Structure-process-performance: A systematic framework for deconstructing constructed wetland models

Numerical models of constructed wetlands (CWs) have evolved from black-box statistical models to process-based mechanistic approaches, and recently to hybrid frameworks integrating data-driven methods. While these models now underpin scientific design and intelligent operation, significant challenges remain particularly the trade-off between process completeness and model generalizability. The prevalent trend of stacking technological complexity often overlooks how wetland structures fundamentally constrain model frameworks, parameters, and resulting predictions. Furthermore, critical scrutiny is lacking regarding physical explanations for inter-model discrepancies, compensatory effects, and assumption conflicts in sub-model couplings, especially given emerging demands for carbon neutrality and intelligent regulation. This review presents the first systematic, formalized 'structure-process-performance' framework, explicitly translating an implicit empirical principle into a guiding methodology. This study reviews the evolution of CWs models, offering references for CWs model construction. It conducts a thorough analysis of the differences resulting from different CWs structures, the factors influencing solute transport pathways, the differences and interrelationships among various components within CWs, and the influence of system components and environmental parameters. Additionally, it proposes CWs model construction strategies and future directions, with the aim of providing theoretical support for the scientific design and reliable operation of CWs.