Abstract:
Reliable spatial and temporal meteorological estimates are essential for accurately
modeling hydrological, ecological, and climatic processes. High-resolution gridded
datasets can be used for such applications, especially in data-sparse regions. However, the
validation of the accuracy of these products is necessary before their application towards
hydrological modeling for the assessment and management of water resources. In this
study, high resolution (0.08°×0.08°), long-term station-based gridded datasets for
precipitation, maximum and minimum temperature were developed for the Potohar region.
Linear regression analysis was performed against the datasets and nearby stations for gap
filling in the base period. The gap-filled observed data was spatially interpolated using the
Ordinary Kriging technique to get an observed gridded dataset. Two gauge based and one
reanalysis datasets, both for precipitation and temperature, were selected for evaluation in
this study. GPCC, APHRODITE, and ERA5-Land for precipitation, while CPC, CRU, and
ERA5 for temperature were selected. Performance evaluation was made using widely used
statistical parameters (KGE, R2, MAE, and RMSE). Bias correction was performed by
selecting the better technique between Linear Scaling and Quantile Mapping. The results
revealed GPCC and ERA5 as the better performing datasets for precipitation and
temperature, respectively, among all the selected datasets. Furthermore, linear scaling
performed better than quantile mapping for bias correction. Finally, the GPCC and ERA5
datasets were bias corrected to develop the final gridded dataset products for precipitation
and temperature, respectively. This dataset would be helpful in future hydrological and
climate change impact studies. Moreover, it would be utilized to forecast the Potohar
region's water availability
