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

📄 arXiv: 2607.23983v1 📥 PDF

作者: Qingyi Yang, Siqian Qiu, Bing Li, Xu Shan, Jia Feng, Shunan Zhou, Xudong Zhou, Tiantian Xing, Jiale Guo, Xiaoyi Dong, Gaoyu Liu, Xiaohuan Liu, Haiqing Pu, Qingwen Deng, Xun Zhang, Zhongrun Xiang, Haiyang Qian, Ying Yan, Yongkang Xu, Nuo Lei, Tianlong Jia, Baoying Shan, Carlo De Michele

分类: physics.geo-ph, cs.LG

发布日期: 2026-07-27


💡 一句话要点

提出HydroAgent以解决洪水预报中专家经验难以形式化的问题

🎯 匹配领域: 支柱九:具身大模型 (Embodied Foundation Models)

关键词: 洪水预报 人工智能 大型语言模型 专家经验 技能协调 模型驱动 决策支持 可审计性

📋 核心要点

  1. 现有洪水预报方法未能有效形式化和转移专家的隐性经验,导致决策过程不够透明和可审计。
  2. HydroAgent通过将大型语言模型嵌入洪水预报工作流程,利用明确的规则约束模型推理,从而实现技能协调的预报。
  3. 实验结果显示,HydroAgent在多个事件中成功捕捉洪水峰值流量和体积,且与基线相比,KGE指标显著提升。

📝 摘要(中文)

操作性洪水预报依赖于难以形式化、审计和转移的隐性专家经验。尽管人工智能方法在洪水预测和模型误差修正方面取得了进展,但大多数现有研究未能明确表示连接模型输出与操作警报决策的隐性专家规则、审查检查点和工作流程约束。为了解决这一问题,我们提出了HydroAgent,一个技能协调的代理框架,将大型语言模型(LLMs)嵌入模型驱动的洪水预报工作流程中,每个技能编码明确的规则以约束LLM推理。我们在南亚姆希尔河流域验证了其有效性,结果表明,先前判断在14个事件中有10和11个事件的观测峰值流量和洪水体积在5%的容忍度内,129个事件的5折交叉验证得到了0.62和0.84的Pearson相关性。HydroAgent并不旨在取代人类预报员,而是将他们的隐性专业知识转化为可审计和可重复的工作流程,简化分析步骤并支持更明智的决策。

🔬 方法详解

问题定义:论文要解决的问题是如何将隐性专家经验形式化,以增强洪水预报的透明度和可审计性。现有方法在连接模型输出与操作决策时缺乏明确的规则和审查机制,导致预报的可靠性和准确性不足。

核心思路:论文提出的核心思路是构建HydroAgent框架,将大型语言模型(LLMs)与明确的规则结合,形成一个技能协调的洪水预报工作流程。这种设计旨在通过规则约束LLM的推理过程,从而提高预报的准确性和可解释性。

技术框架:HydroAgent的整体架构包括多个模块:首先是数据输入模块,接着是LLM推理模块,随后是规则约束模块,最后是决策输出模块。每个模块通过明确的接口连接,确保信息的流动和处理的顺畅。

关键创新:HydroAgent的关键创新在于将隐性专家知识转化为可审计的规则,并将其嵌入到LLM的推理过程中。这与现有方法的本质区别在于,后者通常缺乏对专家经验的明确表示和利用。

关键设计:在设计过程中,HydroAgent采用了多种参数设置和损失函数,以优化模型的推理效果。具体而言,使用了高基线方案库(平均KGE 0.890)作为基础,并通过引导方案选择进一步提升KGE指标,确保模拟的峰值流量和洪水体积符合先前判断范围。

🖼️ 关键图片

fig_0
fig_1
fig_2

📊 实验亮点

实验结果显示,HydroAgent在14个洪水事件中,成功捕捉到10和11个事件的峰值流量和洪水体积,且5折交叉验证的Pearson相关性达到0.62和0.84。通过引导方案选择,KGE指标提升了0.023至0.154,显示出显著的性能改进。

🎯 应用场景

HydroAgent的研究成果具有广泛的应用潜力,尤其在洪水预报、气候变化适应和水资源管理等领域。通过将专家经验形式化,该框架能够提高预报的可靠性,支持决策者在面对极端天气事件时做出更为明智的选择,进而减少潜在的经济损失和人员伤亡。

📄 摘要(原文)

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.