Comparison of Different Artificial Intelligence Techniques to Predict Floods in Jhelum River, Pakistan

Floods are among the major natural disasters that cause loss of life and economic damage worldwide. homes, crops, roads, basic infrastructure, forcing people to migrate from high flood-risk areas. However, due a lack information about effective variables in forecasting, development an accurate flood forecasting system remains difficult. The flooding process is quite complex as it has nonlinear relationship with various meteorological topographic parameters. Therefore, there always need develop regional models could be used effectively for water resource management particular locality. This study aims establish evaluate data-driven Jhelum River, Punjab, Pakistan. performance Local Linear Regression (LLR), Dynamic (DLLR), Two Layer Back Propagation (TLBP), Conjugate Gradient (CG), Broyden–Fletcher–Goldfarb–Shanno (BFGS)-based ANN were evaluated using R2, variance, bias, RMSE MSE. values best-performing LLR model 0.908, 0.009205, 1.018017 training 0.831, −0.05344, 0.919695 testing. Overall, performed best both validation periods can prediction floods River. Moreover, provides baseline early warning area.