Using Remotely Sensed Indices of Fire Severity and Vegetation Structure to Identify Patterns of Avian Occurrence Following Changes in Fire Management Policy within Great Smoky Mountains National Park.

Rose, Eli

Fire management policy has the potential to benefit breeding bird species associated with both early succession and fire adapted plant communities. Many of these species are thought to be declining in the southern Appalachian region. Fire management policies within Great Smoky Mountains National Park changed in 1996, allowing managers to incorporate both prescribed and wild fires to achieve habitat management objectives. Over the past 20 years, an increasing amount of fire on the landscape has changed the biological communities within the Park’s fire adapted pine-oak forests. I use a combination of field-derived and remotely sensed indices of both fire severity and vegetation structure to quantify changes in patterns of bird occurrence following fire within these habitats. LandsatTM derived Difference Normalized Burn Ratios provide an index of fire severity by quantifying changes in photosynthetic biomass from before to after a fire. National Park personnel currently use these indices to identify changes in vegetation within the Park. In collaboration with NPS personnel, I developed a rapid assessment tool to evaluate fire severity and contrast this field-based index with Difference Normalized Burn Ratio (DNBR) indices to identify potential strengths and weaknesses of using DNBR to measure fire effects. I developed species occurrence models using Difference Normalized Burn Ratios summarized at four different spatial scales to explore the possibility that species respond to fire at different spatial scales. Agreement between field derived and Difference Normalized Burn Ratio fire severity indices was lower within the xeric pine-oak forests of the Park than has been shown in other studies. The level of agreement was a function of both the vegetation type and the threshold values used to define fire severity classes. Fire severity influenced patterns of species occurrence all six of the species considered, and my results suggest that different species respond to variation in fire severity at different spatial scales. One way that variation in fire severity can influence patterns of species occurrence is by altering the structural characteristics of the vegetation. In recent years Light Density and Ranging (LIDAR) has been used to provide an index of the vertical distribution of vegetation present on the landscape. In 2011, the National Park Service acquired discrete return, leafoff LIDAR products for the Tennessee side of the Park. I compared field-derived estimates of canopy cover for three strata (shrub, midstory, and overstory) with LIDAR derived indices to explore the utility of incorporating leaf-off LIDAR data in models of bird species occurrence. My results suggest that discrete-return LIDAR products acquired during leaf-off conditions can be used to identify broad scale patterns of breeding bird occurrence, especially for species showing strong associations with the upper canopy of coniferous forests. I modeled occurrence on burned and unburned sites for twenty four bird species to quantify changes associated with the prevalence of fire on the landscape. Overall, thirteen species occurred more frequently on burned sites and two species occurred more frequently on unburned sites....