Development of Level-2 Products for the Infrared Sounder on the Himawari-10

JMA/MSC plans to start operating the geostationary meteorological satellite Himawari-10 in Japanese fiscal year 2029. Himawari-10 will be equipped with an infrared sounder (GHMS: Geostationary HiMawari Sounder) in addition to the imager that previous series have been equipped with.

The infrared sounder on the meteorological satellite has a much higher spectral resolution than the imager, so it is possible to retrieve information on the vertical profile of the Earth's atmosphere and the state of the Earth's surface from the infrared sounder data.

JMA/MSC has begun investigating and developing methods to retrieve geophysical quantities from the GHMS. Currently, the most widely used methods for retrieving physical quantities from infrared sounder data are the maximum likelihood estimation method based on the radiative transfer process and an empirical method that makes extensive use of threshold processing and JMA/MSC is trying to use an AI/ML model that further develops the latter method to retrieve cloud masks, cloud types, cloud heights, vertical profiles of temperature and water vapor, and precipitable water. Since real GHMS data is not available at this stage, we are investigating pairs of GHMS pseudo-observation data and ERA5 physical quantities calculated by radiative transfer calculations from ERA5 reanalysis data as training data.

In this presentation, we will introduce the accuracy and method of retrieving vertical profiles of temperature and water vapor, cloud masks, and precipitable water using AI/ML methods with the above dataset as training data, as well as current challenges and future plans.