantic shelves, Persian Gulf, Alaska downwelling, and west India coastal waters. As expected, lowest POC concentrations were consistently seen in the subtropical and tropical gyre systems of the Trades and Westerlies biomes. Our results nonetheless demonstrated that the Westerlies and Trades biomes accounted for 80–82% of global POC, even greater than the percentage of global PIC that comes from those same biomes. Polar regimes accounted for only 4.7–8.5% of the global POC, less than the fraction of total global PIC that is produced in the same region (Table 7). Coastal biomes supplied 11.8–13.9% of the global POC. Interestingly, the period from January to March and October to December had the highest global concentrations of POC (699 and 682 Mt, respectively, as opposed to 631 and 647 Mt during April–June and July–September (Table 8)).

Table 7. Surface POC Aerially Integrated Within the Biogeochemical Provinces Defined by Longhurst [1998]a
Surface POCJan.–MarApr–JunJul–SepOct–Dec
ProvinceBiomeTot. POC, MtAvg. Surf. POC, mg/m3SD, mg/m3PixelsTot. POC, MtAvg. Surf. POC, mg/m3SD, mg/m3PixelsTot. POC, MtAvg. Surf. POC, mg/m3SD, mg/m3PixelsTot. POC, MtAvg. Surf. POC, mg/m3SD, mg/m3Pixels
  • a

    Provinces are grouped into their respective biome: Polar, Westerlies, Trades, and Coastal. Calculations have been made for each 3 month period of 2002. The algorithm used to derive POC concentrations was according to Morel [1988] and is described in the text. Results are given for each province for (1) total POC (essentially, the total POC found in the top meter of the water column, aerially integrated over the province), (2) average surface POC (mg m−3), (3) standard deviation of the average POC (mg m−3), and (4) the total number of pixels available for the analysis. Any values of 0.00 in the first three columns of the polar biome for each 3 month period represent regions where no satellite radiance data were available, mainly due to low Sun angle. At the bottom of the table global statistics are provided as well as statistics for each of the four biomes.

Boreal Polarpolar0.00000.000.0000.115078.3913.7927860.270564.4613.0473600.00000.000.000
Atl Arcticpolar0.000030.520.0010.221070.599.9449970.191460.778.0149970.00000.000.000
Atl Subarcticpolar0.00000.000.0000.175087.757.7832270.136068.167.6932280.00000.000.000
N Pac Epicontinentalpolar0.046754.6310.317840.336192.3116.6340310.258765.418.3744260.038156.286.15608
Antarcticpolar0.781438.416.00268160.00000.000.0000.00000.000.0000.580739.674.5318713
Austral Polarpolar0.223857.1210.5563700.00000.000.0000.00000.000.0000.019440.974.02719
N Atl Drift(WWDR)westerlies0.073034.955.1220040.235366.157.8936120.192154.036.7436090.050638.242.151239
Gulf Streamwesterlies0.055648.7810.119770.062653.8610.839960.035930.954.449960.045038.875.28994
N Atl Subtrop Gyre (W)westerlies0.171128.635.6146220.157226.308.9846220.113719.022.1546220.142623.853.804622
Mediterranean Seawesterlies0.109542.057.7321300.085832.947.6521310.068926.454.8121300.087533.716.552124
Black Seawesterlies0.033986.2114.543510.033183.6316.013530.028271.606.713510.027671.598.59343
N Atl Subtrop Gyre (E)westerlies0.149933.795.4835590.200145.1116.0835590.115926.125.2635590.126328.474.303559
Pac Subarctic Gyres (E)westerlies0.060738.892.5815180.185056.909.7333710.167452.815.7732930.030745.503.07638
Pac Subarctic Gyres (W)westerlies0.084641.613.5919420.139964.4811.2321050.132660.018.1421390.062948.015.211218
Kuroshio Currentwesterlies0.136042.3312.5125040.172253.3619.9325170.110634.419.5325060.121537.989.372493
N Pac Polar Frontwesterlies0.298640.804.8762140.313943.828.4760800.230131.675.5461690.267836.634.556207
N Pac Subtrop Gyrewesterlies0.291521.116.9498390.285120.656.3598390.245017.751.6798390.272519.743.709839
Tasman Seawesterlies0.052931.315.5913880.078646.694.5713830.076745.504.4813840.082148.574.641388
S Pac Subtrop Gyrewesterlies0.642217.846.88259070.851823.677.29258900.917225.586.95258090.769521.527.7325743
S Subtrop Convergencewesterlies0.627836.4111.53147920.624542.787.14123630.622937.205.02143280.710441.217.5614790
Subantarcticwesterlies1.227139.057.87315540.156843.905.5031620.312234.623.7782321.340442.446.3331734
N Atl Trop Gyretrades0.190423.609.5655770.213726.5515.7155620.192323.929.7755570.194724.138.275578
W Trop Atltrades0.146927.265.1635390.187434.797.3535370.187934.877.9635390.164530.524.453539
East Trop Atltrades0.163232.3411.7233090.196538.1510.9133790.215742.939.6732950.179734.876.773380
S Atl Gyretrades0.358119.876.63131460.513428.497.84131460.567531.867.82130050.496527.8111.0113032
Caribbeantrades0.145635.0017.0728910.132931.9815.7028870.133832.1814.3928890.143834.5813.862891
Indian Monsoon Gyrestrades0.419226.836.56103010.463129.655.47102980.549235.178.37102950.465129.775.0010298
Indian S Subtrop Gyretrades0.305117.733.47124070.464727.005.34124070.519430.184.93124070.387622.524.7912407
N Pac Trop Gyretrades0.331122.145.11106220.322121.694.81105500.305320.472.55105940.318121.272.7910625
N Pac Eq Countercurrenttrades0.297933.1610.2059480.318235.4213.7559480.264329.487.7659340.250927.937.965947
Pac Eq Divergencetrades0.456031.484.8494920.515635.595.1794920.496734.665.3793870.441430.625.509444
W Pac Warm Pooltrades0.248319.965.5281880.273221.955.4581910.268621.584.4881910.262421.083.918191
Archipelagic Deep Basinstrades0.250927.0715.4463480.289631.2511.6463470.327635.3910.0763400.267428.8211.216356
NE Atl Shelvescoastal0.008673.7711.681150.0848130.2718.507950.057888.3717.508000.001248.449.1723
Canary Coastal (EACB)coastal0.050976.4432.915020.063495.4052.605010.035154.2223.234900.035453.3220.43501
Guinea Current Coastalcoastal0.046257.4322.725310.060560.8721.356550.036051.2112.924620.050351.4015.47645
Guianas Coastalcoastal0.074459.5963.038230.093373.7171.018340.079562.9261.178330.073257.4558.78840
NW Atl Shelvescoastal0.108964.9220.5214690.175482.8322.7319080.122657.8116.8719110.084956.7717.101289
Brazil Current Coastalcoastal0.038738.7123.117300.042542.6115.507270.043143.2917.657260.042042.0825.12728
SW Atl Shelvescoastal0.120892.4228.1713150.010475.0712.321220.023862.7213.013470.113787.0525.111315
Benguela Current Coastalcoastal0.077872.0548.787830.084778.3931.237830.069566.0318.157640.077872.3225.58780
E Africa Coastalcoastal0.122934.7721.0625680.161845.8116.2925670.157044.4311.9125680.132437.4514.192569
Red Seacoastal0.024450.3221.263390.020541.8416.353430.018838.7314.043400.023748.5116.63342
Persian Gulfcoastal0.0171100.2112.561250.014786.259.331250.011164.497.841260.013175.948.43126
NW Arabian Upwellingcoastal0.219663.3720.1723370.157645.5921.6623310.207965.3119.7821410.180852.1614.982337
E India Coastalcoastal0.057158.9149.506620.054157.6344.016400.048254.7029.396010.056257.5747.16666
W India Coastalcoastal0.066378.5452.775750.059771.6051.305670.056382.5538.514600.059069.8540.80575
Australia Westcoastal0.094234.0414.2619780.129546.8013.4919760.128146.2311.4019790.101936.8911.591974
Alaska Downwelling Coastalcoastal0.000654.016.22110.0572117.3513.995690.037878.916.895590.00000.000.000
California Upwelling Coastalcoastal0.126647.1319.6221440.147757.4030.7820700.122545.1619.7821720.104042.3913.481914
Centr American Coastalcoastal0.078467.0827.847940.074964.5526.847890.052845.9018.627820.058450.2719.41790
Chilecoastal0.148660.8028.6918910.143769.3326.5215220.110056.6215.4814690.144259.6619.911874
China Sea Coastalcoastal0.055379.1329.155180.044663.7324.005190.035250.2713.555200.041859.9220.07517
Sundacoastal0.160641.8017.6925310.160041.2714.9125540.159641.3813.0625410.149838.7112.852549
East Australian Coastalcoastal0.026025.147.097540.035534.327.527530.038036.626.657540.030529.419.50755
New Zealand Coastalcoastal0.056346.679.5111120.025355.276.083960.036144.695.467210.063252.489.931110
Unclassified 0.013630.6124.313600.015437.7437.863180.014934.7028.703370.012329.3115.57327
Summary 10.173043.6621.5626000710.236352.1326.052131329.955844.1518.762248139.997739.5918.56243205
Polar 1.05250.058.95339700.84782.2612.03150410.85764.709.28200110.63845.644.9020040
Westerlies 4.01438.927.401093013.58246.959.84819833.36937.855.40889664.13738.425.52106931
Trades 3.31326.378.44917683.89030.219.10917444.02831.067.76914333.57227.837.1391688
Coastal 1.78059.8825.58246071.90266.8724.71240461.68755.7617.95240661.63851.3119.8124219
Table 8. Euphotic Zone POC Aerially Integrated Within the Biogeochemical Provinces Defined by Longhurst [1998]a
Int. POC Over Euphotic ZoneJan–MarApr–JunJul–SepOct–Dec
ProvinceBiomeInt. POC, Mt% TotalAvg. Int. POC, mg/m2SD, mg/m2PixelsInt. POC, Mt% TotalAvg. Int. POC, mg/m2SD, mg/m2PixelsInt. POC, Mt% TotalAvg. Int. POC, mg/m2SD, mg/m2PixelsInt. POC, Mt% TotalAvg. Int. POC, mg/m2SD, mg/m2Pixels
  • a

    Provinces are grouped into their respective biome: Polar, Westerlies, Trades, and Coastal. Calculations have been made for each 3 month period of 2002. Results are given for each province for (1) total euphotic zone POC (essentially, the total POC found above the 1% light depth of the water column, aerially integrated over the province), (2) average integrated POC (mg m−2), (3) standard deviation of the average POC (mg m−2), and (4) the total number of pixels available for the analysis. Any values of 0.00 in the first three columns of the polar biome for each 3 month period represent regions where no satellite radiance data were available, mainly due to low Sun angle. At the bottom of the table the global statistics are provided as well as statistics for each of the four biomes.

Boreal Polarpolar0.0000.000.000.0004.1580.662834.36362.72278611.2691.742685.57318.2973600.0000.000.000.000
Atl Arcticpolar0.0020.002200.790.0018.6611.372766.42318.6549978.3791.292660.96309.5749970.0000.000.000.000
Atl Subarcticpolar0.0000.000.000.0005.8670.932941.31207.4532275.4750.852743.88219.9832280.0000.000.000.000
N Pac Epicontinentalpolar2.2020.312575.12195.7178410.8121.712969.20264.13403110.7051.652706.74187.2644261.7640.262604.19112.39608
Antarcticpolar47.3526.772327.40141.38268160.0000.000.000.0000.0000.000.000.00034.5525.072360.06140.4618713
Austral Polarpolar10.0751.442571.39146.1463700.0000.000.000.0000.0000.000.000.0001.1250.172375.89151.20719
N Atl Drift(WWDR)westerlies4.7540.682277.60110.4220049.6871.532722.87162.4136129.1221.412566.18107.9036093.1010.452341.6242.361239
Gulf Streamwesterlies2.8360.412487.58133.559772.9670.472555.10117.319962.5550.392200.4464.429962.7180.402345.2464.92994
N Atl Subtrop Gyre (W)westerlies12.8241.832145.6787.93462212.4101.972076.44126.68462211.4991.781924.0432.44462212.2181.792044.4144.984622
Mediterranean Seawesterlies6.2260.892391.3792.3621305.8060.922228.89104.6821315.4750.852103.0480.7021305.8390.862248.7198.742124
Black Seawesterlies1.1520.162927.53130.013511.1480.182899.48151.303531.0970.172786.8579.513511.0720.162784.57103.67343
N Atl Subtrop Gyre (E)westerlies9.9911.432251.75104.81355910.6201.682393.36152.5335599.2791.432091.2083.5135599.5231.402146.3265.903559
Pac Subarctic Gyres (E)westerlies3.6710.522351.8958.6815188.4381.342594.7896.4233718.0971.252553.92105.8732931.6600.242457.3458.64638
Pac Subarctic Gyres (W)westerlies4.8710.702395.8391.5719425.8490.932696.34152.1221055.8460.902645.96123.3821393.2650.482492.0091.941218
Kuroshio Currentwesterlies7.6001.092365.41157.3725048.0461.272492.88179.0325177.2131.112243.09120.0025067.3811.082306.48109.822493
N Pac Polar Frontwesterlies17.4272.492381.34114.89621417.3212.742418.31136.65608016.0572.482209.5675.06616916.8702.472307.6182.796207
N Pac Subtrop Gyrewesterlies27.0113.861956.54117.82983926.9324.271950.7892.49983926.0834.031889.3355.73983926.7643.921938.6453.979839
Tasman Seawesterlies3.7250.532204.6988.3513884.1630.662472.7161.3013834.1390.642456.9495.7513844.2230.622499.8860.371388
S Pac Subtrop Gyrewesterlies67.1869.611866.10243.412590773.06011.582030.29202.652589074.45711.502076.38187.992580970.35810.321967.25224.9425743
S Subtrop Convergencewesterlies39.1575.602271.23176.101479235.1025.562404.58109.121236338.8276.002318.79111.651432840.9876.012377.68130.3514790
Subantarcticwesterlies73.55310.522340.87175.85315548.6721.372428.4688.73316220.5203.172275.3185.27823275.86111.132402.09155.4431734
N Atl Trop Gyretrades16.3252.332023.40200.95557716.6372.642067.60277.55556216.3312.522031.45202.93555716.4882.422043.22179.405578
W Trop Atltrades11.4631.642126.8591.26353912.2531.942274.75111.82353712.2531.892273.42136.64353911.8371.742196.3184.313539
East Trop Atltrades11.1651.602213.09197.42330911.9831.902325.84165.61337912.0991.872407.98153.17329511.7281.722275.68120.603380
S Atl Gyretrades34.7284.971926.99149.061314638.4726.102134.73136.811314639.3096.072206.85150.141300537.5035.502100.67199.9813032
Caribbeantrades9.3431.342246.15214.8128919.1131.442193.64224.3328879.1541.412202.00216.7028899.3601.372250.26186.952891
Indian Monsoon Gyrestrades32.9924.722111.54141.061030133.9975.392176.50103.041029835.5585.492277.07142.641029534.0464.992179.67102.8610298
Indian S Subtrop Gyretrades32.4114.631883.55123.461240736.3995.772115.35116.301240737.6155.812185.97122.261240734.5905.072010.21121.0712407
N Pac Trop Gyretrades29.8984.271999.56117.781062229.5754.691991.43140.031055029.3214.531966.22106.701059429.7114.361986.46101.0110625
N Pac Eq Countercurrenttrades20.0632.872233.18174.36594820.4053.232271.23193.02594819.4063.002165.19154.10593419.1342.812130.06158.905947
Pac Eq Divergencetrades32.0874.592215.1687.95949233.2095.262292.6686.89949232.5905.032274.51100.38938731.6394.642195.06113.879444
W Pac Warm Pooltrades24.1513.451941.19155.11818824.8473.941996.41140.49819124.7673.831989.97121.80819124.6343.611979.25106.558191
Archipelagic Deep Basinstrades19.3792.772090.99270.60634820.3553.232196.63190.29634721.0843.262277.89176.77634019.9072.922145.13206.436356
NE Atl Shelvescoastal0.3270.052796.14129.121152.1280.343270.12216.097951.9070.292916.59209.708000.0600.012484.89112.2923
Canary Coastal (EACB)coastal1.8590.272793.26431.665021.9410.312922.04594.615011.6430.252534.72355.804901.6820.252532.84339.31501
Guinea Current Coastalcoastal2.0800.302586.17268.465312.6200.422636.38239.516551.7750.272525.54170.324622.4680.362522.80191.04645
Guianas Coastalcoastal3.0760.442463.86572.998233.3280.532630.68591.458343.2050.502536.05527.538333.1400.462464.00524.20840
NW Atl Shelvescoastal4.4850.642673.03254.0114696.0260.952845.67213.0219085.4700.842580.15207.9219113.8550.572577.25212.701289
Brazil Current Coastalcoastal2.2650.322262.93228.127302.3580.372365.56183.627272.3580.362368.91231.987262.3000.342304.64245.52728
SW Atl Shelvescoastal3.8440.552941.61316.8713150.3890.062815.05143.361221.0130.162668.73181.733473.7890.562900.22316.481315
Benguela Current Coastalcoastal2.8740.412661.00506.247833.0370.482811.76340.017832.8490.442707.85246.107642.9710.442762.49288.90780
E Africa Coastalcoastal7.8551.122222.48277.0525688.5961.362433.30206.6525678.5711.322425.24178.8225688.1121.192294.32188.212569
Red Seacoastal1.2070.172492.32292.633391.1610.182370.48272.693431.1280.172324.09252.563401.2100.182477.71258.50342
Persian Gulfcoastal0.5230.073058.40106.921250.5020.082934.3485.931250.4660.072705.6697.061260.4880.072832.2585.84126
NW Arabian Upwellingcoastal9.2431.322666.56193.0423378.3741.332422.13242.0623318.5791.332695.33197.4221418.7881.292535.68152.042337
E India Coastalcoastal2.4480.352524.15416.266622.3670.382522.44393.756402.2310.342531.68289.606012.4560.362517.41395.78666
W India Coastalcoastal2.3170.332742.57432.235752.2340.352680.41434.885671.9380.302844.00351.564602.2750.332693.67346.93575
Australia Westcoastal6.1870.882235.10288.3519786.8001.082457.30248.0919766.8081.052457.28241.6219796.3550.932299.92257.841974
Alaska Downwelling Coastalcoastal0.0270.002572.5582.46111.5540.253187.39151.605691.3720.212860.84121.895590.0000.000.000.000
California Upwelling Coastalcoastal6.5160.932426.11274.2221446.5001.032526.50336.6020706.4891.002392.06256.1521725.8140.852369.83176.901914
Centr American Coastalcoastal3.1450.452692.80267.217943.0960.492667.76261.057892.8030.432437.73216.387822.9030.432498.29220.68790
Chilecoastal6.3720.912607.76440.9118915.6650.902733.11395.9015225.0450.782597.59329.6814696.3380.932621.48372.011874
China Sea Coastalcoastal1.9590.282805.53249.295181.8450.292637.65221.055191.7550.272502.63176.495201.8210.272611.67218.59517
Sundacoastal9.1011.302368.44227.2725319.1911.462370.54193.1825549.1671.422376.30174.2425419.0201.322331.11173.002549
East Australian Coastalcoastal2.1360.312061.66116.237542.3320.372253.8495.917532.3840.372299.9874.527542.2250.332144.41110.25755
New Zealand Coastalcoastal2.9670.422458.43177.4911121.1850.192591.1286.563961.9730.302443.12124.817213.0680.452547.11201.791110
Unclassified 0.9480.142127.81358.393600.9120.142234.12437.313180.9420.152198.27393.833370.8930.132128.84279.66327
Summary 699.380100.002290.22522.24260007631.105100.002407.60555.00213132647.453100.002321.53511.77224813681.892100.002192.01648.06243205
Polar 59.6308.532491.30161.073397029.4974.672877.82288.241504135.8275.532699.29258.772001137.4425.492446.71134.6820040
Westerlies 281.98340.322307.69125.54109301230.22236.482424.35128.9081983240.26737.112289.4093.9588966281.84041.332310.6692.59106931
Trades 274.00639.182084.30160.3291768287.24645.512169.73157.1891744289.48844.712188.21148.6991433280.57741.152124.33140.1691688
Coastal 82.81111.842570.12284.742460783.22913.192655.89267.292404680.93012.502553.57226.692406681.14011.902405.39234.3024219

[38] The PIC:POC ratio was estimated from the data shown in Tables 5 and 7 (note, because of how the PIC calculations were made, the ratios are identical whether they are based on surface or euphotic integrated estimates; thus, results are only shown in Table 5). The global average PIC:POC ratio varied from 0.041 (April–June) to 0.051 (July–September). Such values are highly consistent with average global export ratios estimated by Sarmiento et al. [2002] based on a biogeochemical transport box model of the top 100 m, using measurements of the vertical gradients of potential alkalinity and nitrate. The reader should note that our PIC:POC estimates would not be expected to match PIC:POC export ratios at 100 m. This is mainly because there can be mineralization of both components between the surface optical depth (the top few meters visible to MODIS/Terra [Gordon and McCluney, 1975]) and the 100 m depth horizon used by Sarmiento et al. [2002]. Nonetheless, their area-weighted global mean export ratio (0.056 ± 0.004; based on ocean chemistry) was strikingly close to our average global value of 0.047 ± 0.004 for all seasons, based on MODIS water-leaving radiance measurements. Note, while these PIC:POC and export ratios are consistent, they are both considerably less than the often-used value of 0.25 (see Sarmiento et al. [2002] for further discussion about this discrepancy).

[39] Sarmiento et al. [2002] further found their predicted export ratios to be everywhere less than 0.1. This was not true in our case; the highest PIC:POC ratios that we observed (exceeding 0.1 (see Table 5)) were observed in the Black Sea, NE Atlantic shelves, SW Atlantic shelves, Persian Gulf and China Sea Coastal provinces. All Northern Hemisphere polar provinces had PIC:POC ratios >0.05 from July to September. Other provinces with PIC:POC ratios >0.05 for at least one 3 month period were North Atlantic Drift, Black Sea, northeast Atlantic Shelves, Guianas coastal, northwest Atlantic shelves, southwest Atlantic shelves, Benguela Current coastal, Persian Gulf, east India Coastal, west India coastal, Australia West, China Sea coastal, and New Zealand coastal provinces. Regions where our PIC:POC ratios were less than the export ratios of Sarmiento et al. [2002] were in the Equatorial regions of the Pacific, Atlantic and Indian Oceans. Moreover, we found higher values of PIC:POC in the North Atlantic subarctic gyre (summer values of 0.093 (Table 5)), compared to export ratios of 0.023 ± 0.02 [Sarmiento et al., 2002]. The reason for this disparity is not immediately obvious, unless calcite dissolution in the top 100 m is responsible for remineralizing a larger proportion of the PIC than POC [Milliman et al., 1999]. Given the frequent blooms in the North Atlantic subarctic gyre, higher PIC concentrations (Tables 5 and 6) and higher PIC:POC ratios clearly would be expected.

[40] Another cross check of the above PIC estimates is to estimate the rate of global PIC turnover in the sea by dividing the average global, MODIS-derived PIC value (18.8 Mt PIC ± 2.56) by the typical annual global calcification rate of ∼1 Gt PIC yr−1 [Archer and Maier-Reimer, 1994; Archer, 1997; Milliman et al., 1999; Morse and Mackenzie, 1990; Wollast, 1994] which gives a quotient of 0.0188 years (= 6.86 days). This value is reasonably consistent with the average turnover time of PIC particles based on direct shipboard measurements of 14C calcification, combined with atomic absorption analyses of PIC standing stock. For example, in the equatorial Pacific, the PIC turnover time typically was ∼6.5 days (±3.5 days) [Balch and Kilpatrick, 1996] while in the Arabian Sea, the mean turnover time was ∼13 days (±6 days), depending on season [Balch et al., 2000]. While these observations reveal internal consistencies between PIC standing stock and field measurements of PIC turnover, one must remember that the turnover time of PIC particles can be highly variable due to factors such as growth and grazing. Moreover, there are few field observations of PIC turnover available to constrain global models of the calcium carbonate cycle. Verification of the global PIC estimates must await further in situ validation measurements from ship. Specifically, the large PIC-rich features observed by the MODIS sensors in the southern ocean, must be verified, as they have a large effect on the total global PIC standing stock.

Acknowledgments

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Methods
  5. 3. Results
  6. 4. Discussion
  7. Acknowledgments
  8. References
  9. Supporting Information

[41] Katherine Kilpatrick (University of Miami) helped with algorithm implementation and acquisition of PIC data from several cruises. Amanda Ashe and Jennifer Fritz also provided assistance with some of the validation activities. We gratefully acknowledge several sources of funding over the years. Primary support was generously provided by NASA (NAS5-31363). Support for algorithm development and validation activities also was provided by NASA (NAGW2426; NAS5-97268; NAG5-10622; NASA EPSCOR EP-02-14), the Office of Naval Research (ONR N00014-91-J-1048, N00014-97-1-0034; N00014-98-1-0882; N00014-99-1-0645; N00014-01-1-0042), the National Science Foundation (OCE-9022227; OCE-9596167; OCE-0136541), and NOAA (NA56RM0258; 40-AA-NE-005996).

References

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Methods
  5. 3. Results
  6. 4. Discussion
  7. Acknowledgments
  8. References
  9. Supporting Information
  • Ackleson, S., W. M. Balch, and P. M. Holligan (1988), White waters of the Gulf of Maine, Oceanography, 1, 1822.
  • Ackleson, S., W. M. Balch, and P. M. Holligan (1994), Response of water-leaving radiance to particulate calcite and chlorophyll a concentrations: A model for Gulf of Maine coccolithophore blooms, J. Geophys. Res., 99, 74837499.
  • Archer, D. E. (1997), A data-driven model of the global calcite lysocline, Global Biogeochem. Cycles, 10, 511526.
  • Archer, D., and E. Maier-Reimer (1994), Effect of deep-sea sedimentary calcite on preservation on atmospheric CO2 concentration, Nature, 367, 260263.
  • Archer, D., H. Kheshgi, and E. Maier-Reimer (1998), Dynamics of fossil fuel CO2 neutralization by marine CaCO3, Global Biogeochem. Cycles, 12, 259276.
  • Armstrong, R. A., C. Lee, J. I. Hedges, S. Honjo, and S. G. Wakeham (2002), A new, mechanistic model for organic carbon fluxes in the ocean based on the quantitative association of POC with ballast minerals, Deep Sea Res., Part II, 49, 219236.
  • Balch, W. (1991), Erratum, Limnol. Oceanogr, 36, 1462.
  • Balch, W. (2001), Validation of surface bio-optical properties in the Gulf of Maine as a means for improving satellite primary production estimates, in SIMBIOS Project 2000 Annual Report, edited by G. S. Fargion, and C. R. McClain, pp. 2633, NASA Goddard Space Flight Cent., Greenbelt, Md.
  • Balch, W. M. (2004), Re-evaluation of the physiological ecology of coccolithophores, in Coccolithophores: From Molecular Processes to Global Impact, edited by H. R. Thierstein, and J. R. Young, pp. 165190, Springer, New York.
  • Balch, W. M., and D. T. Drapeau (2004), Backscattering by coccolithophorids and coccoliths: Sample preparation, measurement and analysis protocols, in Ocean Optics Protocols For Satellite Ocean Color Sensor Validation, Revision 5: Biogeochemical and Bio-Optical Measurements and Data Analysis Protocols, edited by J. L. Mueller, G. S. Fargion, and C. R. McClain, pp. 2736, NASA Goddard Space Flight Space Cent., Greenbelt, Md.
  • Balch, W. M., and K. A. Kilpatrick (1996), Calcification rates in the equatorial Pacific along 140°W, Deep Sea Res., Part II, 43, 971993.
  • Balch, W. M., R. W. Eppley, M. R. Abbott, and F. M. H. Reid (1989), Bias in satellite-derived pigment measurements due to coccolithophores and dinoflagellates, J. Plankton Res., 11, 575581.
  • Balch, W. M., P. M. Holligan, S. G. Ackleson, and K. J. Voss (1991), Biological and optical properties of mesoscale coccolithophore blooms in the Gulf of Maine, Limnol. Oceanogr., 36, 629643.
  • Balch, W. M., P. M. Holligan, and K. A. Kilpatrick (1992), Calcification, photosynthesis and growth of the bloom-forming coccolithophore, Emiliania huxleyi, Cont. Shelf Res., 12, 13531374.
  • Balch, W. M., K. Kilpatrick, P. M. Holligan, D. Harbour, and E. Fernandez (1996a), The 1991 coccolithophore bloom in the central north Atlantic. II, Relating optics to coccolith concentration, Limnol. Oceanogr., 41, 16841696.
  • Balch, W. M., K. A. Kilpatrick, P. M. Holligan, and C. Trees (1996b), The 1991 coccolithophore bloom in the central north Atlantic. I. Optical properties and factors affecting their distribution, Limnol. Oceanogr., 41, 16691683.
  • Balch, W. M., D. T. Drapeau, T. L. Cucci, R. D. Vaillancourt, K. A. Kilpatrick, and J. J. Fritz (1999), Optical backscattering by calcifying algae: Separating the contribution by particulate inorganic and organic carbon fractions, J. Geophys. Res., 104, 15411558.
  • Balch, W. M., D. Drapeau, and J. Fritz (2000), Monsoonal forcing of calcification in the Arabian Sea, Deep Sea Res., Part II, 47, 13011337.
  • Balch, W. M., D. Drapeau, J. Fritz, B. Bowler, and J. Nolan (2001), Optical backscattering in the Arabian Sea: Continuous underway measurements of particulate inorganic and organic carbon, Deep Sea Res., Part I, 48, 24232452.
  • Balch, W. M., D. T. Drapeau, B. C. Bowler, E. S. Booth, J. I. Goes, A. Ashe, and J. M. Frye (2004), A multi-year record of hydrographic and bio-optical properties in the Gulf of Maine: I. Spatial and temporal variability, Prog. Oceanogr., 63, 5798.
  • Bates, N. R., A. F. Michaels, and A. H. Knapp (1996), Alkalinity changes in the Sargasso Sea: Geochemical evidence of calcification? Mar. Chem., 51, 347358.
  • Berge, G. (1962), Discoloration of the sea due to Coccolithus huxleyi “bloom,”, Sarsia, 6, 2740.
  • Birkenes, E., and T. Braarud (1952), Phytoplankton in the Oslo fjord during a “Coccolithus huxleyi summer,”, Oslo I. Mat. Naturv. Kl., 2, 123.
  • Blackburn, S., and G. Cresswell (1993), A coccolithophorid bloom in Jervis Bay, Australia, Aust. J. Freshwater Res., 44, 253260.
  • Boyle, E. A. (1988), The role of vertical chemical fractionation in controlling late Quaternary atmospheric carbon dioxide, J. Geophys. Res., 93, 15,70115,714.
  • Broecker, W. S., and T.-H. Peng (1982), Tracers in the Sea, 660 pp., Lamont-Doherty Geol. Obs. Columbia Univ., Palisades, New York.
  • Broecker, W. S., and T.-H. Peng (1989), The cause of the glacial to interglacial atmospheric CO2 change: A polar alkalinity hypothesis, Global Biogeochem. Cycles, 3, 215239.
  • Broecker, W. S., and T. Takahashi (1977), Neutralization of fossil fuel CO2 by marine calcium carbonate, in The Fate of Fossil CO2in the Oceans, edited by N. R. Anderson, and A. Malahoff, pp. 213241, Springer, New York.
  • Broerse, A. T. C., T. Tyrrell, J. R. Young, A. J. Poulton, A. Merico, and W. M. Balch (2003), The cause of bright waters in the Bering Sea in winter, Cont. Shelf Res., 23, 15791596.
  • Brown, C. W., and J. A. Yoder (1994), Coccolithophorid blooms in the global ocean, J. Geophys. Res., 99, 74677482.
  • Brussaard, C. P. D., R. S. Kempers, A. J. Kop, R. Riegmen, and M. Heldal (1996), Virus-like particles in a summer bloom of Emiliania huxleyi in the North Sea, Aquat. Microbial Ecol., 10, 105113.
  • Brzezinski, M. A., D. M. Nelson, V. M. Franck, and D. E. Sigmon (2001), Silicon dynamics within an intense open-ocean diatom bloom in the Pacific sector of the Southern Ocean, Deep Sea Res., Part II, 48, 39974018.
  • Buitenhuis, E., J. van Bleijswijk, D. Bakker, and M. Veldhuis (1996), Trends in inorganic and organic carbon in a bloom of Emiliania huxleyi in the North Sea, Mar. Ecol. Prog. Ser., 143, 271282.
  • Burkill, P. H., S. D. Archer, C. Robinson, P. D. Nightingale, S. B. Groom, G. A. Tarran, and M. V. Zubkov (2002), Dimethyl sulphide biogeochemistry within a coccolithophore bloom (DISCO): An overview, Deep Sea Res., Part II, 49, 28632885.
  • Charlson, R. J., J. E. Lovelock, M. O. Andreae, and S. G. Warren (1987), Oceanic phytoplankton, atmospheric sulphur, cloud albedo and climate, Nature, 326, 655661.
  • Clark, D. K. (1981), Phytoplankton algorithms for the Nimbus-7 CZCS, in Oceanography From Space, edited by J. R. F. Gower, pp. 227237, Springer, New York.
  • Cokacar, E. A. (2001), Structure of Emiliania huxleyi blooms in the Black Sea surface waters as detected by SeaWIFS imagery, Geophys. Res. Lett., 28, 46074610.
  • Cullen, J. J. (1990), On models of growth and photosynthesis in phytoplankton, Deep Sea Res., Part A, 37, 667683.
  • Eppley, R. W., W. G. Harrison, S. Chisholm, and E. Stewart (1977), Particulate organic matter in surface waters off southern California and its relationship to phytoplankton, J. Mar. Res., 35, 671696.
  • Fagerbakke, K. M., M. Heldal, S. Norland, B. R. Heimdal, and H. Batvik (1994), Emiliania huxleyi: Chemical composition and size of coccoliths from enclosure experiments and a Norwegian fjord, Sarsia, 79, 349355.
  • Feely, R. A., C. L. Sabine, K. Lee, W. Berelson, J. Kleypas, V. J. Fabry, and F. J. Millero (2004), Impact of anthropogenic CO2 on the CaCO3 system in the oceans, Science, 305, 362366.
  • Fernández, E., P. Boyd, P. M. Holligan, and D. S. Harbour (1993), Production of organic and inorganic carbon within a large scale coccolithophore bloom in the northeast Atlantic Ocean, Mar. Ecol. Prog. Ser., 97, 271285.
  • Francois, R., S. Honjo, R. Krishfield, and S. Manganini (2002), Factors controlling the flux of organic carbon to the bathypelagic zone of the ocean, Global Biogeochem. Cycles, 16(4), 1087, doi:10.1029/2001GB001722.
  • Fukushima, H., and J. Ishizaka (1993), Special features and applications of CZCS data in Asian waters, in Ocean Colour: Theory and Applications in a Decade of CZCS Experience, edited by P. Schlittenhardt, and V. Barale, pp. 213236, Elsevier, New York.
  • Garcia-Soto, C., E. Fernandez, R. D. Pinigree, and D. S. Harbour (1995), Evolution and structure of a shelf coccolithophore bloom in the western English Channel, J. Plankton Res., 17, 20112036.
  • Gayoso, A. M. (1995), Bloom of Emiliania huxleyi (Prymnesiophyceae) in the western South Atlantic Ocean, J. Plankton Res., 17, 17171722.
  • Geider, R. J. (1987), Light and temperature dependence of the carbon to chlorophyll a ratio in microalgae and cyanobacteria: Implications for physiology and growth of phytoplankton, New Phytol., 106, 134.
  • Gordon, H. R. (1997), Atmospheric correction of ocean color imagery in the Earth Observing System era, J. Geophys. Res., 102, 17,08117,106.
  • Gordon, H. R., and T. Du (2001), Light scattering by nonspherical particles: Application to coccoliths detached from Emiliania huxleyi, Limnol. Oceanogr., 46, 14381454.
  • Gordon, H. R., and W. R. McCluney (1975), Estimation of the depth of sunlight penetration in the sea for remote sensing, Appl. Opt., 14, 413416.
  • Gordon, H. R., and A. Y. Morel (1983), Remote Assessment of Ocean Color for Interpretation of Satellite Visible Imagery: A Review, 114 pp., Springer, New York.
  • Gordon, H. R., and M. Wang (1994), Retrieval of water-leaving radiance and aerosol optical thickness over the oceans with SeaWiFS: A preliminary algorithm, Appl. Opt., 33, 443452.
  • Gordon, H. R., O. B. Brown, R. H. Evans, J. W. Brown, R. C. Smith, K. S. Baker, and D. K. Clark (1988), A semi-analytic radiance model of ocean color, J. Geophys. Res., 93, 10,90910,924.
  • GREPMA (1981), Satellite (AVHRR/NOAA-9) and ship studies of a coccolithophorid bloom in the western English Channel, Mar. Nature, 1, 114.
  • Head, R. N., D. W. Crawford, J. Egge, D. Lesley, S. Kristiansen, E. Maranon, D. Pond, D. A. Purdie, and R. P. Harris (1998), The hydrography and biology of a bloom of the coccolithophorid Emiliania huxleyi in the northern North Sea, J. Sea Res., 39, 255266.
  • Holligan, P. M., M. Viollier, D. S. Harbout, P. Camus, and M. Champagne-Philippe (1983), Satellite and ship studies of coccolithophore production along a continental shelf edge, Nature, 304, 339342.
  • Holligan, P. M., et al. (1993), A biogeochemical study of the coccolithophore, Emiliania huxleyi, in the North Atlantic, Global Biogeochem. Cycles, 7, 879900.
  • Iglesias-Rodrigues, D., R. Armstrong, R. Feely, R. Hood, and J. Kleypas (2001), SMP Workshop assesses role of calcification in ocean carbon cycle, U.S. JGOFS Newsl., 11(3), 910.
  • Kai, M., T. Hara, H. Aoyama, and N. Kuroda (1999), A massive coccolithophorid bloom observed in Mikawa Bay, Japan, J. Oceanogr., 55, 395406.
  • Keller, M. D., W. K. Bellows, and R. R. L. Guillard (1989), Dimethyl sulfide production in marine phytoplankton, in Biogenic Sulfur in the Environment, edited by E. S. Saltzman, and W. J. Cooper, pp. 167182, Am. Chem. Soc., Washington, D. C.
  • Keller, M. D., D. W. Townsend, S. G. Ackleson, P. M. Holligan, W. M. Balch, and P. A. Matrai (1992), Observations of coccolithophore blooms in the Gulf of Maine, in Gulf of Maine Scientific Workshop, edited by J. Wiggin, and C. Mooers, pp. 259263, Univ. of Mass. Urban Harbors Inst., Boston.
  • Lavrentyev, P. J., D. A. Stockwell, and T. E. Whitledge (2001), The distribution of microbial plankton in the Bering Sea during the anomalous coccolithophore bloom in summer 1998 (abstract), in 2001 Aquatic Sciences Meeting, edited by J. D. Ackerman, and S. Twombly, p. 86, Am. Soc. of Limnol. and Oceanogr., Albuquerque, N. M.
  • Longhurst, A. R. (1998), Ecological Geography of the Sea, Elsevier, New York.
  • Maier-Reimer, E. (1996), Dynamic vs. apparent Redfield radio in the oceans: A case for 3D models, J. Mar. Syst., 9, 113120.
  • Malin, G., S. Turner, P. Liss, P. Holligan, and D. Harbour (1993), Dimethylsulphide and dimethylsulphoniopropianate in the northeast Atlantic during the summer coccolithophore bloom, Deep Sea Res., Part I, 40, 14871508.
  • Matrai, P. A., and M. D. Keller (1993), Dimethylsulfide in a large-scale coccolithophore bloom in the Gulf of Maine, Cont. Shelf Res., 13, 831843.
  • Milliman, J. D. (1993), Production and accumulation of calcium carbonate in the ocean: Budget of a nonsteady state, Global Biogeochem. Cycles, 7, 927957.
  • Milliman, J., P. J. Troy, W. Balch, A. K. Adams, Y.-H. Li, and F. T. MacKenzie (1999), Biologically-mediated dissolution of calcium carbonate above the chemical lysocline? Deep Sea Res., Part I, 46, 16531669.
  • Morel, A. (1988), Optical modeling of the upper ocean in relation to its biogenous matter content (case 1 waters), J. Geophys. Res., 93, 10,74910,768.
  • Morse, J. W., and F. T. Mackenzie (1990), Geochemistry of Sementary Carbonates, Elsevier, New York.
  • Mueller, J. L., et al. (2003), Ocean Optics Protocols for Satellite Ocean Color Sensor Validation, vol. 3, Radiometric Measurements and Data Analysis Protocols, rev. 4, 78 pp., NASA Goddard Space Flight Cent., Greenbelt, Md.
  • Najjar, R. G., J. L. Sarmiento, and J. R. Toggweiler (1992), Downward transport and fate of organic matter in the ocean: Simulations with a general circulation model, Global Biogeochem. Cycles, 6, 4576.
  • Napp, J. M., and G. L. J. Hunt (2001), Anomalous conditions in the southeastern Bering Sea 1997: Linkages among climate, weather, ocean and biology, Fish. Oceanogr., 10, 6168.
  • Paasche, E. (1962), Coccolith formation, Nature, 193, 10941095.
  • Paasche, E. (1998), Roles of nitrogen and phosphorus in coccolith formation in Emiliania huxleyi (Prymnesiophyceae), Eur. J. Phycol., 33, 3342.
  • Paasche, E. (1999), Reduced coccolith calcite production under light-limited growth: A comparative study of three clones of Emiliania huxleyi (Prymnesiophyceae), Phycologia, 38, 508516.
  • Paasche, E. (2002), A review of the coccolithphorid Emiliania huxleyi (Prymnesiophyceae), with particular reference to growth, coccolith formation, and calcification-photosynthesis interactions, Phycologia, 40, 503529.
  • Paasche, E., S. Brubak, S. Skattebøl, J. R. Young, and J. C. Green (1996), Growth and calcification in the coccolithophorid Emiliania huxleyi (Haptophyceae) at low salinities, Phycologia, 35, 394403.
  • Peng, T.-H., R. Wanninkhof, J. L. Bullister, R. A. Feely, and T. Takahashi (1998), Quantification of decadal anthropogenic CO2 uptake in the ocean based on dissolved inorganic carbon measurements, Nature, 396, 560563.
  • Riebesell, U., D. A. Wolf-Gladrow, and V. Smetacek (1993), Carbon dioxide limitation of marine phytoplankton growth rates, Nature, 361, 249251.
  • Riebesell, U., I. Zondervan, B. Rost, P. D. Tortell, R. E. Zeebe, and F. M. M. Morel (2000), Reduced calcification of marine plankton in response to increased atmospheric CO2, Nature, 407, 311313.
  • Riebesell, U., I. Zondervan, B. Rost, and R. E. Zeebe (2001), Effects of increasing atmospheric CO2 on phytoplankton communities and the biological carbon pump, Global Change Newsl., 47, 1215.
  • Robertson, J. E., C. Robinson, D. R. Turner, P. Holligan, A. J. Watson, P. Boyd, E. Fernandez, and M. Finch (1994), The impact of a coccolithophore bloom on oceanic carbon uptake in the northeast Atlantic during summer 1991, Deep Sea Res., Part I, 41, 297314.
  • Sabine, C. L., et al. (2004), The oceanic sink for anthropogenic CO2, Science, 305, 367371.
  • Samtleben, C., and T. Bickert (1990), Coccoliths in sediment traps from the Norwegian Sea, Mar. Micropaleontol., 16, 3964.
  • Sarmiento, J. L., J. Dunne, A. Gnanadesikan, R. M. Key, K. Matsumoto, and R. Slater (2002), A new estimate of the CaCO3 to organic carbon export ratio, Global Biogeochem. Cycles, 16(4), 1107, doi:10.1029/2002GB001919.
  • Seibold, E., and W. H. Berger (1982), The Sea Floor: An Introduction to Marine Geology, 288 pp., Springer, New York.
  • Shaw, G. E. (1983), Bio-controlled thermostasis involving the sulfur cycle, Clim. Change, 5, 297303.
  • Steele, J. H., and I. E. Baird (1962), Further relations between primary production, chlorophyll and particulate carbon, Limnol. Oceanogr., 7, 4247.
  • Stockwell, D. A., T. E. Whitledge, T. Rho, P. J. Stabeno, K. O. Coyle, J. M. Napp, S. I. Zeeman, and G. L. Hunt (2000), Field observations in the southeastern Bering Sea during three years with extensive coccolithophorid blooms (abstract), Eos Trans. AGU, 81, Ocean Sci. Meet. Suppl., Abstract OS12D-03.
  • Stumm, W., and J. J. Morgan (1981), Aquatic Chemistry, John Wiley, Hoboken, N. J.
  • Sundquist, E. T. (1993), The global carbon dioxide budget, Science, 259, 934941.
  • Takahashi, W., T. Hiwatari, H. Fukushima, M. Toratani, and T. Akano (1995), High-reflectance waters of possible coccolithophore blooms in NW pacific: Analysis of 1979–86 Nimbus-7/CZCS data set, Japanese, 6, 477486.
  • Townsend, D. W., M. D. Keller, P. M. Holligan, S. G. Ackleson, and W. M. Balch (1994), Blooms of the coccolithophore Emiliania huxleyi with respect to hydrography in the Gulf of Maine, Cont. Shelf Res., 14, 9791000.
  • Troy, P. J., Y.-H. Li, and F. T. MacKenzie (1997), Changes in surface morphology of calicite exposed to the oceanic water column, Aquat. Geochem., 3, 120.
  • Tyrell, T., P. M. Holligan, and C. D. Mobley (1999), Optical impacts of oceanic coccolithophore blooms, J. Geophys. Res., 104, 32233241.
  • Van der Wal, P., R. S. Kempers, and M. J. W. Veldhuis (1995), Production and downward flux of organic matter and calcite in a North Sea bloom of the coccolithophore Emiliania huxleyi, Mar. Ecol. Prog. Ser., 126, 247265.
  • Volk, T., and M. I. Hoffert (1985), Ocean carbon pumps: Analysis of relative strengths and efficiencies in ocean-driven atmospheric pCO2 changes, in The Carbon Cycle and Atmospheric CO2: Natural Variations Archean to Present, Geophys. Monogr. Ser., vol. 32, edited by E. T. Sundquist, and W. S. Broecker, pp. 99110, AGU, Washington, D. C.
  • Wal, P. D., R. S. Kempers, and M. J. W. Veldhuis (1995), Production and downward flux of organic matter and calcite in a North Sea bloom of the coccolithophore Emiliania huxleyi, Marine Ecol. Prog. Ser., 126, 247265.
  • Wollast, R. (1994), The relative importance of biomineralization and dissolution of CaCO3 in the global carbon cycle, report, edited by F. Doumenge, D. Allemand, and A. Toulemont, pp. 1335, Mus. Oceanogr., Monaco.
  • Young, J. R., and P. Ziveri (1999), Calculation of coccolith volume and its use in calibration of carbonate flux estimates, Deep Sea Res., Part II, 47, 16791700.

Supporting Information

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Methods
  5. 3. Results
  6. 4. Discussion
  7. Acknowledgments
  8. References
  9. Supporting Information
FilenameFormatSizeDescription
jgrc9764-sup-0001-t01a.txtplain text document0KTab-delimited Table 1a.
jgrc9764-sup-0002-t01b.txtplain text document0KTab-delimited Table 1b.
jgrc9764-sup-0003-t02.txtplain text document1KTab-delimited Table 2.
jgrc9764-sup-0004-t03.txtplain text document1KTab-delimited Table 3.
jgrc9764-sup-0005-t04.txtplain text document0KTab-delimited Table 4.
jgrc9764-sup-0006-t05.txtplain text document10KTab-delimited Table 5.
jgrc9764-sup-0007-t06.txtplain text document10KTab-delimited Table 6.
jgrc9764-sup-0008-t07.txtplain text document8KTab-delimited Table 7.
jgrc9764-sup-0009-t08.txtplain text document10KTab-delimited Table 8.

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.