暴雨灾害 (Feb 2024)

Verification and analysis of IMERG during the precipitation process of Typhoon In-Fa in 2021

  • Hongbing LI,
  • Xuebin MEI,
  • Yu XIA

DOI
https://doi.org/10.12406/byzh.2023-046
Journal volume & issue
Vol. 43, no. 1
pp. 73 – 83

Abstract

Read online

To study the applicability of the Integrated Multi-satellitE Retrievals for GPM (IMERG) precipitation products to typhoon precipitation, taking Typhoon In-Fa in 2021 as an example, this paper evaluates the IMERG precipitation estimation capabilities and the cause of errors using hourly precipitation data from 563 national meteorological stations in the eastern region of China, FY-4A AGRI infrared brightness temperature data, and Global Precipitation Measurement Dual-Frequency Precipitation Radar (GPM DPR) data. The results indicate that IMERG precipitation estimation shows good spatial consistency compared to observations, with a correlation coefficient of 0.75. IMERG precipitation products exhibit an overall consistent underestimation, the performance varies with different precipitation intensities, As precipitation intensity increases, the degree of underestimation also grows. IMERG products slightly overestimate weak precipitation. For precipitation below 10 mm·h-1, IMERG estimations are slightly larger than the observations, but for precipitation above 40 mm·h-1, IMERG estimations are significantly smaller than the observations. However, an overall underestimation of the IMERG products can be indicated that the line was below the diagonal, which calculated by the least square method between the IMERG estimations and observations. The error of infrared-derived precipitation within IMERG is larger than that from microwave retrieval, with a root mean square error of7.8 mm for hourly precipitation, and it overlooks warm cloud precipitation in Typhoon In-Fa. The accuracy of microwave retrieval for different types of precipitation also varies, which shows better performance in stratiform precipitation with concentrated particle number concentration (dBNw) and particle effective radius (D0) compared to convective precipitation with dispersed dBNw and D0. The high-intensity precipitation estimates retrieved from both infrared and microwave are weaker than those of observations, resulting in underestimated precipitation by IMERG precipitation estimations.

Keywords