Assessment of ERA-Interim and ERA5 reanalysis data on atmospheric corrections for InSAR

Abstract

• Corrections with NWMs are strongly affected by season, climate, and topography. • ERA5 has almost no improvement over ERA-Interim in nonlinear component correction. • Neither reanalysis is sufficient to reproduce the small-scale turbulence. Interferometric synthetic aperture radar (InSAR) has achieved remarkable success in deformation monitoring and many related fields but ordinarily suffers from inherent drawbacks stemming from atmospheric effects. In the meantime, the continuously progressing numerical weather models (NWMs) offer a feasible approach to this problem. This paper aims to further investigate the usability of the NWMs. The latest European Centre for Medium-range Weather Forecast (ECMWF) reanalysis ERA5 and its predecessor ERA-Interim (ERAI) were evaluated using 48 pairs of Sentinel-1 A/B interferograms located in nine subregions of China, covering different seasons, climate types, and landforms. In particular, we focus on the nonlinear component of the InSAR atmospheric error, which is the most difficult to correct. The results show that two reanalyses are the most effective in winter and the least effective in summer. In addition, the correction effect varies between climate types and is further complicated by local topography, water vapor resources, etc. Regions at higher latitude, farther from the sea, and with more moderate topography are more suitable for correction with the two reanalyses. Furthermore, ERA5 shows comparable performance to ERAI in correcting for nonlinear atmospheric effects, and superior performance to ERAI if the spatial linear trend component is retained. Nevertheless, some detailed features of the atmospheric impact, particularly small-scale local turbulence, remain difficult to capture and spatial resolution may not determine their correction effects. Further improvements to NWM are needed in the future, particularly through the assimilation of more accurate meteorological observations with higher spatial and temporal resolution.