H13H-1643
Ground Validation of IMERG and TMPA 3B42 Rainfall Products Based on an Ultra-High Density Network of Rain Gauges over Mountainous Region
Monday, 14 December 2015
Poster Hall (Moscone South)
Ran Xu, Tsinghua University, Beijing, China
Abstract:
Satellite precipitation begins a new era since the first release of IMERG products. The finer spatio-temporal resolution of IMERG rainfall products has great potentials for the improvement of regional hydrology studies related to flood forecast and water resources management, especially over regions with sparsely gauged observations, e.g., mountainous regions. However, previous studies found large discrepancy among satellite rainfall products over regions with high altitudes. The main objective of this study is to provide a first evaluation of IMERG products (Final run, with a latency of 18 hours) over mountainous region and its inter-comparison with its predecessor TMPA 3B42 (V7). We center this study over Yarlung Tsangbo River Basin in Tibetan Plateau region (with mean altitude exceeding 4600m). Over 500 rain gauges are distributed over the basin, constituting an ultra-high density network of rain gauges, which enables us to evaluate the performance of satellite rainfall products (IMERG and TMPA) in capturing space-time rainfall structure in this region. We focus on the warm season (May to October) in 2014. Preliminary results show that IMERG resemble TMPA in spatial rainfall distribution over downstream region, whereas IMPEG present a remarkable smaller estimation than TMPA over upstream region. IMERG outperforms TMPA in all statistical indices (mean bias, correlation coefficient, root mean square error, etc.) used in this study. There is a general dependency of IMERG performance on altitude. The consistency (correlation coefficient) between IMERG and rain gauge improves with increased altitude (over 5000 m.a.s.l), but at the cost of increasing mean bias in the meantime. We evaluate the conditional bias of IMERG products on rainfall intensity and time of the day (daytime and nighttime). Results in this study provide useful reference for the improvement of IMERG rainfall calibration algorithm over high-altitude regions.