Predicting CO2 emissions through partitioned land use change simulations considering urban hierarchy

Abstract

An understanding of the relationship between CO2 emissions and land use can provide a deeper understanding of how urbanization affects climate change, provide insights into sustainable urban development, and reveal CO2 emission dynamics in the context of urbanization. However, the accuracy of these predictions is limited due to the poor understanding of the influence of differences in urban hierarchy on land use transition rules. This study proposes an integrated CO2 emission prediction method that utilizes a newly developed future land use simulation (partitioned-FLUS) model and a Long Short-Term Memory (LSTM) model. Notably, the partitioned-FLUS model considers regional heterogeneities based on urban hierarchies. Based on data from the Yangtze River Delta, zonal partitioning was found to improve Kappa coefficients by at least 13% compared to conventional FLUS models. The results indicated that a significant urbanization belt was likely to form both along and south of the Yangtze River over the next 30 years; resulting in areas of concentrated CO2 emissions. The relationship between land use and CO2 emissions varied greatly between cities of different urban hierarchies. This study suggests that construction land and farmland in Level 1 and 2 cities should be reallocated to Level 3 cities, thereby reducing carbon emissions.

Keywords

• Land-use change
• carbon emissions prediction
• partitioned-FLUS model
• urban hierarchy
• urbanization