Streamflow Rank Estimation (SRE) Model
June 12, 2025
This repo contains the source code for the Streamflow Rank Estimation (SRE) model, which is a deep learning model designed to estimate streamflow using timelapse imagery. The model is built using PyTorch and can be run on any system with or without a GPU.
This model has been integrated into the USGS Flow Photo Explorer, a cloud-based machine learning platform for training the model and running predictions on large image datasets from multiple users.
Citation Information
| Publication Year | 2025 |
|---|---|
| Title | Streamflow Rank Estimation (SRE) Model |
| DOI | 10.5066/P1YSANVM |
| Authors | Jeffrey (Contractor) D Walker, Amrita Gupta, Jennifer B Fair, Phillip J Goodling, Benjamin Letcher |
| Product Type | Software Release |
| Record Source | USGS Asset Identifier Service (AIS) |
| USGS Organization | Eastern Ecological Science Center at the Leetown Research Laboratory |
| Rights | This work is marked with CC0 1.0 Universal |