Evaluating Remote Sensing Precipitation Products Using Double Instrumental Variable Method

• Published in: IEEE geoscience and remote sensing letters Vol. 19; pp. 1 - 5
• Main Authors: Wei, Zhihao, Long, Xunjian, Gai, Yingying, Yang, Zekun, Sui, Xinxin, Chen, Xi, Kan, Guangyuan, Fan, Wenjie, Cui, Yaokui
• Format: Journal Article
• Language: English
• Published: Piscataway IEEE 2022; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Double instrumental variable (DIV); Error analysis; error estimation; Errors; Estimation; Global precipitation; Ground-based observation; Instruments; Integrated Multisatellite Retrievals for Global Precipitation Measurement Mission (IMERG); Mathematical models; Methods; Moisture effects; Numerical models; Precipitation; Precipitation estimation; Quality assessment; Rainfall; Remote sensing; Soil moisture; Soil Moisture to Rain (SM2RAIN); Spatial distribution; Terrain; Water resources
• ISSN: 1545-598X; 1558-0571
• DOI: 10.1109/LGRS.2022.3192644

Error estimation of precipitation products is an important procedure in the data quality evaluation. It is a challenging task due to the lack of the in situ ground observations and the variations of the geophysical characteristics in regions with complex terrain. Compared with the traditional methods, the double instrumental variable (DIV) method has the merits of being able to estimate the errors between two products. In this study, the DIV method for data error estimation is applied and validated on precipitation products in regions with complex terrain. The DIV-based errors for two state-of-the-art precipitation products Integrated Multisatellite Retrievals for Global Precipitation Measurement Mission (IMERG) and Soil Moisture to Rain (SM2RAIN) are being further verified by using another high-accuracy ground-based precipitation products China Merged Precipitation Analysis (CMPA). The results indicate that the DIV-based errors of IMERG and SM2RAIN range from 0 to 25 mm per day and from 0 to 15 mm per day, respectively. The root-mean-square errors (RMSEs) of IMERG and SM2RAIN compared with CMPA, which are defined as CMPA-based errors, are ranging from 0 to 23 and 0 to 22 mm, respectively. It is concluded that the spatial distribution of the DIV-based errors shows the consistency with the CMPA-based errors, which further demonstrates the potential of using the DIV method for precipitation products fusion.