Assessment of long-term water stress for ecosystems across China using the maximum entropy production theory-based evapotranspiration product

Water demand growth coupled with its high spatial-temporal mismatch of water resources will lead to an increasing water scarcity worldwide. In order to investigate a robust long-term water stress for ecosystems and regions across China, the improved maximum entropy production (MEP) method was utilized to obtain a reliable evapotranspiration (ET) product during 1982-2015. Afterwards four water stress indices were constructed based on the MEP, Penman, Priestley-Taylor and complementary relationship model. The MEP estimated ET showed a close agreement with measurements at eddy covariance sites, with R2 = 0.89 and RMSE ranged from 5 to 12 mm/month. All ecosystems were indicated to suffer from a high risk of water stress, and were ranked by desert (0.67-0.93), grassland (0.60-0.78), settlement (0.49-0.63), farmland (0.48-0.63), and forest ecosystem (0.45-0.58) with four indices. Patterns of water stress at the provincial levels were revealed. Provinces including Xinjiang, Qinghai, Inner Mongolia, and Gansu in the northern regions displayed the highest water stress, and months from December to February were most vulnerable to extreme water stress. Overall, results revealed that the MEP model-based water stress index can well characterize the water stress footprints for all ecosystems and regions in China. This study can support the policy-making for improving water use efficiency and optimizing water resource management to alleviate water stress on large scales.