Processing of daily Radar and gauge precipitation data for use in semi-distributed rainfall-runoff modelling in Australia

[摘要] Inadequate spatial characterisation of rainfall made from point observations is one of the highest sources of error in catchment water yield models, due to the high spatial and temporal variability of precipitation.To overcome this issue, the spatial and temporal density of Radar rainfall data can be used to provide continuous representation of rainfall patterns in-between rain gauges, and is the strength of Radar data compared to using only rain gauge data to force hydrological models.Despite their great promise, however, Radar rainfall data are not widely used in Australia outside of the Bureau of Meteorology (BoM) where they have primarily been used in urban environments to assist with flood forecasting.Hence there is a unique scientific opportunity for innovative science in the new BoM/CSIRO alliance known as the Water Information Research And Development Alliance (WIRADA) to assess the impact of using Radar rainfall data in water yielding forestry and agricultural catchments.This report describes and comprehensively documents the initial processing steps of daily Radar and gauge precipitation data for use in semi-distributed rainfall-runoff modelling. As such, it serves as a basis for future hydrological research using Radar rainfall data within WIRADA.This report, therefore, does not verify the usefulness of the Radar data in hydrological applications, rather it describes how the data can be accessed and processed in order to do so in the future.Specifically, this report describes: (1) the processing and reformatting of gauge precipitation data necessary for linking it to Radar imagery; (2) the BoM’s Mapbase Radar data format; (3) processing of Radar precipitation imagery; (4) blending of gauge and Radar precipitation; and (5) finally provides a demonstration of semi-distributed rainfall runoff modelling from the blended gauge and Radar product.The computer code written in the interactive data language (IDL) to perform all of the processing is also provided in appendices.The key findings here show that over two years of Radar data from the Yarrawonga station could be semi-automatically processed in order to be blended with rainfall gauges and used in a semi-distributed rainfall-runoff model.Future research should concentrate on maximising the strengths between the Radar and gauge rainfall data sources in order to quantify and subsequently minimise error in daily stream flow predictions.

[发布日期] 2008-12-01 [发布机构] CSIRO

[学科分类] 地球科学(综合)