Ionospheric correction of InSAR data for accurate ice velocity measurement at polar regions

Interferometric synthetic aperture radar (InSAR) has become an essential tool for measuring ice sheet velocity in the Polar Regions. At low radar frequencies, e.g. L-band (1.2 GHz) but also at higher frequency, e.g. C-band (5.6 GHz), the ionosphere has been documented to be an important source of noise in these data. In this paper, we employ a split-spectrum technique and investigate its performance for correcting ionospheric effects in InSAR-based ice velocity measurements in Greenland and Antarctica. Three case studies using ALOS PALSAR data are used to assess the performance of the split spectrum technique for ionosphere correction over a range of environmental parameters. We employ several approaches to evaluate the results, including visual inspection, profile analysis, comparison of experimental and theoretic errors, comparison with reference data from other sources, generation of double difference interferograms, and analysis of time series of multi-temporal data. Our experiments show that ionospheric distortions are observed regularly, and in our analyzed Greenland dataset and Antarctic dataset the ionospheric noise reaches 14 m/yr and 10 m/yr, respectively, which exceeds the signal associated with ice motion. Our analysis using several different approaches demonstrates that the split-spectrum technique provides an effective correction. The split spectrum technique is also found to be superior to currently used approaches such as baseline fitting and multi-temporal averaging. The noise level is reduced by a factor of 70% in Greenland test areas and 90% in Antarctic test areas. •Split-spectrum methods are effective in correcting ionospheric noise from InSAR data.•Errors in ice velocity estimates were reduced by ~70% (Greenland)–90% (Antarctic).•Split-spectrum-based correction outperforms multi-temporal averaging techniques.