High-Precision AI-Enabled Flood Prediction Integrating Local Sensor Data and 3rd Party Weather Forecast

Area of Science:

• Environmental Science
• Computer Science
• Urban Planning

Background:

• Sea-level cities face significant flooding risks.
• Kristianstad, Sweden, utilizes a vast network of wireless sensors for real-time environmental monitoring via an Internet of Things (IoT) portal.
• Existing monitoring systems lack predictive capabilities for early flood warnings.

Purpose of the Study:

• To develop a real-time flood forecast system.
• To leverage massive sensor data and third-party weather forecasts for accurate predictions.
• To enable early response mechanisms for decision-makers.

Main Methods:

• Development of a smart flood forecast system using machine learning and artificial neural networks.
• Integration of data from multiple sources, including a city's IoT portal and external weather services.
• Implementation as a software product integrated with the existing IoT infrastructure.

Main Results:

• The developed system accurately forecasts floods at distributed locations for the coming days.
• Successful integration of multi-source data into a unified forecasting model.
• Significant extension of the city's IoT infrastructure capabilities beyond basic monitoring.

Conclusions:

• The AI-enabled flood forecast system provides accurate, real-time predictions.
• This represents the first large-scale, AI-powered, real-time flood forecast system deployed in a real-world urban environment.
• The system enhances urban flood risk management and decision-making capabilities.