Error analysis of multi-satellite precipitation estimates with an independent raingauge observation network over a medium-sized humid basin

This study focuses on quantifying the error characteristics of four widely utilized satellite precipitation products (i.e. TMPA 3B42RTV7, TMPA 3B42V7, CMORPH and PERSIANN-CDR) for a five-year period (2005–2009) using an independent raingauge network over the upper-middle Huai River basin in central-eastern China. Assessment results show that CMORPH generally exhibits the best performance with slight underestimation, while 3B42RTV7 has the worst performance with large positive biases. Additionally, 3B42V7 and PERSIANN-CDR tend to have an approximate accuracy. The monthly gauge adjustment applied to 3B42V7 and PERSIANN-CDR significantly reduces their systematic bias and in particular it makes these two research products maintain a stable skill level during winter. As for the heavy rainfall events (>50 mm/d) in summer, 3B42V7 and CMORPH exhibit a relatively better degree of agreement to the gauge observations. Overall, our study suggests that the satellite-based precipitation estimates all have their own pros and cons at different spatiotemporal scales. We expect the results reported here will provide a better understanding of current mainstream satellite precipitation products over similar medium-sized humid basins.Editor Z.W. Kundzewicz; Associate editor N. Verhoest