HydroBlox: AI-Assisted Visual Programming Framework for Enhanced Scientific Reproducibility in Hydrology

Carlos Erazo Ramirez, Ibrahim Demir

Scientific workflow reproducibility for hydrological and environmental analyses remains a challenge due to the heterogeneity of data sources, analysis protocols, and evolving visualization needs. This study introduces HydroBlox, a client-side browser-based framework that supports the creation, execution, and export of hydrological workflows using a visual programming interface. The platform integrates modular web libraries to perform data retrieval, statistical analysis, and visualization directly in the browser. Two case studies are presented in the study includes analyzing precipitation-streamflow response relationships in the Iowa River Basin and computing the Standardized Precipitation Index using a WebAssembly-enhanced drought analysis workflow. Results demonstrate the system’s capacity to facilitate reproducible, portable, and extensible hydrological analyses across spatial and temporal scales. The study discusses the architecture, implementation, and capabilities of the system and explores its implications for collaborative research, education, and low-code scientific computing in hydrology.