Validating infrared-based rainfall retrieval algorithms with 1-minute spatially dense raingage measurements over Korean peninsula

摘要

This study compares and contrasts six infrared-based satellite rain estimation techniques and their validation during a 2-month period from June 20-August 20, 1998 over the Korean peninsula. Two probability matching techniques (PMM1, PMM2), a look-up table technique (LUT), a convective-stratiform technique (CST), the Negri-Adler-Wetzel technique (NAWT), and the Arkin technique (ARKT) are applied to hourly infrared GMS imagery. Retrieved rainrates are compared against one-minute reporting raingage observations from the dense Automated Weather Station (AWS) network of Korea. The high spatial resolution and fine temporal resolution of the AWS measurements provide a unique and effective means to validate rain estimates derived from instantaneous spacemeasurements, which is a main scientific focus of this study. Validation results indicate that all techniques exhibit better performance for more evenly spread rain events while exhibiting lesser performance for weak and sporadic rains for which validation sampling becomes more of a problem. Validation statistics show that climatologically-local techniques such as the PMM and LUT algorithms perform better than techniques developed in climatologically different regimes, indicating the well-known dependence of rain physics on the immediate environment. Nevertheless, the validation results suggest how the rain determination parameters including attributed rain and threshold brightness temperature could be optimized locally before application. As others have found, the most difficult problem with satellite infrared techniques is in the detection and quantification of heavy rainfall events arising from uncertainties in discriminating non-precipitating anvil clouds from convective clouds. However, for theset of algorithms under examination here, given the sharp resolution of the validation measurements, it is evident that the CST algorithm exhibits superior performance in differentiating between non-precipitating anvil and heavy rain.