Advisor

Ervin, Gary N.

Committee Member

Wallace, Lisa

Committee Member

Schauwecker, Timothy

Date of Degree

1-1-2011

Document Type

Graduate Thesis - Open Access

Degree Name

Master of Science

Abstract

Mississippi’s Blackland Prairie has been reduced below 10% of pre-Columbian extent, with few conservation practices in place. To determine efficacy of current restoration practices, plant species at remnant sites were compared with those at restoration sites. Analyses using multivariate statistical approaches revealed no generalizable patterns among four available remnants versus two available restoration sites. Thus, the aim of this project shifted to evaluating methods of identifying Blackland Prairie remnants or potential restoration sites. Location data for Blackland Prairie plant species and potentially informative environmental variables were used to develop geographic information system (GIS)-based habitat models. The best models were selected for validation against a second set of data collected from random points on public lands across the survey region. Validation surveys also were used to explore trends in predictive success and to aid in increasing accuracy through inclusion of other variables. Models incorporating soil characteristics had the highest predictive success.

URI

https://hdl.handle.net/11668/17300

Share

COinS