Hydrometeorological Assessment of Satellite and Model Precipitation Products over Taiwan

Satellite and model precipitation such as the Global Precipitation Measurement (GPM) data are valuable in hydrometeorological applications. This study investigates the performance of various satellite and model precipitation products in Taiwan from 2015 to 2017, including data derived from the Integrated Multi-satellitE Retrievals for GPM Early and Final Runs (IMERG_E and IMERG_F), Global Satellite Mapping of Precipitation_near-real-time (GSMaP_NRT), and the Weather Research and Forecasting (WRF) model. We assess these products by comparing them against data collected from 304 surface stations and gauge-based gridded data. Our assessment emphasizes factors influential in precipitation estimation, such as season, temperature, elevation, and extreme event. Further, we assess the hydrological response to each precipitation product via continuous flow simulation in two selected watersheds. The results indicate that the performance of these precipitation products is subject to seasonal and regional variations. The satellite products (i.e., IMERG and GSMaP) perform better than the model (i.e., WRF) in the warm season and vice versa in the cold season, most apparently in northern Taiwan. For selected extreme events, WRF can simulate better rainfall amount and distribution. The seasonal and regional variations in precipitation estimation are also reflected in flow simulation: IMERG in general produces the most rational flow simulation, GSMaP tends to overestimate and be least useful for hydrological applications, while WRF simulates high flows that show accurate time to the peak flows and are better in the southern watershed.

Evaluation of CMIP6 Model Precipitation Variability Through Compositing in Cloud-defined Weather States

<p>In order to understand the mechanisms determining precipitation variability and to evaluate model skill in simulating those mechanisms, it is important to partition the precipitation field into regimes that include distinct sets of processes. In the past, dynamic fields like omega and SLP have been used to define regimes and study cloud, radiation, and precipitation variability. More recently, cloud-defined weather states were derived and used for similar analyses. Here, we apply a new cloud-defined Weather State dataset derived from the higher-resolution ISCCP-H data to examine precipitation variability at global scales and evaluate CMIP6 model precipitation simulations . In addition, precipitation partitioning using mid-tropospheric vertical velocity is performed, and the differences between the results of the two compositing methodologies are discussed.</p>