Main content area

Error analysis and correction of the daily GSMaP products over Hanjiang River Basin of China

Deng, Pengxin, Zhang, Mingyue, Guo, Haijin, Xu, Changjiang, Bing, Jianping, Jia, Jianwei
Atmospheric research 2018 v.214 pp. 121-134
calibration, data collection, hydrometeorology, rain, rain gauges, rain intensity, remote sensing, satellites, watershed hydrology, watersheds, China
The Global Satellite Mapping of Precipitation (GSMaP) product is an important satellite precipitation product of Global Precipitation Measurement (GPM) mission with high-precision and high-resolution. To quantify the error features and correction of GSMaP estimates and understand their hydrological potentials, the statistical indices were calculated to investigate the errors of three widely used GSMaP products (GSMaP_NRT, GSMaP_MVK and GSMaP_Gauge) at the daily and 0.1° × 0.1° resolutions over Hanjiang River Basin. Assessment results show that both GSMaP_NRT and GSMaP_MVK apparently overestimate the low rainy events (<10 mm/d), while underestimate the mid and high rainy events (>35 mm/d). False bias and hit bias are the main source of the total errors which mainly emerged in the upstream of Hanjiang River Basin. GSMaP_Gauge are in a position to substantially monitor the precipitation events with better error stability and distribution. But its unexpected systematic biases and high false alarm ratio, which mainly come from the overestimation of lower rain rates especially in the Tangbai river basin, still need to strengthen further in future developments. In addition, its mean errors are strongly related to rainfall intensity of rain gauge measurements that could be characterized by a quadratic polynomial fitting, which provides useful information on calibration. In practice, the errors of GSMaP_Gauge estimates have been reduced after correction by the polynomial fitting, while it causes some rain events to be lost. Overall, our study reveals the GSMaP_Gauge dataset can well reflect the precipitation characteristics of Hanjiang River Basin and offer the opportunity to be used in hydrometeorological applications. We expect the results documented here can provide a deep insight to the GSMaP estimates' error features correction and potential utilization for various hydrological applications.