3:00 PM - 3:15 PM
[ACG39-06] Toward near-real-time delivery of global 1°x1° sea-air CO2 flux product
Keywords:carbon cycle, sea-air CO2 flux, G3W, operational GHG monitoring
Recently, governments and policymakers around the world have come to recognize the ocean’s critical role in the global carbon cycle originating from its capacity to absorb anthropogenically emitted CO2. Under such a condition, The World Meteorological Organization (WMO) newly founded a program relevant to global carbon cycle monitoring, Global Greenhouse Gas Watch (G3W), thereby building an operational GHG emission/uptake monitoring system. In its implementation plan published in 2024 (WMO, 2024), G3W recommended a routine supply of global 1°x1° resolution earth surface–atmosphere CO2 flux with maximum a delay of one month. Nowadays, most surface CO2 flux products are provided from June to September of the following year to contribute to the Global Carbon Budget activity, whereas this schedule seems tight due to observation data availability and computational resources.
Japan Meteorological Agency (JMA) releases an annual update of the sea-air CO2 flux product based on the method using surface ocean CO2 observation (https://www.data.jma.go.jp/kaiyou/english/co2_flux/co2_flux_en.html; Iida et al. 2021). Materials necessary to calculate the flux are 1) results of machine learning using observed surface ocean carbonate variables and 2) gridded global surface data; salinity, temperature, chlorophyll-a concentration, surface pressure, wind speed, and atmospheric pCO2 (pCO2a), obtained from the satellite observation and/or reanalysis. Surface ocean pCO2 (pCO2s) measurement data are updated annually owing to the efforts of the SOCAT group (Bakker et al. 2016), and we can use the pCO2s in the last year every June. At that point, we can perform machine learning using the latest data. By extending the results of machine learning and inputting the newest gridded surface data, we can routinely calculate the last month 1°x1° gridded surface ocean carbonate variables and sea-air CO2 flux within 1 week of latency. To use only recent data in the SOCAT database in the learning process is necessary to capture the latest ocean state, and then more likely distributions of surface ocean pCO2 can be obtained. In this way, we calculated the year 2024 CO2 flux and obtained a globally integrated anthropogenic flux value of 3.1 ± 0.6 PgC/yr. The calculated sea-air CO2 flux can be immediately used for atmospheric inversion systems to obtain global earth surface–atmosphere CO2 flux.
Japan Meteorological Agency (JMA) releases an annual update of the sea-air CO2 flux product based on the method using surface ocean CO2 observation (https://www.data.jma.go.jp/kaiyou/english/co2_flux/co2_flux_en.html; Iida et al. 2021). Materials necessary to calculate the flux are 1) results of machine learning using observed surface ocean carbonate variables and 2) gridded global surface data; salinity, temperature, chlorophyll-a concentration, surface pressure, wind speed, and atmospheric pCO2 (pCO2a), obtained from the satellite observation and/or reanalysis. Surface ocean pCO2 (pCO2s) measurement data are updated annually owing to the efforts of the SOCAT group (Bakker et al. 2016), and we can use the pCO2s in the last year every June. At that point, we can perform machine learning using the latest data. By extending the results of machine learning and inputting the newest gridded surface data, we can routinely calculate the last month 1°x1° gridded surface ocean carbonate variables and sea-air CO2 flux within 1 week of latency. To use only recent data in the SOCAT database in the learning process is necessary to capture the latest ocean state, and then more likely distributions of surface ocean pCO2 can be obtained. In this way, we calculated the year 2024 CO2 flux and obtained a globally integrated anthropogenic flux value of 3.1 ± 0.6 PgC/yr. The calculated sea-air CO2 flux can be immediately used for atmospheric inversion systems to obtain global earth surface–atmosphere CO2 flux.