Random forest model-assisted evaluation of spatiotemporal differentiation of China's agricultural net carbon sink and evolution of influencing factors
Given the current climate goals of carbon peaking and carbon neutrality,it is crucial to investigate the spatiotemporal patterns and drivers of agricultural net carbon sink in China.This study aims at promoting high-quality agricultural development and achieving the"dual carbon"goal.Based on the total amount and intensity of China's provincial net carbon sink from 2002 to 2021,a random forest model was used to identify the primary drivers of the agricultural net carbon sink and its nonlinear response relationship.The results show that:(1)China's agricultural sector is experiencing a slow but consistent carbon surplus growth,demonstrating an attained carbon neutrality and transitioning towards an elevated carbon surplus state in most provinces;(2)the total net carbon sink tends to decline from the east to the west,with an increase in high-value agglomeration and a decrease in low-value agglomeration;(3)the spatial agglomeration and non-equilibrium of net carbon sink intensity tend to be obvious,with a higher intensity in the northeast,central,and southwest;and(4)the key factors driving agricultural net carbon sink include irrigation conditions,mechanized straw return,mechanized no-tillage sowing,grain yield,etc.,exhibiting nonlinear effects on the net carbon sink.Specifically,there is a U-shaped relationship between education level and agricultural net carbon sink,and mechanization level has a significant inhibitory effect,while other factors demonstrate a positive fluctuation effect.
agricultural net carbon sequestrationspatial-temporal variationinfluencing factorsrandom forest modelChina
万方数据
维普
