Woody plants in fire-frequented ecosystems commonly resprout from underground organs after fires. Responses to variation in characteristics of fire regimes may be a function of plant physiological status or fire intensity. Although these hypotheses have been explored for trees in southeastern longleaf pine (Pinus palustris) savannas, responses of other life forms and stages have not been studied. We examined effects of fire season and frequency, geography, habitat, and underground organ morphology on resprouting of shrubs. In 1994, we located replicated sites, each containing two habitats, upslope savannas and downslope seepages, in Louisiana and Florida. Each site, which contained quadrats located along transects within a 30 ϫ 60 m plot, was burned either during the dormant or growing season and then reburned similarly two years later. Maximum fire temperatures were measured, and densities of shrub stems were censused in quadrats before and after fires.Shrubs collectively resprouted more following dormant than growing-season fires, regardless of habitat or geographic region. After repeated dormant-season fires, collective densities in seepages of both regions and densities of root-crown-bearing shrubs in Florida seepages were greater than those initially and after repeated growing-season fires. Shrub responses were generally unrelated to fire temperatures, supporting the hypothesis that resprouting of shrubs may be more dependent on their physiological status at the time of fires. There was, nonetheless, an inverse relationship between collective and root-crownbearing shrub densities following repeated fires and maximum fire temperatures in Florida seepages. Anthropogenic dormant-season fires over many decades may have resulted in increases in shrub densities in longleaf pine savannas, especially seepages. Repeated growing-season fires, however, neither increased nor reduced densities of established shrubs. Long-term shifts in characteristics of fire regimes, even in fire-frequented habitats, may produce effects that are not reversible in the short term (Ͻ10 yr) by simply reintroducing prescribed fires that resemble those that occurred naturally during the growing season.

Differences in initial large-scale disturbances might change effects of subsequent large-scale disturbances. We explored possible effects of prior fire regimes on subsequent hurricane-related mortality of south Florida slash pine (Pinus elliottii var. densa) in remnant Everglades pine savannas that were unburned, burned during the wet (lightning fire) season, or burned during the dry (anthropogenic fire) season in the decade before Hurricane Andrew (1992). We measured direct mortality during Andrew (snapped trees) and extended mortality over the subsequent 24-30 mo (mainly insect attacks on damaged trees). We used Bayesian model averaging to obtain probabilities of different models of survival based on fire regime and site characteristics (remnant area, distance to the Atlantic Ocean, depth to water table in the dry season, sustained wind speeds, tree sizes). Most likely models for direct and extended mortality included large negative effects of tree size and dry-season fire regime, and positive effects of stand area (direct mortality) and wetseason fire regime (extended mortality). Depth to water table and distance to the ocean had less certain effects. Our results, not predicted from fires or hurricanes alone, suggest that anthropogenic changes to dry-season fires strongly influence the effects of subsequent hurricanes on the mortality of pines in subtropical savannas.

We examined the demographic responses of Fagus grandifolia to Hurricane Kate (1985) in an old-growth Southern mixed-hardwood forest in northern Florida. Matrix population models were used to contrast pre-and post-hurricane population trends (1978)(1979)(1980)(1981)(1982)(1983)(1984)(1985)(1986)(1987)(1988)(1989)(1990)(1991)(1992) with predictions of four hypotheses of the persistence of this shade-tolerant species in hurricane-frequented forests (resistance, recovery, release, and complementation). Although growth and mortality schedules changed as a result of the hurricane, the asymptotic population growth rate did not change significantly and did not depart significantly from ϭ 1 (stability) in either the pre-or the posthurricane periods. Long-term trends simulated for varying hurricane frequencies also projected stability in population size. These results supported the resistance hypothesis. Elasticity analysis showed that, while recruitment had a minor influence, survival of medium-sized trees was the most critical character influencing population growth. Similar patterns of life history sensitivity are common among other trees, suggesting that traits related to survival of large understory individuals would often be under high selection pressure. Among these traits, those associated with hurricane resistance could also be advantageous in forests subjected to light to moderate disturbance. Thus, resistance to canopy disruption could have remained linked to shade tolerance during expansions and contractions of the distribution of F. grandifolia.

No abstract

No abstract