LiDAR-Based Gully Detection and Bare-Earth Modelling in Australian Rangelands

Rangeland gullies are key drivers of sediment export and landscape change, yet their detection is often constrained by vegetation cover and coarse topography. We acquired high-density airborne LiDAR over ~50 ha of Australian rangelands and produced a bare-earth Digital Terrain Model (DTM) and Digital Surface Model (DSM) and derived DEM products to characterise the gully morphology and near-surface hydrology. Point clouds were quality-checked, georeferenced, and classified into ground/non-ground classes using a terrain-adaptive workflow; terrain grids were generated at a sub-meter resolution and coupled with co-registered optical imagery processed via photogrammetry to derive complementary surface models and orthomosaics. Gully features were mapped from terrain derivatives (slope, curvature, openness, topographic wetness) and drainage enforcement, with targeted manual validation against ortho-imagery and field observations. The results show that LiDAR reliably resolves gully heads, thalwegs, banks, and interfluves beneath dense grass and scattered woody cover, where image-based photogrammetry alone is limited by canopy occlusion; in open patches, photogrammetry improves planform delineation and texture cues. The LiDAR-derived DTM provides a robust foundation for repeat-survey geomorphic change detection (cut/fill and volumetrics) and for hydrological and hydrodynamic modelling of runoff generation, flow concentration, and sediment pathways at the paddock to hillslope scales. We outline a practical decision framework for selecting the model resolution, filtering parameters, and derivative metrics to balance noise suppression against feature preservation. Although developed and validated at our case-study site, the decision framework is intended to be generalisable to rangeland systems more broadly, requiring only modest parameterisation to local terrain, vegetation, and data-availability conditions. The workflow supports monitoring and evaluation of gully remediation, erosion-risk screening, and the design of nature-based interventions in data-limited rangelands and is readily extensible to larger areas using tiled processing and cloud-hosted data services.