Spatiotemporal dynamics and environmental drivers of gross primary productivity in mangrove ecosystems of Hainan Island

Abstract

Mangrove ecosystems are crucial for blue carbon sequestration (carbon captured and stored by oceans and coastal ecosystems), with gross primary productivity (GPP) acting as a key indicator of their carbon capture potential. However, estimating regional-scale GPP for mangrove ecosystems and their response to environmental factors remains challenging. In this study, we focused on the mangrove ecosystems of Hainan Island, utilizing high-resolution Sentinel-2 imagery to derive the fraction of absorbed photosynthetically active radiation (fAPAR). We combined machine learning techniques with the mangrove vegetation photosynthesis light use efficiency (MVP-LUE) model to retrospectively estimate historical fAPAR and project GPP over two decades. Spatiotemporal distribution patterns were analyzed, and environmental driving factors were quantified using partial differential attribution. Our findings can be summarized in three major points. First, the optimized ensemble model demonstrated high accuracy in estimating mangrove ecosystem fAPAR (R² = 0.77, RMSE = 0.08). Second, the mangrove ecosystems of Hainan Island exhibited an overall increasing trend in annual GPP at a rate of 3.46 gC m⁻² y⁻² over the study period. Peak GPP occurred during the transition from dry to rainy season (213.73 gC m⁻² mo⁻¹), with minimum GPP in the mid-dry season (76.07 gC m⁻² mo⁻¹). Spatially, GPP was higher in the eastern and northern regions. Finally, changes in the GPP of mangrove ecosystems were primarily driven by fAPAR. The quantification of environmental effects on mangrove GPP enhances our understanding of carbon cycling dynamics within these highly productive ecosystems, informing protection and restoration efforts.