National Exposed Sediment Search and Inventory (NESSI): Utilizing Satellite Imagery and Machine Learning to Identify Dredged Sediment Placement Site Recovery

Thomas P. Huff, Emily R. Russ, Todd M. Swannack

Anthropogenic activity leads to changes in sediment dynamics, creating imbalances in sediment distributions across the landscape. These imbalances can be variable within a littoral system, with adjacent areas experiencing sediment starvation and excess sediment. Historically, sediments were viewed as an inconvenient biproduct destined for disposal; however, beneficial use of dredge material (BUDM) is a practice that has grown as a preferred methodology for utilizing sediment as a resource to help alleviate the sediment imbalances within a system. BUDM enables organizations to adopt a more innovative and sustainable sediment management approach that also provides ecological, economic, and social co-benefits. Although location data are available on BUDM sites, especially in the US, there is limited understanding on how these sites evolve within the larger landscape, which is necessary for quantifying the co-benefits. To move towards BUDM more broadly, new tools need to be developed to allow researchers and managers to understand the effects and benefits of this practice. The National Exposed Sediment Search and Inventory (NESSI) was built to show the capability of using machine learning techniques to identify dredged sediments. A combination of satellite imagery data obtained and processed using Google Earth Engine and machine learning algorithms were applied at known dredged material placement sites to develop a time series of dredged material placement events and subsequent site recovery. These disturbance-to-recovery time series are then used in a landscape analysis application to better understand site evolution within the context of the surrounding areas.