Building a better burn: Fuel characterization and fire effects in silvicultural systems in the southern United States

Abstract

Landscape-scale fire suppression and poor timber harvesting practices with little regard for post-harvest conditions have contributed to a decline in xerophytic pine (Pinus spp.) and oak (Quercus spp.) forests in the eastern United States and in the Appalachian Mountain region. The success of restoration efforts involving prescribed fire is heavily influenced by the arrangement and composition of wildland fuels, which influence wildland fire behavior and effects. Fire regime alterations influence shifts in stand dynamics via multiple processes, including tree mortality and or regeneration. Over time, these alterations to stand composition and structure further modulate the characteristics of surface fuels. Therefore, understanding how the interactions between fire regimes, stand composition and structure, and wildland fuel complexes influence fire effects, such as site productivity and habitat quality, is critical to meeting diverse economic and ecological silvicultural objectives. In Chapter 2, we evaluated the effect of pine species dominance on soil fertility under a high-frequency fire regime on the Coastal Plain of southeastern Virginia. We compared three frequently burned stands dominated by 1) longleaf pine (P. palustris Mill.), 2) loblolly pine (P. taeda L.), or 3) pond pine (P. serotina Michx.), and one infrequently burned mixed pine-hardwood stand. We found that dominance by longleaf pine was associated with elevated relative abundance of carbon in the O horizon and mineral soil. For forest managers, fuel loading estimates are useful for predicting fire behavior and effects, such as smoke emissions and fuel consumption, which can help determine optimal timing and locations for prescribed fire. Managers can also use a variety of software products to model fire effects based on fuel parameters. However, conventional methods for sampling fuel loads are typically time-consuming, expensive, and vulnerable to observer bias. In Chapter 3, we compared indirect and destructive loadings of litter, duff, and fine woody material (1-hour and 10-hour fuels) for six sequential prescribed burns on the Bankhead National Forest in northcentral Alabama, US. We found that indirect sampling underestimated loading of 1-hour fuels, litter, and duff, but overestimated 10-hour loading. We also tested a free-to-use, decision-support tool, the First Order Fire Effects Model, which consistently overpredicted fuel consumption in our system, regardless of sampling methodology. Finally, in Chapter 4, we present a series of empirical models for downed woody material for upland pine-hardwood forests in the Ridge-and-Valley Province of Virginia and West Virginia. We found that terrestrial LiDAR data can effectively model woody fuels, especially 10-hour fuels, depending on fuel structure.