River–Coast Connectivity Controls Ecosystem Services and Blue Carbon of Coastal Nature-Based Solutions: An Integrated Study Coupling Emergy–Carbon Footprint Accounting and Neural Network Modeling

This study develops an integrated framework combining emergy analysis, carbon footprint accounting, and long short-term memory neural network modeling to investigate the effects of nature-based solutions on coastal ecosystem services and blue carbon functions from the perspective of river–coast connectivity. Three transects along a connectivity gradient were established in the Yellow River Delta, a typical large river delta in temperate China, covering riparian zones, estuarine transition areas, intertidal wetlands, and seagrass beds, with multi-source data collected over three consecutive hydrological years. Emergy–carbon coupling analysis based on this case study indicates that the high-connectivity transect shows a higher emergy yield ratio and net carbon sink compared to the low-connectivity transect, with salt marshes being most sensitive to connectivity change. Threshold analysis, specific to this delta, identifies a three-phase response pattern of carbon burial rate with increasing sediment connectivity, and reveals that wave attenuation efficiency declines notably when hydrological connectivity falls below approximately 0.5, although this value may vary across different coastal settings. A higher sea level rise rate raises the critical connectivity level required to maintain carbon sink function. The long short-term memory neural network trained on observational data achieves better prediction accuracy for blue carbon accumulation rates than traditional statistical methods, and SHAP value analysis suggests the possible existence of synergistic effects among connectivity dimensions. Based on these findings, three optimization strategies including tiered restoration, a dynamic pathway, and spatial configuration are proposed as case-specific recommendations for the Yellow River Delta. Framework-based simulations indicate the potential for connectivity-informed strategy adjustments to improve restoration efficiency under local conditions. This study concludes that river–coast connectivity represents an important lever regulating the ecological benefits of nature-based solutions, but emphasizes that all quantitative thresholds and benefit magnitudes reported here are case-specific estimates that require recalibration when applied to other coastal systems.