Ji et al. (2025) Response characteristics of vegetation net primary production to cascade hydropower development and climate change in the dry-hot valleys of the Jinsha River

Identification

• Journal: Journal of Hydrology Regional Studies
• Year: 2025
• Date: 2025-10-28
• Authors: Chenxu Ji, Yong Huang, Jiaxin Liu, Xinyan Wu, Liding Chen
• DOI: 10.1016/j.ejrh.2025.102880

Research Groups

• School of Ecology and Environmental Sciences, Yunnan University, Kunming, China
• Ministry of Education Key Laboratory for Transboundary Ecosecurity of Southwest China, Yunnan University, Kunming, China
• State Key Laboratory of Urban and Regional Ecology, Research Center for Eco-Environmental Sciences, Chinese Academy of Sciences, Beijing, China
• Institute of International Rivers and Eco-Security, Yunnan University, Kunming, China

Short Summary

This study quantifies the ecological impacts of cascade hydropower development on Net Primary Production (NPP) in the dry-hot valleys of the Jinsha River, finding an overall increase of approximately 12% in NPP compared to pre-development levels due to altered local hydrology and climate. Reservoir impoundment significantly enhanced NPP during the dry season by alleviating water stress, while in the rainy season, its influence shifted to intensifying heat-related factors.

Objective

• To quantify the ecological impacts of cascade hydropower development on Net Primary Production (NPP) in the dry-hot valleys of the Jinsha River.
• To systematically assess the spatiotemporal variations in NPP before and after hydropower development.
• To identify the primary factors affecting NPP.
• To isolate the ecological effects of hydropower development from those of global climate change using a non-reservoir scenario simulation framework.

Study Configuration

• Spatial Scale: The entire dry-hot valley of the Jinsha River, located in the middle and lower reaches, flowing across northern Yunnan Province and southern Sichuan Province, Southwestern China.
• Temporal Scale: Data primarily spans from January 2001 to December 2022, with some soil moisture data from 2000 to 2020.

Methodology and Data

• Models used:
  • Pettitt mutation test: To identify the starting time point of vegetation impacts caused by reservoir impoundment.
  • Long Short-Term Memory (LSTM) model: To construct a predictive model for NPP changes under a no-reservoir scenario.
  • Mann–Kendall (MK) trend test: To detect trends in time series data.
  • Extreme Gradient Boosting (XGBoost) algorithm: To analyze the characteristics of the effects of various factors on NPP changes.
  • SHapley Additive exPlanations (SHAP): To interpret the XGBoost model and assess the contribution of each feature to NPP prediction.
• Data sources:
  • NPP data: MODIS satellite remote sensing product MOD17A2H (8-day integrated, 500 m spatial resolution).
  • Fractional vegetation cover (FVC) data: National Tibetan Plateau Data Center (monthly, 250 m spatial resolution).
  • Meteorological data (Precipitation, Temperature, Potential Evapotranspiration): Monthly datasets for China (Peng, 2019, 2020, 2022) (0.0083333° / ~1 km spatial resolution).
  • Soil moisture data: Qinghai–Tibet Plateau Data Center, "Soil Moisture of China from In Situ Data, Version 1.0" (Shangguan et al., 2022) (daily, 1 km spatial resolution, 10 cm measurement depth).
  • Land surface temperature (LST) data: NASA’s MODIS MOD11A2 product (8-day, 1 km spatial resolution).

Main Results

• A statistically significant change-point in NPP was detected around May 2015 by the Pettitt test, coinciding with the primary impoundment period of major reservoirs.
• The no-reservoir scenario simulation revealed that observed monthly average NPP after impoundment was approximately 11.8% higher than predicted (60.3 gC/m² observed vs. 53.9 gC/m² predicted).
• During the dry season, reservoir impoundment led to a 26.3% increase in monthly average NPP (39.9 gC/m² observed vs. 31.6 gC/m² predicted), significantly greater than in the rainy season.
• Interannual NPP trends showed an increase from 2001–2007, a decrease from 2008–2013 (coinciding with construction), and a significant increase from 2014–2022 (after completion and water storage).
• Spatially, NPP in central and southern regions decreased markedly during construction (2008–2013) but exhibited significant recovery and growth (up to 50% in some areas) after impoundment (2014–2022).
• Temperature (TMP) was identified as the single most dominant, non-linear driver of NPP changes (SHAP value: 12.27), followed by potential evapotranspiration (PET, SHAP value: 6.69), land surface temperature (LST), precipitation (PRECIP), and soil moisture content index (SMCI, SHAP value: 4.05).
• Reservoir impoundment reshaped local climate, leading to increased air temperature, precipitation, and soil moisture, coupled with a decrease in land surface temperature.
• In the dry season, impoundment shifted the correlations of water-related factors (SMCI and PRECIP) with NPP from negative to positive, indicating significant alleviation of water limitations.
• In the rainy season, impoundment strengthened the positive correlation of TMP and PET with NPP, while negative correlations of PRECIP and SMCI remained or slightly intensified, suggesting enhanced influence of heat-related factors and potential adverse effects of excessive water.

Contributions

This study fills a notable gap by systematically assessing the ecological impacts of cascade hydropower development on vegetation NPP in the dry-hot valleys of the Jinsha River. It uniquely isolates the effects of reservoir impoundment from global climate change using a novel non-reservoir scenario simulation framework, providing a deeper understanding of the complex interactions between hydropower operations, climate factors, and vegetation dynamics in this fragile ecosystem.

Funding

• Innovative Project of Caiyun Postdoctoral Program of Yunnan Province (C615300504081)
• Technical service/consulting project of China Three Gorges Corporation (202403240)
• Yunnan University (CZ22621211)

Citation

@article{Ji2025Response,
  author = {Ji, Chenxu and Huang, Yong and Liu, Jiaxin and Wu, Xinyan and Chen, Liding},
  title = {Response characteristics of vegetation net primary production to cascade hydropower development and climate change in the dry-hot valleys of the Jinsha River},
  journal = {Journal of Hydrology Regional Studies},
  year = {2025},
  doi = {10.1016/j.ejrh.2025.102880},
  url = {https://doi.org/10.1016/j.ejrh.2025.102880}
}