Impact of urban expansion on land surface temperature and carbon emissions using machine learning algorithms in Wuhan, China

Abstract

The impact of the rapid expansion of urban land on the urban thermal environment and carbon chain has attracted widespread attention. This paper uses artificial neural network-cellular automata (ANN-CA) and long short-term memory model of the improved whale optimization algorithm (IWOA-LSTM) models to predict the changes of LULC and LST, and explores the correlation between LST and carbon emissions in Wuhan.The results show that urban land will occupy >70.05% of the central urban area, while green land and water areas will continue to decrease to varying degrees in 2030 and 2040. The area of the high temperature area (LST > 30 °C) is expanding in the urban land, while the green land and the low temperature area of the water body are gradually shrinking. The area of high temperature is expanding, and the area with LST > 30 °C accounts for 67.84% in summer, and the area with LST at 10 °C ∼ 15 °C accounts for 96.32% in winter. The fitting results of correlation regression show that there is a significant correlation between carbon emissions and LST. The R2 of linear fitting between LST and carbon emissions in summer and winter of 2000 are 0.6227 and 0.6143, respectively. The R2 of linear fitting in summer and winter is higher, both of which are >0.85 in 2010 and 2020. This research may provide new clues for future urban development and thermal environment governance and carbon emission mitigation.