Vegetation Height in Open Space in San Diego County, Derived from 2014 NAIP Imagery and 2014/2015 Lidar

Shrublands have seen large changes over time due to factors such as fire and drought. As the climate continues to change, vegetation monitoring at the county scale is essential to identify large-scale changes and to develop sampling designs for field-based vegetation studies. This dataset contains two raster files that each depict the height of vegetation. The first layer is restricted to actively growing vegetation and the second is restricted to dormant/dead vegetation. Both layers cover open space areas in San Diego County, California. Height calculations were derived from Lidar data collected in 2014 and 2015 for the western two-thirds of San Diego County. Lidar point clouds were pre-classified into ground and non-ground. Rasters for the Digital Elevation Model (DEM) and Digital Surface Model (DSM) were calculated using ArcGIS software using ground classified points and last returns for the natural surface (DEM) and non-ground first returns for the surface model (DSM). The spatial resolution for both layers is 1 meter and aligns with 2014 National Agriculture Imagery Program (NAIP) imagery. Object height was calculated by subtracting the DEM from the DSM in meters. To remove structures or non-natural objects from the imagery, layers were clipped to open space areas using the National Land Cover Database, building footprints, roads, and railways. This ensures that objects above the natural surface are vegetation, even when Normalized Difference Vegetation Index (NDVI) numbers are very low. NDVI measures the amount of photosynthetically active vegetation in the raster cell. Healthy vegetation reflects high levels of near-infrared and low levels of red electromagnetic radiation. NDVI ranges from -1 to 1 with low values indicating little or no presence of healthy vegetation and higher values indicating the presence of healthy vegetation. The NDVI was calculated from the 2014 NAIP imagery and a cutoff of 0.1 was used to separate photosynthetically active vegetation from non-vegetated or dormant/dead vegetation areas. The imagery was collected during 2014, an exceptional drought year. It is not possible to separate extremely water-stressed plants from truly dead plants using only NDVI. The natural surface was verified using established National Geodetic Survey (NGS) benchmarks and exceeded 98 percent accuracy. Vegetation structure was validated using visual assessments of high-resolution aerial imagery to verify the vegetation form and greenness. Vegetation form and health (NDVI) had an accuracy of 82 percent.