HaorFloodAlert: Deseasonalized ML Ensemble for 72-Hour Flood Prediction in Bangladesh Haor Wetlands
A new machine learning model called HaorFloodAlert has been developed to predict flash floods in Bangladesh's haor wetlands. This model forecasts flood probabilities for up to 72 hours and addresses the unique backwater dynamics of the region. It has achieved high accuracy rates and includes a damage estimator for the boro rice harvest.
- ▪HaorFloodAlert is designed to predict flash floods in Bangladesh's haor wetlands, which are prone to sudden flooding.
- ▪The model utilizes a deseasonalized machine learning ensemble to forecast flood probabilities for the Sunamganj Haor region.
- ▪It has achieved an operational accuracy of 89.6 percent and includes a calibrated damage estimator for the boro rice crop.
Opening excerpt (first ~120 words) tap to expand
Computer Science > Artificial Intelligence arXiv:2605.20167 (cs) [Submitted on 19 May 2026] Title:HaorFloodAlert: Deseasonalized ML Ensemble for 72-Hour Flood Prediction in Bangladesh Haor Wetlands Authors:Salma Hoque Talukdar Koli, Fahima Haque Talukder Jely, Md. Samiul Alim, Md. Zakir Hossen View a PDF of the paper titled HaorFloodAlert: Deseasonalized ML Ensemble for 72-Hour Flood Prediction in Bangladesh Haor Wetlands, by Salma Hoque Talukdar Koli and 3 other authors View PDF HTML (experimental) Abstract:Flash floods in Bangladesh's haor wetlands show up with almost no warning. They wreck the annual boro rice harvest. Current setups, built for riverine floods, miss backwater dynamics entirely. These basins are flat. Water does not behave like it does on the Brahmaputra.
…
Excerpt limited to ~120 words for fair-use compliance. The full article is at arXiv cs.AI.
Discussion
0 commentsMore from arXiv cs.AI
