Herbaceous-layer impoverishment in a post-agricultural southern Appalachian landscape.
Globally, protected areas are important elements of plant and animal conservation strategies (Brooks et al., 2004). However, even within protected areas, species loss and ecosystem simplification resulting from altered or suppressed disturbance regimes have been documented (Fauzi and Buchary, 2002; Rooney et al., 2004), suggesting traditional conservation efforts focused primarily on habitat protection may not be completely successful.
Many protected areas are managed with a "natural regulation" ethic, which limits direct human interference with biological processes (Soukup et al., 1999). However, due to historic and contemporary anthropogenic alterations of protected areas and their environs, biological systems in these protected areas often require management beyond natural regulation to ensure ecosystem integrity (Porter and Underwood, 1999; Weisberg and Coughenour, 2003; Ervin, 2004; Turner and Beazley, 2004). For example, in the eastern and midwestern United States, forest fragmentation, row crop agriculture and the extirpation of large predators have facilitated the expansion of white-tailed deer (Odocoileus virginianus) populations to unprecedented levels (Alverson et al., 1988; Anderson, 1997; Augustine and deCalesta, 2003). Consequently, intense herbivory associated with locally overabundant deer populations poses a major threat to native plant communities, especially where hunting is limited to conservative buck-only harvesting or prohibited entirely (Rooney and Dress, 1997; Horsley et al., 2003; Campbell et al., 2004; Rossell et al., 2005; Barrett and Stiling, 2006; Tremblay et al., 2007).
Leopold (1933) was among the fist to articulate the importance of edge habitats for species such as the white-tailed deer. Edges between mature and successional forests or between forest and field are often areas of enhanced biological diversity, but an increasing awareness has emerged in recent years that edge effects are not inherently positive (Saunders et al., 1991; Guthery, 1997; Lidicker, 1999). For example, forest edges in landscapes dominated by agricultural row crops may concentrate mammalian predators, which constitute an ecological trap for nesting songbirds in these habitats (Heske et al., 1999). Similarly, herbivores that use edges may profoundly modify the structure and composition of these habitats (Cadenasso and Pickett, 2000). Highly mobile ungulates, such as white-tailed deer whose populations respond positively to fragmented landscapes (Anderson, 1997), may alter woody plant communities distal from recognizable edges (Alverson et al., 1988; Cadenasso and Pickett, 2000). Nevertheless, little is known about the potential consequences of herbivore-mediated edge effects on the long-term stability of herbaceous plant communities.
We examined long-term changes in the composition and structure of understory vegetation within and around Cades Cove, Great Smoky Mountains National Park (GSMNP), USA. In the 1930s, all property in Cades Cove was acquired for inclusion in GSMNP, which was dedicated in 1940 (Campbell, 1994). At that time, the Cove was a mosaic of small fields, pastures and woodlots surrounded by continuous secondary forest. Following acquisition, many old fields were planted with non-native grasses (e.g., Lolium pratense (Huds.) S.J. Darbyshire), while others were allowed to succeed to forest. Woodlots and succeeding fields have not been managed or grazed by livestock since acquisition. Combined with the absence of hunting and large predators, this habitat mosaic provided ideal habitat for white-tailed deer. Consequently, Cades Cove has a long and well-documented history of deer overabundance, and at times has contained some of the highest reported deer densities in the Southeast (Wathen and New, 1989). A recent study in Cades Cove identified pronounced changes in the spring flora attributable to foraging by the eruptive deer population, which reached a peak density of 43 deer [km.sup.-2] during the late 1970s (Webster et al., 2005a). Recent exclosure studies in Cades Cove have shown that contemporary deer densities, while lower than peak levels, are sufficient to inhibit the regeneration of most tree species (Fig. 1; Griggs et al., 2006; Webster et al., 2008) and reduce the stature and fecundity of early-blooming, liliaceous species (Webster et al., 2005a; Jenkins et al., 2007 ).
We reinventoried 30 long-term monitoring plots that were established in western Great Smoky Mountains National Park between 1977-1979. Nineteen of these plots were located in the woodlots and surrounding secondary forests of Cades Cove. These plots were the dominant focus of this study and received the most in-depth analysis. The remaining 11 plots served as a reference to assess whether broad changes in composition observed on the Cades Cove plots were occurring in areas of similar vegetation but without a history of chronic herbivory.
This well-chronicled history of deer abundance in conjunction with long-term permanent plot data (26 y interval) from the 19 plots in Cades Cove provided a unique opportunity to examine the long-term consequences of deer overabundance and chronic herbivory on forest plant communities recovering from past agricultural use and settlement. Additionally, the geographic isolation of the Cove (surrounded by mountains and embedded within a large forested matrix) helped reduce the influence of other potential park boundary effects.
Nineteen of our study plots were located in western Great Smoky Mountains National Park (GSMNP), in Cades Cove, TN, a broad valley managed as a cultural landscape (i.e., 19th Century agrarian settlement). Cades Cove covers approximately 2700 ha and is surrounded by mountainous parkland, which protects the Cove from boundary effects between protected natural areas and other ownerships. Although Cades Cove was incorporated into GSMNP in 1940, some fields were leased for crops and grazing until the early 1990s. Cattle were prevented from grazing in woodlots through an extensive network of fences. The contemporary landscape of Cades Cove consists of woodlots interspersed within old-fields that are mowed and burned to represent a 19th Century agrarian settlement. Presently, Cades Cove is surrounded on all sides by closed-canopy forest.
Following incorporation of GSMNP, white-tailed deer numbers in Cades Cove increased dramatically in the absence of large predators, harsh winters and hunting. Pre-settlement deer densities in mountainous regions of the southeastern United States have been estimated at between 3-4 deer [km.sup.-2] (Knox, 1997), but were probably even lower during the settlement period due to heavy hunting and habitat destruction (Dunn, 1988). However, by the late 1970s, it was estimated that deer densities in Cades Cove were 43 deer [km.sup.-2] (Kiningham, 1980; Wathen and New, 1989). Eruptive population growth during the 1970s was punctuated by a crash due to a disease outbreak, but spotlight surveys conducted during the early and mid-1980s indicated that the population had rebounded and stabilized at a relatively high density [Wathen and New, 1989; see Griggs et al. (2006) for a detailed chronology of deer abundance in Cades Cove].
The remaining 11 plots were located in areas of similar vegetation, but without a history of chronic herbivory. All reference plots were at least 3 km linear distance from Cades Cove and physically separated by mountainous terrain. These plots were likewise located in stands with a history of human disturbance including logging of varying intensity, ground fires and livestock grazing (Pyle, 1988). Reference plots exhibited a wider range of elevations (500-1180 m) than the Cades Cove plots (520-680 m). Soils underlying both Cades Cove and reference plots were largely derived from metasedimentary rock such as metasandstone and phylite (Natural Resources Conservation Service, 2007). In Cades Cove, soils were formed from metasandstone derived from debris fans of varying thickness overlying carbonite bedrock (Southworth et al., 2003). Soils on the reference plots were formed from metasedimentary bedrock (Southworth et al., 2005).
[FIGURE 1 OMITTED]
Detailed vegetation data were collected from the 30 plots during the summers of 1977-1979 as part of a larger vegetation study by GSMNP (White and Busing, 1993). The nineteen permanent plots, randomly located in woodlots and forests in and around the Cove, encompassed three forest community associations as defined by NatureServe (White et al., 2003):
(1) Linodendron tulipifera--Betula lenta--Tsuga canadensis / Rhododendron maximum Forest (hereafter "acid cove") stands are generally located along upland streams at the edge of the Cove. Contemporary overstories of these stands are dominated by T. canadensis L. Carr. and L. tulipifera L. (Webster et al., 2005b). R. maximum L. typically dominates the understory. We sampled four plots in this association.
(2) Pinus strobus--Quercus alba (Carya alba) / Gaylussacia ursina Forest (hereafter "mixed conifer-hardwood") is typically located along the edge of the Cove and on slopes and low ridges within the interior woodlots. Contemporary overstories of these stands are dominated by Quercus species, P. strobus L. and T. canadensis (Webster et al., 2005b). We sampled seven plots in this association.
(3) Pinus virginiana Successional Forest (hereafter "successional pine") stands are generally located on gentle slopes between older forest and pasturelands. Contemporary overstories are dominated by P. virginiana Mill. (a short-lived pioneer species), Oxydendrum arboreum (L.) DC. and P. strobus. We sampled eight plots in this association.
Based on aerial photo interpretation, acid coves and mixed conifer-hardwood stands were forested prior to National Park Service acquisition, and successional pine stands were abandoned agricultural fields (Webster et al., 2005b). Six of the reference plots were classified as acid cove forests and five were classified as mixed conifer hardwood forests. No reference plots were available within successional pine forests.
Because heavy cover of subcanopy, evergreen ericaceous shrubs (e.g., Rhododendron maximum) and conifers (i.e., Tsuga canadensis) can reduce herbaceous-layer cover and diversity (Baker and Van Lear, 1998), we tested for differences in the density of this group of species between mixed conifer-hardwood and acid cove plots in Cades Cove and reference areas. While reference plots contained greater density of these species than the Cades Cove plots (1642 [+ or -] 494 vs. 698 [+ or -] 298 stems [ha.sup.-1]), the difference was not significant (one-way ANOVA, [F.sub.1.22] = 3.08, P = 0.094) due to high variability among plots.
Within each of the thirty 20 x 50 m permanent plots, a grid of ten 10 x 10 m subplots was established and permanently marked with steel rod (White and Busing, 1993). Within every-other subplot (n = 5), a diagonal transect (14.14 m) was established at a random corner along which five 1 x 1 m herbaceous cover plots were established. A total of 25 herbaceous cover subplots (25 [m.sup.2] total) were sampled in each plot (White and Busing, 1993). During the 1977-1979 sampling period, all woody stems [greater than or equal to] 1 cm dbh (1.37 m) were measured on the 20 x 50 m plot and percent cover was estimated for each herbaceous species on each of the 25 subplots. Percent cover was estimated ocularly into 10 cover-abundance scale classes (Peet et al., 1998; class midpoints were used for analysis). All herbaceous species occurring on the 20 X 50 m plot were identified and recorded for determination of species richness.
Overstory vegetation on the Cades Cove plots was reassessed between 1995 and 2002 (Webster et al., 2005b). We reinventoried herbaceous vegetation on each subplot in 2004 using the same methods as the original inventory conducted in 1977-1979 (described above). Overstory and herbaceous vegetation on the reference plots were reinventoried between 1995 and 2001 using the same methods as the original 1977-1979 survey. To avoid seasonal variation in species coverage/representation, plots were reinventoried as close as possible to the initial sampling date. Mean absolute and relative percent cover (cover of species i / total coverage of all species) of individual species were calculated by plot. Species nomenclature follows Kartesz (1999).
To evaluate compositional differences through time along environmental gradients in Cades Cove, we used non-metric multidimensional scaling (NMS; Kruskal, 1964; Mather, 1976) as implemented in PCADRD version 4.24 (McCune and Mefford, 1999). Compositional dissimilarity was assessed using a square root transformed and 'zero-adjusted' Sorensen's (Bray-Curtis) coefficient. Zero-adjusted Sorensen's is appropriate for denuded assemblages where impoverishment is likely due to a common cause and not simply a consequence of sampling error that accompanies inadequate sample sizes (Clarke et al., 2006). The zero-adjustment has the advantage that it has little effect on dissimilarities between non-denuded samples, but has a desirable homogenizing effect on dissimilarities between blank or near-blank samples that would otherwise be undefined or demonstrate erratic behavior. A square-root transformation reduces the influences of samples with high cover or abundance which otherwise dominate dissimilarity values (Field et al., 1982). For the NMS, we used PC-ODRD's autopilot mode, with default settings of 0.00001 for the stability criterion, 40 real runs and 50 randomized runs of 500 iterations each for each dimensionality (or 50 continuous iterations within the stability criterion), and a starting dimensionality of six (McCune and Grace, 2002). Autopilot uses a random starting configuration and selects appropriate dimensionality for the ordination by (1) minimizing stress (a measure of lack of fit; Kruskal, 1964; Mather, 1976) for each dimensionality, (2) considering additional dimensions only if they substantially reduced stress and (3) performing Monte Carlo tests, whereby the final stress for each dimensionality must be lower than that for 95% of randomized runs, thus assuring that additional dimensions capture new information. We assessed the role of underlying environmental gradients by joint-plotting environmental variables and species scores in sample unit space. Correlations of environmental variables to ordination axes were calculated with Pearson's r, and environmental vectors were scaled to the maximum absolute value obtained by the axes in the ordination plot.
Because the NMS vector plot suggested an increase in similarity among plots since 1977-1979, we used the program PERMDISP (Anderson, 2004; Anderson et al., 2006) to test the null hypothesis of no differences in D diversity between sampling periods. PERMDISP tests for homogeneity in multivariate dispersions by doing permutational ANOVA on average distances to group centroids, as defined in the principle coordinate space of an appropriate dissimilarity measure. Significant decreases in average distances to group centroids between time periods would strongly suggest biotic homogenization. Alternatively, we could simply compare average dissimilarities between all pairs of samples, but this approach does not permit statistical evaluation due to partial dependencies of individual dissimilarity values (Anderson et al., 2006).
For comparisons of species richness and total percent cover between time periods, we used a paired t-test. This test is appropriate for dependant observations, such as time series data (Zar, 1999). Means are reported plus or minus one standard error. Pearson correlation analysis was used to examine relationships between environmental and stand structure variables (Zar, 1999). To help identify species characteristics related to changes in community composition, we evaluated changes in the representation of various life history traits (USDA PLANTS Database, 2004). Growth forms were selected to characterize browse susceptibility and impact (low-growing: low susceptibility; bunch: high susceptibility, low impact; single-crown: high susceptibility, high impact; multiple crowns: high susceptibility, low impact). Shade tolerance was examined to help isolate the effects of forest succession and stand development from herbivory. Distance measures (distance from plot edge to road and forest/field edge) were determined with Arcview (version 3.2). A one-way ANOVA with a Tukey's pairwise comparison was used to compare the mean distance from edge among plots in the three forest types located around Cades Cove.
Since they were resampled 4-9 y prior to the reinventory of the Cove plots, span a greater range of elevations and were not a formal control, reference plot data were not compared statistically to the data from Cades Cove. As a point of reference, we summarized species loss and changes in percent cover on reference plots and compared the fates of woodland species lost from the Cades Cove plots over the sampling interval. Using the procedures outlined above we also evaluated differences in [beta] diversity between sampling periods. Simple linear regression was used to test the influence of elevation on percent cover and species richness since the greater range of elevation across reference plots could influence these attributes.
SPECIES LOSS AND ABUNDANCE WITHIN THE CADES COVE SETTLEMENT AREA
Overall in Cades Cove, 46 herbaceous species recorded on plots during the 1970s were absent in 2004, and 10 species not recorded during the 1970s were recorded in 2004 (Table 1). Mean coverage of herbaceous species declined significantly (from 18.4 [+ or -] 7.9% to 6.8 [+ or -] 3.8%, P < 0.001). Declines within forest types ranged from 57.2% in mixed conifer-hardwoods to 85.1% in successional pine stands. Both successional pine and mixed conifer-hardwood forests contained exotic species during both sampling intervals. The mean coverage of these species also declined (from 7.6 [+ or -] 5.2% to 3.9 [+ or -] 3.5%). In fact, five exotic species were lost, whereas only one exotic was gained (Table 1). The exotics that were lost included both shade-tolerant and shade-intolerant species.
Absolute coverage of almost all individual species life forms declined between the sampling intervals (1977-1979 and 2004; Table 2). The only increase observed was in the coverage of annual plants, which is associated with an increase in the coverage of the exotic grass Microstegium vimineum (Trin.) A. Camus. The greatest declines were observed for species with single crowns and bunch growth forms. The decline in species with bunch growth forms (mostly graminoids) was associated with a general decline in species less tolerant of shade. Correspondingly, an increase in relative coverage of shade-tolerant species was observed; however, absolute coverage declined by 49% (Table 2). Low-growing species also increased in relative coverage while declining in absolute coverage (Table 2).
COMPOSITIONAL CHANGE ALONG ENVIRONMENTAL GRADIENTS WITHIN THE CADES COVE SETTLEMENT AREA
A three dimensional NMS solution (final stress = 12.81, normalized and scaled from 0100; P-value for Monte Carlo test = 0.0196) provided the best ordination of the multidimensional species data. The three NMS axes combined accounted for 75% of the variance in species data ([r.sup.2] for axes 1, 2 and 3 were 0.22, 0.11, and 0.42, respectively). Of the environmental gradients examined (Table 3), distance to edge had the strongest overall correlation to the NMS ordination (Table 4). Stem density, A in basal area, and tree age had moderate correlations to ordination axes 1 and 2, suggesting that these axes, in part, describe successional gradients, whereas axis 3 describes edge effects. Distance variables were not correlated (P > 0.05) with overstory attributes (Table 5). Similarly, while acid cove plots were more distal from edges than either mixed conifer-hardwood or successional pine plots (P = 0.011 and 0.003, respectively), there was no significant difference in the proximity of plots in the latter two forest types to field edges (P = 0.782).
Life forms sensitive to deer browsing (i.e., lilies and orchids) displayed a positive association with distance whereas more browse tolerant species (i.e., ferns, graminoids and legumes) displayed a negative association with distance (Fig. 2). Exotic species were also associated with shorter distances from the Cove and its edges (Fig. 2). The percent cover of this group, however, declined by approximately 50% over the course of the study.
UNDERSTORY HOMOGENIZATION WITHIN THE CADES COVE SETTLEMENT AREA
Visual interpretation of the NMS vector plot of samples through time in species-space (Fig. 3) revealed that compositional change was greatest along axis-3 and that plots have become more homogeneous in composition. For all plots considered together and for the mixed conifer-hardwoods habitat type, PERMDISP results showed that average distances to group centroids significantly declined over time (Table 6). The composition of successional pine plots also became more similar between sampling intervals, but this result was not statistically significant (P = 0.18). Acid cove plots, on the other hand, showed no trend towards increasing similarity.
Plots close to forest/field edges showed the greatest reduction in heterogeneity based on differences in distance to group centroids between sampling periods (Fig. 4). Ninety one percent of the plots that were within 200 m of an edge declined in heterogeneity. At greater distances, the distribution of increases and decreases was more evenly divided, but sample sizes were too low to identify a threshold distance for edge-mediated homogenization.
PLANT COMMUNITY CHANGE ON REFERENCE PLOTS
In general, reference plots accumulated woodland species over the course of the sampling interval and mean species richness increased by 106 [+ or -] 28% (Table 7). Only one plot exhibited a net loss of species--the majority of which were early successional (e.g., Dichanthelium commutatum (J.A. Schultes) Gould, Lespedeza hirta (L.) Hornem. and Schizachyrium scoparium (Michx.) Nash). Across both forest types, percent cover increased by 183 [+ or -] 76% (Table 7). While four plots exhibited declines in percent cover over the sampling interval, none experienced a net loss of species, and three of the four experienced an increased in species richness. For all reference plots considered together and for the acid cove habitat type, PERMDISP results showed that average distances to group centroids did not decline significantly over time (Table 6). Mixed conifer-hardwood plots did, however, increase in similarity, but in contrast to the Cades Cove plots this increase in similarity was associated with an increase in species richness (mean increase, 68 [+ or -] 36%). Species gained on reference plots were typically interior woodland species such as Goodyera repens (L.) R. Br. ex Ait. f., Prosartes lanuginosa (Michx.) D. Don, Dioscorea quaternata J.F. Gmel., Polygonatum biflorum (Walt.) Ell., Lilium michauxii Poir. and Trillium undulatum Willd. In other words, homogenization on mixed conifer-hardwood reference plots was driven by species gain rather than loss. While the reference plots spanned a greater range of elevations than the Cove plots, no significant trends between richness or cover and elevation were observed within this elevation range (P = 0.561 and P = 0.390, respectively).
Our results suggest that over the last 26 y the understory plant communities across three distinct forest types within Cades Cove have become increasingly impoverished. In 2004, the herbaceous layer on the sample plots contained an average of nine species, which had a total cover of less than 7%. However, during the 1970s inventory these same plots contained an average of 11 species and had a total percent cover of over 18%. These results are in contrast to those observed on reference plots; where species richness increased, understory plant communities on different forest types did not become more similar and within forest type homogenization was predominately associated with species gain rather than loss. At the time of the 1970s survey, Cades Cove contained one of the highest documented deer densities in the southeast at 43 deer [km.sup.-2] (Kiningham, 1980; Wathen and New, 1989). Since that time, deer densities have declined, but remain much higher than those found elsewhere in GSMNP (Griggs et al., 2006). Vegetation surveys conducted in the Cove during the 1970s documented significant reductions in herbaceous coverage and species richness due to intense deer herbivory (Bratton, 1979). Our results suggest that the chronic, but presumably less intense, herbivory associated with the contemporary deer herd has been sufficient to inhibit plant community recovery and further homogenize the woodland flora of the Cove.
Over the course of the last 26 y, 46 species disappeared from our sample plots while only 10 new species were gained. During this interval, the stands examined were either in the understory reinitiation phase of stand development or beginning their transition out of the stem exclusion stage into the understory reinitiation phase (Webster et al., 2005b). The understory reinitiation phase is typically associated with an increase in the coverage and diversity of understory plant species as growing space is released through the death of canopy trees and changes in crown morphology (Peet and Christensen, 1988; Oliver and Larson, 1996). Therefore, we expected an increase in the diversity and cover of late-seral woodland herbs. However, the cover of this guild declined by nearly 50% (shade-tolerant herbs, Table 2), and several common woodland species (16 shade-tolerant and 13 intermediate tolerant species) were lost from the plots over the 26-year interval [e.g., Stylosanthes biflora (L.) BSP, Platanthera clavellata (Michaux) Luer and Prosartes lanuginosa (Michx.) D. Don]. We did not observe a similar loss of common woodland herbs in the mixed-conifer hardwood and acid cove reference plots, where mean plot species richness and percent cover have increased by nearly one- and two-fold, respectively since the 1970s. For example, Prosartes lanuginose, which was lost from the Cove plots, doubled in frequency on reference plots between the 1970s and 1995-2002 inventories. Similarly, Medeola virginiana L. which was recorded on three Cove plots during the 1970s, but was found only on one plot during the reinventory, increased in frequency from 36% to 45% on reference plots over approximately the same interval.
[FIGURE 2 OMITTED]
[FIGURE 3 OMITTED]
[FIGURE 4 OMITTED]
Forest herbs that disappeared from the Cades Cove plots typically were from families that contain preferred browse species and/or are disproportionately impacted by browsing (e.g., Liliaceae and Orchidaceae; Anderson, 1994; Augustine and Frelich, 1998; Fletcher et al., 2001; Webster et al., 2005a; Wiegmann and Waller, 2006). Several late blooming woodland aster species, as well as some shade-tolerant grasses, which typically are not browse sensitive (Wiegmann and Waller, 2006), also disappeared. While some later-seral species were gained, they tended to be on plots distal from the Cove and were usually only represented by a few specimens (e.g., Sanguinaria canadensis L. was found on only one plot, which was 343 m from the field/forest edge, where it had < 1% cover). Early-successional and old-field species that had persisted from earlier successional stages declined as expected.
A surprising result of our investigation was the decline in the number and cover of invasive exotic plants, such as Lonicera japonica Thunb. Over the sampling interval, the coverage of exotic plants declined by 50% and five species recorded during the 1970s were absent in 2004. Numerous studies have implicated invasive exotic plants in the simplification of native plant communities (Gordon, 1998; Levine et al., 2003; Qian and Ricklefs, 2006). The only exotic species to increase in coverage was the invasive grass Microstegium vimineum, which does not appear to be consumed by deer (Griggs et al., 2006). This species has been linked to changes in soil microbial communities and competitive exclusion of native species (Kourtev et al., 2002; Cole and Weltzin, 2004). While highly abundant in some Cove woodlots (Griggs et al., 2006; Webster et al., 2008), Microstegium vimineum was a minor species on our sample plots, and the aggregate cover of exotic species during the 2004 inventory was -4%. Consequently, exotic species do not appear to be driving species loss and homogenization on the sample plots examined in this study.
Our results are consistent with Rooney et al. (2004) who observed compositional erosion and floristic impoverishment of the herbaceous layer in areas of high deer abundance, that were otherwise protected from habitat degradation and loss. Herbaceous layer recovery from agricultural abandonment in the Cades Cove stands we examined has been largely arrested since stands have lost rather than gained late-seral species as they enter the understory reinitiation phase of stand development. Over the course of 26 y, the herbaceous layer in Cades Cove has become increasingly homogeneous, with three distinct forest types decreasing significantly in compositional dissimilarity. Conversely, reference plots across two distinct forest types have not decreased significantly in compositional dissimilarity. In both the Cades Cove and reference mixed conifer-hardwood plots, we observed significant declines in dissimilarity; however, the decline in Cades Cove was associated with species loss whereas the decline on reference plots was associated with species gain. Consequently, measures of homogenization should be viewed in light of other changes in the plant community, and some level of homogenization, especially if associated with the accrual of native late-seral species, may indicate recovery from past disturbance.
Across the three forest types examined, plant community composition was strongly associated with proximity to the Cove's edge environments. This relationship appears to represent a biotic edge effect, attributable, at least in part, to the foraging behavior of the Cove's resident deer population. Our evidence for this conclusion is as follows. First, all of these sample plots were within the bounds of the Cades Cove settlement area and were equally subjected to relatively intense anthropogenic disturbance prior to acquisition (Pyle, 1988). Second, none of the distance measures were significantly correlated with stand structure (Table 5), which has been shaped by past landuse (Webster et al., 2005b). Third, all sample plots were at least 30 m from the forest edge (Table 3), reducing the influence of abiotic edge effects. Consequently, within the range of distances examined, distance was not a proxy for anthropogenic disturbance intensity. Finally, similar patterns of species loss and cover decline were not observed on reference plots >3000 m from the edge of a Cove field. The magnitude of this effect and our ability to establish a threshold distance for it, however, were likely compromised by the fact that the Cove contained a substantial deer herd during the 1970s when the plots were established. Research conducted around that time documented that intense deer herbivory by the resident herd had already led to drastic reductions in the cover of many species (Bratton, 1979).
While the spatial dependence in plant community change may be attributable, at least in part, to a deer behavioral response to forest edges, given the spatial extent of the Cades Cove study area it is unlikely that plant community change was influenced by variation in deer density. The mean summer home-range of adult female white-tailed deer in the southern Appalachians ranges from 378 to 1100 m in diameter (McShea and Schwede, 1993; Campbell et al., 2004). Consequently, all of the distances examined in the Cades Cove plots (maximum distance from field edge, 894 m) could be nested within a single home-range. Reference plots, on the other hand, were well beyond the influence of the Cove's resident deer herd [>3 km; see also Bratton (1979)]. Deer are known to shift their activity patterns within their home-range in response to seasonally available pulses of forage (e.g., acorns and agricultural crops; Nixon et al., 1970; Murphy et al., 1985; Nixon et al., 1991; McShea and Schwede, 1993), hiding cover (Murphy et al., 1985) or changes in hunting pressure or predation risk (Martin and Baltzinger, 2002). In agricultural mosaics, habitat use generally shifts from woodlands in the fall and winter to fields and open habitats in the spring (Murphy et al., 1985). Consequently, forest herb populations near field edges would be expected to be more heavily utilized than those distal from edges.
Biotic edge effects associated with herbivore foraging behavior may result in either subtle or dramatic changes in vegetation. For example, meadow vole (Microtus pennsylvanicus) foraging along forest/field edges inhibits recruitment of some tree species within 30-40 m of an edge (Cadenasso and Pickett, 2000). Conversely, deer browsing may be more spatially variable along forest edges (Cadenasso and Pickett, 2000; Campbell et al., 2006) since deer favor edge environments (Montgomery, 1963), but use them in a rather coarse grained manner (VerCauteren and Hygnstrom, 1998). Edge effects associated with deer foraging have been proposed for forested regions in the Northern Lake States, with the implication that large contiguous blocks of forest are less susceptible to deer mediated tree regeneration failures (Alverson et al., 1988). However, as Augustine and deCalesta (2003) point out, large forested reserves in regions that lack large predators and harsh winters may be especially susceptible to severe deer browsing since natural mortality may be low in the absence of hunting. Consequently, the influence of forest fragmentation on deer abundance and impact may vary both locally and regionally (Milne et al., 1989; Campbell et al., 2006) with contrasting effects on local protected areas.
Long-term assessment of plant community composition provides valuable insights into the ecological consequences of chronic herbivory in forested ecosystems. As Morellet et al. (2007) point out, ungulate management across wide spatial scales requires a synthetic assessment of the ecological status of the ungulate-habitat system. Therefore, the contemporary health of a free-ranging ungulate population cannot be assessed in isolation of its habitat.
Acknowledgments.--We thank the National Park Service and the research staff at Great Smoky Mountains National Park, especially Keith Langdon, Kristine Johnson and Bill Stiver. Susan Bratton was integral in developing baseline studies in the 1970s. Hanni Muerdter, Jon Harrod, Mark Whited, Paul Allyn, Travis Martin, Art vonLehe, Meryl Rose, Heather MacCulloch and Catherine Harney assisted with field work and data entry, and Michael Kunze reformatted the 1977-1979 dataset. Mark Ford, Robert Scheller, Becky Nichols, Keith Langdon, Nancy Finley and two anonymous reviewers provided helpful comments on earlier drafts of the manuscript. Support for this project was provided by the School of Forest Resources and Environmental Science at Michigan Technological University. Internships for seasonal volunteers were provided through the Student Conservation Association.
SUBMITTED 4 JANUARY 2008
ACCEPTED 30 SEPTEMBER 2008
ALVERSON, W. S., D. M. WALLER AND S. L. SOLHEIM. 1988. Forests too deer: edge effects in northern Wisconsin. Conserv. Biol., 2:348-358.
ANDERSON, M. J. 2004. PERMDISP: a FORTRAN computer program for permutational analysis of multivariate dispersions (for any two-factor ANOVA design) using permutation tests. Department of Statistics, University of Auckland, New Zealand.
--, K. E. ELLINGSEN AND B. H. McARDLE. 2006. Multivariate dispersion as a measure of beta diversity. Ecol. Lett., 9:683-693.
ANDERSON, R. C. 1994. Height of white-flowered trillium (Trillium grandiflorum) as an index of deer browsing intensity. Ecol. Applic., 4:104-109.
--. 1997. Native pests: the impacts of deer in highly fragmented habitats, p. 117-134. In: M. W. Schwartz (ed.). Conservation in Highly Fragmented Landscapes. Chapman and Hall, New York.
AUGUSTINE, D. J. AND D. DECALESTA. 2003. Defining deer overabundance and threats to forest communities: from individual plants to landscape structure. Ecoscience, 10:472-486.
-- AND L. E. FRELICH. 1998. Effects of white-tailed deer on populations of an understory forb in fragmented deciduous forests. Conserv. Biol., 12:995-1004.
BAKER, T. T. AND D. H. VAN LEAR. 1998. Relations between density of rhododendron thickets and diversity of riparian forests. For. Ecol. Manag., 109:21-32.
BARRETT, M. A. ANN P. SAALING. 2006. Effects of Key deer herbivory on forest communities in the lower Florida Keys. Biol. Conserv., 129:100-108.
BRATTON, S. P. 1979. Impacts of white-tailed deer on the vegetation of Cades Cove, Great Smoky Mountains National Park. Proceed. Ann. Conf. Southeast. Assoc. Game Fish Commis., 33:305-312.
BROOKS, T. M., M. I. BAKARR, T. BOUCHER, G. A. B. DAFONSECA, C. HILTON-TAYLOR, J. M. HOEKSTRA, T. MORITZ, S. OLIVIERI, J. PARRISH, R. L. PRESSEY, A. S. L. RODRIGUES, W. SECHREST, A. STATTERSFIELD, W. STRAHM AND S. N. STUART. 2004. Coverage Provided by the Global Protected-Area System: Is It Enough? BioScience, 54:1081-1091.
CADENASSO, M. L. AND S. T. A. PICKETT. 2000. Linking forest edge structure to edge function: mediation of herbivore damage. J. Ecol., 88:31-44.
CAMPBELL, C. C. 1994. Birth of a national park in the Great Smoky Mountains, reprint edition. University of Tennessee Press, Knoxville, Tennessee. 184 p.
CAMPBELL, T. A., B. R. LASETER, W. M. FORD AND K. V. MILLER. 2004. Topographic home ranges of white-tailed deer in the central Appalachians. Southeast. Nat., 3:645-652.
--, --, W. M. FORD, R. H. ODOM AND K. V. MILLER. 2004. Feasibility of localized management to control white-tailed deer in forest regeneration areas. Wildl. Soc. Bull., 32:1124-1151.
--, --, --, --, AND --. 2006. Abiotic factors influencing deer browsing in West Virginia. North. J. Appl. For., 25:20-26.
CARMAN, J. B. 2001. Wildflowers of Tennessee. Highland Rim Press, Tullahoma, Tennessee. 427 p.
CLARKE, K. R., P.J. SOMERFELD AND M. G. CHAPMAN. 2006. On resemblance measures for ecological studies, including taxonomic dissimilarities and a zero-adjusted Bray-Curtis coefficient for denuded assemblages. J. Exper. Marine Biol. Ecol., 350:55-80.
COLE, P. G. AND J. F. WELTZIN. 2004. Environmental correlates of the distribution and abundance of Microstegium vimineum, in east Tennessee. Southeast. Nat., 3:545-562.
DUNN, D. 1988. Cades Cove: the life and death of a southern Appalachian Community 1818-1837. University of Tennessee Press, Knoxville, Tennessee.
ERVIN, J. 2004. Rapid assessment of protected area management effectiveness in four countries. Bioscience, 53:833-841.
FAUZI, A. AND E. A. BUCHARY. 2002. A socioeconomic perspective of environmental degradation at Kepulauan Seribu Marine National Park, Indonesia. Coast. Manag., 30:167-181.
FIELD, J. G., K. R. CLARKE AND R. M. WARWICK. 1982. A practical strategy for analyzing multispecies distribution patterns. Marine Ecol. Prog. Series, 8:37-52.
FLETCHER, J. D., L. A. SHIPLEY, W.J. MCSHEA AND D. L. SHUMWAY. 2001. Wildlife herbivory and rare plants: the effects of white-tailed deer, rodents, and insects on growth and survival of Turk's cap lily. Biol. Conserv., 101:229-238.
GORDON, D. R. 1998. Effects of invasive, non-indigenous plant species on ecosystem processes: lessons from Florida. Ecol. Applic., 8:975-989.
GRIGGS, J. A., J. H. ROCK, C. R. WEBSTER AND M. A. JENKINS. 2006. Vegetative legacy of a protected deer herd in Cades Cove, Great Smoky Mountains National Park. Nat. Areas Journ., 26:126-136.
GUTHERY, F. S. 1997. A philosophy of habitat management for northern bobwhites. J. Wildl. Manag., 61:291-301.
HESKE, E.J., S. R. ROBINSON AND J. D. BRAWN. 1999. Predator activity and predation on songbird nests on forest-field edges in east-central Illinois. Landscape Ecol., 14:345-354.
HORSLEY, S. B., S. L. STOUT AND D. S. dECALESTA. 2003. White-tailed deer impact on the vegetation dynamics of a northern hardwood forest. Ecol. Applic., 15:98-118.
JENKINS, M. A., C. R. WEBSTER AND J. H. ROCK. 2007. Effects of chronic herbivory and historic land use on population structure of a forest perennial, Trillium catesbaei. Appl. Veg. Sci., 10:441-450.
KARTESZ, J. T. 1999. A Synonymized Checklist and Atlas with Biological Attributes for the Vascular Flora of the United States, Canada, and Greenland, first ed., J. T. Kartesz and C. A. Meacham (eds.). Synthesis of the North American Flora, Version 1.0. North Carolina Botanical Garden, Chapel Hill, North Carolina.
KININGHAM, M.J. 1980. Density and Distribution of White-tailed Deer (Odocoileus virginianus) in Cades Cove, Great Smoky Mountains National Park. M.S. thesis, The University of Knoxville, Knoxville, Tennessee.
KNOX, W. M. 1997. Historical changes in the abundance and distribution of deer in Virginia, p. 27-36. In: W.J. McShea, H. B. Underwood and J. H. Rappole (eds.). The science of overabundance: deer ecology and population management. Smithsonian Institute Press, Washington, DC.
KOURTEV, P. S., J. G. EHRENFELD AND M. HADDBLOM. 2002. Exotic plant species alter the microbial community structure and function in the soil. Ecology, 83:3152-3166.
KRUSKAL, J. B. 1964. Nonmetric multidimensional scaling: a numerical method. Psychometrika, 29:115-129.
LEOPOLD, A. 1933. Game management. Charles Scribner's Sons, New York. 520 p.
LEVINE, J. M., M. VILA, C. M. D'ANTONIO, J. S. DUKES, K. GRIGULIS AND S. LAVOREL. 2003. Mechanisms underlying the impacts of exotic plant invasions. Proc. Royal Soc. London B, 270:775-781.
LIDICKER, W. Z., JR. 1999. Responses of mammals to habitat edges: an overview. Landscape Ecol., 14:333-343.
MARTIN, J. L. AND C. BALTZINGER. 2002. Interaction among deer browsing, hunting, and tree regeneration. Canad. J. For. Resear., 32:1254-1264.
MATHER, P. M. 1976. Computational methods of multivariate analysis in physical geography. John Wiley and Sons, London. 532 p.
McCUNE, B. AND J. B. GRACE. 2002. Analysis of ecological communities. MjM Software Design, Gleneden Beach, Oregon. 300 p.
-- AND M. J. MEFFORD. 1999. PC-ORD, Multivariate analysis of ecological data, Version 4.0. MjM Software Design, Gleneden Beach, Oregon. 237 p.
McSHEA, W.J. AND G. SCHWEDE. 1993. Variable acorn crops: response of white-tailed deer and other mast consumers. J. Mammal., 74:999-1006.
MILNE, B. T., K. M. JOHNSTON AND R. T. T. FORMAY. 1989. Scale-dependent proximity of wildlife habitat in a spatially-neutral Bayesian model. Landscape Ecol., 2:101-110.
MONTGOMERY, G. G. 1963. Nocturnal movements and activity rhythms of white-tailed deer. J. Wildl. Manag., 27:422-427.
MORELLET, N., J.-M. GAILLARD, A. J. M. HEWISON, P. BALLON, Y. BOSCARDIN, P. DUNCAN, F. KLEIN AND D. MAILLARS. 2007. Indicators of ecological change: new tools for managing populations of large herbivores. J. Appl. Ecol., 44:634-643.
MURPHY, R. K., N. F. PAYNE AND R. K. ANDERSON. 1985. White-tailed deer use of an irrigated agricultural-grassland complex in central Wisconsin. J. Wildl. Manag., 49:125-128.
NATURAL RESOURCES CONSERVATION SERVICE. 2007. United States Department of Agriculture. Official Soil Series Descriptions [Online WWW]. Available URL: http://soils.usda.gov/technical/classification/ osd/index.hmal [Accessed 13 Apr. 2007].
NIXON, C. M., M. W. MCCLAIN AND K. R. RUSSELL. 1970. Deer food habits and range characteristics in Ohio. J. Wildl. Manag., 34:870-886.
--, L. P. HANSEN, P. A. BREWER AND J. E. CHELSVIG. 1991. Ecology of white-tailed deer in an intensively farmed region of Illinois. Wildl. Monog., 118:1-77.
OLIVER, C. D. AND B. C. LARSON. 1996. Forest stand dynamics, update edition. John Wiley and Sons, New York. 520 p.
PEET, R. K. AND N. L. CHRISTENSEN. 1988. Changes in species diversity during secondary forest succession on the North Carolina Piedmont, p. 233-245. In: H. J. During, M. J. A. Werner and J. H. Williams (eds.). Diversity and Pattern in Plant Communities. SPB Academic Publishing, Hague.
--, T. R. WENTWORTH AND P. S. WHITE. 1998. A flexible multipurpose method for recording vegetation composition and structure. Castanea, 63:262-274.
PORTER, W. F. AND B. UNDERWOOD. 1999. Of elephants and blind men: deer management in the U.S. National Parks. Ecol. Appl., 9:3-9.
PYLE, C. 1988. The type and extent of anthropogenic vegetation disturbance in the Great Smoky Mountains before National Park Service acquisition. Castanea, 55:225-235.
QIAN, H. AND R. E. RICKLEFS. 2006. The role of exotic species in homogenizing the North American flora. Ecol. Lett., 9:1293-1298.
ROONEY, T. P. AND W. J. DRESS. 1997. Species loss over sixty-six years in the ground-layer vegetation of Heart's Content, an old-growth forest in Pennsylvania, USA. Nat. Areas Journ., 17: 297-305.
--, S. M. WEIGMANN, D. A. ROGERS AND D. M. WALLER. 2004. Biotic impoverishment and homogenization in unfragmented forest understory communities. Consero. Biol., 18:787-798.
ROSSELL, C. R., B. GORSIRA AND S. PATCH. 2005. Effects of white-tailed deer on vegetation structure and woody seedling composition in three forest types on the Piedmont Plateau. For. Ecol. Manag., 210:415-424.
SAUNDERS, D. A., R. J. HOBBS AND C. R. MARGULES. 1991. Biological consequences of ecosystem fragmentation: a review. Conserv. Biol., 5:18-32.
SOUKUP, M., M. K. FOLEY, R. HEIBERT AND D. E. HUFF. 1999. Wildlife management in US national parks: natural regulation revisited. Ecol. Applic., 9:1-36.
SOUTHWORTH, S., A. SCHULTZ AND D. DENENNY. 2005. Geologic Map of the Great Smoky Mountains National Park Region, Tennessee and North Carolina. U.S. Geological Survey Open File Report 2005-1225, scale 1:100,000.
AND J. TRIPLETT. 2003. Surficial Geologic Map of the Great Smoky Mountains National Park Region, Tennessee and North Carolina. U.S. Geological Survey Open File Report 03-381, scale 1:100,000.
TREMBLAY, , J.-P., J. HUOT AND F. POTVIN. 2007. Density-related effects of deer browsing on the regeneration dynamics of boreal forests. J. Appl. Ecol., 44:552-562.
TURNER, K. AND K. BEAZLEY. 2004. An exploration of issues and values inherent in the concept of ecological integrity. Environments, 32:43-64.
USDA, NRCS. 2004. The PLANTS Database, Version 3.5. National Plant Data Center, Baton Rouge (http://plants.usda.gov)
VERCAUTEREN, K. C. AND S. E. HYGNSTROM. 1998. Effects of agricultural activities and hunting on home ranges of female white-tailed deer. J. Wildl. Manag., 62:280-285.
WATHEN, W. G. ANDJ. C. NEW. 1989. The white-tailed deer of Cades Cove: population status, movements, and survey of infectious disease. U.S. Department of the Interior, National Park Service, Research/Resources Management Report SER-89/01. Southeast Regional Office, Atlanta, Georgia.
WEBSTER, C. R., M. A. JENKINS ANN J. H. ROCK. 2005a. Long-term response of spring flora to chronic herbivory and deer exclusion in Great Smoky Mountains National Park, USA. Biol. Conserv., 125:297-307.
--, -- AND --. 2005b. Twenty years of forest change in the woodlots of Cades Cove, Great Smoky Mountains National Park. J. Torrey Bot. Soc., 132:280-292.
--, J. H. ROCK, R. E. FROESE AND M. A. JENKINS. 2008. Drought-herbivory interaction disrupts competitive displacement of native plants by Microstegiura vimineum, 10-year results. Oecologia, 157:497-508.
WEISBERG, P.J. AND M. B. COUGHENOUR. 2003. Model-based assessment of aspen responses to elk herbivory in Rocky Mountain National Park, USA. Environ. Manag., 32:152-169.
WHITE, P. S. AND R. T. BUSING. 1993. LTERM: Long-term monitoring and research in Great Smoky Mountains National Park, vegetation monitoring and an assessment of past studies. Technical Report NPS/sergrsm/nrtr-93/10.
WHITE, R. D., K. D. PATTERSON, A. WRAKLEY, C.J. ULREY AND J. DRAKE. 2003. Vegetation classification of Great Smoky Mountains National Park. Reports submitted to BRD-NPS Vegetation Mapping Program. NatureServe, Durham, North Carolina.
WIEGMANN, S. M. AND D. M. WALLER. 2006. Fifty years of change in northern upland forest understories: identity and traits of "winner" and loser" plant species. Biol. Conserv., 129:109-123.
ZAR, J. H. 1999. Biostatistical analysis, fourth ed. Prentice Hall, Inc, Upper Saddle River, New Jersey. 929 p.
JENNIFER A. THIEMANN AND CHRISTOPHER R. WEBSTER (1)
School of Forest Resources and Environmental Science, Michigan Technological University, 1400 Townsend Drive, Houghton, Michigan 49931
MICHAEL A. JENKINS
Department of Forestry and Natural Resources, Purdue University, 715 West State Street, West Lafayette, Indiana 47907
PETER M. HURLEY (2)
School of Forest Resources and Environmental Science, Michigan Technological University, 1400 Townsend Drive, Houghton, Michigan 49931
JANET H. ROCK
Inventory and Monitoring Program, Twin Creeks Science and Education Center, Great Smoky Mountains National Park, 1316 Cherokee Orchard Rd, Gatlinburg, Tennessee 37738
PETER S. WHITE
Department of Biology, Campus Box 3280, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina 27599
(1) Corresponding author: Telephone: 906-487-3618; FAX: 906487-2915; e-mail: firstname.lastname@example.org
(2) Current address: Department of Biology, School of Arts and Sciences, Saint Francis University, 2701 Spring Street, Fort Wayne, Indiana 46808
TABLE: 1.--Herbaceous species lost from Cades Cove sample plots between 1977/79 and 2004 as well as species that were first recorded in 2004 (gained). Non-native species are indicated by an asterisk Scientific name Common name Amorpha glabra Desf. ex Poir. mountain false indigo Andropogon virginicus L. broom sedge Arnoglossum atriplicifolium (L.) pale Indian plantain H.E. Robins. Asplenium platyneuron (L.) B.S.P. ebony spleenwort Bidens frondosa L. devil's pitchfork Brachyelytrum erectum (Schreb. ex bearded shorthusk Spreng.) Beauv. Calamagrostis coarctata (Torn.) Eat. Nuttall's reed grass Campanulastrum americanum (L.) Small American bellflower Cerastium fontanum Baumg. * mouse-eared chickweed Chaerophyllum tainturieri Hook. hairyfruit chervil Chamaecrista nictitans (L.) Moench partridge pea Chrysopsis mariana (L.) Ell. Maryland golden aster Chrysosplenium americanum Schwein. American golden ex Hook. saxifrage Danthonia compressa Austin ex Peck flattened wild oat grass Dichanthelium boscii (Poir.) Gould Bosc's rosette grass & C.A. Clark Dichanthelium dichotomum (L.) Gould cypress witchgrass Dichanthelium sabulorum (Lam.) Gould hemlock rosette grass & C.A. Clark Prosartes lanuginosa (Michx.) D. Don yellow fairybells Eleocharis obtusa (Wind.) J.A. Schultes blunt spike rush Euphorbia corollata L. flowering spurge Eurybia surculosa (Michx.) Nesom creeping aster Geum canadense Jacq. white avens Hieracium venosum L. rattlesnake weed Houstonia purpurea L. Large blunt Hypericum hypericoides (L.) Crantz St. Andrew's cross Juncus tennis Willd. poverty rush Leersia virginica Willd. cut grass Lindernia dubia (L.) Pennell yellow-seed false pimpernil Lobelia truberula Michx. downy lobelia Melilotus officinalis (L.) Lam. * yellow sweet clover Mentha X piperita L. (pro sp.) peppermint [aquatics X spicata] * Muhlenbergia schreberi J.F. Gmel. nimblewill Oxalis montana Raf. mountain woodsorrel Oxalis stricta L. yellow wood sorrel Panicum anceps Michx. beaked panicgrass Parthenocissus quinquefolia (L.) Planch. Virginia creeper Paspalum laeve Michx. field crown grass Phytolacca americana L. American pokeweed Platanthera clavellata (Michx.) Luer small green wood orchid Polygonatum pubescens (Wind.) Pursh hairy Solomon's seal Polygonum caespitosum Blume * Oriental lady's thumb Prunella vulgaris L. * selfheal Pycnanthemum muticum (Michx.) Pers. mountain mint Rumex acetosella L. * common sheep sorrel Sanguinaria canadensis L. bloodroot Sanicula marilandica L. Maryland black snakeroot Sanicula smallii Bickn. small's black snakeroot Schizachyrium scoparium (Michx.) Nash little bluestem Sericocarpus asteroides (L.) B.S.P. toothed white-topped aster Solanum carolinense L. Carolina horse nettle Solidago curtisii Torr. & Gray Curtis' goldenrod Solidago puberula Nutt. downy goldenrod Solidago rugosa P. Mill. wrinkle-leaf goldenrod Sorghastrum nutans (L.) Nash Indian grass Stylosanthes bilora (L.) B.S.P. side-beak pencil flower Symphyotrichum lateriflorum calico aster (L.) A.& D. Love Scientific name Lost Gained Amorpha glabra Desf. ex Poir. x Andropogon virginicus L. x Arnoglossum atriplicifolium (L.) x H.E. Robins. Asplenium platyneuron (L.) B.S.P. x Bidens frondosa L. x Brachyelytrum erectum (Schreb. ex x Spreng.) Beauv. Calamagrostis coarctata (Torn.) Eat. x Campanulastrum americanum (L.) Small x Cerastium fontanum Baumg. * x Chaerophyllum tainturieri Hook. x Chamaecrista nictitans (L.) Moench x Chrysopsis mariana (L.) Ell. x Chrysosplenium americanum Schwein. x ex Hook. Danthonia compressa Austin ex Peck x Dichanthelium boscii (Poir.) Gould x & C.A. Clark Dichanthelium dichotomum (L.) Gould x Dichanthelium sabulorum (Lam.) Gould x & C.A. Clark Prosartes lanuginosa (Michx.) D. Don x Eleocharis obtusa (Wind.) J.A. Schultes x Euphorbia corollata L. x Eurybia surculosa (Michx.) Nesom x Geum canadense Jacq. x Hieracium venosum L. x Houstonia purpurea L. x Hypericum hypericoides (L.) Crantz x Juncus tennis Willd. x Leersia virginica Willd. x Lindernia dubia (L.) Pennell x Lobelia truberula Michx. x Melilotus officinalis (L.) Lam. * x Mentha X piperita L. (pro sp.) x [aquatics X spicata] * Muhlenbergia schreberi J.F. Gmel. x Oxalis montana Raf. x Oxalis stricta L. x Panicum anceps Michx. x Parthenocissus quinquefolia (L.) Planch. x Paspalum laeve Michx. x Phytolacca americana L. x Platanthera clavellata (Michx.) Luer x Polygonatum pubescens (Wind.) Pursh x Polygonum caespitosum Blume * x Prunella vulgaris L. * x Pycnanthemum muticum (Michx.) Pers. x Rumex acetosella L. * x Sanguinaria canadensis L. x Sanicula marilandica L. x Sanicula smallii Bickn. x Schizachyrium scoparium (Michx.) Nash x Sericocarpus asteroides (L.) B.S.P. x Solanum carolinense L. x Solidago curtisii Torr. & Gray x Solidago puberula Nutt. x Solidago rugosa P. Mill. x Sorghastrum nutans (L.) Nash x Stylosanthes bilora (L.) B.S.P. x Symphyotrichum lateriflorum x (L.) A.& D. Love Shade-tolerance Scientific name rank ([double dagger]) Amorpha glabra Desf. ex Poir. Intermediate Andropogon virginicus L. Intolerant Arnoglossum atriplicifolium (L.) Intolerant H.E. Robins. Asplenium platyneuron (L.) B.S.P. Tolerant Bidens frondosa L. Intermediate Brachyelytrum erectum (Schreb. ex Tolerant Spreng.) Beauv. Calamagrostis coarctata (Torn.) Eat. Intolerant Campanulastrum americanum (L.) Small Tolerant Cerastium fontanum Baumg. * Intolerant Chaerophyllum tainturieri Hook. Intolerant Chamaecrista nictitans (L.) Moench Intolerant Chrysopsis mariana (L.) Ell. Intolerant Chrysosplenium americanum Schwein. Tolerant ex Hook. Danthonia compressa Austin ex Peck Tolerant Dichanthelium boscii (Poir.) Gould Intermediate & C.A. Clark Dichanthelium dichotomum (L.) Gould Tolerant Dichanthelium sabulorum (Lam.) Gould Intermediate & C.A. Clark Prosartes lanuginosa (Michx.) D. Don Tolerant Eleocharis obtusa (Wind.) J.A. Schultes Intolerant Euphorbia corollata L. Intermediate Eurybia surculosa (Michx.) Nesom Intolerant Geum canadense Jacq. Tolerant Hieracium venosum L. Intermediate Houstonia purpurea L. Intermediate Hypericum hypericoides (L.) Crantz Intermediate Juncus tennis Willd. Intermediate Leersia virginica Willd. Tolerant Lindernia dubia (L.) Pennell Intermediate Lobelia truberula Michx. Intermediate Melilotus officinalis (L.) Lam. * Intolerant Mentha X piperita L. (pro sp.) Intolerant [aquatics X spicata] * Muhlenbergia schreberi J.F. Gmel. Tolerant Oxalis montana Raf. Tolerant Oxalis stricta L. Tolerant Panicum anceps Michx. Intermediate Parthenocissus quinquefolia (L.) Planch. Tolerant Paspalum laeve Michx. Intolerant Phytolacca americana L. Tolerant Platanthera clavellata (Michx.) Luer Tolerant Polygonatum pubescens (Wind.) Pursh Tolerant Polygonum caespitosum Blume * Tolerant Prunella vulgaris L. * Intermediate Pycnanthemum muticum (Michx.) Pers. Tolerant Rumex acetosella L. * Intolerant Sanguinaria canadensis L. Tolerant Sanicula marilandica L. Tolerant Sanicula smallii Bickn. Tolerant Schizachyrium scoparium (Michx.) Nash Intolerant Sericocarpus asteroides (L.) B.S.P. Tolerant Solanum carolinense L. Intolerant Solidago curtisii Torr. & Gray Tolerant Solidago puberula Nutt. Intolerant Solidago rugosa P. Mill. Intermediate Sorghastrum nutans (L.) Nash Intolerant Stylosanthes bilora (L.) B.S.P. Tolerant Symphyotrichum lateriflorum Intermediate (L.) A.& D. Love ([double dagger]) Shade-tolerance rank as reported in the USDA PLANTS Database (2004). If this trait was not listed in the database, then rank was based on the habitat description in an applicable field guide (e.g., Carman 2001) as follows: Tolerant = woodland or forest; Intermediate =open woods; Intolerant = old field, waste places, or prairie TABLE 2.--Absolute and relative percent cover of understory species on sample plots in Cades Cove grouped by life history traits (mean [+ or -] 1 SE) Absolute cover (%) 1977/79 2004 Growth form Low-growing ([double dagger]) 2.6 [+ or -] 1.2 2.1 [+ or -] 1.0 Bunch ([paragraph]) 11.6 [+ or -] 6.2 4.2 [+ or -] 3.5 Single crown ([section]) 2.5 [+ or -] 1.3 0.5 [+ or -] 0.3 Multiple 0.2 [+ or -] 0.1 0.1 [+ or -] 0.04 crowns ([pounds sterling]) Shade tolerance Intolerant 3.6 [+ or -] 2.1 0.2 [+ or -] 0.1 Intermediate 1.5 [+ or -] 0.9 0.5 [+ or -] 0.2 Tolerant 11.8 [+ or -] 6.6 6.0 [+ or -] 3.6 Relative cover (% of total) 1977/79 2004 Growth form Low-growing ([double dagger]) 50.2 [+ or -] 7.8 74.7 [+ or -] 7.0 Bunch ([paragraph]) 23.7 [+ or -] 7.0 17.6 [+ or -] 6.8 Single crown ([section]) 19.5 [+ or -] 4.6 6.8 [+ or -] 2.1 Multiple 6.6 [+ or -] 3.5 0.9 [+ or -] 0.4 crowns ([pounds sterling]) Shade tolerance Intolerant 14.3 [+ or -] 4.4 5.06 [+ or -] 2.2 Intermediate 20.2 [+ or -] 3.8 15.2 [+ or -] 4.0 Tolerant 65.5 [+ or -] 5.9 79.8 [+ or -] 4.6 ([double dagger]) Low-growing includes: "prostrate," "creeping," and plants that have low basal leaves at maturity (e.g. Viola spp.) ([paragraph]) Bunch species were mostly graminoids ([section]) Single crown includes plants with a single apical meristem/crown ([pounds sterling]) Multiple crowns refer to plants with more than one apical meristem, shrub-like TABLE 3.--Attributes of Cades Cove (n = 19) and reference plots (n = 11) Variable Mean ([+ or -] 1 SE) Median Min Cades Cove Mean overstory 83.5 [+ or -] 5.6 71.0 53.0 tree age (yr) Distance from 279.1 [+ or -] 66.1 144.0 33.0 field edge (m) Distance from road (m) 312.4 [+ or -] 67.6 208.0 42.0 Basal area 1977/79 32.5 [+ or -] 2.4 34.62 6.21 ([m.sup.2] [ha.sup.-1]) Basal area 2000 * 37.7 [+ or -] 2.3 37.8 10.2 ([m.sup.2] [ha.sup.-1]) Stem density 1977/79 617.9 [+ or -] 43.5 590.0 300.0 (trees [ha.sup.-1]) Stem density 2000 (trees 617.4 [+ or -] 48.7 600.0 340.0 [ha.sup.-1]) Reference Distance from field edge 6113 [+ or -] 1195 4400 1090 (m) ([dagger]) Basal area 1977/79 35.2 [+ or -] 2.2 34.2 24.3 ([m.sup.2] [ha.sup.-1]) Basal area 2000 35.3 [+ or -] 2.2 35.4 22.8 ([m.sup.2] [ha.sup.-1]) Stem density 1977/79 707 [+ or -] 119 670 170 (trees [ha.sup.-1]) Stem density 2000 (trees 696 [+ or -] 102 720 120 [ha.sup.-1]) Variable Max Cades Cove Mean overstory 135.0 tree age (yr) Distance from 894.0 field edge (m) Distance from road (m) 995.0 Basal area 1977/79 55.4 ([m.sup.2] [ha.sup.-1]) Basal area 2000 * 53.0 ([m.sup.2] [ha.sup.-1]) Stem density 1977/79 1040.0 (trees [ha.sup.-1]) Stem density 2000 (trees 1030.0 [ha.sup.-1]) Reference Distance from field edge 13,302 (m) ([dagger]) Basal area 1977/79 50.2 ([m.sup.2] [ha.sup.-1]) Basal area 2000 46.2 ([m.sup.2] [ha.sup.-1]) Stem density 1977/79 1670 (trees [ha.sup.-1]) Stem density 2000 (trees 1440 [ha.sup.-1]) Note: Overstory attributes are for trees [greater than or equal to] 10 cm diameter at breast height (1.37 m) * Overstory data were collected between 1995 and 2002 ([dagger]) All plots were >3000 m from the edge of a field in Cades Cove; shorter distances were associated with other clearings distal from the Cove TABLE 4.--Pearson correlations between NMS ordination axes and environmental and stand structure variables for sample plots in Cades Cove NMS antis: 1 2 Variable r [r.sup.2] r [r.sup.2] Tree age 0.325 0.106 0.045 0.002 Basal area 1977/79 0.074 0.005 0.073 0.005 Density 1977/79 -0.097 0.009 0.349 0.122 Basal area 2000 * -0.308 0.095 -0.303 0.092 Density 2000 -0.402 0.162 -0.027 0.001 Change in basal area -0.375 0.141 -0.370 0.137 Change in density -0.278 0.078 -0.299 0.090 Distance to edge 0.386 0.149 -0.132 0.017 Distance to road 0.199 0.040 -0.143 0.021 3 Variable r [r.sup.2] Tree age -0.169 0.029 Basal area 1977/79 -0.153 0.023 Density 1977/79 -0.344 0.118 Basal area 2000 * -0.046 0.002 Density 2000 -0.245 0.060 Change in basal area 0.107 0.011 Change in density 0.055 0.003 Distance to edge -0.420 0.177 Distance to road -0.400 0.160 * Overstory data were collected between 1995 and 2002 TABLE 5.--Pearson correlation matrix of environmental and stand structure variables (correlation coefficient followed by P value) for Cades Cove sample plots Distance from Mean overstory Field tree age (yr) edge Road Distance from field 0.287 edge (m) 0.233 Distance from road (m) 0.210 0.942 0.388 0.000 Basal area 1977/78 -0.079 -0.037 -0.094 ([m.sup.2] [ha.sup.1]) 0.749 0.881 0.701 Basal area 2000 0.045 0.212 0.255 ([m.sup.2] [ha.sup.1]) 0.854 0.384 0.291 [DELTA] Basal area 0.123 0.244 0.344 ([m.sup.2] [ha.sup.1]) 0.617 0.313 0.149 Stem density 1977/78 (trees -0.378 -0.028 0.033 [ha.sup.1] 0.111 0.911 0.892 Stem density 2000 -0.348 0.232 0.425 (trees [ha.sup.1]) 0.145 0.340 0.070 4 Stem density -0.009 0.226 0.349 (trees [ha.sup.1]) 0.972 0.352 0.144 Basal area 1977/79 2000 [DELTA] Distance from field edge (m) Distance from road (m) Basal area 1977/78 ([m.sup.2] [ha.sup.1]) Basal area 2000 0.487 ([m.sup.2] [ha.sup.1]) 0.034 [DELTA] Basal area -0.518 0.494 ([m.sup.2] [ha.sup.1]) 0.023 0.031 Stem density 1977/78 (trees 0.478 0.052 -0.425 [ha.sup.1] 0.038 0.833 0.070 Stem density 2000 -0.212 0.280 0.485 (trees [ha.sup.1]) 0.383 0.246 0.035 4 Stem density -0.565 0.206 0.764 (trees [ha.sup.1]) 0.012 0.397 0.000 Stem density 1977/79 2000 Distance from field edge (m) Distance from road (m) Basal area 1977/78 ([m.sup.2] [ha.sup.1]) Basal area 2000 ([m.sup.2] [ha.sup.1]) [DELTA] Basal area ([m.sup.2] [ha.sup.1]) Stem density 1977/78 (trees [ha.sup.1] Stem density 2000 0.289 (trees [ha.sup.1]) 0.230 4 Stem density -0.534 0.655 (trees [ha.sup.1]) 0.019 0.002 * Overstory data were collected between 1995 and 2002 TABLE 6.--Average distance to centroid ([+ or -] 1 se), based on square-root transformed Bray-Curds dissimilarity values, for each habitat type and sampling interval (1977-1979 and 2004) Distance to centroid 1970s 2004 Cades Cove All Plots 59.06 [+ or -] 1.76 50.74 [+ or -] 2.79 Acid Cove 38.43 [+ or -] 3.14 39.73 [+ or -] 5.43 Successional Pine 53.14 [+ or -] 3.55 43.63 [+ or -] 5.57 Mixed Conifer-Hardwoods 58.81 [+ or -] 1.44 50.80 [+ or -] 2.04 Reference All Plots 50.42 [+ or -] 1.79 47.87 [+ or -] 2.74 Acid Cove 48.32 [+ or -] 1.52 50.46 [+ or -] 1.62 Mixed Conifer-Hardwoods 43.36 [+ or -] 3.28 34.56 [+ or -] 2.76 n F ([dagger]) P Cades Cove 19 6.352 0.02 All Plots 4 0.043 0.87 Acid Cove 7 2.075 0.18 Successional Pine 8 10.248 0.01 Mixed Conifer-Hardwoods Reference 11 0.605 0.44 All Plots 6 0.926 0.35 Acid Cove 5 4.227 0.07 Mixed Conifer-Hardwoods * Reference plots were resampled between 1995 and 2001 ([dagger]) F statistic for permutational ANOVA on PCoA scores TABLE 7.--Herbaceous-layer attributes of reference plots (n = 11) Variable Mean ([+ or -] 1 SE) Median % Cover (1977/79) 7.0 [+ or -] 2.5 4.3 % Cover (1995/2000) 11.5 [+ or -] 6.1 5.2 [DELTA] % Cover (%) 183 [+ or -] 76 41 Richness (1977/79) 14.7 [+ or -] 3.3 13 Richness (1995/2000) 24.6 [+ or -] 5.1 24 [DELTA] Richness (%) 106 [+ or -] 28 85 Variable Min Max % Cover (1977/79) 0.08 23.8 % Cover (1995/2000) 0.43 68.2 [DELTA] % Cover (%) -45 ([double dagger]) 688 Richness (1977/79) 1 36 Richness (1995/2000) 4 52 [DELTA] Richness (%) -27 ([paragraph]) 300 [double dagger]) Four plots declined in percent cover (mean decline = 32%); however, none of those plots exhibited a net loss of species, three of which experienced net a gain in species richness ([paragraph]) Only one plot displayed a net loss of species, the majority of which were early successional
|Printer friendly Cite/link Email Feedback|
|Author:||Thiemann, Jennifer A.; Webster, Christopher R.; Jenkins, Michael A.; Hurley, Peter M.; Rock, Janet H|
|Publication:||The American Midland Naturalist|
|Date:||Jul 1, 2009|
|Previous Article:||Spatial and temporal patterns of roost use by western long-eared bats (Myotis evotis).|
|Next Article:||Spatial variation in top-down direct and indirect effects on white oak.|