Development of data-driven estimation models of village carbon emissions by built form factors: The study in Huaihe River Basin, China

Abstract

Decarbonization of built environments is important for climate change mitigation, while most studies have been contextualized in urban areas. In China, village population accounts for about one-third of national population, demonstrating large potential of decarbonization. Focusing on the relationship between village carbon emissions and built form, this study aims to develop data-driven estimation models of village electricity carbon emissions and explore the contributions of built form factors. The models were developed upon linear regression, random forest, eXtreme gradient boosting, artificial neural networks, and deep neural network algorithms, based on the built form factors of 120 villages in Huaiyuan county, in Anhui Province, China. The results verified that the deep neural network had the best estimation capacity for village electricity carbon emissions. This model was further adopted to estimate village electricity carbon emissions in Huaihe River Basin, China. The results showed that the total annual village electricity carbon emissions in Huaihe River Basin were 633,753 ktCO₂, and the average annual village electricity carbon emissions were 2224 ktCO₂. Moreover, Shandong had the largest proportion of villages belonging to the primary carbon reduction villages with the highest carbon reduction needs. Land area was one of the key factors affecting village electricity carbon emissions. Overall, this study helps clarify village carbon emission in Huaihe River Basin and enables the formulation of village planning and design strategies for decarbonization.