INTERACTIONS BETWEEN LANDSCAPE FEATURES … book.pdfaverage temperature during the winter ranging between -7o C and 19 o C and temperatures during the summer ranging between -2o C

In: Forest Ecology Research Horizons ISBN: 1-60021-490-8 Editor: Nole C. Verne, pp. 147-166 © 2007 Nova Science Publishers, Inc.

Chapter 4

INTERACTIONS BETWEEN LANDSCAPE FEATURES, DISTURBANCE AND VEGETATION IN FREQUENTLY

BURNED APPALACHIAN OAK FORESTS

Stephen A. Signell1 and Marc D. Abrams 2* 1 Adirondack Ecological Center, SUNY Environmental

Science and Forestry, 6312 State Route 28N, Newcomb, NY 12852. 2 307 Forest Resources Building, School of Forest Resources,

Pennsylvania State University, University Park, Pennsylvania, 16802 USA

ABSTRACT

Disturbance regime and topographic features can exert a strong influence on vegetation dynamics. Here, we present a case study illustrating how repeated fire and rocky landscape features interact to influence the distribution and regeneration of tree species in the central Appalachian Mountains of the eastern USA. Forests in this region have undergone profound forest changes since European settlement, in part because forest fragmentation and the 20th century practice of active fire suppression have reduced fire frequency from presettlement levels. The site of this study, located on a military reservation in the Ridge and Valley physiographic region of Pennsylvania, is regionally unique in that it has experienced numerous wildfires in the latter half of the 20th century. Stands with strongly contrasting fire histories often exist in close proximity to each other, e.g. on opposite sides of a road. In addition, several stands contain considerable variability in surface rock cover, with small-scale rocky patches (0.03-0.5ha) embedded in a matrix of non-rocky terrain. The juxtaposition of contrasting fire regimes and variable rock cover constituted a "natural experiment," allowing us to study the influence of both fire and rocky patches on spatial patterns of vegetation dynamics.

In burned stands, we found that rock cover had a strong influence on vegetation, with rocky patches supporting greater density of fire-sensitive species such as Acer rubrum, Sassafras albidum and Nyssa sylvatica, but lower abundance of tree regeneration and groundcover than did the surrounding non-rocky matrix. Quercus had fewer fire scars and catfaces (open, basal wounds) on rocky patches, indicating that rocky features

* Tel: 518-582-4551; Fax: 518-582-2181; email: [email protected]

No part of this digital document may be reproduced, stored in a retrieval system or transmitted in any form orby any means. The publisher has taken reasonable care in the preparation of this digital document, but makes noexpressed or implied warranty of any kind and assumes no responsibility for any errors or omissions. Noliability is assumed for incidental or consequential damages in connection with or arising out of informationcontained herein. This digital document is sold with the clear understanding that the publisher is not engaged inrendering legal, medical or any other professional services.

Stephen A. Signell and Marc D. Abrams 148

mitigated fire severity. Under a sustained fire regime, therefore, heterogeneity in rock cover created a mosaic where fire-adapted species such as Quercus dominated the landscape, but where fire-sensitive species persisted in isolated pockets of lower fire severity. In unburned stands, differences between rocky and non-rocky patches were less distinct, with both patch types having sparse groundcover, little tree regeneration, and high understory densities of relatively shade tolerant A. rubrum, N. sylvatica and Betula lenta. Spatial patterns in unburned stands indicate competition has played the key role in stand development. The uniform spatial pattern of overstory stems in unburned stands points to density-dependent mortality, which drives clustered and random stem distributions toward uniformity.

In burned stands, fire appears to have had two important and related impacts on stand structure that have important implications for Quercus reproduction: (1) it reduced overall tree density, particularly in the understory; and (2) it thinned some areas more than others, creating distinct patchiness in tree density. Overstory stems in burned stands had clustered distribution. When overstory or understory tree density exceeded 400 and 200 trees/ha respectively, Quercus saplings were virtually absent. Thus, fire stimulates Quercus regeneration only when it reduces both understory and overstory density. All plots in unburned stands exceeded one or both of these density thresholds. In the absence of fire, Quercus and other early to mid-successional species may eventually be replaced, resulting in declining landscape and species richness.

INTRODUCTION Disturbance and topographic features can exert a strong influence on vegetation

dynamics (Swanson et al. 1988, Veblen et al. 1994). In this chapter we present a case study on how repeated fire and landscape features interact to influence vegetation dynamics and tree regeneration in Quercus-dominated forests in the central Appalachian Mountains of the eastern USA. The site of this study, located on a military reservation in the Ridge and Valley physiographic region on Pennsylvania, is regionally unique in that several stands have experienced numerous fires in the latter half of the 20th century (Signell et al. 2005, Signell and Abrams 2006).

Quercus first became abundant in the Central Appalachians about 8,000 years ago and has maintained dominance well into the 20th century (Delcourt et al. 1984; Webb 1988; Watts 1979; Loeb 1989, Abrams and Ruffner 1995; Black and Abrams 2001; Cogbill et al. 2002; Whitney and DeCant 2003). During this period fire-sensitive genera such as Acer, Betula, and Nyssa were typically present at low levels. Low intensity surface fires are thought to have been common in eastern Quercus forests during this time (Abrams 1992; Lorimer 2001; Brose et al. 2001). Presettlement fires would have been ignited either by lightning or by Native Americans who used fire to create habitat for game, improve hunting conditions, and to clear land for agriculture (Byrd 1728, Maxwell 1910, Day 1953, Russell 1983, Whitney 1994). In some areas, fires were frequent enough to create savannahs or even scrublands. For example, Budd [1685]1966 described areas in Pennsylvania and New Jersey where fire had removed all understory species except the most fire-tolerant grasses and herbs. In other areas, such as the scrubby "grubenlands" of Lancaster County, PA, Indian fires had completely destroyed the overstory, leaving only dense thickets of oak saplings (Fletcher 1950). While it is hard to glean quantitative information about fire frequency from historical accounts such as these, analysis of fire scars from several old growth trees in the Mid-Atlantic region estimate

Interactions Between Landscape Features, Disturbance and Vegetation... 149

fire return intervals of 4-20 years for the presettlement and early settlement eras (Buell et al. 1954, Shumway et al. 2001). Frequent fire is thought to have facilitated regeneration of relatively light-demanding Quercus by removing fire-sensitive competitors and allowing more light to reach the understory (Lorimer 1993).

The reduction in fire frequency during the 20th century fire-suppression era has been linked to the regionally pervasive and widespread problem of oak regeneration failure (Lorimer 1989, Alerich 1993, Steiner et al. 2004). While oaks are still prevalent in forest overstories throughout much of the region, more shade-tolerant competitors such as red maple (Acer rubrum, L.), black birch (Betula lenta, L.), blackgum (Nyssa sylvatica, Marsh, var sylvatica.) and black cherry (Prunus serotina, Ehrh.) often dominate the understory, and oak saplings are often rare or absent (Abrams and Nowacki 1992, Abrams and Downs 1990, Alerich 1993, Steiner et al. 2004). This competing understory vegetation likely plays an important role in oak regeneration failure. Oaks are characterized as having low to intermediate shade tolerance (Barnes and Wagner 1981), and a “sapling bottleneck” is often observed in shaded conditions where oak seedlings are unable to survive after their acorn energy reserves are exhausted (Lorimer 1981, Pubanz et al. 1989, Nowacki et al. 1990). Successful oak advance regeneration is now primarily relegated to dry, low-quality sites where competitive pressure is low, for example, ridge tops (Nowacki and Abrams 1992, Orwig and Abrams 1994, Collins and Carson 2004).

Because few sites in the eastern U.S. have experienced fires during the 20th century, most fire research in central Appalachian forests has focused on prescribed burns (Wendell and Smith 1987, Arthur et al. 1998, Blake and Schuette 2000, Blake and Schuette 2000, Brose et al. 2001, Kuddes-Fischer and Arthur 2002). However, because prescribed fires are strongly controlled, their ecological impact is unlikely to resemble presettlement fires. In uncontrolled conditions, spatial variability in aspect, slope, moisture content, edaphic and other conditions exert a strong influence over fire behavior, creating patchiness at multiple spatial scales (Swanson et al. 1988, Pickett and Cadenasso 1999). Relationships between landscape features, fire regime and vegetation have been established in other regions where wildfires still occur (Turner et al. 1994, Veblen et al. 1994, Heyerdahl et al. 2001, Taylor and Skinner 2003). Unfortunately, no such work has been possible in the east because of the lack of sites where fires have been able to behave naturally over a large temporal and/or spatial scale. Thus, very little is known about how landscape features might interact with fire to influence vegetation dynamics, tree regeneration and the spatial distribution of species in eastern forests.

The forests at the National Guard Training Center at Fort Indiantown Gap (NGTC-FIG) near Harrisburg, Pennsylvania, offer a unique opportunity to investigate interactions between fire, landscape features and vegetation in central Appalachian Quercus forests. An abundance of ignition sources has resulted in multiple forest fires in parts of NGTC-FIG during the past 70 years. Most NGTC-FIG forest fires have originated in the "impact area", a restricted access zone where training exercises often include use of live ammunition. Historically, fires that escaped the impact area were not immediately suppressed, but were allowed to burn until reaching a firebreak such as a road or ridge top. Thus, stands with strongly contrasting fire histories often exist in close proximity, e.g., on opposite sides of a road. On ridges, stands also contain considerable heterogeneity in rock cover, where a matrix of non-rocky terrain surrounds discrete, isolated rocky patches ranging in size from 0.03 to 0.5 ha. The presence of rocky and non-rocky areas within burned and unburned stands constituted a "natural


experiment," enabling us to test the hypothesis that rocky patches serve as fire refugia for fire sensitive species.

SITE DESCRIPTION This research was conducted in oak stands within the National Guard Training Center at

Fort Indiantown Gap (NGTC-FIG), located in Lebanon and Dauphin Counties in south central Pennsylvania (figure 1). NGTC-FIG lies across the southernmost extent of the Ridge and Valley physiographic province of the Appalachian Mountain chain. The climate at NGTC-FIG is humid continental, with average annual temperature of 10 o C for the region, average temperature during the winter ranging between -7o C and 19 o C and temperatures during the summer ranging between -2o C and 30 o C. January is the coldest month with an average range of -7 o to 4 o C, while July and August are the hottest months of the year with an average maximum of 31 o C and an average minimum of 16 o C. Total annual precipitation averages 107 cm with around 60 % of precipitation falling during the growing season (April to October). High average precipitation occurs in May at 11.2 cm, while a low of 6.3 cm falls during February (Anonymous 2002).

Figure 1. Locations of NGTC-FIG and sampled stands with respect to the northeastern USA (inset).

The forest stands sampled for this study lie on the south-facing slope of Second Mountain and the valley to the south. The valley was formed from the more easily weathered shales and siltstones, while the ridges formed from the more resistant shales, sandstones, and conglomerates (USDA 1981). The soils of the ridge sites are of the Laidig and Buchanan series (mesic Typic and Hydric Fragidults, respectively), derived in colluvium from sandstone, siltstone or shale. These very deep soils are most often found in gently sloping areas on lower slopes or mountain terraces, and are characterized as well drained to very well-drained. This soil series is not generally suitable for agriculture and most areas remain forested. The soils of the valley sites are of the Leck Kill and Bedington series (Typic


Hapudults), derived in residuum from red shale, siltstone or sandstone (USGS Soil Survey Maps). These series occur on convex uplands and hill sideslopes, and are characterized as moderately deep to very deep, well-drained soils derived from residuum. A majority of Leck Kill and Bedington soils have been cleared for agriculture while the remainder is in woodland or other uses.

SITE HISTORY During the last glacial maximum (18,000-20,000 bp), NGTC-FIG lay about 100 km south

of the ice sheets, and the landscape was dominated by tundra (Watts 1979). Following the retreat of the glaciers, the area gradually reforested, first by boreal species and then by temperate deciduous forests (Delcourt et al. 1984). By 8,000 years ago, the oak forests that currently dominate much of the region had become established (Delcourt et al. 1984, Foster et al. 2002). Humans inhabited the area during this period of vegetation and climate change. The Shoop archaeological site, one of the oldest Paleo-Indian (12,000 B.C-7,000 BC) sites in North America, is located just 15 km from NGTC-FIG. The NGTC-FIG property itself contains several archaeological sites dating from the Woodland Period of 1,000 - 1550 A.D., most being remnants of hunting camps used by the Lenni-Lenape people (Anonymous 2002). The fact that no permanent settlements were found in the less productive shale-derived Ridge-and-Valley portion of NGTC-FIG, the site of our research, indicates that Indians used the area primarily for hunting and travel. Their permanent villages and croplands were located in the more fertile limestone valleys to the south in Paxtang (Harrisburg) and Indiantown Gap (Wallace 1965).

With the arrival of the first European settlers in the early 1700s, the Lenni-Lenape were displaced from their piedmont settlements and took up full-time residence in the Ridge-and-Valley region to the north. The study site received the most intense use by Native Americans during the period 1720-1763, when the area between Blue and Second mountains was used by the Lenni-Lenape as a base of operations for frequent raids on the European settlements to the south (Smoker 2003). Incidence of anthropogenic fire was probably at a prehistoric high during this time. Following their defeat in the French and Indian war in 1763, the Lenni-Lenape abandoned eastern Pennsylvania, and this paved the way for European settlement of ridge-and-valley region.

Early settlers cleared the valleys for agriculture and managed the ridge slopes as woodlots, which were likely harvested one or more times during the 1800s. Several large charcoal-powered iron furnaces operated in the area, one of which, Manada Furnace, lay only 2 km to the southwest of NGTC-FIG property. Manada Furnace operated from 1841 until 1875, employing 75 men at its peak and consuming up to 0.4 ha of hardwood forest per day (Dove 2004). Many of the ridge forests at NGTC-FIG established during this period, with fires occurring in some areas (Signell et al. 2005).

The decline of the charcoal industry initiated a period of gradual emigration as farmers left their marginally productive farms. By the 1930s, most of the farmland around NGTC-FIG had been abandoned, and in 1932 the State of Pennsylvania purchased land between Blue and Second Mountains for use as a military training facility, now known as NGTC-FIG. While much of the abandoned farmland has reverted to forest, large portions of the valley have been


kept clear for military training operations, including tank, artillery and infantry exercises. The facility also contains a restricted "impact" area where live-ammunition training occurs. Ridge sites are mostly forested.

PRESETTLEMENT FORESTS Presettlement forest composition of the NGTC-FIG area was reconstructed using original

land survey maps for Dauphin and Lebanon counties. We were able to quantify forest change on several landforms by comparing witness tree species distributions with present-day species distributions derived from 1,982 forest inventory plots arranged in a regular grid across the NGTC-FIG property. Witness tree analysis confirmed that the pre-European settlement forests in the NGTC-FIG vicinity were an oak-chestnut type, as classified by Braun (Braun 1950). White oak (Quercus alba) and black oak (Quercus velutina) dominated valley locations while chestnut oak (Quercus montana) and American chestnut (Castanea dentata) dominated ridge sites (table 1). Scarlet oak (Quercus coccinea), hickory (Carya, spp.) and pine (Pinus, spp.)were also relatively common on both valley and ridge sites. The finding that oak species were dominant across all landforms in the presettlement forests of FIG is consistent with other witness tree studies in the region (Abrams and Ruffner 1995, Black and Abrams 2001). Five percent of witness trees were identified as oak, hickory, pine and chestnut "grubs" or “saplings”, suggesting a lack of sizeable trees in some locations. This fact along with the occurrence of shade intolerant and short-lived species such as black locust and scarlet oak provides further evidence that the presettlement landscape was a patchwork of forests in different stages of successional development.

FOREST CHANGE Species composition has changed dramatically since European settlement. Former forest

dominants such as American chestnut, hickory, black oak and white oak have declined on most or all landforms, while species such as red maple, black birch, blackgum, black cherry, tulip poplar, red oak and hemlock have increased (table 1). Many of these increasing species were either absent or very rare in the witness tree record, perhaps relegated to talus slopes, coves and/or drainages (Abrams and McCay 1996). Change has been most pronounced in valleys where human activity has been concentrated. Old fields that were allowed to reforest following widespread agricultural abandonment in the early 1900s have largely succeeded to pine (mostly Pinus rigida and Pinus virginiana), which are often common on former agricultural sites (Stover and Marks 1998, Motzkin et al. 1999). Unburned valley forests have seen large increases in maple and birch, as is typical of the region (Nowacki and Abrams 1992, Steiner et al. 2004).

Change has been less dramatic in burned valley stands, riparian areas and ridge tops. With no significant increases in maple, birch, hemlock or cherry and substantial numbers of white oak, burned valley forests were more similar to presettlement forests than unburned or old-field valley communities. However, the lack of chestnut, hickory and black oak in burned stands indicates that even these forests bear little resemblance to the presettlement forests.

Table 1. Presettlement and present-day species frequency (%) on various landforms based on corner trees of original warrants (1766-1825) and the 2003 NGTC-FIG forest inventory. Valley plots were subdivided by disturbance history with unburned plots

(no evidence of fire), old field plots (1938 field) and burned plots (visible evidence of fire). Significant changes from presettlement

to present-day are shown in the column marked “∆”

Species Witness

Unburned

2003

Old Field

2003

Burned

2003 Witness 2003 Witness 2003 Witness 2003 Witness 2003

Acer, spp. 1.6 14.1 + 15.0 + 9.1 11.5 17.3 18.8 + 12.4 + 20.6 +

Betula, spp. 0.8 17.1 + 2.5 5.2 1.9 13.9 + 1.9 25.0 + 1.6 13.1 + 3.3 14.8

Carya, spp. 7.0 1.5 - 0.5 - 0.8 - 7.7 0.7 - 5.1 0.2 - 15.4 5.0 - 13.3 5.4

Castanea dentata 10.5 - - - 1.9 - 25.0 - 17.0 - 3.3 -

Liriodendron tulipifera 0.8 3.4 + 0.5 0.1 1.9 6.1 3.2 3.0 1.1 5.2 + 1.3

Nyssa sylvatica 2.9 5.2 0.7 - 3.1 3.8 5.5 3.2 5.2 2.2 6.7 6.7 2.3

Pinus, spp. 5.6 1.8 - 49.7 + 2.2 3.8 3.4 5.1 1.5 - 7.1 - 3.3 0.6

Prunus, spp. 1.2 + 13.0 + 0.6

Quercus alba 26.5 4.5 - 2.5 - 20.9 28.8 9.1 - 9.0 0.3 - 4.4 1.4 - 0.1

Quercus coccinea 4.6 4.7 3.1 12.4 3.8 2.8 4.5 1.3 - 7.1 3.7 - 6.7 1.3

Quercus prinus 11.3 22.7 + 0.7 - 33.8 + 9.6 7.6 23.1 22.4 20.3 30.3 36.7 30.0

Quercus rubra 0.3 4.3 + 0.7 1.2 1.7 0.6 7.0 + 4.8 + 8.6

Quercus velutina 22.3 1.3 - 0.7 - 5.4 - 5.8 0.9 - 10.3 0.5 - 18.1 2.2 - 13.3 2.9 -

Sassafras albidum 0.5 2.3 0.5 6.5 1.0

Tsuga canadensis 0.8 9.4 + 2.0 1.3 1.9 23.5 + 1.3 7.7 + 2.2 4.0 6.7 2.4

Other 5.0 8.8 8.4 4.0 17.3 7.0 7.7 4.9 2.8 4.8 6.7 8.6

Total # changes 12 9 4 6 10 10 3

Ridgetop Valley Riparian North-facing Ridge South-facing Ridge


Ridgetop forests have undergone least change of any landform, with only 3 species changing significantly in frequency. Ridgetops and other xeric, unproductive sites can serve as refugia from competition for relatively shade intolerant species such as oak and hickory (Collins and Carson 2004). Riparian areas had a greater mix of species during the settlement era, perhaps because they served as refuges from fire. Thus, 20th century decreases in fire frequency would have had less impact in these areas. Witness tree analysis indicates that while a major species shift from oak/hickory/chestnut to oak/maple/birch communities is occurring across most of the landscape, this shift is less pronounced in areas where current disturbance regimes more closely resemble presettlement regimes.

Unburned Ridge

0

100

200

300

400

500

600

5 10 15 20 25 30 35 40 45 50 55 60 65 70 75Diameter Class (cm)

trees

/ ha

Quercus

Non-Quercus

a.

Burned Ridge

0

50

100

150

200

5 10 15 20 25 30 35 40 45 50 55 60 65 70 75

Diameter Class (cm)

trees

/ ha

Quercus

Non-Quercus

b.

Figure 2. Diameter distributions for A) unburned and B) burned ridge stand at NGTC-FIG. Note different y-axis scale for B.


FIRE ECOLOGY To assess the ecological effect of fire, five recently burned stands were compared with

nearby stands that had no recent fire history (for details on materials and methods, see Signell et. al. 2005 and Signell and Abrams 2006). All stands were 50-60 years old, and burned stands had experienced three or more fires in the last 30 years. Fires have thinned burned stands to the advantage of oak, which was better represented in the smaller diameter classes than in unburned stands (figure 2).

Frequently burned stands also had a higher percentage of oak in the overstory and as understory saplings. One reason for this difference is that competitors such as red maple, sassafras and black birch sustained significantly greater fire-related injury than oaks, and presumably experienced higher mortality rates (table 2). Because oak develops resistance to fire more quickly than most competitors, the fact that fire commenced fairly early in stand development was likely vital to increased oak importance (Harmon 1984).

Table 2. Proportion of trees with catfaces (open injuries) in burned stands at Fort

Indiantown Gap. Values with the same letter are not statistically different at the 0.05 level of significance

n

proportion of catfaces

Acer rubrum 62 0.71 a Sassafras albidum 26 0.50 a,b Betula lenta 60 0.45 b Nyssa sylvatica 27 0.44 b,c Quercus alba 121 0.31 c Quercus coccinea 77 0.18 d Quercus velutina 36 0.17 d Quercus montana 256 0.14 d

Understory density in burned stands fell within the ranges reported in Midwestern oak

forests before the fire suppression era, where small stem (5-10 cm) densities were 16-20 stems/ha (Schnur 1937), and 44-64/ha (Gevorkiantz and Scholz 1948). As a consequence of lower tree density, oak regeneration was much more abundant in burned stands, which had mean sapling density (stems >1.22 m tall, <7.6cm dbh) of 901 saplings/ha, whereas oak sapling density was only 1 sapling/ha in unburned stands. Oak regeneration in burned NGTC-FIG stands was much more successful than in other oak forests in the region (Abrams and Downs 1990; Abrams and Nowacki 1992; Steiner et al. 2004). For example, in a study of 52 mature Pennsylvania State Forest stands, mean oak sapling density was only 155.7/ha (Steiner et al. 2004).


SPATIAL PATTERNS OF OAK REGENERATION Within burned stands, oak regeneration was clustered, with saplings concentrated in areas

of low canopy and sub-canopy tree density. Fire had thinned some areas more than others, producing a spatially clumped stem pattern (figure 3A), and a bell-shaped rather than reverse-J diameter distribution (figure 2B). Such a pattern likely resulted from heterogeneity in fire intensity (Rebertus et al. 1989). Fire in pre-settlement oak forests is thought to have aided oak recruitment by removing understory competitors (Abrams 1992, Lorimer 1993). Our data supports this assertion with the caveat that the overstory must also be thinned sufficiently in order for oak to regenerate successfully. Oak saplings were virtually absent if overstory or understory tree density exceeded 400 or 200 trees/ha respectively, (figure 4). Most plots in burned stands had understory density less than 200/ha. Therefore, in burned stands, oak sapling distribution was largely determined by overstory rather than understory density, with saplings largely restricted to canopy gaps.

-3

-11

3L̂

A)

0 5 10 15

-1.5

0.0

1.0

Distance (m)

L^

B)

Figure 3. Stem maps and point pattern analyses for overstory trees in a burned plot (A) and nearby unburned plot (B). X-axes show the range of distances (1 to 15 m) at which L(d) was calculated. Solid lines are observed values and dotted lines define the 95 percent confidence envelopes. Statistically significant (p < 0.05) clustering of stems is indicated by observed values that fall above the upper confidence bound and statistically significant uniformity is depicted by observed values that fall below the lower confidence bound.

High tree density may explain the relative lack of oak saplings in unburned stands, where spatial patterns indicate that competition rather than disturbance has played the key role in stand development. The uniform spatial pattern of overstory stems (Figure 3B) points to density-dependent mortality (increased mortality in higher density patches), which drives clustered and random stem distributions towards uniformity (Leps and Kindlmann 1987, Kenkel 1988). In a highly competitive understory environment, advantage is conferred to shade tolerant plants. As is typical across the region, unburned stands supported high densities of relatively shade tolerant competitors, in this case Acer rubrum, Betula lenta and Nyssa sylvatica. Unburned forests also had much lower shrub cover than burned forests. We found evidence that this lack of shrub cover was due partially to overhead competition. Shrub


cover was greater beneath the sparse crowns of shade intolerant Q. coccinea and S. albidum then under the denser crowns of A. rubrum, B. lenta and N. sylvatica (Canham et al. 1994, Signell and Abrams 2006).

A)

B)

Figure 4. Mean sapling density in relation to A) overstory and B) understory tree density of plots in burned and unburned stands at NGTC-FIG. Tolerant saplings include Acer rubrum, Amalanchier, spp. and Nyssa sylvatica; intolerant saplings include Betula lenta, Prunus serotina, Sassafras albidum, Robinea pseudoacacia and Castanea dentata.

INFLUENCE OF ROCKY FEATURES ON VEGETATION

Riparian areas, coves and ridgetops have been suggested as potential spatial refuges for

fire-sensitive species during presettlement fires in the Eastern U.S. (Abrams and McCay 1996; Abrams 2003). Another candidate is small-scale rocky features, which are widely distributed across ridge sites in the central Appalachians (Hack and Goodlett 1960; Clark 1968; Middlekauff 1991). Rocky patches have been identified as spatial refuges in other ecosystems, for example in the western U.S. and Australia (Burkhardt and Tisdale 1976; Clarke 2002; Price et al. 2003). Following fires, species are able to expand their range


outward into the surrounding fire-adapted communities (Burkhardt and Tisdale 1976). We investigated the role of rocky patches as refugia using two way analysis of variance (ANOVA) to study the main and interaction effects of fire and rock cover on vegetation. Our balanced design consisted of 40 pairs of rocky and non-rocky plots: 20 in recently burned stands, and 20 in stands with no evidence of recent fire ("unburned" stands). We controlled for aspect, slope and soil by selecting only stands with low slope, south aspect and Laidig or Buchanan soil series.

At NGTC-FIG, many small-scale rocky patches were periglacial features, having formed during the last ice-age as a result of frost activity (Clark 1968, Ciolkosz et al. 1986). In areas where resistant sandstone was underlain by relatively unresistant rock such as shale, repeated freezing and thawing pushed up and reoriented subsurface sandstone blocks so that they become oriented vertically and anchored firmly in the subsoil (figure 5; Ciolkosz et al. 1986, Clark et al. 1992). Such periglacial "patterned ground" is relatively widespread throughout the central Appalachians south of the glacial border (Clark 1968). The distribution of periglacial features is often uncorrelated with larger topographic features; for instance they can occur seemingly at random along otherwise unvarying side slopes (Hack and Goodlett 1960).

Figure 5. Photograph of rocky patch. Vertically aligned boulders in the foreground are characteristic of periglacial patterned ground.

Our evidence indicates that rocky patches provided refuge for fire-sensitive species during fires (Signell and Abrams 2006). Three of the four non-Quercus species analyzed (A. rubrum, S. albidum and N. sylvatica) showed both a negative fire effect and a significant interaction term, indicating that rocks and fire together effected the distribution of these species in burned stands (Table 3). Reduced fire severity on rocky patches is responsible for this interaction effect: Quercus on rocky patches had fewer fire scars and catfaces than in matrix forests. The close grouping of burned and unburned rocky plots in the DCA ordination provides further evidence that fire had less impact on rocky patch communities (figure 6).


Table 3. F-values and significances (P) from the ANOVA tables, showing the effects of substrate (rocky/non-rocky), fire (burned/unburned), and their interaction on the

density of selected species. Cells in bold type indicate significant effects of the factor on each species

Effect Substrate Fire Interaction Acer rubrum F 2.22 17.82 7.64 P 0.1407 0.0001 0.0072 Betula lenta F 15.45 21.32 1.58 P 0.0002 <0.0001 0.2120 Nyssa sylvatica var. sylvatica F 15.24 22.06 5.76 P 0.0002 <0.0001 0.0188 Sassafras albidum F 2.13 9.75 4.95 P 0.1486 0.0025 0.0290 Quercus coccinea F 6.66 2.65 0.88 P 0.0118 0.1076 0.3503 Quercus montana F 0.03 1.84 1.31 P 0.8562 0.1790 0.2564 Quercus rubra F 9.16 1.65 1.68 P 0.0034 0.2024 0.1995

Several factors may contribute to reduced severity on rocky patches. First, by confining

fuels to interstices, rocks create circuitous fuel routes which slow the rate of spread (Veblen et al. 1994). Second, soil in rock interstices is shadier and cooler than in non-rocky areas, which increases moisture content and reduces flammability (Swan 1961; Lauer and Klaus 1975; Pearson and Theimer 2004). Third, by relegating fire to a relatively small percentage of the soil surface, rocks may reduce root mortality in susceptible species such as B. lenta (Niering et al. 1970).


Figure 6. Diagram of 80 plots, and major tree species along detrended correspondence analysis (DCA) axes one and two.

B. lenta was unique in that it was negatively impacted by fire, but did not have a significant interaction effect. For this species, therefore, fire may merely amplify a preexisting preference for rocky patches; the cool, damp, humus-rich microenvironments of rock interstices provide ideal germination sites for B. lenta (Swan 1961; Lauer and Klaus 1975; Pearson and Theimer 2004; Carlton 1993).

While rocky patches served as refuges for fire-sensitive trees, the burned matrix was a safe haven for early to mid-successional species such as Quercus and Sassafras. Of the major species, Quercus were the best adapted to burned matrix forests. As a result of lower scarring rates, none of the Quercus species tested showed a significant fire effect (Table 3). Conditions in the burned matrix were favorable for Quercus regeneration, which, along with Sassafras dominated the seedling and sapling classes (table 4). The shade intolerant shrubs Gaylusaccia and Vaccinium also thrived in the burned matrix, forming a nearly continuous layer. Fire benefits these plants via stimulation of root-suckering and through overstory thinning (Matlack et al. 1993). Ericaceous shrubs such as these are important surface fuels in eastern forests (Brose et al. 2004), providing fuel in and of themselves, but also increasing flammability by trapping and suspending leaf litter above the forest floor (Dighton et al. 2000).

In burned forests, heterogeneity in rock cover creates a mosaic where fire-adapted species dominate the landscape, but where other species may also persist in isolated pockets of lower fire severity. This finding provides some insight into historical forest dynamics of the region. For example, the relegation of species like A. rubrum, B. lenta and N. sylvatica to small but widely distributed rocky refugia during presettlement fires may help explain both their low frequency in ridge forests during European colonization, as well as their rapid increase during


the 20th century fire suppression era (Nowacki and Abrams 1992; Abrams and Ruffner 1995; Abrams and McCay 1996; Steiner et. al. 2004).

Table 4. Seedling and sapling regeneration (stems/ha) for each patch type in study

stands at Fort Indiantown Gap

Species Burned Matrix Burned Rocky Unburned Matrix Unburned Rocky

Seedlings Saplings Seedlings Saplings Seedlings Saplings Seedlings Saplings

Quercus alba 618 - - - 124 - - -

Quercus coccinea 371 371 - - - - - -

Quercus montana 13,837 618 7,660 - 2,347 - 618 -

Quercus rubra 741 - 618 - 1,853 - 1,359 -

Quercus velutina 741 - - - - - 247 -

Acer rubrum 9,884 - 13,220 - 12,849 - 5,436 -

Carya sp. - - - 124 - - - -

Betula lenta 2 347 - 1,853 - 494 - 1,482 - Liriodendron tulipifera 618 - - - 2,471 - - -

Nyssa sylvatica 1 606 - 618 - 741 248 124 124

Pinus strobus 124 - 124 - 247 - 618 124 Populus grandidentata - 124 - - - - - -

Prunus serotina 124 - - - - - 247 -

Sassafras albidum 27,181 4,695 11,984 - 2,965 - 988 -

Tsuga canadensis - - 124 - - - 124 -

Total 58,193 5,807 36,200 124 24,092 248 11,243 248 Indeed, unburned NGTC-FIG stands contained high densities of relatively shade tolerant

A. rubrum, B. lenta and N. sylvatica, leaving few opportunities less tolerant plants to regenerate. No Quercus or Sassafras saplings were found in unburned stands (table 4). Some have suggested that rocky patches may provide refuge for early successional species in the face of increasing competition from shade-tolerant competitors (Frelich and Reich 2002). We found no evidence of this however, at least at this stage of stand development; Quercus and Sassafras saplings were absent on unburned rocky patches despite lower tree density.

CONCLUSION Over time, fire suppression may lead to a loss of patchiness and a subsequent decline in

species richness as disturbance-adapted species such as Quercus are replaced by shade tolerant, later successional species (Loucks 1970). Long periods without fire can result in a loss of landscape diversity as fire-adapted plant communities disappear. For example, Niklasson et al. found that Quercus-Pinus forests in Sweden had been completely replaced by


Fagus-Picea forests following the cessation of fires during the late 18th century (Niklasson et al. 2002). Fagus had survived fires in a refuge made up of large boulders in the center of the study area, expanding outward when fires stopped. Once such fundamental shifts in landscape pattern occur, the original patchwork of communities may be difficult, if not impossible to restore (Heinselmann 1973; Heyerdahl et al. 2001).

Growing concern over the future of Quercus forests has led many to consider prescribed fire as a means of ecosystem and oak restoration (Abrams 1992; Baker 1994; Ford et al. 1999; Blake and Schuette 2000; Brose et al. 2001). However, most prescribed fires in this region have resulted in little or no benefit to oak regeneration due to low mortality rates of competing species (Wendell and Smith 1987; Arthur et al. 1998; Kuddes-Fischer and Arthur 2002). Most eastern forests have aged to the point where even fire-sensitive species have grown large enough to survive moderate surface fires (Harmon 1984). Decreases in shrub cover, such as those documented here, may also make forests less flammable, decreasing the efficacy of prescribed fires even further (Mutch 1970).

In eastern U.S. Quercus forests, as elsewhere, management prescriptions should be based on an understanding of current forest conditions in the context of historic disturbance regimes (Heinselmann 1973; Romme and Despain 1989; Heyerdahl et al. 2001; Motzkin and Foster 2002). Such an understanding would be incomplete without considering the factors controlling spatial variability in fire regimes and regeneration of key tree species such as oak.

REFERENCES

Abrams, M. D. Fire and the Development of Oak Forests in Eastern North America; Oak Distribution Reflects a Variety of Ecological Paths and Disturbance Conditions. Biosci. 1992, 42, 346-353.

Abrams, M. D.; Downs, J. A. Successional replacement of old-growth white oak by mixed mesophytic hardwoods in southwestern Pennsylvania. Can. J. For. Res. 1990, 20, 1864-1870.

Abrams, M. D.; McCay, D. M. Vegetation-site relationships of witness trees (1780-1856) in the presettlement forests of eastern West Virginia. Can. J. For. Res. 1996, 26, 217-224.

Abrams, M. D.; Nowacki, G. J. Historical Variation in Fire, Oak Recruitment, and Post-Logging Accelerated Succession in Central Pennsylvania. J. Tor. Bot. Soc. 1992, 119, 19-28.

Abrams, M. D.; Ruffner, C. M. Physiographic Analysis of Witness-Tree Distribution (1765-1798) and Present Forest Cover through North Central Pennsylvania. Can. J. For. Res. 1995, 25, 659-668.

Alerich, C. L. Forest Statistics for Pennsylvania:1978-1989. Resour. Bull. NE-126, U.S. Department of Agriculture. Forest Service, Northeastern Forest Experiment Station: Radnor, PA.1993.

Anonymous, D. NGTC-FIG Final Integrated Natural Resources Management Plan. Ft. Indiantown Gap National Guard Training Center: Annville, Pa.2002.

Arthur, M. A.; Paratley, R. D.; Blankenship, B. A. Single and repeated fires affect survival and regeneration of woody and herbaceous species in an oak-pine forest. J. Torr. Bot. Soc. 1998, 125, 225-236.


Baker, W. L. Restoration of landscape structure altered by fire suppression. Cons. Biol. 1994, 8, 763-769.

Barnes, B. V.; Wagner, W. H. Michigan Trees; University of Michigan Press: Ann Arbor, MI, USA.1981.

Black, B. A.; Abrams, M. D. Analysis of temporal variation and species-site relationships of witness tree data in southeastern Pennsylvania. Can. J. For. Res. 2001, 31, 419-429.

Blake, J. G.; Schuette, B. Restoration of an oak forest in east-central Missouri. Early effects of prescribed burning on woody vegetation. For. Ecol. Manage. 2000, 139, 109-126.

Braun, E. L. Deciduous forests of eastern North America; Hafner Press: New York.1950. Brose, P.; Schuler, T.; Ward, J. S. Integrating prescribed fire into management of mixed-oak

forests of the mid-Atlantic region: developing basic fire behavior and fuels information for the SILVAH system (abstract). GTR-NE-316, USDA Forest Service: Wooster, OH.2004.

Budd, T. Good Order Established in Pennsylvania; University Microfilms: University Park, PA.[1685]1966.

Buell, M. F.; Buell, H. F.; Small, J. A. Fire in the history of Mettler's Woods. J. Tor. Bot. Soc. 1954, 81, 253-255.

Burkhardt, J. W.; Tisdale, E. W. Causes of juniper invasion in southwestern Idaho. Ecology. 1976, 57, 472-484.

Byrd, W. The Writings of Colonel William Byrd; Burt Franklin: New York, NY.1728. Canham, C. D.; Finzi, A. C.; Pacala, S. W. Causes and consequences of resource

heterogeneity in forests: interspecific variation in light transmission by canopy trees. Can. J. For. Res. 1994, 24, 337-349.

Carlton, G. C. Effects of microsite environment on tree regeneration following disturbance. Dissertation. Harvard University, Cambridge, MA, USA.1993.

Ciolkosz, E. J.; Cronce, R. C.; Sevon, W. D. Periglacial Features in Pennsylvania. 92, Agronomy Department, The Pennsylvania State University: University Park, PA.1986.

Clark, G. M. Sorted patterned ground: new Appalachian localities south of the glacial border. Science. 1968, 161, 355-356.

Clark, G. M.; Behling, R. E.; Braun, D. D.; Ciolkosz, E. J.; Kite, J. S.; Marsh, B. Central Appalachian periglacial geomorphology. Agronomy Department, The Pennsylvania State University: University Park, PA.1992.

Clarke, P. J. Habitat islands in fire-prone vegetation: do landscape features influence community composition? J. Biogeog. 2002, 29, 677-684.

Collins, R. J.; Carson, W. P. The effects of environment and life-stage on Quercus abundance in the eastern deciduous forest, USA: are sapling densities most responsive to envirnonmental gradients? For. Ecol. Manage. 2004, 201, 241-258.

Day, G. M. The Indian as an ecological factor in the northeastern forest. Ecology. 1953, 34, 329-346.

Delcourt, P. A.; Delcourt, H. R.; Webb III, T. Atlas of mapped distribution of dominance and modern pollen percentages for important tree taxa of eastern North America; American Association of Stratigraphic Palyologists Foundation: Dallas, TX.1984.

Dighton, J.; Bonilla, A. S. M.; Jiminez-Nunez, R. A.; Martinez, N. Determinants of leaf litter patchiness in mixed species New Jersey Pine Barrens forest and its possible influence on soil and soil biota. Biology and Fertility of Soils. 2000, 31, 288-293.

Dove, J. (2004). Manada Furnace. http://www.easthanoverhistory.com


Fletcher, S. W. Pennsylvania Agriculture and Country Life (1640-1840); Pennsylvania Historical and Museum Commission: Harrisburg, PA.1950.

Ford, W. M.; Menzel, M. A.; McGill, D. W.; Laerm, J.; McCay, T. S. Effects of a community restoration fire on small mammals and herpetofauna in the southern Appalachians. For. Ecol. Manage. 1999, 114, 233-243.

Foster, D. R.; Clayden, S.; Orwig, D. A.; Hall, B. Oak, chestnut and fire: climatic and cultural controls of long- term forest dynamics in New England, USA. J. Biogeog. 2002, 29, 1359-1379.

Gevorkiantz, S. R.; Scholz, H. F. Timber yields and possible returns from the mixed-oak farmwoods of southwestern Wisconsin. Publ. No. 521, U.S. Department of Agriculture, Forest Service, Lake States Forest Experimental Station.1948.

Hack, J. T.; Goodlett, J. C. Geomorphology and forest ecology of a mountain region in the Central Appalachians. Geological Survey Professional Paper 347, United States Department of the Interior Geological Survey: Washington, D.C.1960.

Heinselmann, M. L. Fire in the virgin forests of the Boundary Waters Canoe Area, Minnesota. Quat. Res. 1973, 3, 329-382.

Heyerdahl, E. K.; Brubaker, L. B.; Agee, J. K. Spatial controls of historical fire regimes: A multiscale example from the interior west, USA. Ecology. 2001, 82, 660-678.

Kenkel, N. C. Pattern of self-thinning in jack pine. Ecology. 1988, 69, 1017-1024. Kuddes-Fischer, L. M.; Arthur, M. A. Response of understory vegetation and tree

regeneration to a single prescribed fire in oak-pine forests. Nat. Areas J. 2002, 22, 43-52. Lauer, W.; Klaus, D. Geoecological investigations of the timberline of Pico de Orizaba,

Mexico. Arctic and Alpine Research. 1975, 7, 315-330. Leps, J.; Kindlmann, P. Models of the development of spatial pattern of an even-aged plant

population over time. Ecol. Modell. 1987, 39, 45-57. Lorimer, C. G. Survival and growth of understory trees in oak forests of the Hudson

Highlands, New York. Can. J. For. Res. 1981, 11, 689-695. Lorimer, C. G. The oak regeneration problem: new evidence on causes and possible solutions.

Forest Resources Analyses No. 8, Univ. of Wisconsin Publ.: Madison, WI.1989. Lorimer, C. G. Causes of the Oak Regeneration Problem. USDA Forest Service: Knoxville,

TN.1993. Loucks, O. L. Evolution of diversity, efficiency, and community stability. American

Zoologist. 1970, 10, 17-25. Matlack, G. R.; Gibson, D. J.; Good, R. E. Clonal Propagation, Local Disturbance, and the

Structure of Vegetation - Ericaceous Shrubs in the Pine-Barrens of New-Jersey. Biol. Conserv. 1993, 63, 1-8.

Maxwell, H. The use and abuse of forests by the Virginia Indians. William and Mary Quarterly Historical Magazine. 1910, 19, 73-103.

Middlekauff, B. D. Probable paleoperiglacial morphosequences in the Northern Blue Ridge. Physical Geography. 1991, 12, 19-36.

Motzkin, G.; Foster, D. R. Grasslands, heathlands and shrublands in coastal New England: historical interpretations and approaches to conservation. J. Biogeog. 2002, 29, 1569-1590.

Motzkin, G.; Patterson, W. A.; Foster, D. R. A historical perspective on pitch pine-scrub oak communities in the Connecticut Valley of Massachusetts. Ecosystems. 1999, 2, 255-273.


Niering, W. A.; Goodwin, R. H.; Taylor, S.Prescribed burning in Southern New England: Introduction to long-range studies; Tall Timbers Fire Ecology Conference; Fredericton, New Brunswick, 1970, 267-286

Niklasson, M.; Lindbladh, M.; Bjorkman, L. A long-term record of Quercus decline, logging and fires in a southern Swedish Fagus-Picea forest. J. Veg. Sci. 2002, 13, 765-774.

Nowacki, G. J.; Abrams, M. D. Community, Edaphic, and Historical-Analysis of Mixed Oak Forests of the Ridge and Valley Province in Central Pennsylvania. Can. J. For. Res. 1992, 22, 790-800.

Nowacki, G. J.; Abrams, M. D.; Lorimer, C. G. Composition, structure, and historical development of northern red oak stands along an edaphic gradient in northcentral Wisconsin. For. Sci. 1990, 36, 276-292.

Orwig, D. A.; Abrams, M. D. Land-Use History (1720-1992), Composition, and Dynamics of Oak Pine Forests within the Piedmont and Coastal-Plain of Northern Virginia. Can. J. For. Res. 1994, 24, 1216-1225.

Pearson, K. M.; Theimer, T. C. Seed-caching responses to substrate and rock cover by two Peromyscus species: implications for pinyon pine establishmen. Oecologia. 2004, 141, 76-83.

Pickett, S. T. A.; Cadenasso, M. L. In Patch Dynamics and the ecology of disturbed ground: a framework for synthesis; L. R. Walker, Ed.; Elsevier: Amsterdam, 1999; pp 707-721

Pubanz, D. M.; Lorimer, C. G.; Guries, R. P.Effects of understory control on survival and vigor of red oak seedlings beneath a shelterwood; Seventh Central Hardwood Conference; Carbondale, IL, 1989, 300-301

Rebertus, A. J.; Williamson, G. B.; Moser, E. B. Fire-Induced Changes in Quercus-Laevis Spatial Pattern in Florida Sandhills. J. Ecol. 1989, 77, 638-650.

Romme, W. H.; Despain, D. G. Historical Perspective on the Yellowstone Fires of 1988. Biosci. 1989, 39, 695-699.

Russell, E. W. B. Indian-set fires in the forests of the northeastern United States. Ecology. 1983, 64, 78-88.

Schnur, G. L. Yield, stand, and volume tables for even-aged upland oak forests. Technical Bulletin 560, U.S. Department of Agriculture.1937.

Shumway, D. L.; Abrams, M. D.; Ruffner, C. M. A 400-year history of fire and oak recruitment in an old-growth oak forest in western Maryland, USA. Can. J. For. Res. 2001, 31, 1437-1443.

Signell, S. A.; Abrams, M. D. Influence of rocky landscape features and fire regime on vegetation dynamics in Appalachian Quercus forests. J. Veg. Sci. 2006, 17, 675-684.

Signell, S. A.; Abrams, M. D.; Hovis, J. C.; Henry, S. W. Impact of multiple fires on stand structure and tree regeneration in central Appalachian oak forests. For. Ecol. Manage. 2005, 218, 146-158.

Smoker, F. H. 2003. Back at the Gap. in Lebanon Daily News, Lebanon, Pa. Steiner, K. C.; Finley, J. C.; McDill, M. D.; Fei, S.; Gould, P. J.; Phelps, T. R.; Grinsted, S. C.

Report of 2004 Regeneration Assessment. Penn State Regeneration Project, The Pennsylvania State University: University Park, PA.2004.

Stover, M. E.; Marks, P. L. Successional vegetation on abandoned cultivated and pastured land in Tompkins County, New York. J. Torr. Bot. Soc. 1998, 125, 150-164.

Swan, L. The ecology of the high Himalayas. Scientific American. 1961, 205, 68-78.


Swanson, F. J.; Kratz, T. K.; Caine, N.; Woodmansee, R. G. Landform Effects on Ecosystem Patterns and Processes. Biosci. 1988, 38, 92-98.

Taylor, A. H.; Skinner, C. N. Spatial patterns and controls on historical fire regimes and forest structure in the Klamath Mountains. Ecol. Appl. 2003, 13, 704-719.

Turner, M. G.; Romme, W. H.; Gardner, R. H. Landscape Disturbance Models and the Long-Term Dynamics of Natural Areas. Nat. Areas J. 1994, 14, 3-11.

Veblen, T. T.; Hadley, K. S.; Nel, E. M.; al., e. Disturbance regime and disturbance interactions in a Rocky Mountain subalpine forest. J. Ecol. 1994, 82, 125-135.

Wallace, P. A. W. Indian Paths of Pennsylvania; Pennsylvania Historical and Museum Commission: Harrisburg, PA.1965.

Watts, W. A. Late Quartenary vegetation of the central Appalachian and New Jersey Coastal Plain. Ecological Monographs. 1979, 49, 427-469.

Wendell, G. W.; Smith, H. C. Effects of a Prescribed Fire in a Central Appalachian Oak-Hickory Stand. USDA, Forest Service, Northeastern Forest Experiment Station, NE-RP-594.1987.

Whitney, G. G. From Coastal Wilderness to Fruited Plain; Cambridge University Press: Cambridge, United Kingdom.1994.

Documents

INTERACTIONS BETWEEN LANDSCAPE FEATURES … book.pdfaverage temperature during the winter ranging between -7o C and 19 o C and temperatures during the summer ranging between -2o C