Forecasting carbon emissions in Chinese coastal cities based on a gated recurrent unit model

Scientific forecasting of carbon emission trends is imperative to understand carbon emission levels and achieve dual carbon goals. Carbon emission prediction models based on backpropagation neural networks and multiple linear regression methods have yielded relatively favourable results. Recently, the scope of deep learning has broadened, and its incredible plasticity and effectiveness compared to traditional methods have been verified. This is particularly important in China, the world’s most rapidly developing economy, and a spatial study on carbon emission trends and predictions in coastal cities in China has yet to be extensively investigated. Herein, a carbon emission forecasting model was built based on the gated recurrent unit model to empirically evaluate the data of 16 coastal cities in China from 1997 to 2021. Carbon emissions in Chinese coastal cities were forecasted under eight scenarios from 2022 to 2031 to investigate the combined effects of the factors driving carbon emissions. Our results predict increasing carbon emissions from 2022–2031, with a significantly slower growth compared with the change in carbon emissions from 1997–2021. Furthermore, changes in driving factors greatly impact carbon emissions, and the size of the economy most significantly affected carbon reduction. Therefore, the marine economic growth model must be redesigned towards carbon emission reduction to establish a virtuous cycle of low-carbon development.