Adaptive wetland management: Insights from nonlinear dynamics between vegetation and hydrological variables

Wetlands provide essential ecosystem services but are increasingly impacted to water scarcity. Understanding the synergistic impacts of multiple hydrological factors on vegetation is crucial for effective wetland management and restoration. This study investigates the nonlinear and asymmetric responses of wetland vegetation, represented by the Normalized Difference Vegetation Index (NDVI), to key hydrological factors: precipitation, inflow, and water depth. Using the NaoLiRiver Wetlands as a case study, we developed a comprehensive framework integrating spatiotemporal image fusion, Generalized Additive Modeling (GAM), and Random Forest analysis to analyze 20 years of remote sensing and hydrological data. Results showed that hydrological factors collectively explained up to 79.4 % of NDVI variation during dry periods, compared to 59.4 % during wet periods. Critical thresholds were identified at 100 mm for precipitation, 75 m³/s for inflow, and 0.6 m for water depth, beyond which NDVI responses began to decline. Time effects, including a 1-month lag and accumulation, were also significant. The Random Forest model further validated the dominance and synergy of hydrological factors. This research not only enhances our understanding of wetland ecohydrology but also offers actionable recommendations for adaptive wetland management in the face of growing climatic challenges.