MetacouplingLLM: Integrating geographic adjacency, flow indicators, and LLM-assisted analysis
The metacoupling framework examines interactions within coupled human and natural systems and between adjacent or distant systems, but its application is constrained by inconsistent geographic classifications and limited standardized quantitative workflows. We present MetacouplingLLM, an open-source Python package integrating adjacency databases, reproducible flow indicators, and large language model (LLM)-assisted qualitative analysis. The databases combine administrative boundaries, hydrography, and fixed-crossing evidence, with configurable adjacency policies and optional country-level maritime adjacency. Metacoupled Flow Shares, Metacoupled Flow Evenness, and the Metacoupled Flow Concentration Index characterize flow composition, balance across coupling types, and partner concentration within each type. The qualitative workflow combines literature retrieval, optional web evidence, and geographic checks to draft structured analyses. Applications to global marine fishing and Indonesian palm oil demonstrate these capabilities; the fisheries reanalysis reveals differing partner concentration among regions with similar telecoupled flow shares. MetacouplingLLM facilitates consistent framework application and comparative sustainability research.
Authors
- Yingjie Li
- Nan Jia
- Zihan Zheng
- Jianguo Liu
- Xin Lan
- Xiang Yu
Publication Details
- Journal
- arXiv (Cornell University)
- Published
- 2026-11-18
- DOI
- https://doi.org/10.48550/arxiv.2611.23920
- Primary Topic
- Sustainability and Ecological Systems Analysis
- Type
- preprint