DEVELOPING HIGH RESOLUTION MONTHLY GRIDDED PRECIPITATION DATASET FOR AFGHANISTAN

Abstract

Being a developing country, Afghanistan faces many challenges in managing its water resources. The analysis and management of water resources have been greatly hampered by the absence of high-resolution precipitation data. Using observational data from 68 stations over a 24-year period (1961-1984), this study attempts to produce a long-term (60 years) high spatial resolution 0.08 ° monthly bias-corrected gridded precipitation dataset for Afghanistan. In addition to taking the relationships between adjacent stations into account, gridded and reanalysis products containing data records for the relevant period were chosen for the monthly statistical comparison with the observed data. Through the application of concerned linear regression equations, the datasets that produced the best relationships were employed to close the gap in the base period (1961–1984). With a grid resolution of 0.08 degrees, the gape-filled data was interpolated on the GIS interface using the ordinary kriging (OK) spatial interpolation tool. Following the creation of the observed reference dataset, the GPCC, Aphrodite, ERA5- Land, and University of Delaware gridded datasets were subjected to commonly used statistical measures for the evaluation of their correctness on a grid-to-grid basis for the selection of the best dataset for bias-correction of the onwards data. Before applying monthly bias-correction factors to the remaining data, linear scaling and quantile mapping approaches were statistically assessed. Additionally, the study intends to assess the accuracy of two recently observed data from the Afghanistan Meteorological Department (AMD) and Agromet utilizing records from 21 and 16 stations during periods of 14 and 10 years, respectively. Using the Mann-Kendel test and Sen's slope methodology, the created long-term observed precipitation record was used to identify the annual and seasonal spatial trend. The findings revealed that the GPCC outperformed alternative datasets across the entirety of Afghanistan, with the exception of a few regions in the central west, mainly Bamiyan and Daikundi, as well as eastern Badakhshan and, southern Paktika. The average biased R2 and KGE between the observed and GPCC increased from 0.80 and 0.51 to 0.85 and 0.88, respectively, thanks to linear scaling, which was also able to reduce the mean absolute error from 9.25 to 6.35. The dataset developed was further used for the Annual and seasonal trend analysis over Afghanistan. In eastern, southern, and central parts of Afghanistan, 40.40 % of the area produced a significant increase in annual precipitation ranging from (0.78 - 16) mm/year, while decreases from (-0.31 -11.85) mm/year were observed in Baghlan, Bamian, Sari pul, Ghor Zabul, Badghis, Southern Paktika, and eastern Badakhshan. Along the Pakistani border, where the monsoon rains fall, and in elevatedvii places, where the most precipitation falls, a growing tendency was noted. Mostly Afghanistan's eastern, central, and southern regions have significant increasing trends during the summer (JJA) and autumn (SON), ranging from 0.15 to 5.2 mm/year, while other parts of the nation experience little to no decreasing trends in these seasons. The winter months in Kunar saw the greatest seasonal rise in precipitation, up to 5.25 mm/year, with the summer months in eastern Paktika seeing an increase of 5.15 mm/year. Engineers and scientists may use the dataset created, evaluation of gridded datasets, and trend analysis as a tool to predict future climatic scenarios, address water management concerns, and analyse water availability, supply, and demand using various hydro climatological tools.