[A-CG37] Satellite Earth Environment Observation

[ACG37-11] Improvement in global all-sky infrared radiance assimilation and assessments of its procedures

*Kozo OKAMOTO¹, Toshiyuki Ishibashi¹, Izumi Okabe¹ (1.Meteorological Research Institute)

Keywords:satellite data assimilation, geostationary satellite, all-sky radiance assimilation

Infrared all-sky radiance (ASR) assimilation is believed to be more beneficial than traditional clear-sky radiance (CSR) assimilation because it enhances the observation coverage, reduces dry bias due to a sampling error, and extracts more observation information in meteorologically sensitive regions. Despite many challenges, we successfully developed ASR assimilation for Himawari-8 in a global data assimilation system. Essential assimilation procedures are cloud-dependent quality control (QC), bias correction (BC), and observation error (OE) models. The ASR assimilation brought better forecast skills than CSR assimilation by modestly utilizing cloud-affected radiances in low and thin ice cloud conditions in addition to CSRs.

We also separately assessed the impacts of several settings of the assimilation procedures using data assimilation cycle experiments. Important findings are as follows: (1) Single-band ASR assimilation that was often employed in previous studies is inferior to multiband CSR assimilation, not to mention multi-band ASR assimilation (2) OE correlation and cloud-dependent OE standard deviation are important, but cloud-dependency of OE correlation structure is not so much, (3) cloud-dependent BC predictors are essential in the presence of (negative) biases in observation-minus-simulation.

We started extending the development to other satellites such as GOES and MSG. The results of assimilating these new developments will also be presented.