Impacts of aquaculture on the area and soil carbon stocks of mangrove: A machine learning study in China

Abstract

Mangrove is an important carbon sink, as it can achieve climate regulation by sequestering carbon dioxide in the atmosphere. However, 50 % of mangrove species are threatened with extinction in China, and the carbon stocks in vegetation has also dropped by 53.1 %. Here, we couple remote sensing data with Random Forests, Support Vector Machines, and XGBoost to analyse mangroves in mainland China from 1986 to 2019. We find that aquaculture has crucial impacts on mangroves and prediction error. Future predictions indicate that the changes of mangroves in different cities range from -5.09E+06 m² to 2.30E+06 m², and soil carbon(C) stocks is "-1.90E+05 Mg ~ 8.57E+04 Mg". To protect mangroves, exploring the balance between aquaculture and mangroves and paying attention to the sustainable transformation of aquaculture are urgently required. In this way, mangroves can fully play the role of carbon sequestration and contribute to China's dual carbon goals.

Keywords: Aquaculture development; Carbon stocks; Machine learning model; Mangrove forests; Sustainable development.