Comparative Analysis of the Observation Bias and Error Characteristics of AGRI and AHI Data for Land Areas in East Asia

Abstract

Observation bias and error characteristics are the preconditions for the effective assimilation of observation data. In this paper, the bias and error characteristics of the AHI (Advanced Himawari Imager (AHI) and AGRI (Multi-Channel Advanced Geostationary Orbit Radiation Imager (AGRI) data are compared and analyzed, with an emphasis on the observations obtained from land areas. The results show that the observation errors of the two instruments for the ocean area have a good channel consistency over ocean areas, which are all with errors of about 0.6 K; however, the bias and error are significantly affected by the land-surface types and terrain heights over land. For most of the AHI channels, the bias in urban-area bias is smaller than that of those of other surface types, while that of the AGRI data exhibits just the opposite trend, with obviously larger biases in urban areas. However, the observation errors of these two instruments in urban areas are significantly smaller than those of other surface types. The biases of the two instruments do not extensively change much with the terrain height, only slightly decreasing when the height is above 1000 m; however, the observation errors increase obviously with the increase of terrain heights. The difference between the two instruments is that the observation error of the AHI data tends to be stable and stabilizes above 1000 m, while that of the AGRI data is relatively stable below 500 m. The observation errors of the CO2-sensitive channels of the two instruments over the land areas are obviously smaller than those of other near-surface channels, which may indicate that the data obtained in these two CO2 channels have good application prospects for assimilation over land.