Comparison of bias correction methods in the arid region of Pakistan

Abstract This study presented the effectiveness of univariate quantile mapping (QM) bias correction and multivariate (MBCn) methods by comparison with the European Land Reanalysis (ERA5) gridded dataset in the arid region of Pakistan. The study Girst bias-corrected rainfall (Pr), maximum temperature (Tmax), and minimum temperature (Tmin) and compared bias-corrected variables with ERA5 variables. The climate indices such as the potential evapotranspiration (PET), aridity index (AI), and multivariate drought index obtained using both methods were also compared with ERA5-based indices. The results revealed the MBCn method performed well in bias-correcting temperature and rainfall in comparison to QM. It also performed better estimating PET and aridity in the region and found a similar spatial pattern to ERA5. However, QM performed better than MBCn in the time-dependent index (multivariate drought index) at 3-, 6-, and 12-months and showed similar patterns of frequency of occurrence for moderate drought. The major Gindings of this study indicated that the multivariate method (MBCn) is more reliable in climate indices independent of temporal properties in the region, whereas the QM method performed well in time-dependent climate indices. This study would be helpful for future researchers to select suitable methods for bias correction of climate variables in the region.