The water year based on minimizing the terrestrial water storage variation and its validity

15 catchments in the Chinese Loess Plateau (CLP) The concept of the water year has been widely used in analyzing the coupled water-energy balance at the annual scale. However, a comprehensive quantitative investigation about its determination and usability in catchment water balance analysis remains absent. Leveraging long-term hydrometeorological observations and terrestrial water storage ( TWS ) data, we ascertained the commencement month of the water year for 15 catchments in the CLP and explored the feasibility of the Budyko framework at the water year scale concerning evapotranspiration ( ET ) estimation and its change attribution. (1) TWS recharge commenced predominately in July, with the minimal TWS variation (∆ S ) consistently observed for years starting from July. Consequently, July marked the start of the water year for CLP catchments. (2) Disregarding ∆ S at the calendar year scale caused significant errors in estimating ET (MAE = 51.16 mm yr −1 ), while the MAE decreased by 35.40% at the water year scale. (3) The relative contributions of precipitation ( P ), potential evapotranspiration ( ET 0 ), and w to ET change deviated significantly from the actual values when ignoring ∆ S at the calendar year scale. Conversely, these contributions aligned closely with actual values at the water year scale. Overall, the improved accuracy in ET estimation and enhanced consistency in ET change attribution results collectively justify the water year temporal framework. • Annual |∆ S | decreases to its minimum only at the water year scale. • The water year begins in July for the CLP catchments. • ∆ S affects ET change less at the water year scale due to its smaller variation. • The water year is adoptable for the Budyko framework in the interannual process.