Skip to main content
ARS Home » Northeast Area » Beltsville, Maryland (BARC) » Beltsville Agricultural Research Center » Hydrology and Remote Sensing Laboratory » Research » Publications at this Location » Publication #356429

Research Project: Integrating Remote Sensing, Measurements and Modeling for Multi-Scale Assessment of Water Availability, Use, and Quality in Agroecosystems

Location: Hydrology and Remote Sensing Laboratory

Title: A double instrumental variable method for geophysical product error estimation

item DONG, J. - US Department Of Agriculture (USDA)
item Crow, Wade

Submitted to: Remote Sensing of Environment
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 12/15/2018
Publication Date: 2/1/2019
Citation: Dong, J., Crow, W.T. 2019. A double instrumental variable method for geophysical product error estimation. Remote Sensing of Environment. 9:1657.

Interpretive Summary: Satellite-based rainfall and soil moisture estimates can be used for a wide range of agricultural applications including: drought forecasting, yield monitoring and flood prediction. Recently, NASA has launched the Soil Moisture Active/Passive (SMAP) and Global Precipitation Measurements (GPM) missions. These two missions provide a significant step forward in our ability to globally monitor water resources within agricultural regions. However, rainfall and soil moisture products generated by these missions must first be validated via comparisons against ground-based observations. These comparisons can be challenging due to the severe lack of available ground observations in many areas of the world and the sharp contrast between the coarse-scale nature of satellite retrievals (with resolutions > 10 km) and the point-scale nature of ground-based soil moisture and rainfall measurements. This paper presents a new mathematical strategy for overcoming these challenges. It will be applied to validate SMAP and GPM products. This credible validation will, in turn, help to promote the full use of SMAP and GPM data products for important agricultural applications.

Technical Abstract: The global validation of remotely sensed and/or modelled geophysical products is often complicated by a lack of suitable ground observations for comparison. By cross-comparing three independent collocated observations, triple collocation (TC) can solve for geophysical product errors in error-prone systems. However, acquiring three independent products for a geophysical variable of interest can be challenging. In this study, a double instrumental variable based algorithm (IVd) is proposed to tackle this challenge. This new approach can estimate product error standard deviation (s) and product-truth correlation (R) using only two independent products. The IVd method is developed and demonstrated by an example analysis conducted using a precipitation error analysis based on reanalysis (i.e., ERA-Interim) and remotely sensed (i.e., Tropical Rainfall Measuring Mission precipitation 3B42RT) data products. Results show that, when sampled over a global extent, the spatial correlations of TC- and IVd-estimates are 0.94 (ERA-Interim), and 0.97 (TRMM) for s and 0.68 (ERA-Interim) and 0.78 (TRMM) for R. Compared with TC-based results, the biases of IVd are generally less than 5% for both s and R. These results are shown to be superior to those obtained by an existing (single) instrumental variable approach. Given the consistency of TC- and IVd-derived results, we conclude that the IVd method is a viable tool for estimating geophysical product errors for cases where only two independent products are available.