Terrestrial ecosystem is one of the most important carbon pools in the world and it plays an important role in mitigating global warming.The estimation of terrestrial ecosystem carbon flux is a key part to evaluate the change of terrestrial ecosystem carbon sink.This study used Convolutional Neural Network(CNN)to estimate the Net Ecosystem Productivity(NEP)of global terrestrial ecosystems based on flux observation data,combined with meteorological data,remote sensing data and soil data.The resutls showed that CNN can effectively simulate the spatial distribution and total amount of global NEP.It was estimated that the global terrestrial ecosystem NEP is around(22±2)pgC/a,with approximately two-thirds of the spatial area being carbon sink regions,predominantly driven by carbon sequestration.The total NEP of China is approximately 1.17 pgC/a,accounting for around 5.3%of the global NEP.The change trend of the global average NEP was not obvious from 2000 to 2018,indicating that the global carbon sink is in dynamic equilibrium.
维普
万方数据