All‐sky infrared radiance assimilation of a geostationary satellite in the Japan Meteorological Agency's global system

Abstract

All‐sky assimilation of infrared (IR) radiances has been developed for water vapor bands of the geostationary satellite Himawari‐8 in the operational global data assimilation system. Cloud‐dependent quality control, bias correction, and observation error modeling are essential developments to effectively utilize the all‐sky radiances (ASRs). ASR assimilation increases the assimilated number of observations by 2.8 times and improves the coverage relative to the traditional clear‐sky radiance (CSR) assimilation. The additional observations better alleviate model dry biases in the middle and upper tropospheric humidity. ASR assimilation brings statistically significant improvements in the background (first guess) in humidity, temperature, and wind over the CSR assimilation. It also better improves short‐range forecasts of the middle and upper tropospheric temperature and humidity up to day 3 in the Tropics. A mixed impact in the stratospheric temperature is under investigation. The impacts of various aspects of the ASR assimilation configuration are evaluated with sensitivity assimilation experiments. The interband correlation and cloud‐dependent standard deviation of the observation error are crucial, whereas the cloud dependency of the correlation is not so important. Although ASRs at a single band were assimilated in many previous studies targeting severe weather using research‐based regional assimilation systems due to decreasing independent information in the presence of clouds, they are distinctly inferior to not only ASRs at multiple bands but also CSRs at multiple bands in a global data assimilation system that contains fewer cloud‐affected scenes. The cloud‐dependent bias correction predictors are essential in the presence of observation‐minus‐background bias that increases with cloud effects.

The Impact of Assimilating Cirrus‐Effected Infrared Satellite Radiance From the FY‐4A AGRI on Water Vapor Analysis and Rainstorm Forecasting

“…Previous works by Okamoto et al. (2014, 2019, 2023) and Okamoto (2017) are good examples. They found that cloud models are not very accurate for high clouds (with improved cloud ice optical properties (Vidot et al., 2015)), which may explain the large observation‐minus‐background (O‐B) standard deviation and the insufficiently low brightness temperature in the intertropical convergence zone.…”

“…In practical use, there are still significant challenges associated with all-sky radiance (ASR) assimilation, mainly due to the extreme uncertainties of the simulated clouds in the background field of numerical weather prediction (NWP) models, and the limited ability of fast RTM to simulate cloud structures and physical properties, especially for vertically complex clouds (Li et al, 2016(Li et al, , 2022. Previous works by Okamoto et al (2014Okamoto et al ( , 2019Okamoto et al ( , 2023 and Okamoto (2017) are good examples. They found that cloud models are not very accurate for high clouds (with improved cloud ice optical properties (Vidot et al, 2015)), which may explain the large observation-minusbackground (O-B) standard deviation and the insufficiently low brightness temperature in the intertropical convergence zone.…”

Enhancing Severe Weather Prediction With Microwave All‐Sky Radiance Assimilation: The 10 August 2020 Midwest Derecho

“…However, the radiances affected by hydrometeors contain information about clouds and precipitation, and they are expected to bring additional benefits to NWP. Therefore, effective assimilation of all‐sky radiances—both the clear‐sky radiances and the cloud‐ and precipitation‐affected radiances—has been the goal of many recent studies and is being actively developed at major NWP centers (e.g., Eyre et al., 2022; Geer et al., 2018, 2019; Gustafsson et al., 2018; Li et al., 2022; Okamoto et al., 2023; Shahabadi & Buehner, 2021; Tong et al., 2020; Y. Zhu et al., 2016, 2019).…”