Southeastern Naturalist A.K. Steketee, P. Bohall Wood, and I.D. Gregg 2015 Vol. 14, Special Issue 7 331 Canaan Valley & Environs 2015 Southeastern Naturalist 14(Special Issue 7):331–343 American Woodcock Habitat Changes in Canaan Valley and Environs Ann K. Steketee1,2,*, Petra Bohall Wood3, and Ian D. Gregg4,5 Abstract - Canaan Valley contains important habitat for Scolopax minor (American Woodcock) in the mid-Atlantic states, especially in West Virginia. Throughout the eastern United States, however, this species has experienced significant population declines since the US Fish and Wildlife Service began monitoring its populations in 1968. Losses of early successional habitats through urbanization and forest-stand maturation have been identified as probable causes for the decrease in population. During 1995–1997, we sampled American Woodcock presence and measured microhabitat and landscape characteristics in a variety of early successional habitats on plots located in and around Canaan Valley. Habitat characteristics related to soil moisture differentiated sites in and outside of Canaan Valley. Sites used by American Woodcock in Canaan Valley generally occurred in or near shrubby wetlands. To identify long-term changes in quality and quantity of American Woodcock habitat, we also compared current availability of appropriate habitat to similar data collected by researchers in the 1970s. We found that almost all of the sites in Canaan Valley that were originally classified as exceptional were still good American Woodcock habitat. Land development and succession, however, have reduced the quality of habitat. Active management and protection are needed, particularly because Canaan Valley is the only place in West Virginia that consistently offers exceptional American Woodcock habitat. Introduction Canaan Valley (hereafter, the Valley) is about 14.9 miles (24 km) long, 3 miles (4.8 km) wide, and over 3100 ft (945 m ) elevation. It is the highest valley of its size east of the Rocky Mountains. The Valley’s climate and habitats are typical of areas much farther north, resulting in the occurrence of plants and animals that are unusual for this latitude. Many of the Valley’s species, including Scolopax minor Gmelin (American Woodcock, hereafter Woodcock), are at or near the southernmost edge of their ranges. Because West Virginia is on the southern edge of the Woodcock’s major breeding range and the rugged terrain with steep slopes limits suitable habitat, statewide populations have always been low compared 1West Virginia Cooperative Fish and Wildlife Research Unit and West Virginia University, Division of Forestry and Natural Resources, PO Box 6125, Morgantown, WV 26506. 2Current address - USDA Forest Service, 180 Canfield St., Morgantown, WV 26505. 3US Geological Survey, West Virginia Cooperative Fish and Wildlife Research Unit, and West Virginia University, PO Box 6125, Morgantown, WV 26506. 4West Virginia University, Division of Forestry and Natural Resources, PO Box 6125, Morgantown, WV 26506. 5Current address - Pennsylvania Game Commission, Bureau of Wildlife Management, Waterfowl and Migratory Game Bird Section, HC 1, Box 144-K, Hawley, PA 18428. *Corresponding author - asteketee@fs.fed.us. Southeastern Naturalist A.K. Steketee, P. Bohall Wood, and I.D. Gregg 2015 Vol. 14, Special Issue 7 332 with those in other states within its breeding range (Kletzly 1976). The Valley, however, supports the largest freshwater wetland (9.4 mi2 [~2438 ha]; USFWS 1979) throughout the central and southern Appalachians. With 40 distinct plant communities (USFWS 1994) including swamp forests, Alnus incana (L.) Moench (Gary Alder) thickets, marshes and bogs, the Valley contains habitat suitable for Woodcock. Consequently, the Valley has been nationally recognized as an important breeding and fall migration area for Woodcock, and protecting Woodcock habitat was a primary objective in establishing the Canaan Valley National Wildlife Refuge (USFWS 1979). Before the 1970s, there were no reliable estimates of the extent of Woodcock habitat in West Virginia, although it was thought to be limited. In the early 1970s, the US Fish and Wildlife Service (USFWS), West Virginia Division of Natural Resources (WVDNR), and West Virginia University inventoried and categorized diurnal Woodcock habitat in the state (Fenwood 1976, Webb 1978). Habitat was defined as having a dense, low, shrubby canopy often associated with a meandering stream; alders were generally the most important shrubs present (Webb 1978, Webb and Samuel 1982). Using aerial photographs and site visits, Fenwood (1976) and Webb (1978) identified and categorized approximately 42,000 ac (17,000 ha) of known and potential Woodcock habitat. The state’s most extensive and best habitat was in the Valley, although suitable areas were found elsewhere statewide. Ten sites comprising 7694 acres (3114 ha) were rated as exceptional habitat and all of them were located in The Valley. Trend analysis from the USFWS’s North American singing-ground survey (SGS) indicated that significant declines in Woodcock populations occurred during 1968–2001 in West Virginia and throughout the entire eastern region (Kelley 2001). This widespread decline was attributed to direct habitat losses from urbanization and industrialization, and indirect losses through natural succession to mature forest stands that are unsuitable for Woodcock (Dwyer et al. 1983, Gutzwiller et al. 1983). The apparent decline of suitable habitat in West Virginia may have been due, in part, to land development (Fenwood and Webb 1981, Kletzly 1976, Webb and Samuel 1982). Along SGS routes in West Virginia, Woodcock counts decreased as the amount of mature deciduous forest increased (Steketee et al. 2000), suggesting that forest maturation may have been an important factor in the population decline. Our study had two objectives: 1) to compare the quantity and quality of Woodcock habitat in the Valley during the mid-1970s with what was available in the mid-1990s and 2) to describe microhabitat and landscape characteristics for the species. We also discuss Woodcock population status and habitat in relation to the Valley’s ecology. Study Area The study area included most of the Valley and selected sites in the nearby counties of Preston, Randolph, Pendleton, Grant, and Tucker (Fig. 1). We Southeastern Naturalist A.K. Steketee, P. Bohall Wood, and I.D. Gregg 2015 Vol. 14, Special Issue 7 333 confined our searches to public lands and to private lands for which we obtained written permission for access. We searched areas included in the 1970s studies, areas that seemed to be likely Woodcock habitat, and areas suggested by local hunters and WVDNR personnel. Figure 1. Woodcock-flush points in and around Canaan Valley, WV. Some flush points are close together, therefore symbols overlap. Southeastern Naturalist A.K. Steketee, P. Bohall Wood, and I.D. Gregg 2015 Vol. 14, Special Issue 7 334 Methods Summary of the 1970s studies Fenwood (1976) and Webb (1978) initially located potential habitat by examining stereo pairs of 9.6 x 9.6-in (23 x 23-cm), black and white, 1:20,000-scale aerial photographs to identify lowland shrubby areas and other suitable sites. These researchers included only sites on level or near level ground because they considered shrubby habitats on steep slopes to be too dry for Woodcock. Fenwood (1976) and Webb (1978) determined approximate site area with a dot grid, and they transferred the site outlines to the appropriate 7.5' or 15' US Geological Survey (USGS) topographical map. They field-checked approximately half of the sites for Woodcock signs, including presence of probings, splashes, and flushed birds. Fenwood and Webb used their knowledge of the species to place field-checked sites into one of four subjective quality categories—poor, fair, good, exceptional—based on presence of Woodcock or their sign, site size, percent ground cover, percent shrub density, soil moisture, land use, and reported use. Using Fenwood’s (1976) and Webb’s (1978) data, we gave all sites not fieldchecked a qualitative rating of questionable, possible, or probable Woodcock habitat based on our interpretation of the aerial photographs. These categories roughly corresponded to the poor, fair, and good ratings assigned to field-checked sites (Fenwood 1976). We identified 1261 potential woodcock habitat sites (~42,000 ac [~17,000 ha]) statewide. Of the 659 sites (30,394 ac [12,300 ha]) that were field-checked, we rated only 124 (14,688 ac [5944 ha]) as good or exceptional. This represented 19% of the total number of sites and 48% of the area surveyed. All exceptional sites (10 sites, 7695 ac [3114 ha]) were located in the Valley. We identified the remaining 602 sites (11,438 ac [4629 ha]) on aerial photographs but we did not field-check them. We classified only 160 sites (27%), containing 3207 ac (1298 ha) or 28% of the total acreage, as possible or probable Woodcock habitat. 1990s data collection We initially located potential habitat in and around the Valley by using a combination of expert opinion, resurveying sites from the 1970s survey, and by examining aerial photographs. We field-checked sites identified as potential habitat for the presence of Woodcock and their sign. We used a variety of methods to determine Woodcock use, including searches with trained pointing or flushing dogs. We used a global positioning system (GPS) to record the exact locations of the Woodcock when dogs discovered them, and we entered the coordinates into an Arc/Info geographic information system (ESRI 1996). When we found a Woodcock, we collected non-scale-dependent habitat data, including vegetation type, percent canopy closure, land use, and soil characteristics, at the flush point and in the area immediately surrounding it (Steketee 2000). We recorded characteristics considered to be important for Woodcock: height of herbaceous vegetation; percent of herbaceous ground cover and litter cover; canopy and shrub cover; shrub-stem density; and soil moisture, texture, and compaction Southeastern Naturalist A.K. Steketee, P. Bohall Wood, and I.D. Gregg 2015 Vol. 14, Special Issue 7 335 (Gutzwiller et al. 1983, McCoy 1987). In general, we used the methods described by James and Shugart (1970) with slight modifications (Steketee 2000). GIS analysis Using ArcGIS software, we established two circular buffers with radii of 492 and 3937 ft (150 and 1200 m) around each flush point. Because of the 98 x 98- ft (30 x 30-m) resolution of Landsat TM data, we chose 492 ft (150 m) as the smallest practical buffer distance. We chose the greater distance because research suggests that Woodcock make daily movements of up to 0.6 mi (1 km; Sepik and Derleth 1993), and therefore, may be affected by and respond to habitat composition at this scale. We used the Multi-Resolution Land Characteristic (MRLC) database (Vogelmann et al. 1996) and FRAGSTATS spatial analysis program (McGarigal and Marks 1995) to quantify habitat variables within each point’s circular buffer. The MRLC is a land-cover database for the entire mid-Atlantic region derived primarily from the Land Remote Sensing Satellite (Landsat) Thematic Mapper data. Because of its 98 x 98-ft (30 x 30-m) pixel resolution, the MRLC cannot be used to identify long, narrow tracts like utility rights-of-way, secondary roads, lowerorder rivers and streams, and all features smaller than 98 x 98 ft (30 x 30 m). We calculated four landscape-level FRAGSTATS metrics: Shannon’s diversity index (SDI), patch richness (PR), area-weighted mean patch fractaldimension (AWMPFD), and mean shape index (MSI). We defined landscape as the area within the circular buffer around each flush point, and the landscape was composed of patches of the 12 MRLC land-cover categories. FRAGSTATS landscape metrics quantify landscape composition or configuration. Landscape configuration refers to the physical distribution of patches within the landscape. Composition, which refers to the presence and amount of each patch type, is not spatially explicit. Landscape composition was quantified with two landscape-level metrics: SDI and PR. SDI represents the probability that any two patches picked at random will be different types. Thus, higher SDI values indicate greater diversity. Because it is a probability, SDI places more weight on common patch types; a single patch of a rare type contributes relatively little to the diversity index. PR, which is sensitive to rare patch types, is simply the number of different land-cover types present within a buffer. It provides no information on the number or size of patches. Landscape configuration was quantified with MSI and AWMPFD. MSI is a measure of patch-shape complexity. When a patch is square, i.e., the simplest patch shape, the MSI equals 1. MSI becomes larger with increasing patch-shape complexity. AWMPFD, another measure of patch-shape complexity, approaches a value of 1 for shapes with simple perimeters and approaches 2 as shapes become more complex (McGarigal and Marks 1995). Because larger patches tend to have more complex shapes than smaller patches, AWMPFD adjusts for patch area and determines patch-shape complexity independent of size. Southeastern Naturalist A.K. Steketee, P. Bohall Wood, and I.D. Gregg 2015 Vol. 14, Special Issue 7 336 Results During May–August 1995–1997, we identified 40 Woodcock-use points in the Valley and 32 in surrounding areas (Fig. 1). Most of the Valley’s Woodcock-use points (27 of 40) occurred in alder thickets, but outsidethe Valley only eight of 32 use points were in alder thickets. 1970s vs. 1990s habitat suitability We rated the 1970s sites as a whole based on current conditions, although the quality of each site may not have been uniform throughout. Some of the larger sites had portions unsuitable for Woodcock but we still rated them as useable habitat if the majority of the site was suitable or if Woodcock were found on part of the site. Fenwood (1976) and Webb (1978) identified 14,688 ac (5944 ha) of good and exceptional habitat statewide. Most of the area (7692 ac [3113 ha]) was located in the Valley, all of which was rated as exceptional. We identified an additional 1235 ac (500 ha) in the Valley as suitable for Woodcock. Our 1990s resurveys indicated that almost all of the Valley sites (90.0% of sites, 95.1% of area) originally classified as exceptional were still good Woodcock habitat. We classified them as probable or definite habitat in the 1990s. We did not reclassify any previously suitable sites as unlikely Woodcock habitat. Because we evaluated sites as a whole and at least part of every 1970s site was still considered suitable Woodcock habitat, it is difficult to quantify the amount of habitat that has been lost in the Valley since the 1970s surveys. Statewide, of the area originally rated as good or exceptional habitat in the 1970s, we rated 93.0% of it as definite Woodcock habitat. We confirmed that sites classified as poor habitat in the 1970s were still unsuitable. We reclassified only two sites as probable (8.0% of total site number, 49 ac [20 ha], 6.3% of total area) that were originally classified as poor. We did not reclassify any sites as definite. These two sites represented a small gain in suitable Woodcock habitat, but this did not balance the loss in habitat statewide. Forty-four percent of resurveyed sites and 19.4% of resurveyed area had become unsuitable for Woodcock since the mid-1970s. Reasons for statewide habitat loss included suburban and industrial development (6.6% of sites resurveyed); flooding due to dams (1.1%); conversion to agricultural use (22%), e.g., open fields, pastures, or row crops; and seral advancement of forest vegetation beyond what is suitable for Woodcock (14.3%). Microhabitat variables Microhabitat characteristics of Woodcock-flush points in the Valley differed from those in surrounding areas (Table 1). The Valley’s flush points featured higher soil pH, moister soils, a greater percent of hydric soils, and closer proximity to wetlands than flush points in surrounding areas. The Valley points also supported greater shrub-stem density and a lower density of large trees. Flush points in the Valley and surrounding areas did not differ in amount of herbaceous Southeastern Naturalist A.K. Steketee, P. Bohall Wood, and I.D. Gregg 2015 Vol. 14, Special Issue 7 337 plant cover or leaf-litter cover, although herbaceous plants at flush sites in the Valley were not as tall as herbaceous plants at flush sites outside the Valley. Soil texture and soil compaction did not differ between flush points inside and outside of the Valley. Landscape variables In comparisons of flush points within and outside of the Valley at both buffer scales (Table 2), all landscape variables except Shannon’s diversity index differed. Flush sites in the Valley at both buffers had fewer (patch richness) and larger patches than flush sites outside The Valley. Patch complexity (AWMPFD) Table 1. Means of microhabitat variables measured at American Woodcock-flush sites in Canaan Valley (n = 40) and surrounding areas (n = 32). Soil moisture is based on a scale of 1–3 where 1 = moist, 2 = medium, and 3 = dry. Soil texture is based on a scale of 1–3 where 1 = coarse, 2 = medium, and 3 = fine. Soil compaction is based on a scale of 1–3 where 1 = difficult, 2 = medium, and 3 = easy. Canaan Valley Outside Canaan Valley flushes flushes P > t Soil pH 4.91 4.58 P < 0.001 Soil moisture 1.95 1.44 P < 0.003 Soil texture 2.05 2.07 P < 0.847 Soil compaction 1.95 2.26 P < 0.208 Hydric soil (%) 34.1 8.9 P < 0.001 Distance to wetland area (m) 38 311 P < 0.028 Elevation (m) 973 1009 P < 0.043 Shrub-stem density (#/ha) 1508 320 P < 0.053 Trees > 9m tall (#/ha) 18 210 P < 0.002 Leaf-litter cover (%) 73 61 P < 0.287 Herbaceous cover (%) 71 66 P < 0.554 Herbaceous height (cm) 25 37 P < 0.039 Table 2. Mean of landscape variables measured within a 150-m and a 1200-m buffer around American Woodcock-flush sites in Canaan Valley (n = 40) and surrounding areas (n = 32). Canaan Valley Outside Canaan Valley flushes flushes P > t 150-m buffer Patch richness 4.3 4.9 P < 0.022 Mean patch size (ha) 0.78 0.36 P < 0.026 Mean shape index 1.42 1.46 P < 0.066 Shannon’s diversity index 0.84 1.65 P < 0.218 Area weighted mean patch fractal dimension 1.35 1.37 P < 0.024 1200-m buffer Patch richness 8.8 9.4 P < 0.043 Mean patch size (ha) 0.60 0.39 P < 0.001 Mean shape index 1.35 1.36 P < 0.040 Shannon’s diversity index 1.65 1.65 P < 0.495 Area weighted mean patch fractal dimension 1.40 1.43 P < 0.001 Southeastern Naturalist A.K. Steketee, P. Bohall Wood, and I.D. Gregg 2015 Vol. 14, Special Issue 7 338 was lower for the Valley’s flush sites within both the 492- and 3937-ft (150- and 1200-m) buffers. Discussion Woodcock require wet woodlands with young forest embedded with scattered openings (Sheldon 1967). Good habitat is generally defined as a dense, low, shrubby canopy often associated with a meandering stream. Alder is generally the most important shrub species present (Webb 1978, Webb and Samuel 1982). Webb (1978) described a good quality habitat site in West Virginia as: “… found in a gently sloping bottomland somewhere in the mountain region and was 45 ac (18.2 ha) in size. It had been used as pastureland sometime in the past but was now abandoned and shrubs were taking over; alder had come in along a small stream and hawthorn was present on the higher ground. The height of the shrub cover was variable, but was generally 9.8–20 ft (3–6 m); canopy closure also was variable but averaged 60%. Ground-cover density was intermediate and moisture content of the soil was damp.” This description is still applicable to Woodcock habitat in West Virginia (Steketee 2000), particularly in the Valley. Additionally, good but atypical habitat was found in the vicinity of the Valley. For example, we found good habitat on the Monongahela National Forest in young Acer pensylvanicum L. (Striped Maple) and Crataegus spp. (hawthorn) stands sited on flat benches and ridge tops (Gregg 1997, Gregg et al. 2000, Steketee 2000,). Our comparison of Woodcock sites within and outside the Valley suggests that characteristics related to soil moisture are the primary differences between them. Additionally, the Valley’s flush points were close to shrubby wetlands; 68% of flush points, vs. only 25% of flush points outside the Valley, occurred directly in shrubby wetlands, although this habitat type comprised less than 2% of the total land area in the study area. This result indicates that Woodcock preferentially choose shrubby wetlands. Distance to shrubby wetlands is a good predictor of Woodcock presence in West Virginia (Steketee 2000). Woodcock are known to make heavy use of wooded wetland areas, particularly those with alder (Sepik et al. 1989, Sheldon 1967), probably because they contain the moist, rich soils that harbor high densities of earthworms (Parris 1986), the Woodcock’s primary food (Sheldon 1967). In heavily forested, mountainous areas, like West Virginia, these soil types are largely limited to wetlands, riparian areas, and small, scattered seeps. Most stream banks and riparian areas in the study area were, however, covered with forest vegetation rather than shrubs, and thus may not have been used by Woodcock (Webb and Samuel 1982). Woodcock in West Virginia are occasionally associated with seeps or other damp areas too small to map with aerial photographs (Webb 1978) or satellite imagery (Steketee 2000). Shrub-stem densities were higher in the Valley than in surrounding areas. However, the 3770 stems/ac (1508 stems/ha) value we calculated for our study Southeastern Naturalist A.K. Steketee, P. Bohall Wood, and I.D. Gregg 2015 Vol. 14, Special Issue 7 339 area was well below the 25,000–125,000 stems/ac (10,000–50,000 stems/ha) that Straw et al. (1994) indicated was typical and the 18,833–289,163 stems/ac (7533–115,665 stems/ha) reported by Sepik et al. (1989). Because of these large differences among the studies, Sepik et al. (1989) concluded that it is difficult to recommend an optimal stem density for Woodcock. Many sites in West Virginia, particularly those outside of the Valley, probably contain suboptimal shrubstem- densities. High shrub-stem density moderates soil temperature, maintains moisture for earthworms (Parris 1986, Reynolds et al. 1977), minimizes herbaceous growth that would inhibit Woodcock access to worms (Sepik et al. 1989), and provides cover to protect the birds from predators. Although all of West Virginia is usually included in its breeding range, in-state Woodcock habitat is limited in both its total acreage and distribution (Fenwood 1976, Steketee 2000, Webb 1978). Because of the state’s mountainous terrain, there are virtually no large expanses of Woodcock habitat. Rather, large patches of the shrubby lowland areas preferred by Woodcock are uncommon. Woodcock habitat areas tends to be small, scattered, patchy, and often irregularly shaped (Fenwood 1976, Steketee 2000, Webb 1978). The exception to this pattern is in the Valley. Our landscape analysis indicated that sites used by The Valley’s Woodcock contained fewer patch types and larger patch sizes than occurrences outside the Valley. We expected this result because Woodcock prefer shrubby wetlands, a habitat type found in large patches only in the Valley. The Valley is the only place in West Virginia that contains large patches of wetlands with the high shrub-stem densities and moist soils that Woodcock require; these features appear to be relatively constant over time. Areas used by Woodcock outside the Valley are highly variable and include some shrub wetlands, but also regenerating timber cuts, reverting old fields, and reclaimed surface mines. Some of these areas, although used by Woodcock, seem to be suboptimal. For example, Gregg et al. (2003) found that although Woodcock used reclaimed surface mines in West Virginia, many of these sites did not provide ideal habitat due to inferior soils and/or insufficient stem densities. Moreover, many areas of Woodcock habitat outside the Valley rapidly become unsuitable due to forest succession. Ongoing management is necessary to create and maintain habitat if net habitat losses are to be avoided. Woodcock habitat is declining range-wide. Several researchers have associated this loss with increasing amounts of urban/developed land and mature forest stands, and with limited availability of regenerating stands dominated by seedlings and saplings (Coulter and Baird 1982, Dwyer et al. 1983, Gutzwiller et al. 1983, USFWS 1990). In our statewide habitat re-inventory, Steketee (2000) reported that 35.9% of resurveyed sites and 19.4% of resurveyed area became unsuitable for Woodcock from the mid-1970s to the 1990s because of forest-stand succession and/or human causes including land development, agriculture, and flooding. Conversion to agricultural uses, like crops, open fields, and animal pasture, accounted for the majority of sites and area lost during our statewide resurvey Southeastern Naturalist A.K. Steketee, P. Bohall Wood, and I.D. Gregg 2015 Vol. 14, Special Issue 7 340 (Steketee 2000). Lowlands and floodplains provide not only prime Woodcock cover, but also an inviting location for urban or agricultural development. This is especially true in West Virginia, where, because such lands are at a premium (Kletzly 1976), humans have diverted much of this flat land for their own uses (Webb 1978). Statewide, forest-stand maturation and urban development were responsible for 20.9% of the sites and 10.9% of the area lost as Woodcock habitat between the 1970s and our resurveys (Steketee 2000). Others have also connected forest-stand succession and urbanization with a loss of Woodcock habitat. In Pennsylvania, Gutzwiller et al. (1983) determined that increases in the amount of mature forest (e.g., sawtimber with ≥1/2 trees of DBH > 11.7 in [27.9 cm]) and urban area were consistent with an apparent decline of Woodcock populations. Dwyer et al. (1983) analyzed habitats along SGS routes in the northeastern US and concluded that declines in Woodcock numbers were related to increases in urban development. Earlier researchers in West Virginia (Fenwood and Webb 1981, Kletzly 1976) noticed and commented on a pattern of human encroachment into Woodcock habitat, a trend that appears to be continuing. Forest-stand succession is also a factor in Woodcock habitat loss. West Virginia’s forests have become increasingly mature since the last major cutting cycle at the turn of the 19th–20th centuries (DiGiovanni 1990). Along SGS routes in West Virginia, Steketee et al. (2000) found that the amount of mature deciduous forest increased as Woodcock counts decreased. The Valley has historically contained the best habitat and largest Woodcock populations in West Virginia (Goudy et al. 1970, Webb 1978, Webb and Samuel 1982). It is also an important migration staging area for the species (USFWS 1990). Within the Valley, Goudy et al. (1970) estimated that about 7900 ac (3200 ha) were suitable for Woodcock in the 1960s. In the 1970s, Fenwood (1976) and Webb (1978) identified 7692 ac (3113 ha) of exceptional Woodcock habitat in the Valley, and scored an additional 1235 ac (500 ha) as suitable. Our resurveys in the 1990s indicated that almost all of the Valley’s sites (90.0% of sites, 95.1% of area) originally classified as exceptional were still good Woodcock habitat. However, Michael (1993) stated, “Woodcock habitat is rapidly disappearing throughout the Valley as development and ecological succession affect many areas that once supported large numbers of breeding and migrating Woodcock. It is estimated that as many as 1300 Woodcock nested in the Valley during the 1960s and 1970s. Although no recent estimates have been made, the resident population has declined significantly, probably to fewer than 450 birds.” Michael (1993) further stated that “prime Woodcock [breeding] habitat in the alder and aspen stands … is presently somewhat overgrown at ground level [due to removal of cattle grazing], while old-field and wet-meadow habitats used for courtship are abundant in the Valley.” Management Opportunities and Ecological Implications Before extensive logging, the Valley’s plant communities were dominated by large Picea rubens Sarg. (Red Spruce) with a dense understory of Rhododendron Southeastern Naturalist A.K. Steketee, P. Bohall Wood, and I.D. Gregg 2015 Vol. 14, Special Issue 7 341 maximum L. (Great Laurel). Natural openings or glades in this forest consisted of dry grass balds or bogs that were too wet to support most tree species. Early successional habitats likely occurred in small patches throughout the Valley as a result of small-scale disturbances such as windthrows. Pre-settlement-era Woodcock populations likely occurred in these small early-successional habitats. Consequently, we speculate that Woodcock populations in the pre-logging era were low. However, the generally moist, humus-rich acidic soil would have provided ideal conditions for earthworms, the primary food of Woodcock. Provision of adequate Woodcock habitat is imperative if the current negative trend in the species’ populations is to be reversed. Given the relatively ephemeral nature and sometimes suboptimal quality of some types of Woodcock habitat, there is a particular need to identify, protect, and in many cases actively manage areas that currently provide or are capable of providing stable, high-quality Woodcock habitat. Sites used by Woodcock outside the Valley tend to be in earlysuccessional habitat associated with regenerating clearcuts or reverting old fields, and thus were suitable for a relatively short period. Alder thickets, however, such as those found within the Valley, tend to persist for longer periods and provide a more consistent habitat. Currently, the Valley includes 40 different wetland and upland plant communities (USFWS 1994) that provide a mix of habitat types. Although habitat conditions may differ from those that were historically present, we believe that protection and active management of Woodcock habitat in the Valley are high priorities. The Valley is critically important to breeding and migrating Woodcock and several of the currently available habitat types can be managed for this declining species. The Canaan Valley National Wildlife Refuge now has over 15,000 acres (6070 ha), which includes most of the Woodcock’s available habitat (Ken Sturm, USFWS, pers. comm.). Refuge lands are protected from development and provide numerous opportunities for managing early successional habitats, which will benefit not only Woodcock but also many other plant and animal species (Denmon 1998). Literature Cited Coulter, M.W., and J.C. Baird. 1982. Changing forest land-uses and opportunities for Woodcock management in New England and the Maritime Provinces. Pp. 75–85, In T.J. Dwyer and G.L. Storm (Technical Coordinators). Proceedings of the Seventh Woodcock Symposium. US Fish and Wildlife Research Report 14. Washington, DC. Denmon, P. 1998. Early-successional habitat use by nongame wildlife species in American Woodcock breeding habitat in West Virginia. M.Sc. Thesis. West Virginia University, Morgantown, WV. 120 pp. DiGiovanni, D.M. 1990. Forest statistics for West Virginia 1975 and 1989. US Department of Agriculture, Forest Service Resource Bulletin NE114, Radnor, PA. 28 pp. Dwyer, T.J., D.G. McAuley, and E.L. Derleth. 1983. Woodcock singing-ground counts and habitat changes in the northeastern United States. Journal of Wildlife Management 47:772–779. ESRI. 1996. The Arc/Info Geographic Information System, Version 7.1. Redlands, CA. Southeastern Naturalist A.K. Steketee, P. Bohall Wood, and I.D. Gregg 2015 Vol. 14, Special Issue 7 342 Fenwood, J.D. 1976. An inventory of Woodcock habitat in southern West Virginia. M.Sc. Thesis. West Virginia University, Morgantown, WV. 110 pp. Fenwood, J.D., and L.O. Webb. 1981. An aerial photo inventory of Woodcock habitat in West Virginia. Wildlife Society Bulletin 9:306–307. Goudy, W.H., R.C. Kletzly, and J.C. Rieffenberger. 1970. Characteristics of a heavily hunted Woodcock population in West Virginia. Transactions of the North American Wildlife and Natural Resources Conference 35:183–195. Gregg, I.D. 1997. American Woodcock use of reclaimed surface mines in West Virginia. M.Sc. Thesis. West Virginia University, Morgantown, WV. 112 pp. Gregg, I.D., P.B. Wood, and D.E. Samuel. 2000. American Woodcock use of reclaimed surface mines in West Virginia. Pp. 9–22, In D.G. McAuley, J.G. Bruggink, and G.F. Sepik (Eds.). Proceedings of the Ninth American Woodcock Symposium. US Geological Survey, Biological Resources Division Information and Technology Report USGS/BRD/ITR-2000-0009, Patuxent Wildlife Research Center, Laurel, MD. Gregg, I.D., P.B. Wood, and A K. Steketee. 2003. Comparison of reclaimed and unmined Woodcock diurnal habitat in West Virginia. Proceedings of the Annual Conference of the Southeast Association of Fish and Wildlife Agencies 55: 488–497. Gutzwiller, K.J., and J.S. Wakeley. 1982. Differential use of Woodcock singing-grounds in relation to habitat characteristics. Pp. 51–54, In T.J. Dwyer and G.L. Storm (Technical Coordinators). Woodcock ecology and management. US Fish and Wildlife Service Wildlife Research Report 14, Washington, DC. Gutzwiller, K.J., K.R. Kinsley, G.L. Storm, W.M. Tzilkowski, and J.S. Wakeley. 1983. Relative value of vegetation structure and species composition for identifying American Woodcock breeding habitat. Journal of Wildlife Management 47:535–540. James, F.C., and H.H. Shugart, Jr. 1970. A quantitative method of habitat description. Audubon Field Notes 24:727–736. Kelley, J.R., Jr. 2001. American Woodcock population status, 2001. US Fish and Wildlife Service, Laurel, MD. 20 pp Kletzly, R.C. 1976. American Woodcock in West Virginia. West Virginia Division of Natural Resources Bulletin No. 8, Charleston, WV. McCoy, R.W. 1987. PAM HEP habitat suitability index model for Woodcock. US Fish and Wildlife Service, State College, PA. 4 pp. McGarigal, K., and B.J. Marks. 1995. FRAGSTATS: Spatial pattern analysis program for quantifying landscape structure. General Technical Report PNWGTR351. USDA Forest Service, Pacific Northwest Research Station, Portland, OR . 122 pp. Michael, E. 1993. An evaluation including southern Canaan Valley in the proposed Canaan Valley National Wildlife Refuge. Report submitted to US Fish and Wildlife Service and US Environmental Protection Agency. Parris, R.W. 1986. Forest vegetation, earthworm, and Woodcock relationships. Ph.D. Dissertation. State University of New York, Syracuse, NY. 240 pp. Reynolds, J.W., W.B. Krohn, and G.A. Jordan. 1977. Earthworm populations as related to Woodcock habitat usage in central Maine. Proceedings of the Sixth Woodcock Symposium 6:13–146. Sepik, G.F., and E.L. Derleth. 1993. Habitat use, home-range size, and patterns of moves of the American Woodcock in Maine. Pp. 41–49, In J.R. Longcore and G.F. Sepik (Eds.). Proceedings of the Eighth Woodcock Symposium. US Fish and Wildlife Service Biological Report 16. Washington, DC. Southeastern Naturalist A.K. Steketee, P. Bohall Wood, and I.D. Gregg 2015 Vol. 14, Special Issue 7 343 Sepik, G F., R.B. Owen, Jr., T.J. Dwyer, and E.L. Derleth. 1989. Habitat requirements and management of Woodcock in the Northeast: Assessment of knowledge and needs. Pp. 97–109, In J.C. Finley and M.C. Brittingham (Eds.). Timber management and its effects on wildlife. Proceedings of the 1989 Pennsylvania state forest resources issues conference, University Park, PA. Sheldon, W.G. 1967. Book of the American Woodcock. University of Massachusetts Press, Amherst, MA. 250 pp. Steketee, A.K. 2000. Predicting habitat suitability for American Woodcock and landscape- level assessment of habitat in West Virginia. Ph.D. Dissertation. West Virginia University, Morgantown, WV. 132 pp. Steketee, A.K., P.B. Wood, J.G. Bruggink, D.E. Samuel, and J.I. Cromer. 2000. Land-use/ land-cover changes along Woodcock singing-ground survey routes in West Virginia. Pp. 107–116, In D.G. McAuley, J.G. Bruggink, and G.F. Sepik (Eds.). Proceedings of the Ninth American Woodcock Symposium. US Geological Survey, Biological Resources Division Information and Technology Report USGS/BRD/ITR-2000-0009, Patuxent Wildlife Research Center, Laurel, MD. Straw, J.A., D.G. Krementz, M.W. Olinde, and G.F. Sepik. 1994. American Woodcock. Pp. 97–114, In T.C. Tacha and C.E. Braun (Eds.). Migratory shore and upland gamebird management in North America. International Association of Fish and Wildlife Agencies, Washington, DC. United States Fish and Wildlife Service (USFWS). 1979. Final environmental impact statement: Acquisition of lands for the Canaan Valley National Wildlife Refuge West Virginia. US Department of the Interior, Washington, DC. 20 pp. USFWS. 1990. American Woodcock management plan. US Fish and Wildlife Service, Washington, DC. 11 pp. USFWS. 1994. Canaan Valley National Wildlife Refuge Station management plan. US Fish and Wildlife Service, Region 5. Hadley, MA. Vogelmann, J.E., S.M. Howard, and T. Sohl. 1996. Regional landcover characterization using multiple sources of intermediate-scale data. Proceedings 1996 International Geoscience and Remote Sensing Symposium 1:246–247. Webb, L.O. 1978. Woodcock habitat inventory of northern West Virginia using aerial photographs. M.Sc. Thesis. West Virginia University, Morgantown, WV. 61 pp. Webb, L.O., and D.E. Samuel. 1982. Woodcock habitat inventory and wetlands in northern West Virginia. In B. McDonald (Ed.). Proceedings of a Symposium on Wetlands of the Unglaciated Appalachian Region, Morgantown, WV.