Abstract Detail


Caruso, Kathryn [1], Lampley, Jayne [1], Rock, Monika [1], Siddhi, Ashna [1], Zhu, Seamore [1], Randall, John L. [2], Waitt, Damon [2], Weakley, Alan [2], Payne, Sarah [1], Madden, Marguerite [3].

GIS Spatial Analysis and Habitat Suitability Modeling Using Remote Sensing Data to Help Identify At-risk Populations of US Carolina Coastal Plain Endemic Venus Flytrap (Dionaea muscipula).

Although the iconic, carnivorous plant species Venus flytrap (Dionaea muscipula, Droseraceae) is identifiable worldwide, its fragmented distribution is restricted to only a small portion of the Carolina Coastal Plain of the southeastern United States. Furthermore, D. muscipula is considered a vulnerable species by botanists; within its limited range, populations are threatened by habitat loss, fire suppression, and plant poaching. Habitat suitability modeling and predictive modeling of future land use change are useful tools in conservation planning for rare and threatened species. NASA DEVELOP partnered with the North Carolina Botanical Garden, the University of North Carolina Herbarium, and the Natural Heritage Programs of North and South Carolina to identify suitable D. muscipula habitat for reintroduction efforts, areas of enduring suitable habitat for management prioritization, and potential locations of previously undocumented populations. A present-day map of likely suitable D. muscipula habitat is presented based on models run using the Software for Assisted Habitat Modeling (SAHM) within the VisTrails platform. Five separate models within SAHM were evaluated: Boosted Regression Tree (BRT), Generalized Linear Model (GLM), MaxEnt, Multivariate Adaptive Regression Spline (MARS), and Random Forest (RF). The input data processed by SAHM included 23 environmental predictor variables and species occurrence data from herbaria records. A 2050 land-use change map is also presented which forecasts areas at high risk of development in the near future. This predictive modeling was performed with TerrSet Land Change Modeler software and was based on a time series of remotely sensed Earth observation satellite imagery and predictor variables such as distance to urban areas and the coast. The results show that approximately 39% of highly suitable habitat, i.e., the area of highest-confidence suitable habitat (probability >80%), occurs within currently protected state or federal land and is at minimal-risk of development. Of the remaining unprotected land, nearly 50% is highly suitable habitat that is forecasted to be developed by 2050 (probability >50%). In conclusion, potential conflict zones occur largely along the coast where development is most likely, and reveal the D. muscipula populations which are at the greatest risk of local extirpation. In contrast, the lower risk areas farther inland may serve as enduring D. muscipula habitat and a safer investment of conservation efforts. Modeling methods and choice of appropriate environmental predictor variables are also discussed.

1 - NASA Develop, Science Systems and Applications, Inc., 10210 Greenbelt Road, Suite 600, Lanham, MD, 20706, USA
2 - University of North Carolina at Chapel Hill, North Carolina Botanical Garden, CB 3375, Chapel Hill, NC, 27599, USA
3 - University of Georgia, Department of Geography, Center for Geospatial Research (CGR), Athens, GA, 30602, USA

Species Distribution Modeling
remote sensing
native species
Endemic Species
Coastal Plain
Ecological Forecasting
threatened species
carnivorous plant.

Presentation Type: Oral Paper
Number: BIOG III011
Abstract ID:643
Candidate for Awards:None

Copyright © 2000-2022, Botanical Society of America. All rights reserved