Abstract
Flash-floods are recurrent natural hazards producing substantial impacts to the human communities. The increase in the frequency and magnitude of flash-floods is associated with the last decades of climate change at a global scale as well as changes affecting the Land Use/Land cover (LULC) characteristics. We employ Machine Learning technniques (Random Forest (RF), Multi layer Perceptron based Artificial Neural Network (MLP-ANN) and Geographical Information System (GIS) framework to deduce the spatial patterns of relationship between the flash-flood potential changes across the study area and the changes in the modified LULC during 2000-2019. High resolution Digital Elevation Model (DEM) and daily precipitation from AgERA5 will be utilized to detect the spatial variability of flash floods through the modified Founier Index. Identifying the probable changes in specific roughness coefficient (through the modified LULC), a geospatial risk based (flash flood potential of high, moderate and low categories) indices will be developed to assess future hydrological hazards.  
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call