Objective The aims were to explore the relationship between land use change and carbon storage, in order to provide scientific basis of enhancing the carbon storage of regional ecosystems.

Method Taking Poyang Lake region as an example, based on land use data in 2000, 2005, 2010,2015 and 2020, the InVEST model was used to estimate the ecosystem carbon storage in Poyang Lake region from 2000 to 2020, and Bayesian Network model was established to explore the influence factors of ecosystem carbon storage and its spatial variation characteristics.

Result The results showed that land use change in Poyang Lake region was mainly manifested as the decreases of cultivated land, forest land, grassland and wetland area, and the increases of construction land and water area from 2000 to 2020. Transfer of land use mainly showed that the cultivated land area decreased by 687.26 km², and the cultivated land area increased by 1333.83 km². In the past 20 years, the carbon storage and average carbon density in Poyang Lake region decreased by 1.38 × 10⁷ t and 2.70 t hm⁻², and the conversion of cultivated land and forest land to construction land leaded to a decrease of carbon storage. The forest land and cultivated land accounted for 57.81% and 33.85%. The spatial pattern of carbon storage showed a significant distribution feature of "high around and low in the middle". It was consistent with regional land use types. The carbon storage negative effect of converting cultivated land and forest land to construction land were strong, with contribution rates of -8.83 and -6.87, respectively. Soil type, night light index, NDVI, elevation and ecological red line were the main factors that easily led to the decrease of carbon storage.

Conclusion Land use change is the main reason of ecosystem carbon storage reduction. Strictly implementing urban development boundaries, protecting ecological red lines, improving NDVI levels, and reducing nighttime light indices can enhance the carbon sequestration capacity in the Poyang Lake region.