The overall objective of this project was to develop and test an ecological assessment framework for recovery of understory plant communities in longleaf pine savannas. This framework provides a means to: (1) determine the degree to which a community is degraded and thus the need for and feasibility of recovery; (2) aid in selecting a successful recovery strategy; and (3) assess progress towards recovery. The first technical objective was to assess sites and define recovery goals by using comprehensive surveys of vegetation in degraded and reference communities at three separate Department of Defense (DoD)/Department of Energy (DOE) sites. The second technical objective was to use distributed experiments to evaluate the determinants of recovery. The project team conducted factorial experiments that manipulated herbivore access, seed addition, and competition across a gradient of common management regimes (e.g., fire frequency, timber density) and land-use histories (i.e., historical use of land for agriculture vs. land never used for agriculture) to determine how multiple, interacting factors might dictate recovery. The third technical objective was to develop the roadmap to recovery, a model of management recommendations to provide informed guidance to managers for implementing recovery given the unique set of starting conditions at their area.

Technical Approach

To understand recovery of longleaf pine (Pinus palustris) stands on DoD military installations, this project followed a three-phase process that directly parallels the three technical objectives. In Phase 1, longleaf pine savanna understory plant communities were assessed using ecological reference models parameterized with available datasets related to historic and contemporary drivers of degradation. In Phase 2, the importance of ecological mechanisms (seed dispersal limitation, plant consumers, plant competition, and microsite availability) were experimentally evaluated for recovery of the savannas identified in Phase 1. In Phase 3, a recovery assessment model was created based on the results of Phases 1 and 2. All three phases were conducted simultaneously at three separate locations: Fort Bragg (DoD) near Fayetteville, North Carolina; Fort Stewart (DoD) near Hinesville, Georgia; and the DOE Savannah River Site near Aiken, South Carolina.


Phase 1. Historical agriculture, low fire frequency, and high tree density all contributed to the degradation of longleaf pine understory plant communities when compared to community-appropriate reference stands. Notable differences occurred among the three geographic locations that depended largely on historical human activities (i.e., intensity of agriculture and related soil compaction), as well as on contemporary management activities (i.e., frequency of prescribed burns and tree density). The seed bank in all geographic locations and especially in post-agricultural sites contained mostly ruderal plant species and, as a result, the seed bank has low potential to assist recovery of plant communities indicative of reference sites. Interactions among local degrading factors (e.g., historical agricultural, fire suppression, high overstory tree density) and edaphic conditions structure longleaf pine understory communities; whereas, the role of the surrounding landscape is negligible.

Phase 2. Seed additions promoted the recovery of species that are indicators of reference plant communities, regardless of the initial level of site degradation (i.e., degradation class); however, degradation class and other restoration techniques interacted to determine how effective seed additions were in recovering understory plants. Reduction of competitors increased the effectiveness of seed additions, but only in degraded sites with few preexisting indicator species. Similarly, removing leaf litter also increased the effectiveness of seed additions, but mainly in fire suppressed sites and sites with high tree density. Fencing to eliminate large mammalian herbivores can also increase the effectiveness of seed additions in some cases, although the effects of fencing are more subtle and variable. Finally, the results from small plots (1 m2) scale up to larger spatial scales (25 m2) and, in many cases, recovery effectiveness strengthens at larger spatial scales, suggesting that the results of small experiments can provide valid recovery guidance.

Phase 3. A model of recovery recommendations was developed that is applicable to a variety of sites throughout southeastern longleaf pine savannas: the Roadmap to Recovery. The goal of this document is to provide a field-ready means to implement and assess recovery that is easily accessible to land managers. To maximize the utility of the Roadmap to Recovery, climatic and biophysical conditions at other DoD/DOE installations in the Southeast were compiled to provide land managers with a means to quantitatively evaluate which of the recovery models is suited to their location.


The DoD and DOE are faced with the challenge of implementing the successful recovery of southeastern plant communities in a way that is consistent with continued execution of mission-related activities. Achieving this goal requires: (1) quantifying ecological reference models; (2) developing a means to assess how degraded stands differ from reference stands; (3) determining which recovery strategies will be effective; and (4) devising metrics to quantify recovery progress. Through a unique combination of historical land-use data, contemporary vegetation data, and large-scale field experiments at three separate installations, the outcomes of this project provide a means of assessment and recovery of southeastern longleaf pine plant communities.

  • Plant Ecology,

  • Pine,

  • Southeastern,