Multivariate copula quantile mapping for bias correction of reanalysis air temperature data

Reanalysis data retrieved from the European Centre for Medium-range Weather Forecasts (ECMWF) are commonly used for hydrological studies. Their use requires bias correction, defined as the difference between reanalysis values and measurements. We propose three multivariate copula quantile mappings (MCQMs) to predict bias-corrected values at unvisited locations. We apply the methods to the Qazvin Plain, Iran, for daily air temperature retrieved from weather stations and the ECMWF archive. Results showed that MCQMs reduced bias by 46% as compared with classical quantile mapping. The study concludes that MCQMs are well able to represent the spatial and temporal variation of air temperature.