Integrating Multicriteria Environmental Factors and Spatially Cross-Validated Machine Learning for Synthetic Erosion Vulnerability Assessment in the Betsiboka Region, Madagascar
Domain:
environment and energygeospatialRecord type:
paper
Creator:
RasRakTovRaz
Publisher:
Int
Host:
This study integrates multicriteria environmental analysis and spatially cross-validated machine learning to assess
synthetic erosion vulnerability in the Betsiboka Region, Madagascar. The vulnerability index combined low NDVI, slope,
hydrographic proximity, elevation, and available geological, soil, rainfall, and land-cover factors on a 1 km grid.
Hydrographic proximity was modelled using an exponential distance-decay function with a characteristic distance of 3 km,
while geology was represented through normalized susceptibility scores ranging from 0 to 1. Three regression algorithms—
Random Forest, Extra Trees, and Histogram Gradient Boosting—were evaluated using five-fold spatial block crossvalidation.
Similar
Integrating Machine Learning and Geospatial Data for Mapping Socioeconomic Vulnerability to Urban Natural HazardMalaria risk assessment through Remote Sensing and MultiCriteria Evaluation in MadagascarDeforestation, Erosion, and Fire: Degradation Myths in the Environmental History of MadagascarLandslide hazard assessment in the heterogeneous geomorphological and environmental context of the rif region, morocco – A machine learning approachPerformance Assessment of Machine Learning Techniques for Gully Erosion Mapping in South-East NigeriaMachine learning models for gully erosion susceptibility assessment in the Tensift catchment, Haouz Plain, Morocco for sustainable development