Northeastern Naturalist Vol. 21, No. 3 T.P. Diggins and R.G. Catterlin 2014 337 2014 NORTHEASTERN NATURALIST 21(3):337–350 Topographic Patterns in Forest Composition and Diversity on Slopes of Zoar Valley Canyon, Western New York Thomas P. Diggins1,* and Richard G. Catterlin1 Abstract - We explored topographic patterns in forest composition and diversity on unlogged slopes along ≈3 km of Zoar Valley Canyon, a 150-m deep east–west gorge in western New York. We catalogued all trees along 3 north- and 3 south-facing 20-m wide vertical belt transects on slopes of up to 50°. North-facing exposures were the more mesic, and were dominated by Acer saccharum (Sugar Maple) on the lower slopes and by Tsuga canadensis (Eastern Hemlock) on the upper slopes. South-facing slopes were more xeric and displayed distinct upper- and lower-slope species assemblages. Lower slope-forest composition was generally similar to that across the canyon, but upper slopes and ridges supported sparse and stunted Quercus prinus (Chestnut Oak), Quercus rubra (Northern Red Oak), Pinus resinosa (Red Pine), and Pinus strobus (Eastern White Pine). Canopy trees on upper slopes were typically shorter (less than 6 m vs. 30 m) and had smaller diameter at breast height (40 cm vs. 80 cm) than those on lower slopes. However, some upper-slope trees exceeded 165 years of age. Results of non-metric multidimensional scaling ordination distinguished between north- and south-facing upper slopes, and showed a broad similarity among forests on lower slopes of both aspects and fluvial terraces at slope bases. The elevationally diverse southfacing slopes contributed more to site-wide (i.e., slopes and terraces) species richness (34) and gamma diversity (Shannon H' = 2.462), than did north-facing slopes. Introduction Associations between plant communities and topographic factors including elevation, slope aspect, and slope angle have long interested ecologists, and were the focus of some seminal monographs of the early and mid-20th century (e.g., Braun 1935, Cantlon 1953, Mowbray and Oosting 1968, Nichols 1914). Recent quantitative studies of vegetation patterns and processes in eastern North American forests have also considered topographic variables (Hwang et al. 2011, Lewin 1974, Murphy and McCarthy 2012, Zimmerman and Runkle 2010). Topographic gradients can be particularly influential at or near the elevational, latitudinal, and climatic limits of dominant species (Cottle 1932, Goldblum and Rigg 2003, Ziegler 1995), but they can also shape community composition well within species’ ranges and under otherwise typical conditions (Fekedulegn et al. 2003, Fralish 1994, Searcy et al. 2003). Topographic heterogeneity often enhances vegetational diversity across multiple scales, and can yield rich mosaics of varied forest types even within relatively small geographic ranges. Braun’s (1935) description of Pine Mountain, KY offers a premier example of plant diversity enhanced by topographic and geologic gradients. 1Department of Biological Sciences, Youngstown State University, One University Plaza, Youngstown, OH 44555. *Corresponding author - tpdiggins@ysu.edu. Manuscript Editor: Elizabeth Hane Northeastern Naturalist 338 T.P. Diggins and R.G. Catterlin 2014 Vol. 21, No. 3 We conducted the present study along the Zoar Valley Canyon of 5th–6th-order Cattaraugus Creek in western New York (Figs. 1, 2), where diverse and very minimally disturbed Tsuga canadensis (Eastern Hemlock)-northern hardwood and mixed deciduous forests line both slopes of an east–west-trending gorge at the margin of the Allegheny Plateau (Hunt et al. 2002). Topgraphic patterns there reflect only modest geographic and geologic gradients because the gorge is only ~500 m wide, ~150 m deep, and its entire linear and vertical extent comprises a single formation of Devonian shales of the Canadaway group (Hunt et al. 2002). However, these features likely provide conditions that contribute to biodiversity and conservation values of the site—values that are already among the region’s highest based on the riparian corridor alone (Diggins 2013, Diggins and Kershner 2005, Pfeil et al. 2007). The objectives of the present study were to: 1) determine species composition and diversity on Zoar Valley’s canyon slopes in relation to slope aspect, elevation, and steepness (Catterlin 2010); and 2) evaluate the contribution of these heretofore unexplored stands to site-wide gamma diversity, which also includes previously studied fluvial terrace-forests below the slopes (see Diggins 2013). We also described semi-xeric stunted elfin Quercus (oak)-Pinus (pine) woodlands (Fig. 3a) on south-facing upper slopes that appear highly divergent from the mesic forest types that characterize the rest of the canyon and the broader western New York region (Gordon 1940, Wang 2007). Field Site Description The Zoar Valley Canyon (N42º26', W78º52') encompasses 11 km of the east– west Main Branch (the subject of the present study) and 8 km of the north–south South Branch of Cattaraugus Creek, the largest tributary to eastern Lake Erie. The ~150-ha study area (Fig. 1) is located entirely within the 620-ha state-owned Zoar Valley Unique Area, which gained state constitutional conservation status in 2007. Main Branch canyon depths are 60–150 m and streambed elevations 266–300 m. Nearby hills rise to >500 m. The Canadaway shales are very friable where exposed, although some less erodible strata create narrow shelves that can support vegetation on even the steepest cliffs. Zoar Valley is by far the most intact ecosystem within the Lake Erie Gorges province of western New York, northwestern Pennsylvania, and northeastern Ohio (Hunt et al. 2002), where a series of canyons dissect the 300-m drop from the Allegheny Plateau to the Lake Erie Plain. Diggins and Kershner (2005), Pfeil et al. (2007), and Diggins (2013) described extensive and potentially unlogged hardwood-dominated old growth forests within Zoar Valley’s riparian zone, with stand ages often exceeding 200 years; they suspected that the canyon slopes on both sides of the Main Branch (Fig. 2) were also old growth. Stunted Quercus prinus (Chestnut Oak) on south-facing upper slopes and moderate-sized Eastern Hemlock on north-facing upper slopes were non-marketable timber and were likely spared from harvest (Figs. 3a, b), but several lines of evidence suggest that exemplary timber stands on lower slopes may also be unlogged. Increment coring has yielded breast-height ages up to 410 years (Eastern Hemlock), and canopy dominants of Northeastern Naturalist Vol. 21, No. 3 T.P. Diggins and R.G. Catterlin 2014 339 multiple species display qualitative features consistent with ages of 150–250 years on adjacent fluvial terraces (see Diggins 2013). Large and well-formed individuals of valuable timber species such as Fraxinus americana (White Ash) and Prunus serotina (Black Cherry) are present but generally scattered, suggesting that they are gap colonizers rather than disturbance-generated cohorts. There are no stumps that might indicate logging, and the few multi-stem coppices present are mostly Tilia americana (American Basswood), which is naturally multi-stemmed. Finally, most lower slopes are near fluvial upper terraces that are closer to the river and where old growth has been quantified previously (Diggins 2013, Diggins and Kershner 2005, Pfeil et al. 2007). Methods During 2008–2009, we established three 20-m-wide vertical belt transects on both north- and south-facing slopes (a total of 6 transects) on inner bends of Zoar Valley Canyon where scalable slopes of <50º supported contiguous forest and allowed access with caution. Transects are denoted, along with adjacent fluvial terraces, in Fig. 1. Rappelling and/or climbing narrower but more numerous transects was not an option due to the instability of the shale strata, and the strict prohibition against climbing of any kind within the study area. Because there were no obvious, consistent geomorphic breaks or thresholds along Figure 1. Location of Zoar Valley, Cattaraugus Creek, in western New York, with locations and orientations of north-facing (N-1–N-3) and south-facing (S-1–S-3) belt transects, and upper and lower fluvial terraces, along the Main Branch Canyon. Flow is east to west. Satellite image was taken in 2008. Northeastern Naturalist 340 T.P. Diggins and R.G. Catterlin 2014 Vol. 21, No. 3 transects, we defined the boundary between upper and lower slopes as the elevation mid-point of each transect. We surveyed transects both from the canyon bottom moving upward, and from the top moving downward, traversing the slope as far as possible. We used a Suunto bubble clinometer (Suunto, Vantaa, Finland) to measure slope angle and a Nikon 400 laser range finder (Nikon, Melville, NY) to determine tree distance from each vantage point. The range finder’s minimum operating distance was 10 m, so we estimated distance of all trees <10 m from any vantage point and converted the distance values to vertical elevations above the canyon bottom by sine triangulation. At each vantage point, we noted prominent trees and/or other features that could be recognized from other locations along the transect to avoid double counting trees. We categorized trees based on their apparent canopy position as canopy, mid-story (not extending to the crowns of canopy trees), and understory (rarely more than 3 m in height). These classifications were based on relative positions within specific stands, e.g., canopy trees on south-facing upper slopes were often shorter and smaller in diameter at breast height (DBH) than midstory trees on lower slopes. We recorded all trees of sapling or greater size (>10 cm DBH). Ultimately, we surveyed 1.35 and 1.27 ha and recorded 441 and 556 canopy + midstory trees on north- and south-facing slopes, respectively. We did not follow Figure 2. Contrast of north- (right) vs. south-facing (left) slopes in Zoar Valley Canyon. Picture taken from transect S-2, looking due east from ~90 m above river channel. Flow is toward the camera. Seen here are north-facing slopes (including profile N-3) to center right, channel and fluvial landforms including an old-growth-forested upper terrace in lower center, and south-facing slopes (including profile S-3) extending from center left to center. Most conifers on north-facing slope are Eastern Hemlock; most on south-facing slope are Red and Eastern White Pines. Eastern Hemlock-dominated north-facing slopes can be seen in the distance extending ~2 km farther upstream. Northeastern Naturalist Vol. 21, No. 3 T.P. Diggins and R.G. Catterlin 2014 341 Diggins’ (2013) systematic approach for measuring tree height and DBH; rather, we measured accessible, representative trees to estimate height and diameter ranges for various stands. Diameter at breast height and thus b asal area also was not systematically recorded, but again some accessible trees were measured to estimate DBH ranges. Authorities for trees recorded in survey transects are given in Tables 1 and 2. We calculated total canopy + midstory species diversity (Shannon H' = -Σpilnpi, where pi is the proportion of species i in each sample and ln is the log base e) for north- and south-facing slopes, and for 5 upper and 5 lower fluvial terraces that were formerly surveyed and located within ~300 m of transects. Terrace data represented 2.41 ha of 30 m x 30m (upper terraces) and 10 m x 10 m (lower terraces) quadrats, including 587 trees >10 cm DBH. The coefficient of variation for the number of trees catalogued among all 10 terraces and 12 slope transects (upper and lower slopes treated separately) was 49%, suggesting only moderate variation in tree-count-based sampling effort. Also, transects and quadrats captured 78–100% of species identified within swaths extending at least 20 m laterally from slope transects, and on whole fluvial landforms below, with only rare species occasionally missed. Thus, comparisons of diversity among the different plot types are Table 1. Percentage by stem count of canopy, midstory, and understory species on north-facing Zoar Valley Canyon slopes. A dash (-) indicates species not found. Lower slopes Upper slopes Species Canopy Mid Understory Canopy Mid Understory Tsuga canadensis (L.) Carr. (Eastern Hemlock) 13.5 38.7 29.3 47.3 58.0 20.5 Acer saccharum Marsh. (Sugar Maple) 44.2 41.9 13.8 10.8 10.9 47.7 Fagus grandifolia L. (American Beech) 7.1 12.9 47.7 7.4 12.6 4.5 Quercus rubra L. (Northern Red Oak) 3.8 - - 8.1 3.4 - Quercus prinus L. (Chestnut Oak) 2.6 - - 8.8 1.7 - Betula alleghaniensis Britton (Yellow Birch) 4.5 2.4 - 3.4 1.7 4.5 Fraxinus americana L. (White Ash) 8.3 0.8 - 2.7 - 4.5 Liriodendron tulipifera L. (Tuliptree) 6.4 - - 2.7 1.7 - Ostrya virginiana (Mill.) K. Koch (Eastern 0.6 2.4 0.6 - 5.9 2.3 Hop Hornbeam) Tilia americana L. (American Basswood) 4.5 - - 1.4 - - Populus grandidentata Michx. (Bigtooth Aspen) - - - 3.7 - - Pinus strobus L. (Eastern White Pine) - - - 2.0 - - Magnolia acuminata L. (Cucumbertree) - - - - 1.7 - Carya cordiformis (Wangenh.) K. Koch 0.6 - - - 0.8 - (Bitternut Hickory) Acer rubrum L. (Red Maple) - 0.8 - - - - Betula lenta L. (Black/Sweet Birch) 0.6 - - - - - Prunus serotina Ehrh. (Black Cherry) 0.6 - - - - - Acer pensylvanica L. (Striped Maple) - - 6.3 - - 9.1 Hamamelis virginiana L. (Witch-hazel) - - 1.1 - - 4.5 Amelanchier arborea (Michx. F.) Fern - - 0.6 - - - (Shadbush) Carpinus caroliniana Walt. (American - - 0.6 - - - Hornbeam) Northeastern Naturalist 342 T.P. Diggins and R.G. Catterlin 2014 Vol. 21, No. 3 defensible here. Understory-tree cover was patchy and we exluded this stratum from our diversity calculations to avoid skewing distribution and abundance patterns of canopy + midstory species, which we suggest better represent stand structure. We also calculated canopy + midstory diversity for the whole site (gamma diversity), for terraces + north-facing slopes only, and for terraces + south-facing slopes only. Because basal area data were available only for the canyon-bottom terraces (Diggins 2013), we based our diversity values reported on stem coun ts. We analyzed percent of canopy + midstory by species based on stem counts (trees >10 cm DBH) for upper and lower slopes and on upper and lower fluvial terraces with non-metric multidimensional scaling (NMDS) ordination to assess patterns of similarity/dissimilarity among different environments. Contribution of the 12 most abundant species to NMDS axes was determined by Spearman rank correlation, reflecting the non-parametric nature of NMDS. Use of correlation does not imply that relationships were necessarily linear. Although we could not access 60°–90° outside-bend slopes, vantage points above, below, and/or across the canyon allowed qualitative assessment of their species composition. We made these assessments during spring and autumn, when variable leaf and bud/flower coloration facilitated tree identification from a distance. Table 2. Percentage by stem count of canopy, midstory, and understory species on south-facing Zoar Valley Canyon slopes. A dash (-) indicates species not found. Mid = midstory. Lower slopes Upper slopes Species Canopy Mid Understory Canopy Mid Understory Chestnut Oak 20.0 - 0.6 65.3 26.4 16.0 Sugar Maple 29.6 35.1 32.1 0.5 6.4 22.2 Northern Red Oak 14.4 10.3 0.6 19.0 25.6 9.8 Eastern Hemlock 8.0 26.8 6.3 - 4.0 2.7 American Beech 8.0 15.5 42.8 - 1.6 1.8 Red Maple 0.8 2.1 1.3 0.5 10.4 2.7 Witch Hazel - 3.1 6.3 0.9 8.0 8.0 Shadbush - 2.1 3.1 1.4 6.4 8.0 Pinus resinosa Ait. (Red Pine) - - - 4.2 4.8 1.8 White Ash 8.0 - - - - 0.9 Bigtooth Aspen - - - 6.5 0.8 - Eastern White Pine - 1.0 1.3 0.9 4.0 0.9 Quercus coccinea Muenchh. (Scarlet Oak) 1.6 2.1 - - - - Bitternut Hickory 3.2 - - - - - American Basswood 3.2 - - - - - Quercus alba L. (White Oak) 0.8 1.0 - 0.9 - - Eastern Hop Hornbeam 0.8 1.0 3.1 - - 2.7 Juniperus virginiana L. (Eastern Red Cedar) - - - - 0.8 0.9 Yellow Birch 0.8 - - - - - Carya ovata (Mill.) (Shagbark Hickory) 0.8 - - - - - Black Cherry - - - - 0.8 - Crataegus sp. (hawthorn sp.) - - 0.6 - - 1.8 Cornus florida L. (Flowering Dogwood) - - 1.3 - - - Striped Maple - - 0.6 - - - Northeastern Naturalist Vol. 21, No. 3 T.P. Diggins and R.G. Catterlin 2014 343 Results and Discussion Although north-facing canyon walls were slightly steeper than south-facing walls (Table 3), they supported closed canopies that occasionally exceeded 30 m in height, even on upper slopes (Fig. 3b). Some trees exceeded 80 cm DBH on lower slopes (Liriodendron tulipifera [Tuliptree], Fagus grandifolia [American Beech], and Sugar Maple) and along the gorge rim (Quercus prinus [Chestnut Oak] and Quercus rubra [Northern Red Oak]), but we visually determined that upper-slope trees otherwise rarely exceeded ~60 cm. In contrast, south-facing upper slopes were dominated by stunted open-canopy oak-pine groves that were most often less than 6 m in height and 40 cm DBH (Fig. 3a). North-facing slopes on outside canyon bends were more wooded than south-facing slopes, where long stretches of canyon wall were largely barren (Figs. 1, 2). North-facing slopes surveyed by transect supported extensive Eastern Hemlocknorthern hardwood stands, but also included components of mesophytic forest types including Sugar Maple, American Beech, White Ash, Yellow Birch, Tuliptree, Chestnut Oak, and Northern Red Oak (Table 1). There was an almost complete reversal of canopy dominance in terms of percent stem density moving from lower (44% Sugar Maple) to upper slopes (47% Eastern Hemlock) (Table 1). Other species were more evenly distributed (Table 1) except Chestnut Oak, which was confined largely to the gorge rim (often as >80-cm DBH edge trees), with only scattered individuals found on lower slopes but never less than 50 m above the terrace below. Canopy composition of some lower-slope stands suggested rich mesophytic, beech-maple mesic, and/or maple-basswood rich mesic forest (see Edinger et al. 2002) in addition to the Eastern Hemlock-northern hardwood type . The understory along north-facing transects (Table 1) was strongly dominated by Sugar Maple, American Beech, and Eastern Hemlock, which are among the most shade-tolerant trees of eastern forests (Burns and Honkala 1990). Other components included Acer pensylvanicum (Striped Maple) and Hamamelis virginiana (Witch-hazel), which often remain in the under- and midstories even as reproductively mature trees, and saplings of White Ash and Betula alleghaniensis (Yellow Birch) which are usually canopy species. Table 3. Height, length, and slope angle of 6 belt transects on Zoar Valley Canyon slopes. Steepest = steepest 10-m (vertical) segment. Slope (degrees) Transect Height (m) Length over ground (m) Overall Steepest North-facing N-1 105 219 28.7 42 N-2 113 205 33.5 40 N-3 128 221 35.6 40 South-facing S-1 64 139 27.5 43 S-2 102 255 23.4 30 S-3 136 283 28.5 41 Northeastern Naturalist 344 T.P. Diggins and R.G. Catterlin 2014 Vol. 21, No. 3 The inaccessible north-facing outside canyon bends were dominated by Eastern Hemlock, and supported more Eastern White Pine than we observed on the quantitatively surveyed slopes. Occasional Northern Red Oak, Yellow Birch, Amelanchier arborea (Shadbush), and Populus tremuloides Michx. (Quaking Aspen) were the only other trees able to persist on north-facing cliffs steeper than 60˚ (i.e., east of transect N-2, as seen in Fig. 1). Much like the stands across the canyon, the south-facing lower-slope stands (Fig. 3c) were also comprised of Eastern Hemlock-northern hardwood and mesic Figure 3. Representative forest types on the different exposures studied here: A) Southfacing upper slope along transect S-2 supporting oak (Chestnut Oak, Northern Red Oak) – pine (Red Pine, Eastern White Pine) stands; some trees here exceed 165 years old. B) North-facing upper slope at top of N-3 supporting predominantly Eastern Hemlock, with scattered northern hardwoods. C) South-facing lower slope along S-2 increasingly dominated by mesic species, and generally compositionally similar to north-facing lower slopes. D) Diverse mid-successional rich mesophytic stands 100–130 years of age below N-1 and N-2 on riverside terrace. Abundant species include Sugar Maple, American Beech, Tuliptree, White Ash, and Bitternut Hickory. Geomorphically older terraces support uneven-aged Eastern Hemlock-American Beech-Sugar Maple old-growth. Northeastern Naturalist Vol. 21, No. 3 T.P. Diggins and R.G. Catterlin 2014 345 forest types. Sugar Maple, Eastern Hemlock, Northern Red Oak, American Beech, White Ash, and Carya cordiformis (Bitternut Hickory) were abundant in the canopy and midstory (Table 2). However, Chestnut Oak was also abundant on lower southfacing slopes whereas it was scarce on the north-facing lower slopes. Also unlike our observations for north-facing transects, there was a major compositional shift between lower and upper slopes, not simply a shift in dominance. Sugar Maple and other mesic components such as American Beech, White Ash, and Eastern Hemlock declined notably with increasing elevation and were replaced by Chestnut Oak and Northern Red Oak (63% and 19% of canopy stems, respectively), eventually yielding the open-canopy stands pictured in Figure 3a. Red Pine, Eastern White Pine, Red Maple, Shadbush, Witch-hazel, and Populus grandidentata (Bigtooth Aspen), also occurred predominantly on the upper slopes (Table 2). South-facing understories, particularly on upper slopes, were more species-rich than those of north-facing slopes (compare Tables 1 and 2), likely a reflection of a more open canopy that allowed shade-intolerant species such as Red Pine and Eastern White Pines, Juniperus virginiana (Eastern Red Cedar), and Crataegus spp. (hawthorn) to become established. Zoar Valley’s south-facing upper exposures were characterized by distinctive oak-pine stands (Chestnut Oak and Northern Red Oak with Red Pine and Eastern White Pine) within which canopy trees typically reached no more than 6 m in height and 40 cm DBH (see Fig. 3a). Chestnut Oak and Red Pine both exceeded 165 y here (we cored stunted trees below breast height) and on eroded glacial outwash deposits along some canyon rims (where, unlike on slopes, trees grow to >20 m). Thus, trees on south-facing upper slopes reach ages comparable to those recorded for old-growth mesic species on lower slopes and fluvial terraces (Diggins 2013, Pfeil et al. 2007). This particular oak-pine assemblage is unusual in the eastern deciduous biome. Various open-canopy oak and oak-pine communities described elsewhere in New York State (Edinger et al. 2002, Lewin 1974, McIntosh 1959) and Pennsylvania (Fike 1999) are dominated by Pinus rigida Mill. (Pitch Pine) rather than Red Pine, or lack pines altogether. Farther south, in addition to Pitch Pine, oak-pine forest types may also include Pinus virginiana Mill. (Virginia Pine) (Braun 1935, Whittaker 1956). Conversely, Red Pine ridge and summit communities in New York State are reported only from the Adirondacks and Catskills, and do not include Chestnut Oak (Edinger et al. 2002). Variation in overall density of canopy + midstory trees among transects and fluvial landforms was moderate (209–525 stems per ha; Table 4), especially in light of the diversity of forest types and/or successional stages within the study area. However, several tends were still evident. Stem density was highest in the oak-pine stands on south-facing upper slopes, even though they represented the shortest and sparsest canopy of any landform studied here. Also, as demonstrated by Diggins (2013) and Van Pelt et al. (2006), early-successional lower fluvial terraces had higher stem-densities than adjacent late-succesional/old-growth upper terraces (Table 4). Northeastern Naturalist 346 T.P. Diggins and R.G. Catterlin 2014 Vol. 21, No. 3 Ordination of percent stem density by species (Fig. 4a) revealed a general clustering of north- and south-facing lower slopes with mature fluvial terraces, but divergence among north- and south-facing upper slopes and the youngest of lower terraces, which had stand ages of 29 and 31 years in 2009 (Diggins 2013). Because it was based on only a single transect, the north-facing upper slope divergence may not represent a real trend (Fig. 4a). The contribution of species important to the ordination (Fig. 4b) reflected transect/terrace trends clearly: Red Oak, Chestnut Oak, and Acer rubrum (Red Maple) for the south-facing upper slopes; Eastern Hemlock the north-facing slopes; and Populus deltoides (Eastern Cottonwood), Platanus occidentalis (American Sycamore), and Robinia pseudoacacia L. (Black Locust) the young fluvial terraces. Mesic hardwoods such as Sugar Maple, American Beech, and Tuliptree, trending to the left of the y-axis, revealed an association with upper terraces (Fig. 3d) and north-facing lower slopes (Fig. 4b). South-facing slopes were higher in species richness and diversity than northfacing slopes (21 and 18 species, respectively; Shannon’s H' diversity values of 2.100 and 1.872, respectively; Table 5), reflecting the south-facing elevational shift in species composition. In contrast, the strong dominance of Sugar Maple and Eastern Hemlock caused the lower diversity values observed on north-facing slopes. Fluvial terraces had higher diversity, but only if both old-growth upper terraces and riparian-pioneer lower terraces were included in the analysis. Interestingly, the combination of fluvial terraces + south-facing slopes (Table 5) slightly exceeded Table 4. Abundance per ha of canopy and midstory trees on slopes and of >10-cm-DBH trees on riverside terraces. The distinction between canopy and midstory was not made on the terraces. Canopy Midstory Total North-facing slopes Upper 162 131 293 Lower 231 169 400 South-facing slopes Upper 332 192 525 Lower 176 137 313 Terraces Upper 209 Lower 470 Table 5. Species richness and diversity (Shannon H') by stem count on Zoar Valley Canyon slopes and adjacent fluvial terraces. Richness Diversity (H') North-facing transects 18 1.872 South-facing transects 21 2.100 Fluvial terraces (all) 23 2.399 Fluvial terraces (upper/old growth only) 15 1.815 Terraces (all) + north-facing 28 2.251 Terraces (all) + south-facing 32 2.505 Whole site (gamma diversity) 34 2.426 Northeastern Naturalist Vol. 21, No. 3 T.P. Diggins and R.G. Catterlin 2014 347 whole-site gamma diversity (2.505 vs. 2.462). In addition to the total of 34 species reported from canyon slopes (Tables 1 and 2) and fluvial terraces (Diggins 2013), we observed 5 additional tree species within the general study area but not in a transect or quadrat: Nyssa sylvatica var. sylvatica Marsh. (Black Tupelo) and Quercus velutina Lam. (Black Oak) on south-facing canyon rims, and Alnus incana subsp. rugosa (Du Roi) R.T. Clausen (Speckled Alder), Populus balsamifera L. (Balsam Poplar), and Quercus palustris Muenchh. (Pin Oak) on terraces. Figure 4. Non-metric multidimensional scaling (NMDS) ordination of slope transects (separated into upper and lower reaches) and fluvial terraces based on percent stem density of canopy + midstory trees. Distribution of slope transects and terraces is shown in panel A; and contribution of selected species to NMDS axes (i.e. Spearman correlation coefficients) is shown by arrows in B. Values of coefficients in B are doubled to improve readability. Four-letter species codes in B (the 1st two letters of genus and species names) represent the following: ACRU = Red Maple, ACSA = Sugar Maple, FAGR = American Beech, FRAM = White Ash, LITU = Tuliptree, PLOC = American Sycamore, PODE = Eastern Cottonwood, QUPR = Chestnut Oak, QURU = Northern Red Oak, ROPS = Black Locust, TIAM = American Basswood, and TSCA = Eastern Hemlock. Northeastern Naturalist 348 T.P. Diggins and R.G. Catterlin 2014 Vol. 21, No. 3 The topography-related vegetation patterns we catalogued at Zoar Valley largely conformed to expectations for the eastern US, with mesic/mesophytic assemblages on north-facing slopes (e.g., Braun 1935 in KY, Fralish 1994 in IL, Potzger 1939 in IN,) and semi-xeric stands on south-facing slopes, particularly on upper slopes (Fekedulegn et al. 2003 in WV, Vankat et al. 1977 in OH). In narrow gorges surrounding central New York State’s Finger Lakes, Lewin (1974) likewise noted elevational gradients from lower-slope mesic hardwoods to either Eastern Hemlockor oak-dominated forests at gorge rims. However, Lewin (1974) found no obvious vegetation patterns related to slope or apect, in contrast to Zoar Valley where the broad ~500-m-wide canyon provides a clear north- vs. south-facing difference in sun exposure. Vegetational gradients in Zoar Valley were especially pronounced given the modest geographic and geologic range represented by the site (see also Mowbray and Oosting 1968, Shanks and Norris 1950, Thomas and Anderson 1993). Additionally, although Zoar Valley is located entirely within the broadly defined eastern deciduous biome (Braun 1950), both climatic and ecotypic gradients within the immediate region may enhance the topographically associated diversity described here. Cattaraugus Creek’s watershed straddles the boundary of the Eastern Hemlock-northern hardwood region of the Allegheny Plateau and the climatically moderated Lake Erie Plain into which mesophytic forest types extend from the south and west (see e.g., Braun 1950). It is also positioned near the latitudinal limits (see Burns and Honkala 1990) of a number of species more widely distributed either to the south (Tuliptree, Chestnut Oak, Quecus coccinea Muenchh. [Scarlet Oak], Juglans nigra L. [Black Walnut]), or to the north (Red Pine, Balsam Poplar). Zoar Valley’s canyon-bottom riparian forests are recognized as among the East’s finest hardwood stands in terms of canopy and emergent heights, advanced age, and diversity of species and forest types (Diggins 2013, Diggins and Kershner 2005, Hunt et al. 2002). The site’s canyon-slope woodlands are also considered exemplary but had not been quantitatively surveyed prior to our project. The results of our study demonstrated that both north- and south-facing slopes were diverse in their own right, and that the elevationally divergent south-facing slopes, in particular, enhance site-wide species richness and diversity (Table 5). We also added an oak-pine woodland type (not yet fully described) to those already listed for the study area (see also Hunt et al. 2002)—i.e., the four previously mentioned mesic/mesophytic forests and a Populus deltoides Marsh. (Eastern Cottonwood)–Platanus occidentalis L. (American Sycamore) lowland forest along the river . Acknowledgments Partial support for this project was provided by the University Research Council of Youngstown State University (to Richard Catterlin). A number of students aided in the collection of data, most notably Sara Paloski and Sean Satterlee. Northeastern Naturalist Vol. 21, No. 3 T.P. Diggins and R.G. Catterlin 2014 349 Literature Cited Braun, E.L. 1935. The vegetation of Pine Mountain, Kentucky: An analysis of the influence of soils and slope exposure as determined by geological structure. American Midland Naturalist 16:517–565. Braun, E.L. 1950. Deciduous Forests of Eastern North America. Hafner Publishing, New York, NY. 596 pp. Burns, R.M., and B.H. Honkala. 1990. Silvics of North America, Volumes 1 and 2. Agriculture Handbook 654, US Department of Agriculture-Forest Service, Washington, DC. Cantlon, J.E. 1953. Vegetation and microclimates on north and south slopes of Cushetunk Mountain, New Jersey. Ecological Monographs 23:241–270. Catterlin, R.J. 2010. Composition of canyon-slope woodlands in Zoar Valley, western New York, as associated with slope orientation and elevation. M.Sc. Thesis. Youngstown State University, Youngstown, OH. 75 pp. Cottle, H.J. 1932. Vegetation on north and south slopes of mountains in southwestern Texas. Ecology 13:121–134. Diggins, T.P. 2013. A 300-year successional sequence in an eastern United States riparian hardwood forest. Journal of the Torrey Botanical Society 140:65–88. Diggins, T.P., and B. Kershner. 2005. Canopy and understory composition of old-growth riparian forest in Zoar Valley, New York, USA. Natural Areas Journal 25:219–227. Edinger, G.J., D.J. Evans, S. Gebauer, T.G. Howard, D.M. Hunt, and A.M. Olivero. 2002. Ecological Communities of New York State. Second Edition. New York Natural Heritage Program, New York State Department of Environmental Conservation, Albany, NY. 136 pp. Fekedulegn, D., R.R. Hicks, Jr., and J.J. Colbert. 2003. Influence of topographic aspect, precipitation, and drought on radial growth of four major tree species in an Appalachian watershed. Forest Ecology and Management 177:409–425. Fike, J. 1999. Terrestrial and palustrine plant communities of Pennsylvania. Pennsylvania Natural Diversity Inventory, Harrisburg, PA. 79 pp. Fralish, J.S. 1994. The effect of site environment on forest productivity in the Illinois Shawnee Hills. Ecological Applications 4:134–143. Goldblum, D., and L.S. Rigg. 2003. Age structure and regeneration dynamics of Sugar Maple at the deciduous/boreal forest ecotone, ON, Canada. Physical Geography 23:115–129. Gordon, R.B. 1940. The primeval forest types of southwestern New York. New York State Museum Bulletin No. 321, Albany, NY, 102 pp. Hunt, D.M., G.J. Edinger, J.J. Schmid, D.J Evans, G. Novak, A.E. Olivero, and S.M. Young. 2002. Lake Erie Gorges biodiversity inventory and landscape integrity analysis. New York Natural Heritage Program, Albany, NY. 131 pp. Hwang, T., C. Song, J.M. Vose, and L. E. Band. 2011. Topography-mediated controls on local vegetation phenology estimated from MODIS vegetation index. Landscape Ecology 26:541–556. Lewin, D.C. 1974. The vegetation of the ravines of the southern Finger Lakes, New York, region. American Midland Naturalist 91:315–342. McIntosh, R.P. 1959. Presence and cover in Pitch Pine-oak stands of the Shawangunk Mountains, New York. Ecology 40:482–485. Mowbray, T.B., and H.J. Oosting. 1968. Vegetation gradients in relation to environment and phenology in a southern Blue Ridge gorge. Ecological Monographs 38:309–344. Northeastern Naturalist 350 T.P. Diggins and R.G. Catterlin 2014 Vol. 21, No. 3 Murphy, S.J., and B.C. McCarthy. 2012. Evidence for topographic control of tree spatial patterning in an old-growth, mixed mesophytic forest in southeastern Ohio, USA. Journal of the Torrey Botanical Society 139:181–193. Nichols, G.E. 1914. The vegetation of Connecticut III. Plant societies on uplands. Torreya 14:167–194. Pfeil, E.K., N. Casacchia, J.G. Kerns, and T.P. Diggins. 2007. Distribution, composition, and orientation of down deadwood in riparian old-growth woodlands of Zoar Valley Canyon, western New York State, USA. Forest Ecology and Management 239:159–168. Potzger, J.E. 1939. Microclimate and a notable case of its influence on a ridge in central Indiana. Ecology 20:29–39. Searcy, K.B., B.F. Wilson, and J.H. Fownes. 2003. Influence of bedrock and aspect on soils and plant distribution in the Holyoke Range, Massachusetts. Journal of the Torrey Botanical Society 130:158–169. Shanks, R.E., and F.H. Norris. 1950. Microclimatic variation in a small valley in eastern Tennessee. Ecology 31:532–539. Thomas, R.L., and R.C. Anderson. 1993. Influence of topography on stand composition in a Midwestern ravine. American Midland Naturalist 130:1–12. Vankat, J.L., D.S. Anderson, and J.A. Howell. 1977. Plant communities and distribution factors in Abner’s Hollow, a south-central Ohio watershed. Castanea 42:216–226. Van Pelt, R., T.C. O’Keefe, J.J. Latterell, and R.J. Naiman. 2006. Riparian forest stand development along the Queets River in Olympic National Park, Washington. Ecological Monographs 76:277–298. Wang, Y-C. 2007. Spatial patterns and vegetation-site relationships of the presettlement forests in western New York, USA. Journal of Biogeography 34:500–513. Whittaker, R.H. 1956. Vegetation of the Great Smoky Mountains. Ecological Monographs 26:1–69. Ziegler, S.S. 1995. Relict Eastern White Pine (Pinus strobus L.) stands in southwestern Wisconsin. American Midland Naturalist 133:88–100. Zimmerman, C.L., and J.R. Runkle. 2010. Using land units for conservation planning in a southwestern Ohio watershed. Natural Areas Journal 30:27–38.