Informing Seagrass Management and Restoration along the Florida Gulf Coast through Remote Sensing and Spatiotemporal Analyses of Seagrass Distribution

Seagrasses are valued for the range of ecosystem services they provide such as nursery and adult habitat for commercially and recreationally fished species, a critical food resource for threatened species, and filtration systems for improving water quality. In Florida, seagrasses are estimated to contribute $20 billion annually in benefits to the Florida Gulf Coast (FGC) region. However, even with the numerous benefits that seagrasses have been shown to provide they are still at risk due to anthropogenic pressures. Given the established importance of and threats to seagrasses, seagrass mapping and monitoring programs in the FGC have worked to better understand and predict the patterns of seagrass decline. Traditional methods for monitoring or mapping seagrasses, such as intensive field surveys (scuba/snorkeling) or costly aerial surveys often result in inconsistencies in the data necessary to better understand the spatial and temporal dynamics of seagrasses. This project used Landsat 5 satellite imagery to classify seagrasses in the Big Bend region of the FGC using a random forest model. Four years of data over a 15 year timeframe were classified and accuracies between 78-88% were achieved for the seagrass class. Counter to the prevailing narrative of seagrass degradation throughout the Gulf coast, 291 km2 of seagrass were gained between 1996-2011, with only 116 km2 lost in this same time period. Both seagrass losses and gains were positively spatially autocorrelated. Results from this project indicate that both accurate classifications and spatiotemporal analyses can be conducted using remotely sensed data for areas with limited or inconsistent field survey data.