Traditionally, marine habitats such as seagrass and coral reefs have been mapped by collecting photography from airplanes (aerial photography) and manually drawing lines around habitat types based on interpretations of features in the photos. This traditional mapping process is costly and time consuming. In recent years, satellite-based imaging technology has progressed significantly, offering image quality comparable to aerial photography. Satellite imagery has also become more affordable with advances in technology and a growing market of applications. Along with satellite imagery, information processing methods have also advanced considerably, making it possible to automate parts of the image interpretation process for habitat mapping.
Satellite imagery can provide more information over time in a more efficient, continuous and consistent manner than aircraft-based photography. These modern data and methods can make habitat mapping in Florida more affordable and objective. Research staff at FWRI are working to leverage and build upon these new technologies to inform new tools and research methods that support natural resource management. The following examples illustrate how these modern approaches have been applied in both marine and terrestrial environments.
Mapping Seagrass and Colonized Hard Bottom in Springs Coast, Florida Using WorldView-2 Satellite Imagery
Springs Coast Florida is known for its expansive seagrass meadows that extend up to 15 miles from the coast. The area is characterized by a gradually sloping depth where seagrass slowly transition to large swaths of hard bottom colonized by soft corals, sponges and macroalgae. Seagrass offshore in deeper waters live on the fringe of their light requirements and are particularly sensitive to changes in water clarity. These offshore seagrass and hard bottom habitats are important corridors for inshore-offshore migration of economically important fish and shellfish such as shrimp, grouper, and snapper species, but these offshore habitats lack mapping information needed for species status and trends monitoring. Much of these expansive coastal marine habitats can be captured within a single continuous satellite image, which accommodates image processing techniques.
A recent pilot project conducted by FWRI research staff proved that satellite imagery and image processing techniques, such as Spectral Classification and Object Based Image Analysis, can be used to quantify the relative distribution and abundance of offshore seagrass and hard bottom habitats. Imagery for this 600 square kilometer study area in Springs Coast was captured by a WorldView-2 satellite. Rather than manually drawing lines around habitat type areas, these new methods quantitatively classify each image pixel based on spectral and geographic location information, producing a higher resolution map. It is expected that modern imagery data sources and image processing methods will over time improve our ability to map and track changes in coastal marine habitats.
This detailed comparison of image processing and photo-interpretation mapping results on the west coast of Florida highlights the higher resolution detail provided with image processing methods. The image processing line-work follows the image features more closely than the photo-interpretation line-work.
The Florida Cooperative Land Cover Map
The Florida Cooperative Land Cover dataset (CLC) is a partnership between the FWC and Florida Natural Areas Inventory (FNAI) to develop an ecologically-based statewide land cover data set. The CLC was developed using existing federal, state, and local data sources and expert review of aerial photography and ground conditions. The CLC uses the Florida Land Cover Classification System, which was developed to address the need for a single classification system that incorporates the level of detail and flexibility needed by the FWC and its conservation partners. The classification schema creates a system that uses well-defined land cover classes that are unique to the state of Florida but can also be incorporated with systems in neighboring states, as well as regionally. This classification system is limited to terrestrial, wetland, and inland aquatic classes.
The Cooperative Land Cover Map is continuously revised, with new versions being released every 6-12 months. Overall revisions to classes are conducted to correct for boundary errors, mislabeled classes, and hard edges between classes. Vector data are compared against high resolution Digital Ortho Quarter Quads and Google Earth imagery.
Areas with complex land cover mixtures undergo further revisions using remote sensing techniques, including image sharpening, pixel and object based classifications utilizing SPOT V (a commercial high resolution satellite) imagery, LiDAR, panchromatic imagery, and other available data sets. Spectral and texture signatures are obtained using the corrected polygon data as a guide for feature classification. Further visual inspections of classified areas are conducted for consistency, error detection, and edge matching between image footprints.
The Cooperative Land Cover dataset is available on FWC’s data library and public GIS data site. FWC now takes lead on updating and maintaining the dataset, while FNAI continues to provide guidance for the classification of natural communities and site-specific data sources, based on their mapping and revision efforts.