Dataset from Mannerfelt et al. (2025): Tracking glacier surge evolution using interferometric SAR coherence — examples from Svalbard

Attached are output data for Mannerfelt et al. (2025): Tracking glacier surge evolution using interferometric SAR coherence — examples from Svalbard. Please cite this paper for any data usage. https://doi.org/10.1017/jog.2025.27 Further explanations for how these data were used/created are in the code of the repository: https://github.com/erikmannerfelt/IncoherentSurges Glacier front positions (front_positions.zip) Glacier centerlines (centerlines.geojson) Maximum extent glacier outlines (largest_extent_glacier_outlines.geojson) Low-coherence zone polygons (low_coh_zones.geojson) Low-coherence boundary lines (low_coh_boundaries.geojson) Glacier surge statistics table (surge_stats_with_geometric.csv) Glacier surge statistics points (surge_stats_with_geometric.geojson) Glacier location points (glaciers.geojson) Full-resolution figures (mannerfelt2025_tracking_fig<number>.{jpg,png}) Notes: The low-coherence zone polygons were only used for plotting (Fig. 2). The boundary lines are the data from which statistics were derived The glacier points are translatable to RGI ids through spatial joins. Glacier front positions are derived from Sentinel-1 SAR data if the date corresponds to a S1 acquisition date. Otherwise it is from Sentinel-2 optical data. The surge statistics table/points are outputs from the code (level-2 data). The rest are made from image interpretations in QGIS (level-1-ish data). The unit of the *_rate variables in the surge statistics table/points is meters per day. The variables zrange, area_km2, lmax_m, slope_deg and tidewater are taken from RGI7. February 2026 update Newly added are 75th percentile 12-day coherence mosaics covering Svalbard for the years 2015–2025 (inclusive). They follow the pattern: coh_q75_<year>_HH.tif and are JPEG compressed with a modified range from 0-255 (scaled from 0-1). The products were not used in the publication but are subsequently added as a simpler dataset to get new users started. The years represent hydrological years (starting 1 October), i.e. they are 120 days ahead of the normal calendar. They only consist of Sentinel-1A data.