Flood Modeling Using Artificial Intelligence and Machine Learning Techniques

(North Kabir River Case Study in Syria)

Abstract

The North Kabir River Basin has been subjected to frequent floods that have led to environmental, social, and economic disasters in the region. This research aims to identify important factors in flood risk management, in the context of accelerated climate change. The Geographic Information System (GIS) was used to analyze spatial data related to floods, where Machine Learning techniques, which are part of Artificial Intelligence, were integrated to analyze these data and extract patterns and predictions, through the AutoML environment available in ArcGIS Pro, which allows for optimal model selection and automatic parameter tuning (parameter optimization), and using a set of machine learning algorithms including: Linear regression, decision tree, random trees, absolute gradient enhancement, light gradient enhancement, and additional trees, and the analysis relied on sixteen environmental and geographical parameters that represent the most important factors contributing to the generation of floods, namely: Elevation, steepness, linear surface curvature, runoff accumulation, slope length, slope-length factor, length-slope factor, rainfall amount, topographic moisture index, terrain ruggedness index, terrain condition index, current strength index, drainage network density, distance from riverbeds, Normalized Difference Vegetation Index, and land use and coverage, and the performance of the different models was evaluated using multiple accuracy measures such as sensitivity, recall, and F1 score. The results showed the superiority of the linear regression model in recognizing patterns associated with flood risk, while the resulting maps showed that high-risk areas are concentrated in the south and southwest of the basin, while low-risk areas are distributed in the north and northeast. This study reflects the importance of employing advanced scientific methods to analyze floods, contributing to the promotion of sustainable water resources management strategies.