Kaufman, Kristen A., et al. “The Use of Imagery and GIS Techniques to Evaluate and Compare Seagrass Dynamics across Multiple Spatial and Temporal Scales”

Kaufman, Kristen A., et al. “The Use of Imagery and GIS Techniques to Evaluate and Compare Seagrass Dynamics across Multiple Spatial and Temporal Scales” Estuaries and Coasts, doi:10.1007/s12237-020-00773-6.

image

Abstract

Using aerial imagery and GIS techniques, a retrospective study conducted fine-scale (1.0 m2 resolution) seagrass mapping to document the spatial heterogeneity within seagrass habitats previously mapped at a broad scale (0.202 ha resolution). Thirty randomly selected estuarine habitats in Tampa Bay, Florida, were manually interpreted and digitized using 0.3 m resolution imagery from 2004, 2006, and 2008. Seagrass patches and patterns of seagrass change were quantified at multiple levels of spatial organization and multiple temporal scales. Habitats classified as patchy seagrass were found to contain, on average, 52.7% less seagrass when mapped using the patch-based, fine-scale approach compared to broad-scale map data. In higher-resolution mapping, seagrass was increasingly differentiated from bare sediment and the amount of seagrass quantified within the study’s extent was reduced compared to broad-scale mapping, which included some bare area in estimates of seagrass cover. Additionally, fine-scale mapping detected the presence of seagrass on tidal flat habitat that was previously considered “unvegetated” in broad-scale mapping. These findings can be used to improve the definitions of broad-scale mapping classes and suggest the selection of smaller ratio digitizing scales when mapping complex or sparsely covered seagrass habitats. Overall, fine-scale mapping at a large spatial extent provided insights into the patterning and rate of seagrass gap formation and fragmentation, appears to be a reasonable method for local studies or resource disturbance assessments, and has potential applications as a training tool for further advancement of semi-automated classification techniques.