| Abstract:The implementation of greenhouse gas emission accounting on the plant scale can provide data support for formulating industry energy conservation and emission reduction policies. The commonly used emission factor method relies on enterprise energy consumption data for calculation. However, due to the difficulty in collecting enterprise energy consumption data, it is challenging to achieve large-scale greenhouse gas emission accounting for enterprises.Satellite images can cover extensive geographical areas and provide stable observational data, which helps to identify various facilities and activities within the plant area, contributes to analyzing the distribution of greenhouse gas emissions of enterprises in the industry, and reveals the characteristics of high-emission and low-emission enterprises, thus offering a new approach to realizing greenhouse gas emission accounting on the plant scale.Taking the pulp and paper industry as an example, this study integrated satellite images and machine learning methods to construct a greenhouse gas emission accounting model on the plant scale. It then conducted greenhouse gas emission accounting on the plant scale for 112 key emitting paper-making enterprises across the country. The results indicate that there are significant differences in greenhouse gas emissions among key emitting units, with the minimum emission being 44,500 tons of equivalent CO? and the maximum emission reaching 4,260,100 tons of equivalent CO?. |