Background/Question/Methods   Anthropogenic landscape changes have markedly altered fire-return patterns in many fire-prone ecosystems. Fire-management plans are aimed at restoring fire-return patterns that match those with which the organisms composing the ecosystem have evolved. Despite the common use of fire-management plans, little effort has gone into the assessment of the floral and faunal responses to prescribed treatments. The goal of our study is to assess the successes and shortcomings of fire-management planning using long-term vegetation change data. We used a 31-yr data set (1977-2008) to assess the consequences of the Archbold Biological Station fire-management plan on plant-species composition and cover at 13 permanent, 200-m line transects located in rosemary scrub, scrubby flatwoods, hickory sandhill, flatwoods, seasonal ponds, wet prairie/bayhead associations that have experienced differing fire occurrences. We used nonmetric multidimensional scaling (NMS) ordination to assess vegetative changes over time.

Results/Conclusions   Vegetation recovery following fires was rapid with associations expressing considerable resiliency. Species responses and NMS-determined community-structure changes confirmed that associations at some sites were stable in response to fire treatments but other associations diverged appreciably from their desired state. Conclusions include the importance of monitoring and fine-tuning of fire-return intervals. Long fire-return-interval associations embedded in shorter return-time matrix associations suffer if matrix fire-return intervals are applied in burn units with multiple vegetation associations. In our study, rosemary scrub with 20-60 yr fire-return intervals is embedded within a scrubby flatwoods matrix, which has a shorter (6-9 yr) fire-return interval. Three fires within 3 decades resulted in rosemary scrub strongly shifting towards scrubby flatwoods. Florida rosemary, the foundation species, declined from 50% cover in 1977 to only 6% by 2008, while oak coverage markedly increased as a consequence of short-interval fire returns. Similarly, frequent fire returns (7 fires during 3 decades) coupled with spring dry-season fires at a wet prairie site (2-5 yr return time) prevented the development and maturation of an embedded bayhead association (60-100 yr return time). Many fire-management plans provide relatively narrow ranges of fire-return intervals and extend prescribed burning beyond the natural fire season. If wrong, narrow intervals and burning outside the natural fire season have the potential to alter species composition and abundance. Our results suggest that best practice should use broad variation in fire-return times, continual comparison of management goals with actual vegetation cover, careful management of embedded associations, and use of the natural fire season for prescription burns.