这个工作是在我们研究团队之前研发的空间全分布式流域模拟框架SEIMS的基础上改进，从之前全流域可灵活定制一个模型结构，升级为允许在流域内对各子流域设定不同的流域模型结构，以满足流域内部应用场景（比如不同子流域的特点、数据可用性）的空间异质性。

这是我一个前年毕业直博生王玉靖的博士论文工作的第一部分，我想做这个问题的时间那就更早了，今年初终于在《Environmental Modelling & Software》发表出来，期待那篇博士论文最核心的创新点和研究工作能早日整理、发表出来。

Wang Y-J, Zhu L-J*, Qin C-Z, Zhu A-X. A spatially hybrid hydrological modeling approach using subbasin-specific model structures. Environmental Modelling & Software, 2026, 200, 106944. https://doi.org/10.1016/j.envsoft.2026.106944

Abstract

Hydrological models adopting spatially consistent model structures are often ill-suited for complex application contexts with significant spatial heterogeneity. Although existing hydrological modeling frameworks support spatially varying lumped or semi-distributed conceptual structures, they face challenges in integrating distributed physically-based structures. This paper proposes a new spatially hybrid hydrological modeling approach combining compatible spatial units and simulation algorithms to construct distinct model structures for individual subbasins within a watershed, integrated through channel routing for watershed-scale simulation. Implemented within the open-source Spatially Explicit Integrated Modeling System (SEIMS), this approach was evaluated through a proof-of-concept case study using two distinct model structures: a lumped conceptual model for gently sloping subbasins and a fully-distributed semi-physically-based model for mountainous subbasins. Comparative experiments demonstrate that the hybrid model inherits both the strengths and limitations of its constituent structures, achieving a balanced performance improvement. This approach enhances modeling flexibility, advancing knowledge-driven intelligent modeling in spatially heterogeneous environments.