RAINFALL DETECTION USING THE INFRARED BANDS OF HIMAWARI-8 ADVANCED HIMAWARIIMAGER IN THE PHILIPPINES

摘要

This study examined the utilization of infrared bands (IR) from Himawari-8 Advanced Himawari Imager (AHI) for rainfall area detection in the Philippines. The parameters used include the IR brightness temperature (BT) at 10.4 microns (T10.4) and several IR BT differences (BTD), which were inferred as proxies for cloud properties such as cloud top height, cloud optical thickness, cloud water path, and cloud phase. These cloud properties were used to create a Probability of Rain (POR) look-up table (LUT) together with processed horizontal reflectivity radar data gathered on rainy periods of July to September 2015. Combinations of BTD's for POR LUT construction includes the simple BTD's, BTD's based on cloud property, and selected BTD's. Rain detection assessment using different skill scores showed an increased performance of POR LUT with selected BTD combinations compared to with simple BTD's in training and testing stage. This is attributed to the increase in the sensitivity of detecting rain due to additional BTD. In visual comparison with Global Satellite Mapping of Precipitation (GSMaP) hourly rainfall maps, the POR LUT dominated by cloud-top height produced rain maps with high similarity. Prominent thin cirrus clouds and cold cloud dense overcast of the typhoons in testing stage were successfully classified as no-rain and rain, respectively. However, high false alarm rates were obtained when compared to radar data which means an over estimation of the method on rain detection. This is due to the weakness of weather radar on detecting rainfall in partially or fully blocked areas by mountains present in the domain, and the high variability of rain occurrences. The POR LUT using the selected BTD combinations dominated by cloud-top height is a viable method for satellite-based rainfall rate estimation due to its superiority to simple BTD combinations.