Zou et al. (2026) Spatio-temporal characteristics and driving factor identification of precipitation use efficiency: Based on machine learning and SHAP analysis

Identification

• Journal: Agricultural Water Management
• Year: 2026
• Date: 2026-09-10
• Authors: Shuai Zou, Fanxiang Meng, Ennan Zheng, Tianxiao Li, Gang Li, M. C. Li
• DOI: 10.1016/j.agwat.2026.110763

Research Groups

• School of Hydraulic and Electric Power, Heilongjiang University, Harbin, Heilongjiang, 150080, China
• School of Public Administration, Northeast Agricultural University, Harbin, Heilongjiang, 150030, China
• Postdoctoral Research Workstation of Harbin Surveying College Surveying Engineering Company, Harbin, Heilongjiang, 150050, China
• College of Hydraulic Science and Engineering, Northeast Agricultural University, Harbin, Heilongjiang, 150030, China

Short Summary

This study investigates the spatiotemporal characteristics and driving factors of precipitation use efficiency (PUE) in the cold regions of Northeast China using machine learning and SHAP analysis. The results show that PUE exhibits a gradient pattern with higher values in the northwest and lower values in the southeast, and is influenced by multiple individual factors and their interactions.

Objective

• To reveal the spatial distribution patterns, temporal evolution, and future trends of PUE in the cold regions of Northeast China from 2001 to 2024.
• To quantify the individual contributions of climatic factors (e.g., PRE, TEMP, PET, SR, WS), vegetation characteristics (e.g., FVC), and geographic attributes (e.g., ELEV, SLOPE, ASPECT) to spatiotemporal variations in PUE.

Study Configuration

• Spatial Scale: The study area covers the cold regions of Northeast China, with a total area of approximately 1.265 million km².
• Temporal Scale: The study period is from 2001 to 2024.

Methodology and Data

• Models used: LGB (Lightweight Gradient Booster) and RF (Random Forest) models were selected for PUE driver analysis in the temporal and spatial dimensions, respectively.
• Data sources: Multi-source remote sensing data from 2001 to 2024, including precipitation, temperature, wind speed, solar radiation, potential evapotranspiration, net primary production, and vegetation cover.

Main Results

• The spatial distribution of PUE exhibits a gradient pattern with higher values in the northwest and lower values in the southeast.
• The temporal evolution of PUE shows a phased change from 2001 to 2024, with a peak value of 0.96 gC⋅m⁻²⋅mm⁻¹ recorded in 2014.
• The Hurst index analysis indicates that approximately 6% of the regions are likely to maintain their current PUE trends in the future, while approximately 94% suggest that their future trends may reverse past patterns.

Contributions

• This study provides a systematic analysis of the spatiotemporal dynamics of PUE and its driving factors in the cold regions of Northeast China.
• The results reveal the multiscale factors driving PUE and identify key regulatory thresholds, providing quantitative evidence for regional “carbon-water” integrated management, precision-based vegetation restoration, and climate change adaptation strategies.

Funding

• This study was funded by the National Natural Science Foundation of China (Grant No. 52179015) and the Heilongjiang Provincial Natural Science Foundation (Grant No. YQ2021C004).

Citation

@article{Zou2026Spatiotemporal,
  author = {Zou, Shuai and Meng, Fanxiang and Zheng, Ennan and Li, Tianxiao and Li, Gang and Li, M. C.},
  title = {Spatio-temporal characteristics and driving factor identification of precipitation use efficiency: Based on machine learning and SHAP analysis},
  journal = {Agricultural Water Management},
  year = {2026},
  doi = {10.1016/j.agwat.2026.110763},
  url = {https://doi.org/10.1016/j.agwat.2026.110763}
}