Evaluation of GPM- and Model-Based Precipitation over Taiwan

Acquiring real-time precipitation is vital for most hydro-meteorological applications, including estimating river discharge and flood conditions. Owing to various advances in satellite, remote sensing, and modeling techniques, precipitation estimates can be acquired from spaceborne data or model simulations; such gridded, wide-coverage precipitation estimates seem particularly indispensable to regions lacking ground observations. Since 2014, the first release of the Integrated Multi-satellitE Retrievals (IMERG) data for Global Precipitation Measurement (GPM), evaluation studies on GPM-based products and their utility in hydrological applications for Taiwan were rare; however, complex terrain and mixed weather patterns that induce considerable rainfall variations over Taiwan should only promote comprehensive evaluation and improved usage of satellite- and model-based precipitation. Existing literature has highlighted the complementary performance of GPM- and model-based precipitation at the continental scale: The former (latter) product performs better during the warm (cold) seasons. This study aims at verifying the complementary behavior in Taiwan. GPM-based products include GPM-IMERG early- and final-runs and the Global Satellite Mapping of Precipitation (GSMaP) data, whereas Model-based precipitation is obtained from the Weather Research and Forecasting (WRF) Model. Evaluation of the two classes of products is made relative to a reference high-resolution gridded data set obtained from the Taiwan Climate Change Projection and Information Platform (TCCIP) (1-km/daily) as well as a dense network of rain gauges during multiple seasons and selective extreme events (e.g., typhoons and convective storms) from 2015 to 2018. Further, quality of river discharge and floods driven by these precipitation products will be assessed using hydrological and hydrodynamic models.