Li et al. (2026) Evapotranspiration dynamics and climatic-land-use controls in the Hanjiang River Basin, 2000-2018

Identification

• Journal: Journal of Hydrology Regional Studies
• Year: 2026
• Date: 2026-01-15
• Authors: Rui Li, Zhijie Zhang, Wanchang Zhang, Ping Rao
• DOI: 10.1016/j.ejrh.2026.103125

Research Groups

• Key Laboratory of Digital Earth Science, Aerospace Information Research Institute, Chinese Academy of Sciences, Beijing, China
• International Research Center of Big Data for Sustainable Development Goals, Beijing, China
• University of Chinese Academy of Sciences, Beijing, China
• School of Geospatial Artificial Intelligence, East China Normal University, Shanghai, China
• Guizhou University of Engineering Science, Bijie, China

Short Summary

This study investigated the spatiotemporal dynamics and climatic-land-use controls of evapotranspiration (ET) and its components in the Hanjiang River Basin (2000-2018) using a physically-based model and advanced machine learning. It found significant spatial and seasonal shifts in ET partitioning, with vegetation-related components increasing, and identified temperature, solar radiation, and LAI as primary drivers, emphasizing the dominant role of mid-depth root water uptake (5–30 cm) in canopy transpiration.

Objective

• To investigate the spatiotemporal variations of total evapotranspiration (ET) and its components (dry canopy transpiration, wet canopy evaporation, saturated soil evaporation, moist soil evaporation) in the Hanjiang River Basin from 2000 to 2018.
• To identify the dominant meteorological, soil, land use, and vegetation drivers of ET and understand how their influences differ across ET components.
• To explore how interactions among these drivers shape ET dynamics, including nonlinear and hierarchical regulatory mechanisms.

Study Configuration

• Spatial Scale: Hanjiang River Basin, China (159,000 km²), with a unified spatial resolution of 1 km.
• Temporal Scale: 2000-2018 (19 years), with daily outputs.

Methodology and Data

• Models used: ESSI-3 (physically-based distributed hydrological model), Random Forest (RF), SHapley Additive exPlanations (SHAP), Structural Equation Modeling (SEM).
• Data sources:
  • Meteorological: China Meteorological Forcing Dataset (CMFD) (wind speed, specific humidity, precipitation, air temperature, solar radiation, surface pressure) at 0.1º daily resolution.
  • Soil property: SoilGrids250m (sand, silt, clay content, bulk density, hydraulic parameters) at 250 m resolution.
  • Vegetation parameter: GLOBMAP-based Leaf Area Index (LAI) at 8 km (16-day/8-day), MODIS MCD12Q1 V061 (LULC) at 1 km, PCR-GLOBWB2 (Tree cover) at 400 m.
  • Topography: SRTMDEM (DEM) at 90 m.
  • Validation/Calibration: China Hydrological Yearbook (streamflow at Hanzhong and Baihe stations, monthly, 2007–2015), GLEAM dataset (monthly actual evaporation, interception loss, transpiration, 2000–2018).

Main Results

• Total ET exhibited a slight increasing trend (0.59 mm/year by linear regression, 0.54 mm/year by Theil-Sen) from 2000 to 2018.
• Vegetation-related ET components (ETCanopy, ECanopy) showed upward trends (0.58 mm/year and 0.41 mm/year, respectively), while soil-related components (ETSoil, ESoil) declined (−0.20 mm/year and −0.25 mm/year, respectively).
• ET displayed strong spatial heterogeneity (increasing from northwest to southeast) and pronounced seasonal shifts: ETCanopy dominated in spring, summer, and autumn (94.9%, 96.4%, 87.7% of area, respectively), while ESoil prevailed in winter.
• The second soil layer (5–30 cm) contributed the largest fraction (72.1%) of total canopy transpiration (ET_Canopy), highlighting the crucial role of mid-depth root zones.
• Temperature, solar radiation, and Leaf Area Index (LAI) were identified as the primary drivers of ET dynamics, exhibiting nonlinear and interactive effects.
• Precipitation and soil water showed limited direct influence on the daily scale but contributed through lagged effects.
• RF models achieved high predictive accuracy for ET_Canopy (R² = 0.92) and total ET (R² = 0.79). SEM models demonstrated excellent fit (CFI > 0.98, TLI > 0.97, RMSEA < 0.04).

Contributions

• Developed a replicable framework integrating a physically-based distributed hydrological model (ESSI-3) with advanced machine learning (Random Forest, SHAP) and causal inference (Structural Equation Modeling) to comprehensively analyze ET components and their drivers at the basin scale.
• Provided a detailed spatiotemporal analysis of total ET and its four components, revealing distinct seasonal shifts in dominance (ETCanopy in growing season, ESoil in winter) and contrasting trends between vegetation-related (increasing) and soil-related (decreasing) components.
• Quantified the dominant role of mid-depth root water uptake (5–30 cm) in sustaining canopy transpiration, showing it contributes 72.1% of total ET_Canopy.
• Disentangled nonlinear, interactive, and hierarchical controls of meteorological, soil, land use, and vegetation factors on ET components, offering mechanistic insights into ecohydrological processes.
• Informed adaptive water management strategies by emphasizing component-specific regulation of ET in monsoon-sensitive basins.

Funding

• National Key R & D Program of China [Grant No. 2023YFC3209102]
• National Natural Science Foundation of China [Grant No. 42471344]
• Financial support from Utah State University, USA, for Dr. Zhijie Zhang.

Citation

@article{Li2026Evapotranspiration,
  author = {Li, Rui and Zhang, Zhijie and Zhang, Wanchang and Rao, Ping},
  title = {Evapotranspiration dynamics and climatic-land-use controls in the Hanjiang River Basin, 2000-2018},
  journal = {Journal of Hydrology Regional Studies},
  year = {2026},
  doi = {10.1016/j.ejrh.2026.103125},
  url = {https://doi.org/10.1016/j.ejrh.2026.103125}
}