Involving Turc-Budyko formula in evaluating gridded precipitation datasets in glaciated catchments

Abstract

Recently, gridded precipitation datasets (GPDs) have become an essential input in hydrological modeling. Evaluation of their performance is a priori step before their application. When they are applied to a glaciated catchment, the glacier mass balance adds another dimension to the evaluation. This study has attempted to introduce the Turc-Budyko formula for validating the GPDs in glaciated catchments. This approach can identify the over- or under-estimation of catchment precipitation input by the GPDs. As a case study, the approach was applied in Upper Indus Basin (UIB), located in High Mountain Asia. The Turc-Budyko formula and other commonly used evaluation techniques found that APHRODITE underestimated the precipitation by 40 %, although it captured most of inter- and intra-annual variations of local climatology (CC > 0.6). CFSR and HAR were overestimated (106 % and 77 %) and characterize UIB as “Leaky” catchment, whereas underestimated GPDs characterize UIB as “Gaining” catchment. In leaky conditions, glacier storage changes were positive (0.21–0.37 m w.e. yr−1); while gaining conditions made negative glacier storage changes (-0.44 to −0.34 m w.e. yr−1) in UIB. None of the GPDs represented true conditions of glacio-hydrology in UIB. Introducing the Turc-Budyko formula confirmed that direct application of GPDs in hydrological modeling is implausible because GPDs generally cannot produce rational output of streamflow and mass balance simultaneously. It is recommended to locally correct the GPDs before any hydrological application in glacierized catchments.