Precision Modeling of Forest Eco-hydrological Process via LiDAR and AI Integration: Advances and Application Framework

Forest ecosystems play a fundamental role in sustaining global water cycles and ecological security, and accurate modeling of their eco-hydrological processes is essential for forest management, watershed regulation, and eco-engineering design. To address the limitations of conventional methods regarding high-precision quantification of the spatial heterogeneity of forest structures and its nonlinear coupling with hydrological processes, this study reviews the advances and engineering application potentials of light detection and ranging (LiDAR) and artificial intelligence (AI) fusion technologies in forest eco-hydrological modeling. The study first analyzes the limitation of homogeneity assumption in conventional models and the challenge of insufficient spatiotemporal data coverage. It highlights the technological breakthroughs of terrestrial and spaceborne LiDAR in constructing high-fidelity forest three-dimensional (3D) structures, as well as the critical role of AI in leaf-wood separation from point clouds, complex hydrological process simulation, and physics-informed machine learning (PIML). Using the National Park of Hainan Tropical Rainforest as a representative case, the study evaluates the application performance of LiDAR + AI fusion in three key areas: estimation of rainfall interception by forest canopies based on fine-scale 3D structures, quantitative analysis of the effects of forest age and structure on evapotranspiration and runoff, and refined accounting of forest water-related ecosystem services. The results demonstrate that LiDAR + AI fusion significantly improves the spatial resolution and predictive accuracy of models, effectively revealing the micro-scale regulation mechanisms of forest vertical structures on rainfall redistribution and evapotranspiration processes. Finally, a “data‒model‒platform‒feedback” fusion modeling framework for forest eco-hydrological processes is proposed. This framework explores the implementation of deep integration between multi-source data and mechanistic models, and identifies current challenges in data standardization, fusion complexity, and computational cost. This study further provides a technical pathway toward high-precision, interpretable forest eco-hydrological modeling with potentials for engineering application, offering scientific support for forest water-related ecosystem service assessment and ecological engineering management.