Pre-event composite of the EarthDEM Strips for the 2026-08-26 Rasuwa (Langtang Lirung) ice-rock avalanche, Nepal

This record provides pre-event digital elevation model (DEM) composites for the area affected by the 2026-08-26 ice-rock avalanche on the north flank of Langtang Lirung, Rasuwa District, Nepal, and its downstream flood along the Chhochen Khola, Lhende Khola, Bhote Koshi and Trishuli rivers. The composites are the pre-event reference surface for the post-event DEMs of this event (Shean, 2026a; Shean and Bhushan, 2026a, 2026b). The composites combine 109 EarthDEM strip DEMs, version 4.1 (Porter et al., 2024), from WorldView-1/2/3 and GeoEye-1 in-track stereo images acquired 2009-03-16 to 2024-01-05. Each strip was co-registered by translation on stable ground, and the co-registered stack was tied to ICESat-2. Two sets are provided. The 2 m set is clipped to the affected corridor and has 1032 km2 of valid data. The 8 m set covers a 45.7 x 40.2 km rectangle around the source, runout and upper corridor, with 1721 km2 of valid data. Both sets include median and weighted-mean DEMs, per-pixel strip count and normalized median absolute deviation (NMAD), hillshades, an ordered temporal layer set and a strip footprint index. The ordered layers are earliest- and latest-strip-on-top DEMs with source-strip and acquisition-date maps, their elevation difference and the apparent rate. No individual strip DEMs and no source imagery are distributed. All products are cloud-optimized GeoTIFFs in EPSG:32645 (WGS 84 / UTM zone 45N) with heights above the WGS 84 ellipsoid. License: CC BY-NC 4.0. DEMs provided by the Polar Geospatial Center under NSF-OPP awards 1043681, 1559691, 1542736, 1810976, and 2129685. Original Level-1B Imagery © 2009-2024 Vantor. 1. Products Nodata is -9999 in elevation, date, change, NMAD and count layers, 0 in hillshade layers and -1 in index layers. NMAD layers are nodata where fewer than two strips overlap. 00_*.png files are quick-look figures. 1.1 Corridor 2 m composite (earthdem_corridor_2m.zip: earthdem_corridor_2m_<layer>_v1.0.tif) The 2 m set is clipped to the affected-corridor polygon plus a 500 m buffer, and the strips were clipped before compositing. The 88 contributing strips were acquired 2009-03-16 to 2024-01-05. The median of date_median is about 2016.7, and the median of date_wmean is about 2017.5. layer description wmean weighted-mean composite DEM (edge-distance weights) -- recommended surface for relief and as a co-registration reference median per-pixel median DEM count, nmad per-pixel strip count and NMAD (m) wmean_hs, median_hs multidirectional hillshades first, last earliest- and latest-strip-on-top DEMs, with first_hs, last_hs index_first, index_last index of the strip supplying each first / last pixel, into earthdem_corridor_2m_strip_order_v1.0.csv (0 = oldest) date_first, date_last, dt_yr acquisition date (decimal year) of the first / last pixel, and dt_yr = date_last minus date_first (decimal years) date_median, date_wmean acquisition date of the strip nearest the median value, and the weight-averaged acquisition date of the wmean dh_m elevation difference last minus first (m), where the two dates differ dhdt_myr apparent rate dh_m / dt_yr (m/yr), only where dt_yr is at least 2 years Also in the zip: earthdem_corridor_2m_strip_order_v1.0.csv: the strips in the product, oldest first. earthdem_strip_index_v1.geojson: footprints, acquisition dates and co-registration shifts of the 224 strips considered. in_mosaic marks the 208 strips in the co-registered stack. The shifts exclude the common tie to ICESat-2 (dE +0.07 m, dN -2.44 m, dU -0.10 m) that was applied to every strip. earthdem_corridor_2m_ordered_stats_v1.0.json: coverage and stable-ground statistics of dh_m and dhdt_myr. 00_earthdem_corridor_2m_ordered_qa_v1.0.png: quick-look figure of the ordered layers. 1.2 Source-rectangle 8 m composite (earthdem_source_8m.zip: earthdem_source_8m_<layer>_v1.0.tif) The 8 m set covers a 45.7 x 40.2 km rectangle around the source area, runout and upper corridor, and is not clipped to the corridor. The rectangle is the same as in the companion HMA pre-event record (Shean and Bhushan, 2026c). The 8 m composite combines 66 co-registered strips acquired 2012-01-15 to 2023-12-02. The median of date_median is about 2016.0, and the median of date_wmean is about 2016.5. The layers and sidecar files are the same as in the 2 m set (section 1.1). In this set the wmean has smoother relief across strip boundaries, and the median has a slightly lower NMAD against ICESat-2 (0.90 m vs 0.94 m). 1.3 Validation (validation_v1.0.zip) The validation zip holds the ICESat-2 comparison for each DEM: a summary table, and a figure and a statistics file per product (section 3). 2. Methods Each strip was co-registered by translation in three passes. The first pass used the Copernicus GLO-30 DEM (European Space Agency and Airbus, 2022), converted to ellipsoid heights, as the reference. The second and third passes used the median of the aligned stack. Each pass solved the horizontal offset with the method of Nuth and Kääb (2011) on exposed stable ground with slopes of 1-45 degrees. Stable ground is the ESA WorldCover 2021 bare, grass, shrub, moss and cropland classes, excluding Randolph Glacier Inventory (RGI) 6.0 glaciers with a 100 m buffer. Before each solve, differences more than 100 m from the median difference and outliers beyond 3 NMAD were removed. The vertical offset is the median residual on stable ground. A refit was accepted only where it reduced the residual aspect signature. The strips were composited with the NASA Ames Stereo Pipeline dem_mosaic tool (Beyer et al., 2018). SETSM (Noh and Howat, 2015) is the EarthDEM stereo software. The 2 m composites use only stereo-matched pixels. Pixels that SETSM interpolated were masked along with the strip bitmask, and they are about 2 % of the valid strip area. The 8 m composites apply the bitmask only. GLO-30 was used only as the first-pass registration reference. No GLO-30 values enter any product, and voids are left as nodata. Before compositing, obvious blunders of up to hundreds of meters were removed from the strips over the source area. The 2019 Pléiades DEM (Berthier, 2026) and other independent DEMs served as the reference there. Coordinates are ITRF2020 at epoch 2022.27, the median date of the ICESat-2 data. The composite surfaces mostly date from about 2016, and plate motion of about 4 cm/yr toward the northeast moves the area about 0.25 m between 2016 and 2022. The 2015-04-25 Gorkha earthquake displaced the area by about 1.5 m horizontally, partway through the strip archive. Each strip was co-registered separately, so strips from before and after the earthquake are aligned to the same stack median. 3. Accuracy The ICESat-2 reference elevations are 10 m segment fits to all ATL03 photons, limited to exposed stable terrain. Against these segments, the four DEMs have median differences of -0.02 to +0.02 m and NMAD of 0.81 to 0.94 m. The 2 m DEMs overlap 43,031 segments and the 8 m DEMs overlap 72,773. Per-product statistics and figures are in validation_v1.0.zip. The figures include a diagnostic Nuth and Kääb (2011) horizontal fit. 4. Known issues The DEMs have voids on snow, shadow and glacier surfaces. median and wmean combine strips acquired 2009-2024, so on glacier, snow and vegetated surfaces they do not represent a single date; date_median and date_wmean give the approximate date of each pixel. first and last take each pixel from a single strip, identified in index_first / index_last and dated in date_first / date_last. The first and last mosaics can have larger errors and more blunders than median and wmean. The latest strip is from 2024-01-05. Along the Lhende Khola and Bhote Koshi, differences against post-event DEMs include changes from the separate 2025-07-08 Lhende Khola flood as well as the 2026-08-26 event. Blunders were removed only over the source area, inside the 2019 Pléiades footprint (section 2). Elsewhere, isolated blunders can remain where few strips overlap. Check the count and nmad layers before interpreting a large elevation change. The companion HMA pre-event record (Shean and Bhushan, 2026c) is co-registered to the same frame; the 8 m medians of the two records agree on stable ground to 0.8 m NMAD. 5. License and acknowledgments CC BY-NC 4.0. DEMs provided by the Polar Geospatial Center under NSF-OPP awards 1043681, 1559691, 1542736, 1810976, and 2129685, derived from EarthDEM Strips, Version 4.1 (Porter et al., 2024). As the strip license requests, please cite the SETSM software (Noh and Howat, 2015) in any publication or presentation that uses these products, and let Ian Howat (The Ohio State University) know of the use. This work utilized data made available through the NASA Commercial Satellite Data Acquisition (CSDA) Program. Original Level-1B Imagery © 2009-2024 Vantor; the NGA Commercial Imagery Public Release Team concurred with the public release of the derived composites on 2026-09-28. ICESat-2 data via NASA/NSIDC, processed with SlideRule (Shean et al., 2023). ESA WorldCover 2021 v200 (Zanaga et al., 2022; © ESA WorldCover project 2021 / Contains modified Copernicus Sentinel data (2021) processed by ESA WorldCover consortium). Glacier outlines from the Randolph Glacier Inventory 6.0 (RGI Consortium, 2017). Copernicus DEM GLO-30, used as a registration reference only: Copernicus DEM © DLR e.V. 2010-2014 and © Airbus Defence and Space GmbH 2014-2018 provided under COPERNICUS by the European Union and ESA; all rights reserved. Funding: NASA award 80NSSC24K1633 (HiMAT-2B) and NASA award 80NSSC26K0856 (DSI24). Products were generated on a rapid disaster-response timeline and are provided without warranty. 6. Citation Shean, D. (2026). Pre-event composite of the EarthDEM Strips for the 2026-08-26 Rasuwa (Langtang Lirung) ice-rock avalanche, Nepal (v1.0) [Data set]. Zenodo. https://doi.org/10.5281/zenodo.22149436 References Berthier, E. (2026). 13 October 2019 Pleiades (CNES) pre-event DEMs for the rock-ice avalanche source area of the August 2026 Trishuli catastrophic flood (v1) [Data set]. Zenodo. https://doi.org/10.5281/zenodo.22147118 Beyer, R. A., Alexandrov, O., and McMichael, S. (2018). The Ames Stereo Pipeline: NASA's open source software for deriving and processing terrain data. Earth and Space Science, 5, 537-548. https://doi.org/10.1029/2018EA000409 European Space Agency and Airbus (2022). Copernicus DEM. https://doi.org/10.5270/ESA-c5d3d65 Noh, M.-J., and Howat, I. M. (2015). Automated stereo-photogrammetric DEM generation at high latitudes: Surface Extraction with TIN-based Search-space Minimization (SETSM) validation and demonstration over glaciated regions. GIScience & Remote Sensing, 52(2), 198-217. https://doi.org/10.1080/15481603.2015.1008621 Nuth, C., and Kääb, A. (2011). Co-registration and bias corrections of satellite elevation data sets for quantifying glacier thickness change. The Cryosphere, 5, 271-290. https://doi.org/10.5194/tc-5-271-2011 Porter, C., Howat, I., Husby, E., Noh, M.-J., Khuvis, S., Danish, E., Tomko, K., Gardiner, J., Negrete, A., Yadav, B., Klassen, J., Kelleher, C., Cloutier, M., Bakker, J., Power, D., Enos, J., Arnold, G., Bauer, G., and Morin, P. (2024). EarthDEM - Strips, Version 4.1. Harvard Dataverse, V1. https://doi.org/10.7910/DVN/YJAH5X RGI Consortium (2017). Randolph Glacier Inventory - A Dataset of Global Glacier Outlines, Version 6. NSIDC. https://doi.org/10.7265/4m1f-gd79 Shean, D. (2026a). Post-event SAR stereo DEMs (Umbra) for the 2026-08-26 Rasuwa (Langtang Lirung) ice-rock avalanche, Nepal (v1.0) [Data set]. Zenodo. https://doi.org/10.5281/zenodo.22842744 Shean, D., and Bhushan, S. (2026a). Post-event optical stereo DEMs (WorldView-2, WorldView Legion 2/3, WorldView-3) for the 2026-08-26 Rasuwa (Langtang Lirung) ice-rock avalanche, Nepal (v1.0) [Data set]. Zenodo. https://doi.org/10.5281/zenodo.22842746 Shean, D., and Bhushan, S. (2026b). Post-event multi-source DEM composite and elevation-change maps for the 2026-08-26 Rasuwa (Langtang Lirung) ice-rock avalanche, Nepal (v1.0) [Data set]. Zenodo (in preparation). Shean, D., and Bhushan, S. (2026c). Pre-event DEM composites for the 2026-08-26 Rasuwa (Langtang Lirung) ice-rock avalanche, Nepal: NASA HMA 8 m (2012-2016) and WorldView/GeoEye 2 m (2015-2017) (v1.0) [Data set]. Zenodo. https://doi.org/10.5281/zenodo.22842748 Shean, D., Swinski, J. P., Smith, B., Sutterley, T., Henderson, S., Ugarte, C., Lidwa, E., and Neumann, T. (2023). SlideRule: Enabling rapid, scalable, open science for the NASA ICESat-2 mission and beyond. Journal of Open Source Software, 8(81), 4982. https://doi.org/10.21105/joss.04982 Zanaga, D., Van De Kerchove, R., Daems, D., De Keersmaecker, W., Brockmann, C., Kirches, G., Wevers, J., Cartus, O., Santoro, M., Fritz, S., Lesiv, M., Herold, M., Tsendbazar, N.-E., Xu, P., Ramoino, F., and Arino, O. (2022). ESA WorldCover 10 m 2021 v200. Zenodo. https://doi.org/10.5281/zenodo.7254221