The USDA Forest Service, Forest Health Protection, is responsible for maintaining an ORACLE database named the Southern Pine Beetle Information System (SPBIS). SPBIS was designed to store survey data for Southern Pine Beetle (SPB) infestations on federal land across the southern United States. The main purpose of this project is to improve the SPBIS database to aid in management of SPB infestations to reduce losses, to harvest compromised timber while it still has value, and to assist resource managers in preventing further infestations from becoming established. The SPBIS mobile mapping system addresses current problems with the database and offers viable solutions to each. Mobile mapping technology is a versatile tool, which is used to collect field data, with unique geospatial time tags and attributes, for integrating into or updating a GIS (Rasher 2001). The ability to efficiently collect data is essential to developing a useable database. The time required to manually enter data into the database is substantial. Ranger district personnel often do not have the time or the desire to enter data. Currently, the database is lacking years of survey and infestation data due to these negligent data entry practices, limiting the usefulness database. Currently, SPBIS data is recorded on a paper survey form. This system introduces a digital version of this form. A time study conducted to define the efficiency of each survey showed that the digital form to be more efficient with a Mann-Whitney p- value of 0.004. A comparative study was conducted to define the difference between currently estimated SPB infestation acreage and those measured using GPS. A Mann-Whitney pvalue of 0.000 shows the significant difference between the two acreage values. GPS measured acreage proved to be more accurate thus promoting the use of GPS for measuring acreage. Navigation us ing GPS coordinates was successful and will greatly decrease the time required to locate a SPB infestation on the ground. This thesis describes a mobile mapping system designed specifically to remedy the problems associated with SPBIS. This system eliminates the need for manual entry of field-collected data, while improving field data collection in terms of efficiency and accuracy.

Thesis--Virginia Polytechnic Institute and State University. / Vita. Abstract. Includes bibliographical references (leaves 49-53). Also available via the Internet.

The overall goal of this study was to develop a framework for using airborne lidar to derive inputs for the SPB infestation growth model TAMBEETLE. The specific objectives were (1) to estimate individual tree characteristics of XY location, individual bole height (IBH), diameter at breast height (DBH), length of crown (CrHT), and age for use in TAMBEETLE; (2) to estimate individual tree age using lidar-estimated height and site index provided by the United States Department of Agriculture (USDA) Natural Resources Conservation Service (NRCS) Soil Survey Geographic Database (SSURGO); and (3) to compare TAMBEETLE simulation results using field measurements and lidarderived measurements as inputs. Diameter at breast height, individual bole height, and crown length were estimated using lidar with an error for mean measurements at plot level of 0.16cm, 0.19m, and 1.07m, respectively. These errors were within root mean square error (RMSE) for other studies at the study site. Age was estimated using the site index provided by SSURGO and the site index curves created for the study area with an RMSE of 4.8 years for mean plot age. Underestimation of tree height by lidar and error in the site index curve explained 91% of the error in mean plot age. TAMBEETLE was used to compare spot growth between a lidar-derived forest map and a forest map generated by TAMBEETLE, based on sample plot characteristics. The lidar-derived forest performed comparably to the TAMBEETLE generated forest. Using lidar to map forests can provide the large spatial extents of the TAMBEETLE generated forest while maintaining the spatially explicit forest characteristics, which were previously only available through field measurements.

M. S.

SPBMODEL, a legacy southern pine beetle (SPB) simulation model, was translated into a new JavaTM model called Javahog. The Javahog output was verified to be essentially identical to SPBMODEL output by means of standard and paired t-tests. Javahog was placed online and is currently accessible via a servlet. Genetic algorithms (GAs) were applied to the Javahog model. GAs are a type of optimization heuristic that operate as an analog to evolution. GAs "evolve" a very good solution to a complex problem. In this case, GAs were intended to evolve a very good version of SPBMODEL. GAs were applied in part to improve upon the SPBMODEL design, and in part to demonstrate that GAs are effective tools for recalibrating legacy simulation models. Beyond simply recalibrating model parameters, the GA was used to select optimal functional forms for the development rates of each SPB life stage. The GA evolved a model that performed better than SPBMODEL at predicting observed field data, according to a balanced fitness function and according to sums of squared errors. However, from a visual comparison of the output of both models versus observed field data, neither model achieved satisfactory performance. / Master of Science

The southern pine beetle (SPB) (Dendroctonus frontalis Zimm.) is one of the most destructive forest insect pests in the southeastern United States and has historically had a large impact on the forests of North Carolina. Many characteristics of a forest can contribute to SPB susceptibility including stand density, growth rate, age, soil type, and position on the landscape. This work was undertaken in an effort to assist and improve on the current federal SPB hazard modeling being conducted for North Carolina by the USDA Forest Service – Forest Health Protection's Forest Health Technology Enterprise Team (FHTET). In our study, predictive SPB susceptibility models were developed for each physiographic region in North Carolina using two variables not currently included in the FHTET modeling, mean stand age and the in-stand percentage of sawtimber-sized pines. These variables were obtained from USDA Forest Service – Forest Inventory and Analysis (FIA) data and North Carolina Forest Service historical SPB records creating a dataset of both infested and non-infested stands and the models were developed using the CART® classification tree approach. Two model-derived age classes (older than and younger than 22 years) were identified on the landscape using current Landsat 5 Thematic Mapper (TM) imagery chronosequences of disturbance index (DI) â transformed scenes to identify stand-replacing disturbances, resulting in a kappa statistic of 0.6364 for the younger than 22 year age class and 0.7778 for the older than 22 years age class. A kappa value of 1 is ideal. The CART® modeling effort produced valid models in all three physiographic regions of North Carolina, though the complexity of the piedmont model makes it impractical for use in the field. The dependent variable in the classification tree was presence or absence of SPB outbreak and the test sample error percentages were similar across regions, with errors ranging between 23.76 - 34.95 percent. Overall prediction success, based on the software's internal cross-validation procedure, was likewise comparable across the regions with 72.28 - 89.56 percent correctly predicted. Based on our modeling, stand age and percent sawtimber should be included in future FHTET SPB hazard modeling efforts for the coastal plain and mountains, respectively. Age classes can be reasonably estimated using Landsat or other multispectral imagery. / Master of Science

Qualitative models were developed to assess SPB's impact on wildlife, wildfire, and grazing. Some of the relationships in the wildlife model were quantitatively expressed, yet dollar impacts were not obtainable. The results indicated that SPB had a positive impact on woodpeckers, quail, rabbits, deer, small mammals, and other birds. No negative impacts on any of the wildlife species groups were found. These impacts were very small because of SPB's small spot size and low spot incidence. The wildfire impact assessment was strictly qualitative. The results indicated that fire rate of spread and intensity are increased on SPB spots, however, this impact is small because these increases would be restricted to SPB spots which are very small and highly dispersed. Only 64 acres of SPB spots were estimated to occur within burned areas in one year, hence, this impact was too small to consider in SPB control decisions. All of SPB's grazing impacts were qualitatively assessed due to the large amount of information available in this area. A computer program was written which estimated grazing impacts given SPB spot size and incidence, the herbage production before attack, and the residual basal area in hardwoods. The results indicated that SPB grazing impacts were also too small to be considered in SPB control decisions. / M. S.

Master of Science

This study investigates the pheromone preference of a bark beetle predator Thanasimus dubius between latent and intermediate phases of southern pine beetle. Two trap treatments were set up in each SPB phase. Standard Lindgren funnel traps were baited with either SPB lures or Ips lures. The number of T. dubius caught in each trap was recorded and data was analyzed using both the Mann Whitney U test and a two-way factorial ANOVA. Thanasimus dubius showed no variation in SPB pheromone preference but did show a slight increase in preference for Ips pheromones in intermediate phase areas. A protocol was developed to identify prey DNA within gut contents of T. dubius to understand prey preference in relation to pheromone preference. Primers were developed to amplify CO1 gene sequences from five different bark beetles. All primers were specific to their own DNA and able to detect at least 0.2 picograms of DNA.

A three-phase study was performed to examine abiotic and biotic metrics at southern pine beetle infestation sites in northern Georgia in 2002 to find early indicators that can be leveraged by forest managers to mitigate the effects of future outbreaks: creation of a 2003 Final Impact Map, determining if MODIS MOD13Q1 EVI 16-day image composites can distinguish differences in biomass indicators among healthy and infested loblolly pine and hardwood forests, and creation of an Infestation Risk Map derived from significant climate and physical variables at known infestation sites.Three land cover classification techniques (change vector analysis, enhanced wetness differencing index and standard land cover classification analysis of Landsat 5 TM) were compared to determine which would provide the best estimate of final infestation damage. Classification accuracy results indicated that the latter provided the most reliable site damage information and it became the reference map against which outbreak model results were compared.Using time series analysis of MODIS composites acquired March 2000 - December 2006 to measure 11 phenology metrics for infested and healthy loblolly and hardwood stands showed that the imagery differentiated between forest classes. Results indicated the lowest base vegetation biomass in 2001 for infested loblolly, relative to healthy loblolly, with many metrics trending towards hardwood values following infestation.Abiotic influences included those related to landscape position and climate. Statistical testing showed increased beetle success: 1) along ridge tops at maximum solar exposure, 2) in areas with canopy density>60%, 3) in areas experiencing cooler summers and warmer winters, and 4) where precipitation was significantly lower at infested sites in the 2 years preceding outbreak.The Infestation Risk Map was developed from significant physical and climate indicator variables using the fuzzy theory modeling approach. Comparison of model output to infestation sites resulted in Chi-squared and Cramér's V values of 55.4 and 0.16, respectively, indicating that infestation risk distributions strongly paralleled site infestation. Comparison of model output and low, medium and high infestation density clusters resulted in Chi-squared and Cramér's V values of 241.24 and 0.66, respectively, indicating a more substantive relationship between infestation density and risk classes.