抄録
Landscape Evolution Models (LEMs) are increasingly challenged to move beyond morphological reconstruction toward verifiable, multiscale representations of Earth-surface dynamics under strong spatiotemporal heterogeneity, nonlinear feedbacks, and multiprocess coupling. This review synthesizes the theoretical development, process formulations, and computational architectures of LEMs, with emphasis on representative models including SIBERIA, CHILD, Landlab, FastScape, and Badlands, and their treatment of fluvial incision, hillslope transport, gravitational processes, and source-to-sink sediment transfer. We focus on four unresolved issues that constrain predictive capability: the scale dependence of effective parameterizations, the transmission of transient signals across temporal and spatial scales, the storage and remobilization of sediment along source-to-sink pathways, and the identifiability of process mechanisms from incomplete observations. We further assess how multitemporal LiDAR/SfM observations, data assimilation, machine learning and physics-informed learning, GPU/HPC computing, and interoperable computational platforms are reshaping parameter inversion, process emulation, and dynamic prediction, while highlighting persistent limitations associated with model complexity, equifinality, observational incompleteness, and the representation of vegetation, anthropogenic forcing, and nonstationary environmental change. To address these challenges, we propose an integrated framework of “observation constraints–heterogeneity representation–cross-scale coupling–computational infrastructure”, in which multisource observations, process-based modeling, data-driven methods, and scalable computing are jointly organized around conservation principles and uncertainty-aware validation. This framework provides a coherent pathway for advancing LEMs from morphology-based reconstruction toward process-verifiable, cross-scale, and predictive Earth-surface system models, strengthening their capacity to resolve landscape responses to global change and Anthropocene disturbances.
