Climate Effects and Driving Mechanisms of Primary Energy Consumption in Sichuan Province: An Integrated Framework Combining Time Series Analysis with Causal Inference

Abstract

This study quantifies the impact of primary energy consumption on climate change in Sichuan Province by analyzing 2012–2022 datasets through integrated time series analysis and causal inference methods. Key methodologies include anomaly detection via boxplot analysis, data correction using Lagrange interpolation and moving averages, and modeling via polynomial fitting, multiple linear regression, Pearson correlation, and Granger causality tests. Results demonstrate: (1) Significant positive correlations between fossil fuel consumption (coal, petroleum, natural gas) and CO₂ emissions (correlation coefficient: 0.9654), with coal exhibiting the strongest driving effect (regression coefficient: 0.9055); (2) CO₂ emissions act as a Granger cause of temperature variation (lag order: 3, p-value <0.05), while no causal link with precipitation is observed; (3) Weak nonlinear climate interactions, evidenced by a marginal positive correlation between CO₂ and temperature (Pearson: 0.2106) and a weak negative correlation with precipitation (-0.3616). As a critical energy hub and ecological barrier in western China, Sichuan’s fossil fuel dependency underscores the urgency of transitioning to renewable energy and implementing climate-resilient water management. The proposed "time series-causal inference" framework offers scalable solutions for regional energy-climate governance, aligning with global sustainability goals and providing actionable insights for policymakers to prioritize coal reduction and enhance adaptive strategies in ecologically vulnerable regions.