Study on the Projection of Net Carbon Emissions in Anhui Province Based on Ensemble Forecasting

Abstract

To achieve precise planning for Anhui Province’s carbon neutrality pathway, this paper constructs a combined forecasting model using Grey Model (GM (1,1)), exponential smoothing, and ARIMA time series, and aggregates the forecasting results using the inverse of the Mean Absolute Percentage Error (MAPE). Empirical results indicate that the combined forecasting model reduced the average MAPE to 2.49% for the 2005-2022 forecast period, representing a decrease of 4.05%, 1.56%, and 0.58% compared to the average MAPEs of the grey system, exponential smoothing, and ARIMA time series models, respectively, outperforming all three individual forecasting models. The ensemble forecast results for 2023-2027 indicate that Anhui Province’s net carbon emissions will increase from 532.99 million tons to 554.99 million tons, with an average annual growth rate of 1.41%. Although the growth rate shows a declining trend, the absolute increase still reaches 22 million tons, suggesting that the current development trajectory is closer to a high-carbon model and highlighting the challenges of achieving the carbon peak target. This paper provides a decision-making benchmark with controllable error margins for regional carbon neutrality pathways, confirming the paradigmatic value of ensemble forecasting in mitigating the extreme risks associated with single-model predictions. Additionally, it proposes a dynamic weighting update mechanism, coordinated emission reductions across the industrial, energy, and building sectors, and a carbon offset system for northern and southern Anhui, thereby offering a data-driven decision-making framework for Anhui Province’s carbon peaking by 2030.