Uncertainty features of weather radar rainfall
Rainfall information plays a crucial part in predicting the occurrences of debris flows. Due to the limitation of its spatial resolution, the traditional rain gauge network is increasingly challenged by the weather radar in rainfall measurements. However, despite its popularity among academic water engineers, the weather radar has not been recognised by many applied water engineers as a viable alternative rainfall device to rain gauges. The authors believe that the uncertainty nature of the weather radar is one of the major inhibitors in its wider applications. This paper describes a study on the uncertainty features of weather radar rainfall based on its error characteristics (bias, variance and temporal autocorrelation). Adense rain gauge network over Brue catchment of SWEngland (49 gauges over 136 km2) and two nearby C-band radars are used to derive these three important features required for the development of a realistic error model of radar rainfall data. It has been found that error bias and variance are variable depending on the rainfall magnitude. The error time series shows a significant autocorrelation among its data points. It has been found that no existing error models are able to address these features, hence there is an urgent need for more research work in this area.
rainfall measurement weather radar error characteristics modelling
D. Han I. Cluckie
Water and Environmental Management Research Centre, Department of Civil Engineering, University of Bristol, Bristol, UK
国际会议
成都
英文
23-30
2007-05-17（万方平台首次上网日期，不代表论文的发表时间）