Advances and challenges in multi‑scale water environment system modeling: from process simulation to a novel simulator architecture

Water ecological environment models serve as essential scientific tools for watershed ecological governance and management, yet they still exhibit notable limitations in systematicity, accuracy, and adaptability when addressing complex multi‑media and cross‑scale ecosystems. Current research lacks a systematic synthesis of the evolutionary pathways of multi‑scale models and has not fully integrated the strengths of artificial intelligence (AI) and mechanistic modeling, which constrains breakthroughs in water ecological system simulation from methodology to application. This paper systematically reviews the development trajectories and typical applications of water ecological environment models across different scales—including watersheds, rivers, lakes/reservoirs, urban water systems, and marine environments—proposes a “source‑flow‑network‑sink” multi‑process coupled systemic architecture, and explores pathways for integrating AI and environmental foundation models into simulation and prediction. The study finds that water ecological simulation in China urgently needs to shift from imported applications toward independent innovation and standardized development. Priority should be given to developing multi‑model coupling architectures with independent intellectual property, establishing localized parameter databases, and deeply incorporating AI and big‑data methods in model calibration, prediction, and uncertainty quantification. Furthermore, the research highlights that building intelligent simulator systems and promoting their operational application is a critical pathway for enhancing ecological risk early‑warning and decision‑support capabilities.