Spatio-temporal characteristics of blended radar/gauge data and the estimation of design rainfalls

Radar data provide consistent spatial and temporal information that cannot be readily gleaned from gauge information alone. As part of the revision of Australian Rainfall and Runoff, and under the auspices of Engineers Australia, the Bureau of Meteorology is currently undertaking a project to evaluate the realism of downscaled rainfall extremes for the Greater Sydney region. This evaluation is based on spatio-temporal characteristics derived from blended radar and gauge rainfall accumulations (Rainfields). Information about the spatio-temporal characteristics of precipitation - especially for large events - is desirable for a multitude of other activities related to the analysis of design rainfall estimates and potential changes in these estimates under climate change. These applications include the choice of realistic parametrisation schemes in the downscaling of rainfall extremes from Global Climate Models, the choice of spatial parameters in the statistical modelling of extremes and in assessing the realism of areal reduction factors depending on the choice gridding technique. This paper outlines algorithms used in blending radar and gauge data, the analyses undertaken to remove artefacts, and the analysis of errors in the resulting product of radar rainfall accumulations to provide information to users on the quality of these products. We discuss spatio-temporal characteristics of radar data suitable for the analysis of rainfall extremes. To help quantify the contribution radar data can make in analysing and specifying models for rainfall extremes we will undertake comparisons of these characteristics based on the blended data and from gauge data alone. Finally, we provide information about how to acquire blended radar rainfall accumulations from the Bureau of Meteorology.