Processing methodology for the ITS_LIVE Sentinel-1 ice velocity products

Abstract. The NASA MEaSUREs Inter-mission Time Series of Land Ice Velocity and Elevation (ITS_LIVE) project seeks to accelerate understanding critical glaciers ice sheet processes by providing researchers with global, low-latency, comprehensive state the art records surface velocities elevations as observed from space. Here we describe image-pair velocity product processing methodology for ESA Sentinel-1 radar data. We demonstrate improvements core algorithm dense offset tracking, “autoRIFT”, that provide finer resolution (120 m instead previous 240 used version 1) higher accuracy (20 % 50 improvement) data products significantly enhanced computational efficiency (>2 orders magnitude) when compared earlier versions “dense ampcor” routine in JPL ISCE software. In particular, disparity filter is upgraded handling grid overlapping search chip sizes, oversampling ratio subpixel cross-correlation estimation adaptively determined matching precision measured displacement based on size used. A novel calibration applied correct Sentinel-1A/B subswath full-swath dependent geolocation biases caused systematic issues instruments. C-band images are affected variations total electron content ionosphere results large errors azimuth (along-track) direction. To reduce these effects, slant range (line sight or LOS) accompanied LOS parameters support map coordinate (x/y) inversion ascending descending measurements, derived two image pairs. After proposed correction errors, uncertainties reduced 9 %–61 %. further validate ITS_LIVE Version 2 products, 6-year time series composed thousands epochs, over three typical test sites covering globe: Jakobshavn Isbræ Glacier Greenland, Pine Island Antarctic, Malaspina Alaska. By comparing other similar (PROMICE, FAU, Annual Antarctic Map products), well Landsat-8 Sentinel-2 data, find an overall variation between around 100 yr−1 fast-flowing glacier outlets, where both mean order km yr−1, increases up 300–500 (3 %–6 %) fastest Glacier. magnitude uncertainty calculated be uniformly distributed 60 regions investigated. described methods comprise Image-Pair Sheet Surface Velocities: (DOI: https://doi.org/10.5067/0506KQLS6512, Lei et al., 2022).