Statistical corrections of spatially interpolated missing precipitation data estimates

摘要: Spatial interpolation methods used for estimation of missing precipitation data generally under and overestimate the high low extremes, respectively. This is a major limitation that plagues all spatial as observations from different sites are in local or global variants these data. study proposes bias-correction similar to those climate change studies correcting estimates provided by an optimal method. The applied post-interpolation using quantile mapping, variant equi-distant matching new single best estimator (SBE) scheme. SBE developed mixed-integer nonlinear programming formulation. K-fold cross validation correction carried out 15 rain gauges temperate climatic region U.S. Exhaustive evaluation bias-corrected several statistical, error, performance skill score measures. differences among methods, effectiveness their limitations examined. method based on recommended. Post-interpolation bias corrections have preserved site-specific summary statistics with minor changes magnitudes error were found be statistically insignificant parametric nonparametric hypothesis tests. improved scores minimal extreme indices. estimated also brought serial autocorrelations at lags transition states (dry-to-dry, dry-to-wet, wet-to-wet wet-to-dry) close observed series. Bias provide better serially complete time series useful variability comparison uncorrected filled Copyright © 2013 John Wiley & Sons, Ltd.