Evapotranspiration and its components partitioning based on an improved hydrological model: historical attributions and future projections

Estimation and attribution of evapotranspiration (ET) and its components under changing environment is still a challenge but is essential for understanding the mechanisms of water and energy transfer subsequently regional water resources management. In this study, an improved hydrological model is developed to estimate evapotranspiration and its components i.e., evaporation (E) and transpiration (T) by integrated the advantages of hydrological modeling constrained by water balance and the water-carbon close relationships. Results show that the improved hydrological model could captures ET and its components well in the study region. During the past years, annual ET and E increase obviously about 2.40mm/a and 1.42mm/a, particularly in spring and summer accounting for 90%. T shows less increasement and mainly increases in spring while it decreases in summer. Precipitation is the dominant factor and contributes 74.1% and 90.0% increases of annual ET and E, while the attribution of T changes is more complex by coupling of the positive effects of precipitation, rising temperature and interactive influences, the negative effects of solar diming and elevated CO₂. In the future, ET and its components tend to increase under most of the Shared Socioeconomic Pathways (SSP) scenarios except for T decreases under the very high emissions scenario (SSP5-8.5) based on the projections. From seasonal perspective, the changes of ET and the components are mainly in spring and summer accounting for 75%, while more slightly changes are found in autumn and winter. This study highlights the effectiveness of estimating ET and its components by improving hydrological model within water-carbon coupling relationships, and more complex mechanisms of transpiration changes than evapotranspiration and evaporation changes under the interactive effects of climate variability and vegetation dynamics. Besides, decision makers should pay attention to the more increases in the undesirable E than desirable T.