PEP2026: USGS HIVIS Observation of Hurricane Debby 2024 in Raleigh, NC on Walnut Creek at South State Street; Image and Sensor Data with GRIME-AI Processing

This resource was developed as a participatory exercise during the Pixels to Enviro Patterns 2026 workshop at the University of Nebraska, Lincoln.

The authors retrieved imagery and water depth data from the USGS monitoring station Walnut Creek at South State Street at Raleigh, NC - USGS-0208735460. The specific time period targeted for analysis was the flood hydrograph of Hurricane Debby (2024), which occurred in Raleigh between August 8, 2024 and August 11, 2024.

189 images were retrieved and processed using GRIME-AI to extract relevant imagery characteristics (entropy, green chromatic coordinate, and greenness-redness index). These colors were of interest as they were expected to correspond to vegetation (primarily green) and flood flows (primarily red, due to abundant red clay soils). Values were inspected for changes across the flood hydrograph and compared across whole-image values and specific regions of interest (water and streambank).

A subset of five images was used to train a machine learning model to distinguish between vegetation and water, and this model was applied to the entire 189-image set. The generated heatmaps and model predictions are included herein.