nena masthead
NENA Home Staff & Editors For Readers For Authors

The Effects of a Tornado Disturbance and a Salvaged Timber Extraction on the Seed-rain and Recruitment Community of an Eastern Temperate Deciduous Forest
Alexander C. Curtze, Tomás A. Carlo, and John W. Wenzel

Northeastern Naturalist, Volume 25, Issue 4 (2018): 627–645

Full-text pdf (Accessible only to subscribers. To subscribe click here.)

 

Access Journal Content

Open access browsing of table of contents and abstract pages. Full text pdfs available for download for subscribers.



Current Issue: Vol. 30 (3)
NENA 30(3)

Check out NENA's latest Monograph:

Monograph 22
NENA monograph 22

All Regular Issues

Monographs

Special Issues

 

submit

 

subscribe

 

JSTOR logoClarivate logoWeb of science logoBioOne logo EbscoHOST logoProQuest logo

Northeastern Naturalist Vol. 25, No. 4 A.C. Curtze, T.A. Carlo, and J.W. Wenzel 2018 627 2018 NORTHEASTERN NATURALIST 25(4):627–645 The Effects of a Tornado Disturbance and a Salvaged Timber Extraction on the Seed-rain and Recruitment Community of an Eastern Temperate Deciduous Forest Alexander C. Curtze1, Tomás A. Carlo1,*, and John W. Wenzel2 Abstract - Understanding forest-regeneration pathways following salvage-logging operations and windthrow disturbances is a step towards the sustainable management of forests. To accomplish this, assessment of the relative contribution of seed dispersal and post-dispersal abiotic and biotic factors in shaping post-disturbance seedling recruitment is necessary. In a temperate deciduous forest located in western Pennsylvania, we measured the seed rain over 2 years and the resulting woody-plant recruitment in 2 tornado-disturbed stands. Our data shows that the salvage logging operation and the dispersal mode of plants (wind, animal) affected the composition of the seed rain. Communities of plant recruits, in contrast, were more strongly determined by soil-cover variables than by the net inputs of seed rain. Our results indicate that salvage logging following natural windthrow events has minimal negative impacts on forest regeneration capacity. Introduction Salvage logging is a forestry practice in which the standing and fallen dead trees in a forest are removed following natural disturbances such as tornadoes, wildfires, floods, and catastrophic insect outbreaks (Lindenmayer and Noss 2006). This practice is common throughout the northeastern US and in other countries as a means to ameliorate economic losses in forestry associated with unexpected disturbances (Lindenmayer and Noss 2006, Nappi et al. 2004, Smith-Ramírez et al. 2014). Strong wind is the principal type of disturbance in the temperate deciduous forests of the northeastern US (Fischer et al. 2013, Stueve et al. 2011), yet few studies have evaluated the effects of strong-windthrow events, such as tornadoes, and the subsequent salvage-logging disturbances on the regeneration ecology of northeastern temperate deciduous forests (e.g., Royo et al. 2016). Especially lacking are studies that account for the potential limitations imposed by seed-dispersal processes and modes of dispersal in the regeneration of northeastern deciduous forests following salvage-logging operations. Tornado disturbances are unique compared to other natural and anthropogenic disturbances in that they increase the heterogeneity of forest stands by snapping and uprooting many trees, while leaving some trees standing, which survive and maintain local fruit and seed supplies (Chazdon 2014; Lindenmayer and Noss 2006; Waldron et al. 2013, 2014). Tornadoes increase the density of coarse and fine 1 Biology Department, The Pennsylvania State University, University Park, PA 16802. 2Powdermill Nature Reserve, Carnegie Museum of Natural History, 1847 Route 381, Rector, PA 15677. *Corresponding author - tac17@psu.edu. Manuscript Editor: Elizabeth Hane Northeastern Naturalist 628 A.C. Curtze, T.A. Carlo, and J.W. Wenzel 2018 Vol. 25, No. 4 woody debris and form a pit-and-mound microtopography from the uprooted trees, which creates a high diversity of microsites (Lindenmayer and Noss 2006; Waldron et al. 2013, 2014). Thus, increased heterogeneity of microtopography, soil, light penetration, moisture, and woody debris result in a higher availability of optimal safe sites for plant recruitment (Puerta-Piñero et al. 2010, Terborgh et al. 2011, Thiffault et al. 2011). The extra woody debris also has the benefit of accelerating recolonization by animals in disturbed areas by increasing the availability of cavities, breeding sites, and food resources (Lindenmayer and Noss 2006, Puerta-Piñero et al. 2010, Wermelinger et al. 2017). Increased animal activity, in turn, boosts plant–animal interactions such as herbivory, pollination, seed predation, and seed dispersal, resulting in effects on plant communities that remain little understood, but which are likely to be important. Salvage logging, in contrast, creates high-light conditions following the removal of all standing and fallen dead trees, which promotes colonization by light-demanding species (Burns and Honkala 1990, Coomes and Grubb 2000). However, the loss of coarse woody debris, competition by allelopathic weeds, and the loss, degradation, and compaction of soils during and after salvage logging could offset any benefits of increased light (Coomes and Grubb 2000, de Jesus Jatoba et al. 2016, DeLuca et al. 2012, George and Bazzaz 1999, Kruegler and Peterson 2009). Due to concomitant reductions in both coarse woody debris and structural heterogeneity, some authors have suggested that salvage-logging sites are less attractive to animals and may hamper animal-mediated seed dispersal due to the lack of perches, shelter, and food resources (Carlo and Morales 2016, Puerta-Piñero et al. 2010, Wunderle 1997). Seed dispersal (i.e., seed transport away from the parent tree) is a key ecological process affecting forest-regeneration dynamics, gene flow, species’ range expansions, and species’ responses to climate change (Cain et al. 2003, Davis and Shaw 2001). Seed dispersal plays a key role in the spatial distribution of plants within a forest community based on the success or failure of seeds to arrive in particular sites (Hurtt and Pacala 1995, Levine and Murrell 2003, Muller-Landau 2008). Despite the importance of seed dispersal in forest-regeneration dynamics, it has remained a relatively understudied aspect of forest regeneration (Chazdon 2014). A handful of studies have evaluated the role of seed dispersal in Mediterranean Quercus ilex L. (Holly Oak) and Pinus spp. (pines) forests following wildfires (Castro et al. 2012, Rost et al. 2009), but there are no studies that assess the role of seed dispersal in the regeneration of temperate forests following tornado and salvage-logging disturbances. Thus, there is a lack of empirical studies analyzing how the success or failure of seed arrival at potential recruitment sites (i.e., dissemination limitation; Muller-Landau 2008) affects temperate deciduous-forest regeneration following tornado disturbances and salvage-logging operations. The goal of this study was to investigate the relative contribution of the seed rain and the seed-dispersal mode (i.e., by wind or animal vectors) to the abundance of recruits from common tree and shrub species following tornado damage and salvage- logging disturbances in a temperate deciduous forest in western Pennsylvania. Northeastern Naturalist Vol. 25, No. 4 A.C. Curtze, T.A. Carlo, and J.W. Wenzel 2018 629 Our first objective was to determine how the seed rain and woody-plant recruitment community (i.e., germinant and established seedlings) varies by type of disturbance (i.e., salvaged versus unsalvaged) for the selected plant species. For this, we hypothesized that the seed-rain intensity would be higher in the unsalvaged compared to salvaged areas, due to the influence of increased distance to sources in the salvaged areas (dispersal limitation). We also hypothesized that the density of recruited seedlings would be higher in the salvaged areas relative to the unsalavged areas due to the increased availability of light. Our second objective was to examine the relationship between the seed rain and the recruitment of plant species across different microhabitats. For this, we tested the null hypothesis that the microhabitat factors are not correlated with the incoming seed rain and seedling communities. Field-site Description We conducted our study at Powdermill Nature Reserve located in Rector, PA (40°09'36.3"N, 79°16'19.6"W), ~ 400 m above sea level. The vegetation is a mixed deciduous forest dominated by Liriodendron tulipifera (Tulip-poplar), Acer rubrum (Red Maple), Betula lenta (Black Birch), Acer saccharum (Sugar Maple), Prunus serotina (Black Cherry), Fagus grandifolia (American Beech), Quercus spp. (oaks), and Carya spp. (hickories). The topography is mountainous with very stony silt loam or loam soils on moderate slopes (0–30% slopes) (NRCS 2017, PADCNR 2006–2008). Precipitation during the study years, 2014 and 2015, was ~800 mm and ~1147 mm, respectively (NOAA 2017). Straight-line winds, associated with an EF 0 tornado on 1 June 2012, blew down trees in several hardwood stands that resulted in more than 90% canopy loss on several large patches (>40 ha in total). Following the disturbance, we mapped 2 adjacent areas with a submeter Trimble Geoexplorer 6000 XT Garmin GPS. We divided each tornado-damaged area (hereafter Site A and Site B) in half using ArcGIS. We randomly assigned one half of each site to be salvage-harvested in typical clearcut fashion, while the other was left unsalvaged (Fig. 1a). The salvage-logging operations in both sites were conducted in the Fall of 2013. Each site has an elevation gradient (Site A = ~100 m, Site B = ~50 m) with average slopes of 9.25 ± 3.85% in Site A and 8.99 ± 4.57% in Site B that both vary from ~0–30% (Fig. 1b). Aspect for Site A is predominately northeast with a slight tail at the base facing north to northwest and Site B is predominately northwest. There are streams at the bottom of both sites, whereas the higher elevation areas were drier and had no surface water. Each study site was bordered by undisturbed mature forest. Methods Measuring the seed rain For our study, we measured the seed rain at Sites A and B by establishing 7 transects, each containing 8 seed traps, for a total of 56 seed traps per site. Seed-trap transects, spaced ~30 m from each other, were perpendicular to the long axis of each salvaged area and traversed all 4 habitats (Fig. 1a). Prevailing winds blew due east generally accross the long axis of each salvaged area (Fig. 1a) (NOAA 2017). Northeastern Naturalist 630 A.C. Curtze, T.A. Carlo, and J.W. Wenzel 2018 Vol. 25, No. 4 In each transect, we placed 2 seed traps in each of 4 habitat categories traversed by each transect: salvaged, clearcut edge, forest, and unsalvaged. We placed the seed traps in the clearcut edge right along the boundary between the forest and salvaged habitats and the boundary between the forest and unsalvaged habitats (Fig. 1a). The traps were circular laundry baskets 44 cm in diameter (0.15 m2) with a suspended net made from screen cloth and 1 cm x 1 cm chicken-wire mesh to protect the net from seed predators removing captured seeds (see Carlo and Tewksbury 2014). The traps are effective at intercepting wind-dispersed seeds, bird-dispersed seeds, and seeds dispersed by small mammals (e.g., squirrels), but typically do not capture seeds dispersed by large mammals (e.g., bears) or nuts hoarded and scattered by rodents. We collected the seed rain once each in April, July/August, and November of 2014 and 2015. For each seed trap, we noted the number of dispersed seeds per species, including the number of seeds of each species destroyed by predators. In the case of animal-dispersed seeds, we obtained separate counts for seeds that Figure 1. (a) Layout map of the 2 study sites with prevailing winds and seed-trap transects. (b) Slope-surface map of the 2 study sites (data source: P ADCNR 2006–2008). Northeastern Naturalist Vol. 25, No. 4 A.C. Curtze, T.A. Carlo, and J.W. Wenzel 2018 631 fell from fruit above (i.e., not dispersed by animals), and those truly dispersed by animals. For analyses we pooled the annual seed-rain per trap, then calculated the mean across 2014 and 2015. Woody-plant recruitment Between 15–17 May 2015, we placed 0.25-m2 square plots around each seed trap (excluding clearcut edge) in the 4 cardinal directions 1 m from the seed trap. Within each plot, we noted the abundance of all tree and woody species (plus Phytolacca americana L. [Pokeweed]) that were <75 cm tall and characterized them as either “germinant” (i.e., <1 y old) or “established seedlings” (i.e., >1 y old). Although we did not track individual seedlings, we believe this method provides a good snapshot of the early regenerating community in the site. Besides seedling recruitment, we estimated fern, rock, log, Smilax spp. (greenbriers), and herbaceous plant cover (excluding ferns) into percentage classes (0, 1–10, 11–25, 26–50, 51–75, and 76–100%). We visually estimated overstory-canopy cover into percentage-cover categories using the same cover-class scale as above. For analysis, we used the mid-point value of each cover class. Seed-producing plant community Around each seed trap in the forest and unsalvaged areas in sites A and B, we estimated (1) seed-producing overstory tree (trees >10 cm dbh) density and (2) seed-producing understory woody-plant (trees <10 but >2 cm dbh, shrubs >1 cm average stem diameter) species density using the point-quarter method (Table 1; Mitchell 2015). We based plant fecundity on experience and existing fruits on the plants, which we standardized for all plots. Statistical analyses To examine general patterns in the similarity of the seed rain and seedling composition among habitats, we conducted non-metric multidimentional-scaling ordinations with varimax rotation in PC-ORD 6 (McCune and Mefford 2011). We Table 1. The density (individuals ha-1) of woody plants in the forest and unsalvaged habitats. Includes only individual species that were analyzed (see Table 1). Density (ha-1) Species Forest Unsalvaged Acer rubrum (Red Maple) 78 60 Betula lenta (Black Birch) 31 34 Fagus grandifolia (American Beech) 26 6 Lindera benzoin (Spicebush) 604 1553 Liriodendron tulipifera (Tulip-poplar) 62 46 Nyssa sylvatica (Blackgum) 5 14 Phytolacca americana (Pokeweed) 0 259 Prunus serotina (Black Cherry) 13 10 Rubus spp. (blackberries) 0 2243 Sassafras albidum (Sassafras) 0 0 Smilax spp. (greenbriers) 1028 5434 Vitis spp. (grapes) 16 0 Northeastern Naturalist 632 A.C. Curtze, T.A. Carlo, and J.W. Wenzel 2018 Vol. 25, No. 4 conducted our ordination analyses at the level of individual seed traps to assess patterns in the seed rain and recruitment communities (next to each seed trap) in relation to habitat, site, and groundcover variables. We used a matrix of 28 plant-seed species x 96 seed traps for the seed rain, a matrix of 11 plant species x 68 plots for the germinant seedlings; and a matrix of 18 plant species x 68 plots for the established seedlings. Seed traps and seedling plots with no species data were excluded from analysis. Each ordination employed a random starting configuration for 100 runs with real data and 0 runs with randomized data. The instability criterion for accepting a solution was 0.000001 over the last 10 iterations. We then conducted univariate analyses to examine how responses of individual species (seed rain, recruitment) responded to habitats and selected habitat covariates (e.g., canopy and understory cover). For analysis, we used generalized linear models (GLM) performed in JMP 10 Pro (SAS Institute 2012). Each GLM used overdispersion tests and intervals and included site (A and B), habitat (salvaged, clearcut edge, forest, and unsalvaged) and an interaction term as model effects. Models were fitted with Poisson error distributions suitable for count data. We used the Bonferroni correction to set thresholds for significant P-values. We calculated Bonferroni correction values by dividing the specified alpha (a) of 0.05 by the number of models (n) run. We stepped the models to remove the least significant parameter(s) based on the chi-square value(s) and removed the specified parameter(s) if the Akaike’s information criterion (AIC) value decreased by 2 or more (Burnham and Anderson 2002). We determined significant differences in mean seed rain and seedling recruitment between habitats by using the GLM contrast function in JMP (SAS Institute 2012). The model for seedling emergence was similar except that it included only 3 levels for habitat (salvaged, forest, and unsalvaged). Covariates for seedling emergence included canopy cover, total groundcover, mean annual seed-predation (by species), and mean annual seed-rain (by species). We used a correlation analysis in JMP 10 to determine collinearity between cover parameters and excluded any predictor variables that had a correlation coefficient of |r| > 0.7 (Dormann et al. 2013). We included all cover variables in the GLMs because there was little collinearity (i.e., |r| > 0.7) between the variables (i.e., canopy, fern, log, rock, Smilax spp., and herb cover). The seedling-establishment GLM contained the same parameters listed above except that it excluded the mean annual seed predation and seed-rain parameters. Detailed parameter-estimates for the seed rain and recruitment GLMs can be viewed in Supplemental Tables S1 and S2 (see Supplemental File F1, available online at https://www.eaglehill.us/NENAonline/suppl-files/n25-4-N1562-Carlos1, and, for BioOne subscribers, at https://dx.doi.org/10.1656/N1562.s1). Table 2 outlines all species that we recorded in the 2 y of seed-rain and 1 y of recruitment data. This table also outlines the common seed-dispersal mechanisms (not all-inclusive) by each plant species. In subsequent discusion, we aggregate wind and mechanical (i.e., autochory) dispersal into the category of “wind-dispersed seeds” and bird- and rodent-dispersed seeds into the category of “animaldispersed seeds”. Northeastern Naturalist Vol. 25, No. 4 A.C. Curtze, T.A. Carlo, and J.W. Wenzel 2018 633 Results Nonmetric multidimensional ordinations The nonmetric multidimensional (NMS) ordination of the seed rain arranged the seed-trap data from both sites along 2 axes that accounted for 87% of the variation (Fig. 2a). Site A traps were located along the upper space of both axis 1 (accounting for 28% of the variation) and axis 2 (accounting for 59% of the variation), while Site B traps had lower scores on both axes. Thus, this ordination shows that the plant community represented in the seed rain varied by site and habitat type. The seed-rain samples in the salvaged areas formed a distinct cluster compared to samples from the forest and unsalvaged habitats in both sites. In contrast, the seed rain in the forest and unsalvaged habitats formed distinct clusters at Site B, but not Table 2. All recorded species in seed-rain and recruitment surveys with abbreviations used and common seed-dispersal mechanism. † denotes species included in Tables 3 and 4. However, all listed species were included in the NMS ordination. Scientific name Common name Seed dispersal mechanismA, B, D Acer pensylvanicum L. Striped Maple Wind †Acer rubrum L. Red Maple Wind Acer saccharum Marsh Sugar Maple Wind Amelanchier spp. Juneberries Mammal, bird †Betula lenta L. Black Birch Wind Carya spp. Hickories Mammal, bird Crataegus spp. Hawthorns Bird †Fagus grandifolia Ehrh. American Beech Mammal, bird Fraxinus spp. Ashes Wind Hammamelis virginiana L. Witch-hazel MechanicalC,D Ligustrum obtusifolium Sieb. and Zucc. Privet Bird †Lindera benzoin (L.) Blume. Spicebush Bird †Liriodendron tulipifera L. Tulip-poplar Wind Lonicera spp. Honeysuckles Bird Malus spp. Apples Mammal, bird †Nyssa sylvatica Marsh Blackgum Mammal, bird Parthenocissus quinquefolia (L.) Planch. Virginia Creeper Bird †Phytolacca americana L. Pokeweed Bird †Prunus serotina Ehrh. Black Cherry Bird, mammal Quercus spp. Oaks Mammal, bird Robinia pseudoacacia L. Black Locust Wind, mammal, bird Rosa multiflora Thunb. Multiflora Rose Bird †Rubus spp. blackberries Bird, mammal †Sassafras albidum (Nutt.) Nees. Sassafras Bird †Smilax spp. Greenbriers Bird Toxicodendron radicans (L.) Kuntze Poison Ivy Bird Viburnum spp. Viburnums Bird †Vitis spp. Wild grapes Bird ABurns and Honkala (1990). BCraves and Wloch (2012). CAnderson and Hill (2002). DRhoads and Block (2005). Northeastern Naturalist 634 A.C. Curtze, T.A. Carlo, and J.W. Wenzel 2018 Vol. 25, No. 4 Figure 2. (a) Nonmetric multidimensional (NMS) ordination of the mean annual seed rain (2 years of data) in 2 sites grouped by habitat. For NMS axis 1, R2 = 0.281; for NMS axis 2, R2 = 0.592. (b) Nonmetric multidimensional (NMS) ordination of the mean seedling emergence in 2 sites grouped by habitat. For NMS axis 1, R2 = 0.573; for NMS axis 2, R2 = 0.392. (c) Nonmetric multidimensional (NMS) ordination of the mean established seedlings in 2 sites grouped by habitat. For NMS axis 1, R2 = 0.522; for NMS axis 2, R2 = 0.148. Northeastern Naturalist Vol. 25, No. 4 A.C. Curtze, T.A. Carlo, and J.W. Wenzel 2018 635 Site A (Fig. 2a). The seed rain at the clearcut edge overlapped extensively with all of other habitats (Fig. 2a). The NMS ordination of germinant seedlings (i.e., <1 y old) arrayed the plots in both sites along 2 axes that accounted for 96.5% of the variation with 4.43% stress (Fig. 2b). Seedling plots in Site B had low scores on both axis 1 (57.3% of the variation) and axis 2 (39.2% of the variation) as compared to Site A. Some differences were that axis 1 was significantly correlated with the amount of herbaceous plant cover (r = 0.413, df = 67), while axis 2 was not significantly correlated with any of the measured covariates and may reflect unaccounted factors. Unlike the community of plants represented in the seed rain (Fig. 2a), the community of germinant seedlings was less differentiated across habitats and sites (Fig. 2b). The NMS ordination for the established seedlings (i.e., >1 y old) arrayed plots in both sites along 2 axes that accounted for 98.5% of the variation with 5.3% final stress. Axis 1 accounted for 83.7% of the variation and axis 2 accounted for 14.8% of the variation. Similar to the germinant-seedling ordination, plots from both sites were well mixed (Fig. 2c). Axis 1 was significantly positively correlated with percent canopy cover (r = 0.522, df = 67) and axis 2 was significantly positively correlated with Smilax spp. cover (r = 0.210, df = 67). Similar to the germinant-seedling community, the established-seedling communities in the forest and unsalvaged areas overlapped extensively, but unsalvaged formed a less distinct cluster than the forest plots. However, unlike the germinant-seedling community, the established-seedling community formed a distinct cluster in the salvaged habitat with lower scores along axis 1 while the unsalvaged and forest communities scored higher along axis 1 (Fig. 2). Seed-rain and recruitment patterns Wind-dispersed seeds of 6 tree species dominated the seed rain (min–max = 0–3355 seeds m-2 y-1; average = 720.3 ± 48.2 seeds m-2 y-1), accounting for 92.8 % of all seeds collected from seed traps. Black Birch, Red Maple, and Tulip-poplar accounted for 98.5% of the wind-dispersed seed rain and were widespread across habitats types (Fig. 3). Animal-dispersed seeds composed a small fraction of the total seed-rain (7.2%; range 0–1353 seeds m-2 y-1; average. = 56.12 ± 24.15 seeds m-2 y-1), was more species-rich, and their densities more variable than the winddispersed species (Fig. 3). When examining the seed rain across habitats, we found that the seed rain of 5 of the 11 dominant wind- and animal-dispersed species varied significantly by habitat type (Table 3, Fig. 3). For example, Tulip-poplar was the only species with a higher seed rain in the forest habitat, while Pokeweed was higher in the salvaged and clearcut-edge habitats compared to the forest areas. Seed-rain from Sassafras albidum (Sassafras) and from Rubus spp. (blackberries) were both greater in the clearcut-edge habitat, though not significanly so, while Black Cherry had more seed rain in the clearcut-edge and lowest in the salvaged habitat (Fig. 3. Table 3). Tulip-poplar and Pokeweed had more seeds arriving in 1 of the sites (A), but the effect of habitat for Tulip-poplar was site-dependent (Table 3; Supplemental Table S3, available online at https://www.eaglehill.us/NENAonline/suppl-files/ Northeastern Naturalist 636 A.C. Curtze, T.A. Carlo, and J.W. Wenzel 2018 Vol. 25, No. 4 n25-4-N1562-Carlo-s1, and, for BioOne subscribers, at https://dx.doi.org/10.1656/ N1562.s1). Both habitat and site variables significantly affected the richness of Figure 3. Graph showing the mean annual seed-rain (seeds m-2 y-1) for wind- and animaldispersed seeds (based on 2 y of data at 2 sites) and the mean seedling recruitment (seedlings m-2). We included the 9 most common species based on the mean seed rain and total seedling recruitment. Top row of graphs represent wind-dispersed species, other rows are animaldispersed species. Error bars represent 1 standard error. A value of 0.1 signifies no recorded seeds and/or seedlings in that habitat. Different letters above bars indicate significantly different values for that category for that species. Northeastern Naturalist Vol. 25, No. 4 A.C. Curtze, T.A. Carlo, and J.W. Wenzel 2018 637 the seed rain; we detected the highest richness in the clearcut edge and unsalvaged habitats, followed by the forest, and lowest in the salvaged habitat (Table 3). The germinant-seedling community was dominated by Red Maple, Tulip-poplar, and Vitis spp. (grapes), which accounted for 90.0% of all recorded germinant seedlings (min–max = 0–18 seedlings m-2 y-1; average = 2.22 ± 0.41 seedlings m-2 y-1). When examining the effects of habitat and habitat covariates on seedling emergence, we found that emergence varied in significance, magnitude, and direction depending on the observed plant species (Table 4; Supplemental Table S4, available online at https://www.eaglehill.us/NENAonline/suppl-files/n25-4-N1562-Carlo-s1, and, for BioOne subscribers, at https://dx.doi.org/10.1656/N1562.s1). The amount of cover of canopy, ferns, and rocks had a significant negative relationship to emergence in Black Birch, while the amount of Greenbrier cover was negatively related with the emergence of grapes. (Table 4). In addition, the emergence of Black Birch and grape seedlings differed by habitat type (Table 4). Emergence of Red Maple and grapes varied by site (Table 4). None of the measured covariates was found to be significantly related with the species richness of germinant seedlings (Table 4). The intensity of the seed rain and the rates of seed-predation were unrelated to the emergence of seedlings in all studied plant species (Table 4, Fig. 3). We also recorded seedlings in the salvaged and unsalvaged areas for some species that were undetected in seed traps such as greenbriers. (Fig. 3). The established-seedling community was dominated by Red Maple, Tulippoplar, and greenbriers, which accounted for 86.4% of all recorded established seedlings (min–max = 0–137 seedlings m-2 y-1; average = 19.0 ± 2.90 seedlings m-2 y-1). Densities of established seedlings of Tulip-poplar, Nyssa sylvatica (Blackgum), and Fagus grandifolia (American Beech) were affected by habitat, but in different ways. For example, Tulip-poplar had significantly more established seedlings in the salvaged areas (on average, 4.9 times higher compared to unsalvaged, and 92.9 Table 3. Summary of the effects of habitat on the seed rain of plant species at Powdermill Nature Reserve. Signs denote significant correlations (positive, negative) with the seed rain. For more details on the parameter-estimate values see Supplemental Table S1 (available online at https://www.eaglehill. us/NENAonline/suppl-files/n25-4-N1562-Carlo-s1, and, for BioOne subscribers, at https://dx.doi. org/10.1656/N1562.s1). Unsalvaged habitat taken as the intercept of GLM model and not shown. Clearcut Site * Site* Site* Species Site Salvaged edge Forest salvaged clearcut edge forest Acer rubrum 0 0 0 0 0 0 0 Betula lenta 0 0 0 0 0 0 0 Fagus grandifolia 0 0 0 0 0 0 0 Liriodendron tulipifera + 0 0 + 0 0 + Nyssa sylvatica 0 − + 0 0 0 0 Phytolacca americana 0 0 + − 0 0 0 Prunus serotina 0 − + 0 0 0 0 Rubus spp. 0 0 0 0 0 0 0 Sassafras albidum + 0 0 0 0 0 0 Smilax spp. 0 0 0 0 0 0 0 Vitis spp. 0 0 + 0 0 0 0 Species richness + − + 0 0 0 0 Northeastern Naturalist 638 A.C. Curtze, T.A. Carlo, and J.W. Wenzel 2018 Vol. 25, No. 4 Table 4. Summary of the effects of habitat on seedling recruitment of plant species at Powdermill Nature Reserve. Signs denote significant correlations (positive, negative) with seedling recruitment. § denotes excluded from model. For more details on the parameter-estimate values see Supplemenatal Table S2 (available online at https://www.eaglehill.us/NENAonline/suppl-files/n25-4-N1562-Carlo-s1, and, for BioOne subscribers, at https://dx.doi. org/10.1656/N1562.s1). Unsalvaged habitat taken as the intercept of GLM model and not shown. Site* Site* Fern Log Rock Smilax spp. Herb Seed Annual Species Site Salvaged Forest salvaged forest Canopy cover cover cover cover cover predation seed-rain Germinant seedlings Acer rubrum - 0 0 § § 0 0 § 0 0 0 0 0 Betula lenta § - + § § - - 0 - + 0 0 0 Liriodendron tulipifera § 0 0 § § 0 0 0 0 0 0 § 0 Prunus serotina § 0 0 § § + 0 + 0 0 0 0 § Vitis spp. + + − § § + 0 0 + - + § 0 Species richness 0 0 0 0 0 0 0 0 0 0 0 § § Established seedlings A. rubrum 0 0 0 0 0 - - 0 0 0 0 § § B. lenta 0 § 0 § § 0 0 0 0 0 0 § § Fagus grandifolia - 0 0 § § 0 - 0 - + 0 § § Lindera benzoin § 0 0 § § 0 0 0 0 0 § § § Liriodendron tulipifera § + - § § 0 - 0 § § 0 § § Nyssa sylvatica § + - § § + 0 0 0 0 0 § § P. serotina 0 0 0 0 0 0 0 0 0 0 0 § § Rubus spp. § 0 0 0 0 0 0 0 0 § 0 § § Sassafras albidum § 0 0 § § 0 - 0 0 0 - § § Smilax spp. - 0 0 0 0 0 0 0 0 § 0 § § Vitis spp. § § 0 § § § 0 0 0 0 0 § § Species richness - + - § § 0 - § 0 § § § § Northeastern Naturalist Vol. 25, No. 4 A.C. Curtze, T.A. Carlo, and J.W. Wenzel 2018 639 times higher compared to forest), American Beech had more established seedlings in the forest (2.9 times higher compared to unsalvaged), whereas Blackgum had low numbers of established seedlings in the forest (see Fig. 3, Table 4). When comparing the species richness of established seedlings across habitats, we found that it was significantly higher in the salvaged than in the forest and unsalvaged habitats (Table 4). Discussion Our study shows that tornado disturbances followed by salvage-logging operations have a strong effect on the quantity and diversity of the seed rain, affecting both the emergence and establishment of seedling communities. The tornado disturbance had an overall positive or neutral effect on the quantities of dispersed seeds, whereas salvage-logging operations limited dispersal for most species. The seed rain was clearly dominated by wind-dispersed species, and seeds of animaldispersed species were more frequent along edges of the salvaged than in the salvaged or the forest. Also, the opening of the canopy by the tornado disturbance and the salvage-logging operation created conditions that favored recruitment for a handful of species that positively respond to canopy gaps. The tornado may have increased the fecundity of some tree species in the unsalvaged areas by increasing light exposure, compensating for the lower tree densities in this habitat. Despite the differences in seed rain between the habitats, seedling emergence was minimally correlated with seed inputs and cover covariates, suggesting that other factors such as microsite conditions are affecting establishment (e.g., leaflitter depth, aspect; see Holl 1999, Kostel-Hughes et al. 2005) and differences in seed viability (Burns and Honkala 1990). As the seedlings aged, cover and habitat became more important filters for seedling establishment, probably because of light limitation (George and Bazzaz 1999, Kruegler and Peterson 2009), altered nutrient cycling (DeLuca et al. 2012), chemical allelopathy from certain fern species (de Jesus Jatoba et al. 2016, Horsley 1993), herbivory (Nuttle et al. 2013, Terborgh 2012), and/or pest/pathogen attacks (Terborgh 2012). Indeed, the heterogeneity within the disturbed habitats caused many site-specific conditions (e.g., fern cover) to partially outweigh the influence of disturbance type and canopy cover on seedling establishment for most species. Our results align with and expand upon the findings of previous studies in showing that post-dispersal ecological filters have a strong influence on the structure of regenerating plant communities in eastern deciduous forests (Caspersen and Saprunoff 2005, Kostel-Hughes et al. 2005, Terborgh 2012) and that salvage-logging operations generally only impact the regeneration of a few species (Peterson and Leach 2008, Royo et al. 2016). Seed-rain patterns Except for Tulip-poplar, we found no difference in the intensity of the seed rain of overstory-tree species between forest and unsalvaged habitats, suggesting that most species did not become source-limited despite suffering population reductions in the unsalvaged habitat. For example, seed-producing tree densities were 1.3–4.3 Northeastern Naturalist 640 A.C. Curtze, T.A. Carlo, and J.W. Wenzel 2018 Vol. 25, No. 4 times higher in the forest as compared to the unsalvaged areas for all but 2 tree species (Blackgum and Black Birch; Table 1). This result could be caused by 2 mechanisms triggered by the tornado damage that are not mutually exclusive: increased light exposure and more available growing area; these factors likely would boost photosynthesis and fecundity (Peters et al. 2016). Understory plant species (e.g., blackberries and greenbriers), in contrast, were more abundant in the unsalvaged habitat (Table 1), but such differences did not affect the seed rain between these habitats, probably because many were still immature plants. Seed rain in the salvaged areas was significantly reduced for most species regardless of their mode of dispersal (i.e., animal or wind). This result can be explained by a combination of source and/or dispersal limitation (Nathan and Muller-Landau 2000). For example, most species, with the exception of Pokeweed, did not have reproductive individuals close to traps in the salvaged areas, and thus, relied solely on seed dispersal to reach the site, pointing to limitations in dispersal due to distance from seed sources. In terms of animal dispersal, both the clearcut edge and occasionally unsalvaged areas received more seeds than the other habitats. This finding can be explained by the higher heterogeneity of vegetation layers (unsalvaged), abundance of biological legacies (unsalvaged), and light environments (clearcut edge and unsalvaged) that result in these areas being more attractive to animal dispersers (Chazdon et al. 2009, Wunderle 1997). The high density of biological legacies in the unsalvaged habitat may also have attracted and augmented wildlife activity by increasing the availability of food resources (e.g., insects, lichens), nesting locations, and hiding locations that were created by decaying, coarse, woody debris and standing dead snags (Lindenmayer and Noss 2006, Rost et al. 2009, Terborgh et al. 2011, Thiffault et al. 2011, Wermelinger et al. 2017). Castro et al. (2012) documented this phenomenon in a previously burned and salvage-logged pine stand in southeastern Spain, and found that Eurasian Jays preferentially selected the unsalvaged stands and avoided the salvaged stands. Other studies, such as Rost et al. (2009), similarly found that woody debris and wood erosion-barrier piles intensified bird-mediated seed dispersal and abundance in logged and burned pine stands in northeastern Iberia. As a transition zone, birds likely preferentially used the clearcut edge as a travel corridor, and thus, intensified dispersal and animal activity along edges (Levey et al. 2005). All of these attributes likely increased animal activity in the unsalvaged and clearcut edge habitats, and consequently, increased the frequency of dispersal events in these areas compared to the salvaged and, to a limited extent, the undisturbed forest habitats. Recruitment patterns We found a minimal effect of disturbance type (salvaged versus unsalvaged) and groundcover variables on seedling emergence. Even though other studies have demonstrated that groundcover can improve seed germination by protecting the seeds from predation and improving moisture retention (see Lindenmayer and Noss 2006, Terborgh et al. 2011, Thiffault et al. 2011), it appears that either the types of cover or other microsite conditions tended to offset the influence of cover on the Northeastern Naturalist Vol. 25, No. 4 A.C. Curtze, T.A. Carlo, and J.W. Wenzel 2018 641 emergence of most species. Furthermore, dominant and fecund trees, such as Tulippoplar, can have low seed-viability (5–15%), so potentially a large proportion of the recorded seed-rain failed to germinate because of unviable seeds, and not because of habitat and/or soil conditions (Burns and Honkala 1990, Hille Ris Lambers et al. 2005). Following the successful emergence of seedlings, patterns of seedling establishment (i.e., >1 y old) were strongly hampered in areas of high fern cover likely because of the high shading (George and Bazzaz 1999, Kruegler and Peterson 2009), alteration of the cycling and distribution of nitrogen (DeLuca et al. 2012), and/or chemical allelopathy of certain fern species (de Jesus Jatoba et al. 2016, Horsley 1993). For instance, several fern species present at our study sites, such as Pteridium spp. (bracken ferns) and Dennstaedtia punctilobula (Michx.) T. Moore (Hay-scented Fern), may have lowered seedling establishment through the chemical inhibition of plant growth (de Jesus Jatoba et al. 2016, Horsley 1993) and/or possibly through the alteration of nitrogen nutrient cycling that benefits the growth and development of bracken ferns (DeLuca et al. 2012). Other biotic interactions such as herbivory, negative density-dependent conspecific and heterospecific competition, and predation from pests and pathogens (e.g., fungi, insects) might also be acting to shape the spatial distribution of plants in the study site (Hille Ris Lambers et al. 2003, Terborgh 2012). Although we didn’t study their direct effects, Odocoileus virginianus (Zimmermann) (White-tailed Deer) were present at the study site, and we observed considerable deer browsing of seedlings in the site. Additional seedling damage may be due, in part, to predation from fungal pathogens responding to the higher seedling densities (Terborgh 2012) and/or increases in insect pest abundances as a result of the tornado disturbance (Stadelmann et al. 2013, Wermelinger et al. 2017). Thus, mortality caused by deer browsing and other organisms may have partially offset the benefits of large canopy-gaps created by the tornado and the salvage-logging operation on seedling recruitment (Long et al., 2007, Nuttle et al. 2013). The establishment of some species in the absence of detectable seed-rain inputs at seed traps can be explained by (1) the seed bank, and (2) dispersal events that happened before we deployed seed traps. Additional studies have demonstrated that many of the species we studied (e.g., Blackgum, greenbriers, and grapes) germinate from the seed bank (see Hille Ris Lambers et al. 2005, Small and McCarthy 2010). Thus, a persistent soil seed-bank may have further contributed to the abundance of seedlings encountered in salvaged stands and naturally disturbed sites, such as tornado blowdowns. For Tulip-poplar, Blackgum, and American Beech, salvage logging significantly affected the abundance of their established seedlings. Despite receiving fewer seed-dispersal inputs, Tulip-poplar—a shade intolerant tree—had much more establishment in the salvaged habitat likely because the species prefers high-light environments in which it grows rapidly (Burns and Honkala 1990). Similarly, abundance of Red Maple seedlings was negatively related with the amount of canopy cover, irrespective of the habitat type. Other species like American Beech and Northeastern Naturalist 642 A.C. Curtze, T.A. Carlo, and J.W. Wenzel 2018 Vol. 25, No. 4 Blackgum exhibit clonal growth, which makes them less dependent on dispersal to recruit at local scales (Burns and Honkala 1990). These species can capitalize on resources from parent trees and utilize clonal establishment in response to increased light conditions (Burns and Honkala 1990). Our results demonstrate that salvage-timber extractions may not significantly alter forest regeneration post-disturbance, with the notable exception of a handful of species such as Tulip-poplar and Blackgum. These findings align with the conclusions of Peterson and Leach (2008), Palik and Kastendick (2009), and Fidej et al. (2016) who all found that a few shade-intolerant species were more abundant in salvaged compared to unsalvaged stands. Interestingly, our results do indicate higher species richness in the salvaged compared to the unsalvaged habitats, conflicting with the results of Waldron et al. (2014) and Fidej et al. (2016). Although seedling species richness significantly varied between habitat types, the majority of species appeared to be relatively unaffected by salvage operations in tornadodisturbed stands, which is similar to the conclusions of the studies of Royo et al. (2016), Peterson and Leach (2008), and Fidej et al. (2016) conducted in various temperate deciduous forests in the Appalachians and Europe. In summary, our findings show that disturbances created by tornadoes and salvage- logging operations can have a strong influence on seed-rain patterns, and that seedling emergence appears to be strongly influenced by site-specific conditions. Avian seed-dispersal increased in association with the tornado disturbance, as well as along the edges of the created by the logging operation. In particular, salvagetimber extractions may improve the recruitment of light-demanding species and accelerate forest regeneration. The overall species richness of the regenerating plant community also increases as a result of the wind and logging disturbances, at least in the context of a forested-landscape matrix. Thus, salvage-timber extractions following windthrow events in the northeastern US may be an effective way of recouping economic losses associated with windthrow events while maximizing the regeneration of forests. Acknowledgments We thank the Penn State Biology Department, the Carnegie Museum of Natural History, and the staff of Powdermill Nature Reserve for sponsoring this research. Lastly, we thank B. Boyer, J. Salazar, M. Caraballo-Ortiz, O. Bonilla, and C. Venable for their help collecting samples in the field. Two anonymous reviewers made constructive comments on an earlier version of this manuscript. Literature Cited Anderson, G.J., and J.D. Hill. 2002. Many to flower, few to fruit: The reproductive biology of Hamamelis virginiana (Hamamelidaceae). American Journal of Botany 89(1):67–78. Burnham, K.P., and D.R. Anderson. 2002. Model Selection and Multimodel Inference: A Practical Information Theoretic Approach, 2nd Edition. Springer, New York, NY. 487 pp. Burns, R.M., and B.H. Honkala (Technical Coordinators). 1990. Silvics of North America. US Forest Service, Washington, DC. 877 pp. Northeastern Naturalist Vol. 25, No. 4 A.C. Curtze, T.A. Carlo, and J.W. Wenzel 2018 643 Cain, M.L., R.N. Nathan, and S.A. Levin. 2003. Long-distance dispersal. Ecology 84(8):1943–1944. Carlo, T.A., and J.M. Morales. 2016. Generalist birds promote tropical-forest regeneration and increase plant diversity via rare-biased seed dispersal. Ec ology 97(7):1819–1831. Carlo, T.A., and J.J. Tewksbury. 2014. Directness and tempo of avian seed-dispersal increases emergence of wild chiltepins in desert grasslands. Journal of Ecology 102(1):248–255. Caspersen, J.P., and M. Saprunoff. 2005. Seedling recruitment in a northern temperate for - est: The relative importance of supply and establishment limitation. Canadian Journal of Forest Research 35(4):978–989. Castro, J., C. Puerta-Piñero, A.B. Leverkus, G. Moreno-Rueda, and A. Sánchez-Miranda. 2012. Post-fire salvage logging alters a key plant–animal interaction for forest regeneration. Ecosphere 3(10):1–12. Chazdon, R.L. 2014. Second Growth: The Promise of Tropical Forest Regeneration in an Age of Deforestation. The Chicago University Press, Chicago, IL. 449 pp. Chazdon, R.L., C.A. Peres, D. Dent, A.E. Lugo, D. Lamb, N.E. Stork, and S.E. Miller. 2009. The potential for species conservation in tropical secondary forests. Conservation Biology 23(6):1406–1417. Coomes, D.A., and P.J. Grubb. 2000. Impacts of root competition in forests and woodlands: A theoretical framework and review of experiments. Ecological Monographs 70(2):171–207. Craves, J.A., and D. Wloch. 2012. Fruit seeds of southern Michigan, an online guide. Available online at http://seedguide.blogspot.com/p/index.html. Accessed 5 December 2015. Davis, M.B., and R.G. Shaw. 2001. Range shifts and adaptive responses to quaternary climate change. Science 292(5517):673–679. de Jesus Jatoba, L., R.M. Varela, J.M.G. Molinillo, Z.U. Din, S.C.J. Gualtieri, E. Rodrigues- Filho, and A.M. Francisco. 2016. Allelopathy of Bracken Fern (Pteridium arachnoideum): New evidence from green fronds, litter, and soil. PLoS ONE 11(8):e0161670. DOI:10.1371/journal.pone.0161670 DeLuca, T.H., S.A. Zewdie, O. Zackrisson, J.R. Healey, and D.L. Jones. 2012. Bracken Fern, (Pteridium aquilinum) (L.) Kuhn, promotes an open nitrogen cycle in heathland soils. Plant and Soil 367:521–534. Dormann C.F., J. Elith, S. Bacher, C. Buchmann, D. Carl, G. Carré, J.R. García Marquez, B. Gruber, B. Lafourcade, P.J. Leitão, T. Münkemüller, C. McClean, P.E. Osborne, B. Reineking, B. Schröder, A.K. Skidmore, D. Zurell, and S. Lautenbach. 2013. Collinearity: A review of methods to deal with it and a simulation study evaluating their performance. Ecography. 36(1):27–46. Fidej, G., A. Rozman, T.A. Nagel, L. Dakskobler, and J. Diaci. 2016. Influence of salvage logging on forest recovery following immediate-severity canopy disturbances in mixedbeech– dominated forests of Slovenia. iForest 9:430–436. Fischer, A., P. Marshall, and A. Camp. 2013. Disturbances in deciduous temperate-forest ecosystems of the northern hemisphere: Their effects on both recent and future forest development. Biodiversity Conservation 22:1863–1893. George, L.O., and F.A. Bazzaz. 1999. The fern understory as an ecological filter: Emergence and establishment of canopy-tree seedlings. Ecology 80(3):833–845. Hille Ris Lambers, J., J.S. Clark, and M. Levine. 2005. Implications for recruitment of southern woody species. Ecology 86(1):85–95. Holl, D.H. 1999. Factors limiting tropical rain-forest regeneration in abandoned pasture: Seed rain, seed germination, microclimate, and soil. Biotropica 31(2):229–242. Northeastern Naturalist 644 A.C. Curtze, T.A. Carlo, and J.W. Wenzel 2018 Vol. 25, No. 4 Horsley, S.B. 1993. Role of allelopathy in Hay-scented Fern interference with Black Cherry regeneration. Journal of Chemical Ecology 19(11):2737–2755. Hurtt, G.C., and S.W. Pacala. 1995. The consequences of recruitment limitation: Reconciling chance, history, and competitive differences between plants. Journal of Theoretical Biology 176(1):1–12. Kostel-Hughes, F., T.P. Young, and J.D. Wehr. 2005. Effects of leaf-litter depth on the emergence and seedling growth of deciduous-forest tree species in relation to seed size. The Journal of the Torrey Botanical Society 132(1):50–61. Kruegler, L.M., and C.J. Peterson. 2009. Effects of woody debris and ferns on herb-layer vegetation and deer herbivory in a Pennsylvania forest blowdown. Ecology 16(4):461–469. Levey, D.J., B.M. Bolker, J.J. Tewksbury, S. Sargent, and N.M. Haddad. 2005. Effects of landscape corridors on seed dispersal by birds. Science 309:146 –148. Levine, J.M., and D.J. Murrell. 2003. The community-level consequences of seed-dispersal patterns. Annual Reviews of Ecology, Evolution, and Systematics 34:549–574. Lindenmayer, D.B., and R.F. Noss. 2006. Salvage logging, ecosystem processes, and biodiversity conservation. Conservation Biology 20(4):949–958. Long, Z.T., T.H. Pendergast IV, and W.P. Carson. 2007. The impact of deer on relationships between tree growth and mortality in an old-growth beech–maple forest. Forest Ecology and Management 252(1–3):230–238. McCune, B., and M.J. Mefford. 2011. PC-ORD. Multivariate analysis of ecological data. Version 6. MjM Software, Gleneden Beach, OR. Mitchell, K. 2015. Quantitative analysis by the point-centered quarter method. Available online at https://arxiv.org/abs/1010.3303v2. Accessed 15 April 2017. Muller-Landau, H.C. 2008. Colonization-related trade-offs in tropical forests and their role in the maintenance of plant species diversity. Pp. 182–195, In W.P. Carson and S.A. Schnitzer (Eds.). Tropical Forest Community Ecology, Wiley-Blackwell, Oxford, UK. 536 pp. Nappi, A., P. Drapeau, and J.-P.L. Savard. 2004. Salvage logging after wildfire in boreal forest: Is it becoming a hot issue for wildlife? The Forestry Chronicle 80(1):67–74. Nathan, R., and H.C. Muller-Landau. 2000. Spatial patterns of seed dispersal, their determinants, and consequences for recruitment. Trends in Ecology & Evolution 15:278-285. National Oceanic and Atmospheric Administration (NOAA). 2017. Local climatological data-station details. Available online at https://www.ncdc.noaa.gov/cdo-web/datasets/ LCD/stations/WBAN:04726/detail. Accessed 16 July 2017. Natural Resources Conservation Service (NRCS). 2017. Published soil surveys of Pennsylvania. Available online at https://www.nrcs.usda.gov/wps/portal/nrcs/surveylist/soils/ survey/state/?stateId=PA. Accessed 15 April 2016. Nuttle, T., A.A. Royo, M.B. Adams, and W.P. Carson. 2013. Historic-disturbance regimes promote tree diversity only under low browsing regimes in eastern deciduous forest. Ecological Monographs 83(1):3–17. Palik, B., and D. Kastendick. 2009. Woody-plant regeneration after blowdown, salvage logging, and prescribed fire in a northern Minnesota forest. Forest Ecology and Management 258:1323–1330. Pennsylvania Department of Conservation and Natural Resources (PADCNR). 2006–2008. PAMAP program-3.2 ft digital elevation model. Available online at http://maps.psiee. psu.edu/ImageryNavigator/. Accessed 22 July 2017. Peters, V., T.A. Carlo, M.A.R. Mello, R.A. Rice, D.W. Tallamy, S.A. Caudill, and T.H. Fleming. 2016. Selection of species for Neotropical tree-based agroecosystems: Perspectives for decision-making. Bioscience 66(12):1046–1056. Northeastern Naturalist Vol. 25, No. 4 A.C. Curtze, T.A. Carlo, and J.W. Wenzel 2018 645 Peterson, C.J., and A.D. Leach. 2008. Limited salvage-logging effects on forest regeneration after moderate-severity windthrow. Ecological Applications 18(2):407–420. Puerta-Piñero, C., A. Sánchez-Miranda, A. Leverkus, and J. Castro. 2010. Management of burnt wood after fire affects post-dispersal acorn predation. Forest Ecology and Management 260(3):345–352. Rhoads, A.F., and T.A. Block. 2005. Trees of Pennsylvania: A Complete Reference Guide, 1st Edition. University of Pennsylvania Press, Philadelphia, PA. 407 pp. Rost, J., P. Pons, and J.M. Bas. 2009. Can salvage logging affect seed dispersal by birds into burned forests? Acta Oecologica 35:763–768. Royo, A.A., C.J. Peterson, J.S. Stanovick, and W.P. Carson. 2016. Evaluating the ecological impacts of salvage logging: Can natural and anthropogenic disturbances promote coexistence? Ecology 97(6):1566–1582. SAS. 2012. JMP, Version 10 pro. SAS Institute Inc., Cary, NC. Small, C.J., and B.C. McCarthy. 2010. Seed bank variation under contrasting site-quality conditions in mixed-oak forest of southeastern Ohio, USA. International Journal of Forest Research Volume 2010:1687–9368. Smith-Ramírez, C., V. Maturana, A. Gaxiola, and M. Carmona. 2014. Salvage logging by indigenous people in a Chilean conifer forest. Forest Science 60(6):1100–1106. Stadelmann, G., Bugmann, H., Meier, F., Wermelinger, B., and C. Bigler. 2013. Effects of salvage logging and sanitation felling on bark beetle (Ips typographus (L.)) infestations. Forest Ecology and Management 305:273–281. Stueve, K.M., C.H. Perry, M.D. Nelson, S.P. Healey, A.. Hill, G.G. Moisen, W.B. Cohen, D.D. Gormanson, and C. Huang. 2011. Ecological importance of intermediate windstorms rivals large, infrequent disturbances in the northern Great Lakes. Ecosphere 2(1):1–21. Terborgh, J. 2012. Enemies maintain hyperdiverse tropical forests. American Naturalist 179(3):303–314. Terborgh, J., P. Alvarez-Loayza, K. Dexter, F. Cornejo, and C. Carrasco. 2011. Decomposing dispersal limitation: Limits on fecundity or seed distribution? Journal of Ecology 99(4):935–944. Thiffault, E., K.D. Hannam, D. Paré, B.D. Titus, P.W. Hazlett, D.G. Maynard, and S. Brais. 2011. Effects of forest-biomass harvesting on soil productivity in boreal and temperate forests: A review. Environmental Reviews 19:278–309. Waldron, K., J. Ruel, and S. Gauthier. 2013. Forest structural attributes after windthrow and consequences of salvage logging. Forest Ecology and Management 289:28–37. Waldron, K., J. Ruel, S. Gauthier, L. De Granpré, and C.J. Peterson. 2014. Effects of salvage logging on microsites, plant composition, and regeneration. Applied Vegetation Science 17:323–337. Wermelinger, B., M. Moretti, P. Duelli, T. Lachat, G.B. Pezzatti, and M.K. Obrist. 2017. Impact of windthrow and salvage-logging on taxonomic and functional diversity of forest arthropods. Forest Ecology and Management 391:9–18. Wunderle, J.M., Jr. 1997. The role of animal dispersal in accelerating native-forest regeneration on degraded tropical lands. Forest Ecology and Manageme nt 99(1–2):223–235.