[A-CG36] 衛星による地球環境観測

[ACG36-02] Validation of SGLI/GCOM-C cloud products using surface and MODIS observation data

*Pradeep Khatri¹、Tadahiro Hayasaka¹、Hitoshi Irie²、Tamio Takamura² (1.Center for Atmospheric and Oceanic Studies,Tohoku University、2.Center for Environmental Remote Sensing, Chiba University )

キーワード：SGLI, validation , sky radiometer, global radiative flux, MODIS

The recent version 2.0 cloud properties (cloud optical thickness COT and cloud-particle effective radius CER) of both water and ice clouds obtained from the Second-generation Global Imager (SGLI) aboard the Global Change Observation Mission-Climate (GCOM-C, “Shikisai” in Japanese) were validated using sky radiometer cloud properties and global radiative flux data observed at different SKYNET sites within Japan (Chiba, Hedo-misaki, Fukuejima, Miyakojima, and Sendai) for the period of January-November, 2020. Accuracies of SGLI cloud properties were further confirmed by comparing them with MODIS cloud properties observed for the period of June-July, 2020 over the region of 20⁰N-40⁰N and 106⁰E-156⁰E. The SGLI COTs agreed quite well with sky radiometer (r = 0.73 and MBE = 1) and MODIS (r >= 0.97 and MBE < 2 over land and ocean) values by further reproducing global radiative fluxes modeled using SGLI cloud properties comparable to observed values (r = 0.85 and MBE = 25 Wm^-2), where r and MBE are correlation coefficient and absolute value of mean bias error, respectively. The comparison between SGLI and sky radiometer based CER is relatively poor (r=0.12, MBE = 1.5 µm), possibly due to the effect of inhomogeneous vertical profile of CER. But, MODIS and SLGI CERs agreed very well (r >= 0.93 and MBE < 0.5 µm). On the other hand, the agreement for ice cloud is relatively poor: r = 0.5 and MBE = 4.5 for comparison with sky radiometer COTs; r >=0.97 and MBE < 1 for comparison with MODIS COTs over land and ocean; r >=0.73 and RE < 8 µm for comparison with MODIS CERs over land and ocean; r =0.89 and RE = 16 Wm^-2 for comparison of modeled and observed global radiative fluxes at SKYNET sites. The relatively poor agreement for ice cloud properties is likely due to the difference in cloud phase detection method and single scattering properties for ice cloud in different cloud retrieval methods, suggesting the necessity of more validation studies for ice clouds in future.