Spatiotemporal characteristics and dynamic prediction of agricultural carbon compensation potential in the middle and lower reaches of the Yellow River Basin.

Abstract

Climate change, driven by carbon emissions, has emerged as a pressing global ecological and environmental challenge. Here, we leverage the panel data of five provinces and above prefecture-level cities in the middle and lower reaches of the Yellow River Basin to estimate the agricultural carbon emissions (CEs), carbon sinks (CSs), carbon compensation rate (CCR), and carbon compensation potential (CCP) from 2001 to 2022 and investigate the spatiotemporal evolution characteristics for this region. We propose an improved GLM-stacking ensemble learning method for CE prediction with limited sample data. The findings indicate the following: (i) From 2001 to 2022, the overall CEs show a trend of "development - decline - stabilization" and reach a peak of 172.54 Mt in 2005. CCR first exceeded the "CCR = 1" in 2008, which also indicates that reducing CEs and increasing CSs are the paths to achieving agricultural carbon neutrality. (ii) Although each province has achieved "net-zero emissions," the CCP of most urban agglomerations is about 0.5 and shows a certain agglomeration trend, indicating significant room for further carbon offset. (iii) The novel GLM-stacking model has higher prediction accuracy when compared to a single model. These findings provide scientific and technological support to realize the provincial dual carbon goals in China.