Topographic Change from Lidar in Eastern Kentucky, USA (2012–2023, 2017-2023): Datasets covering the July 2022 Storm

This dataset contains lidar change detection products developed to support the Geophysical Research Letters manuscript: “Reconsidering the magnitude of convective storms in triggering landslide events in the Appalachian Plateau, USA” by Scheip, et al. 1) Summary This repository contains lidar change detection (LCD) products quantifying topographic change to support studying the 26–30 July 2022 Eastern Kentucky storm (Crawford et al., 2023). We provide surface-normal change rasters and derived layers for three pre/post pairs. 2012→2023 (independent 2012 baseline - M3C2_ICP_2012_vs_2023.tif) 2012/2017→2023 (state blend over the 2012/2017 overlap area - M3C2_ICP_2012_2017_vs_2023.tif) 2017→2023 (independent 2017 baseline - M3C2_ICP_2017_vs_2023.tif) Additional, we provide a difference of digital elevation model (DoD) change raster for one pre/post pair. 2017→2023 (independent 2017 baseline, iterative closest point alignment - DoD_ICP_2023_minus_2017.tif) All inputs are classified lidar point clouds, processed to 0.5 m raster products. Point density varies by year (see Table 1). Methods follow Weidner et al. (2023) as implemented by Cambio Earth (https://www.cambioearth.com). 2) Spatial coverage AOI: Processing Footprint: lcd-area.geojson CRS: EPSG:32617 - WGS 84 / UTM zone 17N 3) Source lidar datasets (Table 1) Source lidar datasets used for LCD comparisons Metadata summarized from state records (Kentucky DGI); all rasters gridded at 0.5 m. Year Nominal date Platform Original format Procured by Collected by Bare earth Point density (pts/m²) DEM res used 2012 2012-01-01 Fixed-wing aircraft Point cloud Kentucky DGI Quantum Spatial Yes 1.2 0.5 m 2012/2017 (blended) Blended Fixed-wing aircraft Point cloud Kentucky DGI Quantum Spatial Yes 2.4 0.5 m 2017 2017-01-01 Fixed-wing aircraft Point cloud Kentucky DGI Quantum Spatial Yes 3.0 0.5 m 2023 2023-01-01 Fixed-wing aircraft Point cloud Kentucky DGI Quantum Spatial Yes 4.7 0.5 m 4) Processing summary Processing was based on methods from Lato and Ferrier, 2019 and Weidner et al., 2023. Pre-processing: source point clouds → bare-earth filtering. Coregistration: Iterative closest point (ICP: Besl and McKay, 1992) based on stable ground with removal of outliers; residuals evaluated over non-changing control areas. Change computation: Surface normal multi-scale model-to-model cloud comparison (M3C2, Lague et al., 2013) or DEM generation and subtraction. Deliverables: TIF files of change magnitudes 5) Quality, uncertainty, and recommended use Limit of detection: Recommended to use +/-13 cm for the limit of detectable change (M3C2 files) and +/- 40 cm (DoD file) Known limitations & artifacts: Mixed baselines (2012/2017): boundaries between 2012 and 2017 swaths can introduce small bias. Hydro & structures: bridges/buildings may produce false positives if not fully masked. Point density differences: earlier baselines (1.2–3.0 pts m⁻²) have coarser effective resolution than 2023 (4.7 pts m⁻²), influencing detection limits. Surface mining: Large magnitude changes near active surface mines may influence results and impede interpretation. Use with caution in heavily modified areas. 6) References Lato, M., and Ferrier, A. 2022. Systems and methods for evaluating changes in terrain topography over time. US 11,288,826 B1. U.S. Patent and Trademark Office. Weidner, L., Ferrier, A., van Veen, M., & Lato, M. J. (2023). Rapid 3D lidar change detection for geohazard identification using GPU-based alignment and M3C2 algorithms. Canadian Geotechnical Journal, 61(5), 896-91, https://doi.org/10.1139/cgj-2023-0073.