What question did this study set out to answer?

The central aim is to enhance river water level forecasting accuracy using advanced machine learning techniques in resource-limited settings.

March 14, 2026Open Access

From Climate Signals to Smart Predictions: Advanced Machine Learning and Deep Learning for Hydrological Modeling in Data-Scarce Bangladesh

Key Points

The central aim is to enhance river water level forecasting accuracy using advanced machine learning techniques in resource-limited settings.
Evaluated five machine learning and deep learning models including RF and LSTM.
Used 26 years of hydrological and meteorological data for training and testing.
Optimized model hyperparameters with grid search and 5-fold cross-validation.
Incorporated lagged variables to account for temporal dependencies.
Assessed model performance using six metrics and ranked them via principal component analysis.
Random Forest achieved the highest accuracy in water level-only scenarios (MAE 0.1445 m, R² 0.9916).
LSTM showed superior performance in climate-driven scenarios with R² values of 0.8145 and 0.8064 for rainfall-temperature and rainfall-only, respectively.
Temperature-only scenarios had limited predictive capability (R² 0.5346).
Cross-station performance without recalibration demonstrated strong spatial transferability (LGBM: MAE 0.1205 m, R² 0.9917).
The framework supports effective flood forecasting in data-scarce environments like Bangladesh.

Abstract

Reliable river water level forecasting is crucial for flood management and sustainable water resource planning in climate-vulnerable deltaic regions like Bangladesh, where conventional hydrological models face challenges due to data scarcity and complex monsoon-driven dynamics. This study evaluates five machine learning and deep learning models—Linear Regression (LR), Random Forest (RF), XGBoost, Light Gradient Boosting Machine (LGBM), and Long Short-Term Memory (LSTM)—to predict daily water levels in the Old Brahmaputra River under four scenarios. Using 26 years (1999–2024) of hydrological and meteorological data, we optimized hyperparameters via grid search and 5-fold cross-validation, incorporating lagged variables (1–5 days) for temporal dependencies. Performance was assessed using six metrics, with principal component analysis for model ranking. Results revealed exceptional accuracy in the water level-only scenario (RF: MAE 0.1445 m, R² 0.9916, NSE 0.9916). Climate-driven scenarios demonstrated LSTM's superiority, achieving R² of 0.8145 (rainfall-temperature), 0.8064 (rainfall-only), while temperature-only scenarios showed limited predictive capability (R² 0.5346). Spatial transferability assessment at Sarishabari station validated robust cross-station performance without recalibration (LGBM: MAE 0.1205 m, R² 0.9917 for water level scenario). The study provides operational frameworks for climate-driven forecasting in data-scarce deltaic environments, highlighting LSTM's exceptional capability for ungauged catchments. • RF excels with hydrological memory, while LSTM leads in climate-scenario accuracy • LSTM enables operational forecasts in ungauged basins using weather data • Multivariate rainfall-temperature inputs beat single-variable models • Spatial transferability confirmed with robust cross-station performance without recalibration • Framework supports flood forecasting in data-scarce networks like Bangladesh

Read Full Paperexternally

KI fragen

Bookmark

View Full Paper

Cite This Study

Islam et al. (Sun,) studied this question.

synapsesocial.com/papers/69b4fa6fb39f7826a300b304 https://doi.org/https://doi.org/10.1016/j.htopen.2026.100005

Also Consider

Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context:

KI fragen

Bookmark

View Full Paper