HydroAgent: Formalizing Forecaster Expertise into Skill-Orchestrated Flood Forecasting Workflows

2026-07-27

AI Summary

Existing AI-based flood forecasting methods struggle to formalize forecasters’ tacit expertise and lack explicit modeling of expert rules, review checkpoints, and workflow constraints. This work proposes HydroAgent, a novel framework that introduces a skill orchestration paradigm by embedding large language models (LLMs) into a model-driven forecasting pipeline. Explicit rules are encoded within modular skill components to constrain LLM reasoning, enabling the system to assist—rather than replace—human decision-making grounded in physical simulations. Experimental results across 129 flood events under five-fold cross-validation show Pearson correlation coefficients of 0.62 for peak discharge and 0.84 for flood volume. Scenario selection improved the Kling–Gupta Efficiency (KGE) by 0.023–0.154, and in 13–14 out of 14 events, simulated outcomes fell within expert-defined prior ranges. Across five LLMs, judgment accuracy ranged from 40% to 80%.

Abstract

Operational flood forecasting depends on tacit forecaster expertise that is difficult to formalize, audit, and transfer. Although artificial intelligence methods have advanced flood prediction and model-error correction, most existing studies have not explicitly represented the tacit expert rules, review checkpoints, and workflow constraints that connect model outputs to operational warning decisions. To address this issue, we propose HydroAgent, a skill-orchestrated agent framework that embeds Large Language Models (LLMs) into a model-driven flood forecasting workflow, where each skill encodes explicit rules to bound LLM reasoning. We validated its effectiveness using five state-of-the-art LLMs in the South Yamhill River basin. Our results demonstrate that prior judgment captures observed peak flow and flood volume within 5% tolerance in 10 and 11 out of 14 events, with 5-fold cross-validation over 129 events yielding Pearson correlations of 0.62 and 0.84. Building on a high-baseline scheme library (average KGE 0.890), the guided scheme selection further improves KGE by 0.023-0.154, with simulated peak flow and flood volume falling within the prior judgment ranges for 14 and 13 out of 14 events. All five tested LLMs successfully execute the HydroAgent workflow with comparable judgment accuracy (40%-80%), while showing moderate performance variation and substantial cost differences. HydroAgent does not aim to replace human forecasters; instead, it translates their tacit expertise into an auditable and reproducible workflow, streamlining analytical steps and supporting more informed decision-making. This skill-orchestrated paradigm demonstrates how explicit rule boundaries can guide language model reasoning to complement physically based simulation in next-generation flood forecasting.

Problem

• tacit expertise
• flood forecasting
• operational decision-making
• workflow formalization
• model-output interpretation

Innovation

• skill-orchestrated agent
• tacit expertise formalization
• Large Language Models (LLMs)
• flood forecasting workflow
• rule-bounded reasoning