Abstract:
With the advent of numerous global precipitation estimates (GPEs) in the recent decades, dependability of hydrologists has lessened on the station data as the GPEs can be readily availed and utilized. Since the skills of GPEs may differ from region-to-region, it is vital to analyse their ability in resolving the regional precipitation climatology using appropriate statistical methods. In this study, a total of five GPEs, viz., APHRODITE, PERSIANN-CDR, CHIRPS, CMORPH, and IMERG were evaluated for their abilities in resolving regional precipitation climatology of WHR with respect to gridded precipitation product of India Meteorological Department (IMD). Different performance indicators i.e., Probability of Detection (POD), False Alarm Ratio (FAR), Normalised Root Mean Square Deviation (NRMSD), Pearson Correlation Coefficient (CC) and Skill Score (SS) were used for evaluating the GPEs. Multicriterion Decision Making (MCDM)approaches i.e., Compromise Programming (CP), Cooperative Game Theory (CGT), Technique for Order of Preference by Similarity to Ideal Solution (TOPSIS), Weighted Average Technique (WAT), and Fuzzy TOPSIS were used for ranking the GPEs across different grids in WHR. Entropy based weight assignment to NRMSD, CC, and SS were performed while applying them in MCDM methods. Group Decision Making (GDM) approach utilizing spearman correlation coefficient and additive ranking rule was employed to obtain the final ranking of GPEs from multiple rankings assigned through different MCDM methods. Across 115 grids, APHRODITE exhibits superior performance compared to other GPEs in 89 grids. Conversely, CHIRPS and CMORPH emerge as the least favorable products among the five GPEs across more than 70 grids, being consistently ranked either 4th or 5th. Notably, IMERG was identified as the best-performing product in 14 grids and as the second-best product in 63 grids, positioning it as the second most suitable option after APHRODITE for monthly rainfall time series analysis. Similar results, as detailed in the paper, were also obtained for month-wise rainfall time series analysis
