Assessing the performance of GSMaP and IMERG in representing the diurnal cycle of precipitation in the Philippines during the southwest monsoon season

Abstract

The Philippines faces the challenge of having a limited number of rain-gauge stations, which are a vital source of observation data. Satellite-based precipitation data is a viable alternative; however, it is necessary to assess the strengths and weaknesses of these products over various regions in the Philippines. This study analyzes the performance of two leading products from the Global Precipitation Measurement (GPM) mission, namely: GSMaP_G v08 and IMERG_F v07. In particular, the diurnal cycle of precipitation in the Philippines during the southwest monsoon season (May to September) from 2013 to 2018 is assessed, in terms of precipitation amount (PA), frequency (PF) and intensity (PI), which has not been done in previous works. Results show that GSMaP_G outperforms IMERG_F in capturing the diurnal cycle for PA and PF, whereas IMERG_F better captures PI as GSMaP_G tends to underestimate PI consistently. In terms of timing, peak values for PA (PPA) and PF (PPF) have similar patterns for both station data and satellite products where these events develop during the afternoon to evening hours (15 to 18 Local Standard Time). Further examination of the possible drivers of these observed patterns suggests that for both station data and IMERG_F, short-duration events have the highest frequency contribution (≥ 60 %), while GSMaP_G has higher contribution (≥ 40 %) from long-duration precipitation events. However, in analyzing the contributions where intensity is considered, both station data and satellite products were observed to be dominated by short-duration light precipitation events (≥ 60 %). These results indicate that during the southwest monsoon season in the Philippines, the main drivers for PA and PF are short-duration light precipitation events, whereas PI may be driven more by localized rain showers. This may indicate that GSMaP_G is more suitable for examining hourly precipitation events that are more frequent but have a lower intensity, whereas IMERG_F may prove to be more useful when analyzing hourly precipitation events with higher intensities such as localized rain showers.