Effects of overstory structure and fire regime upon diversity and abundance of selected understory species in longleaf pine (Pinus palustris Mill.) forests in southeastern Georgia

We examined the impacts of variation in overstory structure and burning regime on understory vegetation in the longleaf pine (Pinus palustris Mill.) forest of the Okefenokee National Wildlife Refuge in southeastern Georgia, U.S.A. On sandy upland sites surrounding the Okefenokee Swamp and on islands, we established five randomly-located 0.04-ha plots within each of six study areas. The plots varied in overstory density, past management and fire regime. We measured overstory tree size of longleaf pines in the 0.04-ha plots and percent cover of ground cover species in five 1-m2 subplots nested within each 0.04-ha plot. We also calculated diversity indices for the ground cover species. There was no significant relationship between overstory basal area and any diversity measure. Even at the subplot level, there was no significant relationship observed between the diversity measures and distance to the nearest tree, presumably a surrogate for overstory density. The diameter (DBH) of the nearest tree exhibited a slight influence over the Shannon Index, and the nearest tree's height was significantly related to evenness in each subplot. There were varied relationships between individual species and fire regime or nearest tree measurements. There was a definite, but statistically insignificant, trend of decreasing diversity with increasing years since the last burn. Height of the nearest tree positively influen­ced evenness, probably reflecting the influence of stochasticity and amount of needle deposition (influencing fire behavior) over a wide area. Within the generally equal fire-return interval across sites, individual fire-events and site-specific behavior apparently influence understory diversity. It is not surprising that the last burn would affect species richness, even more so than the number of growing-season burns, other studies support this observation. The presence of obligate seeding and (clonal) sprouting as dominants in the ground cover suggests that the fire-return interval is not regular. The cycling of fire-return intervals, even within the manage­ment-prescribed 2- to 4-year range, and the variation by microsite appear to provide sufficient variability of disturbance to create diversity in the ground-level cohort. Site-specific relationships between particular species and the nearest tree suggest that even a narrowly-prescribed fire management regime can provide sufficient diversity.