GEOECOLOGICAL MODEL OF LANDSLIDE RISK WITH CONSIDERATION OF FOREST-COVER STRUCTURE AND SPATIAL EXPOSURE (A CASE STUDY OF A SITE IN THE TEREBLIA RIVER BASIN, UKRAINIAN CARPATHIANS)

Abstract

The article proposes and generalizes a geoecological approach to landslide risk assessment for a study area located in the upper Tereblia River basin within the Ukrainian Carpathians, based on a set of landslide development factors with consideration of the vegetation cover of the territory. Special attention is given to forest cover as a factor in the spatial development of landslides; in particular, its spatial structure, density, spatiotemporal dynamics, and position relative to landslide occurrence and slope location were assessed. The relevance of the study is determined by the fact that most regional landslide susceptibility models rely primarily on morphometric, geological, and anthropogenic factors, whereas the condition of vegetation cover in landslide-prone areas is often assessed only in a binary way, considering merely its presence. For the Carpathian region, such simplification is insufficient, since forest edge, fragmentation, logging, cover disturbance, and vegetation changes significantly affect surface and near-surface runoff, infiltration, the erosional preconditioning of slopes, and, ultimately, the spatial localization of hazardous landslide areas. The aim of the study is to construct a baseline geoecological landslide risk model for a site in the Tereblia River basin on the basis of the developed Landslide Susceptibility Assessment (LSA) model, which, in addition to a range of factors, takes into account the characteristics of forest cover and its exposure. The study used cadastral landslide data from DNVP “Geoinform of Ukraine,” the ALOS AW3D30 digital surface model, the MERIT Hydro dataset, ESA WorldCover maps, Hansen Global Forest Change products, and Landsat and Sentinel-2 satellite data. The modeling was carried out in Google Earth Engine and the QGIS geographic information system. The landslide susceptibility model of the territory includes a morphometric block, a hydrological block, and a forest block. Forest cover was parameterized through distance to forest edge, cover density, NDVI amplitude, forest loss before 2001 and during 2001–2023, as well as the proportion of forest located upslope. It was found that 73% of landslides are confined to a belt within 500 m of the forest edge. The constructed LSA map showed that 35.3% of the modeled area falls within zones of high and very high landslide susceptibility. The scientific novelty lies in the adaptation of an algorithm for calculating landslide susceptibility and landslide risks for the study area within the Ukrainian Carpathians, with explicit consideration of forest cover condition parameters. The practical significance lies in the possibility of using the risk map to prioritize territorial monitoring, plan forestry measures, support sustainable territorial development, and identify sites requiring priority engineering-geological investigation.