Juglans Cinerea L

Home , Juglans, Juglans ailantifolia, Juglans cinerea, Liriodendron tulipifera, Tilia americana

R E S E A R C H A R T I C L E ABSTRACT: To mitigate the loss of native tree species threatened by non-native pathogens, managers need to better understand the conservation status of remaining populations and the conditions that favor successful regeneration. Populations of Juglans cinerea L. (butternut), a wide-ranging riparian species, • have been devastated by butternut canker, a disease caused by a non-native fungal pathogen. We assessed J. cinerea within Great Smoky Mountains National Park (GSMNP) to determine post-disease survivorship and health, recruitment history, environmental conditions associated with survival, and the Conservation Status extent of hybridization with a non-native congener. Monitoring records were used to locate and collect data for 207 J. cinerea trees in 19 watersheds. Tree cores were collected from a subset of individuals to of a Threatened Tree assess recruitment history. We sampled vegetation plots within areas that contained J. cinerea to assess site conditions and overstory species composition of characteristic habitat. We collected leaf samples for Species: Establishing genetic analysis to determine the frequency of hybridization. Our reassessment of monitoring records suggests that J. cinerea abundance in GSMNP has declined due to butternut canker and thirty years of a Baseline for poor regeneration. Populations displayed continuous recruitment following Park establishment (1934) until around 1980, after which regeneration declined drastically. Ordination analysis revealed that J. Restoration of Juglans cinerea in the contemporary forest was associated with greater distance from homesites and reduced basal area of competing species. Hybrids comprised a small portion of sampled trees. The presence of cinerea L. in the healthy trees and low rate of hybridization suggest that these trees may contribute to the development of disease-resistant genotypes for future restoration efforts.

Southern Appalachian Index terms: butternut, cohort structure, disturbance regime, forest disease, fungal pathogen, hybridiza- Mountains, USA tion, mortality, recruitment history

Amanda M. Parks1 INTRODUCTION in tree species (Lovett et al. 2006; Loo 1Department of Forestry and Natural Re- 2009; Holzmueller et al. 2010). While sources Exotic diseases are an increasing threat research has focused on the effects of Purdue University to native biodiversity as global movement diseases on common species, less is known 715 West State St. of materials introduces non-native patho- about the effects of introduced pathogens West Lafayette, IN, 47907, USA gens, and changing land use facilitates on less-common species, for which the their spread (Lovett et al. 2006). Fungal consequences of disease may be ampli- pathogens have devastated North Ameri- fied because of inherently low population Michael A. Jenkins1,4 can forests in the last century, resulting density and limited reproduction of the Keith E. Woeste2 in fundamental alterations to ecosystems host species. 3 (Ellison et al. 2005; Lovett et al. 2006; Loo Michael E. Ostry 2009; Holzmueller et al. 2010). Baseline The four-decade program to produce a 2Northern Research Station (pre-disease) data are necessary to exam- blight-resistant hybrid of C. dentata is USDA Forest Service ine the overall effects of species loss, but nearing the stage of active restoration Hardwood Tree Improvement and often extensive mortality occurred before (Jacobs 2007). As the program progressed, Regeneration Center detailed information was collected, making other researchers have focused on the Purdue University it difficult to determine the true effects of restoration of other tree species affected 715 West State St. disease on ecosystem processes (Ellison by exotic disease (Windham et al. 1998; West Lafayette, IN, 47907, USA et al. 2005). Griffin 2000; Ostry and Moore 2008). For example, Juglans cinerea L. (butternut), a 3Northern Research Station species whose range extends across much USDA Forest Service Several eastern tree species have been virtu- 1561 Lindig Ave. ally eliminated by introduced pathogens; of eastern North America, has experienced St. Paul, MN 55108, USA others have experienced such severe decline extensive mortality from butternut canker that they no longer retain their pre-disease caused by the non-native fungal pathogen ecological function (Loo 2009). Chestnut Ophiognomonia clavigignenti-juglan- • blight, caused by Cryphonectria parasitica dacearum (Nair, Kostichka, & Kuntz) Murrill (Barr) and perhaps the best example Broders & Boland (Woeste et al. 2009). 4 Corresponding author: of a fungal disease that led to the loss of a While estimates of mortality caused by but- [email protected] keystone species, functionally eliminated ternut canker are as high as 92% (Carlson Castanea dentata [American chestnut, and Guthmiller 1993), evidence suggests (Marsh.) Borkh] from eastern forests in that genetic resistance exists and may the early twentieth century. Other fungal potentially be exploited for conservation diseases, including beech bark disease, and restoration efforts (Thompson et al. Natural Areas Journal 33:413–426 Dutch elm disease, and dogwood anthrac- 2006; Ross-Davis et al. 2008). Efforts are nose, have caused widespread declines underway to identify resistant individuals

Volume 33 (4), 2013 Natural Areas Journal 413 (Ostry and Moore 2008) and establish a dominance by other early-successional Wisconsin (Carlson and Guthmiller 1993), framework for the development of resistant species; and (4) What was the temporal within GSMNP, numerous scattered indi- genotypes (Michler et al. 2006). recruitment pattern of surviving J. cinerea viduals and groups of trees remain (Figure populations and when did contemporary 1) within a continuous (non-fragmented) The persistence of J. cinerea in contempo- competitors establish? forest that may limit exposure to non-na- rary forests is precarious not only because tive genes and where high ridges further of the effects of butternut canker, but also isolate watersheds. METHODS due to habitat loss, poor regeneration, and genetic dilution from hybridization with Juglans cinerea has a range of historic and non-native Juglans ailantifolia Carr. (Japa- Study Area modern uses: Native Americans and early nese walnut; Hoban et al. 2009). Juglans American settlers had several uses for the ailantifolia was historically planted within Great Smoky Mountains National Park nuts, bark, and sap, including dyes, foods, agricultural landscapes (Hoban et al. 2009) is an internationally renowned center of and medicines (Ostry and Pijut 2000; and has been present within the range of biological diversity in North America, lead- Schultz 2003). Historically in areas where J. cinerea for multiple generations. As ing to its designation as an International J.cinerea was common, its wood was nearly with other uncommon species that co-oc- Biosphere Reserve in 1976 and World equal in value to that of Juglans nigra L. cur with a non-native congener, such as Heritage Site in 1983 (Jenkins 2007). Be- (black walnut) (Nicholls 1979). Although Celastrus scandens (American bittersweet) cause of its large size (over 210,000 ha), J. cinerea is largely a riparian species and C. orbiculatus (oriental bittersweet), biological diversity, and protected status, that rarely dominates stands, the loss of genetic dilution as a result of hybridization GSMNP serves a vital role in conservation J. cinerea could have ecological impacts presents a serious threat to the persistence within the southeastern United States and because it is a highly cold-tolerant hard of the species (Pooler et al. 2002). When observed changes within the biota of GS- mast species whose nuts are preferred by hybridization occurs in conjunction with MNP serve as baselines for comparison to many species of wildlife (Ostry and Pijut an invasive pathogen, the threat to a native other state and federal lands (Jenkins 2007). 2000). In addition, J. cinerea litter has high species is amplified. GSMNP was once extensively settled and concentrations of nitrogen and calcium 75% of the Park was logged prior to its compared to most other riparian species Successful restoration of J. cinerea not establishment in 1934 (Pyle 1985), but for- (Ricklefs and Matthew 1982). only depends upon mitigating the effects est communities in GSMNP have received of disease and developing resistant geno- minimal direct human disturbance over the Survivorship, Health, and types, but also requires understanding the last 75 years. Distribution regeneration dynamics of J. cinerea and the community context in which it his- Focal Species To assess survivorship and health, we torically occurred and currently persists used a National Park Service (NPS) (Thompson et al. 2006). In this study, we Juglans cinerea typically grows in high- monitoring database to locate J. cinerea examined the contemporary health, genetic light environments along streams in the trees throughout GSMNP, systematically integrity, canopy associates, and recruit- moist, fertile soils of riparian forests (Rink sampling watersheds to represent the full ment history of J. cinerea populations 1990), although the species has been ob- distribution of J. cinerea. We used site (defined in this study as all individuals directions and geographical coordinates to occurring within a watershed; Figure 1) served on drier, rocky sites (Cogliastro et al. 1997). The loss of riparian habitats caused search for individual trees, and those not within Great Smoky Mountains National found at specified locations were assumed Park (GSMNP), a protected area with nu- by agriculture, development, alteration to stream courses, and installation of dams to have died since the initial NPS surveys. merous well-documented populations of We searched for 268 trees previously identi- J. cinerea. Based upon this examination, has greatly diminished habitat suitable for J.cinerea (Clark et al. 2008). When fied in the NPS records, of which we found we address four primary questions: (1) 178 that were still living. We sampled a marginal farmland was abandoned in the What is the survivorship and health of J. total of 207 mature trees in 19 watersheds. early-mid twentieth century, J. cinerea trees cinerea across an unfragmented landscape The 29 additional mature trees we sampled likely established in the resulting early- represented by GSMNP?; (2) What is the were newly identified individuals that were extent of hybridization with J. ailantifolia? successional habitat. However, as forest discovered in the course of our survey. For We hypothesized that the isolation and lack canopies developed, subsequent cohorts of each tree, UTM coordinates, diameter at of fragmentation within GSMNP buffers this highly shade-intolerant species were breast height (dbh; 1.4 m), and canopy posi- J. cinerea populations from invasion by unable to successfully regenerate on sites tion (dominant, codominant, intermediate, non-native genes; (3) What is the overstory with less frequent and intense disturbance suppressed) were recorded. Although not composition in contemporary forests that (Ostry et al. 1994). Although J. cinerea has all locations in the database were visited, contain surviving J. cinerea trees? We experienced extensive mortality ranging we sampled as many as was logistically hypothesized that J. cinerea occurs with from 77% in the southeastern United States possible for a thorough representation of greater abundance in forests with reduced (USDA Forest Service 1995) to 92% in J. cinerea habitat.

414 Natural Areas Journal Volume 33 (4), 2013 Figure 1. (a) Locations of 207 Juglans cinerea trees surveyed across (b) 19 watersheds in Great Smoky Mountains National Park.

We used several metrics to characterize below dbh, and percent of basal area girdled To assess regeneration, we initiated sur- disease severity of individual trees: percent by cankers. In addition to ease of evalua- veys near adult trees and systematically live crown (the estimated proportion of the tion, the number of cankers below dbh is expanded our search within a circle of a consistent measure of disease because approximately 100-m radius around each total tree length to the living crown length), infection typically involves spores from the adult. For each seedling or sapling we vigor (percent of crown with dieback; see upper branches washing down the trunk and found, we recorded UTMs and collected Figure 3 for class deﬁnitions), presence causing cankers at the tree base (Tisserat a leaf sample for genetic analysis to dis- of epicormic sprouts, number of cankers and Kuntz 1983). tinguish J. nigra from J. cinerea. Because

Volume 33 (4), 2013 Natural Areas Journal 415 seedlings of these two species are very (Thermo Fisher Scientific, Wilmington, areas and roadsides). Nevertheless, plots similar, morphological identification is MA) and all samples were diluted to a were distributed throughout GSMNP and difficult. common working concentration of 10ng/μL included a range of elevations. We recorded prior to PCR reactions. the dbh by species of all stems ≥ 10 cm We compiled spatial summary data using dbh within a 500-m2 plot. These data were ArcMap (ESRI 2009) and GIS layers ac- We assessed hybridization status using used to calculate relative basal area (species quired from the NPS database (http://www. one chloroplast marker and three nuclear basal area/total basal area of all species), nps.gov/gis/data_info/) and determined el- markers (Zhao and Woeste 2010) to dif- relative density (species density/total evation, vegetation association (as defined ferentiate J. cinerea genotypes from J. density), and importance value (relative by White et al. 2003), disturbance history, ailantifolia and hybrids of the two species density + relative basal area/2) for each and historic forest type for each tree loca- (Table 1). DNA from each individual was species on each plot. tion. We derived contemporary vegetation examined by performing polymerase chain associations from the GSMNP vegetation reaction (PCR) as described by Zhao and map, which was developed using aerial Woeste (2010). The resulting product size Vegetation Data Analysis photography interpretation, on-the-ground (cut or uncut, based on the marker) was We used species importance values calcu- verification, and the national classification assessed using agarose gel electrophoresis. system developed by NatureServe (Madden For each PCR reaction, control DNA was lated from our 20 plots to perform non- et al. 2004). Disturbance history classes amplified from individuals of known spe- metric multidimensional scaling (NMS) were derived from Pyle (1985), which cies origin. In addition, all seedling and based on Sorensen’s distance to assess the classified human disturbance based upon sapling samples and several mature trees distribution of tree species and plots along ownership records and historic maps into that were tagged as morphologically cryptic environmental gradients (McCune and the following categories: corporate logging, were assessed using the ITS marker, which Grace 2002). We used random starting con- diffuse disturbance, diffuse disturbance paired with chloroplast marker trnT-F, figurations and 50 runs with real data with with large residual trees, concentrated identifies J. nigra: individuals with a J. three to five runs performed on each data settlement, and undisturbed. cinerea genotype for the trnT-F marker, set. Field and GIS-derived environmental but a J. ailantifolia genotype at the ITS variables (elevation, distance to stream, marker were considered black walnut and distance to former homesite), overstory Leaf Sampling and Genetic Testing (Zhao and Woeste 2010). Sixteen individu- basal area, and stand age were correlated als identified as J. nigra were removed with each ordination axis and plotted as We collected leaf samples from 174 J. from subsequent analyses, but retained as vectors within ordination space. Distances cinerea trees (including 18 seedlings or controls for genetic tests. between plot locations and nearest features saplings) for genetic analysis. Between in the streams and homesites layers were one and 30 trees were sampled from each calculated using Near analysis in ArcMap watershed using an arborist’s slingshot. Plot Sampling Samples were refrigerated at the end of (ESRI 2009). Stand age was determined J. cinerea each day and shipped weekly to the genet- We installed a total of 20 vegetation plots from a representative tree nearest ics lab at Purdue University where DNA across 13 watersheds in areas containing J. the plot. Two-way cluster analysis (using was extracted using a chloroform-phenol cinerea trees to assess site conditions and Ward’s linkage method and Euclidean dis- extraction method (Zhao and Woeste overstory composition. Some watersheds tance measure) was performed to examine 2010). DNA concentration was measured were not sampled due to lack of appropriate species patterns related to clustering of using a NanoDrop-8000 spectrophotometer sites for plot installation (e.g., developed plots and confirm NMS results.

Table 1. Markers used in hybrid identification. CAPS: cleaved amplified polymorphic sequence; SCAR: sequence characterized amplified region. See Zhao and Woeste (2010) for additional details.

! !! !!% & !!% & !!% & !!% & ( -"*'!,#) ., '!-., ., '!-"*'!-.,'!,#) ., '!-., ++(. ( "). ) ").'! ) ((( * (+ ( ,)) ,)) ((( *-(* ( ++#'!* "'!*+"'!-) ++$'!* " ++$'!++#'* "'!*+"'!-) ((( ( ,).'!+-)'!*$. -,)'!,). -,)'!,).'!+-)'!*$. ,).'!+-)'!*$.

416 Natural Areas Journal Volume 33 (4), 2013 Tree Core Sampling and Analysis centric rings equal to the average growth distributed across all watersheds. of the innermost rings (Applequist 1958). To assess the age of the dominant J. cinerea Because the limited number of surviving Juglans cinerea occurred across a fairly cohort, we collected tree cores from a sub- trees made cross-dating difficult, trees were narrow range of site conditions. Most trees set of the 207 trees we located. Trees were placed into five-year recruitment classes to occurred between 400 and 700 m eleva- selected to represent the spatial distribution account for possible errors in determining tion and within 12 of the 37 associations of living J. cinerea trees across sampled exact year of establishment. described in the Park vegetation map, with watersheds. All trees were cored at 0.5 m RESULTS most trees occurring in floodplain (33%) above the ground using a 4.3 mm incre- and successional hardwood (28%) forest ment borer. The borer was sterilized with associations. Fewer trees occurred in cove alcohol between sampling individual trees Survivorship, Health, and forests and areas of ‘human influence,’ (Clark et al. 2008) to prevent spread of the Distribution which largely consisted of old fields and pathogen. At the request of the NPS, trees margins around front-country development. showing no sign of disease (< 1% of all Based on our resurvey of J. cinerea trees According to the disturbance classes of trees we encountered) were not cored. in the NPS database, we estimate survivor- Pyle (1985), most J. cinerea trees (74%) ship to be 68% over the past two decades. occurred in locations historically domi- To assess recruitment and cohort structure, Although J. cinerea trees ranged in size nated by settlement. we intensively sampled within five water- from 7 – 74 cm dbh, there was a paucity sheds, coring all J. cinerea trees within the of individuals in the small diameter classes forest interior (excluding trees along roads (Figure 2). Most trees (49%) were classified Hybridization and trails), from which an intact sample as codominant, with 9%, 25%, and 13% could be collected. We sampled 8 – 18 in- classified as dominant, intermediate, and Based on analysis of hybrid identification dividuals per watershed (Table 2). For each suppressed, respectively. Only 14 J. cinerea markers, five out of the 174 samples had of these individuals, we also cored the two seedlings and saplings were found. hybrid genotypes at all four markers. These nearest overstory competitors to determine hybrids were found in two locations, Cades the composition of the developing forest Individual J. cinerea trees exhibited a Cove in the Abrams Creek watershed and at the time of J. cinerea regeneration and range of disease severity and often trees the Hazel Creek watershed (Figure 1). Prior examine the temporal pattern of recruit- of similar size within the same area were to the establishment of GSMNP, Cades ment amongst competing species. in different stages of decline. Most trees Cove was a large agricultural community had a vigor class rating of 2, indicating and the Hazel Creek area once contained Cores were mounted on boards, sanded 1% – 25% crown dieback (Figure 3). Most Proctor, a town with large logging opera- with a belt sander, and then hand-sanded crowns were small in relation to total tree tions. Only one of five trees sampled in with increasingly finer grit sand paper height, presumably due to disease effects Cades Cove was a hybrid, but all four (Clark et al. 2008). The annual rings and loss of canopy due to shading. Cankers mature trees sampled in Hazel Creek were were counted using a dissecting scope. ranged in size from several centimeters to hybrids. We cored only one hybrid (from For cores that did not include the pith, over 1 meter in length. Epicormic sprouts Hazel Creek), which established in 1971. we estimated the number of missing years occurred on 58% of trees. Healthy and No seedlings or saplings were identified using a graphic with equally-spaced con- severely diseased individuals were evenly as hybrids.

Table 2. Number of J. cinerea trees surveyed, number of leaf samples, and core samples collected per watershed for the ﬁve intensively sampled watersheds. In addition to J. cinerea trees, cores were collected from the two nearest canopy competitors in these watersheds. The numbers of samples from the 14 non-intensive watersheds are grouped (Other watersheds; competitors not sampled). Total trees and leaf sample numbers include samples from 14 seedling/saplings that were not used in health assessments of mature trees.

" !## !## # ! !### ! ## # ,% ,) *+ # #' ( ,* +$ *% ## # +- *% *, # *, & *) # &%% # *)-%++)

Volume 33 (4), 2013 Natural Areas Journal 417 Figure 2. Diameter distribution in 5 cm DBH classes for 207 Juglans cinerea trees surveyed across Great Smoky Mountains National Park.

Vegetation Composition plots, it was absent or had low basal area on surge of regeneration that continued from all four of the differentiated plots. Another the time of Park establishment in the NMS ordination resulted in a two-dimen- common species, Carpinus caroliniana mid-1930s to around 1980, after which sional solution with a final stress value of Walter (musclewood), was absent from all regeneration declined dramatically: only 13.86. A Monte Carlo test revealed that but one of these plots. Both of these species one sampled tree established after 1980, were also separated by the first division of the stresses of both Axis 1 (p = 0.02) and and only 14 immature trees were located the dendrogram (Figure 5). Three out of Axis 2 (p = 0.05) were less that than that in our survey. Generally, individual water- produced by randomized data. The cumula- four of these plots contained large basal sheds exhibited trends similar to the pooled tive variance explained by the ordination area of J. cinerea, and two exhibited the was 88% with Axis 1 explaining the most greatest J. cinerea importance values of any watersheds, with recruitment spanning the variance (R2 = 0.618) followed by Axis 2 plot. Bubble plots of species importance same time frame but with some variability (R2 = 0.261). Distance to homesite and onto the species ordination plot confirmed in initial J. cinerea establishment or peak total overstory basal area exhibited the these results: J. cinerea importance value year (Figure 8). The establishment of strongest relationships with Axis 1 (R = was inversely related to the importance competitors peaked around the same time -0.478 and 0.527 respectively), while eleva- of L. tulipifera, Liquidambar styraciflua as J. cinerea, but competitor recruitment tion (R = -0.609) exhibited the strongest L. (sweetgum), and C. caroliniana and was generally more continuous through relationship with Axis 2 (Figure 4, Table was associated with greater distance from time (Figure 8). 3). Four plot locations clustered away homesites and lower stand basal area (Fig- from all other plots and were associated ures 4 and 6). with increased distance from homesites DISCUSSION and lower overstory basal area (Figure 4). Two-way cluster analysis clearly separated Recruitment these four plots at the first division of the Survivorship, Health, and dendrogram, supporting the results of the The oldest J. cinerea tree we sampled Distribution NMS ordination (Figure 5). Although on established in 1925 and the youngest es- average Liriodendron tulipifera L. (tulip tablished in 2001 (Figure 7). Data pooled Our survivorship estimate of 68% since the poplar) was the dominant species across our across all watersheds revealed an initial late 1980s was considerably greater than

418 Natural Areas Journal Volume 33 (4), 2013 Figure 3. Health data summary for 207 Juglans cinerea trees surveyed across 19 watersheds in Great Smoky Mountains National Park. Vigor class represents the percent of the crown with dieback or discoloration. Percent live crown represents the proportion of the total tree length to the living crown length. the 23% survivorship observed in North NPS surveys (Ostry et al. 2003), and many had < 20% girdling, suggesting that they Carolina and Virginia between 1966 and individuals likely died prior to the initia- have been able to limit canker growth and 1986 (USDA Forest Service 1995). This tion of monitoring. Consequently, the Park avoid girdling. Since we know the pathogen earlier (USDA Forest Service) estimate database may not capture the full extent of produces a proliﬁc number of spores that may reﬂect mortality since the onset of dis- mortality and the individuals we surveyed can be spread by precipitation, wind, and ease; but the disease had presumably been likely represent a group that survived the living vectors (Kuntz et al. 1979; Tisserat active in the southeastern United States for initial wave of disease. While cankers were and Kuntz 1983), we assume all individu- at least two to three decades before the present on most trees we sampled, most als clustered closely in a watershed had

Volume 33 (4), 2013 Natural Areas Journal 419 about the same time interval of exposure to the pathogen. If true, the variation in health among living trees in our study may reﬂect varying levels of disease resistance. Individuals that remained healthy in close proximity to dead or declining trees may be the best sources for resistance genes since it is unlikely that they escaped infection and may have qualities that enabled them to better tolerate the canker disease.

Our results indicate that J. cinerea occurs with greatest frequency under a fairly limited range of environmental conditions (400 – 700 m elevation, low stand basal area, floodplain forests). Other studies have described J. cinerea as occurring in low elevation alluvial floodplains, but also in upland areas and cove forests with suf- ficient light and well-drained soils (Ostry et al. 2003). When assessing the distribution of J. cinerea, land-use history must be considered because human-mediated dispersal of this species and its congeners may be reflected in the current distribution and hybridization status of populations. In addition, anthropogenic disturbance likely created the site conditions under which J. cinerea historically regenerated and persisted. Although J. cinerea seeds are dispersed naturally by rodents, streams, Figure 4. Non-metric multidimensional scaling ordination of overstory vegetation data from 20 plots or gravity (Schultz 2003), human activity distributed across 13 watersheds in Great Smoky Mountains National Park. Dominant environmental variables are represented as vectors [distance to homesite, elevation, total overstory basal area (TOBA)]. also likely contributed to the movement Hollow squares represent the four plots that clustered away from other plots, and are characterized of J. cinerea in GSMNP because trees by high Juglans cinerea basal area and low Liriodendron tulipifera (LIRITUL) basal area relative to historically were planted at homesites and all other plots. farms (Ross-Davis et al. 2008). While we cannot determine whether the origin of J. cinerea in a given watershed was through

Table 3. Environmental variables (Mean ± 1 standard error) and associated correlations with nonmetric multidimensional scaling (NMS) ordination axis scores for 20 overstory vegetation sampling plots sampled across 13 watersheds.

% & # " " ! ϲϮϯ͘Ϯцϯϱ͘ϰ ϯϰϳ ϵϬϳ Ϭ͘ϭϯϭ Ϭ͘ϱϴϭ ͲϬ͘ϲϬϵ Ϭ͘ϬϬϰ ϰϭ͘Ϯцϲ͘Ϯ ϱ͘ϲ ϵϯ͘ϴ Ϭ͘Ϯϲϵ Ϭ͘ϮϱϮ Ϭ͘Ϭϵϵ Ϭ͘ϲϳϴ ! Ϯϴϵ͘ϭцϭϬϴ͘ϴ ϭϰ͘ϯ ϭϴϴϴ͘ϯ ͲϬ͘ϰϳϴ Ϭ͘Ϭϯϯ ͲϬ͘ϭϰϱ Ϭ͘ϱϰϭ ! ϯϲ͘ϱцϯ͘ϯ ϭϬ͘ϭ ϳϴ͘ϱ Ϭ͘ϱϮϳ Ϭ͘Ϭϭϳ Ϭ͘ϬϮϱ Ϭ͘ϵϭϳ &$! ! ϱϰ͘ϯцϯ͘ϴ Ϯϯ ϴϱ ͲϬ͘ϬϰϮ Ϭ͘ϴϲ ͲϬ͘ϯϭϯ Ϭ͘ϭϳϵ

420 Natural Areas Journal Volume 33 (4), 2013 Figure 5. Two-way dendrogram of plots and overstory species in 20 vegetation plots distributed across 13 watersheds in Great Smoky Mountains National Park. The four plots described in Figure 4 are highlighted in the plot/sample column. Liriodendron tulipifera (LIRITUL), Carpinus caroliniana (CARPCAR), and Juglans cinerea (JUGLCIN) are highlighted in the species row. Other species: Acer rubrum (ACERRUB), Acer saccharum (ACERSAC), Aesculus flava (AESCFLA), Asimina triloba (ASIMTRI), Betula alleghaniensis (BETUALL), Betula lenta (BETULEN), Carya cordiformis (CARYCOR), Cornus florida (CORNFLO), Fraxinus americana (FRAXAME), Fraxinus pennsylvanica (FRAXPEN), Halesia tetraptera (HALETET), Juglans nigra (JUGLNIG), Liquidambar styraciflua (LIQUSTY), Magnolia acuminata (MAGNACU), Nyssa sylvatica (NYSSSYL), Oxydendrum arboreum (OXYDARB), Pinus echinata (PINUECH), Pinus strobus (PINUSTR), Pinus virginiana (PINUVIR), Platanus occidentalis (PLATOCC), Prunus serotina (PRUNSER), Quercus alba (QUERALB), Quercus imbricaria (QUERIMB), Quercus rubra (QUERRUB), Rhododendron maximum (RHODMAX), Tilia americana (TILIAME), Tsuga canadensis (TSUGCAN), Ulmus americana (ULMUAME), Ulmus rubra (ULMURUB). human planting or natural reproduction, have been documented in more fragmented Existing hybrids may also be the progeny most of the contemporary trees we sampled landscapes (Hoban et al. 2009). Within of J. ailantifolia trees that were planted established after original landowners were GSMNP, we theorized that increased within GSMNP. removed from GSMNP. isolation from developed areas provided a buffer from hybridization due to increased Vegetation Composition barriers to invasion by non-native pollen. Hybridization Also, because J. ailantifolia was planted, NMS analysis revealed an association be- Although widespread settlement in the areas with higher historic human popula- tween greater abundance of J. cinerea and southern Appalachians may have promoted tions should have higher proportions of lower importance of other early-succes- the introduction of J. ailantifolia, we hy- hybrids. Anecdotally, this appears to be sional species. Liriodendron tulipifera was pothesized that the isolation, topographic the case. The few hybrids we observed the dominant species across our study sites complexity, and lack of fragmentation in were in historically settled areas near and sustained disturbance associated with GSMNP favored low occurrence of hybrids the present-day Park boundary. However, land use at homesites (soil disturbance, compared to more-fragmented landscapes. other sampled watersheds were formerly lack of litter accumulation) may have Hybrids comprised < 3% of our sampled settled and near the Park boundary, but created seedbed conditions that favored trees, but hybridization rates over 90% exhibited no evidence of hybridization. germination and establishment of this and

Volume 33 (4), 2013 Natural Areas Journal 421 other small-seeded and widely-dispersed silvics, we predicted that contemporary were collected following clearcutting. species (Hosner and Graney 1970; Jenkins J. cinerea trees regenerated following Secondary succession following harvest and Parker 2001). Reduced dominance the abandonment of agricultural land and is typically the result of stump sprouting of L. tulipifera and other species farther that subsequent recruitment in maturing or seeds from the intact seedbank (Oliver away from homesites may have allowed forests would be rare. The overall pattern 1981), but advanced regeneration and J. cinerea to survive because of reduced of recruitment we observed across all seeds of tree species are typically reduced competition. Plantation studies have shown watersheds exhibited an initial peak in the or eliminated following decades of agri- that controlling competing vegetation is late 1940s, approximately 10 – 15 years cultural use (Wijdeven and Kuzee 2000). critical for the successful survival and after Park establishment. A similar lag Since J. cinerea seeds are not stored in the growth of J. cinerea seedlings (Lambert time has been observed for the establish- seed bank, dispersal from nearby sources et al. 1994). ment of other early successional species would be necessary for re-establishment. in abandoned agricultural fields (Peroni Although J. cinerea seedlings may perform 1994; Jenkins and Parker 2000). Clark et well in the open conditions of abandoned Recruitment al. (2008) found most J. cinerea regenera- fields, its large seeds may make re-estab- tion occurred within nine years; but, unlike lishment difficult due to limited dispersal Based on existing knowledge of butternut the abandoned fields in GSMNP, their data (Battaglia et al. 2008).

Figure 6. Bubble plots of overstory species importance value overlaid onto the NMS overstory ordination. Symbol size is proportional to the importance value (relative density + relative basal area/2) of a species on a given plot.

422 Natural Areas Journal Volume 33 (4), 2013 protected from genetic contamination. Although butternut canker was present in all areas, we observed many vigorous trees that displayed symptoms of low disease severity. Natural selection may have increased the frequency of these relatively healthy-appearing trees; butternut canker is thought to have been present in the study area for over 50 years and many highly susceptible trees may have already died, leaving trees that have relatively greater resistance (Ingwell and Preisser 2010). Genetically pure and diverse populations, such as those found in GSMNP, may prove critical to future efforts to breed resistant genotypes of this species by serving as sources of propagation material for restoration efforts centered on intraspeciﬁc (non-hybrid) disease resistance.

Figure 7. Juglans cinerea recruitment within five-year intervals between 1925 and 2005 pooled across all Breeding efforts designed to enhance dis- sampled watersheds in Great Smoky Mountains National Park. Tree cores were collected in 2010-11. ease resistance in J. cinerea may require the introduction of genes from J. ailantifolia in a manner similar to the breeding programs We identified continuous recruitment of et al. 2006). Although J. cinerea often using non-native genes for blight resistance competing species concurrent with recruit- successfully establishes in natural forest in C. dentata (Griffin 2000; Jacobs 2007). ment peaks of J. cinerea, indicating that openings, such as along stream banks, However, additional studies are necessary J. cinerea was able to compete, to some habitat destruction in riparian areas and to determine the ecological implications degree, with other trees in developing reduced population densities may neces- of this strategy. Although it remains to be stands. As the forest matured, competition sitate active management and the creation seen if J. cinerea restoration will require intensified and light became less available of forest openings to supplement natural the incorporation of non-native genes for subsequent regeneration. The precipi- reproduction (Ostry et al. 1994). In addi- from J. ailantifolia, we recommend that tous drop in regeneration after 1980 likely tion to managing disease effects, different any introduction of non-native genes into resulted from continued forest development management prescriptions may be neces- the gene pool of J. cinerea or other native and the effects of disease. Regeneration that sary for different life stages of J. cinerea species should be minimal. Hybrids used did occur likely had low survival due to to promote regeneration, establish new for restoration need to be carefully evalu- shading and disease (Kuntz et al. 1979). cohorts, and facilitate continued recruit- ated for altered traits such as morphology, ment (Woeste et al. 2009). environmental tolerance, and competitive The lack of regeneration over the past ability that will affect their ability to fill several decades has serious implications for the ecological niche of J. cinerea. Conclusions and Management future J. cinerea conservation, regardless Implications of disease impacts. While J. cinerea can Most J. cinerea trees in the watersheds establish in silvicultural openings, over- we sampled were recruited following Other studies have reported frequent hy- story competition in post-harvest stands agricultural land abandonment. Contem- bridization of J. cinerea with its non-native may accelerate decline (Clark et al. 2008). porary populations persist in areas where congener J. ailantifolia. The frequency In our study, the oldest cohort of J. cinerea of hybridization in our study was quite prior disturbance created an environment trees established on abandoned agricultural low, however. Plant populations in un- of reduced interspecific competition. Over land where substrate disturbance and a fragmented landscapes, such as GSMNP, the past 30 years, J. cinerea recruitment lack of vegetation legacies resulted in a may be buffered from gene swamping by has declined severely due, in part, to a reduced rate of successional development invasive species. As such, these landscapes less-intensive contemporary disturbance and, consequently, reduced interspecific may serve as refugia for genetically-un- regime. Following the development of competition. Site alterations that reduce contaminated populations and may play resistant genotypes, active restoration of competition may be needed to promote a critical role in maintaining the diversity this species will require control of com- the establishment and persistence of the of the species. Future restoration efforts peting vegetation to allow planted trees species, even if disease-resistant lineages should focus on unfragmented landscapes to successfully reach the forest canopy are developed in the future (Thompson where newly-established populations are and substrate disturbance combined with

Volume 33 (4), 2013 Natural Areas Journal 423 Figure 8. Recruitment of Juglans cinerea and competitors within five-year intervals in three watersheds with greater than 12 tree core samples: MPLP (Middle Prong Little Pigeon), MPLR (Middle Prong Little River) and Oconaluftee River. canopy thinning or removal to allow suc- imply official endorsement or approval research interests include breeding of fine cessful germination and establishment of by the USDA or the Forest Service of any hardwoods and the conservation genetics seedlings. product to the exclusion of others that may of butternut and yellowwood. be suitable. Michael E. Ostry is a research plant pa- ACKNOWLEDGMENTS thologist with the USDA Forest Service. Amanda M. Parks recently completed her His research is focused on developing We thank the staff of GSMNP for their sup- master’s degree in forest biology at Purdue management guidelines to minimize dis- port, especially Tom Remaley, Janet Rock, University. Her areas of specialization in- ease impacts on trees in native forests and Keith Langdon, Kristine Johnson, Glenn clude forest ecology, conservation genetics, plantations. Taylor, Emily Darling, Emily Guss, and and forest pathogens and pests. Paul Super. We thank Peng Zhao, Michael LITERATURE CITED Saunders, and Robert Morrissey for their Michael A. Jenkins is an Associate Profes- assistance. Kim Steiner, Eric Holzmueller, sor of forest ecology at Purdue University. Applequist, M.B. 1958. A simple pith locator and three anonymous reviewers provided His research interests include plant com- for use with off-center increment cores. Journal of Forestry 56:141. helpful comments on earlier versions of munity responses to disturbance, impacts of the manuscript. This work was supported exotic plants, insects, and disease, and the Battaglia, L.L., P.R. Minchin and D.W. Pritch- by the Hardwood Tree Improvement and ett. 2008. Evaluating dispersal limitation use of fire in community restoration. in passive bottomland forest restoration. Regeneration Center (USDA FS NRS) and Restoration Ecology 16:417-424. the Department of Forestry and Natural Keith E. Woeste is a USDA Forest Ser- Carlson, J.C., and M. Guthmiller. 1993. Inci- Resources at Purdue University. The use vice Research Geneticist stationed at the dence and severity of butternut canker in of trade names is for the information and Hardwood Tree Improvement and Regen- Wisconsin in 1976 and 1992. Phytopathol- convenience of the reader and does not eration Center at Purdue University. His ogy 83:1352 (Abstract).

424 Natural Areas Journal Volume 33 (4), 2013 Clark, S.L., S.L. Brosi, S.E. Schlarbaum, and Jenkins, M.A., and G.R. Parker. 2001. Woody Ostry, M.E., B. Ellingson, D. Seekins, and W. H.D. Grissino-Mayer. 2008. Dendrochro- species composition of disturbed forests in Ruckheim. 2003. The need for silvicultural nology of two butternut (Juglans cinerea intermittent stream bottomlands of southern practices and collection of butternut germ- L.) populations in the southeastern United Indiana. Journal of the Torrey Botanical plasm for species conservation. Pp. 515-55 States. Forest Ecology and Management Society 128:165-175. in J.W. Van Sambeek, J.O. Dawson, F. 255:1772-1780. Kuntz, J.E., A.J. Prey, S. Jutte, and V.M.G. Nair. Ponder, Jr., E.F. Lowenstein, and J.S. Fral- Cogliastro A., D. Gagnon, and A. Bouchard. 1979. The etiology, distribution, epidemi- ish, eds., Proceedings of the 13th Central 1997. Experimental determination of soil ology, histology, and impact of butternut Hardwood Forest Conference. General characteristics optimal for the growth canker in Wisconsin. Pp. 69-72 in K.J. Technical Report NC-234, U.S. Department of ten hardwoods planted on abandoned Kessler, Jr., and B.C. Weber, eds., Walnut of Agriculture, Forest Service, North Central farmland. Forest Ecology and Management Insects and Diseases. General Technical Research Station, St. Paul, Minn. 96:49-63. Report NC-52, Workshop Proceedings, U.S. Ostry, M.E. and M. Moore. 2008. Response Ellison, A.M., M.S. Bank, B.D. Clinton, E.A. Department of Agriculture, Forest Service, of butternut selections to inoculation with Colburn, K. Elliott, C.R. Ford, D.R. Foster, North Central Forest Experiment Station, Sirococcus clavigignenti-juglandacearum. B.D. Kloeppel, J.D. Knoepp, G.M. Lovett, J. St. Paul, Minn. Plant Disease 92:1336-1338. Mohan, D.A. Orwig, N.L. Rodenhouse, W.V. Sobczak, K.A. Stinson, J.K. Stone, C.M. Lambert, F., B. Truax, D. Gagnon, and N. Ostry, M.E. and P.M. Pijut. 2000. Butternut: Swan, J. Thompson, B. Von Holle, and J.R. Chevrier. 1994. Growth and nutrition, an underused resource in North America. Webster. 2005. Loss of foundation species: monitored by enzyme assays, in a hardwood HortTechnology 10:302-306. consequences for the structure and dynamics plantation: effects of mulching materials and Peroni, P.A. 1994. Invasion of red maple (Acer of forested ecosystems. Frontiers in Ecology glyphosate application. Forest Ecology and rubrum L.) during old field succession in the and the Environment 3:479-486. Management 70:231-244. North Carolina Piedmont: age structure of ESRI. 2009. ArcMap 9.3.1. ESRI, Redlands, Loo, J. 2009. Ecological impacts of non-indig- red maple in young pine stands. Bulletin of Calif. enous invasive fungi as forest pathogens. the Torrey Botanical Club 121:357-359. Griffin, G.J. 2000. Blight control and restora- Biological Invasions 11:81-96. Pooler, M.R., R.L. Dix, and J. Feely. 2002. tion of the American chestnut. Journal of Lovett, G.M., C.D. Canham, M.A. Arthur, Interspecific hybridizations between the na- Forestry 98:22-27. K.C. Weathers, and R.D. Fitzhugh. 2006. tive bittersweet, Celastrus scandens, and the introduced invasive species, C. orbiculatus. Hoban, S.M., T.S. McCleary, S.E Schlarbaum, Forest ecosystem responses to exotic pests Southeastern Naturalist 1:69-76. and J. Romero-Severson. 2009. Geographi- and pathogens in Eastern North America. cally extensive hybridization between the Bioscience 56:395-405. Pyle, C. 1985. Vegetation disturbance history of Great Smoky Mountains National Park: forest trees American butternut and Japanese Madden M., R. Welch, T. Jordan, P. Jackson, an analysis of archival maps and records. walnut. Biology Letters 5:324-327. R. Seavey, and J. Seavey. 2004. Digital Research/Resource Management Report Holzmueller, E.J., S. Jose, and M.A. Jenkins. vegetation maps for the Great Smoky Mountains National Park: final report. Center SER-77, U.S. Department of the Interior, 2010. Ecological consequences of an exotic National Park Service, Southeast Region, fungal disease in eastern U.S. hardwood for Remote Sensing and Mapping Science, University of Georgia, Athens. Atlanta, Ga. forests. Forest Ecology and Management 259:1347-1353. McCune, B., and J.B. Grace. 2002. Analysis Ricklefs, R.E., and K.K. Matthew. 1982. Chemi- of ecological communities. MjM Software cal characteristics of the foliage of some Hosner, J.F., and D.L. Graney. 1970. The relative Design, Gleneden Beach, Ore. deciduous trees in southeastern Ontario. Ca- growth of three forest tree species on soils Michler, C.H., P.M. Pijut, D.F. Jacobs, R. nadian Journal of Botany 60:2037-2045. associated with different successional stages Meilan, K.E. Woeste, and M.E. Ostry. Rink, G. 1990. Juglans cinerea L., butternut. Pp. in Virginia. American Midland Naturalist 2006. Improving disease resistance of but- 386-390 in R.M. Burns, and B.H. Honkala, 84:418-427. ternut (Juglans cinerea), a threatened fine tech coords., Silvics of North America, Vol. Ingwell, L.L., and E.L. Preisser. 2010. Using hardwood: a case for single-tree selection 2. Hardwoods, Agricultural Handbook 654, citizen science programs to identify host through genetic improvement and deploy- U.S. Department of Agriculture, Forest resistance in pest-invaded forests. Conserva- ment. Tree Physiology 26:121-128. Service, Washington, D.C. tion Biology 25:182-188. Nicholls, T.H. 1979. Butternut canker. Pp. Ross-Davis, A., M. Ostry, and K. Woeste. 2008. Jacobs, D.F. 2007. Toward development of 73-82 in Proceedings of the Symposium Genetic diversity of butternut (Juglans silvical strategies for forest restoration of on Walnut Insects and Diseases. General cinerea) and implications for conserva- American chestnut (Castanea dentata) Technical Report NC-52, U.S. Department tion. Canadian Journal of Forest Research using blight-resistant hybrids. Biological of Agriculture, Forest Service, North Central 38:899-907. Forest Experiment Station, St. Paul, Minn. Conservation137:497-506. Schultz, J. 2003. Conservation assessment for Jenkins, M.A. 2007. Vegetation communities Oliver, C.D. 1981. Forest development in North butternut or white walnut (Juglans cinerea) of Great Smoky Mountains National Park. America following major disturbances. For- L. U.S. Department of Agriculture,Forest Southeastern Naturalist 6 (Special Issue est Ecology and Management 3:153-168. Service, Eastern Region, Milwaukee, 1):35-56. Ostry, M.E., M.E. Mielke, and D.D. Skilling. Wisc. Jenkins, M.A., and G.R. Parker. 2000. The 1994. Butternut – strategies for managing Thompson, L.M., F.T. van Manen, S.E. Schlar- response of herbaceous-layer vegetation to a threatened tree. General Technical Report baum, and M. DePoy. 2006. A spatial model- anthropogenic disturbance in intermittent NC-165, U.S. Department of Agriculture, ing approach to identify potential butternut stream bottomland forests of southern Indi- Forest Service, North Central Forest Experi- restoration sites in Mammoth Cave National ana, USA. Plant Ecology 151:223-237. ment Station, St. Paul, Minn. Park. Restoration Ecology 14:289-296.

Volume 33 (4), 2013 Natural Areas Journal 425 Tisserat, N., and J.E. Kuntz. 1983. Dispersal sification of Great Smoky Mountains Na- flowering dogwood resistant to dogwood gradients of conidia of the butternut canker tional Park. Report submitted to BRD-NPS anthracnose. HortScience 33:1265-1267. fungus in a forest during rain. Canadian Vegetation Mapping Program. NatureServe, Woeste, K., L. Farlee, M. Ostry, J. McKenna and Journal of Forest Research 13:1139-1144. Durham, N.C. S. Weeks. 2009. A forest manager’s guide USDA Forest Service. 1995. Forest insect and Wijdeven, S.M.J., and M.E. Kuzee. 2000. Seed to butternut. Northern Journal of Applied disease conditions in the United States 1994. availability as a limiting factor in forest re- Forestry 26:9-14. Forest Pest Management, United States covery processes in Costa Rica. Restoration Zhao, P., and K.E. Woeste. 2010. DNA markers Department of Agriculture, Forest Service, Ecology 8:414-424. identify hybrids between butternut (Juglans Washington, DC. Windham M.T., E.T. Graham, W.T. Witte, J.L. cinerea L.) and Japanese walnut (Juglans White, R.D., K.D. Patterson, A. Weakley, C.J. Knighten, and R.N. Trigiano. 1998. Cor- ailantifolia Carr.). Tree Genetics and Ge- Ulrey, and J. Drake. 2003. Vegetation clas- nus florida ‘Appalachian Spring’: a white nomes 7:511-533.

426 Natural Areas Journal Volume 33 (4), 2013 Conservation Status of a Threatened Tree Species: Establishing a Baseline for Restoration of Juglans cinerea L. in the Southern Appalachian Mountains, USA Author(s): Amanda M. Parks , Michael A. Jenkins , Keith E. Woeste and Michael E. Ostry Source: Natural Areas Journal, 33(4):413-426. 2013. Published By: Natural Areas Association DOI: http://dx.doi.org/10.3375/043.033.0404 URL: http://www.bioone.org/doi/full/10.3375/043.033.0404

BioOne (www.bioone.org) is a nonprofit, online aggregation of core research in the biological, ecological, and environmental sciences. BioOne provides a sustainable online platform for over 170 journals and books published by nonprofit societies, associations, museums, institutions, and presses. Your use of this PDF, the BioOne Web site, and all posted and associated content indicates your acceptance of BioOne’s Terms of Use, available at www.bioone.org/page/terms_of_use. Usage of BioOne content is strictly limited to personal, educational, and non-commercial use. Commercial inquiries or rights and permissions requests should be directed to the individual publisher as copyright holder.

BioOne sees sustainable scholarly publishing as an inherently collaborative enterprise connecting authors, nonprofit publishers, academic institutions, research libraries, and research funders in the common goal of maximizing access to critical research.