15:00 〜 15:15
[ACG50-05] Quantification of the factors contributing to decadal variations of air-sea CO2 flux in the North Pacific Ocean
キーワード：大気ー海洋間CO2フラックス、北太平洋、長期変動
We quantified the factors contributing to decadal variations of air-sea CO2 flux in the North Pacific Ocean using several air-sea CO2 flux data products. The North Pacific Ocean is one of the strongest sinks of atmospheric CO2 in the global ocean. The absorption mechanism is attributed to the decrease in the partial pressure of CO2 in seawater due to the changes in SST and high biological production, and to the seasonal wind in winter. This region also shows strong interannual variability like the polar and tropical upwelling regions. Analysis of this interannual variability is an important issue for predicting future changes in atmospheric CO2 uptake in the North Pacific Ocean.
The data products used in this study are based on oceanic CO2 observations obtained by research vessels and voluntary cargo ships and are available in global database SOCAT (Surface Ocean CO2 Atlas). The data products are the statistical processing of the SOCAT data for spatiotemporal homogeneity. We first used the data product by NIES (NIES-ml3), which is an ensemble average of the processing from three machine learning, random forest, gradient boost machine, and feed-forward neural network.
The air-sea CO2 flux is mainly determined by the CO2 fugacity difference between the atmosphere and seawater (ΔfCO2) and the wind speed over the sea surface. In this study, we performed Reynolds decomposition of CO2 flux data based on these two parameters and quantified the spatiotemporal fluctuation of CO2 flux anomaly and the contributions of ΔfCO2 and wind speed.
Figure 1 shows the spatial distribution of the intensities of CO2 flux anomaly and the contributions in the North Pacific Ocean from 1982 to 2023. The high intensity of CO2 flux anomaly was confirmed in the area north of 30°N, and the intensity of ΔpCO2 contribution was high north of 40°N, while the intensity of wind contribution was high south of 40°N. The intensity of cross-term, which is the product of ΔfCO2 and wind speed components, was found to be small and almost negligible in the analysis.
The time-series variations of the CO2 flux anomaly and the contributions show a 4–5-year cycle fluctuations (Figure 2). Prior to 2015, ΔfCO2 contribution tended to be higher, but since then that of wind speed appeared to be higher. In the period from 1995 to 2010, the contributions were offset, and the CO2 flux anomaly was close to zero. The contribution of ΔfCO2 is expected to be affected by oceanic fluctuations in the sub-arctic Pacific, since the area north of 40°N is north of the Kuroshio Extension. The sub-arctic Pacific is strongly influenced by upwelling in the Bering Sea area and freshwater inflow from the Amur River, and comparisons with these phenomena will be an issue in the future. On the other hand, the contribution of wind velocity is expected to be affected by meteorological fluctuations such as ENSO and PDO with a few-decadal cycles. The presentation will include a quantitative comparison with these phenomena as well as a comparative validation with the results of other data products.
The data products used in this study are based on oceanic CO2 observations obtained by research vessels and voluntary cargo ships and are available in global database SOCAT (Surface Ocean CO2 Atlas). The data products are the statistical processing of the SOCAT data for spatiotemporal homogeneity. We first used the data product by NIES (NIES-ml3), which is an ensemble average of the processing from three machine learning, random forest, gradient boost machine, and feed-forward neural network.
The air-sea CO2 flux is mainly determined by the CO2 fugacity difference between the atmosphere and seawater (ΔfCO2) and the wind speed over the sea surface. In this study, we performed Reynolds decomposition of CO2 flux data based on these two parameters and quantified the spatiotemporal fluctuation of CO2 flux anomaly and the contributions of ΔfCO2 and wind speed.
Figure 1 shows the spatial distribution of the intensities of CO2 flux anomaly and the contributions in the North Pacific Ocean from 1982 to 2023. The high intensity of CO2 flux anomaly was confirmed in the area north of 30°N, and the intensity of ΔpCO2 contribution was high north of 40°N, while the intensity of wind contribution was high south of 40°N. The intensity of cross-term, which is the product of ΔfCO2 and wind speed components, was found to be small and almost negligible in the analysis.
The time-series variations of the CO2 flux anomaly and the contributions show a 4–5-year cycle fluctuations (Figure 2). Prior to 2015, ΔfCO2 contribution tended to be higher, but since then that of wind speed appeared to be higher. In the period from 1995 to 2010, the contributions were offset, and the CO2 flux anomaly was close to zero. The contribution of ΔfCO2 is expected to be affected by oceanic fluctuations in the sub-arctic Pacific, since the area north of 40°N is north of the Kuroshio Extension. The sub-arctic Pacific is strongly influenced by upwelling in the Bering Sea area and freshwater inflow from the Amur River, and comparisons with these phenomena will be an issue in the future. On the other hand, the contribution of wind velocity is expected to be affected by meteorological fluctuations such as ENSO and PDO with a few-decadal cycles. The presentation will include a quantitative comparison with these phenomena as well as a comparative validation with the results of other data products.