Evaluation of Atmospheric Profile Retrieval Algorithm for GK2A Satellite with Dropsonde Observations

Abstract

The Korea's Geostationary Multi-Purpose Satellite-2A (GK2A) was launched in December 2018 with its main payload, Advanced Meteorological Imager (AMI). Using the 9 infrared channels of the AMI, an algorithm has been developed to retrieve vertical profiles of temperature and humidity for the clear-sky. The algorithm, named as AMI Atmospheric Profile (AAP), is based on an optimal estimation method with its a priori information from model forecasts. From the retrieved profiles, total precipitable water and atmospheric instability indices are estimated, while total column ozone is derived as a by-product. Comparisons of the AAP products and radiosonde observations over land surface stations show that the AAP humidity profiles improve the a priori information by about 4% of root mean square error (RMSE) between 100 and 1000 hPa. However, as most of radiosonde data used for the validation are also used for the preparation of the a priori data, comparisons with radiosonde data would not reveal full characteristics of the AAP algorithm. Therefore, current study evaluates the AAP using dropsonde data obtained over the ocean. Since the dropsonde data are not used for the preparation of the a priori data, they could be a truly independent reference data set. The validation results with the dropsonde data confirm that the AAP performance over the ocean is almost the same as that on land for the temperature profiles. In the case of the humidity, however, a priori improvements by AAP algorithm is much larger over the ocean than on land (the RMSE is improved by 11% between the surface and 400 hPa). The results also show that the performance of the retrieval algorithm under clear-sky conditions is similar to that at the cloud edges over the sea, suggesting the potential benefits of using AAP temperature and humidity profiles for real-time analysis of the atmospheric conditions over the ocean.