Forecasting of extreme flood events using different satellite precipitation products and wavelet-based machine...

Area of Science:

• Hydrology and Water Resources
• Climate Science
• Artificial Intelligence

Background:

• Accurate forecasting of extreme weather events like floods is crucial for disaster preparedness and impact mitigation.
• Traditional flood forecasting models struggle in regions with limited ground-based rainfall data (sparse rain gauges).
• Satellite precipitation data offers a viable alternative for hydrological monitoring in data-scarce areas.

Purpose of the Study:

• To develop a reliable, simple, and accurate flood forecasting model applicable in data-scarce regions.
• To integrate satellite precipitation data with machine learning for improved extreme flood event prediction.
• To enhance the lead time of extreme flood forecasts.

Main Methods:

• Development of a novel forecasting method combining satellite precipitation products with wavelet-based machine learning models.
• Application and testing of the proposed approach in the flood-prone Vamsadhara river basin, India.
• Comparison of the developed model's performance against benchmark models for extreme flood event forecasting.

Main Results:

• The proposed method demonstrates promising capabilities for forecasting extreme flood events.
• The approach shows potential for achieving longer lead times in flood predictions compared to existing models.
• Validation in the Vamsadhara river basin indicates the model's effectiveness in a real-world scenario.

Conclusions:

• The integration of satellite precipitation and wavelet-based machine learning offers a robust solution for flood forecasting in data-scarce regions.
• The developed model shows significant potential for improving preparedness and mitigating the impacts of extreme flood events.
• This approach provides a valuable tool for enhancing the timeliness and accuracy of hydrological disaster warnings.