City-level carbon emissions accounting and differentiation integrated nighttime light and city attributes

Abstract

The city plays a crucial role in CO2 emissions reduction, and exploring the accounting methods of city-level CO2 emissions is important for formulating effective emission reduction policies. Although previous studies have verified a linear relationship between nighttime light data (NTL) and CO2 emissions at the provincial level, the consistent fitting coefficient adopted by cities ignored the socioeconomic characteristics of the city itself. Therefore, this study proposed a new method of using NTL to account city-level CO2 by dividing all cities into seven types which were mainly characterized by resources-based, energy-produced, light industry-based, heavy industry-based, high-end service-based, general service-based, and others using K-means, and constructing multivariate regression and hierarchical regression model to correct fitting coefficients of NTL of each city. The results showed that the estimation in this study improved the accuracy of city-level direct CO2 emissions compared with fitting results obtained directly by NTL. The high-emission cities concentrated in four urban agglomerations accounted for 32.68% of total emissions. Resources-based cities, energy-produced cities, and heavy industry-based cities showed higher CO2 emissions intensity (CEI) and CO2 emissions per capita (CEC) among seven-type cities. The contribution of population density to the cities with low CEI and CEC was obvious, but the trends of urbanization and GDP per capita were opposite. This study improved and enriched the methods of estimating city-level CO2 emissions with NTL, and provided support for different-type cities to formulate targeted CO2 emission reduction policies.