作者: A. Berne , V. Boudevillain , P. E. Kirstetter , G. Delrieu
DOI:
关键词:
摘要: summary Characterisation of the error structure radar quantitative precipitation estimation (QPE) is a major issue for applications technology in hydrological modelling. Due to variety sources QPE and impact correction algorithms, problem can only be addressed practically by comparison QPEs with reference values derived from ground-based measurements. Using raingauge datasets Bollene-2002 experiment, preliminary investigation this subject has been carried out within geostatistical framework context Cevennes–Vivarais Mediterranean Hydrometeorological Observatory. First, measurements were critically analysed using variograms detect erroneous squared differences between pairs values. The anisotropic block kriging technique was then used compute select most reliable statistical distribution spatial temporal residuals established analysed. variance separation concept also tested estimate estimates true unknown rainfall. A version model 1-km