Ecology--Remote sensing

The Florida Everglades is a vast subtropical wetland that historically spanned over 1,000,000 hectares, but much of the Everglades has changed in the last 100 years due to anthropogenic activity. Collier County was subject to alteration in the form of logging, road building, and canal digging. These actions disrupted the natural sheet flow of water and had large environmental impacts on the region, impacts which are slowly being addressed by Everglades restoration efforts. The aim of this project was to observe the effects of environmental change at a cypress swamp forest in Collier County within the Big Cypress National Preserve. Using sediment core data including charcoal analysis, loss on ignition, and peat humification, as well as remote sensing techniques, this project uses a novel approach to assess local environmental conditions in the modern era. Historical records and contemporary data are used to evaluate change over time, and satellite imagery is used to quantify vegetative health. Modification of the environment related to anthropogenic activity is noted, and evidence of progress from restoration efforts is observed from the last two decades in our study’s data.

Invasive exotic plant species cause a number of problems in native south
Florida ecosystems, and a great deal of effort is being put into controlling the
populations ofthese species. Control efforts require updated information on the
locations of exotic species. This information can be obtained from high resolution
remotely sensed data such as digital orthoimagery and LIDAR. Extraction of
information from these data sources is often problematic using traditional pixel-based
image processing techniques. An object oriented method of image analysis, however,
has been shown to be better suited to this task.
One invasive exotic species that has become widespread in south Florida is
Casuarina equisetifolia, also known as Australian pine. This study develops a semiautomated
procedure for detecting Australian pine over a large, diverse area with high
resolution remotely sensed data using the object oriented method of analysis.

The development of high resolution LIDAR DSM combined with digital infrared ortho-photography data enhances the ability to map canopy structures with a higher degree of accuracy and precision than with either data set alone. The purpose of this thesis is to map Australian Pine (Casuarina equisetifolia with a 85% or greater accuracy by creating a methodology that uses LIDAR and color infrared ortho-photography and to test it within three different landscape types within Broward County. LIDAR features below a determined height threshold (i.e. Deerpoint 25 ft) were eliminated and recoded to 0 to create Mask 1. NDVI technique separated non-vegetative features from vegetative features to create Mask 2. Mask 1 and Mask 2 were merged and overlaid on the raw LIDAR data set to perform isodata clustering, as well as density slicing to identify mature Australian Pines. Careful delineation of study areas is critical to obtain the highest possible accuracy. Density slicing proved to be a faster and less time consuming technique for achieving 85% level of accuracy than compared to isodata clustering.