2009 SOUTHEASTERN NATURALIST 8(1):129–140 Hemlock Susceptibility to Hemlock Woolly Adelgid Attack in the Chattooga River Watershed Mark Faulkenberry1,*, Roy Hedden2, and Joe Culin3 Abstract - Adelges tsugae (Hemlock Woolly Adelgid [HWA]), an introduced pest, is impacting Tsuga canadensis (Eastern Hemlock), and T. caroliniana (Carolina Hemlock) stands throughout the eastern United States. Currently, hemlock stands in the southeast US are on the leading edge of the infestation. This study investigated HWA distributions in the Chattooga River watershed, and examined relationships between site and stand variables and hemlock susceptibility to HWA attack. The following variables were examined: latitude, longitude, elevation, slope, aspect, terrain shape index (TSI), landform index (Lfi), percent infestation, quadratic mean diameter, total basal area (BA), hemlock BA, non-hemlock BA, hemlock BA(%), non-hemlock BA(%), and tree height. Multiple regression with backward selection showed statistically significant relationships of HWA infestation to latitude (P = 0.0006), longitude (P < 0.0001), and TSI (P = 0.0316). The proximity of a hemlock stand to existing HWA infestations appears to be the primary factor infl uencing its susceptibility to attack. Introduction Since Adelges tsugae Annand [Hemlock Woolly Adelgid (HWA)] was accidentally introduced into Virginia from Japan in the 1950s, it has become a serious pest of Tsuga canadensis L. (Eastern Hemlock) and Tsuga caroliniana Engelm (Carolina Hemlock) in the eastern United States (McClure et al. 2001, Orwig and Foster 1998). Carolina Hemlock distribution is limited to the Appalachian Mountains, from southwestern Virginia, south into western North Carolina, northwestern South Carolina, eastern Tennessee, and northeastern Georgia, where it commonly occurs on cliffs and rocky slopes and ridges at elevations greater than 914 m (Caladonato 1993, Humphrey 1989, Little 1975). Eastern Hemlock is more widely distributed, occuring throughout New England, New York, Pennsylvania, the mid-Atlantic states, central New Jersey west to the Appalachian Mountains, and south into Northern Georgia and Alabama (Godman and Lancaster 1990). In the northern and northeastern parts of its range, Eastern Hemlock is commonly found up to 730 m in elevation on benches, fl ats, and swampy borders. In the southern Appalachians, Eastern Hemlock is typically found from 610–1520 m in elevation, on north and east slopes, coves, or cool moist valleys (Godman 1Department of Forestry and Natural Resources, Department of Entomology, Soils and Plant Science, College of Agriculture Forestry and Life Sciences, Clemson University, Clemson, SC 29634. 2Department of Forestry and Natural Resources, Clemson University, Clemson, SC 29634. 3Department of Entomology, Soils and Plant Science, Clemson University, Clemson, SC 29634. *Corresponding author - mfaulke@clemson.edu. 130 Southeastern Naturalist Vol. 8, No. 1 and Lancaster 1990). Both Carolina Hemlock and Eastern Hemlock are slow growing, long lived, and very tolerant of shade, often considered late successional species (Caladonato 1993, Godman and Lancaster 1990, Humphrey 1989). HWA infestations are now found in 17 states in the eastern US, with extensive hemlock mortality being reported in Virginia, Pennsylvania, Connecticut, and New Jersey (Knauer et al. 2002, Orwig and Foster 1998, Skinner et al. 2003, USDA Forest Service 2007). Infestations may cause mortality within four years (McClure et al. 2001). Since 1990, the HWA has been spreading at a mean rate of 12.5 km per year, with rates as high as 15.6 km per year in the southern portion of its range (Evans and Gregoire 2007). Hemlock stands provide habitat for a variety of wildlife and create dense canopies, shading and cooling headwater streams (McClure et al. 2001). In addition, hemlock stands have also been shown to have a powerful infl uence on nutrient cycling (Jenkins et al. 1999, Yorks et al. 2000). HWA-related mortality will have a drastic effect on the composition of habitats currently dominated by hemlock throughout the eastern US, with a shift from conifers to mixed species hardwoods (Jenkins et al. 1999, Kizlinkski et al. 2002, Orwig et al. 1998). The loss of hemlock, and its replacement by hardwood species, is likely to significantly affect stream habitats, lowering invertebrate diversity and altering the trophic structure of fish and invertebrates, as well as nitrogen and other nutrient fl uxes from the canopy to the forest fl oor (Snyder et al. 2005, Stadler et al. 2006). Although much research has been conducted on the relationships of site and stand components to HWA infestations in the northeast, (e.g., Orwig and Foster 1998, Orwig et al. 2002, Royle and Lathrop 2000), few studies have been performed on similar factors in hemlock stands in the southeastern United States. Therefore, the objectives of this study were: 1. to map HWA infestations within the Chattooga River watershed of North Carolina, South Carolina, and Georgia, and 2. to determine the relationship between stand and site characteristics and hemlock susceptibility to HWA attack. Infestation maps from this study may be used to determine the rate of spread of HWA in the Chattooga watershed, as well as the likely pattern of hemlock mortality. Site or stand components found to be correlated with hemlock susceptibility to the HWA could serve as the basis for a risk model to identify high-risk areas for intensive monitoring. Early identification of high-risk areas would also allow resource managers to initiate control efforts before significant damage occurred and optimally allocate resources for HWA management. Methods and Materials Study area and study sites The study was conducted in the Chattooga River watershed, which covers 72,840 ha, in South Carolina, Georgia, and North Carolina. All study sites were located in the Nantahala National Forest in North Carolina, the 2009 M. Faulkenberry, R. Hedden, and J. Culin 131 Chattahoochee National Forest in Georgia, and the Sumter National Forest in South Carolina (Fig. 1). Most of the study sites were established in cool, moist coves and riparian areas, with overstories of Eastern Hemlock, Pinus strobus L. (White Pine), Liriodendron tulipifera L. (Yellow Poplar), mixed mesophytic hardwoods, and understories of Rhododendron maximum L. (Rhododendron), and Kalmia latifolia L. (Mountain Laurel). There were 104 study sites located in the watershed, many of them along existing hiking trails and paths. Trails were chosen at random, and sample plots were inserted at 0.80-km intervals along each trail. Figure 1. Hemlock Woolly Adelgid infestation map for the Chattooga River Watershed in North Carolina, Georgia, and South Carolina. 132 Southeastern Naturalist Vol. 8, No. 1 At each sample plot, a random number between 1 and 12 was generated, corresponding with the numbers on a clock face, with 12 and 6 forming an axis parallel to the trail. Sampling points were established 6 m off of the trail, in the direction of the random number. The random numbers of 6 and 12 were not used in the selection of sample points, to prevent sampling the trails. The following site characteristics were measured for each point: latitude, longitude, elevation above sea level, slope angle, topographic aspect, terrain shape index (TSI), landform index (Lfi), and level of HWA infestation. Variable radius plots were established at each point to measure stand components. A description of each site and stand component measured is given below. Site data Latitude, longitude, elevation. Using a GPS receiver (Garmin™ eTrex®), latitude and longitude and elevation above sea level were determined. The slope angle and topographic aspect were measured for each point. Aspect values were transformed for use in statistical analysis (Beers et al. 1966, Trimble and Weitzman 1956). The aspect transformation formula is: A = sin (A1 + 45) + 1, with “A” equaling the transformed aspect, and “A1” equaling the aspect originally recorded in azimuth degrees. With this transformation, a minimum value of zero is equivalent to southwest (generally more xeric) and a maximum value of 2 is equivalent to northeast (generally more mesic). Topography and landform. Terrain shape index (TSI) is a quantifiable measurement of the shape of a sample plot (McNab 1989). Using a handheld clinometer, eight percent gradients were recorded from the center of the plot starting in the direction of the aspect. At every 45 degrees, a percent gradient was measured at the observer’s eye level, targeting wherever the surface changed shape to plot perimeter. The eight slope percents were then summed and divided by 800. Negative TSIs correspond to convex plots, while positive TSIs indicate concave plots. Landform index (Lfi) is a quantifiable measurement of the position of a sample plot on the landscape (McNab 1993). Using a handheld clinometer, eight percent gradients were recorded from the center of the plot starting in the direction of the aspect. At every 45 degrees, a percent gradient was measured, targeting the area where the landscape and the horizon made contact. High Lfivalues are indicative of a concave landform (e.g., cove), and low Lfivalues are indicative of a convex landform (e.g., ridge). Level of infestation. Starting at plot center, an 8-m radius (0.08-ha) circular plot was measured. Visual surveys were done on five random hemlocks distributed throughout the plot, with attempts to measure trees of varying size classes within each site. On each of the five survey trees, four branches were examined, inspecting one branch for every side of the tree. Attempts were made to measure branches throughout the crowns of the survey trees. However, due to the difficulty of examining upper crowns 2009 M. Faulkenberry, R. Hedden, and J. Culin 133 of hemlocks over 9 m in height, primarily middle to lower crown branches were examined. For each branch that was examined, the presence or absence of HWA was recorded. The percentage of the 20 branches (5 trees, 4 branches each) with the HWA present, was recorded as the percent infestation for the sample plot. Stand data Bitterlich’s variable radius plot method was performed at each sample point, using a wedge prism with a basal area factor (BAF) of 2 m2/ha to estimate stand basal area. The diameter at breast height (1.37 m, DBH) was measured for each tree in the 0.08-ha circular plots, and was used to determine the basal area (BA) for each in-plot tree, as well as the mean BA per plot. The mean BA per tree was used to calculate the quadratic mean diameter (QMD) using the formula: QMD = where QMD is a measure of the average tree diameter in the plot, and is the typical measurement of mean diameter used in forestry (Curtis and Marshall 2000). The quadratic mean is different than the arithmetic mean commonly encountered in statistical analyses, although the differences between mean diameters calculated from both methods usually do not differ greatly (Curtis and Marshall 2000). Basal area per hectare was calculated separately for hemlock and non-hemlock. These values were also used to calculate the percentage of total basal area comprised of hemlock or nonhemlock trees in each stand. Finally, the height was measured for two of the tallest trees within each 0.08-ha circular plot. Tree heights were averaged as a measure of mean tree height within each plot. Statistical analysis Multiple regression analysis was used to identify significant relationships between infestation level and the site and stand variables measured (Ott and Longnecker 2001). In order to meet normality assumptions, percent infestation was arcsin transformed (AINF) as follows: AINF = ARSIN(SQRT(percent infestation / 100)), where AINF is represented in radian units. Backward selection was first used to determine if any site or stand variables were significant in explaining HWA infestation (AINF). Multiple regression analysis was then repeated using all significant site and stand variables. All statistical analyses were performed with SAS statistical software (SAS Institute Inc. 2002). Infestation map An HWA infestation map of the study site was created using Arc Map 9.0 GIS software (ESRI Inc. 2004) and ERDAS IMAGINE 8.7 GIS software (Leica Geosystems 2004). Hillshade and hydrography layers were used to construct the map, in addition to a layer describing the latitude, longitude, and infestation level for each sample point. Infestation levels were divided √ (BA / 0.000007854), 134 Southeastern Naturalist Vol. 8, No. 1 into 4 discrete categories of infestation (Table 1). The hillshade layers were obtained through the South Carolina Department of Natural Resources website (www.dnr.sc.gov), and the hydrography layers were purchased through www.mapmart.com. Results and Discussion Regression of site factor and stand component data Multiple regression analysis, using backward selection on all site and stand variables resulted in a model of latitude, longitude, and TSI that significantly explained percent HWA infestation, (t = 12.47, P = 0.0006), (t = 20.65, P < 0.0001), (t = 4.75, P = 0.0316), respectively (Figs. 2–4). The model was highly significant (F = 31.85, P < 0.001), explaining almost 50% of the variation in the data (R2 = 0.48; Table 2). Of these variables, TSI was Table 1. Criteria for assigning the level of infestation for Hemlock Woolly Adelgid in the Chattooga watershed. Percent of branches infested Infestation level 0% None 1–50% Light 51–80% Medium 81–100% Heavy Figure 2. Correlation of Hemlock Woolly Adelgid infestation with longitude for the Chattooga River watershed (n = 104; ∝ = 0.05). 2009 M. Faulkenberry, R. Hedden, and J. Culin 135 Figure 3. Correlation of Hemlock Woolly Adelgid infestation with latitude for the Chattooga River watershed (n = 104; ∝ = 0.05). Figure 4. Correlation of Hemlock Woolly Adelgid infestation with terrain shape index for the Chattooga River watershed (n = 104; ∝ = 0.05). 136 Southeastern Naturalist Vol. 8, No. 1 the least significant variable in the model, and the only variable that was not singularly correlated with infestation (Table 3). A positive correlation of infestation with elevation was observed (Table 3). All site and stand variables which were eliminated by backward selection and their accompanying F and P values are listed in (Table 2). Infestation map of the Chattooga watershed Infestation level declined from northeast to southwest through the Chattooga watershed. Of 104 sample sites in the watershed, 51 were heavily infested, 16 were moderately infested, 24 were lightly infested, and 13 were not infested (Fig. 1). The northern portion of the watershed, which is also the most heavily infested, was likely the first area where HWA infestations were Table 2. All site and stand variables eliminated from the model using the backward selection method. All variables appear in the order in which they were removed from the model, and the R-square for the model after each variable was removed is given in column four. The test statistic for each variable is given in column two, and the corresponding P values for each variable is given in column three (n = 104). Variables remaining in final regression model: latitude, longitude, TSI. Final R-square = 0.48 Abbreviations are: Lfi= landform index, BA= basal area. Variable F value Pr > F R-square Mean tree height (m) 0.00 0.9636 0.5475 Lfi0.19 0.6635 0.5465 Elevation (m) 0.60 0.4415 0.5436 Quadratic mean diameter (cm) 0.68 0.4116 0.5402 Slope 1.36 0.2469 0.5336 % BA Hemlock 1.56 0.2148 0.5259 % BA non-hemlock 1.29 0.2594 0.5336 BA non-hemlock (m2/ha) 1.32 0.2528 0.2528 Basal area (BA) (m2/ha) 0.93 0.3384 0.5225 Aspect 1.78 0.1851 0.5074 BA hemlock (m2/ha) 3.78 0.0547 0.4886 Table 3. Correlations of all site and stand variables measured in the Chattooga Watershed to hemlock woolly adelgid infestation (n = 104). All correlations appear as correlation coefficients (r), and variables are listed in ascending order to their correlation with infestation. Variable Correlation (r) Longitude -0.632 Quadratic mean diameter (cm) -0.172 Mean tree height (m) -0.127 % BA non-hemlock -0.032 % BA hemlock 0.028 BA hemlock (m2/ha) 0.029 BA non-hemlock (m2/ha) 0.050 Lfi0.054 Basal area (BA) (m2/ha) 0.062 Aspect 0.079 TSI 0.095 Slope 0.169 Latitude 0.616 Elevation (m) 0.631 2009 M. Faulkenberry, R. Hedden, and J. Culin 137 encountered in the watershed as the adelgid moved south and southwest from its initial introduction point in Virginia. The Forest Service reported the presence of the HWA in Macon and Jackson counties (North Carolina), and Oconee County (South Carolina) in 2001 (USFS 2007). Starting in the northeast portion of the infestation map, the HWA traveled approximately 20 km in the Chattooga watershed from 2001–2004, nearly 40% of the length of the watershed. Although wind is one of the primary vectors of the HWA, the movement of the infestation front through the Chattooga watershed is not consistent with the prevailing winds in the area (Koch et al. 2006). The mean prevailing wind directions for two nearby weather monitoring stations in Asheville, NC, and Athens, GA from 1930–1996 were from north-northwest, and west-northwest, respectively (National Climatic Data Center 2007). Site and stand components Of all the stand and site components measured in this study, only longitude, latitude, and TSI significantly predicted levels of HWA infestation. These results suggest that HWA may have the potential to spread throughout the range of hemlock in the southeastern United States, regardless of any particular site or stand component. Latitude and longitude were the only significant predictors of percent infestation, with the trend running northeast to southwest. Since this is an artifact of how the infestation front is moving, the only predictor of susceptibility to HWA attack is the proximity of the hemlock stand to an infested area. HWA is expanding throughout the range of hemlock, and we can expect that all hemlock stands in the southernmost portion of the range, which includes the Chattooga river watershed, will eventually be colonized. A similar trend in infestation levels was observed as HWA moved north from Virgina to Connecticut (Orwig et al. 2002). In a study in Connecticut, site and stand variables were also found to have little effect on hemlock susceptibility to HWA attack, and latitude had the strongest correlation with HWA infestation and hemlock mortality (Orwig and Foster 1998, Orwig et al. 2002). Research by Koch et al. (2006) contradicts the suggestion that all hemlock stands are equally susceptible to the HWA, reporting that the distance of a hemlock stand from the closest road, trail, and stream all infl uence where the HWA is more likely to first appear in the landscape. Streams, roads, and trails create corridors that make hemlock stands more accessible to humans, birds, and wind, three main vectors for the insect. Unlike our research, Koch et al. (2006) found elevation and slope to be significant predictors of the susceptibility of a site to HWA infestation, suggesting that steeper slopes and higher elevations made a site less accessible to vectors of the adelgid. Although elevation was significantly correlated with infestation in our data, it was not a significant predictor of infestation. It should be noted that while TSI was not significantly correlated with infestation (Table 3), it was a significant predictor of infestation in our model. Although TSI was significant in the model, it was the least significant variable of the three, and was only weakly related to infestation level, explaining 3% of the variation in the model. It is unclear why TSI was significant in the model, since the mean 138 Southeastern Naturalist Vol. 8, No. 1 TSI for all plots was -0.003, meaning the plots were almost completely fl at. If the mean terrain shape were more convex or concave, this would likely affect the quality of a site, infl uencing water movement and other factors (Mc Nabb 1989). Small et al. (2005) also found elevation to be a significant predictor of hemlock mortality due to the HWA, yet reported that ledges had a 15.6% greater decline in hemlock basal area than ravines. The significance of elevation to HWA-related hemlock mortality in this study was more closely correlated with the quality of the site in relation to water and other factors, than the accessibility of the site to vectors, as reported by Koch et al. (2006). Due to the fact that many of our sample points were along streams, hiking trails, and paths, this may have led to some inaccuracies regarding the levels of HWA infestation for the study. Since research by Koch et al. (2006) showed that the proximity of a hemlock stand to the nearest road, trail, or stream has a considerable infl uence on where HWA will first appear in a landscape, it is possible that we overestimated the infestation levels for our study site. The decision to use hiking trails and paths was made in order to facilitate navigation through the study area and to increase to amount of sample points that could be completed. Conclusions The objectives of this study were to map HWA infestations within the Chattooga River watershed and to determine the relationship between stand and site characteristics and hemlock susceptibility to HWA attack. A map of HWA infestations within the Chattooga watershed was created, and there appears to be no relationship between the stand and site variables measured and susceptibility to HWA infestation. The only variables of consequence were the latitude and longitude of the study sites, which showed that HWA is spreading south and west through the Chattooga watershed. In the southeastern United States, our research suggests that all hemlock stands appear to be equally susceptible to HWA attack. Acknowledgments We thank Vic Shelburne (Clemson University), Jim Sullivan (Georgia Forestry Commission), Rusty Rhea (US Forest Service), and Buzz Williams (Chattooga Conservancy) for valuable assistance. This project was funded by the US Forest Service, National Forest Foundation, Chattooga Conservancy, and the Jackson Macon County Alliance (JMCA). We also thank Thorlos, Kelty, Patagonia, Cascade Designs, Woolrich, and Wrangler for supplying research gear. This report is publication number 5132 of the South Carolina Agricultural and Forestry Research System of Clemson University. Literature Cited Beers, T.W., Dress, P.E., and L.C. Wensel. 1966. Aspect transformation in site productivity research. Journal of Forestry 64:691–692. Caldadonato, M. 1993. Tsuga caroliniana: Fire Effects Information System, Available online at http://www.fs.fed.us/database/feis. Accessed August 15, 2007. US Department of Agriculture, Forest Service, Rocky Mountain Research Station, Fire Sciences Laboratory, Missoula, MT. 2009 M. Faulkenberry, R. Hedden, and J. Culin 139 Curtis, R., and D. Marshall. 2000. Why quadratic mean diameter. Western Journal of Applied Forestry 3:137–139. ESRI, Inc. 2004. ArcMap 9.0. ESRI Inc., Redlands, CA. Evans, A.M., and T.G. Gregoire. 2007. A geographically variable model of Hemlock Woolly Adelgid spread. Biological Invasions 9:369–382. Godman, G.A., and K. Lancaster. 1990. Tsuga canadensis (L.) Carr. Eastern Hemlock. Pp. 604–612, In R.M.Burns and B.H. Honkala (Technical Coordinators). Silvics of North America. Volume 1. Conifers. Agric. Handb. 654. US Department of Agriculture, Forest Service, Washington, DC. Humphrey, L.D. 1989. Life-history traits of Tsuga caroliniana Engelm. (Carolina Hemlock) and its role in community dynamics. Castanea 54:172–190. Jenkins, J.C., J.D. Aber, and C.D. Canham. 1999. Hemlock Woolly Adelgid impacts on community structure and N-cycling rates in Eastern Hemlock forests. Canadian Journal of Forest Research 29:630–645. Kizlinski, M.L., D.A. Orwig, R.C. Cobb, and D.R. Foster. 2002. Direct and indirect ecosystem consequences of an invasive pest on forests dominated by Eastern Hemlock. Journal of Biogeography 29:1489–1503. Knauer, K., J. Linnane, K. Shields, and R. Bridges. 2002. An initiative for management of Hemlock Woolly Adelgid. Pp. 9–12, In Proceedings: Symposium on the Hemlock Woolly Adelgid In Eastern North America. B. Onken, R. Reardon, and J. Lashomb (Eds.). East Brunswick, NJ, February 5–7, 2002. USDA Forest Service, Rutgers, NJ. 403 pp. Koch F.H., H.M. Cheshire, and H.A. Devine. 2006. Landscape-scale prediction of Hemlock Woolly Adelgid, Adelges tsugae (Homoptera: Adelgidae), infestation in the Southern Appalachian mountains. Environmental Entomology 35:1313–1323. Leica Geosystems, Erdas Imagine 8.7 (2004). Leica Geosystems AG, St. Gallen, Switzerland. Little, E.L., Jr. 1975. Rare and local conifers in the United States. Conservation Research Report No. 19. US Department of Agriculture, Forest Service, Washington, DC. 25 pp. McClure, M.S., S.M. Salom, and K.S. Shields. 2001. Hemlock Woolly Adelgid. USDA Forest Service, Morgantown, WV. FHTET-2001-03. 14 pp. McNab, W.H. 1989. Terrain shape index: Quantifying effects of minor landforms on tree height. Forest Science 35:91–103. McNab, W.H. 1993. A topographic index to quantify the effect of meso-scale landform on site productivity. Canadian Journal of Forest Research 23:1100–1107. National Climatic Data Center (NCDC) 1997. Climatic wind data for the United States. Available online at http://www5.ncdc.noaa.gov/documentlibrary/pdf/ wind1996.pdf. Accessed December 17, 2007. Orwig, D.A., and D.R. Foster, 1998. Forest response to the introduced Hemlock Woolly Adelgid in southern New England, USA. Journal of the Torrey Botanical Society 125:60–73. Orwig, D.A., Foster, D.R., and D.L. Mausel. 2002. Landscape patterns of hemlock decline in New England due to the introduced Hemlock Woolly Adelgid. Journal of Biogeography 29:1475–1487. Ott, L., and M. Longnecker. 2001. An Introduction To Statistical Methods and Data Analysis. Duxbury Press, Belmont, CA. 1152 pp. 140 Southeastern Naturalist Vol. 8, No. 1 Royle, D., and R. Lathrop. 2000. The effects of site factors on the rate of hemlock decline: A case study in New Jersey. P. 103, In K.A. McManus, K.S. Shields, and D.R. Souto. (Eds.). Proceedings: Symposium on Sustainable Management of Hemlock Ecosystems in Eastern North America. Durham, New Hampshire. June 22–24, 1999. USDA Forest Service, Newtown Square, PA. SAS Institute Inc. 2002.Version 9.0. SAS Institute Inc. Cary, NC. Skinner, M., B. Parker, S. Gouli, and T. Ashikaga. 2003. Regional responses of Hemlock Woolly Adelgid (Homoptera: Adelgidae) to low temperatures. Environmental Entomologist 32:523–528. Small M.J., Small C.J., and G.D. Dreyer. 2005. Changes in a hemlock-dominated forest following woolly adelgid infestation in southern New England. Journal of the Torrey Botanical Society 132:458–470. Snyder, C.D, Young, J.A., Ross, R.M., and D.R. Smith. 2005. Long-term effects of hemlock forest decline on headwater stream communities. Pp. 42–55, In B. Onken and R. Reardon (Eds.). Proceedings: Third Symposium On Hemlock Woolly Adelgid in the United States. Asheville, North Carolina. February 1–3, 2005. 384 pp. USDA Forest Service, Morgantown, WV. South Carolina Department of Natural Resources (SC DNR). 2004. SC DNR Home Page. Available online at http://www.dnr.sc.gov. Accessed July 5, 2004. Stadler, B., T. Muller, and D. Orwig. 2006. The ecology of energy and nutrient fl uxes in hemlock forests invaded by Hemlock Woolly Adelgid. Ecology 87:1792–1804. Trimble, G.R., and S. Weitzman, 1956. Site-index studies of upland oaks in the northern Appalachians. Forest Science 2:162–173. USDA Forest Service (USFS). 2007. List of states with known Hemlock Woolly Adelgid infestations. Available online at http://www.fs.fed.us/na/morgantown/ fhp/hwa/infestations.htm. Accessed November 12, 2007. Yorks, T.E., J.C. Jenkins, D.J. Leopold, D.J. Raynal, and D.A. Orwig. 2000. Infl uences of Eastern Hemlock mortality on nutrient cycling. Pp. 126–133, In K.A. McManus, K.S. Shields, and D.R. Souto (Eds.). Proceedings: Symposium on Sustainable Management of Hemlock Ecosystems in Eastern North America. Durham, New Hampshire. June 22–24, 1999. USDA Forest Service, Newtown Square, PA.