Development of a Localized Radar-Rain Gauge Co-Kriging QPE Scheme for Potential Use in Quality Control of Real-time RainfaH Data

摘要

@@ l. Introduction Quality control (QC) is required before automatic rain gauge (RG) data, e.g. typical tippingbucket RG, can be used quantitatively due to various types of systematic and random errors. caused by various factors including wind. wetting, evaporation, splashing, calibration, finite sampling, mechanical failure, funnel blockage, signal transmission interference, power failure, etc (Habib, 2001 ). Among the random errors, those arising from contamination during data transmission through radio telemetry systems and blockage of the rain gauge itself by extemal obstructions, like insects or fou ling by birds. are crucial in he avy rain m onitoring and assessment. While telemetry errors usually contaminates a single observation with an extreme value, a blocked gau ge will retum zero or un reasonably small values that fall well within the climatological range and are therefore difficult to detect. A QC procedure based on spatial consistency checking is proposed in this paper. Major techniques employed include a lo calized quantitative precipitation estimation (QPE) scheme based on the co-Kriging of radar and rain gauge data, as well as statistical analysis of the RG-QPE residuals in search for suitable QC criteria.