Integrated Application of Deep Learning Algorithms and Geographic Information Systems in Forecasting and Intelligent Management of Seasonal Floods: A Case Study of Nooshabad, Kashan

Integrated Application of Deep Learning Algorithms and Geographic Information Systems in Forecasting and Intelligent Management of Seasonal Floods: A Case Study of Nooshabad, Kashan

seyed jamal yadollahi1

1) Graduate of Industrial Management from Kashan Azad University and employee of Noushabad Municipality

IOI Code : ABFF-ABDGZG
Publication : 9th International Congress of Developing Agriculture, Natural Resources, Environment and Tourism of Iran(9icsda.ir)
Abstract :
    Seasonal floods in semi-arid regions of Iran, such as Noushabad downstream of Kashan, pose a serious threat to urban infrastructure, public safety, and the environment.This study presents an integrated model based on deep learning algorithms (LSTM networks) and geographic information systems (GIS) for accurate flood discharge prediction and hazard zoning in the watershed of the seasonal Sok Cham and Sili rivers of Kashan.Using 30 years of hydrological data (1374-1404), Sentinel-2 satellite images, and the HEC-RAS hydrological model, the proposed model achieved 94.6% accuracy (R²=0.946) in predicting flood discharge, which is significantly superior to the traditional HEC-RAS model (R²=0.723).This model, by designing an Early Warning System (EWS) and Health, Safety, and Environment (HSE) protocols, enabled a reduction in infrastructure damage by up to 68% and environmental pollution by up to 75%.The results have practical applications for Noushabad urban services, including targeted dredging and emergency management, and can be generalized to other small cities in Iran.

Keywords : seasonal flood deep learning lstm network crisis management hse