Longleaf pine and oak responses to hardwood reduction techniques in fire-suppressed sandhills in northwest Florida

Restoring fire-suppressed longleaf pine (Pinus palustris Mill.) sandhill communities often includes reducing hardwood structure before re-establishing maintenance fire regimes. Using a randomized complete block design, we compared the effects of three hardwood reduction techniques (spring burning, application of the ULW (R) form of the herbicide hexazinone. and midstory chainsaw felling/girdling) and a no-treatment control on oak and longleaf pine densities in fire-suppressed sandhills at Eglin Air Force Base. FL. Treatments were applied in the spring and summer of 1995. Felling/girdling and herbicide plots were also burned for fuel reduction from March to April in 1997. Frequently burned, high-quality sandhill plots were sampled to establish reference conditions. Pre-treatment diameter distributions of oaks followed a negative-exponential curve in all treatments, but were Aat with low tree densities in reference plots. Oak densities were significantly reduced in the herbicide and felling/girdling plots in 1995. Compared to the controls, growing season fire topkilled up to 20% more hardwoods among smaller trees in 1995, but this value increased to approximately 50% after 1996. In all pears, the greatest reduction of oak juvenile density (<1.4 m high) was caused by herbicide application. Control plots contained significantly fewer oak juveniles than the burn and felling/girdling plots. Reference plots contained the lowest and most variable oak juvenile densities. Size distributions of longleaf pine across all plots were bimodal with modes at 0-4.9 and 25-29.9 cm in diameter. The highest mode was at 0-4.9 cm in treatment plots and at 25-29.9 cm in reference plots, Only fire quantitatively changed the distributions by the attrition of the smallest trees >1.4 m high in all years. Fire caused approximately 50% decreases in longleaf pine juvenile (<1.4 m high) density in 1995 and 1997. By 1997, median juvenile densities converged to 5-6 stems/200 m(2) in all treatments, including the control. Juvenile densities were slightly higher and more variable in reference plots than in treatments. In 1997, fuel reduction burns in the herbicide and felling/girdling plots decreased densities of recently germinated longleaf pines to <5 seedlings/20 m(2), a 90% decrease compared to 1996 densities. Seedling densities dropped by approximately 50% in control and burn plots, although these sites received no manipulations after 1995. Seedling densities only decreased by 22% in reference plots (205 seldlings/20m(2) in 1996), which did experience some fires, (C) 2001 Elsevier science B.V. All rights reserved.