Led by Tony Chang, CSP and colleagues from the USGS have won a substantial award from National Geographic and Microsoft to support the creation and deployment of an innovative deep learning model for estimating streamflow.
CSP will develop this model using an extensive crowd-sourced training dataset of tens of thousands of stream images, matched to USGS stream gauge stations (see below). The goal of this unique modeling effort is to provide a cost-efficient monitoring system for ephemeral and intermittent streams in the northeastern US, prototyping the potential for national applications.
