Quantitative evaluation of the performance of satellite and reanalysis precipitation datasets is challenging but indispensable for hydroclimate research on poorly gauged regions. This study comprehensively assesses satellite precipitation products (IMERG_E, IMERG_L, IMERG_F, CHIRPS, TRMM_3B42, TRMM_3B42RT) and reanalysis product (ERA-Interim) over the eastern Tibetan Plateau from 2008 to 2019. Each dataset's accuracy features are evaluated based on gauge observations (184 rain gauges) at various temporal scales (i.e., daily, monthly, and annual), elevations and precipitation intensities. Especially, on the topography's influence on precipitation, the spatial consistency is analyzed based on the extended triple collocation (ETC) approach along with the gauge-based gridded precipitation dataset. The results show that: (1) IMERG_F is more consistent with gauge observations and achieves the best overall performance. All products overestimate precipitation amounts, they present lower CCs (CC: 0.13–0.42) with gauge observations at daily scale, whereas show higher (CC:0.68–0.93) and acceptable CCs (CC: 0.49–0.79) at monthly and annual scale, respectively. (2) Seven products have ability to detect precipitation occurrence and extreme precipitation, the probability of detection is higher for tiny and light precipitation events and the false alarm rate is larger for extreme precipitation events. In addition, the ability of precipitation products to reproduce heavy precipitation events, both in terms of the occurrence date or precipitation amount, needs to be improved. (3) The accuracy of precipitation products varies with elevation, which has relatively good performance in high elevations. (4) The ETC analysis suggests that the accuracies are increased abruptly from the eastern edge of the Tibetan Plateau to the Sichuan Basin because of the sudden drop of terrain. ERA-Interim exhibits better performance than others, indicating reanalysis product has excellent potential in complex terrain regions. This study demonstrates new findings on the uncertainties of various precipitation products over topographical complex mountainous.

对卫星和再分析降水数据集的性能进行定量评估具有挑战性，但对于测量范围较差的地区的水文气候研究必不可少。这项研究全面评估了2008年至2019年青藏高原东部的卫星降水产物（IMERG_E，IMERG_L，IMERG_F，CHIRPS，TRMM_3B42，TRMM_3B42RT）和再分析产物（ERA-Interim）。每个数据集的准确度特征都是根据量具观测值进行评估的（184各种时间尺度（即每天，每月和每年），海拔和降水强度的雨量计。特别是，在地形对降水的影响上，基于扩展三重配置（ETC）方法以及基于量规的网格降水数据集，分析了空间一致性。结果表明：（1）IMERG_F与量规观测值更加一致，并获得最佳的整体性能。所有产品都高估了降水量，它们在日尺度下具有较低的CC（CC：0.13-0.42），而在月尺度和年尺度下则显示较高的CC（CC：0.68-0.93）和可接受的CC（CC：0.49-0.79），分别。（2）七种产品具有探测降水和极端降水的能力，微小和轻度降水事件的检测概率较高，极端降水事件的误报率较高。另外，无论从发生日期还是在降水量方面，都需要提高降水产物重现强降水事件的能力。（3）降水产物的精度随高度而变化，在高海拔地区具有相对较好的性能。（4）ETC分析表明，由于地形的突然下降，从青藏高原东缘到四川盆地的精度突然增加。ERA-Interim具有比其他产品更好的性能，这表明再分析产品在复杂的地形区域中具有出色的潜力。这项研究证明了关于地形复杂山区各种降水产物不确定性的新发现。