Preliminary Evaluation of GPM‐IMERG Rainfall Estimates Over Three Distinct Climate Zones With APHRODITE

摘要

Evaluation of Global Precipitation Measurement‐Integrated Multi‐satellitE Retrieval for GPM (GPM‐IMERG) final precipitation product is performed over Japan, Nepal, and Philippines regions against further improved APHRODITE‐2 V1801R1 product. The evolution is carried out for nearly two consecutive years 2014–2015. Various qualitative and quantitative statistical indices such as mean bias, root‐mean‐square error, correlation coefficient, false alarming ratio, and probability of detection are considered to evaluate GPM‐IMERG precipitation estimates with APHRODITE‐2. Intraseasonal variability of two products is shown to explore the seasonal dependency of GPM‐IMERG performance. The performance of GPM‐IMERG research product with respect to rainfall intensity is shown by the cumulative probability distribution of target and reference data sets. Percentile‐based statistics is implemented for evaluating the advantages of GPM‐IMERG over Tropical Rainfall Measuring Mission‐3B42 while detecting the light and heavy rainfall events during wet/dry seasons. The overall performance of GPM‐IMERG seems to be good over Japan followed by Philippines and Nepal regions. This feature is clearly evidenced in terms of mean bias, root‐mean‐square error, and correlation magnitudes over three regions. GPM‐IMERG shows ability to follow the intraseasonal variability as shown by APHRODITE‐2 product with minor differences observed in precipitation maximum values during rainy season. Good agreement is seen between GPM‐IMERG and APHRODITE‐2 at different rainfall intensities except underestimation during heavy rainfall events. GPM‐IMERG seems to be improved in detecting light/heavy rainfall event magnitude than TRM‐3B42. However, the performance of both data sets encountered clear dependency on seasons.