Abstract
Radar rainfall estimates are widely used for real-time flood forecasting. An effective and accurate early warning system, with the assistance of radar rainfall estimates and hydrological modelling, is vital for reducing the effects of flood related-hazards. Detailed spatial and temporal data on the distribution of rainfall over a wide range area can be provided by weather radar. However, rainfall data captured at a single point location remains inaccurate. The combination of two datasets, radar and ground gauges, therefore constitutes the optimal method to estimate rainfall in a river basin.
This study evaluated the rainfall data-combining technique of conditional merging (CM) in the Jinzu River Basin. This was then applied as the rainfall input for flood forecasting. A distributed hydrological model, developed by the International Center for Water Hazard and Risk Management (ICHARM) known as the Integrated Flood Analysis System (IFAS), was established and applied to evaluate the performance of the CM method for flood forecasting relative to the original radar simulations. Numerous cases with two sets of input data (radar and conditional methods) were tested during the period of 2011-2014. Here, the CM method was evaluated, and the applicability of a distributed model to flood forecasting in the Jinzu River Basin was quantitatively examined.
