This study utilizes high-resolution suspended sediment data to develop and demonstrate the effectiveness of new computational tools suitable for analysis of watershed sediment dynamics. The combination of advanced analysis of storm-event dynamics with prediction tools will lead to a more robust method for estimating sediment production from watersheds that can be used to inform watershed management.
The Mad River watershed in central Vermont, in the Lake Champlain basin, was instrumented with a network of turbidity sensors and weather stations. Four years of near-continuous data was collected from the sensors along with many water quality samples. Relationships between turbidity and suspended sediment and phosphorus allow the sensors to be used as a surrogate for measuring the water quality parameters.
The high frequency data set of water quality measures allow for advance data analysis tools, such as artificial neural networks, to be applied. Artificial neural networks are used to develop the non-linear relationships between weather and water quality data. They also allow for the recognition and clustering of patterns in the data that lead to better understanding of the dynamics within the watershed and help to discern the sources and availability of sediments.
What do we mean by event sediment dynamics?
Here’s an example of how we identify patterns in the relationship between suspended sediment concentration and discharge to understand the sources and connectivity of sediments in the watershed. We’re working on different computational methods to both recognize and classify these events.
Hamshaw, Scott D., Mandar M. Dewoolkar, Andrew W. Schroth, Beverley C. Wemple, and Donna M. Rizzo. 2018. “A New Machine Learning Approach for Classifying Hysteresis in Suspended Sediment-Discharge Relationships Using High- Frequency Monitoring Data.” Water Resources Research. In Press.
Hamshaw, Scott D., Mandar M. Dewoolkar, Andrew W. Schroth, Beverley C. Wemple, and Donna M. Rizzo. 2018. “Unraveling Sediment Dynamics within Watersheds from Patterns in Suspended-Sediment Discharge Relationships.” March 19, 2018. Presented at the Northeast Geological Society of America Meeting, Burlington, Vermont. Link to Presentation
- National Science Foundation, Graduate Research Fellowship (Grant No. DGE-0925179)
- Vermont Water Resources & Lake Studies Center Grant
- Robert & Patricia Switzer Foundation
- Research on Adaptation to Climate Change (RACC) Program, Vermont EPSCoR (National Science Foundation Grant EPS-1101317)
- Basin Resilience to Extreme Events (BREE) Program, Vermont EPSCoR (National Science Foundation Grant OIA-1556770)
- Unmanned Aircraft Systems Team, UVM Spatial Analysis Lab
- Gund Institute of Environment
- Richard Barrett Foundation