An Effective Flood Forecasting System Based on Web Services

Abstract

A flood forecasting system usually needs to integrate many hydraulic modules, which may be legacy programs written in FORTRAN, running in heterogeneous environments, and differing in execution time. Besides, the data required for each module should be provided in real time, and the programs also need to be executed in a correct sequence. In this paper, we discuss how to build the flood forecasting system based on Web services to support module integration and data exchange. We also propose several workflow strategies for composing services, and perform experiments to determine which strategy is better.