Results and Discussion

3.1 Long-Term Trend of pCO|W in the HNLC Region

As compared with pCO2w distribution in the subtropics showing basically the same pattern in the same time (month) of the year (Inoue et al., 1995), the pCO2w distribution in the equatorial Pacific varied considerably especially associated with the ENSO phenomena (Fig. 2). In the equatorial Pacific, therefore, it is extremely difficult to detect the long-term trend of pCO2w by simply comparing pCO2w data taken several years apart (Goyet and Peltzer, 1994).

Thermodynamics, ocean dynamics (lateral flow, vertical mixing, and up-welling of water), and biological activities are considered to be major processes (Poisson et al., 1993; Lee et al., 1998) affectingpCO2w via changes in total dissolved inorganic carbon (DIC), total alkalinity, pH, temperature, and salinity. One of the approaches for evaluating the long-term trend is to compare the pCO2w data taken at the same thermodynamic, oceanographic and biological conditions. Feely et al. (1999) have reported the long-term trend of pCO2w for the core of upwelled water in the region of temperature minimum near the equator between 140 and 160°W using temperature

140°E 150 160 170 180 170 160 150°W Longitude

Figure 2: Longitudinal distributions of pCO2w, pCOfir, [NO—]+[NO—], SSS and SST along the equator in October/December 1999 (red), January/March 2001 (blue), January/March 2002 (green) and November 2002/March 2003 (yellow). In the upper panel, the thick line shows pCO2w and the thin line pCO|ir. Arrows show the boundary between the HNLC region and western Pacific warm pool (For colour version, see Colour Plate Section).

140°E 150 160 170 180 170 160 150°W Longitude

Figure 2: Longitudinal distributions of pCO2w, pCOfir, [NO—]+[NO—], SSS and SST along the equator in October/December 1999 (red), January/March 2001 (blue), January/March 2002 (green) and November 2002/March 2003 (yellow). In the upper panel, the thick line shows pCO2w and the thin line pCO|ir. Arrows show the boundary between the HNLC region and western Pacific warm pool (For colour version, see Colour Plate Section).

normalized pCO2w values. Feely et al. (2002) also discussed the long-term trend of pCO2w along 150°W between 1979 and 1990s by using a pCO2w—SST relationship.

As pointed out by Wanninkhof et al. (1996), SST by itself is not a robust predictor of pCO2w in the HNLC region. Dandonneau (1995) concluded that most of the variations of pCO2w in the eastern equatorial Pacific were coherent in parallel with nitrate concentration and fluorescence, and inversely with SST. Wanninkhof et al. (1996) reported that the pCO2w showed the best correlation with a combination of NO— and SST:

where A, B, and C are coefficients determined empirically and subscript c means the HNLC region. In this work, we assumed that calculating the pCO2w normalized to the same SST and concentration of nitrate and nitrite ([NO—]+[NO—], hereafter expressed as NO3) is the best approach to determine the pCO2w value at the same conditions of thermodynamics, ocean dynamics and biological activities in the HNLC region. The spatial distribution ofpCO2w in the HNLC region was generally approximated well using equation (2) (Table 2). An example of the least squares fit of equation (2) to the data in January 2001 is shown in Fig. 3, in which the pCO2w has been calculated with a root-mean-squares deviation (rms) of 2 matm. However, the coefficients for this relationship cannot be applied to the data during other phases of the ENSO cycle with the same amount of accuracy (Wanninkhof et al., 1996). For example, the relationship of pCO2w with SST and NO3 in January 2001 gave a deviation (bias) of about 13 matm for the pCO2w in January/February 1999 (Table 2).

In order to evaluate the long-term trend, we determined the coefficients and the intercept of equation (2) for respective cruises (Table 2) and then calculated the pCO2w normalized to the average SST and NO3. The average SST (27.4 °C) and NO3 (3.9 mmol/kg) observed from 1989 to 2003 minimize any effects of changes in thermodynamics, ocean dynamics, and biological activity on the pCO2w. By assuming the linear long-term trend, the growth rate of pCO2w at the average SST and NO3 was estimated to be 1.470.5 matm/yr (Fig. 4) in the HNLC region, which is very similar to the increase in atmospheric CO2. Figure 5 represents the growth rate ofpCO2w expressed as a function of SST and NO3, which varied from 0.6 to 2.1 matm/ yr. If the pCO2w were determined at exactly the same condition of thermodynamics, ocean dynamics, and biological activity, the calculated growth rate of pCO2w would be the same irrespective of changes in SST and NO3. NO3 increases with decrease in SST (Chavez et al., 1996). In this context, the fairly constant growth rate of pCO2w with conditions between low SST—high NO3 and high SST—low NO3 as indicated by an arrow in Fig. 5,

Table 2: Multiple linear regressions to estimate the spatial distribution of pCO2w 1989 2003

in the HNLC region in

Time of observation Longitude

Equation e

Jan.-Feb. 1989 Sep.-Oct. 1990 Jan. 1994 Jan. 1996 Jan. 1997 Jan. 1999 Dec. 1999 Jan. 2001 Jan. 2002 Jan. 2003

pCO2w(±3.8) = 263.9(±78.3)+4.36(±2.59)SST+9.44(±2.08)N03 9 0.948

pC02w(±5.9) = 807.9(7 84.9)-14.65( 7 2.88)SST+2.37( 71.45)N03 23 0.952

pC02w( 73.8) = 1060.9( 7 267.0)-22.35( 7 8.92)SST-2.37( 7 5.75) NO3 7 0.989

pC02w(73.6) = -137.2(7183.5)+18.07( 7 6.14)SST+20.16(7 4.66)N03 8 0.948

pC02w(72.7) = 372.2(756.7)+0.87(71.90)SST+11.86(71.19)N0S 32 0.984

pC02w(73.4) = 168.5(790.6)+7.29(73.17)SST+16.58(72.18)N0S 68 0.969

pC02w( 73.9) = 412.0( 7 41.9)-0.07( 71.44)SST+7.58( 7 0.77)N0s 175 0.941

pC02w(72.6) = 272.1(725.1)+4.42(70.82)SST+12.42(70.60)N03 194 0.978

pC02w(72.1) = 478.3(770.1)-2.41(72.34)SST+8.04(71.25)N0S 59 0.957

pC02w(71.7) = 239.1(741.9)+5.28(71.40)SST+12.33(70.62)N0s 94 0.987

Note: The standard error of analysis (in matm) is listed in parentheses after pC0|w, the standard error for the coefficient and intercept are in parenthesis after coefficient and intercept, n is the number of data, and r is the correlation coefficient. Macronutrients in surface seawater were monitored continuously for cruises conducted after 1999.

410 420 430 440 450 460 470 Observed pC02sw (^atm)

Figure 3: The pCO2w calculated as the function of SST and NO3 against observed pCO2w in January/February 2001.

410 420 430 440 450 460 470 Observed pC02sw (^atm)

Figure 3: The pCO2w calculated as the function of SST and NO3 against observed pCO2w in January/February 2001.

400 I—■_i_■_>_i_'_i_i_i_i_i_i_i_i_i_i_i_i_i—I

1985 1990 1995 2000 2005

400 I—■_i_■_>_i_'_i_i_i_i_i_i_i_i_i_i_i_i_i—I

1985 1990 1995 2000 2005

Year

Figure 4: The pCO2w at average SST (27.4 °C) and NO3 (3.9 mmol/kg) in the HNLC region west of 150°W in 1989-2003. The solid line shows the linear long-term trend ofpCO2w(t) - 1.4(t-1985)+418.0, where t is the elapsed time (year) since January 1985. The standard error of analysis was 7.4 matm, and the standard error for the coefficient was 0.5 matm/yr, intercept 6.7 matm and the correlation coefficient was 0.68.

could suggest the validity of the present approach for estimating the long-term trend of pCO2w.

Our results obtained in the HNLC region also agreed well with the increase of pCO2w in the core of upwelled water, where the entrained

Figure 5: The growth rate of pCO2w that was calculated using SST and NO3 via equation (2). Solid circle shows the growth rate ofpCO2w at average SST and NO3. An arrow shows apparent SST dependence of NO3- observed in the HNLC region.

Figure 5: The growth rate of pCO2w that was calculated using SST and NO3 via equation (2). Solid circle shows the growth rate ofpCO2w at average SST and NO3. An arrow shows apparent SST dependence of NO3- observed in the HNLC region.

subtropical water is injected into the upwelled water at the equator. This suggests that the upwelled water has been recently exposed to, and equilibrated with, the atmosphere (Feely et al., 1999). After upwelling of water, the current atmospheric CO2 level does not largely affect the growth rate of pCO2w in the HNLC region, suggesting a relatively small uptake of anthropogenic CO2 in the HNLC region. The apparent growth rate of total DIC in the HNLC region inferred from pCO2w and buffer (Revelle) factor is estimated to be ^0.8 mmol/kg/yr. This agrees well with the DIC increase of 0.91 + 0.09mol/kg/yr that Ishii et al. (2003) deduced from the apparent temperature dependence of normalized DIC and total alkalinity.

3.2 Long-Term Trend of pCO2w in the Western Pacific Warm Pool

The western Pacific warm pool migrates eastward following the occurrence of El Nino. In this warm pool, high pCO2w is absent due to a deep therm-ocline and nutricline, combined with a barrier layer produced by shallow haloclines (Le Borgne et al., 2002). A barrier layer between a shallow halocline and the top of the thermocline prevents entrainment of cold water, rich in CO2 and nutrients, across the bottom of the mixed layer, indicating the importance of surface salinity field in the warm pool on modulating pCO2w levels. The distribution of pCO2w in the western Pacific warm pool can be approximated well by a linear function of SST and SSS

where A, B, and C are coefficients determined empirically and subscript w means western Pacific warm pool. The relationship of pCO2w with SST and SSS in January/February 2002 is shown as an example in Fig. 6. This relationship of pCO2w with SST and SSS cannot be applied to the data recorded during other periods (Table 3), as is the case for equation (2) in the HNLC region. The growth rate of pCO2w at the average SST and SSS observed in 1987—2003 is calculated to be 1.370.3 matm/yr in the western Pacific warm pool (Fig. 7). Within the range of a standard deviation from the average for SST and SSS, the long-term trend of pCOs2w ranged from 1.0 to 1.6 matm/yr (Fig. 8). With respect to changes in SST and SSS in the western Pacific warm pool, Ando and Kuroda (2002) reported two dominant modes: a positive correlation mode in which the density field is not affected and a negative correlation mode having substantial impact on the density field. The fairly constant growth rate of pCOs2w occurred under conditions of increasing SST and SSS and thus a negligible change in the density field. A stable density field provides favorable conditions for determining the long-term trend ofpCO2w. The growth rate ofpCO2w determined for average SST and SSS in the western Pacific warm pool is equal to that given by the average SST and NO3 in the HNLC region.

In order to evaluate the long-term trend of pCOs2w in the HNLC region and western Pacific warm pool, we used a multiple linear regression

365 370 375 380 385

Observed pC02s" (^atm)

Figure 6: The pCOs2w calculated as a function of SST and SSS against observed pCOs2w in January/February 2001.

365 370 375 380 385

Observed pC02s" (^atm)

Figure 6: The pCOs2w calculated as a function of SST and SSS against observed pCOs2w in January/February 2001.

Table 3: Multiple linear regressions to estimate the spatial distribution ofpCO2w in the western Pacific warm pool in 1987-2003

Time of observation Longitude

Equation e

Jan.-Feb. 1987 Jan.-Feb. 1990 Nov. 1990 Feb. 1991 Jan.-Feb. 1994 Nov.-Dec. 1994 Dec. 1995 Jan.-Feb. 1997 Oct. 1997 Dec. 1997 Feb. 1998 Jan. 1999 Nov. 1999 Jan.-Feb. 2002 Mar. 2002 Jan. 2003

pCO2w(±2.4) = -717.6(±242.7)+10.66(±2.82)SST+21.90(±5.45)SSS 19 0.743

pCO2w( 7 0.9) = —619.9( 7 242.7)+8.54( 7 3.46)SST+21.08( 7 2.59)SSS 8 0.972

pCO2w(71.8) = 88.8(7124.9)—1.36(72.37)SST+8.71(72.62)SSS 10 0.800

pCO2w(73.1) = —605.8(7204.4)+7.47(76.37)SST+21.24(77.96)SSS 11 0.859

pCO2w(76.5) = —20.2(7155.3)—7.57(71.58)SST+17.42(73.75)SSS 192 0.555

pCO2w(73.1) = —82.0(732.8)+13.2(70.88)SST+1.36(71.16)SSS 363 0.710

pCO2w(73.0) = —325.1(793.6)+5.83(72.48)SST+14.87(72.48)SSS 68 0.680

pCO2w(73.6) = —249.2(794.6)—1.44(71.42)SST+19.19(73.46)SSS 43 0.736

pCO2w( 7 2.3) = —579.0( 7 38.3)+10.43( 7 0.67)SST+18.60( 7 0.89)SSS 162 0.890

pCO2w( 7 2.8) = —536.2( 7 22.2)+9.46( 7 0.42)SST+18.21( 7 0.85)SSS 774 0.904

pCO2w( 7 2.6) = —439.5( 7 42.6)-2.44( 7 0.64)SST+25.27( 7 0.97)SSS 184 0.863

pCO2w(71.9) = —575.1(766.9)+7.66(71.14)SST+20.60(71.91)SSS 177 0.733

pCO2w(71.7) = —1344.6(739.8)+9.89(70.89)SST+41.37(71.00)SSS 182 0.959

pCO2w(72.5) = —1665.1(750.0)+9.23(70.48)SST+51.15(71.38)SSS 374 0.905

pCO2w( 7 2.5) = —525.5( 7 58.5)-15.27( 7 0.77)SST+38.91( 71.53)SSS 364 0.871

pCO2w(71.4) = —485.9(710.1)+4.89(70.36)SST+20.67(70.25)SSS 382 0.982

Figure 7: The pCO2w at average SST (29.5 °C) and SSS (34.29) in the western Pacific warm pool in 1987—2003. The solid line shows the linear long-term trend of pCO2w(t) - 1.3(t—1985)+344.0, where t is the elapsed time (year) since January 1985. The standard error of analysis was 5.6 matm, and the standard error for the coefficient was 0.3 matm/yr, intercept 5.6 matm and the correlation coefficient was 0.71.

Figure 7: The pCO2w at average SST (29.5 °C) and SSS (34.29) in the western Pacific warm pool in 1987—2003. The solid line shows the linear long-term trend of pCO2w(t) - 1.3(t—1985)+344.0, where t is the elapsed time (year) since January 1985. The standard error of analysis was 5.6 matm, and the standard error for the coefficient was 0.3 matm/yr, intercept 5.6 matm and the correlation coefficient was 0.71.

Figure 8: The growth rate ofpCO2w that was calculated using SST and SSS via equation (3). Solid circle shows the growth rate ofpCO2w at average SST and SSS.

Figure 8: The growth rate ofpCO2w that was calculated using SST and SSS via equation (3). Solid circle shows the growth rate ofpCO2w at average SST and SSS.

analysis given by equations (2) and (3). With respect to spatial distributions, the pCO2w can be calculated well with rms of 2-3 matm for the HNLC and the western Pacific warm pool. In this work, we examined the long-term trend ofpCO2w at given SST and NO3 in the HNLC region and SST and SSS

in the western Pacific warm pool, which might provide the pCO2w at the same conditions of thermodynamics, ocean dynamics and biological activity. At average values of SST and NO3 in the HNLC region and SST and SSS in the western equatorial Pacific warm pool observed from 1987 to 2003, the long-term trend ofpCO2w was calculated to be 1.4 7 0.5 matm/yr in the HNLC region and 1.370.3 matm/yr in the western Pacific warm pool. The growth rates of pCO2w for both regions were close to that in the core of upwelled water between 140 and 160°W (Feely et al., 1999; Feely et al., 2002) and pCO|ir (1.670.4matm/yr). The growth rates of pCO|ir was calculated from the atmospheric data from Christmas Island (1°420N, 157°100W) and container ships by NOAA/CMDL (Conway et al., 1994). The growth rate of pCO2w in the central and western equatorial Pacific and that of the core of upwelled water suggests a somewhat smaller uptake of anthropogenic CO2 after upwelling than other portions of the North Pacific (Sabine et al., 2004). Takahashi et al. (2003) showed that the pCO2w over the Pacific equatorial zone appears to have changed substantially during the past two decades in coincidence with the PDO phase shift that occurred between 1988 and 1992. They reported that after the PDO phase shift DpCO2 increased by 19 matm in the western area by the year 2001. This includes changes in carbonate system other than uptake of anthropogenic CO2, that is, ocean dynamics and/or biological activity. The extremely high growth rate of pCO2w (2.57 0.3 matm/yr) was reported at the time-series station ALOHA in the North Pacific subtropical gyre (Dore et al., 2003). Keeling et al. (2004) proposed that the large increase in pCO2w resulted mainly from the increased flow from northwestern waters with greater pCO2w and salinity related to a large-scale reorganization of the climate system over the North Pacific. Our data sets are mostly taken after the PDO phase shift (Takahashi et al., 2003) and include data from 5°S to 2°N, 137°E to 170°W in the western Pacific warm pool. In order to examine temporal and spatial variations in growth rate of pCO2w in the equatorial Pacific, it is further necessary to conduct precise and repeated measurements at least over a few decades.

3.3 Air-Sea CO2 Flux in the Equatorial Pacific

If the parameters SST, SSS, and NO3 are measured widely and remotely like SST and chlorophyll, we could easily interpolate/extrapolate the observed pCO2w data in the equatorial Pacific, which would allow us to evaluate the sea-air CO2 flux more precisely. Historically, the pCO2w-SST or DIC-SST relationship has been used in order to estimate temporal and spatial variations in pCO2w in the equatorial Pacific (Lee et al., 1998; Boutin et al., 1999; Loukos et al., 2000; Cosca et al., 2003; Ishii et al., 2003). For example, on the basis of the assumption that, either directly or indirectly, the thermodynamic, transport, and biological effects are correlated with temperature, Lee et al. (1998) examined the interannual variability of pCO2w using the pCO2w—SST relationship. Wanninkhof et al. (1996) and Feely et al. (1999, 2002) calculated CO2 flux along the cruise track by using observed pCOs2w and extrapolated it to the equatorial Pacific for specific latitude and longitude regions.

As discussed above, the pCO2w—SST relationship determined for a certain cruise cannot be applied for long-term changes in pCO2w (Dandonneau, 1995; Wanninkhof et al. 1996). The pCO2w—SST relationship determined by compiling data taken over several years gives us a spatial average estimate of pCO2w variations as a function of SST, because it is determined based on the concept that the observed pCO2w included a "random error.'' With respect to the global carbon cycle, what we would like to know is if the pCOs2w field including the "random error'' varied on time scale of ENSO cycle. The approach by Wanninkhof et al. (1996) and Feely et al. (1999, 2002) is to evaluate this pCOs2w field, though the pCOs2w is extrapolated without using parameters that relate to the carbon system.

In this work, we determined pCO2w—SST relationships on the basis of data collected within a few months to obtain pCOs2w distribution on the basis of MRI/JMA, NOAA/PMEL and NOAA/AOML data (Table 3). We divided the equatorial Pacific rather arbitrarily into regions in which a single pCO2w—SST relationship was applied (see Appendix). We obtained a pCO2w map (Fig. 9) by applying average SST during the observation period (ftp:// ftp.ncep.noaa.gov/pub/cmb/sst/oimonth). The CO2 transfer velocity was calculated by Wanninkhof (1992):

where Uav is the averaged wind speed during the observation, and Sc20 and Sc are the Schmidt number for CO2 at 20 °C and at SST. Data of averaged wind speed were acquired from the Japan Meteorological Agency (GANAL data sets).

Boutin et al. (1999) provided a time series of sea—air CO2 flux of 0.18 Pg C/yr between the equator and 5°S usingpCO2w-SST andpCO2w-SST anomaly relationships, which agreed well with that of Cosca et al. (2003), who gave the mean flux of 0.370.1 Pg/yr for an area that covers approximately half of the Pacific equatorial belt (90°W-165°E, 5°N-10°S) during 1985—2001. Loukos et al. (2000) found an average CO2 flux of 0.5 Pg C/yr for 1982—1993, while recently Ishii et al. (2003) found 0.4 Pg C/yr for 1990—2000.

Table 4 lists the air—sea CO2 flux obtained in this work. A large sea—air CO2 flux of 0.9 7 0.4 Pg C/yr was observed in January/February 2001, and

Jan.-Feb. 1998

-

0

-

10 ____

5

—330 Mc

10

10°S

Jan.-Feb. 1999

Jan.-Feb. 1999

Nov.-Dec. 1999

Nov.-Dec. 1999

Nov. 2001-Feb. 2002

Nov. 2001-Feb. 2002

Oct. 2002-Jan. 2003

140°E 170 160 130

Longitude

Figure 9: The distribution of DpCO2 in the equatorial Pacific in January/ February 1998, January/February 1999, November/December 1999, January/February 2001, November 2001/February 2002, and October 2002/Jan-uary 2003.

140°E 170 160 130

Longitude

Figure 9: The distribution of DpCO2 in the equatorial Pacific in January/ February 1998, January/February 1999, November/December 1999, January/February 2001, November 2001/February 2002, and October 2002/Jan-uary 2003.

Table 4: Estimated sea-to-air CO2 flux in the equatorial Pacific during the last 5 years

Year

Month

SOI

Region of interest

Area ( x 106

km2) ApCO|w(matm)

Annual flux (Pg C/yr)

Reference

1997/98

Spring-spring

10°N-10°S, 135°E-80°W

35

0.2 + 0.14

Feely et al., 2002

1998

10°N-10°S, 135°E-80°W

35

30

0.4 + 0.2

Feely et al., 2002

1998

Jan.-Feb.

-3.0

5°N-10°S, 140°E-90°W

24

12(18)

0.1 + 0.1(0.08 + 0.05)

This work

1999

Jan.-Feb.

1.4

5°N-10°S, 140°E-90°W

24

71(39)

0.6 + 0.3(0.2 + 0.1)

This work

1999

Nov.-Dec.

1.3

5°N-10°S, 140°E-90°W

24

64(43)

0.6 + 0.3(0.2 + 0.1)

This work

2001

Jan.-Feb.

1.3

5°N-10°S, 140°E-90°W

24

88(58)

0.9 + 0.4(0.3 + 0.1)

This work

2001-2002

Nov.-Feb.

0.2

5°N-10°S, 140°E-90°W

24

50(31)

0.5+0.3(0.1+0.1)

This work

2002-2003

Oct.-Jan.

-0.8

5°N-10°S, 140°E-90°W

24

42(18)

0.5 + 0.3(0.05 + 0.03)

This work

Note: DpCO2 and CO2 flux west of 160°W (central and western equatorial Pacific) are in parentheses after those of equatorial Pacific.

Note: DpCO2 and CO2 flux west of 160°W (central and western equatorial Pacific) are in parentheses after those of equatorial Pacific.

e a relatively low CO2 flux (~0.5PgC/yr) in November 2001/February 2002 and October 2002/January 2003. The lowest CO2 flux of 0.1PgC/yr was estimated for the period of January/February 1998, during the period of 1997/98 El Niño. Feely et al. (2002) estimated the sea-air CO2 flux of 0.2±0.14PgC/yr for the 1-year period from spring 1997 to spring 1998. The bimonthly sea-air CO2 flux listed in Table 3 varied more than those estimated by using a pCO2w(TCO2)-SST relationship over the same period. During 1997-1998, the growth rate of atmospheric CO2 increased steeply and reached a level larger than 3ppm/yr in the first half of 1998 and then decreased (http:// www.cmdl.noaa.gov/ccg/index.html). It is thus of particular importance to examine the CO2 flux on time scales of ENSO phenomena (probably shorter than a few months) to figure out the role of the equatorial Pacific in the current global carbon cycle. During the 2002/03 El Nino event (October 2002/January 2003), the pCO2w in the central and western equatorial Pacific decreased considerably, but in the eastern equatorial Pacific remained almost at the same level as during the previous non-El Nino period of November 2001/February 2002 (Fig. 10). This led to the relatively lower sea-air CO2 flux in the central and western equatorial Pacific (west of 160°W) during the 2002/03 El Niño period, which was comparable to that of 1997/98 El Nino. The remarkable contrast between the western and eastern Pacific is attributed to a deeper thermocline and weaker winds in the western Pacific, which causes pCO2w values to be nearly in equilibrium with the atmosphere and, hence, the lower sea-to-air CO2 fluxes. In a remarkable contrast, the eastern Pacific had much higher CO2 fluxes because the thermocline was almost as shallow as during non-El Nino periods, resulting in very high pCO2w values due to upwelling of relatively cold CO2-enriched waters from the equatorial undercurrent. Coupled with strong southeasterly winds, the resulting CO2 fluxes during the 2002/03 El Nino in the eastern equatorial Pacific were very similar to the fluxes observed during a non-El Nino period. The central and western equatorial Pacific (west of 160°W) showed relatively large variability of sea-air CO2 flux as pointed out by Inoue et al. (2002), though changes in sea-air CO2 flux would only affect the growth rate of atmospheric CO2 by about ~0.2ppm/yr. These results suggest the recent phase shift of the 1990s (Chavez et al., 2003; Takahashi et al., 2003) may have resulted in a net increase in the outgasing of CO2 at the equator, causing an enhancement of the global warming impacts of CO2 on the atmosphere.

Acknowledgements

We thank officers and crew of R/V Natsushima, R/V Kaiyo, and R/V Mirai belonging to the Japan Marine Science and Technology Center (JAMSTEC),

Jan.-Feb. 1998

Jan.-Feb. 1998

Jan.-Feb. 1999

Nov.-Dec. 1999

Nov.-Dec. 1999

170 160

Longitude

170 160

Longitude

Jan.-Feb. 1999

Jan.-Feb. 2001

Jan.-Feb. 2001

Nov. 2001-Feb. 2002

Oct. 2002-Jan. 2003

160 Longitude

Nov. 2001-Feb. 2002

Oct. 2002-Jan. 2003

160 Longitude

to o

Figure 10: The distribution of CO2 flux in the equatorial Pacific in January/February 1998, January/February 1999, November/December 1999, January/February 2001, November 2001/February 2002, and October 2002/ January 2003 (For colour version, see Colour Plate Section).

of R/V Hakuho-maru belonging to the Ocean Research Institute/University of Tokyo, of R/V Ryofu-maru belonging to the Japan Meteorological Agency and of the ships Malcolm Aldridge, the Discoverer, and the Ka'imimoana and Ron Brown of the National Oceanic and Atmospheric Administration (NOAA) of USA. RAF and RWalso thank the officers and crew of the NOAA ships Ronald H. Brown and Ka'imimoana for logistics support. HYI, MI, TK, and AM appreciate the staff of Marine Work Japan Ltd., Nippon Marine Enterprise Ltd., Global Ocean Development Inc., and Kansai Environmental Engineering Center Co. for their technical support on board and providing hydrographic and meteorological data sets. Hydrographic and meteorological data observed aboard the R/V Mirai are/will be available from Data Management Office of JAMSTEC and Japan Oceanographic Data Center. This work was supported by the grant from the "Global Carbon Cycle and Related Mapping based on Satellite Imaginary Program (GCMAPS)'' of Ministry of Education, Culture, Sports, Science and Technology, Japan. The Office of National Oceanic and Atmospheric Research of NOAA supported RAF and RW with project support from the Global Carbon Cycle Program under the leadership of Dr. Kathy Tedesco and Dr. Mike Johnson.

Was this article helpful?

0 0

Post a comment