ABSTRACT:Precipitation is the most important weather parameter in the Philippines. Made up of more than 7100 islands, the Philippine archipelago is an agricultural country that depends on rain-fed crops. Located in the western rim of the North West Pacific Ocean, this tropical island country is very vulnerable to tropical cyclones that lead to severe flooding events. Recently, satellite-based precipitation estimates have improved significantly and can serve as alternatives to ground-based observations. These data can be used to fill data gaps not only for climatic studies, but can also be utilized for disaster risk reduction and management activities. This study characterized the statistical errors of daily precipitation from four satellite-based rainfall products from (1) the Tropical Rainfall Measuring Mission (TRMM), (2) the CPC Morphing technique (CMORPH) of NOAA and (3) the Global Satellite Mapping of Precipitation (GSMAP) and (4) Precipitation Estimation from Remotely Sensed information using Artificial Neural Networks (PERSIANN). Precipitation data were compared to 52 synoptic weather stations located all over the Philippines. Results show GSMAP to have over all lower bias and CMORPH with lowest Mean Absolute Error (MAE) and Root Mean Square Error (RMSE). In addition, a dichotomous rainfall test reveals GSMAP and CMORPH have low Proportion Correct (PC) for convective and stratiform rainclouds, respectively. TRMM consistently showed high PC for almost all raincloud types. Moreover, all four satellite precipitation showed high Correct Negatives (CN) values for the north-western part of the country during the North-East monsoon and spring monsoonal transition periods. Corresponding author

ABSTRACT:Precipitation is the most important weather parameter in the Philippines. Made up of more than 7100 islands, the Philippine archipelago is an agricultural country that depends on rain-fed crops. Located in the western rim of the North West Pacific Ocean, this tropical island country is very vulnerable to tropical cyclones that lead to severe flooding events. Recently, satellite-based precipitation estimates have improved significantly and can serve as alternatives to ground-based observations. These data can be used to fill data gaps not only for climatic studies, but can also be utilized for disaster risk reduction and management activities. This study characterized the statistical errors of daily precipitation from four satellite-based rainfall products from (1) the Tropical Rainfall Measuring Mission (TRMM), (2) the CPC Morphing technique (CMORPH) of NOAA and (3) the Global Satellite Mapping of Precipitation (GSMAP) and (4) Precipitation Estimation from Remotely Sensed information using Artificial Neural Networks (PERSIANN). Precipitation data were compared to 52 synoptic weather stations located all over the Philippines. Results show GSMAP to have over all lower bias and CMORPH with lowest Mean Absolute Error (MAE) and Root Mean Square Error (RMSE). In addition, a dichotomous rainfall test reveals GSMAP and CMORPH have low Proportion Correct (PC) for convective and stratiform rainclouds, respectively. TRMM consistently showed high PC for almost all raincloud types. Moreover, all four satellite precipitation showed high Correct Negatives (CN) values for the north-western part of the country during the North-East monsoon and spring monsoonal transition periods. Corresponding author