GeoFEW SIUC

Proof-of-Concept Model and Datasets for an End-to-End Hydro-AI Framework for Participatory Water Resources Planning

This repository contains the data, software, and reproducibility materials supporting the manuscript “An End-to-End Hydro-AI Framework for Participatory Water Resources Planning: Coupling Agentic Language Models with Physics-Informed Surrogates.” The archived materials include the MODFLOW-2005 benchmark configuration, trained physics-informed surrogate, model-training scripts, scenario schema and validation logic, evaluation and figure-generation code, and the 16-phrasing natural-language prompt test set used in the study. These materials support reproduction and evaluation of the computational results reported in the manuscript.