Using multiple satellite-gauge merged precipitation products ensemble for hydrologic uncertainty analysis over the Huaihe River basin

Abstract

Global satellite–gauge merged precipitation (SGMP) products combine the advantages of satellite precipitation estimates with rain gauge data, providing great potential to hydrological applications. However, the inaccuracies of the precipitation products together with hydrologic model limitations, could cause great uncertainty in streamflow predictions. Therefore, this study investigates the hydrological value of three mainstream global SGMP products, including the Tropical Rainfall Measuring Mission (TRMM) Multi-satellite Precipitation Analysis (TMPA) 3B42V7 product, the Climate Prediction Center (CPC) MORPHing technique (CMORPH) satellite–gauge merged product (CMORPH BLD), the Global Satellite Mapping of Precipitation (GSMaP) Gauge-calibrated product (GSMaP Gauge). They are used as the precipitation input of the Variable Infiltration Capacity (VIC) hydrologic model over the Huaihe River basin in China. To better quantify their effects on parameter calibration and streamflow predictions, a newly developed residual error model accompanied with the Bayesian uncertainty analysis are performed. CMORPH satellite-gauge merged precipitation product, recently developed by the China Meteorological Administration (CMA) (CMORPH CMA), is a high-quality regional precipitation product. Thus, this study applies the CMORPH CMA within the same framework to provide a benchmark. The results show that the parameter uncertainty are influenced significantly by the input of various precipitation products. There is a tradeoff between the deterministic streamflow performance and the probabilistic predictive performance for selecting the best input among the three global precipitation products. The streamflow uncertainty intervals of the three global precipitation products are then merged using the Bayesian Model Averaging (BMA) method. The BMA results show satisfying hydrological performance in terms of deterministic streamflow predictions, with the largest Nash-Sutcliffe coefficient of Efficiency (NSCE) values of 0.86 and 0.64, and the smallest absolute relative error (RE) values of 0% and 10.2% in the calibration and validation periods, respectively. In addition, the BMA results also produce much more reliable probabilistic predictions, which even outperform the outcomes of the high-quality CMORPH CMA. Our study demonstrates the potential uncertainty of various SGMP products for model calibration and streamflow predictions. The hydrologic ensemble using multiple global SGMP products provides a promising and advantageous approach to support water management and decision making, especially in ungauged basins.