How well were the early 2017 California Atmospheric River precipitation events captured by satellite products and ground-based radars?

In January and February of 2017, California experienced multiple heavy storms that caused serious destruction of facilities and economic loss, although it also helped to reduce water storage deficit due to prolonged drought in previous years. These extreme precipitation events were mainly associated with Atmospheric Rivers (ARs) and brought about 174 km³ of water to California according to ground observations. This article evaluates the performance of six commonly used satellite-based precipitation products (IMERG, 3B42RT, PERSIANN, CCS, CMORPH and GSMaP), as well as ground-based radar products (Radar-only and Radar-lgc) in capturing the ARs precipitation rate and distribution. It is found that precipitation maps from all products present heavy precipitation in January and February, with more consistent observations over ocean than land. Though large uncertainties exist in quantitative precipitation estimation (QPE) over land, the ensemble mean of different remote-sensing precipitation products over California is consistent with gauge measurements. Among the six satellite-based products, IMERG correlates the best with gauge observations both in the detection and quantification of precipitation, but it is not the best product in terms of root-mean-square error or bias. Compared to satellite products, ground weather radar shows better precipitation detectability and estimation skill. However, neither radar nor satellite QPE products have good performances in quantifying the peak precipitation intensity during the extreme events, suggesting that further advancement in quantification of extremely intense precipitation associated with ARs in the western United States is needed.