Abstract Detail



Conservation Biology

Rominger, Kody [1], Meyer, Susan [2].

Vegetation Mapping with UAVs for Use in Habitat Suitability Modeling.

Astragalus holmgreniorum is a rare plant species endemic to a 60 km2 area extending from extreme southern Washington County, Utah to northern Mojave County, Arizona. It was listed as endangered in 2001 due to habitat loss from the encroachment of St. George, Utah, which is one of the fastest growing metropolitan areas in the U.S., and the associated anthropogenic impacts from that growth. In 2015, a habitat suitability model (Maxent) was developed for A. holmgreniorum using presence/absence of the plant, geology, relative elevation, landforms, slope, and several satellite-derived characters as variables. A major limitation of this model is that it did not include a vegetation variable. We know that A. holmgreniorum does not occupy habitat dominated by the shrub Larrea tridentata (Creosote bush), which is widespread throughout its range. Therefore, some areas predicated as having a high probability of presence were not habitat due to creosote occupancy. For the model to more accurately predict suitable habitat, vegetation must be considered. We decided to try a novel approach to vegetation mapping—drones. The use of small unmanned aerial vehicles (UAVs), or drones, for examining the landscape on a relatively fine-scale is an emerging technology in plant conservation. UAVs can quickly survey moderately large areas and acquire massive amounts of data with minimal on-the-ground impact. We used a consumer grade quadcopter to make 21 flights across roughly 600 acres of designated critical habitat for A. holmgreniorum. Flights were conducted over two four-hour days with the drone flying at 35-meter altitude which resulted in over 5,000 images at 1.3-cm resolution. Automated imagery processing software was used to generate orthomosaics and digital surface models (DSMs). We used ArcGIS maximum likelihood analysis to detect individual creosote plants using color as training data. The resulting raster was then converted to vector data and a buffer applied to outline each individual or group of creosote plants. The vegetation layer was then overlaid across the habitat suitability model, effectively masking areas previously identified as suitable. Additional results are forthcoming; however, initial results have shown that UAVs have the potential to drastically reduce the cost and on-the-ground effort for fine-scale vegetation mapping. 


1 - Utah Valley University, 800 W University Pkwy, Orem, UT, 84058, United States
2 - USFS SHRUB SCIENCES LABORATORY, 735 North 500 East, Provo, UT, 84606, United States

Keywords:
Conservation
Ecology
Vegetation Mapping
Drone
Endangered Species
Habitat Suitability.

Presentation Type: Oral Paper
Number: 0001
Abstract ID:309
Candidate for Awards:None


Copyright © 2000-2018, Botanical Society of America. All rights reserved