Robust Surface Deformation and Tropospheric Noise Characterization From Common-Reference Interferogram Subsets

Interferometric Synthetic Aperture Radar (InSAR) surface deformation estimates often suffer from tropospheric noise errors, and many study sites do not have GPS or in-situ measurements for validating InSAR surface deformation solutions. Here we present a method for characterizing both surface deformation and tropospheric noise from interferogram subsets. Choosing different subsets of interferograms that use a common-reference SAR scene allows us to quantify tropospheric noise and deformation signals. We demonstrate this method using 95 C-Band Sentinel-1 SAR scenes acquired over the Oman Ophiolite and 133 scenes over the Island of Hawaii. For the Oman case, our method suggests that there is no detectable deformation signal. In this scenario, the average of any subset of interferograms with a common-reference SAR scene estimates the tropospheric noise contribution on that reference SAR date. Achieving ~0.5 cm of uncertainty requires a subset size of 40 common-reference interferograms. The observed tropospheric noise is non-Gaussian, and follows a seasonal pattern. Propagation of tropospheric noise leads to up to ~5 cm of false deformation signal when deriving either a stacking or SBAS time series solution. For the Hawaii case, our method shows that the observed InSAR phase on the south flank of Kilauea is due to a secular deformation signal, while the phase over Mauna Loa is mostly associated with tropospheric noise. Our results are validated with independent GPS tropospheric zenith delay and surface deformation time series with sub-cm RMS errors.