MACHINE LEARNING AND REMOTE SENSING APPROACHES TO DEVELOP HIGH-RESOLUTION SPATIAL PREDICTION OF AQUATIC INVASIVE SPECIES IN LAKES

Aquatic invasive species (AIS) have significant negative impacts on lake ecosystems, underscoring the need for improved detection and management. However, on-the-ground monitoring efforts are time- and resource-intensive. Recent advancements in satellite-based technology and machine learning algorithms present a promising pathway to predict AIS presence within and across lakes. The goal of this study was to train a statistical model to predict AIS presence within lakes. First, a large spatial dataset was created that included areas of mapped AIS locations, sonar-derived bathymetric conditions, proximity to anthropogenic points of interest (e.g., boat launches, beaches, campsites), and adjacent land cover. Then, three statistical models (linear regression, tree-based machine learning model, artificial neural network) were tested head-to-head to determine the model that best predicted AIS. The machine learning model (XGBoost) had the best model performance, correctly predicting AIS presence in ¾ of locations. The most important predictor variables were proximity to shoreline, forest, impervious cover, and agricultural land cover. This final model selected also minimized false negatives, an important outcome as this model will be used to guide monitoring activities by The Nature Conservancy. The model was applied to thousands of lakes in Adirondack Park, enabling prioritization of monitoring efforts for early detection surveys and other mitigation measures as well as serve as an invasive species communication tool for stakeholders and the public in the Adirondack region.

Primary Presenter: Kateri Salk, Tetra Tech (kateri.salkgundersen@tetratech.com)

Authors:

Kateri Salk, Tetra Tech  (kateri.salkgundersen@tetratech.com)

Brian Pickard, Tetra Tech  (brian.pickard@tetratech.com)

Mark Fernandez, Tetra Tech  (mark.fernandez@tetratech.com)

Brian Greene, The Nature Conservancy  (brian.greene@tnc.org)

Zachary Simek, The Nature Conservancy  (zachary.simek@tnc.org)

Tammara Van Ryn, The Nature Conservancy  (tammara.vanryn@tnc.org)