2019-02-142019-02-142018-122018-122018-12https://hdl.handle.net/2346/82704In this study, remote sensing and landcover classification were utilized to detect black-tailed prairie dog towns throughout the city of Lubbock without the use of field visits. Using eight high resolution (0.5 meters) aerial images covering roughly 374 square kilometers, training data identifying the location of prairie dog towns, alongside training sites for other basic landcover types, were used to highlight the areas containing black-tailed prairie dogs. Intensity values found within freely acquired LiDAR data were also used in conjunction with the aerial imagery, to potentially improve accuracy. In addition to random-point accuracy analyses, a region-based accuracy assessment—in which large cell clusters were highlighted within classified imagery—was completed. This enabled me to easily mark detected areas, as well as assess the quality of training data. Overall accuracy with just the aerial imagery was 37%, which was later improved to 46% with the inclusion of the LiDAR intensity data. Accuracy was further improved to the range of 62-68% after examining the regions created through classification. Although accuracy throughout the study was not necessarily high, the methods used can potentially reduce field work in conservation or habitat research for large areas and highlighted many previously unknown prairie dog towns.application/pdfengLiDARBlack-Tailed Prairie DogAerial ImageryRemote SensingLandcover ClassificationImage InterpretationLubbockTexasHabitatHigh ResolutionThesis2019-02-14Access is not restricted.