Spatiotemporal Dynamics and Nonlinear Associations of Township-Level Environmental Performance During the Rural Clean Heating Transition

Rural clean heating in China is moving away from large-scale infrastructure retrofits towards long-term operation and optimization. How environmental performance changes over time and among different townships remains insufficiently understood. This study examines 36 townships in Qingdao using multisource data from 2017 to 2024. A composite environmental performance index (CEPI) was constructed from heating-season concentrations of PM2.5, PM10, NO2, and SO2, together with building-sector CO2 emission intensity. CRITIC–TOPSIS, spatial autocorrelation analysis, and an XGBoost–SHAP framework with township-grouped cross-validation were applied to characterize spatiotemporal changes and nonlinear associations with explanatory factors. The results show that: (1) CEPI rose from 0.385 in 2017 to 0.816 in 2024, and all 36 townships improved. SO2 concentrations fell the most, and the reductions in CO2 emission intensity were smaller. Sensitivity analyses supported the robustness of the evaluation. (2) Environmental performance exhibited persistent positive spatial autocorrelation, with relatively stable high-value clusters in northern areas and low-value clusters in central areas. (3) XGBoost achieved a mean R2 of 0.793 in five-fold grouped cross-validation and outperformed the linear benchmark. SHAP analysis revealed nonlinear statistical associations between CEPI and climatic conditions, population concentration, and residential energy use, rather than causal effects attributable to the clean-heating policy. Heating-season surface solar radiation accounted for 42.32% of total SHAP importance, while climatic variables jointly accounted for 67.5%, reflecting their predictive importance rather than independent causal effects. Strong conditional interactions were observed between precipitation and solar radiation and between population density and residential electricity-use density. These findings indicate that long-term clean-heating management should move beyond city-level averages and incorporate township-scale spatial clustering, meteorological conditions, and energy-use characteristics into monitoring and regional diagnosis to support more targeted environmental management and the environmental sustainability of rural clean heating.

Authors

Institutions

Publication Details

Journal
Sustainability
Published
2026-11-12
DOI
https://doi.org/10.3390/su182211547
Primary Topic
Energy and Environment Impacts
Type
article
Field-Weighted Citation Impact
0.00

Funders

Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

Spatiotemporal Dynamics and Nonlinear Associations of Township-Level Environmental Performance During the Rural Clean Heating Transition

Shengxi Zhang, Guanghui Wang, Mengmeng Li, Xueqing Shi et al.
Sustainability
Energy and Environment Impacts
article

Spatiotemporal Dynamics and Nonlinear Associations of Township-Level Environmental Performance During the Rural Clean Heating Transition

Shengxi Zhang, Guanghui Wang, Mengmeng Li, Xueqing Shi, Huiqian Wang, Weiliang Wang, Ziying Ma
article en

Abstract

Rural clean heating in China is moving away from large-scale infrastructure retrofits towards long-term operation and optimization. How environmental performance changes over time and among different townships remains insufficiently understood. This study examines 36 townships in Qingdao using multisource data from 2017 to 2024. A composite environmental performance index (CEPI) was constructed from heating-season concentrations of PM2.5, PM10, NO2, and SO2, together with building-sector CO2 emission intensity. CRITIC–TOPSIS, spatial autocorrelation analysis, and an XGBoost–SHAP framework with township-grouped cross-validation were applied to characterize spatiotemporal changes and nonlinear associations with explanatory factors. The results show that: (1) CEPI rose from 0.385 in 2017 to 0.816 in 2024, and all 36 townships improved. SO2 concentrations fell the most, and the reductions in CO2 emission intensity were smaller. Sensitivity analyses supported the robustness of the evaluation. (2) Environmental performance exhibited persistent positive spatial autocorrelation, with relatively stable high-value clusters in northern areas and low-value clusters in central areas. (3) XGBoost achieved a mean R2 of 0.793 in five-fold grouped cross-validation and outperformed the linear benchmark. SHAP analysis revealed nonlinear statistical associations between CEPI and climatic conditions, population concentration, and residential energy use, rather than causal effects attributable to the clean-heating policy. Heating-season surface solar radiation accounted for 42.32% of total SHAP importance, while climatic variables jointly accounted for 67.5%, reflecting their predictive importance rather than independent causal effects. Strong conditional interactions were observed between precipitation and solar radiation and between population density and residential electricity-use density. These findings indicate that long-term clean-heating management should move beyond city-level averages and incorporate township-scale spatial clustering, meteorological conditions, and energy-use characteristics into monitoring and regional diagnosis to support more targeted environmental management and the environmental sustainability of rural clean heating.

SustainabilityVol. 18(22)
Guilin University of Technology (CN), Qingdao University of Technology (CN), Ocean University of China (CN)
Department of Science and Technology of Shandong Province, Qingdao Municipal Science and Technology Bureau
Climate action, Sustainable cities and communities, Affordable and clean energy
Energy and Environment Impacts
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon