AQUAH: Automatic Quantification and Unified Agent in Hydrology

Abstract

We introduce AQUAH, the first end-to-end language-basedagent designed specifically for hydrologic modelling. Starting from a simple natural-language prompt (e.g., “simulatefloods for the Little Bighorn basin from 2020 to 2022”),AQUAH autonomously retrieves the required terrain, forcing, and gauge data; configures a hydrologic model; runsthe simulation; and generates a self-contained PDF report.The workflow is driven by vision-enabled large-languagemodels, which interpret maps and rasters on the fly andsteer key decisions such as outlet selection, parameter initialisation, and uncertainty commentary. Initial experiments across a range of U.S. basins show that AQUAHcan complete cold-start simulations and produce analystready documentation without manual intervention—resultsthat hydrologists judge as clear, transparent, and physically plausible. While further calibration and validationare still needed for operational deployment, these earlyoutcomes highlight the promise of LLM-centred, visiongrounded agents to streamline complex environmental modelling and lower the barrier between Earth-observationdata, physics-based tools, and decision makers