Developing of High-Resolution Monthly Gridded Temperature Dataset for Afghanistan

Abstract

Without precise temperature data, hydro-climatology analysis is challenging. This thesis developed high-resolution (0.08°x 0.08°) gridded maximum and minimum datasets for Afghanistan spanning 60 years (1961–2021). First, the linear regression technique filled in the observed data gaps using neighboring stations and the gridded datasets (CRU, ERA5, and TerraClimate). Then, with a 0.08° grid resolution, the gap-filled data was extrapolated using the GIS interface and the ordinary kriging (OK) interpolation spatial approach. This study assessed the accuracy of multiple gauge-based and reanalysis datasets, including CRU, ERA5, PGFv2.1, NCEP/NCAR Reanalysis1, and TerraClimate. Standard statistical measures like KGE (Kling Gupta Efficiency), R2 (Coefficient of Determination), and MAE (Mean Absolute Errors) were used to complete the assessment. The more accurate approach—among quantile mapping and linear scaling—was selected to generate a longterm result and bias correct the dataset that performed better. Using the generated dataset, we investigated spatiotemporal annual and seasonal trends using Man Kendel and Sen's Slop estimator techniques. It was clear from the results that CRU was the most accurate available dataset. Moreover, the average of MAE dropped from 3.84 to 1.15 for the maximum temperature, while the average of R2 and KGE increased from 0.93 and 0.33 to 0.94 and 0.95. These results indicate that linear scaling was superior to quantile mapping in the setting of bias correlations. The MAE dropped from 3.48 to 1.17 at the minimum temperatures, but the average of R2 and KGE increased from 0.84 and 0.21 to 0.86 and 0.89. In addition, the annual trend analysis results showed that both the minimum and maximum temperatures were trending upward. The maximum temperatures in Winter (DJF), Spring (MAM), Autumn (SON), and Summer (JJA) increased, according to seasonal trend studies, with the exception of the east and south-west regions, which include Nangarhar, Kunar, Laghman, Nuristan, and the northern districts of Paktika. On the other side, the season trend research shows that the minimum temperature is trending upward for all four seasons: winter (DJF), spring (MAM), summer (JJA), and autumn (SON).