Improvements in rainfall estimation over Bangkok, Thailand by merging satellite, radar, and gauge rainfall datasets with the geostatistical method

Abstract

Bangkok is located in a low land area, and floods frequently occur from rainfall, river discharge, and tides. High-accuracy rainfall data are needed to achieve high-accuracy flood predictions from hydrological models. The main objective of this study is to establish a method that improves the accuracy of precipitation estimates by merging rainfall from three sources: an infrared channel from the Himawari-8 satellite, rain gauges, and ground-based radar observations. This study applied cloud classification and bias correction using rain gauges to discriminate these errors. The bias factors were interpolated using the ordinary kriging (OK) method to fill in the areas of estimated rainfall where no rain gauge was available. The results show that bias correction improved the accuracy of radar and Himawari-8 rainfall estimates before their combination. The merged algorithm was then adopted to produce hourly merged rainfall products (GSR). Compared to the initial estimated product, the GSR is significantly more accurate. The merging algorithm increases the spatial resolution and quality of rainfall estimates and is simple to use. Furthermore, these findings not only reveal the potential and limitations of the merged algorithm but also provide useful information for future retrieval algorithm enhancement. © 2023 The Author(s). Published by Taylor & Francis Group and Science Press on behalf of the International Society for Digital Earth, supported by the International Research Center of Big Data for Sustainable Development Goals, and CASEarth Strategic Priority Research Programme.