INTRODUCTION

Forests of eastern North America experienced a dramatic increase in mesophytic and generally shade tolerant tree species during the twentieth century (Abrams 1998; Brose et al. 2001; Nowacki and Abrams 2008). These species have become nearly ubiquitous across subxeric sites throughout the temperate forest biome. This is thought to be related, at least in part, to the exclusion of fire in most forests, particularly those dominated by pyrogenic species, such as oak (Quercus), hickory (Carya), and pine (Pinus) (Buell et al. 1954; Lorimer 1985; Abrams 1992). In addition, differential deer (Odocoileus virginianus) browsing on oak (favored by deer) versus those species avoided by deer (such as red maple; Acer rubrum) has also acted to promote certain oak-replacement species (Abrams 1998; Abrams and Johnson 2012). Presently, later successional, mesophytic trees dominate the understory and mid-canopy of many forests, and it appears that they have the potential to cause widespread replacement of the historically dominant trees in the eastern United States. This process is also changing the forest microenvironment and reducing the likelihood of restoring the natural fire cycle in these ecosystems; thus, it has been termed “mesophication” (Nowacki and Abrams 2008).

Periodic burning of pyrogenic forests by Native Americans was arguably a key (but not the only) factor in their long-term sustainability, while limiting mesophytic tree dominance in the presettlement forests (Brose at el. 2001; Guyette et al. 2002; Abrams and Nowacki 2008). However, given current conditions of the ecology and management of eastern forests, the increasing importance of red maple and other species seems inevitable. The loss of oak, hickory, and pine dominance in eastern U.S. forests will be a primary consequence of the continued expansion of later successional tree species. Forest managers at nature preserves and natural areas throughout the eastern U.S. are concerned about the lack of regeneration of desired, native tree species that will form the next generation of canopy individuals (Abrams and Hayes 2008; Abrams and Sands 2010). The use of periodic understory prescribed burning to reduce forest floor litter depth and plant competition and increase understory light intensity, coupled with other ecosystem management practices (e.g., canopy thinning), may be an effective tool for restoring the ecological integrity of these forests while halting the “mesophication” process (Wang et al. 2005; Royse et al. 2010; Fan et al. 2012). These practices may also help prevent the formation of unwanted, novel ecosystems with no historic precedent (Hobbs et al. 2009).

This research investigated the status of oak regeneration on dry sites at the Mohonk Preserve in the Shawangunk Mountains of eastern New York, with a particular emphasis on red oak and chestnut oak forests (Q. rubra and Q. montana). After a century or more of fire suppression, prescribed fire was reintroduced to several sites at the Mohonk Preserve in 2009 and 2011. In addition, a mast year with abundant red oak acorn production occurred at Mohonk in the fall of 2010, which likely impacted oak seedling density. This research investigated the impacts mast year and prescribed fires, including possible synergism, by comparing temporal variation in tree regeneration on burned and unburned sites.

Site Description

This 2009 to 2012 research was conducted at the Mohonk Preserve (41°45′29″N, 74°09′28″W), which was established in 1869 in the Shawangunk Mountains of eastern New York (Figure 1). Pollen extracted from lake sediment showed an increase in oak and pine dominance throughout southeastern New York over the past 9000 years (Maenza-Gmelch 1997). This coincided with an influx of charcoal in lake sediments, suggesting that fire played an important role in the development of woodlands in this area. Delaware Indians inhabited the Hudson Valley and were known to deliberately burn forest areas (Buell et al. 1954; Josephson 2002). Throughout the nineteenth and early twentieth centuries, more extensive logging of forests occurred to supply wood to local sawmills for charcoal iron production and the tanning industry (Snyder and Beard 1981; Josephson 2002). This was followed by the harvesting of American chestnut (Castanea dentata), killed by the chestnut blight (Cryphonectria parasitica) in the early 1900s. The first outbreak of the gypsy moth (Lymantria dispar) on the Mohonk Preserve occurred in 1957, resulting in oak mortality on the ridges. More recently, fires have been common on the Shawangunk Ridge (Laing 1994). Chestnut oak and red oak forests cover approximately 1420 ha of the Mohonk Preserve.

Figure 1.

Location of Mohonk Preserve in southeastern New York (inset) and the seven burned and unburned study stands within the preserve (A = Old Stage 1; B = Glory Hill 2; C = Old Stage 2; D = Old Minnewaska 2; E = Old Minnewaska 1; F = Oakwood 2; G = Oakwood 3).

We chose seven study sites that occurred on the Martinsburg shale bedrock of the Nassau-Manilus soil type, a shaly silt loam that is moderately deep and well drained (USDA Soil Conservation Service 1979; Table 1; Figure 1). These sites ranged from 267 to 357 m in elevation and occupied ridgeline to upper and middle side slope positions. Their aspect varied from east (mesic) to west (more xeric) and they had a slope of about 5% to 25%. Three of the study sites (Old Stage 1, Oakwood 2 (in 2009), and Oakwood 3) are unburned, whereas the understory of the five other sites were burned in the spring of 2009 (Oakwood 2) or spring 2011 (Table 2). Oakwood 2 is considered an unburned site in 2009 and a burned site in 2011 and 2012. The red oak acorns produced in the fall of 2010 did not germinate until after the spring 2011 burns.

Table 1.

Study site names, burn status, and other physical features of the seven burned and unburned study sites on shale soils on the Mohonk Preserve..

METHODS

During July 2011 and June 2012, regeneration surveys were conducted at the seven study sites within the Mohonk Preserve. Additionally, the three unburned stands were surveyed in May 2009. Ten permanently marked (in either 2009 or 2011), fixed-area plots located at approximately 15-m intervals along transects (typically 200 – 300 m in length) through the forest interior were used for vegetation sampling. The location of these plots was recorded by GPS, which was also used to locate a few of the plots with missing markers in 2012. Saplings and seedlings (excluding first year germinants) were counted in nested circular plots of 20 m² and 10 m², respectively, located at each plot center. Saplings are classified as tree species > 1.4 m in height but < 10.0 cm in dbh and seedlings were < 1.4 m. Tree species (> 10 cm dbh) density was recorded by species in a 200-m² circular plot at each understory plot center in July 2011. A general description of each site was recorded, including landscape position, soil type, elevation, and aspect.

RESULTS AND DISCUSSION

A total of 13 tree species were recorded in the overstory on the seven study sites in 2011 (Table 2). Tree density across all stands was fairly similar, averaging 553 trees per ha and ranging from 455 – 675 trees per ha. The 2009 and 2011 prescribed fires appeared to have little impact on overstory tree density via mortality. Overall the study sites were dominated by different combinations of red oak, red maple, chestnut oak, and white pine (P. strobus; Table 2 and Figure 2). The stands with the highest tree densities overall (Glory Hill 2 and Oakwood 2) were dominated by these two oaks. Red oak averaged 39% of the total tree density and dominated five stands. Red maple tree density averaged 21% (range from 0% – 46%) and dominated two stands. Chestnut oak averaged 18% of the total density and was not the number one dominant in any of the study sites. Pignut hickory (C. glabra) was present at low to moderate levels across all study sites.

A total of 13 tree species were recorded as seedlings with an average density of 19,300 per ha on the three unburned study sites in 2009 (including Oakwood 2; Table 3). This compares to an average density of 32,300 seedlings/ha in 2011 and 17,300 seedlings/ha in 2012. The large increase in seedlings in 2011 is due primarily to an increase in red oak following the 2010 mast year. Seedling density for the other oak species did not increase in 2011, which suggests that 2010 was a mast year for red oak only. Old Stage 1 and Oakwood 3 experienced a large drop in seedling density (mainly red oak) between 2011 and 2012 not seen at Oakwood 2, by which time the latter site had been burned (Tables 3 and 4). The other dominant seedling on unburned sites was chestnut oak, although it was almost nonexistent at Old Stage 1. In contrast, red maple, striped maple (A. pensylvanicun), sweet birch (Betula lenta), witch hazel (Hamamelis virginiana), white pine, and hophornbeam (Ostrya virginiana) seedlings were more important at Old Stage 1 than the other two unburned sites. Hophornbeam seedling density dramatically increased, while witch hazel seedlings decreased, at Old Stage 1 between 2009 and 2011.

A total of 15 tree species were recorded as seedlings on the five burned study sites (Table 4; Figure 2). Seedling density was highest at Old Minnewaska 2 in 2011 (60,500/ha) and lowest at Old Stage 3 in 2012 (10,000/ha). Red oak and chestnut oak were the dominant seedlings across all sites and years, with the exception of red maple at Old Stage 3 in 2011. Between 2011 and 2012, seedling density remained stable at Oakwood 2, increased by about 20% at Glory Hill 2 (mainly hophornbeam and chestnut oak), declined by 33% – 35% at Old Minnewaska 1 and 2, and declined by 68% at Old Stage 1. The decline in seedling density was mainly due to large reductions in red oak and red maple. This suggests the ephemeral nature of the many newly established seedlings produced as a result of a mast year or prescribed fire. This mortality has been attributed to low understory light levels and changes in the litter depth (Green et al. 2010; Royse et al. 2010; Greenberg et al. 2012), although deer browsing and insect and disease may also be casual factors at the Mohonk Preserve. It is interesting to note a relatively high number of red maple seedling on OS3 and OM1 that survived the 2011 fires also declined by 2012.

Table 2.

The total (#/ha) and percent density of trees species on the seven burned and unburned (Old Stage 1 and Oakwood 3) Mohonk study sites surveyed in 2011. OS = Old Stage; OW = Oakwood; GH = Glory Hill; OM = Old Minnewaska.

A total of eight tree species were present as saplings on the three unburned sites, dominated by witch hazel, striped maple, and red maple (Table 5). A total of ten tree species occurred as saplings at the five burned sites (Table 6). Old Minnewaska 2 had the highest sapling density ( 1950/ha in both 2011 and 2012), whereas Old Stage 3 in 2012 had the lowest density (350 saplings/ha). The dominant saplings were pignut hickory, striped maple, witch hazel, and red maple, depending on site. Across all seven sites, total sapling density was highest on the unburned Old Stage 1 (4850/ha) due to the amount of witch hazel. The low number of oak saplings at all burned and unburned sites is notable.

Average seedling density was higher on burned sites (31,030/ha versus 23,189/ha), while sapling density was higher on the unburned sites (2077 versus 995/ha; Tables 3–6). Red oak seedling density was generally higher on burned sites (mean = 19,400 versus 9578/ha, respectively) mainly due to its abundance on Old Minnewaska 1 and 2. Red maple seedling mean density was higher on the burned than unburned sites (2740 versus 778/ha). The increase in red maple on burned sites was also reported in the Cumberland Plateau in Kentucky, which was attributed to its vigorous sprouting ability (Chiang et al. 2005; Green et al. 2010). Lower sapling density on burned sites suggests, at least in part, they were killed or transformed to seedling sprouts by fire. Fire may act to increase the proportion of oak seedlings, but may cause an overall decline in total seedling density (Green et al. 2010; Royse et al. 2010). An increase in oak seedlings or seedling sprouts (either relative or total density) may occur following burning due to improved seedbed conditions for germination (e.g., exposed mineral soil, reduced litter layer), increased light and space, vigorous sprouting ability, and reduced competition (Lorimer 1985; Reich et al. 1990; Fan et al. 2012). Using repeated prescribed fire, coupled with canopy thinning, may be more effective at stimulating oak regeneration compared with using only a single burn (McEwan et al. 2011).

Table 3.

Tree seedling density (#/ha) in the unburned Old Stage 1 (OS1), Oakwood 3 (OW3) and Oakwood 2 (OW2) study sites at Mohonk in 2009, 2011, and 2012. Oakwood 2 was burned after our May 2009 survey.

Table 4.

Tree seedling density (#/ha) at the five burned study sites at Mohonk in 2011 and 2012. OW = Oakwood; OS = Old Stage; GH = Glory Hill; OM = Old Minnewaska.

Figure 2.

From top left to bottom right- A) July 2011 overview of the spring 2011 prescribed lire in the forest understory at Glory Hill 2 site; B) Red oak and chestnut oak dominated forest at the unburned Oakwood 3 study site; C) Large cohort of red oak seedlings that germinated after the April burn in Old Minnewaska 2 site following a mast year in 2010; D) Large seedling sprouts of red oak and chestnut oak (in 2011) created after the 2009 fire on the Oakwood 2 site. The photos were taken by Marc Abrams, Penn State University.

Table 5.

Tree sapling density (#/ha) on the three unburned study sites at Mohonk in 2009, 2011, and 2012.

Table 6.

Tree sapling density (#/ha) on the five burned study sites at Mohonk in 2009, 2011, and 2012.

One of the major findings of this study is the dramatic post-mast year increase in red oak seedling density in 2011 (an average of 23,000/ha versus 4267/ha in 2009). Red oak represented the dominant seedling in 2011 at six of seven burned and unburned study sites; whereas in 2009, it was subordinate to chestnut oak and red maple on most sites surveyed in this and a previous study (Abrams and Sands 2010). The prescribed fires in 2009 and 2011 appeared to positively impact red oak seedling numbers, as two burned sites had the highest number of red oak seedlings (mean density of 44,350/ha; Table 4, Figure 2). This suggests a synergistic interaction between fire and mast year for promoting oak seedling density. Indeed, the burned sites had higher average red oak density than the unburned sites, and appeared to be more successful in producing large seedling sprouts in the advance regeneration (Figure 2). However, red oak seedling density also increased by 15,200/ha and 20,900/ha between 2009 and 2011 on two unburned sites, suggesting the overriding importance of mast year on all sites.

The spring 2011 prescribed fires did not negate the beneficial impacts of the 2010 mast years via acorn mortality. Acorn losses likely occurred as a result of the fire, but many of them survived to greatly elevate oak seedling density. A Pennsylvania study reported that about 40% of red oak acorns were killed by a typical fire for the region (Auchmoody and Smith 1993). Significant losses of red oak and white oak acorns following prescribed fires of increasing intensity were also reported in North Carolina forests (Greenberg et al. 2012). On average, nearly 50% of the newly established red oak seedlings in 2011 were lost by 2012 on the burned and unburned sites. Nonetheless, red oak seedling density in 2012 remained 300% higher than that in 2009. Therefore, the impact of the 2010 mast year in stimulating red oak regeneration was still evident two years later. The results of this study indicate the importance of mast years and prescribed fire, both individually and collectively, in stimulating oak regeneration.