Workshop 1 Flood forecasting with reference to global change

Presenters in workshop 1 had discussed a variety of issues and strategies based on their regional issues and opportunities. They chose solutions that reflected their local needs and realities and considered issues in model
complexity.

- In tropical regions where relatively persistent weather patterns may be found, monthly or seasonal precipitation predictions may be generated through studies of teleconnections and extrapolation of longer term cycles of rainfall and other climatic variables.

- For relatively data-rich regions, a framework was presented for assessing risk of flood damages from initial precipitation inputs and generation of runoff through to risk (and implications) of dyke failure.

- Other presenters discussed issues in model selection such as determining optimal resolution requirements and suggesting that soft computing approaches may be considered as a option to avoid unnecessary complexity and
costs.

- Also shown was a method to use downscaled numerical weather prediction model data output instead of historical meteorological records to improve ensemble streamflow forecasts. An example of a coupled
atmospheric-hydrological model which may provide feedback to eventually improve precipitation forecasts and subsequent streamflow predictions was shown as well.

 
 Paul Campbell