Effects of climate change on residential electricity consumption in Guangzhou: evidence, forecasting, and warming scenarios

Introduction: Climate change, rising living standards, and increasing household electrification are expected to make residential electricity demand an important source of future energy growth in emerging economies. Focusing on Guangzhou, a humid subtropical megacity in southern China, this study examines how temperature affects daily residential electricity consumption, evaluates short-term forecasting performance, and assesses demand pressure under counterfactual warming scenarios.

Materials and methods: Using daily weather and electricity data from 2019 to 2025, the analysis combines nonlinear econometric models with Ridge Regression, Random Forest, and Gradient Boosting.

Results: The results reveal a pronounced U-shaped temperature response, with electricity consumption reaching its minimum at approximately 19 °C and increasing sharply under hotter conditions. Relative to days with mean temperatures of 20–24 °C, consumption is about 62.4% higher when the mean temperature reaches 30 °C or above. Extreme hot days, hot nights, and prolonged heat also generate additional demand, while no statistically significant difference is found between urban and rural temperature-response patterns. Gradient Boosting performs best, reducing mean absolute error (MAE) and root mean squared error (RMSE) by around 77% compared with the weekly naïve benchmark. Under warming of 1 °C, 2 °C, and 3 °C, total residential electricity consumption increases by approximately 4.9%, 11.1%, and 18.5%, with larger increases during summer and the hottest days.

Conclusions: These findings support heat-sensitive capacity planning, weather-informed forecasting, demand-response programmes, and improved cooling efficiency.