Spatio-Temporal Dynamics of Carbon Storage in Rapidly Urbanizing Shenzhen, China: Insights and Predictions

Rapid urbanization in developing countries leads to significant land-use and land-cover change (LULCC), which contributes to increased carbon dioxide (CO₂) emissions and the degradation of carbon storage. Studying spatio-temporal changes in carbon storage is crucial for guiding sustainable urban development toward carbon neutrality. This study integrates machine-learning random forest algorithm, CA–Markov, and InVEST models to predict carbon storage distribution in Shenzhen, China, under various scenarios. The findings indicate that, over the past two decades, Shenzhen has experienced significant land-use changes. The transformation from high- to low-carbon-density land uses, particularly the conversion of forestland to construction land, is the primary cause of carbon storage loss. Forestland is mainly influenced by natural factors, such as digital elevation model (DEM) and precipitation, while other land-use and land-cover (LULC) types are predominantly affected by socio-economic and demographic factors. By 2030, carbon storage is projected to vary significantly across different development scenarios, with the greatest decline expected under the natural development scenario (NDS) and the least under the ecological priority scenario (EPS). The RF-CA–Markov model outperforms the traditional CA–Markov model in accurately simulating land use, particularly for small and scattered land-use types. Our conclusions can inform future low-carbon city development and land-use optimization.