Recent developments in remote sensing of SWE using InSAR

Satellite remote sensing of snow water equivalent (SWE) at high spatiotemporal resolutions remains an unsolved challenge in snow hydrology. While accurate and high resolution measurements of snow surface properties (e.g., snow cover, grain size, albedo) can be derived from multispectral and hyperspectral data, these sensors cannot provide direct SWE information. Synthetic aperture radar (SAR) has the potential to measure SWE directly because the radar signal at sufficiently low frequencies can penetrate a dry snowpack. Depending on the SAR frequency used, both backscatter-based and interferometric (InSAR) approaches have been demonstrated. Here we present recent results from several studies that investigate remote sensing of SWE using airborne L-band (1.26 GHz) and spaceborne C-band (5.405 GHz) InSAR data. Because the InSAR technique is sensitive to changes in atmospheric and soil conditions as well as snow, one way to determine where to apply the technique is to incorporate satellite-based optical snow cover maps alongside the InSAR data. We show that careful selection of optical snow data is necessary because differences in the spatial and temporal resolutions between the optical and InSAR products propagate uncertainties into SWE calculations, which can change the final SWE estimates by more than 100%. Additionally, optical sensors can accurately detect snow cover in forested areas with canopy densities up to 60%, but vegetation effects may cause temporal decorrelation in InSAR data over these environments and prevent the retrieval of SWE information. Using data from two field sites in Colorado, USA, we show that InSAR coherence generally remains sufficiently high over temporal baselines of 12 days or more, allowing unbiased SWE estimates to be obtained across landscapes with canopy densities up to 40%. These results show the potential for SWE monitoring with the L-band InSAR sensor on the NISAR satellite, especially when combined with other SAR (e.g. Sentinel-1) and optical (e.g. Landsat 8/9) satellites.