The files linked to this reference are the geospatial data created as part of the completion of the baseline vegetation inventory project for the NPS park unit. Current format is ArcGIS file geodatabase but older formats may exist as shapefiles. A geodatabase containing various feature class layers and tables show the locations of vegetation types and general land cover (vegetation map), vegetation plot samples, AA sites, project boundary extent, and aerial photographic centers. The feature class layer for the APIS vegetation map provides 4,862 polygons of detailed attribute data covering 28,971.9 ha, with an average polygon size of 6.0 ha. Of the area mapped, 4,732 polygons (97.4%) represent natural/semi-natural types in the NVCS, encompassing 17,600.9 ha (60.8%) of the total map extent.

The files linked to this reference are the geospatial data created as part of the completion of the baseline vegetation inventory project for the NPS park unit. Current format is ArcGIS file geodatabase but older formats may exist as shapefiles. Converting delineations to a digital format involved four main procedures: a) preparation of manuscript maps b) input of the spatial data: c) populating of the attribute tables, and d) conversion to GIS. Each step included many quality control steps. Maps were prepared by pin-registering a clean sheet of mylar to each photo and transferring delineations to the new overlay in ink. Each manuscript map was edgematched to the adjoining sheet. Each photo was numbered, and each polygon on the photo was numbered in sequence. An attribute table was created containing a field each for the photo number, polygon sequence number, land use code, layer number, stature type, height, density, and floristic composition attributes.

The files linked to this reference are the geospatial data created as part of the completion of the baseline vegetation inventory project for the NPS park unit. Current format is ArcGIS file geodatabase but older formats may exist as shapefiles. Converting delineations to a digital format involved four main procedures: a) preparation of manuscript maps b) input of the spatial data: c) populating of the attribute tables, and d) conversion to GIS. Each step included many quality control steps. Maps were prepared by pin-registering a clean sheet of mylar to each photo and transferring delineations to the new overlay in ink. Each manuscript map was edgematched to the adjoining sheet. Each photo was numbered, and each polygon on the photo was numbered in sequence. An attribute table was created containing a field each for the photo number, polygon sequence number, land use code, layer number, stature type, height, density, and floristic composition attributes.

The files linked to this reference are the geospatial data created as part of the completion of the baseline vegetation inventory project for the NPS park unit. Current format is ArcGIS file geodatabase but older formats may exist as shapefiles. ECognition software was used to circumscribe visually homogeneous polygons. Multiresolution segmentation (Scale parameter: 25, Shape: 0.2, Compactness: 0.2) was applied to four bands of 1 meter (3.3 ft) National Agricultural Imagery Program (NAIP) imagery and normalized difference vegetation index (NDVI) derived from the imagery. We used eCongition to calculate eighty-one variables and attributed the variables to the polygons. These variables summarized attributes of the topographic relationships of the polygons, reflectance values derived from aerial photography, surficial geologic composition, and vegetation height.

The files linked to this reference are the geospatial data created as part of the completion of the baseline vegetation inventory project for the NPS park unit. Current format is ArcGIS file geodatabase but older formats may exist as shapefiles. ECognition software was used to circumscribe visually homogeneous polygons. Multiresolution segmentation (Scale parameter: 25, Shape: 0.2, Compactness: 0.2) was applied to four bands of 1 meter (3.3 ft) National Agricultural Imagery Program (NAIP) imagery and normalized difference vegetation index (NDVI) derived from the imagery. We used eCongition to calculate eighty-one variables and attributed the variables to the polygons. These variables summarized attributes of the topographic relationships of the polygons, reflectance values derived from aerial photography, surficial geologic composition, and vegetation height.

The files linked to this reference are the geospatial data created as part of the completion of the baseline vegetation inventory project for the NPS park unit. Current format is ArcGIS file geodatabase but older formats may exist as shapefiles. A geodatabase containing various feature class layers and tables was constructed to show the locations of vegetation types and general land cover (vegetation map), vegetation plot samples, AA sites, project boundary extent, and aerial photographic centers. The feature class layer for the CUIS vegetation map provides 1,005 polygons of detailed attribute data covering 15,163.4 hectares, with an average polygon size of 144.7 hectares. Of the area mapped, 816 polygons (81% of all polygons) represent natural/semi-natural vegetation types in the NVCS, encompassing 9,542.8 hectares (62.9%) of the total map extent.

The files linked to this reference are the geospatial data created as part of the completion of the baseline vegetation inventory project for the NPS park unit. Current format is ArcGIS file geodatabase but older formats may exist as shapefiles. A geodatabase containing various feature class layers and tables was constructed to show the locations of vegetation types and general land cover (vegetation map), vegetation plot samples, AA sites, project boundary extent, and aerial photographic centers. The feature class layer for the CUIS vegetation map provides 1,005 polygons of detailed attribute data covering 15,163.4 hectares, with an average polygon size of 144.7 hectares. Of the area mapped, 816 polygons (81% of all polygons) represent natural/semi-natural vegetation types in the NVCS, encompassing 9,542.8 hectares (62.9%) of the total map extent.

The files linked to this reference are the geospatial data created as part of the completion of the baseline vegetation inventory project for the NPS park unit. Current format is ArcGIS file geodatabase but older formats may exist as shapefiles. In 2009, the National Park Service (NPS) Vegetation Mapping Inventory funded the NPS South Florida Caribbean Network (SFCN) to evaluate the accuracy of a vegetation map produced by the University of the Virgin Islands (UVI) Eastern Caribbean Center, Conservation Data Center for Virgin Island National Park (VIIS). The UVI vegetation map of VIIS was completed in 2001 and was based on aerial imagery from 1994. VIIS park staff felt that the UVI vegetation map was relatively accurate, but recognized that no formal accuracy assessment of the product had occurred at the time of its creation. Both the UVI and SFCN vegetation maps of VIIS relied on aerial imagery and photo-interpreters to delineate vegetation communities. However, the SFCN vegetation map had the benefit of having LiDAR data available; a technology and data source not readily available when the UVI vegetation map was produced. In addition, the SFCN vegetation map benefited from technological advances in aerial image acquisition that significantly improved the quality and resolution of imagery used; GPS that allowed precise spatial location determination; and GIS science that permitted the viewing, layering, and manipulation of multiple data sources simultaneously. The SFCN vegetation map also benefitted from the use of digital orthophotographs that take into account the surface elevation, topography, of the earth and camera tilt. The UVI vegetation map has an estimated map accuracy of 45.9% with a lower 90th Percentile Confidence Interval of 38.5% while the SFCN vegetation map accuracy is estimated at 87.9% with a lower 90th Percentile Confidence Interval of 82.0%. The SFCN vegetation map has approximately 2.1 times more detail, in the form of individual patches, than the UVI vegetation map does, 1,430 vs 686 patches, respectively. Mean patch size and maximum patch size are smaller in the SFCN vegetation map than in the UVI vegetation map. This results in the SFCN vegetation map being less homogeneous than the UVI map even though the total number of community types mapped are nearly identical, 27 vs 29.

The files linked to this reference are the geospatial data created as part of the completion of the baseline vegetation inventory project for the NPS park unit. Current format is ArcGIS file geodatabase but older formats may exist as shapefiles. In 2009, the National Park Service (NPS) Vegetation Mapping Inventory funded the NPS South Florida Caribbean Network (SFCN) to evaluate the accuracy of a vegetation map produced by the University of the Virgin Islands (UVI) Eastern Caribbean Center, Conservation Data Center for Virgin Island National Park (VIIS). The UVI vegetation map of VIIS was completed in 2001 and was based on aerial imagery from 1994. VIIS park staff felt that the UVI vegetation map was relatively accurate, but recognized that no formal accuracy assessment of the product had occurred at the time of its creation. Both the UVI and SFCN vegetation maps of VIIS relied on aerial imagery and photo-interpreters to delineate vegetation communities. However, the SFCN vegetation map had the benefit of having LiDAR data available; a technology and data source not readily available when the UVI vegetation map was produced. In addition, the SFCN vegetation map benefited from technological advances in aerial image acquisition that significantly improved the quality and resolution of imagery used; GPS that allowed precise spatial location determination; and GIS science that permitted the viewing, layering, and manipulation of multiple data sources simultaneously. The SFCN vegetation map also benefitted from the use of digital orthophotographs that take into account the surface elevation, topography, of the earth and camera tilt. The UVI vegetation map has an estimated map accuracy of 45.9% with a lower 90th Percentile Confidence Interval of 38.5% while the SFCN vegetation map accuracy is estimated at 87.9% with a lower 90th Percentile Confidence Interval of 82.0%. The SFCN vegetation map has approximately 2.1 times more detail, in the form of individual patches, than the UVI vegetation map does, 1,430 vs 686 patches, respectively. Mean patch size and maximum patch size are smaller in the SFCN vegetation map than in the UVI vegetation map. This results in the SFCN vegetation map being less homogeneous than the UVI map even though the total number of community types mapped are nearly identical, 27 vs 29.

The files linked to this reference are the geospatial data created as part of the completion of the baseline vegetation inventory project for the NPS park unit. Current format is ArcGIS file geodatabase but older formats may exist as shapefiles. The interpreted polygons were manually transferred to overlays that were registered to the base maps. Map unit attributes and appropriate physiognomic modifier codes were added to a second overlay. The overlays were subsequently rechecked for accuracy. Each overlay of transferred data was scanned using a large format sheet fed scanner at a resolution of 400 dots per inch. The resulting Tagged Image File Format (TIFF) images were then converted to a grid using ArcInfo (Version 7.2.1 Patch 2, Environmental Systems Research Institute, Redlands, California). For data produced with the DOQ base maps, the converted grid was projected to UTM Zone 15 using North American Datum of 1983 (NAD83).