| Back

A novel Web-based system for typhoon analysis and prediction

In collaboration with the Department of Computer Science and Engineering at CUHK, Professor Yee Leung and his team have completed a project on the system development for typhoon analysis and prediction. The overall objective of this project is to build a powerful platform for the analysis, tracking, prediction and visualization of typhoon tracks, landfalls, and recurvatures through dynamic modeling and data mining in integrated multisource, multi-scale and multi-level typhoon data. The prototype will be employed as a basis for further development of a full-fledged system for professional use and general public consumption.

The novelty of the system lies on its comprehensive and flexible architecture that integrates seamlessly multisource, multiscale and multilevel typhoon related data with powerful data mining algorithm and dynamic models for typhoon analysis and forecasting with user-friendly interface. The system is developed using Adobe Flash and ActionScript 3.0 on top of the latest Google Map. Tracks of tropical cyclones are queried from database and plotted onto Google Map with fast and frequent displays. Users can access it through web browser at any time and in any place.

Atmospheric data about typhoons such as maximum sustained wind, sea level pressure, wind shear and wind fields, rainfall, sea surface temperature and humidity, etc. collected from JTWC (Joint Typhoon Warning Center), GFS (Global Forecasting System), NCEP (National Centers for Environmental Prediction), NOAA and NASA in the western North Pacific and South China Sea region are comprehensively integrated into one single system (Fig. 1).

Through the system, typhoon analysis can be carried out via super-imposed layers (Fig. 1), query by Similar Path, query by Direction, query by Turning Angle, Genesis Analysis (Fig. 4), Key Area Analysis (Fig. 4), and query by Landing Location. Data mining algorithms (Fig. 2) and dynamic models (MM5) with parallel computation (Fig. 3) are implemented for typhoon prediction.
In brief, novel methods and advanced information technologies are employed to build this typhoon analysis and prediction system suitable for academic research, practical application, and public education. It is a state-of-the-art system with high academic value and great potential for technology transfer.

Fig.1 Super-Imposed Layers	Fig.2 Data mining-based prediction scheme	Fig.3 Dynamic model-based prediction scheme	Fig.4 Key area analysis and genesis analysis