Dynamic correlation analysis of sectoral electricity consumption and urban carbon concentration using machine...

Area of Science:

• Environmental Science
• Climate Change Research
• Data Science

Background:

• Global climate change, driven by carbon dioxide (CO2) emissions, presents a significant environmental challenge.
• Achieving carbon neutrality requires understanding and reducing emissions across diverse industries, yet progress remains uncertain.

Purpose of the Study:

• To develop a dynamic correlation model linking sectoral electricity consumption to urban CO2 column concentration (XCO2).
• To explore the temporal relationships between industry-specific electricity use and atmospheric CO2 levels using advanced machine learning.

Main Methods:

• Employed machine learning algorithms: random forest, extreme gradient boosting (XGBoost), and stacked regression.
• Integrated time-rolling window techniques to dynamically assess correlations.
• Validated the model using data from 16 cities (2017-2021) for sectoral electricity consumption and urban XCO2.

Main Results:

• The dynamic correlation model achieved a high accuracy (R² up to 0.864) and low error (RMSE 1.350).
• The model significantly outperformed traditional methods in capturing complex relationships.
• Revealed significant temporal fluctuations and sector-specific influences on urban CO2 concentrations.

Conclusions:

• Advanced machine learning models can effectively capture dynamic correlations between sectoral electricity consumption and urban CO2.
• Findings provide crucial insights for developing precise, industry-targeted carbon reduction strategies.
• The study highlights the importance of dynamic analysis for effective climate change mitigation policies.