Accounting of far-field CO₂ contribution to XCO₂ calculation using a high spatial resolution regional model simulation

抄録

This study simulated carbon dioxide (CO₂) using regional Weather Research and Forecasting coupled with greenhouse gas modules (WRF-GHG) based on a central grid over Japan and at 27 km spatial resolution for the year 2019. We analyzed the Total Carbon Column Observing Network (TCCON) total column of CO₂ dry air mole fraction (XCO₂) using both global and regional modeling frameworks. XCO₂ was found to be significantly influenced by the CO₂ concentration at higher atmospheric pressure levels (> 400 hPa). We made use of the global Model for Interdisciplinary Research on Climate, version 4.0 based Atmospheric Chemistry-Transport model (MIROC4-ACTM) with a well-resolved stratosphere to better represent variabilities in XCO₂. We analyzed observations from three TCCON sites over Japan: Saga (130.3^oE, 33.2^oN), Tsukuba (140.1^oE, 36.0^oN), and Rikubetsu (143.7^oE, 43.4^oN), finding that correlation improved between observed and model-simulated XCO₂ profiles by using CO₂ concentration profile data produced by a hybrid model that combines WRF-GHG and MIROC4-ACTM; we used MIROC4-ACTM CO₂ concentrations at higher altitudes (< 400 hPa) with the WRF-GHG CO₂ output at lower altitudes. The correlation improvement between observed and simulated XCO₂ concentrations was most prominent over Saga (~35 %), which is near high CO₂ emission/sink regions such as China, Korea and southeast Asia. WRF-GHG simulations show a significant underestimation of XCO₂ over Saga during April and May of 2019. An analysis of WRF-GHG CO₂ spatial plots shows a tropopause fold that brings more depleted CO₂ air to Saga through stratosphere-troposphere exchange (STE) than suggested by the observations. The analysis was also conducted with XCO₂ data from Orbiting Carbon Observatory (OCO)-2 satellite observations, WRF-GHG, MIROC4-ACTM, and the hybrid model for the initial four and six months. The findings indicate that during the initial six months, WRF-GHG slightly underperformed compared to MIROC4-ACTM XCO₂. This suggests the need for precise tuning of the land biosphere in WRF-GHG, as the inclusion of the active land-biospheric period appears to deteriorate XCO₂ calculations from WRF-GHG.