<<

EFFECTIVENESS OF AMUR HONEYSUCKLE

() REMOVAL TREATMENTS IN

RAVINE FORESTS OF CENTRAL OHIO

Thesis

Presented in Partial Fulfillment of the Requirements for

the Degree Master of Science in the Graduate

School of The Ohio State University

By

Edmund M. Ingman, B. A.

*****

The Ohio State University 2009

Advisory Committee: Approved by Dr. David M. Hix, Advisor Dr. Peter Curtis ______Dr. P. Charles Goebel Advisor Environmental Science Graduate Program

ABSTRACT

Central Ohio ravine forests are being subjected to increasing levels of disturbance

due to residential and commercial development. This development has led to increased

fragmentation of these urban forests, allowing non-native species to invade, e.g.,

Lonicera maackii (Rupr.) Herder. L. maackii is detrimental in forest ecosystems due to

its allelopathic effects, fast growth rates, and leaf phenology. In central Ohio, community

groups have conducted removal efforts aimed at eradicating this species from two urban

ravines, Adena Brook and Rush Run.

Stand composition, and L. maackii abundance, height, and response to treatment

were determined following sampling in the summer of 2008. Fifteen 200-m2 plots were established in three ravines in Franklin County, Ohio. Flint Run was the reference for this study, and three plots were established in this relatively undisturbed ravine. In both

Rush Run and Adena Brook, three plots were established where no treatment had occurred and three plots were established where treatment had occurred. It was anticipated that the urban ravines had higher densities of L. maackii and decreased woody species diversity. It was also expected that after removal efforts were completed bank regeneration would result in lower densities of L. maackii compared with planting of woody species.

ii

There was a significant difference between density of L. maackii in the disturbed areas compared with the reference study location, as well as treated areas compared with untreated areas (p < 0.001 for both). There were differences in the density, height, and height distributions. The morality percentage was significantly different between treated plots in Rush Run and Adena Brook (p = 0.001).

In untreated plots across all locations, there were significant differences in the

Shannon-Weiner Diversity Index between the reference study area and the disturbed, untreated areas at the 1- by 1- m level (p < 0.001). There was lower woody plant diversity in the reference area compared with the untreated, disturbed areas.

Finally, there was a significant difference in woody plant density in treated Adena

Brook plots compared with treated Rush Run plots (p = 0.03). The L. maackii reproduction rate was lower at Rush Run compared with Adena Brook, however, this difference was not statistically significant (p = 0.905). Treated Rush Run plots had higher density of L. maackii individuals per square meter compared with both Flint Run and treated Adena Brook plots, however, the difference was also not significant (p = 0.5).

An understanding of the effectiveness of L. maackii removal efforts in urban ravine forests will provide beneficial information to Central Ohio community groups, allowing them to better restore these urban ravine forests.

iii

Dedicated in loving memory of my mother, Phyllis.

Also dedicated to my amazing wife Lindsay.

iv

ACKNOWLEDGMENTS

I would like to thank my advisor, Dr. David Hix, for his gracious assistance in this research. I would also like to thank my committee, Dr. Peter Curtis and Dr. P. Charles

Goebel, for their guidance during the process, and Chris Holloman, Ph. D. for his assistance with the statistical analysis. Heather Dean of Friends of the Lower Olentangy

Watershed, Susan Barrett-Michaels of the Adena Brook Community, Greg Schneider of the ODNR Department of Natural Areas and Preserves, and Elayne Grody of the

Columbus Parks and Recreation department, provided their assistance in establishing the sample plots, were helpful during interviews, and took the time to answer questions.

Their work was invaluable to this study! Finally I would like to thank my wife, Lindsay, for her company and assistance while sampling and writing this thesis.

v

VITA

2003 ……………………………………..… Bachelor of Arts, Ohio Wesleyan University

2005 – Present …………………………….. Graduate Student, The Ohio State University

FIELDS OF STUDY

Major Field: Environmental Science

vi

TABLE OF CONTENTS Page

Abstract ...…………………………………………………………………...…………… ii Dedication ………………………………………………………………………...…….. iv Acknowledgments ……………………………………………………………………..… v Vita ……………………………………………………………………………….…...… vi List of Tables ………………………………………………………………………….. viii List of Figures .………………………………………………………...……………...... ix

Chapters:

1. Introduction ……………………………………………………………………… 1 2. Literature Review ……………………………………………………………….. 6 3. Study Area ………….………………………………………………………….. 16 4. Methods ………………………………………………………………………… 20 5. Results ………………………………………………………………………….. 25 6. Discussion ……………………………………………………………………… 30 7. Conclusions ……...……………………………………………………………... 35

Literature Cited ………………………………………………………………………… 39

Appendix A: Tables and Figures…………………………………………………..…… 44

Appendix B: Woody plant species stand density and relative densities ……………...…71

vii

LIST OF TABLES Table Page 1 Characteristics of the three study locations in central Ohio …..……...……….. 45

2 Native woody plant species planted at Rush Run after treatment of L. maackii by volunteers …………….……………………………………... 46

3 Analysis of variance of L. maackii density among study locations and treatment types ………………………………………………………..…… 47

4 L. maackii sampled at each study location by treatment ……………………..... 48

5 Analysis of variance of L. maackii density amongst treatment………………… 49

6 L. maackii height distributions by plot ...………………………………………. 50

7 L. maackii mortality and reproduction percentages ….………………………... 51

8 L. maackii treatment summary ……………………………………………….… 52

9 Environmental cover noted by square-meter quadrat …..……………………… 53

10 Mean light levels by plot ……………………………………………………….. 54

11 Shannon-Weiner Diversity Index values, based on species densities, by study location ………………………………………….…………. 55

12 Woody plant density and species richness by study location and treatment type ……………………………………………………………… 56

13 Woody plant density by plot …………………………………………………… 57

viii

LIST OF FIGURES Figure Page 1 Map of the three study locations, which are located in Franklin Country, central Ohio …………………………………………………………... 58

2 The three 10- by 20-m plots sampled in Flint Run ravine are shown as dark rectangles within the outlined box ..……………………………………. 59

3 The six Rush Run plots are represented as black rectangles within the box ………………………………………..………………………………… 60

4 Adena Brook sample plots are all located in the box near the center of the map ……………………..……………………………………………….. 61

5 Sampling design ………………………………………………………………... 62

6 Number of L. maackii (stems) in each sample plot ……………………………. 63

7 Total number of individual L. maackii stems sampled ….……………………... 64

8 Mean number of individual L. maackii stems by sample plot condition …...….. 65

9 Mean height (cm) by area and treatment ………….…………………………… 66

10 Height distributions in untreated plots in all three areas …..……………...……. 67

11 Height distributions in treated plots for both areas ……………….……………. 68

12 Shannon-Weiner Diversity Index values for woody plant species found in each sample plot ………………………………………………………………….….. 69

13 Woody plant species density in treated plots …..………………………………. 70

ix

CHAPTER 1

INTRODUCTION

Disturbance levels have been steadily increasing in central Ohio since settlement, and the ravine forests are becoming more prone to invasion by non-native species, which have a detrimental impact on native flora. One such species, Amur honeysuckle

(Lonicera maackii (Rupr.) Herder), is especially invasive. Local community groups have

sought to eradicate this species from the ravine forests. This thesis will examine the

effectiveness of their efforts and, in particular, their use of removal techniques.

L. maackii is an introduced species that is now common throughout the upper

Midwest. This multi-stemmed, is native to the Asian continent, but was

brought to the U. S. in 1896 (Luke and Thieret 1995). Since its escape from cultivation,

it has changed the dynamics of North American deciduous forests. It thrives in highly

fragmented forests (Brothers and Spingarn 1991), near human habitation (Gayek and

Quigley 2001), and is widely dispersed by avian consumers (Borgmann and Rodewald

2005). These forest conditions are very common in many areas of central Ohio.

Community groups have sought to eradicate this species using volunteer labor and implementing recommended removal strategies in an attempt to restore ravine forests ecosystems, i.e., to emulate pre-European settlement stand structures and functions.

1

Central Ohio ravine forests are unique ecosystems that are highly prized for residential development due to their seclusion and abundance of wildlife. These areas have historically been less affected than non-ravine areas due to the difficulty of access.

Most portions have steep slopes and are unsuitable for agricultural use. The high erosion potential and slope instability make them hazardous for typical home construction methods. The residential communities that have developed in the ravines have sought to balance home ownership concerns against ecosystem concerns. This has lead some residents to be actively involved in environmental stewardship by joining community groups.

There are several of these community groups, which have been active in

removing L. maackii from Franklin County ravines. In central Ohio, there are nine

distinct watersheds, all of which eventually lead to the Scioto River. There are eight

organized community groups who seek to support neighborhood-level efforts to monitor

the overall watershed health, raise awareness about watershed issues, and educate the

public. Friends of the Lower Olentangy Watershed (FLOW) is one of the eight groups.

FLOW was established in 1997 in order to support the Olentangy River watershed

from the northern border of Franklin County to the confluence with the Scioto River.

While FLOW is primarily concerned with the organization of monitoring and clean up

efforts around the Olentangy River, they also help coordinate work done surrounding the

tributaries of the river. While there are several tributary groups, two are notable and

central to this study: the Adena Brook Community and the Friends of Rush Run.

2

The Adena Brook Community formed in 2002 in order to encourage biodiversity,

help clean the waterway, provide safer neighborhoods, and remove litter (Anonymous

2008d). Their work has taken place along the Adena Brook tributary (also known as

Overbrook) between Cooke Road and Glenmont Avenue, east of High Street and west of

Interstate 71. Adena Brook has a large group of volunteers, organizes litter and non- native species removal efforts every Saturday during the growing season each year, and advocates for the tributary’s water health to city and regional officials.

The Friends of Rush Run consists of one organizer and a small volunteer pool.

They help protect and maintain Rush Run Creek, which extends west from Interstate 71

to the Olentangy River, south of State Route 161, and north of Walnut Grove Cemetery

and Park Boulevard in Worthington, Ohio. The Friends of Rush Run has been organized

since 1999 and has worked to protect the waterway from development, dumping, and

litter. They have also organized invasive species efforts, most notably an extensive

removal event in 2007. This singular event removed L. maackii from the stream bank as

it passes through Park Boulevard Park. The removal efforts were followed by the

planting of, “…94 native trees and , 50 ferns and over 900 perennial wildflower

plugs.” (FLOW 2008) This study assesses the effectiveness of removal efforts and

suppression techniques implemented by these two neighborhood groups.

L. maackii impacts a forest ecosystem due to its allelopathic effects, unique leaf phenology, and efficient seed dispersal mechanisms. L. maackii has been shown to be allelopathic resulting in decreased species richness (Collier et al. 2001) and leaf extractions have caused diminished growth rates (Trisel 1997). The leaf phenology of L.

3

maackii has a significant shading impact on native flora, affecting their growth rates, and and seed production. L. maackii leafs out earlier in the spring than native species and retains its leaves longer into autumn. This longer leaf phenology allows L. maackii more opportunity for photosynthesis, which constitutes an advantage over native species.

With an average height growth rate of 0.4 m yr-1, L. maackii tends to out-grow most

native species and successfully competes for available light resources (Deering and

Vankat 1999). Combined with its leaf phenology, the fast growth rate can cause L.

maackii to have increased shading over native forest floor flora. Finally, L. maackii

are readily consumed and distributed by Cardinalis cardinalis and Turdus

migratorius (Rodewald and Borgemann 2005). These authors indicate that in central

Ohio these bird species are primary modes of L. maackii seed dispersal.

Given these growing characteristics, once L. maackii is introduced to the forest

ecosystem it will occupy a niche and disrupt the recruitment of native woody plant

species. L. maackii’s allelopathic impacts on surrounding vegetation cause it to have a detrimental impact on native vegetation. Whether mortality results may depend on the species, but disruptions of woody plant species’ growth over time has been shown by

Hartman and McCarthy (2007), and of perennial forest herbs by Miller and Gorchov

(2004). These disruptions may lead to consequences to the forest ecosystem; one such impact is in woody plant species diversity. Removal methods are, therefore, logical responses to realign forest structure. Once removal efforts have been conducted, suppression techniques center on controlling future recruitment of L. maackii individuals.

4

The native seedbank provides a rich seed source that can establish dense growth around

L. maackii and potentially limit its growth through shading.

A literature review was conducted to locate peer-reviewed journal studies of L. maackii. Many studies were found for upland, contiguous forests, or in an urban-to-rural gradient of forests. Studies were found on the specie’s effects on fragmented forests in urban settings, as well. No published study was found on L. maackii removal in urban ravine forests by community groups. Since the ravine systems are unique forest ecosystems and there has been no study of L. maackii within these systems, it is anticipated that this project will inform future studies of L. maackii as well as be important for neighborhood groups in implementing removal events.

Hypotheses and Objectives

In relation to the L. maackii removal efforts by the community groups, two hypotheses were established.

First hypothesis: woody plant species diversity will be higher in Flint Run, which has low densities of L. maackii, compared with untreated areas of Adena Brook and Rush

Run, which have high densities of L. maackii.

Second hypothesis: in a central Ohio ravine forest where L. maackii has been treated with herbicide (after cutting each stem), there will be a greater density of woody plant species when the natural seed bank is allowed to regenerate the treated area compared with a similar central Ohio ravine forest where planting of native woody plant species has occurred following the same treatment of L. maackii.

5

CHAPTER 2

LITERATURE REVIEW

L. maackii Description, History, and Niche L. maackii has, at the time of this writing, invaded 21 Ohio counties, 26 states, and one Canadian province (Natural Resources Conservation Service 2008). This deciduous, multi-stemmed shrub originated in the Amur River valley in China and was introduced into the United States as an ornamental shrub in 1896 (Luken and Thieret

1995). L. maackii has a shallow root system (Czarapata 2005) and it produces abundant small, round, red with yellow seeds in the center, which ripen in early summer

(Deering and Vankat 1999, Luken and Mattimiro 1991, and Bartuszevige et al. 2006).

The fruits are opposite along the outer branches, rotating roughly 90 degrees between fruits (Personal observation). L. maackii can obtain a maximum height of 5.5 m, with an

average height growth at 0.4 m yr-1 (Deering and Vankat 1999). The leaves are ovate-

elliptical, dark green above, lighter below, 5-8 cm long, and pubescent along veins

(above and below) (Braun 1961, Czarapata 2005). It’s are white with peduncles

0.5 cm long, which yellow with age.

While the exact date of L. maackii’s arrival in Ohio is unknown, Hutchinson and

Vankat (1998) report that a garden club in the vicinity of the city of Oxford, Butler

6

County, planted L. maackii in 1960. Braun (1961) reported that the species was present

in Hamilton County pastures and woodlands. L. maackii has since spread to 20 other

counties (Natural Resources Conservation Service 2008).

The realized niche of L. maackii is the edges of forests and fence rows, primarily

due to high light conditions, which tend to be present in these areas. Several authors

(Borgmann and Rodewald 2005 and Gayek and Quigley 2001) report that L. maackii has

a greater percent cover in urban forest landscapes than rural ones. Bartuszevige et al.

(2006) found a positive correlation between the amount of forest edge and abundance of

L. maackii. These authors also found that colonization does not occur at the edge first.

They found individuals of the species throughout woodlots that were of the same age,

with some at edges and others in the interior.

Borgmann and Rodewald (2005) noted that avian seedivores play a primary role

in the dispersal of L. maackii, observing Carinalis cardinalis and Turdus migratorius foraging on L. maackii. These birds are associated with edges and therefore may be contributing to the dispersal of the species in urban forests. These and other species may be dispersing L. maackii into forest interiors, but L. maackii has both reduced growth rates and reproductive capacity due to shade intolerance in the understory (Bartuszevige et al. 2006).

L. maackii also has allelopathic impacts on the forest floor community after initial

invasion. Collier et al. (2002) studied the effects of L. maackii on forest floor

immediately surrounding it’s base. They established two rings, an inner ring of 25.0 cm

and the outer of 83.6 cm around twenty individuals, and sampled herbaceous and woody

7

species within these plots. They found that there was a 53% decrease in species richness

and a 63% decrease in cover within the 25.0 cm ring. These percentages increased with residence time of L. maackii. The data suggests that the presence of L. maackii in second-growth forests decreases plant richness and ground cover. In another study of

how L. maackii impacts forest flora, Miller and Gorchov (2004) studied the effect of shading resulting from the presence of L. maackii on Allium burdickii, Thalictrum thalictroides, and Viola pubescens. They found that all three forest floor herbaceous species demonstrated decreased growth rates, flower production, and seed production, when shaded by L. maackii. Plant mortality was not associated with shading, but all three species demonstrated reduced growth and reproduction, which may eventually affect their long-term viability.

Although there have been studies regarding the allelopathy of L. maackii (Trisel

1997), no known studies have reported a causal mechanism. Hartman and McCarthy

(2007) examined the allelopathic effects of L. maackii on forest overstory productivity.

Dendrochronological analysis was performed ten years prior to and ten years following

invasion. It was found that radial and basal growth rates were significantly decreased

compared to non-invaded areas. The basal area growth rate was reduced by 15.8%,

typically occurring 6.25 ± 1.24 years after invasion. They hypothesized that the most

likely explanation of these aboveground effects were associated with belowground root

competition. While tree roots grow deep into the soil to anchor the plant and search for

nutrients, L. maackii has a shallow and wide-spreading root system. The authors suggested that the stratification of L. maackii root systems may be robbing the deeper tree

8

roots of essential nutrients, water, and may potentially be changing the chemical composition of the soil (Hartman and McCarthy 2007).

Invasive Species Definition The Ohio Department of Natural Resources (2001) defines an invasive species as

having, “fast growth rates, high production, rapid vegetative spread, and efficient

seed dispersal and germination”. Schwartz (1997) further characterized invasive species

using nine criteria. Most relevant to this research, an invasive species is one having

individuals who are without historical documentation (i.e., by earliest records) and have a

documented means of introduction. When taken cumulatively, these criteria act as

guides to identifying and understanding the manner by which a species invades a certain

niche. It can be argued that L. maackii meets all of the above criteria. It has a growth

rate of 0.4 m per year, produces “copious amounts” of fruits (Deering and Vankat 1999),

reaches sexual maturity within five years, has multiple seed dispersal mechanisms, and

has a documented historical introduction.

Invasive species have various ecological and economic impacts on local

communities. L. maackii has been shown to increase the mortality of forest flora (Miller

and Gorchov 2004), reduce overstory productivity (Hartman and McCarthy 2007), reduce

plant richness and abundance (Collier et al. 2002), and has demonstrated resiliency to

removal (Hartman and McCarthy 2004; Luken and Mattimori 1991).

9

Theories of Invasibility While it is well known that L. maackii invaded Ohio’s ecosystems following the actions of homeowners and garden clubs, it is important to understand the underlying causes and processes that support the species’ continued range expansion. There are two competing ideologies concerning the invasibility of communities by non-native species.

The first postulates that disturbed forest communities are more prone to invasion than pristine forest communities (Brothers and Spingarn 1992). Brothers and Spingarn’s research suggests that alien species are abundant on the edges of forest fragments as opposed to in their interiors. They found a dramatic drop in non-native species in forest interiors, which led them to hypothesize that low light availability and low levels of disturbance are the main drivers of this trend.

The second ideology postulates that disturbance is not as important as access

points for invasion. Planty-Tabacchi et al. (1996) found that species-rich riparian forest communities have a higher density of invasive species than species-poor communities and that young stands are more invasible than old stands. They hypothesized that these

trends are related to disturbance regimes and the structures of each community or stand.

The counter-intuitive nature of these results was also examined in a study conducted by

Johansson et al. (1996), which examined the function of rivers in plant dispersal. In their

study, a positive correlation was found between the floating time of diaspores and the

frequencies of the species (Johansson et al. 1996). While it is unknown whether or not L.

maackii seeds have a long float time, this study shows the importance of rivers for

determining flora composition of riparian forest communities. The Planty-Tabacchi et al.

(1996) and Johansson et al. (1996) studies indicate that species-rich riparian communities 10

may be dispersal corridors for invasive species, and therefore may act as another means

for L. maackii to expand it’s range. Avian dispersal, combined with anthropogenic

introduction, leave young riparian forest communities especially vulnerable to invasion

and can inadvertently act as loci for dispersal across large areas.

The conditions that assist in the invasion of Ohio forest communities by L.

maackii have been studied by Hutchinson and Vankat (1997). These authors found that

the presence of L. maackii, in an urban area, is facilitated by, “five variables (in

descending order of importance): tree canopy cover, distance from Oxford, shade

tolerance index (a calculation of recent light regimes based off of the importance

percentage for sapling species multiplied by shade tolerance values), tree basal area, and

time since invasion.” They also found an inverse correlation between L. maackii cover

and the basal area of shade tolerant tree species, e.g., Acer saccharum and Fagus

grandifolia. The results also seem to indicate a positive correlation between L. maackii

cover, tree canopy cover, shade tolerance index, and basal area. Therefore, older stands

are more difficult for the species to invade than younger stands. The authors also

conclude that forests should be managed to reduce the number of canopy disturbances,

because any disturbance would almost certainly be accompanied by a change in the

amount of light reaching the forest floor.

L. maackii Removal Methods

L. maackii removal methods include physical, chemical, or a combination. At this time, there are no known biological controls. Physical removal methods include repeated

11

clipping and use of a weed or root wrench. Repeated clipping in an open area will result

in suppression of L. maackii growth. In a forest interior, it can cause mortality. Weed or

root wrench removal of L. maackii may not result in complete eradication because of the

species’ ability to resprout from remaining roots. This method also can cause disruption

in the growth rates of neighboring plants due to soil disruption.

After clipping, the stumps can be chemically treated with 20% a.i. glyphosate or

12.5% a.i. triclopyr, both of which have been shown to be effective at controlling

resprouting (Czarapata 2005). Basal bark treatments are also available and do not require

clipping. This treatment calls for a 12.5% a.i. triclopyr solution formulated with

penetrating oils to be applied to the bark. Finally, foliar sprays may be used, but they are

toxic to the surrounding flora.

After removal methods have been attempted, suppression methods include

planting woody plant species or allowing seed bank regeneration. Both methods result in

shading of new L. maackii sprouts, which will suppress future growth. The seed bank regeneration method allows treated sites to be reclaimed by seeds already present on the forest floor. Planting native woody plant seedlings attempts to shade the forest floor with woody plant species characteristic of the ecosystem. Seed bank regeneration is expected to take relatively longer to achieve the same shading effects. Czarapata (2005) and

Hartman and McCarthy (2004) recommend planting native species in removal areas to discourage reinvasion, in combination with one of the physical or chemical methods. No studies were found that discussed the merits of seed bank regeneration over planting as a primary suppression method.

12

While the aforementioned methods can be very effective, there are other

considerations that should be taken into account before treatment is attempted. Luken

and Mattimiro (1991) designed a study where yearly clipping was employed for three

years in forest preserves owned by Northern Kentucky University. They studied

resprouting after clipping and measured the ability of the plant to respond to treatment. It was found that open-grown populations of L. maackii demonstrated greater resiliency to the treatment when compared with forest populations and clipped individuals showed decreased net primary productivity per shrub. The authors suggest that a forest is an inherently stressful environment for L. maackii; therefore, in a forest the species is more susceptible to treatment stress than their open-grown counterparts. For forest-grown L. maackii, the authors recommended annually clipping until carbohydrate root reserves are exhausted, while clipping coupled with herbicide treatment provided the most effective results for open-grown individuals (Luken and Mattimiro 1991).

Hartman and McCarthy (2004) examined the effectiveness of two chemical

eradication methods: glyphosate herbicide treatment after cutting, and glyphosate

herbicide injection. They found that at least 94% of the L. maackii died with either type

of treatment. The study also examined the effectiveness of planting native woody plant

species in eradicated areas versus in control plots. After three years, the seedlings planted in the eradicated plots had survivorship between 45 – 51%, compared with 32% in the control plot. No comprehensive strategy has been adopted regarding the treatment of L. maackii as an invasive species; however, either form of herbicide treatment, along

13

with native woody plant planting, may result in the eradication of L. maackii (Hartman and McCarthy 2004).

Post-treatment effects were studied by Gill and Mitsch (2003) following removal

of L. maackii in a bottomland hardwood forest along the Olentangy River, in Franklin

county, Ohio, south of Adena Brook. Their study removed L. maackii through clipping

alone, but clipped individuals were primarily located within a closed forest. The area was

divided into a treated and untreated section, with five, 4- by 4- m quadrats established in

each condition, for a total of ten quadrats. Vegetation cover values were estimated per

species and L. maackii individuals were counted (clipped or living). Gill and Mitsch

found that L. maackii coverage was reduced by 77.2%, with the average number of

species increasing from 6.8 to 8.2. Gill and Mitsch’s results demonstrate that decreases

in L. maackii can be achieved through treatment and that other species tend to respond

positively.

Runkle et al. (2007) conducted a similar study, but examined the vegetative cover

eight years after treatment. The authors examined L. maackii removal on forest floor

flora diversity, cover, and tree seedling density. In their study, the authors established ten pairs of plots in the Sugarcreek Reserve in Greene County, Ohio. Each pair contained an experimental plot where treatment was applied and another where no treatment was attempted. Initial data collected included woody plant species’ abundance and diameter at breast height (DBH). This data was used to classify the forested stand age as young, medium, or old aged. Cut stumps of L. maackii were treated with a 1:10 dilution of

Round-Up to water application. All vegetation was sampled after treatment by species,

14

percent cover, and, if woody, the number of plants was recorded. L. maackii abundance

and height was also recorded within the control plot. After eight years, the authors

resampled the pairs of plots. They found that the mean number of species in treated

versus control plots increased over time. The total percent cover of plant species

increased over time per square meter as well. L. maackii had reinvaded the treated plots,

but the height and densities was less than in the control plots. The authors found that

over time the species richness increased, but that long periods of time are needed for

vegetation to establish after an invasion and treatment. This study demonstrates how disturbed areas may respond to treatment over time; even if no additional follow up treatment is conducted.

15

CHAPTER 3

STUDY AREA

Early land surveys indicate that Ohio was dominated by deciduous forests from the Appalachian Mountains to the Lake Erie basin (Gordon 1966). Braun (1961) notes that most of Ohio lies within the beech- forest region, which is dominated by Fagus grandifolia (Ehrh.) (American beech) and Acer saccharum (Marshall) (sugar maple).

Other species such as Ulmus americana (L.) (American elm) and Fraxinus pennsylvanica

(Marshall) (green ash) occur in floodplain areas, e.g., along the Olentangy River in central Ohio (Friends of the Lower Olentangy Watershed 2001). Today, central Ohio forests have largely been replaced by urban and suburban infrastructure.

The climate of central Ohio is continental with an average yearly temperature of

11.1 °C, and a yearly average precipitation of 7.8 cm (McLoda and Parkinson 1980). The growing season extends from May to October, which translates into approximately 130 frost-free days per year (Braun 1961).

The bedrock of central Franklin County consists of Devonian dolomite, non- calcareous Ohio shale, and Waverly shale (McLoda and Parkinson 1980). The Olentangy

River runs through level till plains and the streams in the ravine forests studied (i.e., Flint

Run, Rush Run, and Adena Brook, Figure 1) flow into the river from east to west.

16

The three study ravines, Flint Run, Rush Run, and Adena Brook (Table 1), are

underlain by non-calcareous Ohio shale. All three ravines are east of the Olentangy

River. The dominant soil series in the study locations are Alexandria, Ross, Cardington,

and Bennington silt loams (McLoda and Parkinson 1980). Alexandria soils are deep,

well-drained soils formed in glacial till. Ross soils form in deep, well-drained alluvium

in floodplains. Cardington soils are similar to Alexandria, except they have slower

permeability. Bennington soils form in deep, somewhat poorly drained till on very gentle

slopes. The ravines all have outcroppings of Ohio shale along eroded portions of their

side slopes and are steeply sloping.

Flint Run

At Flint Run (40°7.4’N, 83°1.3’W; Figure 2), plots were established in

Alexandria silt loam, and slopes averaged 18-25% (Anonymous 2008c). The run also

crosses Ross silt loam and Genesse silt loam before entering the Olentangy River. The

ravine also has Mitiwanga soils and shale outcroppings. The run is periodically flooded, but Alexandria silt loam has a permeability of between 1.5 cm and 5.1 cm per hour

(Anonymous 2008c), and does not remain inundated for long periods of time.

Flint Run is located within Camp Mary Orton, which has been privately owned

and operated by the Godman Guild Association since 1910. This camp lies just north of

I-270 and west of highway U.S.-23, and is bordered on the south by residential communities and Pontifical College Josephinum. High Banks Metro Park is on the north

17

side of the camp along with an office park. Residential development is on the ravines

eastern edge. Despite this development, Flint Run remains relatively undisturbed.

Rush Run

At Rush Run (40°4.8’N, 83°0.9’W; Figure 3), plots were established in Ross silt loam, that has a permeability of 1.5 to 5.1 cm per hour, and experiences occasional flooding (Anonymous 2008b). The slopes were 0 – 2 percent.

Rush Run is located in Worthington, and is bordered on all sides by residential

communities, commercial, and light industrial development. A cemetery is located south

of the run. Dominant canopy tree species are Platanus occidentalis (L.) (sycamore),

Populus deltoides (Bartr.) (eastern cottonwood), and Quercus palustris (Muenchh.) (pin

oak) (Simpson 1993). Removal of L. maackii occurred in Park Boulevard Park in 2007.

The City of Worthington Parks and Recreation Department manages the park.

Adena Brook

At Adena Brook (40°2.8’N, 83°0.8’W; Figure 4), plots were established in

Alexandria silt loam (Anonymous, 2008a). The slopes were 18-25%. The ravine crosses

Ross silt soil before entering the Olentangy River.

Adena Brook is the southernmost ravine in this study. Platanus occidentalis (L.)

(sycamore), Fagus grandifolia (Ehrh.) (American beech), Ulmus americana (L.)

(American elm), and Carya ovata (Mill.) (shagbark hickory) are the dominant canopy tree species (Adena Brook Community 2007b). The ravine area has residential properties

18

and commercial development on all sides. Ownership near the ravine is both public and

private. The Ohio Department of Natural Resources, Division of Natural Areas and

Preserves manages the public land.

The soil series of each area was verified by examining the soil profile and

comparing the properties of the horizons with the published descriptions of the expected

soils (McLoda and Parkinson 1980).

Community Organizations

Community organizations (i.e., Friends of Rush Run and the Adena Brook

Community) have begun treatment of L. maackii. The Friends of Rush Run clipped L.

maackii, applied glyphosate herbicide to the stumps, and planted native species (Table 2)

at Park Boulevard Park in Worthington, Ohio in the spring of 2007. The Friends of

Adena Brook clipped, applied glyphosate herbicide to the stumps, and anticipated that the

seed bank would regenerate the area. Their work along Adena Brook occurred from

2005 to 2007. Flint Run has experienced little anthropogenic disturbance for the past 90

years and there is little L. maackii present (Personal observation).

19

CHAPTER 4

METHODS

Data Collection

Sample plots were established in three ravine forests in order to address the

hypotheses. Three 200-m2 (10-m by 20-m) plots were established in Flint Run. Six 200-

m2 plots were established in both Rush Run and Adena Brook; three 200-m2 plots were

established in treated areas and three in untreated areas. Each plot was further divided into 200, 1- by 1- m quadrats that were sampled during the summer of 2008 (Figure 5).

The sample plots established in each area were at least 10 m apart, and their long

axes were parallel to the water course (i.e., the run). Rectangular plots, instead of square

plots, were selected to better account for the expected heterogeneity among the samples

(Krebs 1989).

The density and the height of all L. maackii individual stems were recorded in all

fifteen plots. If the area had been treated, the result of treatment was noted for each stem

as: missed during treatment, resprouted after treatment, or dead as a result of treatment.

Since it could not be determined whether an individual sprouted after treatment was

applied or was simply missed, all individuals not clipped were categorized as missed. As

volunteer labor was used to conduct the treatments, missed individuals were expected.

20

All woody plants were inventoried by species, crown class (either canopy, i.e.,

dominant, codominant, or intermediate; or overtopped), and it was noted whether the

individual was living or dead. If an individual was taller than 1.37 m, it’s diameter at breast height (DBH) was also recorded. Diameter at breast height is defined as the diameter of the trunk at 1.37 m.

Presence of ground vegetation or other feature was estimated at the quadrat level.

If a 1-m2 quadrat demonstrated a type of cover, it was tallied as present. Therefore, a

single quadrat could have more than one type of feature. Environment types were:

coarse woody debris, grass, sand, bare ground, rocks, moss, and vegetative cover.

Light levels were ocularly estimated at both the quadrat and plot level. Light

levels were estimated using these classes: <1%, 1-5%, 6-10%, 11-20%, 21-40%, 41-

70%, and 71-100%.

Nomenclature is based on Braun (1961). A Garmin Global Positioning System

unit (eTrex Vista HCx) was used to collect spatial data and to permanently locate the plot

corners.

Data Analyses

Comparisons between the three ravines will help support conclusions resulting

from testing the first hypothesis, and comparisons between the treated areas within Rush

Run and Adena Brook will help support conclusions resulting from testing the second

hypothesis.

21

Data was entered into EXCEL by quadrat and analyzed. L. maackii density,

height, and height distribution were compared among study locations and treatments.

Heights of living individuals were taken and broken down into four categories: <1 m, 1.1

– 2.0 m, 2.1 – 3.0 m, and > 3.1 m. L. maackii presence was compared between Flint Run

and untreated, disturbed areas (e.g., Adena Brook and Rush Run combined), as well as

between untreated areas of Adena Brook and Rush Run. Treated areas of Adena Brook

and Rush Run were compared as well.

The percent mortality of L. maackii was calculated for each treated plot. This

percentage was defined as the number of individuals dead due to treatment divided by the

total number of treated individuals in the plot. L. maackii reproduction is defined, for this

study, as individuals that were either missed during treatment or those that resprouted

after treatment was applied. Reproduction percentages were calculated as the number of missed and resprouted, divided by the total number treated.

The Shannon-Weiner Diversity Index (Shannon 1949) was calculated at both the

quadrat (1- by 1-m) level as well as at the plot (10- by 20-m) level. The Shannon-Weiner

Index was calculated using the formula:

H’ = - Σ pi(ln pi) with ί ranging from 1 to the number of plots,

where pί is the relative density (percentage) of a species of the ίth plot.

Species richness was also examined. Species richness is defined as the number of

species per plot.

In order to determine relationships among the independent variables, data was

grouped by plot (10- by 20-m), and by quadrat (1- by 1-m). Analysis at these scales

22

allowed for examination of spatial patterns. The dependent variables were: density of L.

maackii, height of L. maackii, Shannon-Weiner Diversity Index, and woody plant species

density. Density and height of L. maackii are expected to significantly decrease after

treatment. Shannon-Weiner Diversity Index is expected to decrease in response to a

difference in L. maackii density between the reference and untreated, disturbed study

locations. The woody plant species density is expected to be greater in treated areas of

Adena Brook than Rush Run due to the abundance of seeds in the seed bank compared to

the comparatively less dense planted areas.

ANOVAs were used to examine the null hypotheses, primarily to determine

whether differences among the study locations were random or if there was an effect of

the factor being tested. The comparisons were among these dependent variables: L.

maackii density between treated and untreated plots as well as between the reference

study location and the untreated, disturbed areas; L. maackii density (number of individual stems/m2) compared between treated and untreated plots; and light levels (%)

compared between the reference and the untreated, disturbed study locations, as well as

between treated and untreated. Mean height distribution of L. maackii between treated and untreated as well as reference compared to untreated, disturbed study locations;

Shannon-Weiner Diversity Index compared between the reference and the untreated, disturbed areas, and treated compared to untreated plots; and woody plant density

(number of individual stems/m2) compared between treated and untreated were analyzed as well. Finally, a mixed-effect ANOVA model (SPSS version 11.5 2002) was used to

23

examine the relationship between Shannon-Weiner Diversity Index and density of L. maackii.

24

CHAPTER 5

RESULTS

Density of L. maackii: reference area vs. untreated, disturbed areas

There were significant differences in L. maackii density between the reference area (Flint Run) and untreated, disturbed areas (Adena Brook and Rush Run) at the quadrat level (Table 3) (p < 0.001). Two percent of all L. maackii individuals sampled were found at Flint Run (23 individuals) compared with 35% at untreated Rush Run plots

(367 individuals) and 22% at untreated Adena Brook plots (225 individuals) (Figure 6 &

7; Table 4). Flint Run had the lowest density of L. maackii individuals, 0.04 per square meter, with untreated Rush Run plots having the highest levels of invasion with 0.61 individuals per square meter. Flint Run also had the lowest average density with 7.7 individuals per plot (10- by 20- m2), with all untreated plots having an average of 98.7

individuals per plot, and all treated plots having 70 individuals per plot (Figure 8).

Density of L. maackii: treated vs. untreated plots in disturbed areas

There was significantly greater density of L. maackii in the untreated plots of

Adena Brook and Rush Run compared with the treated plots of both (p < 0.001) (Figures

7 and 8, Table 5). Treated plots had 16% fewer L. maackii stems per square meter

25

compared with untreated plots (Table 4). Treated plots had an average density of 0.35

individuals per square meter with untreated having 0.49; this difference was not

significant at the plot level (p=0.31). There was no significant difference at the plot level

between the number of individuals in treated compared with untreated areas (p = 0.50).

Height Relationships of L. maackii: reference area vs. untreated, disturbed areas

Mean height of L. maackii was significantly less (p = 0.06) at the reference area

(0.38 m) compared with Adena Brook and Rush Run, 0.99 m and 1.54 m respectively

(Table 6). For all areas, the greatest proportion of L. maackii (48.2%) was < 1.0 m in height. For all areas, the average height was 0.97 m. Rush Run had 100 individuals with height > 3.0 m, while Flint Run had only one larger than 1.0 m tall. In the untreated,

disturbed areas, 47% of the individuals sampled were less than 1.0 m in height, whereas

96% were of that height in the reference area.

Height Relationships of L. maackii: treated vs. untreated plots in disturbed areas

Treated plots had a significantly lower (p = 0.01) average L. maackii height (0.54

m), while in untreated it was 1.57 m (Figure 9). The average height of individuals in

untreated plots within Adena Brook is 116.05 cm, while the height in treated plots is

60.93 cm (Figure 10 & 11). Rush Run untreated plots had an average height of 191.06 cm, and a height of 48.18 cm in treated plots.

There were equal percentages, 47%, of individuals in the less than 1.0 m category

in treated versus untreated plots. However, in treated plots, 46% of individuals sampled

26

were dead and the rest of the categories accounting for 13% of the individuals sampled.

In untreated plots, 20% of individuals sampled were in the 1.1 to 2.0 m and the greater

than 3.0 m height categories. Treated plots trended towards L. maackii being in the less than 1.0 m height class or dead, whereas in untreated plots it trended toward less than 1.0 m or relatively equal height distribution.

L. maackii Mortality and Reproduction

The percent mortality (Table 7) for treated L. maackii between Adena Brook and

Rush Run were significantly different (p = 0.001), while the amounts of reproduction

were not (p=0.905). Percent mortality of Adena Brook was 88.3%, and 69.6% in Rush

Run. Adena Brook had an average reproduction percentage of 55.51% while Rush Run

had a value of 52.21%. There was also more resprouted individuals at Rush Run (28

individuals) compared with Adena Brook (9 individuals), which contributed to the higher

reproduction rate (Table 8).

Environment Features and Light

Areas with high values of coarse woody debris, bare ground, sand, vegetative, and

grass have significantly higher woody plant species density (p = 0.03, 0.01, 0.03, 0.009,

0.012, respectively; Table 9) than areas with rocks or moss. Shannon-Weiner Diversity

Index values were significantly greater (p = 0.007) when there was bare ground present in

an area, compared to any other feature.

27

There were lower light percentages in Flint Run and in the untreated, disturbed areas compared with the treated areas (Table 10). Light levels averaged 83% in the

reference study location, 82% in untreated plots, and 70% in treated plots.

Hypothesis 1

When the data is considered at the plot (10- by 20-m) level, there was no

significant difference in Shannon-Weiner Diversity Index (p=0.05) values among areas

(reference and untreated, disturbed) and between treated and untreated plots (p= 0.40).

However, at the quadrat (1- by 1-m) level, there was a significant difference between

treated and untreated areas (p < 0.001) (Figure 12).

There was a higher density of L. maackii in untreated, disturbed areas as

compared with the reference area, 1012 and 23 individuals respectively. There is also a

greater number of individuals in untreated plots (592 individuals) compared with treated

plots (420 individuals) in disturbed areas. Finally, there was less woody plant species

diversity in the reference area compared with the disturbed and untreated plots (Table

11). The reference area had a species richness of 10 woody plant species (in a 600 m2

sample area). The disturbed areas’ species richness was 73 species (in a 2400 m2 sample area). Treated areas had similar species richness compared with untreated, 25 species compared with 22 species, respectively. There were no significant difference in species’ relative density among the three study locations (except for Ulmus americana [p=0.063]))

(Appendix A).

28

Hypothesis 2

There was a significant difference in woody plant species density between treated

plots at Adena Brook (1.1 stems/m2) as compared with Rush Run (0.52 stems/m2), and there was a higher number of stems at Adena Brook (Table 12) (p =0.03). When comparing the two disturbed areas, there was a greater density of woody plants of Adena

Brook as compared with Rush Run, both of which had lower woody plant density values than Flint Run (1.80 stems/m2) (Figure 13 & Table 13).

There was a trend towards lower reproduction percentage of L. maackii in Rush

Run (52.21%) treated plots compared with Adena Brook (55.5%) treated plots, although this result was not significant (p = 0.905) (Table 6). Rush Run treated plots had a higher average L. maackii density (0.41 individuals/m2) compared with Adena Brook treated

plots (0.29 individuals/m2), although this difference was not significant (p = 0.5) (Table

3). These mixed results demonstrate a possible trend toward lower reproduction and

higher woody plant species density following the seed bank regeneration method

compared with the woody plant species planting method.

29

CHAPTER 6

DISCUSSION

In central Ohio ravine forests, L. maackii treatment methods have been generally effective. There were fewer L. maackii individuals in treated plots compared with

untreated plots. The presence of L. maackii was related to woody plant diversity in

untreated plots; therefore, the first hypothesis was supported. Given that there were

significant differences apparently due to post-treatment management, it is expected that

treated ravine forests in central Ohio will trend towards the species composition of

similar but undisturbed ravine forests.

In general, L. maackii removal affects their density, height, and height

distribution. When treatment methods were applied, the number was 16% less, the

density was 14.5% less, height was 0.99 m shorter, and the height distribution contained more individuals <1.0 m tall. The reference area had the fewest invasions. The reference area had only 2.2% of all sampled L. maackii, and the other 97.8% was in the disturbed areas. This lends support to previous studies, e.g., Brothers and Spingarn (1992).

The work of the central Ohio community groups and the governmental agencies

that supported them appears to be restoring the two disturbed ravine forests to emulate

the reference area. It is expected that these forests will recover over a time period of 30

greater than five years, after the allelopathic effects of the invasion have diminished.

With L. maackii either removed from the area, or diminished in height, the native flora

will have an opportunity to reestablish at each area. This will allow an increase in shading

of the forest floor and will help to further suppress future L. maackii growth.

The mortality rate was significantly different in the treated plots of Adena Brook

compared with Rush Run. There are many possible reasons for this result: the volunteers who performed the treatments may have made errors; there may have an important environmental variable that was not measured; or the timing of treatments may be influential. In both study locations, volunteer labor performed the removal treatments. L. maackii individuals were missed in all treated plots. It can be assumed that these trained volunteers made some errors in identifying L. maackii individuals for treatment.

Therefore, it is likely that the volunteers did not follow the removal protocol strictly, or mistakes were made.

While the ravines were determined to be very similar to one another, it is possible

that other factors that were not measured impacted mortality rate. All plots were located

within a few meters of the stream, therefore, they have experienced periodic flooding.

No measurements were taken that are directly related to flood frequency, but it is possible

that a flood event might have caused some of the variability in mortality rate.

Finally, the time (season) of treatment is important, and may have resulted in

differences in mortality rate. L. maackii has a longer photoperiod than native plant

species; therefore the Adena Brook community chose to do most of their removal events

during the early spring or late autumn when the species is easier to identify. Friends of

31

Rush Run did their removal efforts during the late spring. The different treatment times may be affecting the mortality percentages in these ravines.

Hypothesis 1

The null hypothesis was accepted for the first hypothesis because in untreated plots there was a significant difference between L. maackii density and woody species diversity, as measured by the Shannon-Weiner Diversity Index. This result lends further support to the belief that disturbance is a major factor in invasion by L. maackii, as has been noted by many authors. This result should be understood in context, however.

Analysis of variance was conducted at the 200-m2 and 1- by 1- m2. The only significant difference was found at the 1- by 1- m2 level in untreated plots. The Shannon-Weiner

Diversity Index is often utilized; however, another index may have better approximated the diversity at these scales. Other diversity indices might also lend further clarity to the relationship between L. maackii density and woody plant species diversity.

Finally, the reference area was found to have lower woody plant species diversity than the untreated, disturbed areas, and the untreated plots were somewhat less diverse than the treated plots. The reference area was selected because it has remained relatively undisturbed for nearly 100 years. It is being maintained as a camp so the proprietors have been practicing little management. The species seen in the disturbed areas, however, are more varied and there are possible reasons for these observations. When the list of species found in the disturbed areas is examined for common native species, there are a few which do not belong. These species (i.e., Metasequoia glyptostroboides

32

[dawn redwood] in Adena Brook) may have been planted by a community member or

escaped from landscaping, thereby artificially increasing the diversity of the area.

Hypothesis 2

The seed bank regeneration strategy was a more effective post-treatment

management strategy than planting. Both strategies have the same fundamental causal

mechanism: where there is more shading, there will be, over time, decreased abundance

of L. maackii. In order to possibly demonstrate a significant difference in woody plant

density between areas, more data would be needed. There was a narrow area (<10 m)

where dense planting was attempted at Rush Run, but this area was not sampled because it was on the north slope of the ravine, the sample area would have been affected by the

grass area of the park, and the area was not subject to the same pressures because it is

heavily maintained by park personnel.

The analyses for testing this hypothesis were further confused by the lower

reproduction rate and the higher average density of L. maackii at Rush Run compared

with Adena Brook. If woody plant density is a determinant of post-treatment plot

management success, then a lower regeneration rate and density at Adena Brook than

Rush Run would be expected. Given the differences in significance levels among the 1- by 1- m2 and 200-m2 plot levels when examining the relationship between the Shannon-

Weiner Diversity Index and L. maackii density, further analyses may yield different

results. Regardless, more data collection and analysis must be done to elucidate the

33 causal mechanisms, and to determine the effectiveness of seed bank regeneration as a post-treatment management strategy.

34

CHAPTER 7

CONCLUSIONS

When individuals from the Adena Brook and Rush Run community groups sought

to eradicate L. maackii their primary objective was to return the ravine forests to pre-

European settlement conditions. They brought experts in to assess their plans and to assist with the treatments. They rallied community members and governmental entities around their causes and they began their treatments. They utilized the recommended

removal methods and performed post-treatment management strategies that they felt

would work well. Adena Brook has on-going removal activities while Rush Run does

not have another removal scheduled at this time. Both community groups took major

steps toward eradicating L. maackii from these urban, ravine forests, and their results have been dramatic. Their efforts must be considered in context, however. First, there is little information about ravine forest ecosystems during pre-European settlement times.

Secondly, the community groups are treating small portions of the ravine.

The intent of the Adena Brook Community and Friends of Rush Run are to help return the forest ecosystem to pre-European settlement conditions. After an extensive

search into scholarly literature and early surveys on riparian forest communities, Goebel

et al. (2003) determined that while broad forest types were known at the time of the 35

earliest settlers, more detailed understory or forest floor composition details were lacking.

This information is essential to provide reference conditions upon which to establish

forest restoration work and determine change over time. Riparian forests are challenging

areas to survey due to changing surface conditions, soil characteristics, geomorphology,

and disturbance regimes. Regardless, having an accurate understanding of forest

composition is a helpful template upon which to base the success of L. maackii removal.

This lack of information makes pre-European settlement conditions difficult to define, let

alone replicate. Therefore, Goebel et al. (2003) recommended that restoration efforts mimic reference sites that are located in remaining old-growth forests in the state.

Six plots were established per ravine. While plots were no closer than 10 m from

one another, they were all within 0.5 km of one another. These small areas had an

average density of 0.61 individuals per square meter in Rush Run and 0.38 individuals

per square meter in Adena Brook. Compared with the reference area, with a density of

0.038 individuals per square meter, these values are much higher. If left untreated, the L.

maackii densities in these ravine forests may increase. In Rush Run, treatment occurred

along the ravine in Park Boulevard Park only. The ravine stretches from the Olentangy

River to the intersection of the CSX railroad and State Route 161, where it appears to end

at a light industrial complex. This is a distance of approximately 3.4 km. The total

length sampled during this study was 5% of the total. Adena Brook stretches from the

Olentangy River east to the Overbrook Drive and Interstate 71, where it is bifurcated.

The total area sampled within Adena Brook was also approximately 5% of the total 2.5 km stretch of the ravine. The work that went into removing plants from those small

36

stretches was significant, but more must be done to remove L. maackii or these community groups will only be keeping up with the invasion and never eradicating the species.

Even though the abundance of L. maackii was significantly lower in treated plots

compared with untreated plots, the plant species was still present in all plots. L. maackii

seedlings may invade areas after removal efforts with either post-treatment management

strategy. Runkle et al. (2008) found L. maackii individuals that reinvaded treated areas

had shorter heights and lower densities. They also found that removal of L. maackii can

have a long-term effect; up to eight years of suppression after one removal event. L.

maackii, however, reaches sexual maturity in five to seven years (Deering and Vankat

1997); therefore, both ravine forests should be treated more frequently. Consistent

removal efforts could help the forest regenerate more successfully by thoroughly

suppressing the growth of future L. maackii individuals, as well as limiting the L. maackii

seed bank. Since L. maackii has an allelopathic effect on neighboring plants, sexual

maturity is not the only consideration. As noted in studies such as Collier et al. (2002)

and Trisel (1997), L. maackii may work to undermine regeneration of native woody plant

species in treated areas, which would shade out L. maackii individuals in the understory.

Runkle et al. (2008) also suggested this in regards to the native seed bank. If L. maackii

persists in an area, the authors argue, it will reduce the overall seed bank, which will lead

to slower regeneration by native species after its removal.

This study sought to support the efforts of local community groups by examining their work in the unique ravine forests of central Ohio. This study sought to determine

37

the effect of disturbance on L. maackii density and how seed bank regeneration will

impact regeneration in treated plots. There were significantly different Shannon-Weiner

Diversity Index values in the reference area compared to untreated, disturbed study

locations. There were significantly different woody plant species densities in treated

areas of Adena Brook, which practices seed bank suppression, compared with Rush Run

treated areas. More analysis on seed bank regeneration and planting strategies may yield

more clear results.

Future work may also include establishing experiments where light levels are related to L. maackii growth in urban ravine forests. If seed bank regeneration yields a greater density of woody plant seedlings, and there is a positive relationship between light levels and L. maackii abundance, we would expect that seed bank regeneration would be a suitable post-treatment strategy. In addition, future studies can continue sampling these plots over a longer time interval (> 2 years). If the vegetative releases documented by Runkle et al. (2008) were observed after eight years, a longitudinal study

would demonstrate how the forest rebounds after several years, which would help determine long-term urban, ravine forest trends after removal efforts.

These community groups’ efforts will eventually be thwarted if the entire ravine

is not treated. If they do nothing else to the treated plots, L. maackii will reinvade, although it may take longer than if no removal methods were attempted at all. Finally, this study was sufficient for attempting to address the two hypotheses that were proposed,

but more analyses will be needed of the post-treatment plots and removal strategies in the

future.

38

LITERATURE CITED

Anonymous. 2007a. About us. URL: http://adenabrook.org/mission.html/ Adena Brook Community. Date accessed: 16 November 2008.

Anonymous. 2007b. Education. URL: http://www.adenabrook.org/education.html Adena Brook Community. Date accessed: 16 November 2008.

Anonymous. 2008a. Custom soil resource report for Franklin County, Ohio: Lower Adena Brook. United States Department of Agriculture, Natural Resource Conservation Service.

Anonymous. 2008b. Custom soil resource report for Franklin County, Ohio: Rush Run, Worthington, Ohio. United States Department of Agriculture, Natural Resource Conservation Service.

Anonymous. 2008c. Custom soil resource report for Franklin County, Ohio. United States Department of Agriculture, Natural Resource Conservation Service.

Anonymous. 2008d. How we work. URL: http://adenabrook.org/howwework.html. Adena Brook Community. Date accessed: 25 November 2008.

Bartuszevige, A. M., D. L. Gorchov, L. Raab. 2006. The relative importance of landscape and community features in the invasion of an exotic shrub in a fragmented landscape. Ecography 29(2): 213-222.

Borgmann, K. L. and A. D. Rodewald. 2005. Forest restoration in urbanizing landscapes. Restoration Ecology 13(2): 334-340.

39

Braun, E. L. 1961. The woody plants of Ohio. Ohio State University Press: Columbus, Ohio.

Brothers, T. S. and A. Spingarn. 1991. Forest fragmentation and alien plant invasion of central Indiana old-growth forests. Conservation Biology 6(1): 91-100.

Collier, M. H., J.L. Vankat, and M. R. Hughes. 2002. Diminished plant richness and abundance below Lonicera maackii, an invasive shrub. American Midland Naturalist 147(1): 60-71.

Czarapata, E. 2005. Invasive plants of the upper midwest. University of Wisconsin Press, Madison.

Deering, R. H. and J. L. Vankat. 1999. Forest colonization and developmental growth of the invasive shrub Lonicera maackii. American Midland Naturalist 141(1): 43-50.

Friends of the Lower Olentangy Watershed. 2008. Rush Run Riparian Restoration Project. URL: http://www.olentangywatershed.org/RushRun/index.html. Friends of the Lower Olentangy Watershed. Date Accessed: 25 November 2008.

Friends of the Lower Olentangy Watershed. 2003. A snapshot: the state of the lower Olentangy River watershed in 2001. Friends of the Lower Olentangy Watershed, Columbus, OH.

Gayek, A. and M. F. Quigley. 2001. Does topography affect the colonization of Lonicera maackii and Ligustrum vulgare in a forested glen in southwestern Ohio? Ohio Journal of Science 101(5): 95 – 100.

Gill, D. P. and W. J. Mitsch. 2003. Response of the invasive shrub, Lonicera maackii (Rupr.), to removal efforts in a bottomland hardwood forest of central Ohio. The Olentangy River Wetland Research Park: Annual Report 03: 165 – 172.

Goebel, P. C., D. M. Hix, and M. E. Semko-Duncan. 2003. Identifying reference conditions for Riparian Areas of Ohio. Ohio Agricultural Research and Development Center Special Circular 192, Wooster, OH.

40

Gordon, Robert B. 1966. Natural vegetation of Ohio at the time of the earliest land surveys. Columbus (OH): Ohio Biological Survey. Map.

Hartman, K. M. and B. C. McCarthy. 2004. Restoration of a forest understory after the removal of an invasive shrub, Amur honeysuckle (Lonicera maackii). Restoration Ecology 12(2): 154-165.

Hartman, K. M. and B. C. McCarthy. 2007. A dendro-ecological study of forest overstorey productivity following the invasion of the non-indigenous shrub Lonicera maackii. Applied Vegetation Science 10(1): 3-14.

Hutchinson, T. F. and J. L. Vankat. 1997. Invasibility and effect of Amur honeysuckle in southwestern Ohio forests. Conservation Biology 11(5): 1117-1124.

Hutchinson, T. F. and J. L. Vankat. 1998. Landscape structure and spread of the exotic shrub Lonicera maackii (Amur honeysuckle) in southwestern Ohio forests. American Midland Naturalist 139(2): 383-390.

Johansson, M. E., C. Nilsson, and E. Nilsson. 1996. Do rivers function as corridors for plant dispersal? Journal of Vegetation Science 7(4): 593-598.

Krebs, C. J. 1989. Ecological methodology. Harper & Row Publishers, New York.

Luken, J. O. and D. T. Mattimiro. 1991. Habitat-specific resilience of the invasive shrub Amur honeysuckle (Lonicera maackii) during repeated clipping. Ecological Applications 1(1): 104-109.

Luken, J. O. and J. W. Thieret. 1995. Amur honeysuckle (Lonicera maackii; Caprifoliaceae): its ascent, decline, and fall. Sida 16(3): 479-503.

McLoda, N. A. and R. J. Parkinson. 1980. Soil survey of Franklin County, Ohio. United States Department of Agriculture, Soil Conservation Service, Washington, DC.

41

Miller, K. E. and D. L. Gorchov. 2004. The invasive shrub, Lonicera maackii, reduces growth and fecundity of perennial forest herbs. Oecologia 139(3): 359-375.

Natural Resources Conservation Service. 2008a. PLANTS profile: county distribution of Lonicera maackii (Rupr.) Herder - Amur honeysuckle in the state of Ohio. URL: http://plants.usda.gov/java/county?state_name=Ohio&statefips=39&symbol=LOMA6/ Date Accessed: 16 November 2008.

Natural Resources Conservation Service. 2008b. PLANTS profile: Lonicera maackii (Rupr.) Herder Amur honeysuckle. URL: http://plants.usda.gov/java/profile?symbol=LOMA6/ Date Accessed: 16 November 2008.

Ohio Department of Natural Areas and Preserves. 2006. Invasive plants of Ohio. URL: http://www.dnr.state.oh.us/tabid/2005/Default.aspx/ Date Accessed: 6 February 2008.

Ohio Department of Natural Resources. 2001. Invasive plants of Ohio: Amur, Morrow, and Tatarian honeysuckle. URL: http://www.dnr.state.oh.us/Portals/3/invasive/pdf/invasivefactsheet1.pdf/ Date Accessed: 3 February 2008.

Planty-Tabacchi, A.M., E. Tabacchi, R.J. Naiman, C. Deferrari, and H. Decamps. 1996. Invasibility of species-rich communities in riparian zones. Conservation Biology 10(2): 598-607.

Runkle, J. R., A. DiSalvo, Y. Graham-Gibson, and M. Dorning. 2007. Vegetation release eight years after removal of Lonicera maackii in west-central Ohio. Ohio Journal of Science 107(5): 125–129.

Schwartz, M. W. 1997. Defining indigenous species, pp. 7-17. In: J. O. Luken (ed.). Assessment and management of plant invasions. Springer-Verlag, New York.

Shannon, C. E. and W. Weaver. 1949. The mathematical theory of communication. University of Illinois Press, Urbana, IL.

42

Simpson, J. 1993. Rush Creek: Broadmeadows Park site analysis. Landscape Architecture 640 Course Report. The Ohio State University, Columbus, Ohio.

SPSS. 2002. SPSS version 11.5. Copyright SPSS Incorporated 1975-2008.

Trisel, D. E. 1997. The invasive shrub, Lonicera maackii (Rupr.) Herder (Caprifoliaceae): factors contributing to its success and its effect on native species. Ph.D. Dissertation, Miami University, Oxford, Ohio.

43

APPENDIX A

TABLES AND FIGURES

44

Flint Run Adena Brook Rush Run

Latitude 40°7.4' N 40°2.8' N 40°4.8' N

Longitude 83°1.3' W 83°0.8' W 83°0.9' W

Private and Ohio Department of Worthington Parks and Ownership Camp Mary Orton Natural Resources Recreation

Soil Series Alexandria silt loam Alexandria silt loam Ross silt loam

Slope 12-25% 12-25% 0 - 2 %

Table 1. Characteristics of the three study locations in central Ohio. All plots were located on south-facing slopes.

45

Species Name Common Name Asimina triloba (L.) paw paw Betula nigra (L.) river birch Platanus occidentalis (L.) sycamore Quercus macrocarpa (Michx.) bur oak Quercus bicolor (Willd.) swamp white oak Amelanchier canadensis (L.) Mendik. shadblow serviceberry trilobum (Marshall) highbush Lindera benzoin (L.) spicebush Cercis canadensis (L.) eastern red Cornus florida (L.) flowering dogwood

Table 2. Native woody plant species planted at Rush Run after treatment of L. maackii by volunteers.

46

Adena Brook & Adena Rush Flint Run Rush Run Flint Run Brook Flint Run Run

H0 H0 H0 F-test H1 H1 H1 Mean 0.04 0.49 0.04 0.37 0.04 0.61 Variance 0.10 0.69 0.10 0.60 0.10 0.76 df 599 1199 599 599 599 599 F 0.15 0.17 0.14 p <0.001 <0.001 <0.001

Table 3. Analysis of variance of L. maackii density among study locations and treatment types. The F-tests were conducted with alpha (α) = 0.05. Flint Run was compared with Adena Brook and Rush Run combined, as well as with each separately.

47

Total No. Density (No./m2) Mean (Avg. No./plot)

Flint Run Total 23 0.04 7.67

Adena Brook Untreated 225 0.38 75.00

Adena Brook Treated 174 0.29 58.00

Rush Run Untreated 367 0.61 122.33

Run Run Treated 246 0.41 82.00

Table 4. L. maackii sampled at each study location by treatment. The density was calculated as the total number sampled divided by the sample size (i.e., 600 m2 or 1200 m2). The mean number of individuals within each study area and treatment was calculated as the number of individuals sampled divided by the number of plots (i.e., 3 or 6).

48

Untreated plots Treated plots

F-test H0

H1 Mean 0.49 0.35 Variance 0.7 0.47 df 1199 1199 F 1.47 p < 0.001

Table 5. Analysis of variance of L. maackii density by treatment in Rush Run and Adena Brook. The F-tests were tested with alpha (a) = 0.05.

49

Flint Run Adena Brook Rush Run

Plot 123123456123456

Treatment No No No No No No Yes Yes Yes No No Yes Yes No Yes

Mean height (m) 0.4 0 0.2 0.6 0.9 2.6 0.6 0.8 0.2 2.1 1.6 0.4 0.3 2.1 0.7

<1.0 m 210 1 97407 115029385750274031

1.01 - 2.0 m 1 0 0 12 10 13 2 6 1 19 38 5 1 27 6

2.01 - 3.0 m000241701081320130

> 3.01 m0 0 0 0 31603 0301900513

Table 6. L. maackii height distributions by plot (number of individual stems sampled in each category).

50

Adena Brook Rush Run

Plot 45 6 346

Reproduction percent (%) 44.8 76.9 44.8 68.7 50.9 37.0

Average 55.5 52.2

Mortality percent (%) 94.1 78.3 92.5 65.0 77.1 66.7

Average 88.3 69.6

Table 7. L. maackii mortality and reproduction percentages. Reproduction percent was determined as the number of individuals missed or resprouted after treatment divided by the total number treated times 100. Mortality percent was calculated as the number of dead individual stems divided by the total number treated times 100.

51

Adena Brook Rush Run

Plot 456346

Missed (No. of individual stems) 12 55 27 43 20 6

Resprout (No. of individual stems) 15314834

Dead (No. of individual stems) 16 18 37 26 27 68

Table 8. L. maackii treatment summary. Within both disturbed study locations, three plots were sampled where treatment was previously applied. All individual stems within each plot were sampled and noted whether missed by treatment, resprouted after treatment, or died.

52

Type (No. of quadrats demonstrating type) Coarse Plot Treated Bare Vegetative woody Sand Grass Rocks Moss ground cover debris

F-test p = 0.03 p = 0.01 p = 0.03 p = 0.009 p = 0.012 p = 0.4 p = 0.001

1No20000 0 0 0 0

Flint Run 2 No 200 41 0 0 117 0 79

3No20080 0 015 9

1 No 200 148 0 4 0 32 1

2No199500 0 043 1

3 No 200 108 0 52 45 73 3 Adena Brook

4 Yes 200 145 0 13 69 52 17

5 Yes 198 80 0 186 0 40 0

6 Yes 164 194 0 113 38 121 39

1 No 200 167 0 168 0 9 22

2 No 200 111 0 113 0 0 0

3 Yes 72 43 0 161 0 0 0 Rush Run

4Yes183946117260 0

5 No 195 34 54 104 1 0 0

6 Yes 200 73 0 183 32 11 0

Average 187.40 86.40 7.67 74.27 21.87 26.40 11.40

Table 9. Environmental type noted by square-meter quadrat. Coarse woody debris was defined as primarily woody plant leaf and stem litter. Bare ground was defined as the soil surface being visible. Vegetative cover was not defined by species, but rather wherever there was forest floor flora that shaded the ground directly beneath it. F-tests were conducted comparing environmental type with L. maackii density.

53

Average by Mean light Plot Treated Average treatment level (%) No Yes

1No88 Flint Run 83 -- -- 2No72

3No88

1No73

2No90

3No90 Adena Brook 77 84 70 4Yes74

5Yes71

6Yes65

1No82

2No74

3Yes77 Rush Run 75 80 70 4Yes65

5No83

6Yes69

Table 10. Mean light levels by plot. Light levels were estimated ocularly for each square meter and placed into categories of <1%, 1-5%, 6-10%, 11-20%, 21-40%, 41-70%, and 71-100%. Average light levels by treatment and study location are also included.

54

Plots Treatment Shannon-Weiner

1 No 0.77

Flint Run 2 No 1.22

3 No 0.73

1 No 1.65

2 No 1.39

3 No 2.03 Adena Brook

4Yes1.95

5Yes1.85

6Yes1.91

1 No 1.00

2 No 1.65

3Yes1.27 Rush Run

4Yes1.25

5 No 1.83

6Yes1.97

Table 11. Shannon-Weiner Diversity Index values, based on species’ densities, by study location. Shannon-Weiner was calculated as: H’ = - Σ pί(ln pί). Pί was the relative density of a species in a plot. There is a significant difference between the Shannon- Weiner Diversity Index values and the density of L. maackii in untreated plots (p < 0.001).

55

Flint Run Adena Brook Rush Run

Treatment Total Total No Yes Total No Yes

No. species 10 25 20 22 17 17 14 Individual stems 1078 1302 404 898 621 211 410

Avg. no. stems/plot (200 m2) 359.33 217.00 134.67 299.33 103.50 70.33 136.67

Stand density (stems/ha) 17967 10850 6733 14967 5175 3517 6833

Table 12. Woody plant density and species richness by study location and treatment type. Density was calculated as the number of individual stems divided by the area sampled (i.e., 600 m2 or 1200 m2).

56

Stand density Plots Treatment (stems/ha)

1 No 1500

Flint Run 2 No 22650 3 No 29750

1 No 15200

2 No 3300

3 No 1650 Adena Brook 4 Yes 4350

5 Yes 16450 6 Yes 24250

1 No 1000

2 No 3000

3 Yes 4850 Rush Run 4 Yes 8100

5 No 6700

6 Yes 7600

Table 13. Woody plant density by plot (area of each was 200 m2).

57

Figure 1. Map of three study locations, which are located in Franklin County, central Ohio. Flint Run is a relatively undisturbed ravine and therefore was the reference site. Rush Run in Worthington, Ohio has been heavily disturbed due to surrounding residential development. Adena Brook is located in Clintonville, a neighborhood within Columbus, Ohio. Adena Brook has also been disturbed due to residential and commercial development.

58

Figure 2. The three 10- by 20-m plots sampled in Flint Run ravine are shown as dark rectangles within the outlined box.

59

Figure 3. The six Rush Run plots are represented as black rectangles within the box. Three of the plots were in areas where treatment was applied, and three where in areas where no treatment was applied.

60

Figure 4. Adena Brook sample plots are all located in the box near the center of the map. Dark rectangles represent the six plots (three treated and three untreated).

61

Figure 5. Sampling design. Each plot (10-by 20-m2) was established along the tributary with the long axis parallel to the forest edge to better account for heterogeneity. Plots were at least 10 m from the forest edge and had a minimum spacing of 10 m. Plots were positioned on south-facing slopes, within the Alexandria silt loam soil mapping unit, and were sampled at the square-meter level.

62

160

140

120 Reference

Untreated 100 Treated 80

60

No. of L. maackii 40

20

0 FR01 FR02 FR03 RR RR R R RR05RR A A AB03AB04AB0 A R R B B02 B 0 0 0 04 0 1 2 3 6 01 5 06

Figure 6. Number of L. maackii (stems) in each sample plot. Rush Run (RR) had the highest number, compared with Adena Brook (AB) and Flint Run (FR). Untreated plots in Adena Brook and Rush Run had significantly more individuals of L. maackii than the reference area (p < 0.001).

63

700

600 Reference

500 Untreated

Treated 400 L. maackii 300 No. of

200

100

0 Reference Untreated Treated

Condition

Figure 7. Total number of individual L. maackii stems sampled. Conditions at both Rush Run and Adena Brook had significantly higher numbers than the reference area (Flint Run) (p < 0.001). The treated plots had significantly lower numbers than the untreated plots (p < 0.001).

64

120

100 Reference

Untreate d

80 Treated

60

40 Mean number of individual stems

20

0 Reference Untreated Treated

Condition

Figure 8. Mean number of individual L. maackii stems by sample plot condition. The mean number of individuals of L. maackii was significantly greater in untreated plots compared with treated plots (p < 0.001), both of which were greater than the reference site (p < 0.001).

65

250

200

Untreated 150 Treated

100 Mean Height (cm)

50

0 Total Adena Brook Rush Run Flint Run

Sites

Figure 9. Mean height (cm) by area and treatment. L. maackii mean height was significantly less in treated areas as compared with untreated areas in disturbed study locations (p = 0.01). Flint Run included for visual reference, and was not included in the analysis.

66

160

140

120 Flint Run 100 Adena Brook 80 Rush Run

60 No. of individual stems 40

20

0 <1.0m 1.01 - 2.0 2.01 - 3.0 >3.01

Height Distribution (m)

Figure 10. Height distributions in untreated plots for all three areas. Without treatment, Rush Run had a relatively even distribution in the >1.0 m categories, while Adena Book had many more individuals in the <1.0 m category.

67

120

100

80

Adena Brook

60 Rush Run

40 No. of individual stems

20

0 <1.0m 1.01 - 2.0 2.01 - 3.0 >3.01

Height distribution (m)

Figure 11. Height distributions in treated plots for both areas. Both areas had many more individuals in the < 1.0 m category than in the larger height categories.

68

2.5

2

Reference

1.5 Untreated

Treated

1 Shannon-Weiner Diversity Index Diversity Shannon-Weiner 0.5

0 FR01 FR02 FR03 A A A AB0 AB05AB06R RR RR RR RR05RR06 B01 B B R 01 0 0 0 02 03 4 2 3 4

Plots

Figure 12. Shannon-Weiner Diversity Index values for woody plant species found in each sample plot. Flint Run (FR) plots generally had the lowest Shannon-Weiner Diversity Index values.

69

3.00

2.50 Adena Brook 2.00 )

2 Rush Run 1.50

Density (No./m 1.00

0.50

0.00 123

Treated plots

Figure 13. Woody plant species density in treated plots. Adena Brook treated plots had a mean density of 1.51 individual stems per square meter. Rush Run had a mean density of 0.69 individual stems per square meter.

70

APPENDIX B

WOODY PLANT SPECIES STAND DENSITY AND RELATIVE DENSITIES

71

Appendix B

Woody plant species stand density (stem/ha) and relative densities (number of stems of species/total number of stems * 100). All stems greater than 1.37 m in height.

Flint Run (reference) Adena Brook Untreated Adena Brook Treated Rush Run Untreated Rush Run Treated Relative Relative Relative Relative Relative density (No. density (No. density (No. density (No. density (No. Density stems of Density stems of Density stems of Density stems of Density stems of (Individual species/Total (Individual species/Total (Individual species/Total (Individual species/Total (Individual species/Total Species stems/ha) no. stems) stems/ha) no. stems) stems/ha) no. stems) stems/ha) no. stems) stems/ha) no. stems)

Acer negundo 0.00 0.00 266.67 4.52 50.00 0.97 200.00 2.54 616.67 9.92

Acer saccharum 616.67 42.53 650.00 11.02 333.33 6.47 366.67 4.66 416.67 6.70

Acer spicatum 0.00 0.00 0.00 0.00 16.67 0.32 0.00 0.00 0.00 0.00

Aesculus glabra 0.00 0.00 50.00 0.85 116.67 2.27 0.00 0.00 33.33 0.54

Asimina triloba 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00

Carya ovata 0.00 0.00 116.67 1.98 0.00 0.00 16.67 0.21 0.00 0.00

Cercis canadensis 0.00 0.00 50.00 0.85 166.67 3.24 50.00 0.64 33.33 0.54

Cornus florida 0.00 0.00 0.00 0.00 16.67 0.32 33.33 0.42 116.67 1.88

Crataegus disperma 0.00 0.00 16.67 0.28 0.00 0.00 0.00 0.00 0.00 0.00

Fagus grandifolia 100.00 6.90 0.00 0.00 0.00 0.00 16.67 0.21 0.00 0.00

Fraxinus americana 100.00 6.90 450.00 7.63 266.67 5.18 400.00 5.08 216.67 3.49

Fraxinus pennsylvanica 0.00 0.00 50.00 0.85 33.33 0.65 0.00 0.00 0.00 0.00

Fraxinus quadrangulata 0.00 0.00 83.33 1.41 300.00 5.83 0.00 0.00 0.00 0.00

Gleditsia triacanthos 0.00 0.00 33.33 0.56 0.00 0.00 16.67 0.21 0.00 0.00

Juglans nigra 0.00 0.00 16.67 0.28 100.00 1.94 83.33 1.06 116.67 1.88

Lonicera maackii 383.33 26.44 3750.00 63.56 2900.00 56.31 6116.67 77.75 4100.00 65.95

Liriodendron tulipifera 0.00 0.00 0.00 0.00 33.33 0.65 0.00 0.00 0.00 0.00

72

Flint Run (reference) Adena Brook Untreated Adena Brook Treated Rush Run Untreated Rush Run Treated Relative Relative Relative Relative Relative density (No. density (No. density (No. density (No. density (No. Density stems of Density stems of Density stems of Density stems of Density stems of (Individual species/Total (Individual species/Total (Individual species/Total (Individual species/Total (Individual species/Total Species stems/ha) no. stems) stems/ha) no. stems) stems/ha) no. stems) stems/ha) no. stems) stems/ha) no. stems) Metasequoia glyptostroboides 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00

Morus rubra 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 33.33 0.54

Ostrya virginiana 66.67 4.60 33.33 0.56 66.67 1.29 0.00 0.00 16.67 0.27

Pinus strobus 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00

Platanus occidentalis 0.00 0.00 0.00 0.00 50.00 0.97 233.33 2.97 0.00 0.00

Populus deltoides 0.00 0.00 0.00 0.00 0.00 0.00 116.67 1.48 0.00 0.00

Prunus serotina 0.00 0.00 0.00 0.00 0.00 0.00 16.67 0.21 0.00 0.00

Quercus alba 250.00 17.24 0.00 0.00 33.33 0.65 33.33 0.42 0.00 0.00

Quercus bicolor 0.00 0.00 0.00 0.00 83.33 1.62 0.00 0.00 0.00 0.00

Quercus palustris 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00

Sassafras albidum 200.00 13.79 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00

Staphylea trifolia 0.00 0.00 0.00 0.00 16.67 0.32 0.00 0.00 0.00 0.00

Tilia americana 0.00 0.00 16.67 0.28 16.67 0.32 0.00 0.00 0.00 0.00

Ulmus americana 0.00 0.00 0.00 0.00 266.67 5.18 50.00 0.64 416.67 6.70

Viburnum prunifolium 16.67 1.15 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00

Zanthoxylum americanum 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00

73