AUTOMATED MONITORING OF RIVERBANK EROSION AND ASSESSMENT OF ANTHROPOGENIC IMPACTS USING REMOTE SENSING DATA

Abstract

The intensification of river deformation due to human activities necessitates the implementation of operational monitoring methods for riparian zones. Traditional hydromorphological surveys are fragmented and costly, making it imperative to transition to automated monitoring technologies based on open Earth remote sensing data. The study examines the processes of monitoring river erosion and identifying factors contributing to its intensification. The purpose of the article is to improve the accuracy of determining the dynamics of riverbank erosion and the objectivity of assessments of anthropogenic impact by automating the analysis of ES data to inform management decisions regarding rational land use. Research objectives: to justify the insufficient informativeness of the traditional channel width indicator for assessing erosion processes; to propose a new method for automated monitoring of shoreline dynamics and the extent of floodplain plowing; to investigate the statistical relationship between the intensity of shoreline changes and agricultural land use in riparian areas, thereby confirming the practical effectiveness of the developed method. The research methods are based on systems and functional analysis of information processes, GIS analysis, object-oriented modeling using UML, and statistical, correlation, and regression analyses. The results include the development of a conceptual model and the formalization of the automated shoreline monitoring method's structure. The method integrates transect analysis algorithms with pixel-level land-cover classification using Sentinel-2 satellite imagery, enabling adaptation of floodplain research to conditions of limited field access. The proposed approach allows for analyzing geomorphological metrics in spatial connection with the average areas of plowed land within the surveyed segments. Practical validation of the method on a test site along the Siverskyi Donets River confirmed its feasibility. In particular, it was mathematically proven that 57.5% of the variance in channel deformation rates is directly attributable to the level of floodplain plowing, with extreme erosion reaching up to 53.1 m. The created thematic maps provide a clear visualization and rapid identification of critical “red zones” of channel erosion. Conclusions. The proposed method automates channel-process detection using GIS tools, minimizing subjective expert assessments in river hydromorphology and human error during satellite data processing, while enabling a transition from periodic field surveys to continuous monitoring. The obtained results serve as an effective practical tool for digitalizing environmental management and creating harmonized national geospatial datasets for Ukraine's river basins.