Search

The Central Water Commission (CWC) revised its estimates of water resources availability (WRA) for river basins of India in June 2019 (CWC19). Despite more than six years of modelling effort, several major drawbacks remain – the model needs extensive inputs on water management and seasonal crop growth; model calibration is cumbersome and not very effective; and, it cannot be easily extended to other years or ungauged basins. This paper proposes a conceptually equivalent, but simpler, alternative to estimate WRAusing evapotranspiration (ET) products based on remote sensing or land surface models (ETRS). In this approach, WRAis formulated in terms of ETand necessary assumptions are made to account for ETfrom irrigated areas. The proposed alternative does not involve any modelling, does not require any inputs on water management or crop growth, and is applicable even in ungauged basins. The analysis presented here has two main parts. In the first part, since ETRSproducts are prone to systematic biases they are evaluated for their validity against basin-scale proxy observations of ET. In the second part, the validated ETRSproducts are used to estimate WRAusing the proposed approach. Over Godavari and Krishna river basins of peninsular India, and also 15 sub-basins within these two basins, five state-of-the-art ETRSproducts (ERA5-Land, GLEAM, NTSG/MOD16, NTSG/PLSH and FEWS/SSEBop) are evaluated. At a basin-scale two of these products, ERA5 and SSEBop, are found to have limited bias and reasonably captured annual basin-scale ET. Using these two products, WRAis estimated for the two basins and their 15 sub-basins. Resulting WRAestimates are found to be similar to, and highly consistent with, those from CWC19. Potential causes of discrepancies in WRAbetween CWC19 and this study are discussed, drawbacks of this approach and potential improvements, and next steps in expanding this work to other river basins of India are also discussed.