(Received 26 July 2001; in revised form 29 November 2001; accepted 29 November 2001)

Transcription

1 Journal of Oceanography, Vol. 58, pp. 227 to 243, 2002 Review Seasonal and Interannual Variability in the Distribution of Surface Nutrients and Dissolved Inorganic Carbon in the Northern North Pacific: Influence of El Niño CHI SHING WONG 1 *, NATHALIE A. D. WASER 2, YUKIHIRO NOJIRI 3, WM. KEITH JOHNSON 1, FRANK A. WHITNEY 1, JOHN S. C. PAGE 1 and JIYE ZENG 4 1 Centre for Ocean Climate Chemistry, Institute of Ocean Sciences, P.O. Box 6000, Sidney, B.C., V8L 4B2, Canada 2 Department of Earth and Ocean Sciences, University of British Columbia, 6270 University Boulevard, Vancouver, B.C., V6T 1Z4, Canada 3 National Institute for Environmental Studies, 16-2 Onogawa, Tsukuba, Ibaraki , Japan 4 Global Environmental Forum Foundation, c/o National Institute for Environmental Studies, 16-2 Onogawa, Tsukuba, Ibaraki , Japan (Received 26 July 2001; in revised form 29 November 2001; accepted 29 November 2001) The seasonal and interannual changes in surface nutrients, dissolved inorganic carbon (DIC) and total alkalinity (TA) were recorded in the North Pacific (30 54 N) from 1995 to This study focuses on the region north of the subarctic boundary (~40 N) where there was extensive monthly coverage of surface properties. The nutrient cycles showed large interannual variations in the eastern and western subarctic gyres. In the Alaska Gyre the seasonal depletion of nitrate ( NO 3 ) increased from 8 14 µmol kg 1 in to 21.5 µmol kg 1 in In the western subarctic the shifts were similar in amplitude but more frequent. The large NO 3 levels were associated with high silicate depletions, indicating enhanced diatom production. The seasonal DIC:NO 3 drawdown ratios were elevated in the eastern and central subarctic due to calcification. In the western subarctic and the central Bering Sea calcification was significant only during 1997 and/or 1998, two El Niño years. Regional C/N stoichiometric molar ratios of 5.7 to 7.0 (>40 N) were determined based on the years with negligible or no calcification. The annual new production (NP a ) based on NO 3 and these C/N ratios showed large interannual variations. NP a was usually higher in the western than in the eastern subarctic. However, values of 84 gc m 2 yr 1 were found in the Alaska Gyre in 2000 which is similar to that in the most productive provinces of the northern North Pacific. There were also large increases in NP a around the Alaska Peninsula in 1997 and Finally, the net removal of carbon by the biological pump was estimated as 0.72 Gt C yr 1 in the North Pacific (>30 N). Keywords: Dissolved inorganic carbon, nitrate, seasonal cycles, new production, interannual variations. 1. Introduction The ocean draws down atmospheric CO 2 mainly due to the photosynthetic activity of the primary producers in the sunlit ocean surface. The lowering of atmospheric CO 2 levels depends to a large extent on the intensity of this biological process. The biological pump removes carbon * Corresponding author. WongCS@pac.dfo-mpo.gc.ca Copyright The Oceanographic Society of Japan. and nutrients from the ocean surface and exports them in the form of dissolved and particulate organic matter to the deep ocean where they are sequestered (Berger et al., 1989). The intensity of the biological pump can be derived from the magnitude of the seasonal DIC depletion as well as that of nitrate at the ocean surface (Codispoti et al., 1986; Codispoti, 1989; Karl et al., 1991; Minas and Minas, 1992; Ishii et al., 1998; Bates et al., 1998; Wong et al., 1998, 2002a). This later method allows the nitrate-based new production to be determined (Eppley, 227

2 1989). Knowledge of the C/N stoichiometric ratios then allows one to determine the net removal of carbon from the ocean surface (Wong et al., 2002a). Many studies have reported variability in the C/N ratio and deviations from the Redfield ratio (Redfield et al., 1963), suggesting that our knowledge is still uncertain (Takahashi et al., 1985; Minster and Boulahdid, 1987; Karl et al., 1991; Sambrotto et al., 1993; Anderson and Sarmiento, 1994; Ishii et al., 1998; Shaffer et al., 1999; Körtzinger et al., 2001; Wong et al., 2002a, b). This is particularly true for studies of surface nutrient drawdown which have often reported non- Redfieldian ratios (Karl et al., 1991; Sambrotto et al., 1993; Ishii et al., 1998; Körtzinger et al., 2001; Wong et al., 2002a, b). Given these uncertainties, it is important to accompany studies of new production based on seasonal nutrient depletions with measurements of C/N stoichiometric ratios. There is great interest in the productivity and the biogeochemical cycles of elements in the subarctic North Pacific and the Bering Sea, both regions being HNLC (High Nitrate-Low Chlorophyll) regions (Martin et al., 1989; Harrison et al., 1999; Banse and English, 1999). The western subarctic and the central Bering Sea are also vast regions with very large seasonal changes in nutrients, DIC and pco 2 (Takahashi et al., 1993; Nojiri et al., 2001; Zeng et al., 2002; Wong et al., 2002a). These regions are sources of CO 2 to the atmosphere in the winter due to vertical mixing and powerful CO 2 sinks in the summer due to the biological pump (Takahashi et al., 1993; Nojiri et al., 2001; Zeng et al., 2002). Time series studies of the surface biogeochemical cycles and productivity have been ongoing for 30 years at station P (50 N, 145 W) in the northeast subarctic Pacific (Wong and Chan, 1991; Wong et al., 1995, 1998, 2002b, 2002c; Boyd and Harrison, 1999; Wong and Matear, 1999; Whitney and Freeland, 1999). In contrast, far fewer surface observations have been made in the west, central subarctic and the Bering Sea (Codispoti et al., 1986; Takahashi et al., 1993; Longhurst et al., 1995; Shiomoto et al., 1998; Sasai et al., 2000; Nojiri et al., 2001; Wong et al., 2002a). The existing studies have revealed large contrasts between the eastern and western subarctic Pacific. First, the amplitude of the seasonal cycles of nutrients, DIC and pco 2 is much larger in the west than in the east (Takahashi et al., 1993; Nojiri et al., 2001; Wong et al., 2002a), resulting in higher rates of new production in the west (Wong et al., 2001a). Second, the C/N ratios were more elevated in the east than in the west due to greater CaCO 3 production in the east (Wong et al., 2002a, b). In contrast, diatom production is much greater in the west (Obayashi et al., 2001; Wong et al., 2002a). Similar observations have been reported from sediment trap studies (Wong et al., 1999; Takahashi et al., 2000). This paper reports the results of a large-scale study of the distribution of surface macronutrients, DIC and TA in the North Pacific (30 54 N) from January 1995 to January The data were collected on board the commercial cargo carrier M/V Skaugran. There was extensive monthly coverage of surface properties in regions north of the subarctic boundary (~40 N), and this study focuses on those regions (in particular the subarctic western and eastern Pacific and the southern central Bering Sea). As in the study of the first two years of the program (Wong et al., 2001a), the northern North Pacific was divided into 12 provinces. New production was calculated from the seasonal depletion of the monthly mean nitrate concentration, defined here as NO 3. Thus, NO 3 was defined as the difference between the maximum and minimum monthly mean nitrate concentration. The stoichiometric C/N ratios were determined from the seasonal DIC:NO 3 drawdown ratios at the surface. The importance of calcification was estimated from TA and DIC data. The net carbon removal via the biological pump was then estimated for the northern North Pacific. Finally, this paper also focuses on the relationship between the interannual variations in nutrient depletions and new production, and the El Niño Southern Oscillation (ENSO) cycle. At Station P, relationships have been established between NO 3, the particulate organic carbon (POC) flux and the particulate inorganic C (PIC) flux on the one hand, and the ENSO cycle on the other (Wong et al., 1998; Wong and Crawford, 2001). In particular, it was shown that during El Niño years, NO 3, POC flux and PIC flux were enhanced. Relationships with the North Pacific Index or NPI (difference in winter pressure between the Aleutian Low and the California coast), which is a better indicator of atmospheric conditions in the North Pacific, have also been examined and a significant correlation was found between NO 3 and NPI (Wong et al., 1998). 2. Materials and Methods The surface data were collected on board the commercial cargo carrier M/V Skaugran during six consecutive years from January 1995 to January This study was part of a bilateral ship-of-opportunity program involving cooperation between the Institute of Ocean Sciences (IOS) in Canada and the National Institute for Environmental Studies (NIES) in Japan. The westbound cruise tracks from Vancouver, Canada to several Japanese ports usually followed the great circle route (the most northern route via the Bering Sea), and were relatively consistent (Fig. 1). The eastbound routes from Japan to Canada varied more with the season. In the winter vessels would sail along a more southerly track to avoid storms at higher latitudes. Overall, about 4000 individual stations were occupied. In the first 3 4 years many more round trips (a round trip would typically take six weeks) 228 C. S. Wong et al.

3 Fig. 1. Transects of the M/V Skaugran cruises in the North Pacific Ocean during the period January 1995 to January The 12 major provinces covered by the tracks are shown. 1-BER: Central Bering Sea, 2-AKP: Alaska Peninsula, 3-AG: Alaska Gyre, 4-WSAG: West Subarctic Gyre, 5-CSP: Central Subarctic Pacific, 6-SUB: Subarctic Current System, 7-EJAP: East Japan, 8-HOK: Hokkaido, 9-DIL: Dilute Domain, 10-TD: Transition Domain, 11-CPG: Central Pacific Gyre and 12-EBER: East Bering Sea. The westbound tracks are shown by the triangles. Most of them are along the great circle route from Vancouver, Canada, to various Japanese ports. The great circle route is the most northern route. The eastbound tracks are shown by the circles. would be made each year, permitting detailed monthly coverage. There was much sparser coverage during the two last years of the program. More specifically, about 50% of the data set was collected during the March 1995 March 1997 period, 35% during April 1997 March 1999, and 15% during April 1999 December Coverage in the more northern regions was also much better, as shown by the great number of tracks along the great circle route as opposed to the more southerly routes. These biases should be kept in mind. 2.1 Sample collections and analyses The M/V Skaugran was equipped for shipboard sampling and measurements. A water intake line was installed at a sea chest near the bottom of the hull of the ship. Seawater was piped into the laboratory for the collection of discrete samples for shore laboratory analysis. The discrete samples included DIC, TA, NO 3 + NO 2 (further referred to as nitrate), PO 4 3, Si(OH) 4 2, chlorophyll-a and salinity. SST data were collected on board by the ship s engineer. Nutrient samples were collected twice a day by the ship s engineer. The other discrete samples were also collected three times a day by the research personnel. The samples were frozen until analysis in the laboratory at IOS. The macronutrient analyses were done by a Technicon Autoanalyzer (Barwell-Clarke and Whitney, 1996). During the M/V Skaugran cruises, the precision of the nitrate, silicate and phosphate measurements was ±0.53 µm, ±1.45 µm and ±0.034 µm, respectively. The DIC samples were preserved with HgCl 2 (Wong, 1970) and analyzed using the coulometric technique of Johnson and Sieburth (1996), adopted by the DOE handbook (DOE, 1994). The precision of the DIC measurements was ±2 µmol kg 1. TA samples were analyzed by two potentiometric methods, the closed cell method of Millero et al. (1998) and the open cell method of Dickson et al. (2001). The precision of the TA measurements was ±7.2 µeq kg 1 in closed cell and µeq kg 1 in open cell. The certified reference material (CRM) was CRM 48 supplied by A. Dickson (Scripps Institution of Oceanography, USA). 2.2 Data analyses Contour maps were made by first obtaining monthly averages in grids of 2 of latitude by 5 of longitude. Seasonal averages were then computed from the monthly averages. The seasonal averages could include 1, 2 or 3 monthly averages depending on data availability. The average seasonal and annual distribution of surface properties for the January 1995 January 2001 period were then derived as follows. The average seasonal value in a grid Seasonal and Interannual Variability 229

4 Fig. 2. Average annual distribution of surface properties during the Skaugran cruises for the January 1995 to January 2001 period. A) Salinity. B) SST. C) PO 4 3 (µmol kg 1 ). D) NO 3 (µmol kg 1 ). E) SiO 4 2 (µmol kg 1 ). F) Salinity-normalized DIC, ndic (µmol kg 1 ). G) Salinity-normalized TA, nta (µeq kg 1 ). H) Chl a (µg L 1 ). 230 C. S. Wong et al.

5 Fig. 2. (continued). Seasonal and Interannual Variability 231

6 for the period was computed from the average of the yearly seasonal averages. The number of yearly seasonal averages could vary from 1 to 6 depending on data availability. The average annual value in each grid for the period was computed from the average of the four average seasonal values (spring, summer, autumn and winter). The average annual value was only computed when all four average seasonal values were available. The Surfer 7.04 contouring program was then used to compute an interpolated grid of 20 lines by 7 lines, which was then spline smoothed. Monthly mean surface properties were also computed in 12 provinces. The division into 12 provinces was done mostly according to SST and surface nitrate gradients. Salinity was used to divide the near-shore regions along the Alaska Peninsula. Low salinity samples collected near islands in the central Bering Sea were eliminated from the data set. The provinces and their denominations were taken from Favorite et al. (1976) and were essentially the major water masses of the north Pacific. Further screening of the data sets was then done based on salinity. Samples with large deviations from the monthly mean salinity were eliminated. The results of the screening are as follows. In the provinces of BER, CSP, AG, SEBER and WSAG, salinities were within ± of the monthly mean. In HOK, DIL and AKP, deviations from the monthly mean salinity were larger at ± However, the monthly data set was not large enough to pursue further screening. Finally, in the provinces of EJAP and SUB, deviations from the monthly mean salinity were sometimes quite substantial, i.e. ± , but usually within ±0.5. There were great latitudinal gradients in these provinces, suggesting the occurrence of more than one water mass. Because of lack of data (and thus insufficient monthly coverage), these provinces could not be subdivided. Thus, again, emphasis is here given to the northern provinces, which correspond to well-defined water masses where there was extensive monthly coverage. 3. Results and Discussion 3.1 Distributions of surface properties The annual distributions of surface properties in the northern North Pacific (30 54 N) are shown in Fig. 2. These are average distributions for January 1995 to January During the time series there were two years which corresponded to El Niño years (1997/ 98) and one year which corresponded to La Niña year (1999). The distributions showed strong latitudinal gradients as well as an east-west contrast in the subarctic North Pacific. The subarctic region extends north of the vertical 34 isohaline (Favorite et al., 1976) corresponding to the subarctic boundary (~40 N). Salinity was low (i.e. 33 or less) in the subarctic Pacific (Fig. 2A). It was lowest along the Alaska Peninsula and along the northern North American Continent due to a large freshwater input. At low latitudes the distribution of salinity was asymmetric, with higher salinity in the west than in the east. This is partly due to the strong Kuroshio current, which transports high salinity water from the subtropical ocean into the west (Favorite et al., 1976) and partly due to greater vertical mixing in the west (Sasai et al., 2000). The asymmetry can be observed in the distribution of all properties. The east-west contrast can be observed in the distributions of many properties. Temperatures were cooler (Fig. 2B), nutrient concentrations were higher (Figs. 2C, D and E) and salinity-normalized DIC values (normalization to S = 35), ndic, were higher (Fig. 2F) in the western subarctic and the Bering Sea than in the eastern subarctic. Salinity-normalized TA, nta, was also higher in the west relative to the east (Fig. 2G). This contrast was due to the greater vertical mixing in the western Pacific, which brings to the surface elevated nutrient concentrations, DIC and TA (Sasai et al., 2000). Chlorophyll a (Chl a) concentrations were usually low, being about µg l 1 in the subarctic, except off Hokkaido (HOK), in the Southeast Bering Sea (SEBER) and along the Alaska Peninsula (AKP) where Chl a could reach high levels of µg l 1 (Fig. 2H). These regions have some of the largest seasonal nutrient cycles, new production (Shiomoto et al., 1998; Sasai et al., 2000; Wong et al., 2002a) and surface pco 2 cycles in the northern North Pacific (Nojiri et al., 2001; Zeng et al., 2002). The spatial variations were also apparent in the average seasonal distribution of nitrate (Fig. 3). In the spring, very high surface nitrate concentrations of 16 µmol kg 1 or more were observed in the Western Subarctic Gyre (WSAG) and the Bering Sea (BER) (Fig. 3A). Nitrate levels were much lower in the Alaska Gyre (AG). Nitrate concentrations were lower everywhere in the summer, but generally remained well above the detection limit north of the subarctic boundary (Fig. 3B). In the western subarctic gyre, nitrate levels decreased to 8 14 µmol kg 1 in the summer and showed some spatial variability. There were summer depletions to below detection levels in the Dilute Domain (DIL) in the east. This phenomenon has been well documented (Whitney and Freeland, 1999). Nitrate concentrations increased again (Fig. 3C) in the autumn. Winter nitrate concentrations were very elevated, averaging µmol kg 1 in the west and the Bering Sea and µmol kg 1 in the eastern subarctic (Fig. 3D). Salinity-normalized DIC was above 2,180 µmol kg 1 in WSAG and BER in the spring compared to the lower values of 2,160 µmol kg 1 in AG (Fig. 4A). In the summer, DIC decreased to low values of 2,100 2,160 µmol kg 1 in WSAG and BER and, 2,100 2,140 µmol kg 1 in AG (Fig. 4B). 232 C. S. Wong et al.

7 Fig. 3. Average seasonal distribution of surface nitrate concentration for the Jan Jan period. A) Spring (April, May and June). B) Summer (July, August and September). C) Autumn (October, November and December). D) Winter (January, February and March). Seasonal and Interannual Variability 233

8 Fig. 4. Average seasonal distribution of surface salinity-normalized DIC concentration from January 1995 to January A) Spring. B) Summer. C) Autumn. D) Winter. 234 C. S. Wong et al.

9 3.2 Time series along the great circle route The 6-year time series of surface properties along the great circle route is shown in Fig. 5. This route connects various Japanese ports at longitudes of about 140 E to Vancouver, Canada via the Bering Sea (Fig. 1) and crosses the provinces of HOK (145 to 150 E), WSAG (150 to 172 E), BER (172 to 190 E), SEBER (190 to 195 E), AKP (195 to 200 E), AG (200 to 218 E), Subarctic Current System (SUB) (218 to 225 E) and finally DIL (225 to 234 E) (Fig. 5). The SEBER and AKP regions appeared clearly with a salinity minimum in the E region (Fig. 5A). This salinity minimum revealed the separation between the Bering Sea and the western Pacific on the one hand and the eastern Pacific on the other hand. Salinity showed some temporal changes in BER where it increased from 1995 to 2000 by about 0.3. There was no long-term salinity trend in WSAG. In AG, salinity increased during the La Niña year Sea surface temperatures (SST) were cooler during the 1999 La Niña, particularly in BER and WSAG where SST values were lower by about 2 C (Fig. 5B). The seasonal nitrate cycle showed both spatial and interannual variations (Fig. 5C). In WSAG (150 to 172 E), while the winter maximum nitrate concentrations were relatively similar each year, the summer minimum exhibited great interannual variations. In 1996 and 1999 the summer minima were as low as 0 2 µmol kg 1 over a wide area in the E region, producing a very large seasonal nitrate drawdown of µmol kg 1 (Table 1). In BER there was a decrease in the seasonal nitrate depletion from µmol kg 1 in 1995, 1996 and 1997 to µmol kg 1 in 1998, 1999 and 2000, respectively (Fig. 5C, Table 1). In the Alaska Gyre (200 to 218 E) the seasonal nitrate depletions were quite similar the first five years of the time series but increased dramatically to 21.5 µmol kg 1 in 2000 (Table 1). The high NO 3 value was due both to higher maximum and lower minimum nitrate concentrations than usual (Fig. 5C). In WSAG, the large NO 3 of 1996, 1999 and 2000 were accompanied by large SiO 4 depletions (Fig. 5D, Table 1). In 1996, 1999 and 2000, the seasonal SiO 4 :NO 3 drawdown molar ratio ranged from 1.5 to 1.9, while it was in 1995, 1997 and The seasonal SiO 4 depletions were also very large in 1996, 1999 and 2000 (Table 1). These observations suggest that the elevated nitrate depletions observed in 1995 and during the 1997/ 98 El Niño years were due to enhanced diatom production. In AG, the seasonal SiO 4 :NO 3 depletion ratios (with values of ) and the seasonal SiO 4 depletions were usually relatively low compared to the western subarctic North Pacific. This was consistent with the knowledge that diatom production was lower on average in AG than in WSAG (Obayashi et al., 2001). However, in 2000, the seasonal SiO 4 :NO 3 depletion ratio was 1.5 in AG, i.e. the highest in the 6-year time series, and the seasonal depletions of SiO 4 and NO 3 were 34.7 and 21.5 µmol kg 1, respectively, indicating increased diatom production. In the subarctic Pacific as in other HNLC regions, the true Si/N uptake ratio of the diatom assemblage is quite elevated (Pondaven et al., 1999; Wong et al., 2002a). In a previous study (Wong et al., 2002a), we combined information from pigment analysis (Obayashi et al., 2001) and seasonal SiO 4 :NO 3 depletion ratios to estimate that the true Si/N uptake molar ratio of diatoms in the Bering Sea was about 2.7. Although there are variations in Si/N uptake ratios in diatom cultures (Brzezinski, 1985), as well as variations in diatom species forming the phytoplankton assemblage, a Si/N uptake ratio of about 3 for the subarctic Pacific and the Bering Sea seems to be consistent with both the observed SiO 4 :NO 3 drawdown ratios and the percentage of diatoms in the various phytoplankton assemblages (taken from Obayashi et al., 2001). In the Bering Sea ( E), the seasonal silicate depletions were greatest in 1995 and 1999 (Fig. 5D, Table 1). During these two years the SiO 4 :NO 3 drawdown ratio was 2.3 and 2.9, respectively. These are the largest ratios that we observed in the time series and we argue that in 1999 the entire seasonal nitrate depletion could be essentially attributed to diatoms. In contrast, during the El Niño years of 1997 and 1998 we observed low SiO 4 :NO 3 drawdown ratios of 1.7 and 2.0, respectively. As seen in the next section, 1998 was a year when significant calcification was found in the Bering Sea. This suggests that the seasonal nitrate depletion was not only due to diatom production but also due to production by calcifying organisms. Blooms of coccolithophores have been observed since 1997 in the Bering Sea with the Seaviewing Wide Field-of-view Sensor (SeaWiFS) ( In 1997, the M/V Skaugran data set shows that both ndic and nta decreased during the growth season (Figs. 5E and F). Analysis of the monthly mean properties showed that nta decreased by 12.6 µeq kg 1 and ndic by 25.0 µmol kg 1 during the July September period. These observations are consistent with the production of CaCO 3 (the ratio of CaCO 3 production to organic C production was estimated to about 25%). The observed changes in nta and ndic were much larger in 1998, as discussed in the next section. The time series of ndic and nta are shown in Figs. 5E and F. The ndic seasonal depletions were larger in the western subarctic and the Bering Sea than in the eastern subarctic, averaging about 90 µmol kg 1 in BER and WSAG compared to 81 µmol kg 1 in AG, respectively. In WSAG, DIC ranged from 70 to 105 µmol kg 1. In BER, ndic was slightly higher at µmol kg 1 during the 1997/98 El Niño years relative to µmol kg 1 in 1995/96. This difference was due to the greater CaCO 3 Seasonal and Interannual Variability 235

10 Fig. 5. Temporal and spatial distribution of surface properties along the great circle route. The great circle route crosses the provinces of HOK (43.1 N; 145 to 150 E), WSAG (49 N; 150 to 172 E), BER (56 N; 172 to 190 E), SEBER (58.5 N; 190 to 195 E), AKP (55 N; 195 to 200 E), AG (52.2 N; 200 to 218 E), SUB (45 N; 218 to 225 E) and DIL (47 N; 225 to 234 E). The latitudes indicated are the average latitudes of a province. A) Salinity. B) SST. C) Nitrate ( µmol kg 1). D) Silicate ( µmol kg 1). E) ndic (µmol kg 1). F) nta (µ eq kg 1). 236 C. S. Wong et al.

11 Fig. 5. (continued). production during the El Niño years. In AG, ndic was also variable, i.e µmol kg 1. It was highest in C/N stoichiometric ratios The C/N stoichiometric ratios associated with biological uptake were calculated on the basis of leastsquares regressions of nno3 and ndic monthly mean concentrations measured during the growth season (approximately late winter to late summer). The ndic:nno3 depletion ratios showed great variations temporally and spatially (Table 1). In the western provinces of HOK and WSAG, the ndic:nno3 drawdown ratios were usually close to or a little higher than the Redfield ratio. However, in WSAG the ratio was elevated at 7.9 in 1997 (Table 1). The time series revealed a large decrease in both nta and ndic in the spring of 1997 and a smaller one in the summer. This was consistent with the precipitation of calcium carbonate by organisms. Calcification produces a decrease in both TA and DIC. We can estimate the ratio of CaCO3 production to organic C production (defined here as α, 0 < α < 1). α can be derived from nta and ndic data as follows: nta = 2α ndic + nno 3 or α = 1/2( nta/ ndic) 1/2( nno3/ ndic). The effect of nitrate uptake on nta is small and thus α can be approximated as 1/2( nta/ ndic). In WSAG we estimated that α was 0.42 and 0.21 during the two periods May June and July August of 1997, respectively. We also found that calcification was a little higher (α = 0.33) than usual in HOK in May June of In the Bering Sea, the ndic:nno3 drawdown ratio was usually low, i.e (Table 1). However, it reached a high of 8.7 in Observations showed that there was a large decrease in both nta (80 µeq kg 1) and ndic (64 µmol kg 1) between June and July of This was consistent with the production of CaCO3 with a very high α value, which was estimated as 0.63 during that time period. Observations also showed that there were nta and ndic decreases between July and September of 1997 (with an α value of 0.25). This was the only event identified during that year and since it was relatively small it presumably did not give rise to very elevated ndic:nno3 depletion ratio (Table 1). Low molar ratios of were observed in the provinces of AKP and SEBER. In contrast, the ratios were usually relatively elevated in the Alaska Gyre, i.e , and Central Subarctic Pacific (CSP), i.e (Table 1). Time series data showed that the high depletion ratios were associated with years when, at times, α was higher than 0.2. In particular, in a previous study of the time series (Wong et al., 2002a), we identified a period with very high CaCO3 production in the spring of 1995 (α = 0.75) and another with lower production in the spring of 1996 (α = 0.23). There was no evidence of calcification in 1998 Seasonal and Interannual Variability 237

12 Table 1. Seasonal and annual new production (NP s and NP a ) and C:N stoichiometric ratios in the 12 provinces. in AG. In CSP, calcification was identified in 1995 but not in 1996 (Wong et al., 2002a). In 1997, there were two calcification events, one in the spring (α = 0.75) and the other in the summer (α = 0.29). In 1998, there was only one large calcification event in CSP and the estimated α was In the subarctic current system, evidence of high calcification was found in 1995, as in AG. In the Dilute Domain we found some evidence of calcification during the growth season during the period (α > 0.30 at times). However, in DIL the relationship between α and the ndic:nno 3 drawdown ratio was not clear and high ratios did not necessarily correspond to years with high α. In 1995, the high ndic:nno 3 ratio of 7.85 could not have been due to calcification because the seasonal decrease in nta relative to that of ndic was much too large. Physical processes were presumably responsible for the high ratios. Overall, we found that there was a relatively good correspondence between high ndic:nno 3 drawdown ratio and high calcification in the subarctic North Pacific. Similar conclusions were drawn in a time 238 C. S. Wong et al.

13 Table 1. (continued). *Indicates when calcification was important (i.e. α > 20%). **Average of the C:N drawdown ratio for 1995 to 1998 when calcification was not important. series study at Station P (Wong et al., 2002b). In East Japan we had previously identified calcification events in 1995 and especially in 1996 (Wong et al., 2002a). There were relatively large events in the spring (α = 0.42) of 1997 and the spring of 1998 (α = 0.37). However, these events could not have produced the large ndic:nno 3 drawdown molar ratios of (Table 1). We suggest that advection was also partly responsible for the high ratios. In EJAP there were relatively large variations in salinity within each month. In particular, there were large latitudinal changes, which suggested that the province should be subdivided further. However, the data set was too small to do the subdivision since it would have led to many months with no coverage. Seasonal and Interannual Variability 239

14 Table 2. Comparison between the average DIC: NO 3 ratio determined from regression of the spring summer monthly mean salinity-normalized concentrations (Ave. DIC: NO 3s ) and average DIC: NO 3 ratio determined from regressions of all monthly mean salinity-normalized concentrations (Ave. DIC: NO 3a ). Province Ave. DIC: NO 3s Ave. DIC: NO 3a molar ratio molar ratio BER AG SEBER HOK WSAG CSP AKP SUB EJAP DIL To calculate the seasonal new production, an average C/N stoichiometric ratio was used for all six years. The average C/N uptake ratio was derived from the first four years of the time series when there was extensive monthly coverage. Furthermore, the average C/N ratio was calculated from the ndic:nno 3 drawdown ratios for the years when calcification was low (i.e. when only one event with α < 0.2 was identified during the seasonal nutrient drawdown). We also assumed that the net effect of gas exchange and dissolved organic matter production was small. This assumption was shown to be valid at Station P (Wong et al., 2002b) and in the North Atlantic (Broström, 1998). Thus, we did not correct any C/N stoichiometric ratio for gas exchange, as is often done, nor for DOM production. Neither did we include the very high ratios of found in EJAP, which were likely largely due to advective processes. In the Transition Domain (TD) and the Central Pacific Gyre (CPG) there was not enough seasonal coverage to determine C/N ratios and thus we used the Redfield ratio of 6.5. The contrast between the western and the Bering Sea on the one hand, and the eastern and central subarctic Pacific on the other can be summarized as follows. Average ndic:nno 3 drawdown ratios based on the ones listed in Table 1 (average of all the years from 1995 to 1998, including years with and without significant calcification) were computed (Table 2). The average molar ratios had values less than about 7.1 (i.e. ranging from 6.2 to 7.0) in the provinces of BER, SEBER, HOK, WSAG and AKP and, higher (i.e. about 7.8) in AG and CSP. The first five provinces were provinces where calcification was found to be occasionally important during the period. For BER, HOK and WSAG, there was evidence of Fig. 6. Seasonal new production (NP s ) calculated for each year from 1995 to 2000 (when data are available) in each of the 12 provinces. NP s is based on the seasonal nitrate drawdown (difference between the winter maximun and the summer minimum monthly mean nitrate concentration). The solid line represents the average seasonal new production for the years 1995 to calcification only during the El Niño years 1997 and There was no evidence of CaCO 3 production in any other years. The last two provinces, AG and CSP, were regions where calcification was shown to be very important during many years in the period. Finally, we should add that similar results were obtained by looking at the relationship between the monthly mean ndic and nno 3 concentrations of all the months, including fall and winter data (Table 2). 240 C. S. Wong et al.

15 Table 3. Average annual new production (Ave. NP a ) and average C removal (R c ) in each province for the years 1995 to Province Ave. NP a Area R c gc m 2 y m gc y 1 Central Bering Sea Alaska Gyre East Being Sea Hokkaido West Subarctic Gyre Central Sub. Pacific Alaska Peninsula Sub. Current System East Japan Dilute Domain Transition Domain Central Pacific Gyre TOTAL Seasonal new production New production was determined as explained previously in Wong et al. (2002a), where the M/V Skaugran data set was analyzed. Briefly, the seasonal new production (NP s ) was derived from the monthly mean nitrate concentrations. The time that elapsed between the maximum and minimum monthly mean nitrate concentrations ( T) was also determined. To calculate the seasonal new production NP n (in mmol N m 2 d 1 ), it was assumed that the seasonal removal of N occurred, on average, in the top 50 m of the ocean in all provinces. This assumption was based on information about the average depth of the euphotic zone in various regions of the North Pacific (Longhurst et al., 1995; Boyd and Harrison, 1999; Harrison et al., 1999). The seasonal new production NP s (in gc m 2 yr 1 ) could then be determined from the C/N stoichiometric ratios. The seasonal new production, NP s, is shown in Fig. 6. It was highest in AKP and SEBER, averaging 320 and 206 gc m 2 yr 1 for the period (Fig. 6). Furthermore, it was very high during the El Niño years of 1997 and 1998 in SEBER and AKP (Fig. 6). The average NP s was high in the western provinces HOK and WSAG. It was also high in AG due to the extremely high NP s of 800 gc m 2 yr 1 in This dramatic increase in NP s was due to a very rapid (39 d) almost complete depletion of surface nitrate (Table 1). For the first five years, NP s averaged 106 gc m 2 yr 1 in AG, which is similar to values in BER and CSP and about half the values in WSAG and HOK. In SUB as in AG, NP s greatly increased in 2000 relative to the previous years. An increase in NP s in 2000 was also observed in DIL. Overall, the seasonal new production showed great interannual variability. 3.5 Annual new production The average annual new production was then calculated from the seasonal new production assuming that there was no biological activity during the fall and winter (Table 3). The net removal of C via the biological pump was then determined in each province. For this we assumed that the results of sometimes relatively limited areas could be extrapolated to an entire province as defined in Fig. 1. The average NP a was highest in HOK at 78 gc m 2 yr 1 and lowest in CPG at 14 gc m 2 yr 1. We then estimated that the biological pump removed an average of 0.72 Gt C per year in the North Pacific (>30 N) for the years 1995 to Zeng et al. (2002), based on surface pco 2 measurements made during the same M/V Skaugran cruises, estimated that the net oceanic uptake of CO 2 was 0.26 Gt C year 1 for the 1995 to 1999 period in the N region. Restricting our analysis to these five years, we estimated that the biological activity of phytoplankton removes on average 0.53 Gt C year 1 in the N region. 4. Conclusion This study has demonstrated that there were large seasonal and interannual variations in the surface nutrient cycles and new production in the northern North Pacific. The relationship between these variations and El Niño was not simple. In the western subarctic and the south central Bering Sea, large nitrate seasonal depletions occurred in 1996, 1999 and These large depletions were linked to increased diatom production. Furthermore, in these regions the production of CaCO 3 (based on TA and DIC data) became important during the 1997/98 El Niño years. There was also a correspondence between increased calcification events and the occurrence of elevated seasonal DIC:NO 3 drawdown ratios in 1997 or 1998 or during both years in the provinces of the western subarctic Gyre, off Hokkaido and the central Bering Sea. Observations also showed that years with high calcification in the western subarctic Gyre coincided with low seasonal nitrate depletions. In the Alaska Gyre, the seasonal nitrate depletions and new production were usually much lower than in the west and the Bering Sea. However, in 2000 the amplitude of the seasonal nitrate depletion was as large as the ones found in the very productive province off Hokkaido and during some years in the western subarctic gyre. That large nitrate depletion was also linked to enhanced diatom production. The importance of calcification varied from year to year in the Alaska gyre and no clear relationship could be established with the occurrence of El Niño. Finally, this study has shown that the 1997/98 El Niño had a great impact on the seasonal cycles in the regions surrounding the Alaska Peninsula. There was a very large Seasonal and Interannual Variability 241

16 increase in the seasonal new production in these areas, not so much due to the magnitude of the seasonal nutrient depletion but due to the rapidity with which the depletion occurred. Acknowledgements We thank the owners, captain, engineers and crew of the commercial cargo carrier M/V Skaugran (owned by Seaboard International Shipping Co., Vancouver, Canada) for their support and participation in this extensive monitoring program. We are very grateful to the staff of the National Institute for Environmental Studies and the Global Environmental Forum Foundation for maintaining and manning the monitoring program. We also thank two anonymous reviewers for their insightful comments. The Canadian work is supported by Fisheries and Oceans Canada for salary and laboratory facilities, and by project grants #52539 (Oceanic CO 2 uptake) and #52540 (CO 2 partition and air-sea CO 2 exchange) from the Panel for Energy & Research Development (PERD) of Natural Resources Canada. References Anderson, L. A. and J. L. Sarmiento (1994): Redfield ratios of remineralization determined by nutrient data analysis. Global Biogeochem. Cycles, 8, Banse, K. and D. C. English (1999): Comparing phytoplankton seasonality in the eastern and western subarctic Pacific and the western Bering Sea. Prog. Oceanogr., 43, Barwell-Clarke, J. and F. Whitney (1996): Institute of Ocean Sciences nutrient methods and analysis. Canadian Technical Report of Hydrography and Ocean Sciences, 182, 43 pp. Bates, N. R., D. A. Hansell, C. A. Carlson and L. I. Gordon (1998): Distribution of CO 2 species, estimation of net community production and air-sea CO 2 exchange in the Ross Sea polynia. J. Geophys. Res., 103, Berger, W. H., V. S. Smetacek and G. Wefer (1989): Ocean productivity and paleoproductivity An overview. p In Productivity of the Ocean: Present and Past, ed. by W. H. Berger, V. S. Smetacek and G. Wefer, John Wiley & Sons Ltd. Boyd, P. W. and P. J. Harrison (1999): Phytoplankton dynamics in the NE subarctic Pacific. Deep-Sea Res. II, 46, Broström, G. (1998): A note on the C/N and C/P ratio of the biological production in the Nordic Seas. Tellus, B50, Brzezinski, M. A. (1985): The Si:N ratio of marine diatoms interspecific variability and the effect of some environmental variables. J. Phycol., 21, Codispoti, L. A. (1989): Phosphorus vs. nitrogen limitation of new and export production. p In Productivity of the Ocean: Present and Past, ed. by W. H. Berger, V. S. Smetacek and G. Wefer, John Wiley & Sons Ltd. Codispoti, L. A., G. E. Friederich and D. W. Hood (1986): Variability in the inorganic carbon system over the southeastern Bering Sea shelf during spring 1980 and spring summer Cont. Shelf Res., 5, Dickson, A. G., J. D. Afghan and G. C. Anderson (2001): Seawater based reference materials for CO 2 analysis: 2. A method for the certification of total alkalinity. Mar. Chem. (submitted). DOE (1994): Handbook of Methods for the Analysis of the Various Parameters of the Carbon Dioxide System in Seawater, version 2, Carbon Dioxide, ed. by A. G. Dickson and C. Goyet, Information Center, Report ORNL/CDIAC-74, Oak Ridge National Laboratory, Oak Ridge, TN, U.S.A. Eppley, R. W. (1989): New production: History, methods, problems. p In Productivity of the Ocean: Present and Past, ed. by W. H. Berger, V. S. Smetacek and G. Wefer, John Wiley & Sons Ltd. Favorite, F. A., A. J. Dodimead and K. Nasu (1976): Oceanography of the subarctic Pacific region, Bull. Inter. North Pac. Fish. Comm., 33, Harrison, P. J., P. W. Boyd, D. E. Varela, S. Takeda, A. Shiomoto and T. Odate (1999): Comparison of factors controlling phytoplankton productivity in the NE and NW Subarctic Pacific Gyres. Prog. Oceanogr., 43, Ishii, M., H. Y. Inoue, H. Matsueda and E. Tanoue (1998): Close coupling between seasonal biological production and dynamics of dissolved inorganic carbon in the Indian Ocean sector and the western Pacific Ocean sector of the Antarctic Ocean. Deep-Sea Res. I, 45, Johnson, K. M. and J. McN Sieburth (1996): Coulometric total carbon dioxide analysis for marine studies: automation and calibration. Mar. Chem., 21, Karl, D. M., B. D. Tilbrook and G. Tien (1991): Seasonal coupling of organic matter production and particle flux in the western Bransfield Strait, Antarctica. Deep-Sea Res., 38, Körtzinger, A., W. Koeve, P. Kähler and L. Mintrop (2001): C:N ratios in the mixed layer during the productive season in the northeast Atlantic Ocean. Deep-Sea Res. I, 48, Longhurst, A., S. Sathyendranath, T. Platt and C. Caverhill (1995): An estimate of global primary production in the ocean from satellite radiometer data. J. Plankton. Res., 17, Martin, J. H., R. M. Gordon, S. E. Fitzwater and W. W. Broenkow (1989): VERTEX: phytoplankton/iron studies in the Gulf of Alaska. Deep-Sea Res., 36, Millero, F. J., K. Lee and M. Roche (1998): Distribution of alkalinity in the surface waters of the major oceans. Mar. Chem., 60, Minas, H. J. and M. Minas (1992): Net community production in High Nutrient-Low Chlorophyll waters of the tropical and Antarctic Oceans: grazing vs iron hypothesis. Oceanol. Acta, 15, Minster, J. F. and M. Boulahdid (1987): Redfield ratios along isopycnal surfaces. A complementary study. Deep-Sea Res., 34, Nojiri, Y., Y. Fujinuma, J. Zeng and C. S. Wong (2001): Measurements in the northern North Pacific from a commercial vessel: 2, Seasonal and spatial characteristics of fco 2, J. Geophys. Res. (submitted). 242 C. S. Wong et al.

17 Obayashi, Y., E. Tanoue, K. Suzuki, N. Handa, Y. Nojiri and C. S. Wong (2001): Spatial and temporal variabilities of the phytoplankton community structure in the northern North Pacific as determined by phytoplankton pigments. Deep- Sea Res. I, 48, Pondaven, P., D. Ruiz-Pino, J. N. Druon, C. Fravalo and P. Tréguer (1999): Factors controlling silicon and nitrogen biogeochemical cycles in high nutrient, low chlorophyll systems (the southern Ocean and the North Pacific): Comparison with a mesotrophic system (the North Atlantic). Deep-Sea Res. I, 46, Redfield, A. C., B. H. Ketchum and F. A. Richards (1963): The influence of organisms on the composition of sea-water. p In The Sea, vol. 2, ed. by M. N. Hill, Interscience, New York. Sambrotto, R. N., G. Savidge, C. Robinson, P. Boyd, T. Takahashi, D. M. Karl, C. Langdon, D. Chipman, J. Marra and L. Codispoti (1993): Elevated consumption of carbon relative to nitrogen in the surface ocean. Nature, 363, Sasai, Y., M. Ikeda and N. Tanaka (2000): Changes in total CO 2 and pco 2 in the surface ocean during the mixed layer development in the northern North Pacific. J. Geophys. Res., 105, Shaffer, G., J. Bendtsen and O. Ulloa (1999): Fractionation during remineralization of organic matter in the ocean. Deep-Sea Res. I, 46, Shiomoto, A., Y. Ishida, M. Tamaki and Y. Yamada (1998): Primary production and chlorophyll a in the northwestern Pacific Ocean in summer. J. Geophys. Res., 103, Takahashi, K., N. Fujitani, M. Yanada and Y. Maita (2000): Long-term biogenic particle fluxes in the Bering Sea and the central subarctic Pacific Ocean, Deep-Sea Res. I, 47, Takahashi, T., W. S. Broecker and S. Langer (1985): Redfield ratio based on chemical data from isopycnal surfaces. J. Geophys. Res., 90, Takahashi, T., J. Olafsson, J. G. Goddard, D. W. Chipman and S. C. Sutherland (1993): Seasonal variation of CO 2 and nutrients in the high-latitude surface oceans: A comparative study. Global Biogeochem. Cycles, 7, Whitney, F. A and H. J. Freeland (1999): Variability in upperocean water properties in the NE Pacific Ocean. Deep-Sea Res. II, 46, Wong, C. S. (1970): Quantitative analysis of total carbon dioxide in sea water: a new extraction method. Deep-Sea Res., 17, Wong, C. S. and Y.-H. Chan (1991): Temporal variations in the partial pressure and flux of CO 2 at ocean Station P in the subarctic northeast Pacific Ocean. Tellus, B43, Wong, C. S. and D. Crawford (2001): Fluxes of particulate inorganic and organic carbon to the deep subarctic Pacific correlates with El Niño. Deep-Sea Res. II (accepted). Wong, C. S. and R. J. Matear (1999): Sporadic silicate limitation of phytoplankton productivity in the Subarctic NE Pacific. Deep-Sea Res. II, 46, Wong, C. S., F. A. Whitney, K. Iseki, J. S. Page and J. Zeng (1995): Analysis of trends in primary productivity and chlorophyll a over two decades at Ocean Station P (50 N, 145 W) in the subarctic northeast Pacific Ocean. p In Climate Change and Northern Fish Populations, vol. 121, ed. by R. J. Beamish, Canadian Special Publication of Fisheries and Aquatic Sciences. Wong, C. S., F. A. Whitney, R. J. Matear and K. Iseki (1998): Enhancement of new production in the northeast subarctic Pacific Ocean during negative North Pacific index events. Limnol. Oceanogr., 43, Wong, C. S., F. A. Whitney, D. W. Crawford, K. Iseki, R. J. Matear, K. W. Johnson, J. S. Page and D. Timothy (1999): Seasonal and interannual variability in particle fluxes of carbon, nitrogen and silicon from time series of sediment traps at Ocean Station P, : relationship to changes in subarctic primary productivity. Deep-Sea Res. II, 46, Wong, C. S., N. A. D. Waser, Y. Nojiri, F. A. Whitney, J. S. Page and J. Zeng (2002a): Seasonal cycles of nutrients and dissolved inorganic carbon at high and mid latitudes in the North Pacific Ocean during the Skaugran cruises: Determination of new production and nutrient uptake ratios. Deep- Sea Res. II (in press). Wong, C. S., N. A. D. Waser, F. A. Whitney, W. K. Johnson and J. S. Page (2002b): Time series study of the biogeochemistry of the North East subarctic Pacific Ocean: Reconciliation of the C org /N remineralization and uptake ratios with the Redfield ratios. Deep-Sea Res. II (in press). Wong, C. S., Z. Yu, N. A. D. Waser, F. A. Whitney and W. K. Johnson (2002c): Seasonal changes in the distribution of dissolved organic nitrogen in coastal and open ocean waters in the North Pacific: Sources and sinks. Deep-Sea Res. II (in press). Zeng, J., Y. Nojiri, Y. Fujinuma, P. Murphy and C. S. Wong (2002): A comparison of pco 2 distribution in the northern North Pacific using results from a commercial vessel in Deep-Sea Res. II (in press). Seasonal and Interannual Variability 243