Statistical Assessments of InSAR Tropospheric Corrections: Applicability and Limitations of Weather Model Products and Spatiotemporal Filtering

Abstract

Tropospheric correction is a crucial step for interferometric synthetic aperture radar (InSAR) monitoring of small deformation magnitude. However, most of the corrections are implemented without a rigorous evaluation of their influences on InSAR measurements. In this paper, we present three statistical metrics to evaluate the correction performance. Firstly, we propose a time series decomposition method to estimate the tropospheric noise and mitigate the bias caused by ground displacement. On this basis, we calculate the root-mean-square values of tropospheric noise to assess the general performance of tropospheric corrections. Then, we propose the use of semi-variograms with model-fitted range and sill to investigate the reduction of distance-dependent signals, and Spearman’s rank correlation between phase and elevation to evaluate the mitigation of topography-correlated signals in hilly areas. The applicability and limitations were assessed on the weather model-derived corrections, a representative spatiotemporal filtering method, and the integration of the two mainstream methods. Furthermore, we notice that the persistent scatter InSAR processing resulted in two components, the primary and secondary images’ contribution to the tropospheric and orbit errors. To the best of our knowledge, this paper for the first time analyzes the respective roles of the two components in the InSAR tropospheric corrections.