Confidential — Stefan Michaelcheck Only

Advancing Hydrologic Knowledge Access through an AI-Powered Conversational Bot for CIROH DocuHub

2026application demonstrationapplicationsystem

Abel Andrés Ramírez Molina, Jiaqi Gong, Austin Schraivogel, Shenglin Li, Arpita Patel, Fabricio Joel Gutiérrez Juárez, Margulan Baizhakyp

https://doi.org/10.1061/9780784486931.004OpenAlex: W7155351289
1
URLs Found
0
Internal Citations
7
Authors
usable
Abstract Quality
GPT-5.5 Abstract Analysis

Problems Identified (2)

Fragmented Hydrologic Knowledge: Hydrologic knowledge resources are distributed across multiple heterogeneous platforms, making timely and accurate information access difficult for researchers.

Fragmented Hydrologic Knowledge: Hydrologic knowledge resources are distributed across multiple heterogeneous platforms, making timely and accurate information access difficult for researchers.

Proposed Solutions (5)

Domain-Specific RAG Chatbot: The paper presents an AI-powered conversational bot for CIROH DocuHub using a Retrieval-Augmented Generation pipeline grounded in hydrologic documentation.

Two-Stage Semantic Retrieval: The system first searches document summaries for relevant high-level pages and then retrieves associated text chunks and neighbors to preserve local context.

Embedded Chat Interface: A floating chat interface embedded in DocuHub supports curated example queries and open-ended user exploration.

Domain-Specific RAG Chatbot: The paper presents an AI-powered conversational bot for CIROH DocuHub using a Retrieval-Augmented Generation pipeline grounded in hydrologic documentation.

Two-Stage Semantic Retrieval: The system first searches document summaries for relevant high-level pages and then retrieves associated text chunks and neighbors to preserve local context.

Results (3)

Improved Grounded Answers:

Scalable Hydrologic Knowledge Framework:

Expanded Coverage Ongoing Work:

Research Domain

Hydrologic knowledge access and environmental data management

← Back to all papers