Witness trees are an important direct source of quantitative data based on surveys conducted in the period 1600–1900. Land surveyors marked and recorded witness (e.g., warrant, bearing, corner) trees to identify property corners and boundaries during the early European settlement of the eastern United States (Lutz 1930, Bourdo 1956). The original land surveys along the Appalachian Mountains and coastal regions from Maine to Georgia were primarily conducted using metes and bounds of property boundaries, which eventually developed into a more regular and systematic land partitioning following the Northwest Ordinance of 1785. These data can be used in conjunction with contemporary surveys of old-growth forests and palaeoecological analysis to assess the composition of forests prior to European settlement (Lutz 1930, Bourdo 1956, Abrams and Downs 1990, Abrams and Ruffner 1995, Delcourt et al. 1998, Ruffner and Abrams 1998, Black and Abrams 2001, Foster et al. 2002). Despite certain biases and inconsistencies, witness tree data remain an important resource in historical ecological studies and reveal consistent species patterns that accord with other lines of evidence, such as historical accounts (Bourdo 1956, Hanberry et al. 2020, Hanberry 2021).

Witness trees have been helpful in quantifying large-scale changes in tree composition over time. When witness tree data are compared with modern-day forests, researchers can assess how and why they have changed over several centuries or more. Applications include impacts of Native American land use and subsequent European settlement and land-clearing activities as well as helping map historic fire patterns and landscape flammability (Abrams 2010, Thomas-Van Gundy et al. 2015, Hanberry et al. 2018).

Witness tree data have been compiled and analyzed for much of the northern half of the eastern United States (Cogbill 2000). One exception to this is the state of Maryland, where only one witness tree paper from the Eastern Shore has been published (Briand and Folkoff 2019). No witness tree data have been compiled for the mountainous regions of western Maryland and none exist on historical databases, such as PalEON ( http://paleonproject.org/). Western Maryland contains a well-documented old-growth forest on Savage Mountain, which was the location of a fire history and dendroecological study spanning four centuries (Shumway et al. 2001). Savage Mountain extends into southwestern Pennsylvania, where historical witness tree records are more accessible. Our purposes for this study are as follows:

Fig. 1.

The Allegheny Mountain Plateau ecological subsection in Pennsylvania and Maryland, which contains Savage Mountain.

Study Area and Methods.

The Allegheny Mountains represent the high-eastern margin (subsection) of the Allegheny Plateau running southwesterly from north central Pennsylvania to southeastern West Virginia (Fig. 1). Fenneman (1938) reported the subsection as a mixture of mildly folded ridges in anti-, mono-, and synclinal structures separated by dendritically eroded cuesta uplands. Our specific study area is located on Big Savage Mountain in Garrett and Allegany counties, Maryland, running northward into Somerset and Bedford counties, Pennsylvania (39830′–40800′N, 79815′–78845′W). It is located on the southern terminus of Big Savage Mountain in the Savage River State Forest in the Allegheny Mountain section of the Appalachian Plateau. Elevation ranges from about 600 to 670 m. This ridgeline area is characterized by extensive sandstone outcroppings and many areas constitute periglacial scree-talus slopes (Fenneman 1938). Although the area was not glaciated, these rock outcrops constitute periglacial screes. Soils belong to the DeKalb-Gilpin-Cookport association, which are very stony, acidic soils derived largely in residuum from sandstone. Ericaceous shrubs dominate the ground layers on such sites. The Allegheny Mountains have a wet and cool climate, with approximately 98 cm of precipitation per year with about 20% falling as snow, an annual average temperature of 7 8C to 11 8C, an average annual maximum temperature of 14 8C to 17 8C, an average annual minimum temperature of 2 8C to 3 8C, and a growing season of 126 to 155 days in the study area (Cleland et al. 1997, Thomas-Van Gundy 2014).

Braun (1950) related the highly varied forest cover of this subsection to the complex erosional surfaces driven by soil and parent material types and ages. Overall, she mapped the larger area as mixed-mesophytic forest but specifically identified localized communities dominated strongly by mixed oak on upper slopes and ridgetops of western Maryland. Survey records indicate that the area was not settled until 1789 (Schlosnagle 1978), and therefore, we designated pre-1780 as the presettlement period. The area appears free of historical logging activity and was acquired as a Maryland State Forest in 1929.

Results.

The Coleman Hollow old-growth forest at Savage Mountain, Maryland, is presently dominated by Acer rubrum L. (red maple) with an importance value of 24.3%, followed by Quercus montana Willd. (chestnut oak), Quercus alba L. (white oak), Quercus rubra L. (red oak), and Quercus velutina Lam. (black oak) (Table 1). South Savage Mountain is dominated by chestnut oak (20.4% importance), followed by Betula lenta L. (black birch), red oak, red maple, black oak, and white oak. A comparison of forest surveys at Savage Mountain with present-day FIA data for the AMP reveals a higher importance of oak at Savage Mountain, 41.8% versus 25.5% in the AMP (Table 2). This is mainly due to the dominance of chestnut oak at Savage Mountain. Black birch and red maple have a higher combined importance at Savage Mountain versus AMP (50.1% and 31.3%, respectively). In contrast, AMP has a higher density of Prunus serotina (black cherry) and Pinus (pine).

Table 1.

Relative stand density (RDN), dominance (RBA), frequency (RFQ), and importance value (RIV) by species, stand totals, and number of plots encountered for Coleman Hollow and South Savage forests at Savage Mountain in western Maryland. Other species included ironwood Carpinus caroliniana Walter, Acer pensylvanicum L., and Amelanchier laevis Wiegand.

Table 2.

Relative density of witness trees and contemporary survey data for Savage Mountain, Maryland, and the Allegheny Mountain Plateau/M221Bf (AMP) in western Maryland and Pennsylvania. Witness tree data for the AMP are to genus only and do not contain data for Maryland.

Witness trees at Savage Mountain were dominated by white oak (26.5%), followed by Carya spp. (hickory), black oak, chestnut oak, and Castanea dentata (chestnut) (Table 2). Witness tree data in the AMP of Pennsylvania were dominated by oak (41.2%), followed by pine, maple, and chestnut (Table 2). The main differences in witness trees between Savage Mountain and the AMP were the greater presence of pine and maple in the AMP in Pennsylvania.

A comparison of witness trees with contemporary surveys indicated the loss of chestnut, oak, and hickory in both study areas. Oak averaged 55% and 41% in the witness tree record of Savage Mountain and the AMP, respectively, versus 42% and 21% present-day. Hickory declined from 17% to 2% at Savage Mountain. In contrast, maple increased from 0.3% to 32% and black birch from 0% to 18% at Savage Mountain, while Nyssa sylvatica Marshall (blackgum) and black cherry increased from 0% to 2% and 10%, respectively, in the AMP.

Discussion.

Loss of fire-tolerant oaks and increase in red maple and other mesophytes, from pre-European settlement to the present day, in this study corresponds with similar trends throughout most of the eastern United States (Nowacki and Abrams 2008, Thompson et al. 2013). Given current management and ecological conditions the loss of pyrogenic trees (e.g., oak, hickory, pine) and continued expansion of later successional tree species (e.g., red maple) is nearly ubiquitous (Abrams 1998, Fei and Steiner 2007). Periodic burning prior to and catastrophic land clearing and fires after European settlement were arguably key factors promoting oak-hickory and limiting mesophytic species in eastern forest stands (Shumway et al. 2001, Aldrich et al. 2009, Howard et al. 2021). This changed after fire suppression in the early 20th century, which facilitated the increase in mesophytic and later-successional, shade-tolerant trees (Abrams 2010, Hanberry et al. 2020). The clear-cut and catastrophic-fire era also resulted in the deconiferization of northeastern forests; thus the loss of pine in our AMP study area (Whitney 1990, 1994; Nowacki and Abrams 2015; Abrams and Nowacki 2016).

Another mesophytic species, black birch, increased from pre-European settlement to the present day. Black birch has also increased across its range, perhaps corresponding to the demise of chestnut from the blight and hemlock (Tsuga canadensis (L.) Carrière) due to the hemlock woolly adelgid (Adelges tsugae Annand; Shumway et al. 2001, Orwig and Foster 1998). Black birch is also highly competitive on north-facing upper slopes and ridges (Nowacki and Abrams 1992) and is benefitting from oak decline on a wide range of sites. Repeated gypsy moth defoliation and eventual death of overstory oaks has also promoted black birch (Kegg 1971, Hix et al. 1980). However, the low shade tolerance and competitive ability of black birch may limit future increases in eastern forests (Leak 1958). The AMP saw a slight increase in shade-tolerant blackgum, from limited presence to 2% of all trees, which will likely continue to slowly increase in eastern forests (Abrams 2007). A large increase in blackgum in the 1930s coincided with fire suppression at Catoctin Mountain in north-central Maryland (Howard et al. 2021). Blackgum is unique in the combination of being both shade-tolerant and resistant to fire from its thick bark and sprouting ability, including root suckers. Thus, it can maintain itself in pyrogenic and nonpyrogenic forests, despite its inherently slow growth and fecundity (Abrams 2007). The large post-logging increase in the disturbance-oriented black cherry in the AMP is consistent with previous studies, although this trend may be reversed due to its diminished establishment, growth, and survival in recent decades (Nowacki and Abrams 1992, Royo et al. 2021).

Our previous research at Savage Mountain reported a very frequent pre-European and early European settlement fire return interval of approximately 8 yr (Shumway et al. 2001). Pine trees on Catoctin Mountain, Maryland, had a mean fire interval of 5.47 yr between 1702 and 1951 (Howard et al. 2021). Prior to European settlement, burning of the forest understory was a common and long-term practice of Native Americans, including at Savage Mountain, as evidenced by archaeological and anthropological records and artifacts, accounts of early settlers, and paleoecological charcoal and fire scar studies (Day 1953, Abrams 1992, Whitney 1994, Foster et al. 2002, Abrams et al. 2022). In northeastern oak forests, Native American burning facilitated travel, improved visibility in the forest, increased browse for game, increased berry and mast production, and cleared and maintained land for agricultural fields, as well as improved hunting, plus other uses (Patterson and Sassaman 1988, Williams 2004, Abrams and Nowacki 2008). The high frequency of surface burning perpetuated oak and hickory, as well as pine, all of which are particularly well adapted to reproduce and grow in high light conditions. Characteristics of oaks and hickories that allow them to be so well adapted to fire and other disturbances are thick bark, resprouting ability, high root:shoot ratios of growth, and the ability to compartmentalize wounds (Lorimer 1985, Abrams 1996). These characteristics apply to most eastern pine species, although sprouting is uncommon.

The contemporary survey data from the old-growth forests on Savage Mountain are more similar to the witness tree record in regard to oak species, but less similar in regard to red maple and black birch, when compared with the regional AMP data. The AMP also has lower oak dominance in present-day forests than that on Savage Mountain, which may be explained by the extensive cutting of forests throughout the AMP. Large increases in red maple and black birch on Savage Mountain may be explained by the chestnut blight and suppression of fire after 1940 (Shumway et al. 2001). However, old-growth and second-growth stands on xeric sites in the Ridge and Valley Province of central Pennsylvania and western Maryland have resisted mesophytic species invasion (Nowacki and Abrams 1992, Heeter et al. 2019).

Following the extirpation of Native Americans in the eastern United States, European settlement brought on intensive land clearing and exploitation that started in the late 1700s, peaked between 1870 and 1930, and moved from east to west (Ruffner and Abrams 1998, 2002; Stambaugh et al. 2018). Prior to 1850, extensive logging on the Allegheny Plateau was typically limited to small sawmills in most towns where white pine was highly favored and areas of charcoal iron production (Whitney 1990, 1994; MacCleery 1992). The increasing demands for timber led to the large-scale commercialization of the forest industry by the middle of the 19th century, including steam-powered saws and railroad logging. Logging escalated, leading to the height of the clear-cut era from 1850 to 1920. The original forests were extensively cut on nearly all but the most difficult sites. In addition, a large loss of forest area to land clearing (agriculture) occurred during this period across most eastern states (Whitney 1994). Areas left in old-growth forests totaled less than 0.5%. The “Great Cutover” produced extensive stump-lands covered in logging debris (slash), which resulted in catastrophic wildfires, which, in extent and intensity were not experienced in the original forest (Pyne 1984). Extensive cutting of forests, the charcoal iron industry, mining, and catastrophic wildfires in the early history of European settlement apparently maintained or elevated fire levels and pyrogenic vegetation in most locations of the eastern United States (Abrams 2010). Exceptions to this are the loss of chestnut in the early 1900s from chestnut blight and the loss of conifer species (deconiferization) due to their inability to compete with hardwoods that can sprout after logging and fire (pitch pine, Pinus rigida Mill., being the exception). Nonetheless, catastrophic wildfires ushered in the fire suppression (Smokey Bear) era in the United States starting in the 1930s (Nowacki and Abrams 2008, Abrams 2010). Fires in the eastern United States became very infrequent and highly localized after the 1930s.

One of the most dramatic changes that has occurred in forests of eastern North America during the 20th century is the increase in red maple, as evidenced in this and other studies (Abrams 1998, Nowacki and Abrams 2015). Red maple has become nearly ubiquitous across most sites, which is in stark contrast to the rather limited distribution of red maple reported in pre-European settlement forests (Abrams and Ruffner 1995). The increase in red maple is thought to be related, at least in part, to the exclusion of fire in most forests, particularly those dominated by oak and pine, and its being avoided by loggers during the major clear-cut era from 1870 to 1930. The increase in red maple and other shade-tolerant, nonpyrogenic species is resulting in the “mesophication” of eastern forests (Nowacki and Abrams 2008). This results in cooling, dampening, and reduced flammability of the ground layer in many eastern oak forests, increasing the difficulty of using prescribed fire. Low- to moderate-intensity understory fire is needed for the long-term sustainability of most eastern oak and pine forests because it is effective in eliminating fire-sensitive, shade-tolerant, mesophytic trees species, while creating the proper light and forest floor conditions for oak seedlings (and other pyrophiles) to regenerate and recruit into the canopy (Brose 2014, Hanberry et al. 2020).

The results of this study help confirm that AMP forests have changed from pyrogenic to more mesophytic conditions following the fire suppression policies of the Smokey Bear era and, perhaps, accelerated succession from extensive forest cutting and the chestnut blight (Keever 1953, Abrams and Downs 1990, Abrams and Nowacki 1992, Fei et al. 2005). This has led to a decline of several historically dominant trees in the AMP and throughout much of the eastern United States, as well as the increase in other tree species that limit the sustainability of eastern oak-hickory-pine forests. This study also demonstrates important differences in the long-term dynamics of old-growth versus second-growth forests in the region, with old-growth forests maintaining a greater oak component, as well as the value of comparing witness tree and contemporary survey data.