IMPACTS OF FOREST MANAGEMENT ON MEDICINAL HERBS

IN THE MISSOURI OZARKS

______

A Thesis

presented to

the Faculty of the Graduate School

at the University of Missouri-Columbia

______

In Partial Fulfillment

of the Requirements for the Degree

Master of Science

______

by

BADGER BALDWIN JOHNSON

Dr. Shibu Jose, Thesis Supervisor

MAY 2017

The undersigned, appointed by the dean of the Graduate School, have examined the thesis entitled

IMPACTS OF SILVICULTURAL MANAGEMENT ON MEDICINAL HERBS IN THE

MISSOURI OZARKS

presented by Badger Baldwin Johnson,

a candidate for the degree of Master of Science

and hereby certify that, in their opinion, it is worthy of acceptance.

______

Shibu Jose, Ph.D.

______

Rose-Marie Muzika, Ph.D.

______

James L. Chamberlain III, Ph.D.

______

Leszek Vincent, Ph.D.

______

Chung-Ho Lin, Ph.D.

DEDICATION I dedicate this research work to all well-intentioned foresters, farmers, scientists, activists, bureaucrats, politicians, teachers and students who help our society care for the earth. May we come together in the ways we need to and up our game in the 21st century.

Thank you Gaia, in your intelligent and loving embrace I move and live and have my being.

Thank you to my Gunga, and my mom Nancy Sullivan, who steadfastly supported my connection with

Gaia from a young age. Thank you to my dad Mac Johnson for loving me and for being a solid job coach these days. Beyond that, I feel so much gratitude to all the teachers who have particularly helped me develop along the way. The list is long, but starts with Eileen Frechette, all the staff at Edge of Appalachia

Reserve, Jack Muir Laws, Misha Golfman, Starhawk, Joe Hollis, Israel Regardie, Charles Griffin, Art

Trese, Frances Gander, John Schmieding, Rudy Nickens and Tim Jackins. Your mentorship has been invaluable.

Finally, thanks to my family of choice. Together we can do anything! ACKNOWLEDGEMENTS

Much gratitude goes out to my thesis committee. An especial thank you to Dr. Shibu Jose for believing in me, Drs. Rose-Marie and Jim Chamberlain for your careful attention to detail on the manuscripts, Dr. Leszek Vincent for your warm words and encouragement, and Dr. Chung-Ho Lin for inspiring me. An extra thanks is due to Shibu, Rose-Marie, Jim and Leszek, as well as to Ryan Dibala and

Michael Borucke for reviewing this manuscript. Each of you helped turn this from a few term papers into what will hopefully be published as a series of articles in peer-refereed journals. Additional thanks are owed to Dr. Mike Gold, Dr. Ben Knapp and Dr. John Kabrick for being there when I needed someone to extrovertly geek out with, or had a question they could efficiently answer.

To our primary collaborator at the Missouri Department of Conservation, Dr. Liz Olson, as well as Dr. Dawn Henderson. The generosity with your time, and your clarifying support has been indispensable. Cal Maginel, thank you for your ongoing help in getting my bearings in Missouri ecosystems. Thank you to Alicia Struckhoff for assisting with our adoption of the Ecosystem Site

Description typology. Thank you to Thomas Fieldan and Doug Ladd of The Nature Conservancy Missouri, your unreserved data sharing was a real boon to the project.

Thank you Elly Lang and Bo Young for your field support at HARC and Doug Allen’s!

Wow! And though I don’t know all of their names, thank you to all the people who were out there botanizing as seed tick bombs went off all around you in 100 ° F weather, at MOFEP and CCMA.

Piggybacking on your efforts allowed us at the Center for Agroforestry to do a lot more than we ever could have on our own. Thanks is due to Calvin Maginel, friend, Missouri native and botanist extraordinaire for helping me understand and adapt to the Ozark context of these forest farming projects. Finally, thanks to my comrades at Fretboard Coffee (Aba, Sarah, Dan, Austin et al.) for creating a supportive atmosphere for writing this thesis.

ii TABLE OF CONTENTS

ACKNOWLEDGEMENTS ...... ii

TABLE OF TABLES ...... v

TABLE OF FIGURES ...... ix

ABSTRACT ...... xiii

CHAPTER ONE: INTRODUCTION ...... 1

PROBLEM STATEMENT ...... 2

GOALS AND OBJECTIVES ...... 4

CHAPTER 2: GROWTH, REPRODUCTION AND PHOTOSYNTHESIS OF

FOUR HERBS IN CANOPY GAPS OF HARDWOOD FORESTS IN

CENTRAL MISSOURI ...... 8

INTRODUCTION ...... 9

MATERIALS AND METHODS ...... 11

RESULTS ...... 15

AVERAGE DAILY PAR AND NTFP PHYSIOLOGY ...... 21

SOIL FERTILITY AND PRECIPITATION DIFFERENCES BY SITE ...... 29

DISCUSSION ...... 30

ACKNOWLEDGEMENTS ...... 34

CHAPTER 3 IMPACTS OF TIMBER HARVEST ON MEDICINAL HERBS IN

THE MISSOURI OZARKS ...... 35

INTRODUCTION ...... 35

MATERIALS AND METHODS ...... 39 iii Study Area ...... 39

Statistical Analyses ...... 40

RESULTS ...... 44

DISCUSSION ...... 53

CHAPTER 4: EFFECTS OF PRESCRIBED FIRE ON MEDICINAL HERBS IN

THE MISSOURI OZARKS ...... 58

INTRODUCTION ...... 58

MATERIALS AND METHODS ...... 61

RESULTS ...... 65

DISCUSSION ...... 71

ACKNOWLEDGEMENTS ...... 77

CHAPTER 5: CONCLUSIONS ...... 78

LITERATURE CITED ...... 81

APPENDIX 1 ...... 98

APPENDIX 2 ...... 99

APPENDIX 3 ...... 102

iv TABLE OF TABLES

Table 1 Summary statistics (mean, standard deviation, range) and description of

explanatory variables in the forest farming study. SD=standard deviation.

Units of measurement indicated in brackets...... 17

Table 2 Results of ANOVA with for canopy gap treatment, by site. Significance value

P=0.05. Sample size n=80 for each site by treatment combination...... 21

Table 3 Accumulated precipitation in cm, annually and during the growing season,

for both sites and treatments...... 29

Table 4 Texture information for site as a whole. Eight soil cores were taken from each

site, one randomly from each plot, and mixed together before analysis...... 29

Table 5 Results of ANOVA of soil fertility samples, which may partially explain inter

site differences in NTFP maximum height, mortality and flowering. Mean at

each site as well as standard error and probability are included in the table.

Sample size n=64 at each site...... 30

Table 6 Summary statistics (mean, standard deviation, range) and description of

explanatory variables. SD=standard deviation. Units of measurement

indicated in brackets. Sample size (total quadrats) containing members of

each species changed by year, but over the course of the study there were 618

v

plots that had at least one quadrat with one study species recorded growing

in it...... 44

Table 7 Number of plots receiving each harvest type in 1996 and 2011...... 45

Table 8 Names and descriptions of ESDs, with number of plots classified as each ...... 46

Table 9 All harvest types, ESDs, years since harvest and interaction terms thereof

with significant (P < 0.05) relationships with percent cover or frequency of

Actaea racemosa...... 47

Table 10 All harvest types, ESDs, years since harvest and interaction terms thereof

with significant (P < 0.05) relationships with percent cover or frequency of

Apocynum cannabinum...... 47

Table 11 All harvest types, ESDs, years since harvest and interaction terms thereof

with significant (P < 0.05) relationships with percent cover or frequency of

Aristolochia serpentaria...... 48

Table 12 All harvest types, ESDs, years since harvest and interaction terms thereof

with significant (P < 0.05) relationships with percent cover or frequency of

Dioscorea quaternata...... 49

Table 13 All harvest types, ESDs, years since harvest and interaction terms thereof

with significant (P < 0.05) relationships with percent cover or frequency of

Echinacea simulata...... 49

Table 14 All harvest types, ESDs, years since harvest and interaction terms thereof

with significant (P < 0.05) relationships with percent cover or frequency of

Geranium maculatum...... 49

Table 15 All harvest types, ESDs, years since harvest and interaction terms thereof

with significant (P < 0.05) relationships with percent cover or frequency of

Hydrastis canadensis...... 50

vi

Table 16 All harvest types, ESDs, years since harvest and interaction terms thereof

with significant (P < 0.05) relationships with percent cover or frequency of

Parthenium integrifolium...... 50

Table 17 All harvest types, ESDs, years since harvest and interaction terms thereof

with significant (P < 0.05) relationships with percent cover or frequency of

Sanguinaria canadensis...... 50

Table 18 All harvest types, ESDs, years since harvest and interaction terms thereof

with significant (P < 0.05) relationships with percent cover or frequency of

Tephrosia virginiana...... 51

Table 19 Summary statistics (mean, standard deviation, range) and description of

explanatory variables in the CCMA study. SD=standard deviation. Units of

measurement indicated in brackets...... 65

Table 20 Guide to significant relationships between NTFPs and ESDs at CCMA.

Significant (P < 0.05) relationships between percent cover and frequency with

ESDs, number of years and number of fires. "NS" stands for "not

significant", and represents a relationship with a P-value= >0.1. "-" indicates

there were no individuals of that species found growing in plots assigned to

the corresponding ESD ...... 67

Table 21 Name of all ESDs at CCMA, with number of plots per ESD...... 68

Table 22 Mean percent cover of NTFPs by and frequency by ESD at CCMA by year.

This table is useful for interpreting the significant relationships between

NTFP percent cover and ESDs as summarized in Table 20. "-" indicates that

there were no individuals of that species found growing in plots assigned to

the corresponding ESD...... 69

vii

Table 23 Mean frequency of NTFPs by ESD at CCMA. This table is useful for

interpreting the significant relationships between NTFP frequency and ESDs

as summarized in Table 20. "-" indicates that there were no individuals of

that species found growing in plots assigned to the corresponding ESD...... 70

Table 24 List of type specimens assessed into the Dunn-Palmer Herbarium, since

relocated to the Missouri Botanical Garden...... 98

viii

TABLE OF FIGURES

Figure 1 Map of planting site at the Doug Allen Project Site (DAPS)and plots. Gap

treatment plots are in gold, shaded control plots are in blue. Orange line

reflects boundaries of soil map unit...... 13

Figure 2 Map of planting site at the Horticulture and Agroforestry Research Center

(HARC), with plot placement. Gap treatment plots are pictured as gold

polygons, shaded control plots as blue polygons...... 14

Figure 3 Bar graphs showing mean for both growing seasons of the maximum height,

as well as mean number of individuals out of 80 by site and treatment that

flowered, and that didn’t appear above soil for the duration of the growing

seasons (“mortality”) for Allium tricoccum...... 18

Figure 4 Bar graphs with mean for both growing seasons of maximum height, number

of individuals out of total 80 at each site and treatment that flowered or died

for Actaea racemosa...... 19

Figure 5 Figure 2. Bar graphs showing mean for both growing seasons of the

maximum height, as well as mean number of individuals out of 80 by site and

treatment that flowered, and that didn’t appear above soil for the duration of

the growing seasons (“mortality”) for Colinsonia canadensis...... 20

Figure 6 Bar graphs showing mean for both growing seasons of the maximum height,

as well as mean number of individuals out of 80 by site and treatment that

flowered, and that didn’t appear above soil for the duration of the growing

seasons (“mortality”) for Hydrastis canadensis...... 21

Figure 7 Mean daily PAR by month and hour, open field control reading at HARC...... 23

Figure 8 Daily PAR by month, hour and treatment at DAPS...... 24

ix Figure 9 Daily PAR by month, hour and treatment at HARC...... 25

Figure 10 A. racemosa light response curves at DAPS...... 26

Figure 11 A. racemosa light response curves at HARC...... 26

Figure 12 C. canadensis light saturation curves at DAPS...... 27

Figure 13 C. canadensis light response curves at HARC...... 27

Figure 14 H. canadensis light response curves at DAPS ...... 28

Figure 15 H. canadensis light response curves at HARC...... 28

Figure 16 Maps of plot design at MOFEP...... 43

Figure 17 Silvicultural systems employed at MOFEP, by compartment...... 43

Figure 18 Map of management unit at CCMA...... 63

Figure 19 Plot design at CCMA...... 63

Figure 20 Figure 9a Graph showing mean percent of Actaea racemosa, by year since

beginning of experiment, on significantly (P < 0.05) related ESDs. A.

racemosa cover had a positive relationship with two ESDs over time, and a

negative relationship with one ESD...... 73

Figure 21 Figure 9b. Graph showing mean percent of Tephrosia virginiana, by year

since beginning of experiment, on significantly (P < 0.05) related ESDs. T.

virginiana cover had a positive relationship with two ESDs over time and a

negative relationship with one ESD...... 74

Figure 22 Distribution of Geranium maculatum frequency at CCMA. For most species,

arcsin, gamma or Log10 transformations were applied to cover and

frequency values in an effort to satisfy the assumption of normality, but this

was not always successful...... 76

Figur’e 23 Difference (P < 0.05) in soil pH by site...... 99 x Figure 24 Bar graph showing significant difference (P < 0.05) in neutralizable acidity

(meq/100 g) by site...... 99

Figure 25 Bar graph showing significant difference (P < 0.05) in Ca (kg/hectare) by site...... 99

Figure 26 Bar graph showing significant difference (P < 0.05) in Mg (kg/hectare) by

site...... 100

Figure 27 Bar graph showing significant difference (P < 0.05) in Fe (ppm) by site...... 100

Figure 28 Bar graph showing significant difference (P < 0.05) in Mn (ppm) by site...... 100

Figure 29 Bar graph showing significant difference (P < 0.05) in NH4-N (ppm) by site...... 101

Figure 30 Bar graph showing significant difference (P < 0.05) in B (ppm) by site...... 101

Figure 31 Bar graph showing the mean percent cover and frequency values for Actaea

racemosa by harvest type at MOFEP...... 102

Figure 32 Bar graph showing the mean percent cover and frequency values for

Apocynum cannabinum by harvest type at MOFEP...... 103

Figure 33 Bar graph showing the mean percent cover and frequency values for

Aristolochia serpentaria by harvest type at MOFEP...... 103

Figure 34 Bar graph showing the mean percent cover and frequency values for

Dioscorea quaternata by harvest type at MOFEP...... 104

Figure 35 Bar graph showing the mean percent cover and frequency values for

Echinacea simulata by harvest type at MOFEP...... 104

Figure 36 Bar graph showing the mean percent cover and frequency values for

Geranium maculatum by harvest type at MOFEP...... 105

Figure 37 Bar graph showing the mean percent cover and frequency values for

Hydrastis canadensis by harvest type at MOFEP...... 105

xi Figure 38 Bar graph showing the mean percent cover and frequency values for

Parthenium integrifolium by harvest type at MOFEP...... 106

Figure 39 Bar graph showing the mean percent cover and frequency values for

Podophyllum peltatum by harvest type at MOFEP...... 107

Figure 40 Bar graph showing the mean percent cover and frequency values for

Sanguinaria canadensis by harvest type at MOFEP...... 107

Figure 41 Bar graph showing the mean percent cover and frequency values for

Tephrosia virginiana by harvest type at MOFEP...... 108

xii ABSTRACT As agroforestry becomes more accepted in the Midwestern US, understanding how non-timber forest product growth is impacted by forest management practice becomes increasingly important. Timber harvest and prescribed fires are common forest management practices in the Central Hardwoods Region, and canopy gaps commonly result from these practices. Three studies were conducted in central and southern Missouri to assess the impacts of forest management.

One of the studies focused on the effects of discrete canopy gaps on the height, reproduction and mortality of transplanted Actaea racemosa L., Allium tricoccum Aiton, Collinsonia canadensis L. and

Hydrastis canadensis L. This study was replicated at the Horticulture and Agroforestry Research Center in

New Franklin, MO as well as the Doug Allen Project Site in Gravois Mills, Missouri. Results indicate that small canopy gaps may prove beneficial for increasing rates of photosynthesis, as well as the height and sexual reproduction of the study species.

Using ground flora cover data from the Missouri Ozark Forest Ecosystem Project, a second study was conducted on the relationships (P < 0.05) between size and abundance of eleven herb species with ecological site types and the following timber harvest types: clearcutting, group selection, single-tree selection and intermediate thinning timber and no harvest (control). The goals were to ascertain whether ecological site descriptions can be used in conjunction with silvicultural management to select and manage sites for forest farming.Study species included Actaea racemosa L., Apocynum cannabinum L.,

Aristolochia serpentaria L., Dioscorea quaternata J.F. Gmel., Echinacea simulata R.L. McGregor,

Geranium maculatum L., Hydrastis canadensis L., Parthenium integrifolium L., Podophyllum peltatum L.,

Sanguinaria canadensis L. and Tephrosia virginiana (L.) Pers. Uneven-aged management harvest types were significantly positively related to the percent cover and/or frequency of A. racemosa, E. simulata, G. maculatum, H. canadensis, P. peltatum and S. canadensis. Percent covers and/or frequencies of A. racemosa, A. cannabinum and A. serpentaria were significantly , positively related to mesic ESDs only, while the percent covers and/or frequencies of D. quaternata, G. maculatum and P. integrifolium were significantly , positively related to specific mesic and specific xeric ESDs

xiii A third study was conducted using the ground flora cover data from The Nature Conservancy’s

Chilton Creek Management Area, to assess significant (P < 0.05) relationships between size and abundance of eight herb species with different ecological site descriptions under prescribed fire management. Goals for the study included identifying forest farming sites and planning prescribed fire management that, paired together, might stimulate growth of these plants. Study species included Actaea racemosa L., Apocynum cannabinum L., Aristolochia serpentaria L., Dioscorea quaternata J.F. Gmel., Geranium maculatum L.,

Parthenium integrifolium L., Sanguinaria canadensis L. and Tephrosia virginiana (L.) Pers. Percent cover values for the following species were positively related to time since beginning of experiment: A. racmosa,

A. serpentaria, A. cannabinum, D. quaternata and T. virginiana, and frequency of A. cannabinum and

D. quaternata was positively related to time since beginning of experiment. Representatives of these species significantly increased in percent cover over time and in the latter subset, had higher frequency over the course of prescribed fire management.

Small canopy gaps may be beneficial to non-timber forest product species in oak hickory forests.

Knowledge of how these species perform on particular ecological site descriptions can influence forest farming site selections. Additionally, decisions about which overstory management activities will be conducted, whether these be timber stand improvement activities involving canopy-gap creation, uneven- aged management timber harvest types, and prescribed fire, can be used to determine whether forest farming operations can be synergistic with the existing silvicultural management plan.

xiv CHAPTER ONE: INTRODUCTION According to the Association for Temperate Agroforestry (AFTA), agroforestry is “an intensive land management system that optimizes the benefits from the biological interactions created when trees and/or shrubs are deliberately combined with crops and/or livestock” (AFTA, 2015). Some goals of agroforestry are: increasing revenue generation for farmers (Nadeau et al., 1998); adding to the land equivalent ratio of the space so it can more efficiently meet the demands of a growing population (Graves et al., 2010), and enhancing ecosystem services and biodiversity agriculture can provide for (De

Beenhouwer et al., 2013).

Forest farming is an agroforestry practice in which crops are cultivated beneath the canopy of an existing forest (Chamberlain, 2013; Naud et al., 2010). The understory crops in a forest farming system are classified non-timber forest products (NTFPs). NTFPs “span a range of wild and semi- domesticated biological resources harvested by local households and communities from around homesteads, fields, grazing lands and relatively intact vegetation, such as grasslands, woodlands and forests” (Shackleton and

Pandey, 2014). In the US, common NTFPs include bee products, fuelwood, maple syrup, fence posts, edible and medicinal plants, craft material, fruits, nuts, and mushrooms (Collins, 2014).

Forest farming happens across a scale of management intensity. NTFP species may be introduced into the understory of a timber stand, or a forest stand can be modified to enhance the value of existing

NTFP crops (Chamberlain et al., 2009). Intentionally, intensively encouraging the proliferation of NTFP species for commercial harvest through compatible silvicultural treatments and/or prescribed fire management treatments (Lake and Long, 2014) already constitute an extensive form of forest farming. For example, NTFPs from the genus Vaccinium L. sp. respond favorably to thinning and low-intensity fires

(Vance et al., 2001). The understory is currently managed in these ways, in parts of the Pacific Northwest, for huckleberry production (Knowledge, 2014).

NTFP harvesting makes large contributions to the US economy. The retail value of commercial harvests of NTFPs from U.S. forest lands was conservatively estimated at $1.4 billion annually in 2011

(Alexander et al., 2011). More than one million six hundred thousand pounds of edible fruits, nuts, berries, and sap were permitted for harvest on USFS land in 2007, which is nearly double the quantity permitted for 1 harvest on federal public land in 1998 (Alexander et al., 2011). The American Herbal Products Association estimated the total average commercial harvest of wild medicinal plants from all US lands was

3,561,462 lbs annually, for years 1997-2005. (Alexander et al., 2011). While estimates of NTFP economic impact vary, there is evidently a sizeable demand for NTFPs. One study found that demand currently exceeded supply for agroecologically-produced NTFPs in Switzerland (Kilchling et al., 2009). If NTFP supply was increased through forest farming, this could contribute to meeting the rising demand for organic and local foods (Starr, 2010).

Most national forest ranger districts report NTFP harvest or “wildcrafting” occurring (McLain and

Pacific Northwest Research Station (Portland, Or.), 2005). Some wildcrafters have a strong conservationist ethic and have managed to safeguard the populations they depend on from overharvest (Price and

Kindscher, 2007), but a lack of information about appropriate methods and rates of harvest can prevent wildcrafting from being sustainably conducted (Hernández-Barrios et al., 2015). In recent decades there has been growing concern about unsustainable harvesting of rare medicinal herbs (Chamberlain et al., 2002;

Young et al., 2011). Forest farming could be an alternative to wildcrafting rare NTFPs from unmanaged populations (Burkhart, 2011a). In the United States, the transition from wildcrafting to forest farming may be hampered by lack of profitability due to labor costs. For example, according to one study, if a forest farmed NTFP with a tiny root sells for less than ~$20/lb it is unlikely to be worth the expenditure of time and money to plant and actively manage (Burkhart and Jacobson, 2009a). If additional, naturally-occurring

NTFP populations were being stewarded at the stand level through compatible management (Chamberlain et al., 2009), this could reduce the need for additional cost-prohibitive interventions.

PROBLEM STATEMENT According to one study, when Missouri farmers ask natural resource professionals about forest farming, they encounter knowledge gaps around best practices; this uncertainty about proper management is a barrier to forest farming adoption by risk-averse farmers (Valdivia and Poulos, 2009). There has been little academic research on forest farming practices in Missouri, although there was a study on the effects of wildcrafting and deer herbivory on Panax quinquefolius L. populations in the Ozarks (Farrington et al.,

2009a). Missouri is at or near the western edge of the native range of many NTFPs in this study, including 2

Actaea racemosa L. (Ranunculaceae), Allium tricoccum Aiton (Liliaceae),

Aristolochia serpentaria L. (Aristolochiaceae), Collinsonia canadensis L. (Lamiaceae), Echinacea simulata

R.L. McGregor (Asteraceae), Geranium maculatum L. (Geraniaceae), Podophyllum peltatum L.

(Berberidaceae), Sanguinaria canadensis L. (Papaveraceae). NTFP tolerance of harvest may vary with climatic conditions (Ticktin, 2004), and these species may respond differently to harvest and other disturbances than in other parts of their ranges.

Forest farmers would benefit from knowing how standard forest management techniques will affect their potential crop species. While habitat destruction and fragmentation can negatively affect NTFPs

(Farnsworth and Ogurcak, 2006), some anthropogenic disturbance may be compatible and can even promote NTFPs. For instance, some NTFP species respond positively to prescribed fire (Pilz et al., 2004).

The conservation status of certain populations of NTFP medicinal and aromatic plant species is threatened,

(Rao et al., 2004) has likely been negatively impacted by the diminished frequency and intensity of forest fires over the last 150 years (Van Sambeek et al., 1997).

In one study of the responses of NTFPs to silvicultural treatments in a mature Canadian boreal forest, stand thinning resulted in significant (P < 0.05) increases in abundance for six target NTFP species

(Clason et al., 2008a), and other studies have shown positive responses by NTFP species to thinning treatments as well (Diamond and Emery, 2011; Kerns et al., 2004).

Forest disturbances such as ice storms or derechos can create canopy gaps, and canopy gaps may impact NTFP species. Canopy gap formation represents the effect of many naturally occurring forest disturbances, as when trees die or are defoliated from natural causes. Canopy gaps can also be created in an uneven aged forest management system with “selection cuts”, where particular trees are harvested for timber while other trees are left to grow. The study in chapter three includes two types of selection cuts, group selection and single-tree selection (Table 6). Late successional, shade-tolerant understory herbs such as many NTFPs may persist in and around canopy gaps. Though some soil disturbance in and around canopy gaps can sometimes be beneficial to certain NTFPs (Sinclair and Catling, 2004), severe soil disturbance may result in the extirpation of rare understory species (Aikens et al., 2007).

3

Independent of soil disturbance, the size of a canopy gap affects understory species differently, depending on the size of the gap and the functional traits of the organism (Zenner et al., 2006). Large canopy gaps created by group selection timber harvests can increase understory herb species diversity at the stand level (Jenkins and Parker, 1999). Problematically, the increase in understory floristic alpha-diversity in canopy gaps may occur by dominance of ruderal and exotic species (Kern et al., 2014). However, changes in understory flora species richness in response to small canopy gaps created by single tree selection harvest have often been insignificant (Jenkins and Parker, 1999, 2000).

Abiotic site factors such as soil depth, parent material and soil series, slope position, steepness and aspect may also play a role in NTFP productivity. It has been shown that the woody-stemmed community composition in conjunction with slope position significantly (P < 0.05) affects on understory herb diversity

(Ellum et al., 2010). Aspect can be a dominant factor in determining species composition of forest understory vegetation communities (Jenkins and Parker, 1999), just as aspect is a determining factor in overstory species identity in the Ozarks (Villwock et al., 2011). New canopy gaps and ground disturbance created by timber harvest have been shown to combine with site factors, including soil type and slope position, to significantly (P < 0.05) affect ground flora diversity (Duguid et al., 2013). So, silvicultural treatments and site factors may be having an impact on NTFP understory herbs.

Prescribed fire is another common forest management tool that may be affecting NTFPs. In a study conducted in the mixed oak forests of southern Ohio, there was no significant interactive effect between prescribed burning and the integrated moisture index of a site (xeric, intermediate or mesic) on its understory species richness (Hutchinson et al., 2005). However such interactive effects might be different when examined at the species level (Elliot and Vose, 2010), and soil moisture may act as a habitat filter in fire-adapted plant communities (Kirkman et al., 2001) such as the Missouri Ozarks.

GOALS AND OBJECTIVES The overall goals for this thesis include identifying combinations of forest management and site types that, when paired with populations of in situ NTFP populations, result in these species growing larger and more numerous and becoming more abundant on the landscape. This knowledge can guide forest management decisions if forest farming or stewarding wild NTFP populations are goals of the land owner. 4

The goals of chapter two, “Growth, Reproduction and Photosynthesis of four Herbs in Small

Canopy Gaps of Hardwood Forests in Central Missouri, USA”, were to see how NTFPs in a forest farming system responded to canopy gaps in terms of size, mortality and reproduction. This study was a common garden experiment using four NTFP species. The study objective is to examine effects of canopy gaps and site on typical forest farming crops. Effects on size, reproduction and mortality of Actaea racemosa

L. (Ranunculaceae), Allium tricoccum Aiton (Liliaceae), Collinsonia canadensis L. (Lamiaceae) and

Hydrastis canadensis L. (Ranunculaceae), are explored.

Goals for chapter three included ascertaining combinations of logging and site type that are conducive to the proliferation of a subset of NTFP species. Titled “Effects of Timber Harvest on Medicinal

Herbs in the Missouri Ozarks”, chapter three included finding positive relationships between particular

NTFP species with specific timber harvest and site types, to find combinations of logging and is an analysis of ground flora cover data on 11 NTFP species growing wild at the Missouri Ozark Forest Ecosystem

Project. Relationships are explored using linear mixed-effects modeling between ground flora size

(“percent cover” within study plots), with ecological site descriptions and timber harvest types over time.

Study species include: Actaea racemosa L. (Ranunculaceae), Apocynum cannabinum L. (Apocynaceae),

Aristolochia serpentaria L. (Aristolochiaceae), Dioscorea quaternata J.F. Gmel. (Dioscoreaceae),

Echinacea simulata R.L. McGregor (Asteraceae), Geranium maculatum L. (Geraniaceae), Hydrastis canadensis L. (Ranunculaceae), Parthenium integrifolium L. (Asteraceae), Podophyllum peltatum

L. (Berberidaceae), Sanguinaria canadensis L. (Papaveraceae) or

Tephrosia virginiana (L.) Pers. (Fabaceae).

The goals of chapter four, “Effects of Prescribed Fire on Medicinal Herbs in the Missouri Ozarks”, were to ascertain whether fire management was related to increases or decreases in size and abundance of a subset of NTFP species, and whether site type mediated this overall trend. Analyses were conducted of ground flora percent cover data on eight NTFP species, found growing wild at The Nature Conservancy’s

Chilton Creek Management Area. Relationships were explored using linear mixed-effects modeling between study species percent cover, with ecological site descriptions and number of fires over time. Study species included: Actaea racemosa L. (Ranunculaceae), Apocynum cannabinum L. (Apocynaceae),

5 Aristolochia serpentaria L. (Aristolochiaceae), Dioscorea quaternata J.F. Gmel. (Dioscoreaceae),

Geranium maculatum L. (Geraniaceae), Parthenium integrifolium L. (Asteraceae),

Sanguinaria Canadensis L. (Papaveraceae) or Tephrosia virginiana (L.) Pers. (Fabaceae).

For chapters three and four, all data are sourced from the Missouri Department of Conservation and The Nature Conservancy employees and contractors. Because harvest treatments, burn treatments, biotic and abiotic site factors exist on a continuum, using a priori typologies and treating them as distinctive categories for an ANOVA is problematic. Multiple regression analyses were conducted using mixed general linear modeling equations in R. Because of unequal number of plots in treatment-level management units, analyses of frequency are nested within management unit, and analyses of percent cover were nested within site and plot. These nesting factors are coded as random effects.

The selection of NTFP species for these studies was based on several criteria. The primary criterion considered was whether the company American Botanicals was purchasing wildcrafted plant matter for the species in 2015. American Botanicals is a large Missouri-based company and one of the largest single purchasers and processors of wildcrafted NTFPs in the US. A lower price cutoff is set at

$1/lb, as paid to wildcrafters for wildcrafted material in 2015. Additionally, priority is given to whether the species had previously been discussed in the forest farming literature. For chapter two, Actaea racemosa L.,

Allium tricoccum Aiton and Collinsonia canadensis L. were selected partially on the basis of reported tolerance for full sun (Renaud, 2001; Thomas et al., 2010; Vasseur and Gagnon, 1994) and therefore potential compatibility with growing in canopy gaps. In addition, Hydrastis canadensis L. was selected because it is a high-value NTFP, has been found growing wild at DAPS, and was hypothesized to grow better in or around canopy gaps than under a closed canopy (Gillespie et al., 2006a). Study species for chapters three and four met the price criterion (worth >$1/lb for sale to American Botanicals), and occur within the study areas in quantities sufficient for the linear mixed effects models to converge.

The data analyzed in chapters three and four were mostly collected by researchers and technicians not directly associated with this thesis, over the course of almost twenty years, and as such it seemed prudent to verify the botanical identities of the study species at the site level. Plots at MOFEP and CCMA were visited, and voucher specimens were collected from near the plots and accessioned to the Dunn-

6 Palmer Herbarium (UMO), University of Missouri in August 2015. The UMO herbarium has since been transferred to the Missouri Botanic Garden Herbarium (MBG) in St. Louis. See Table 24 in APPENDIX 1 for list of accessions.

7 CHAPTER 2: GROWTH, REPRODUCTION AND PHOTOSYNTHESIS OF FOUR HERBS IN CANOPY GAPS OF HARDWOOD FORESTS IN CENTRAL MISSOURI

Abstract: Small anthropogenic canopy gaps are common in Missouri forests, but relationships between canopy gaps and understory medicinal herb size, mortality and reproduction are not well understood. To explore these relationships, a study was conducted in central Missouri with Actaea racemosa L., Allium tricoccum Aiton, Collinsonia canadensis L. and Hydrastis canadensis L. Transplanted rootstocks of each species were planted in small canopy gaps as well as under closed canopy at two sites,

DAPS and HARC. Height and flowering of individual plants were recorded weekly, April-September, for two years. Using ANOVA with a complete randomized design, relationships between canopy gap treatment and height, mortality and reproductive status were assessed (P < 0.05).

• Flowering and height of A. racemosa was significantly greater in canopy gaps at DAPS.

• A. tricoccum mortality was significantly greater in the undisturbed canopy control, at both sites. A.

tricoccum height was significantly greater in canopy gaps, at DAPS.

• C. canadensis height was significantly greater in canopy gaps than in the control, at both sites. C.

canadensis mortality was significantly greater in the control, at HARC.

• H. canadensis mortality was significantly greater in canopy gaps, at both sites.

As expected, mean hourly rates of photosynthetically active radiation (PAR) were higher in canopy gaps compared to closed canopy. Mean PAR at mid-day by month ranged from 175-400 µmol/m2/s in canopy gaps and was around 100 µmol/m2/s in the undisturbed canopy control. Photosynthesis was measured for A. racemosa, H. canadensis and C. canadensis midseason in all treatments, and the light saturation points were typically reached at 250-300 µmol/m2/s. Overall, mean peak mid-day PAR in canopy gaps met field- measured light saturation points of A. racemosa and H. canadensis, but not of C. canadensis. Small canopy gaps may prove beneficial for increasing rates of photosynthesis, as well as the height and sexual reproduction of the study species.

Keywords: non-timber forest product; canopy gaps; forest farming; uneven-age management;

Missouri River Hills; Ozark Highland Plateau. 8 INTRODUCTION Forest farming is the agroforestry practice of cultivating non-timber forest products species

(NTFPs) beneath a canopy of trees (Bernatchez et al., 2013). Understory forbs, including many NTFPs, are able to grow with relatively little light. Light in this case is measured as photosynthetic photon flux density,

µmol/m2/s in the 400 to 700 nm range, and is referred to as photosynthetically active radiation (PAR)).

Understory forbs plants utilize C3 carbon fixation and become light-saturated at low levels and photo- inhibited above species-specific thresholds (Jose et al., 2004a; Naud et al., 2010; Oláh and Masarovičová,

1997). Canopy closure mediates PAR at ground level, and as such the relationship between deciduous forest herb cover with basal area and density of trees is a significantly negative one(Gilliam and Turrill,

1993). Light can still be a limiting factor of growth for understory plants (Chazdon and Pearcy, 1991), so while stands under even-age management which have reached at least the understory re-initiation stage may provide opportunities for forest farming, light transmittance through the canopy may need to be increased if NTFPs are to reach their maximum growth rate (Gillespie et al., 2006b).

Small canopy gaps may not constitute a large enough disturbance to elicit an increase in growth rate by ground flora (Collins and Pickett, 1988). Among understory forbs in the Ozarks of Missouri, there is not always a significant difference in aboveground growth of herbs growing in uncut, single-tree selection, and thinned stands (Zenner et al., 2006). Because the overstory canopy intercepts most incoming solar radiation, the understory light environment is diffuse, often characterized by light levels of less than

50 µmol/m2/s , interspersed with higher intensity sun flecks (Sakai et al., 2005). Reflecting the low light regime to which understory herb species are adapted, these species can low light saturation points, ie the level of light at which a species cannot use additional photons to photosynthesize at a higher rate is relatively low. For example, the light saturation point for Panax quinquefolius L. was measured as

200 µmol/m2/s PAR when grown under shade cloth (Proctor, 1980), and as 500 µmol/m2/s PAR when grown under closed forest canopy (Wagner and McGraw, 2013). Such shade-adapted NTFPs may be light- limited in even-aged forests with unbroken canopies.

Selection cuts, which are typical of uneven-aged management, can mimic the structure of mature oak-hickory woodlands and forests through the creation of small to medium size canopy gaps (Goebel and

Hix, 1996). These canopy gaps enhance light transmittance to the forest floor, which may then support a 9 continuous layer of forbs (Hanberry et al., 2014a). Uneven-aged management has been growing in popularity over the past few decades (Dwyer et al., 2004a; O’Hara, 2002), due to the perceived advantages it has over even-age management, both social (Guldin et al., 2008) and ecological (Duguid and Ashton,

2013). Forest farming may provide a complementary land use to uneven-age timber management

(Chamberlain et al., 2013). For example, uneven-aged selection cuts were found to result in higher rates of survival for P. quinquefolius than populations growing in stands receiving even-aged treatments (Chandler and McGraw, 2015). In another study, the valuable NTFP medicinal herb Ligusticum porteri J.M. Coult. &

Rose was found to produce more flowers per leaf when growing in open areas than under a closed canopy

(Mooney et al., 2015). However, if selection cuts increase the PAR at the ground level above certain thresholds, this may favor other understory plants at the expense of NTFPs; increases in light are known to shift the relative competitive hierarchy of multispecies assemblages of forbs (Fortner and Weltzin, 2007).

PAR is primarily controlled by canopy closure rather than stand density or basal area (Blizzard et al.,

2013). Forest farmers seeking to optimize the understory light environment for NTFP would benefit from knowing whether small canopy gaps are positively or negatively related to NTFP growth, survival and reproduction.

Forest managers seeking to enhance the value of both timber and NTFPs need to know what stage of the silvicultural management cycle is conducive to NTFP cultivation. If NTFP species grow faster under closed canopies, the best window for cultivation would be in the years immediately prior to even-aged regeneration cuts, when the canopy is closed. However, if NTFP species grow faster and reproduce more under a semi-open canopy, integrating forest farming into uneven-aged management may be ideal.

During the growing season (April-September) of 2015 and 2016, mortality status, height and reproductive status were measured weekly for all individual plants. The rationale for measuring height and reproductive status is that the faster each plant grows, the faster it can be harvested and marketed. If these species produce seed more readily in canopy gaps, this could be exploited to supply planting material to sustain production in a forest, even with the destructive harvest of parent plants. Sourcing future planting material from a current planting could defray costs of initiating forest farming plots, which can otherwise be cost-prohibitive (Burkhart and Jacobson, 2009a).

10 It was hypothesized that height and flowering of the study species would be greater in the canopy gap treatment compared to the closed canopy. Additionally, it was hypothesized that mean PAR measured under canopy gaps will meet or exceed the light saturation points of A. racemosa, C. canadensis and

H. canadensis, but mean PAR measured under closed canopy will not.

MATERIALS AND METHODS In the winter of 2014-2015, forest farming planting areas were established across the slope’s contour north-facing backslopes of approximately 30° at two sites. Planting site one was established at the

Doug Allen Project Site (DAPS) in Gravois Mills, MO (38° 14' 25'' N, 92° 48' 0'' W), and matches the ecological site description of “Chert Protected Backslope Forest” (Missouri Department of Conservation,

2016a) on soil classified as series Goss, a very gravelly, very stony silt loam on 15 to 35 percent slopes.

Planting site two was established at the University of Missouri Horticulture and Agroforestry Research

Center (HARC) in New Franklin, MO (39° 1' 8'' N, 92° 45' 41'' W), and matches the ecological site description of “Deep Loess Protected Backslope Forest” (Missouri Department of Conservation, 2016b) on series Menfro, an eroded silt loam on 14 to 35 percent slopes. Four small canopy gaps and four adjacent closed canopy areas were selected at each site, For the canopy gap treatment, overstory trees were removed to create four circular canopy openings of approximately 10.5 m in diameter, with a 12 m uncut buffer in between gaps. Within the gaps, all woody stemmed plants with a height >2 meters were cut down at the level of the forest floor. For the control, no trees were removed. Plots were prepared using a BCS tractor rotary plow attachment see Figure 1 and Figure 2 for site maps with plot locations. Trenches were dug approximately 20 cm deep on the contour of the hill (swales), a practice which can increase rainwater infiltration to crop species (Yuen et al., 2001). The rotary plowing brought rocks to the surface of the soil which were discarded outside of growing area to help increase uniformity of root growing space at all planting locations, as well as reducing competition for moisture between fine roots from adjacent trees and

NTFPs (Grubb, 1998).

Plant location was randomized within plot, and within the restrictions of the site the plot location was randomized at the site level. Eighty plants of each species were planted in each of eight plots per site for a total of 1,280 plants. In all plots understory shrubs and leaf litter were mechanically removed in the 11 fall of 2014, immediately before planting. Humus layer mass can have a strong negative relationship with herb cover and diversity (Vockenhuber et al., 2011), possibly by acting as a mulch and preventing adequate seed to soil contact for germination and recruitment. A. tricoccum and H. canadensis were planted 2 cm below the soil surface; C. canadensis and A. racemosa were planted 4 cm deep with a mean 64 cm left between individual plants.

Figure 1

12 Figure 1 Map of planting site at the Doug Allen Project Site (DAPS)and plots. Gap treatment plots are in gold, shaded control plots are in blue. Orange line reflects boundaries of soil map unit.

Figure 2 13 Figure 2 Map of planting site at the Horticulture and Agroforestry Research Center (HARC), with plot placement. Gap treatment plots are pictured as gold polygons, shaded control plots as blue polygons.

To characterize differences in PAR between control and canopy gap treatment, PAR was measured continuously using HOBO data loggers connected with three PAR sensors each. A mean hourly

PAR measurement was calculated from the values from all PAR sensors. Every week, the data loggers were moved to one of four stations spaced evenly across each treatment area, to help account for possible effects from sunflecks at any single location. There was also a PAR sensor in an open field at HARC, to have full sun measurements of µmol/m2/s PAR measurements for comparison. Photosynthesis was measured mid-season using a Licor-6400 Portable Photosynthesis System on one to four individuals of each species, except A. tricoccum (due to unavailability of instrument during early spring growing season), in each site by treatment combination each year.

14 To explain potential differences in plant response by planting site, soil samples were taken in the spring of 2015 and analyzed by the University of Missouri Soil and Plant Testing Laboratory. In total, 32

15-cm soil cores were taken from each site and treatment combination. These samples were dried, homogenized and assayed for soil organic matter and the following nutrient concentrations: pH, NA

(neutralizable acidity) (meq/100g), Bray IP (kg/hectare), Ca (kg/hectare), Mg (kg/hectare), K (kg/hectare),

Na (kg/hectare), Zn , Fe, Mn, Cu, S, NO3-N, NH4-N and B (ppm). Soil texture was measured at the site level using one sample from each plot, as per the soil testing lab’s suggestion that this would be sufficient.

Estimated accumulated precipitation data was obtained at the site level from the U.S. Climate Data website

(http://www.usclimatedata.com/). Precipitation data was available for New Franklin, MO but not Gravois

Mills, MO, so data collected 16 km away at Lake Ozark, MO was used.

Plant height, flowering and mortality were assessed every week during the growing season (April-

September) for 2015 and 2016, and these data was used to evaluate the productivity of each NTFP species.

Complete randomized design analysis of variance (ANOVA) was conducted in RStudio. The maximum height reached during the growing seasons was used for analysis. The variable “mortality” encompassed individuals that did not emerge from the ground , as well as individuals that were growing and wilted or were browsed to the ground and didn’t regrow during the growing season. The variable “flowering” described whether an individual plant flowered during the course of the growing season. The maximum height of each plant measured in a season was used for analysis. Height data was transformed for normality, using a logarithm of base 10, for A. racemosa, A. tricoccum and H. canadensis (for C. canadensis height was already normally distributed. For all analyses in this study, alpha=0.05.

RESULTS Basic summary statistics for all soil nutrients, as well as initial weight, height, flowering, mortality of the study species are presented (

15 Table 1 Summary statistics (mean, standard deviation, range) and description of explanatory variables in the forest farming study. SD=standard deviation. Units of measurement indicated in brackets.).

Mean, standard deviation, minimum and maximum values are included, except for the population percentages of mortality and flowering.

16 0.00 84.00 20.00 15.00 75.00 39.50 34.50 84.50 329.30 100.00 100.00 100.00 100.00 277.00 179.00 204.50 Maximum 0.18 2.30 0.00 0.00 5.00 0.00 0.00 0.00 0.00 0.00 9.00 2.00 1.50 6.70 0.50 23.00 =standard Minimum SD 6.38 45.86 0.00 8.81 6.84 7.84 7.38 5.39 55.33 45.79 39.91 33.27 28.73 49.11 21.91 31.66 6.60 0.00 0.78 0.47 9.06 6.62 5.48 0.652.56 4.30 0.55 1.90 6.40 3.80 1.06 0.400.49 0.705.01 0.125.46 1.80 1.485.94 0.27 3.220.51 2.50 1.42 0.80 0.74 0.15 3.80 7.50 0.17 14.90 8.60 0.71 Mean 78.03 18.98 84.56 20.34 48.90 62.06 30.00 29.84 19.84 12.66 36.0333.02 9.84 17.70 20.00 11.90 55.50 67.60 306.25 110.47211.19 198.00 33.60358.20 358.20 148.00 607.00 358.20 272.00 358.20 2126.25 911.41 672.00 3944.00 , n=320. , , n=320. , n=320. , n=320. , A. raceomsa A. canadensis C. canadensis H. A. tricoccum, tricoccum, A. individuals, n=320. individuals, individuals, n=320. individuals, individuals, n=320. individuals, individuals, n=320. individuals, individuals, n=320. individuals, individuals, n=320. individuals, individuals, n=320. individuals, n=320. individuals, A. racemosa A. A. racemosa A. collinsonia C. canadensis H. A. racemosa A. racemosa A. collinsonia C. canadensis H. maximum height (cm) of height maximum (cm) of height maximum (cm) of height maximum (cm) of height maximum

Description (g) rhizome at time of planting, , n=320. , time ofat rhizome planting, (g) n=320. time ofat planting, bulb (g) A. racemosa A. tricoccum A. Mean plot-level soil magnesium (kg/hectare) between all plots/sites/treatments, all 2015.(kg/hectare) between plot-levelspring soil in Mean magnesium measured plots/sites/treatments, all plot-level soil 2015.(kg/hectare) between Mean potassium spring in measured plots/sites/treatments, all plot-level soil 2015.(kg/hectare) between Mean sodium spring in measured plots/sites/treatments, all 2015. between plot-levelspring soil (ppm) in Mean zinc measured Individual mean for plant height forseasons forboth height forthe plant mean Individual forforseasons forboth leaflength mean the Individual forseasons forboth height forthe plant mean Individual forseasons forboth height forthe plant mean Individual for of mass initial mean Individual for of mass initial mean Individual n=320. time ofat forrhizome planting, (g) mass initial mean Individual n=320. time ofat forrhizome planting, (g) mass initial mean Individual forPlotpercent seasons/sites/treatmentsflowering all mean of forPlotpercent seasons/sites/treatmentsflowering all mean forPlotpercent seasons/sites/treatmentsflowering all mean forPlotpercent seasons/sites/treatmentsflowering all mean Plotpercent mortalityfor seasons/sites/treatmentsmean all Plotpercent mortalityfor seasons/sites/treatmentsmean all Plotpercent mortalityfor seasons/sites/treatmentsmean all Plotpercent mortalityfor seasons/sites/treatmentsmean all plots/sites/treatments, all of2015.measure soilbetween a spring acidity, in plot-levelmeasured pH, Mean plots/sites/treatments, all (meq/100 2015. acidity between g) spring in plot-level neutralizable measured Mean plots/sites/treatments, all plot-levelsoil 2015. between matterpercentage Mean spring organic in measured plots/sites/treatments, all plot-level soil 2015.(kg/hectare) between Mean phosphorous spring in measured 2.97plots/sites/treatments, plot-levelall soil 2015.(kg/hectare) Mean between calcium spring in measured 1.97 36.75 0.50 20.83plots/sites/treatments, all plot-level2015.soil between spring Mean iron (ppm) in measured 7.00 7.00plots/sites/treatments, all 2015. between spring (ppm) in measured plot-level soil Mean manganese plots/sites/treatments, all plot-level2015.soil between Mean spring copper (ppm) in measured 81.00 plots/sites/treatments, all 2015.plot-level between soil spring Mean (ppm) in sulfur measured plots/sites/treatments, all 2015.plot-level soil between spring Mean nitrate(ppm) in measured plots/sites/treatments, all 2015. between spring (ppm) in plot-levelmeasured soil Mean ammonium plots/sites/treatments, all 2015.plot-level soilbetween spring Mean (ppm) boronin measured Summary Summary statistics (mean, standard deviation, andrange) description explanatoryof variables the in forest farming study. SD deviation. Units of measurement indicated in brackets. . canadensis . A. racemosa A. tricoccum A. canadensis C. canadensis H. A. racemosa A. tricoccum A. canadensis C. canadensis H. H. canadensis H. A. racemosa A. tricoccum A.

A. racemosa A. tricoccum A. canadensis C. canadensis H. VariableName 1 Mg K Na Zn Flowering of Flowering of Flowering Mortalityof Mortalityof Mortalityof Mortalityof pH Initial mass mass Initial mass Initial C mass Initial mass Initial of Flowering of Flowering O.M.% Ca Height Height Height Height Height N.A. P I Bray Fe Mn Cu S NO3-N NH4-N B Table

17

HEIGHT, FLOWERING AND MORTALITY

Flowering and height of A. racemosa was significantly (P < 0.0001) greater in canopy gaps than at the undisturbed-canopy control at DAPS (Figure 3 and Table 2). A. tricoccum mortality was significantly greater in the control at DAPS (P=0.0001) and at HARC (P < 0.0001), while height was significantly

(P < 0.0001) greater in canopy gaps at DAPS (Figure 4 and Table 2). C. canadensis height was significantly (P < 0.0001) higher at both sites in canopy gaps (Figure 5and Table 2). H. canadensis mortality was significantly (P=0.0337) greater in canopy gaps at HARC (Figure 6 and Table 2).

Figure 3

Figure 3 Bar graphs showing mean for both growing seasons of the maximum height, as well as mean number of individuals out of 80 by site and treatment that flowered, and that didn’t appear above soil for the duration of the growing seasons (“mortality”) for Allium tricoccum.

18

Figure 4

Figure 4 Bar graphs with mean for both growing seasons of maximum height, number of individuals out of total 80 at each site and treatment that flowered or died for Actaea racemosa.

19 Figure 5

Figure 5 Figure 2. Bar graphs showing mean for both growing seasons of the maximum height, as well as mean number of individuals out of 80 by site and treatment that flowered, and that didn’t appear above soil for the duration of the growing seasons (“mortality”) for Colinsonia canadensis.

20 Figure 6

Figure 6 Bar graphs showing mean for both growing seasons of the maximum height, as well as mean number of individuals out of 80 by site and treatment that flowered, and that didn’t appear above soil for the duration of the growing seasons (“mortality”) for Hydrastis canadensis.

Table 2 Results of ANOVA with for canopy gap treatment, by site. Significance value P=0.05. Sample size n=80 for each site by treatment combination. Actaea racemosa Allium tricoccum Collinsonia canadensis Hydrastis canadensis DAPS HARC DAPS HARC DAPS HARC DAPS HARC Response St. Error Pr(>|t|) St. Error Pr(>|t|) St. Error Pr(>|t|) St. Error Pr(>|t|) St. Error Pr(>|t|) St. Error Pr(>|t|) St. Error Pr(>|t|) St. Error Pr(>|t|) Flowering 0.0462 <.0001 0.0529 0.5550 0 NA 0 NA 0.0547 0.0682 0.0446 0.8890 0.0403 0.2790 0.0481 0.4360 Height 0.0283 <.0001 0.0264 0.3490 0.7642 0.3490 1.0693 <.0001 2.5870 <.0001 1.8020 <.0001 0.8678 0.0337 0.8311 0.3750 Mortality 0.0139 0.1774 0 NA 0.0201 0.0007 0.0448 <.0001 0.0108 0.0823 0 NA 0.0377 0.0984 0.0243 0.0403

AVERAGE DAILY PAR AND NTFP PHYSIOLOGY For comparison, the hourly PAR values by day, for each month of the growing season are shown from an open field adjacent to the study site at HARC (Figure 7), as well as at both growing sites from within and without the canopy gaps at DAPS (Figure 8) and HARC (Figure 9). In the open field, mean mid- day PAR peaked at 1000-1500 µmol/m2/s, depending on the month. Mean daily peak PAR at both DAPS and HARC gap treatments decreased substantially after May, coinciding with the overstory canopy leafing

21 out. After May, mean PAR at DAPS peaked at around 400 µmol/m2/s in the areas receiving the canopy gap treatment, and around 100 µmol/m2/s PAR in the control. At HARC, mean peak light was 200 PAR in the areas receiving the canopy gap treatment, 100 µmol/m2/s PAR in the control.

Mean peak daytime PAR measured in canopy gaps at DAPS met or exceeded the field-tested light saturation points of A. racemosa and H. canadensis but not of C. canadensis. At DAPS, A. racemosa light saturation point was reached for plants at around 250 µmol/m2/s. At HARC, A. racemosa light saturation point was reached for some but not all plants at around 250 µmol/m2/s. At DAPS, C. canadensis light saturation point was reached for first-year plants at around 500 µmol/m2/s, and for second growing year the light saturation point was reached at approximately 250 µmol/m2/s. C. canadensis light saturation points at

HARC was reached by most plants at 750-1000 µmol/m2/s. At DAPS, H. canadensis light saturation points were reached for most plants at 250 µmol/m2/s. At HARC H. canadensis light saturation points were between 300 and 500 µmol/m2/s.

Unlike in the canopy gaps at DAPS, the mean peak daytime PAR for both treatments at HARC and the closed canopy control at DAPS fell short of the light saturation point of all three species measured.

The light response curves are presented for A. racemosa in Figure 10 and Figure 11, for C. Canadensis

Figure 12 and Figure 13, and for H. canadensis in Figure 14 and Figure 15.

22

Figure 7

23

Figure 7 Mean daily PAR by month and hour, open field control reading at HARC.

Daily PARDaily by month, hour and at DAPS. treatment

8 gure 8 Figure Figure Fi

24

Daily PARDaily by month, hour and at HARC. treatment

9 Figure Figure Figure Figure 9

25 Figure 10

Figure 10 A. racemosa light response curves at DAPS.

Figure 11

Figure 11 A. racemosa light response curves at HARC.

26

Figure 12

Figure 12 C. canadensis light saturation curves at DAPS.

Figure 13

Figure 13 C. canadensis light response curves at HARC.

27

Figure 14

Figure 14 H. canadensis light response curves at DAPS

Figure 15

Figure 15 H. canadensis light response curves at HARC.

28 SOIL FERTILITY AND PRECIPITATION DIFFERENCES BY SITE In order to examine if precipitation or soil fertility might be contributing to differences in height, flowering and mortality at different sites, precipitation and soil fertility information was examined. The accumulated precipitation during the growing season was similar at both sites in both years (Table 3). Soil at DAPS was gravelly, and had an average of 10% higher sand content than at HARC. The higher gravel and sand content probably means that DAPS soils are better drained than soil at HARC (Table 4).

ANOVAs were conducted on soil nutrient concentrations (Table 5). At HARC, mean soil pH, Ca, Mg, Fe and B were significantly higher, while at DAPS, mean neutralizable acidity (N.A.), Mn, and NH4-H were significantly higher (Error! Reference source not found., Error! Reference source not found., Error!

Reference source not found., Error! Reference source not found., Error! Reference source not found., Error! Reference source not found., Error! Reference source not found.,Error! Reference source not found.). The soils at both sites were silt loams.

Table 3 Accumulated precipitation in cm, annually and during the growing season, for both sites and treatments. Precipitation (cm) Site Year Apr.-Sep. Annual DAPS 2015 74.4 133.8 HARC 2015 75.2 120.2 DAPS 2016 80.2 89.0 HARC 2016 76.0 91.0

Table 4 Texture information for site as a whole. Eight soil cores were taken from each site, one randomly from each plot, and mixed together before analysis. Site Sand % Silt % Clay % Texture DAPS 20 65 15 Silt loam HARC 10 72.5 17.5 Silt loam

29

Table 5 Results of ANOVA of soil fertility samples, which may partially explain inter site differences in NTFP maximum height, mortality and flowering. Mean at each site as well as standard error and probability are included in the table. Sample size n=64 at each site.

Site mean Soil fertility metric DAPS HARC St. Error Pr (>|t|) pH 5.09 5.88 0.2654 0.0102 N.A. (meq/100g) 4.25 1.69 0.7556 0.0044 O.M. % 2.63 2.49 0.2848 0.6370 Bray I P (kg/hectare) 41.19 41.19 10.7800 1.0000 Ca (kg/hectare) 1733.68 3032.74 355.7000 0.0057 Mg (kg/hectare) 256.11 430.41 39.2600 0.0014 K (kg/hectare) 230.90 242.52 17.1700 0.5550 Na (kg/hectare) 26.03 24.15 3.1800 0.6067 Zn ppm 0.94 1.19 0.1940 0.2184 Fe ppm 31.25 40.81 4.4040 0.0476 Mn ppm 47.66 18.38 4.7630 <0.0001 Cu ppm 0.52 0.47 0.0598 0.4770 S ppm 4.80 5.21 0.7572 0.5945 NO3-N ppm 5.73 5.15 1.6588 0.7285 NH4-N ppm 7.03 4.86 0.4570 0.0003 B ppm 0.42 0.59 0.0636 0.0213

DISCUSSION This study sought to examine whether the vertical aboveground growth, flowering and mortality of four commonly harvested perennial woodland herbs was significantly higher in canopy gaps than under undisturbed canopy. Differences in plant response between treatment and control may have been confounded by transplant shock. However, all plants were transplanted so any shock that would induce or restrict flowering was distributed across all individuals, though potentially transplant shock and canopy openness may have interacted to impact the growth, survival and reproduction differently in canopy gaps and under the control. One example of distinct effects of light levels after transplanting comes from a reciprocal transplant study on Sanguinaria canadensis, where individuals were moved between sites with low and medium light levels. Unfertilized rhizomes from the sunnier location did not produce any seeds in the more shaded location, whereas unfertilized rhizomes from the shadier location produced seeds in the higher light environment(Marino et al., 1997).

It was expected that height and flowering of the study species would be greater in the canopy gaps compared to under the closed canopy. There was only one positive relationship between gap treatment and 30

flowering and that was for A. racemosa. The positive relationship between canopy gaps and height for A. racemosa, A. tricoccum and C. canadensis supported the hypothesis. H. canadensis had its highest mean height in canopy gaps at HARC, but in the control at DAPS. These findings are similar to those of a boreal forest study in which NTFP herb size was significantly (P < 0.05) higher in plots with canopy thinning than herbs under the closed canopy in unthinned stands of trees(Clason et al., 2008b).

In 2015, the light response curves at HARC for both A. racemosa (Figure 11) and C. canadensis

(Figure 13) indicated higher rates of photosynthesis in the control than in the canopy gap plots. This unexpected result may be explained by the large range in initial weight of the rhizomes of these two species, see Table 1. Light curves were time-consuming to obtain, keeping sample size to one individual plant per species by site by treatment combination. Plants were randomly selected for measurement of photosynthesis, and if initial mass of propagule had a significant impact on photosynthesis rates, selected specimens may not have been representative of a population mean light curve. Plants with larger initial rhizome mass may have had a photosynthetic advantage over smaller individuals. The finding that the light saturation points of all species were less than or equal to the PAR available in canopy gaps is still interesting. It suggests that the light needs of these NTFPs can be completely met, for at least part of the day, in canopy gaps that are created by uneven age management harvests.

In a study on NTFPs cultivated under different light and soil conditions (Naud et al., 2010), specimens of the taxon studied, which included A. racemosa, were significantly (P < 0.05) positively related to concentrations of both acidity-related elements including pH and Fe, as well as elements associated with soil fertility, including Ca and Mg. Those particular indices of fertility were significantly higher at HARC than at DAPS, as was soil B. However, mean soil levels of NA, Mn and NH4-N were significantly higher at DAPS. This complicated picture of differences in soil fertility by site makes it hard to say how soil fertility may have impacted NTFP response. Mean H. canadensis height was higher at

HARC than at DAPS, both in the canopy gap and control plots. C. canadensis height was comparable between both sites within treatment. Mean height of A. tricoccum was significantly greater at HARC than at DAPS in the canopy gap treatment, while A. racemosa height was significantly greater at HARC than at

DAPS in the control. These differences in height by site may point to overall greater soil fertility at HARC.

31

There may have been an interactive effect of light level and soil fertility, considering how hourly mean light levels in canopy gaps at DAPS spiked almost twice as high around mid-day than in canopy gaps at

HARC (Figure 8 and Figure 9). While both sites were positioned on North-facing backslopes, the growing area at HARC was positioned downslope from a shoulder ridge that is cleared for agriculture (Figure 2).

Consequently there is a thick midstory layer upslope, to the South of plots at HARC which may be intercepting more PAR than the sparse mid-story upslope from the plots at DAPS. The lower light levels in canopy gaps at HARC in comparison with measurements taken in canopy gaps at DAPS can be seen in

Figure 8 and Figure 9. Given that the measured light saturation points of all three species occurred at levels higher than the mean PAR levels in canopy gaps at HARC, light may have been a limiting factor for plant growth at HARC even in canopy gaps. Light deprivation at HARC may be masking what would be a postive relationship between NTFP responses and the higher levels of essential soil nutrients at HARC.

Mortality was significantly lower in canopy gap treatments for A. tricoccum at both sites, but significantly greater for H. canadensis in canopy gaps at DAPS. Reducing the overstory percent stocking via gap treatment may have inadvertently reduced the amount of soil moisture available to plants in that treatment, which may have impacted mortality in multiple ways. Soil organic matter may have accumulated less quickly in canopy gaps (Ni et al., 2016), and given the positive relationship between soil organic matter and soil moisture, soil moisture may have been a limiting factor in the canopy gaps. Devising strategies to increase soil moisture in canopy gaps may be beneficial, as the ability to opportunistically increase photosynthesis to take advantage of increased PAR after timber harvest may be moisture-limited.

Increasing soil organic matter, or at least reducing losses associated with forest farming-relating activities, could be one strategy to achieve higher soil moisture levels. Soil organic matter can increase the water- holding capacity of soils, by acting as a reservoir for additional soil moisture. In other species, it has been shown that additional soil moisture can buffer against crop failure during periods of drought-induced stress

(Williams et al., 2016). There has been concern that climate change and related drought events could adversely affect NTFP production in the Midwest (Winkler et al., 2014). Uneven-aged management has been suggested as a silvicultural tool to mitigate the effects of climate change-induced drought on timber production, by increasing soil organic matter. In comparison with Ozark forest stand receiving even-aged

32

management, stands receiving uneven-aged management have accumulated significantly (P < 0.05) higher- on-site soil carbon (C) and nitrogen (N) stores (Li et al., 2007) including soil organic matter. Uneven-aged management can also yield higher aboveground C stocks than clearcutting stands, via both higher stocking rates, and stimulation of C sequestration by small trees from increased canopy openness and PAR

(D’Amato et al., 2011a; Pukkala et al., 2011). Poorly drained soils in Missouri may increase risk of root mortality of A. racemosa due to a Phytophthora–Pythium disease complex (Thomas et al., 2006), but soils were fairly well drained at both sites. Organic matter percentage was not significantly (P=0.6370) different between sites, so it is unlikely this played a key role by itself in the differences observed in NTFP responses between sites.

Differences in flowering for A. racemosa, H. canadensis and C. canadensis by site could be useful information for forest farmers. If herbs growing on sites that match a standardized ecological site descriptions have significantly higher rates of flowering, these populations may produce additional seeds, additional seeds that could be used to establish plantings with shorter payback periods. There may be an effect of transplant shock interacting with the canopy gap treatment, and caution should be used in extrapolating too much from this single study. A follow up study assessing the same response variables, as well as new rametes of the same species in the plots, could help untangle the initial effects of transplant shock from the main treatment effects.

Harvesting too much of a plant’s root structure for sale can eliminate that individual from the landscape, which could be problematic for the viability of uncommon NTFP species. If the creation of canopy gaps can stimulate sexual reproduction, the increase in available propagules may allow the forest farming practitioner to expand a rare NTFP population through seeding, prior to destructive harvest of mature specimens. While canopy gaps may enhance growth of well-established herbs and stimulate seed production of NTFPs, this can also lead to decreases in the quality of microsite conditions for NTFP seed germination (Chandler and McGraw, 2015). Practitioners may establish new patches of NTFP in closed canopy areas, with the seed produced in recently created canopy gaps, prior to destructive harvest of the parent plants in those gaps. This could have the additional advantage of stimulating a growth spurt in the parent plants prior to harvest. Initial propagule mass was positively related to the reproduction of

33 A. racemosa. If facilitating sexual reproduction of A. racemosa at a new forest farming site is a goal, it may be advisable to transplant larger rhizomes to initiate that process more quickly.

Anecdotal observations revealed that competition for space and possibly for water between NTFPs and graminoid species was a problem in the canopy gap treatment at DAPS. Site preparation to control for weedy species may be helpful. Small-statured perennial herbs such as the NTFPs in this study may benefit from ongoing fire treatments (Knapp et al., 2015) or fire and herbicide treatments (Miller and Chamberlain,

2008), as these can reduce competition from graminoids and mid-story woody plants.

ACKNOWLEDGEMENTS We would like to thank the University of Missouri Center for Agroforestry, the Agricultural

Research Service, and ultimately the American public for funding this project. Additionally, the highest gratitude goes out to Ellen Lang and Richard Young for helping monitor and establish these plots, and to

Doug Allen for use of their land and for their passion in promoting agroforestry in rural Missouri.

34 CHAPTER 3 IMPACTS OF TIMBER HARVEST ON MEDICINAL HERBS IN THE MISSOURI OZARKS Abstract: Integrating silvicultural systems with forest farming could result in profitable over- yielding. To calibrate some of these multi-canopy level olycultures for the Central Hardwoods region, relationships were explored between percent cover and frequency of native non-timber forest product

(NTFP) medicinal herbs with timber harvest type and ecosystem site description. Study species were

Actaea racemosa L., Apocynum cannabinum L., Aristolochia serpentaria L., Dioscorea quaternata J.F.

Gmel., Echinacea simulata R.L. McGregor, Geranium maculatum L., Hydrastis canadensis L., Parthenium integrifolium L., Podophyllum peltatum L., Sanguinaria canadensis L. or Tephrosia virginiana (L.) Pers.

NTFP percent cover and frequency was significantly (P < 0.05) related to certain ecological site descriptions. Significant positive relationships were found between NTFP responses with no harvest and uneven-aged harvest types, as well as certain ecological site types. Ecological site descriptions can be used in conjunction with silvicultural management to select and manage sites for forest farming.

Keywords: non-timber forest product; ecological site description; forest farming; medicinal herb; even-aged management; uneven-aged management; Ozark Highland Plateau, USA.

INTRODUCTION Forest farming is an agroforestry system where wild or planted populations of non-timber forest products species (NTFPs) are cultivated beneath a woodland or forest canopy (Chamberlain, 2013; Naud et al., 2010). Over-yielding refers to the phenomenon of when the sum of the percentages of maximum theoretical yields for each species, as measured growing in a monoculture, is greater than 100% when these species are grown together in a polyculture(Jose et al., 2004b). Populations of certain wild NTFPs in the eastern United States are shrinking to threatened levels, partially as a result of overharvesting (Farrington et al., 2009b; Souther and McGraw, 2014). Active cultivation of NTFPs has been suggested as a conservation measure (Burkhart, 2011b), but the cost of labor can be prohibitively expensive if the NTFP is the only saleable product of management activities (Burkhart, 2011a). Silvicultural prescriptions may be tailored to stimulate increased NTFP volume (Clason et al., 2008b). When treatments prescribed for timber management create conditions condusvie to NTFP proliferation, these stands could profitably over-yield

35

with little additional labor, assuming weak competition between timber and NTFP species (Jose et al.,

2004b).

Selection cuts, which are typical harvest types of uneven-aged silvicultural management systems, can mimic the structure of mature oak-hickory woodland and forest (Goebel and Hix, 1996) and may favor understory species that are adapted to a partial shade and sun fleck-based light regime. Adequate light is critical for overstory regeneration as well as ground flora; naturally occurring canopy gaps may have been more important than fire for Quercus alba L. and Pinus echinata Mill. in attaining canopy status (King and

Muzika, 2014a). Mimicking the temporal and spatial heterogeneity of mature woodland canopies, whether through timber harvest (Gillespie et al., 2006b) or fire (Knapp et al., 2015), enhances light transmittance to the forest floor. Whether from the elimination of midstory or the creation of canopy gaps, management that mimics the structure of a climax woodland supports a continuous layer of forbs (Hanberry et al., 2014a).

Since the European conquest and the suppression of wildfire in the 20th century, tree canopies have become more dense on the Missouri plains by an estimated factor of two (Hanberry et al., 2014b), and by an estimated factor of 2.3 on the Missouri Ozark Plateau (Hanberry et al., 2014c). Regional ecological restoration efforts often employ prescribed fire (Ladd, 2014) and thinning harvest types, both of which typically result in increases in size and species richness of understory forbs (Kinkead, 2013; Maginel et al.,

2016a). The current dominance of oak-hickory in Ozark forests is hypothesized to be a legacy of a regional

First Nations’ agroforestry strategy: prescribed fire was used to select for mast-producing Quercus species, which supported up to five times the large mammal game species biomass as today (Jurney, 2012).

Similarly, foresters employ silvicultural techniques to favor Quercus for timber and wildlife benefits today

(Dey and Hartman, 2005; Lupardus et al., 2011; Vickers et al., 2014). Managers have a short-term opportunity to increase the abundance of certain ground flora (non-woody, non-vining forest or woodland plant species typically attaining 2>meter height), with timber harvest, due to increased light levels reaching the forest floor (Zenner et al., 2006). Deliberately steering the forest to increase production of NTFPs would require understanding the relationships between wild NTFP and reductions in overstory percent stocking resulting from timber harvest (Gingrich, 1967). With this information, inferences could be made

36

about what harvest types have the best chance at supporting increases in NTFP cover and frequency. With this information, a forester’s stand inventories could be used to determine site suitability for forest farming.

Forest farming may increase profits from forests between timber harvests (Vaughan et al., 2013) by providing revenue on an annual basis- even from small acreages (Bruhn, 2008). In Missouri, the ability to realize profits from small acreage has grown in importance over the past century as forestland ownership by individual landowners has been replaced by multiple landowners with smaller holdings, resulting in increasing number of owners and decreasing size of holdings; these small, non-industrial forest owners have insufficient acreage to achieve sustained timber yield (Ko and He, 2011). While large industrial forest owners may prefer even-aged management for simplicity’s sake, public and special private industrial forests such as Pioneer Forest often employ uneven-aged management, citing benefits - both social (Guldin et al., 2008) and ecological (D’Amato et al., 2011b; Duguid and Ashton, 2013). Portions of the public have an aversion to clearcutting (Dwyer et al., 2004b) and forest farming may provide a complimentary land use to the uneven-aged management systems that are alternatives to clearcutting; forests with an open-canopy woodland stocking level of 40-65%, such as that following an uneven-aged harvest, may leave growing conditions improved relative to the habitat needs of NTFP species. Uneven-age management was found to result in higher rates of survival for American ginseng (Panax quinquefolius L.), North America’s most profitable NTFP, than populations growing beneath a clearcut harvest type (Chandler and McGraw, 2015).

In this way, forest farming might provide an impetus for small non-industrial forest owners to proactively co-manage their timber and NTFP resources (Chamberlain et al., 2013), with profits from forest farming paying for consulting foresters’ services.

Beside the impact of timber harvest, abiotic site characteristics may also affect NTFP species.

Local variation in soil characteristics can mediate ground level flora response to timber harvest (Roberts and Gilliam, 1995). Soil disturbance in canopy gaps created by timber harvest had differential impacts on ground level species composition based on soil type and slope position (Duguid et al., 2013). Landscape positions that are steeper, tend to have well drained soils, while flatter positions have poorer drainage and may support mesic-adapted plant communities; tree canopies on flatter landscape positions are forced to compete for space in two dimensions, and develop greater crown shyness compared with trees on steeper

37

positions, which allows for a greater portion of the rain making it directly to the surface of the forest floor

(Siegert et al., 2016). Combined with slope position, overstory composition (which is in itself partially an artifact of past management activities) has significant (P < 0.05) effects on herb diversity(Ellum et al.,

2010); Panax ginseng C.A. Mey. has shown greater preference for oak-dominant stands, when compared with oak-pine and pine dominated stands (Woo, 2004). Aspect is a primary factor in determining microclimatic gradients on steep hillsides in the Central Hardwoods Region: South, Southwest and West- facing slopes tend to be drier and warmer than northwest, North and East-facing (Dress and Boerner, 2004).

Because of this, aspect can play a dominant factor in determining composition of ground flora species assemblages (Jenkins and Parker, 1999), just as it does with overstory species assemblages in the Ozarks

(Villwock et al., 2011).

Rather than choosing forest farming sites based on overstory condition and constellations of favorable site characteristics assembled piecemeal, ecological site classifications may provide useful guidance. The Natural Resource Conservation Service typology for terrestrial ecosystems, the Ecological

Site Description (ESD), may be useful for selecting appropriate forest farming sites. ESDs “incorporate interactions between local soil, climate, flora, fauna, and humans into schema for land management decision-making" (Nauman et al., 2015), and takes into consideration a site’s landscape position, soil series, steepness class, aspect class and dominant vegetation type. NTFPs growing in different ESDs may respond differently to timber harvests. This study will evaluate the usefulness of ESDs in forest farming site selection.

This study examines naturally-occurring populations of 11 commercially-harvested NTFP species:

Actaea racemosa L. (Ranunculaceae), Apocynum cannabinum L. (Apocynaceae),

Aristolochia serpentaria L. (Aristolochiaceae), Dioscorea quaternata J.F. Gmel. (Dioscoreaceae),

Echinacea simulate R.L. McGregor (Asteraceae), Geranium maculatum L. (Geraniaceae),

Hydrastis canadensis L. (Ranunculaceae), Parthenium integrifolium L. (Asteraceae),

Podophyllum peltatum L. (Berberidaceae), Sanguinaria canadensis L. (Papaveraceae) and

Tephrosia virginiana (L.) Pers. (Fabaceae). Using data from a long-term ecological research project at the

38 western edge of the ranges of these species, relationships between size (percent cover) and abundance

(frequency) of NTFP ground flora species with ESDs, harvest types and time since harvest were explored.

It was hypothesized that NTFP percent cover and frequency would be positively related to uneven- aged harvest types, and not with even-aged harvest types; decreasing canopy density would increase light availability at the ground level and benefit NTFP production but only up to a certain point. The level of canopy closure that would afford 300 µmol/m2/s PAR, based on results presented in Chapter II, enabled the development of the following hypothesis: Site characteristics which make the soil more hydric, defined as flat landscape positions, protected aspects and at least moderately deep, well-drained and permeable soils, would support NTFP species better than sites with distinctly xeric soils.

MATERIALS AND METHODS Study Area

The study was carried out using data from the Missouri Ozark Forest Ecosystem Project

(MOFEP), a long-term, landscape-scale experiment, designed by the Missouri Department of Conservation and partners to ascertain the effects of common silvicultural systems on an open-ended number of ecological responses (Knapp et al., 2014). MOFEP was laid out in a randomized complete block design. In each of three blocks, there are three compartments, to which one of three treatments was randomly assigned: even-aged management (EAM), uneven-aged management (UAM), or no-harvest management

(NHM) (Figure 17). Compartments are multi-stand management units ranging in size from 311 to 501 hectares each; collectively, the nine MOFEP compartments cover roughly 9,400 acres (Knapp et al., 2014).

Timber harvesting occurred on a 15 year interval; the first harvest was in 1996 and the second harvest was in 2011. There is no timber harvesting or salvage logging in NHM compartments. The two primary EAM silvicultural harvest types examined in this study are clearcut with reserves and intermediate thinning. In stands receiving clearcut with reserves, reserve trees are retained to provide food and cover for wildlife or, in the case of Pinus echinata Mill. Reserve trees, to provide seed for natural regeneration. There are usually ten to twenty reserves trees per hectare and are generally >30.5 cm diameter at breast height. Intermediate thinnings are conducted periodically within stands to increase the growing space for residual trees. Stands chosen for intermediate thinning usually have some mature and over-mature large sawtimber but also an 39

abundance of high quality poles and small sawtimber (Jensen and Kabrick, 2008). Additional EAM harvest types at MOFEP included commercial and non-commercial woodland management, shelterwood, and glade management EAM, but for simplicity’s sake these harvest types are not examined in this study.

Clearcutting affected an average of 11% and 13% of each EAM compartment, in the first and second harvest entries, respectively (Knapp et al., 2014). The primary UAM silvicultural prescription was single- tree selection; group selection was also used in 1996. Harvesting in UAM affected an average of 57% and

40% of each UAM compartment, in the first and second harvest entries, respectively (Knapp et al., 2014).

MOFEP contains 648 individual 0.202 hectare vegetation monitoring plots which were randomly distributed within stands, with each stand including at least one plot (Figure 16) (Dwyer et al., 2004b).

Thirty of these plots had none of the study species in recorded in any quadrat for any data collection year, leaving a total of 618 plots with NTFP data (n=614). Detailed soil information including depth (measured down to 48 inches) and soil series classification were collected at every plot before the first harvest (Shifley and Brookshire, 2000a). Field crews recorded cover of ground flora in sixteen 1-m2 quadrats/plot in the summers of 1995, 1999, 2000, 2001, 2009, 2013 and 2014 (Figure 16). All vascular plants in quadrats were identified to species level. The cover of each was estimated to the nearest 1%; values of <0.5% were assigned a value of 0.1%.

Statistical Analyses

The percent cover of NTFP species within quadrats, and the frequency (fraction of quadrats/plot with that species present) served as response variables. From the MOFEP dataset, 11 NTFP species were chosen for investigation in this study. These were chosen on the basis of being purchased by Missouri- based American Botanicals for >$1/lb and being present in enough plots for the linear mixed-effects models to meaningfully predict their frequency and cover. The numbers of individuals of each medicinal herb species in each quadrat was not tallied, because at the time of the data collection protocol design The

Nature Conservancy did not consider this important data to collect. Rather, the cover for a species within a quadrat represents an approximation of the biomass of an individual or small group of individuals. The frequency is an approximation of the spatial extent and density of that population. Increases in frequency

40

could be the result of already existing individuals sexually reproducing across the landscape; decreases could be the result of mortality or dormancy. Data were logarithmically or gamma-transformed in an effort to satisfy the assumption of normality.

Treatments were randomly assigned within the blocks, but silvicultural prescriptions (also known as ‘harvest type’) occurred at the stand level and were not randomly assigned, since they are based on knowledge of pre-harvest inventory data. The stands in the EAM and UAM compartments that have not been harvested would provide substantial noise to multiple comparison analyses or analysis of variance between compartments. Therefore, unharvested plots in EAM and UAM were given the same designation as stands in the NHM compartments (no harvest) and were compared with plots receiving other harvest type designations. Significant positive relationships between particular ESDs and harvest types, may point to ESDs and harvest types that are compatible with forest farming.

Linear mixed-effects regressions were fit using ggplot2, lme4 and lmerTest packages in RStudio.

Data from years 1995, 1999, 2000, 2001, 2009, 2013 and 2014 were used. These models used data from all complete botany survey years and had harvest type, ESD, years since treatment and interaction terms as predictors. NTFP responses, which had significant relationships with predictor variables, are examined in the results. Examples of the models used for each response variable for A. racemosa are given:

ACTRACcover=lmer(log.cover~ESD*harvest*years+(1|Site/Plot/Quad),data=MOFEPcover)

ACTRACfreq<-lmer(gamma.freq~ESD*harvest*years+(1|Site/Plot),data=MOFEPfreq)

where “lmer” is a function for fitting linear mixed-effects models using the “lme4” package for the

R software program (Bates et al., 2014); “log.cover” is a logarithmic transformation of the percent cover of this species for the sake of normality and “gamma.freq” is a gamma transformation of the frequency; “~” is used to separate the left- and right-hand sides in a model formula ; “ESD*Treatment*years.since” is an interaction term between ESD assigned at the plot level, the timber harvest type received, and years since harvest (interaction terms also call up each component variables for individual variables analysis as well);

41

“(1|Site/Plot/QUADRAT_ID)” and “(1|Site/Plot) ”signify the random-effects terms of management blocking unit, plot and quadrat; “data=ACTRACpercent” and “ACTRACfreq1” refer to the names of the files from which the data were analyzed. For all analyses in this study, alpha=0.05.

42 Figure 24

Figure 16 Maps of plot design at MOFEP.

Figure 25

Figure 17 Silvicultural systems employed at MOFEP, by compartment.

43

RESULTS Significant positive relationships were found between NTFP responses, ESDs and harvest types.

Summary statistics for explanatory variables are reported (Table 6). Number of plots receiving each harvest type by year of entry are summarized (Table 7). A list of all ESDs at MOFEP are summarized (Table 8).

Significant relationships between NTFP frequency and cover with treatment, ESD, years since harvest and interaction terms thereof are listed in Table 9, Table 10, Table 11, Table 12, Table 13, Table 14, Table 15,

Table 16, Table 17 and Table 18, “NS” stands for not significant. Mean percent cover and frequency by species and harvest type are shown (Figure 31, Figure 32, Figure 33, Figure 34, Figure 35, Figure 36,

Figure 37, Figure 38, Figure 39, Figure 40 and Figure 41).

Table 6 Summary statistics (mean, standard deviation, range) and description of explanatory variables. SD=standard deviation. Units of measurement indicated in brackets. Sample size (total quadrats) containing members of each species changed by year, but over the course of the study there were 618 plots that had at least one quadrat with one study species recorded growing in it. Variable Name Description Mean SD Min. Max. Years since Time since harvest/tending/no harvest (years). 9.23 6.7 0 19 Even-age management harvest type, typically leaving few trees for regeneration purposes Clearcut (% stocking). 30 26 0 72 Uneven-aged management harvest type, leaving canopy gaps 70-140 feet wide (% Group Selection stocking). 60 18 23 92 Uneven-aged management harvest type, trees of all size classes removed uniformly throughout stand to promote growth of remaining trees and provide space for Single-Tree Selection regeneration (% stocking). 67 16 29 103 Even-age management "harvest type", in reality a tending treatment to enhance growth, quality, vigor and composition of stand after regeneration and prior to actual harvest (% Intermediate Thin stocking). 69 15 40 118 Leave No trees were harvested during the course of this study (% stocking). 83 15 24 125 Actaea racemosa frequency Frequency (quadrats/plot) of Actaea racemosa in plots containing this species. 3.09 2.43 1 14 Apocynum cannabinum frequency Frequency (quadrats/plot) of Apocynum cannabinum in plots containing this species. 1.27 0.64 1 5 Aristolochia serpentaria frequency Frequency (quadrats/plot) of Aristolochia serpentaria in plots containing this species. 1.97 1.15 1 9 Dioscorea quaternata frequency Frequency (quadrats/plot) of Dioscorea quaternata in plots containing this species. 3.19 2.25 1 12 Echinacea simulata frequency Frequency (quadrats/plot) of Echinacea simulata in plots containing this species. 1.80 1.52 1 7 Geranium maculatum frequency Frequency (quadrats/plot) of Geranium maculatum in plots containing this species. 2.54 2.01 1 14 Hydrastis canadensis frequency Frequency (quadrats/plot) of Hydrastis canadensis in plots containing this species. 1.72 1.04 1 5 Parthenium integrifolium frequency Frequency (quadrats/plot) of Parthenium integrifolium in plots containing this species. 1.54 1.06 1 9 Podphyllum peltatum frequency Frequency (quadrats/plot) of Podophyllum peltatum in plots containing this species. 1.56 1.26 1 8 Sanguinaria canadensis frequency Frequency (quadrats/plot) of Sanguinaria canadensis in plots containing this species. 2.02 1.70 1 8 Tephrosia virginiana frequency Frequency (quadrats/plot) of Tephrosia virginiana in plots containing this species. 1.33 0.68 1 5 Actaea racemosa cover Percent cover (0.01-100%) of Actaea racemosa in plots containing this species. 7.72 9.12 0.10 80.00 Apocynum cannabinum cover Percent cover (0.01-100%) of Apocynum cannabinum in plots containing this species. 3.51 6.24 0.10 65.00 Aristolochia serpentaria cover Percent cover (0.01-100%) of Aristolochia serpentaria in plots containing this species. 0.42 0.53 0.01 11.00 Dioscorea quaternata cover Percent cover (0.01-100%) of Dioscorea quaternata in plots containing this species. 1.89 2.00 0.01 20.00 Echinacea simulata cover Percent cover (0.01-100%) of Echinacea simulata in plots containing this species. 1.68 1.66 0.10 10.00 Geranium maculatum cover Percent cover (0.01-100%) of Geranium maculatum in plots containing this species. 1.67 2.39 0.01 30.00 Hydrastis canadensis cover Percent cover (0.01-100%) of Hydrastis canadensis in plots containing this species. 5.19 7.73 0.10 45.00 Parthenium integrifolium cover Percent cover (0.01-100%) of Parthenium integrifoliumin plots containing this species. 1.54 1.68 0.10 20.00 Podphyllum peltatum cover Percent cover (0.01-100%) of Podophyllum peltatum in plots containing this species. 4.84 5.37 0.10 28.00 Sanguinaria canadensis cover Percent cover (0.01-100%) of Sanguinaria canadensis in plots containing this species. 0.80 0.79 0.10 3.00 Tephrosia virginiana cover Percent cover (0.01-100%) of Tephrosia virginiana in plots containing this species. 2.52 3.94 0.05 48.00

44

Table 7 Number of plots receiving each harvest type in 1996 and 2011. Treatment Plots harvested in 1996 Plots harvested in 2011 Clearcut 28 32 Group Selection 43 0 Single-Tree Selection 75 81 Intermediate Thin 36 23 Leave 429 462

45

Quercus muehlenbergii, Quercus shumardii; Rhus aromatica; Schizachyrium scoparium, scoparium, Schizachyrium Rhusshumardii; Quercus muehlenbergii, Quercus aromatica; Asimina Asarum triloba; canadense,saccharum; Carex Acer sp. rubra, Quercus Asimina rubra; Asarum Quercus triloba; canadense. saccharum, Acer Quercus alba,benzoin; Quercus Lindera bulbosa,Erigenia rubra; Quercus Asarum canadense. canadensis; virginicus, Uvularia Elymus alba; Quercus muehlenbergii, Quercus Cercis Quercus alba,Quercus Desmodium sp. arborea; velutina; Quercus Amelanchier alba,Quercus scoparium. Carex, Schizachyrium velutina; Quercus Rhus aromatica; Description

Quercus alba, Quercus velutina; Cercis canadensis; alba, virginicus, PodophyllumQuercus Elymus velutina; Quercus Cercis peltatum. Danthonia Rhus aromatica; velutina; Quercus Vaccinium, Schizachyrium stellata, spicata, Quercus Platanus Populus occidentalis, canadensis. Laportea Salix deltoides; interior; scoparium. Schizachyrium Rhus muehlenbergii; Quercus aromatica; stellata, Quercus Quercus stellata, Quercus muehlenbergii; Bumelia lanuginosa, muehlenbergii; Quercus stellata, Quercus Rudbeckia Rhus missouriensis, aromatica; Rhusmarilandica; Quercus copallina; Crotonopsis scoparium, Schizachyrium elliptica. alba,Quercus velutina; Quercus Cornus hystrix. florida; Elymus virginicus, Carex Elymus pensylvanica. velutina,Quercus alba; Quercus Vaccinium; Fraxinus macrocarpa, Quercus pennsylvanica; Cornus Cardamine cylindrica, foemina; Boehmeria Vitis, Helianthus scoparium, Schizachyrium Rhusshumardii; hirsutus. Quercus muehlenbergii, Quercus aromatica; Quercus alba, Quercus coccinea; Rhus Cornus alba, Quercus aromatica, Solidagoflorida; Desmodium sp., coccinea; Quercus ulmifolia. Quercus velutina, Pinus echinata; Vaccinium; Carex, Schizachyrium scoparium. velutina, Carex, Schizachyrium Quercus Pinus Vaccinium; echinata; alba, Quercus Cornus rubra; Claytonia Quercus florida; Aristolochia serpentaria, virginica. Quercus stellata, Quercus marilandica; Rhus aromatica; Schizachyrium scoparium. Schizachyrium Rhusmarilandica; Quercus aromatica; stellata, Quercus Quercus stellata, Quercus velutina; Rhus aromatica; Schizachyrium scoparium, Desmodium sp.scoparium, Schizachyrium velutina; Quercus Rhus stellata, aromatica; Quercus scoparium. Carex, Schizachyrium Vaccinium; velutina,Quercus stellata; Quercus Steep backslopeSteep (>15%) woodland & forest, typical floristic association includes: Upland summit, shoulder & backslope (<15%) woodland, typical floristic backslope associationSteep (>15%) woodland & forest, typical includes: floristic associationconcatenata. includes: backslopeSteep (>15%) woodland & forest, typical floristic associationUpland includes: summit, shoulder & backslope (<15%) woodland, typical floristicgrandiflora. association includes: Sinkholes, footslopes, & high terraces, typical floristic association includes: backslopeSteep (>15%) woodland & forest, typical floristic associationUpland includes: summit, shoulder & backslope (<15%) woodland, typical floristicHelianthus hirsutus. association includes: Sinkholes, footslopes, & high terraces, typical floristic association includes:Sinkholes, footslopes, & high terraces, typical floristic association includes:Upland drainageways, low terraces & floodplains, typical floristic backslopeSteep (>15%) woodland association & forest, typical floristic includes: associationscoparium. includes: backslopeSteep (>15%) woodland & forest, typical floristic association includes: Upland summit, shoulder & backslope (<15%) woodland, typical floristic associationUpland includes: summit, shoulder & backslope (<15%) woodland, typical floristic association backslopeSteep includes: (>15%) woodland & forest, typical floristic association includes: backslopeSteep (>15%) woodland & forest, typical floristic association includes: Upland drainageways, low terraces & floodplains, typical floristic associationGlades & cliffs, includes: typical floristic association includes: Glades & cliffs, typical floristic association includes: Steep backslopeSteep (>15%) woodland & forest, typical floristic association scoparium. Schizachyrium includes: backslopeSteep (>15%) woodland & forest, typical floristic association includes: backslopeSteep (>15%) woodland & forest, typical floristic association includes: 6 1 6 1 7 5 1 1 2 3 2 2 7 89 40 18 15 14 59 16 80 106 108 Plots Names andNames descriptions ESDs,of number plotsof with classified eachas ESD

8 Low-base Sandstone Backslope Protected Woodland Sandy/Gravelly Floodplain Forest Shallow Dolomite Upland Glade/Woodland Shallow Sandstone Upland Glade/Woodland Loamy Sinkhole Woodland Loamy Terrace Forest Low-base Exposed Chert Backslope Woodland Low-base Backslope Protected Chert Woodland Low-base Upland Chert Woodland Low-base Loamy Upland Woodland Low-base Sandstone Exposed Backslope Woodland Chert Protected Backslope Protected Chert Forest UplandChert Woodland Dry Footslope Forest Dry Igneous Exposed Backslope Woodland Fragipan Upland Woodland Loamy Footslope Forest Calcareous Dolomite Exposed Backslope Woodland Calcareous Dolomite Backslope Protected Forest DolomiteChert Exposed Backslope Woodland DolomiteChert Backslope Protected Forest DolomiteChert Upland Woodland ExposedChert Backslope Woodland Table

46 Table 9 All harvest types, ESDs, years since harvest and interaction terms thereof with significant relationships with percent cover or frequency of Actaea racemosa.

Actaea racemosa Response cover frequency Predictor Est. St. Error Est. St. Error (Intercept) NS NS -1.024 0.3643 Single-Tree Selection 1.2020 0.2986 NS NS Single-Tree Selection:years since -0.0686 0.0334 NS NS Intermediate Thin 2.6020 1.0980 NS NS Chert Protected Backslope Forest:Intermediate Thin -2.5590 1.1430 NS NS Chert Protected Backslope Forest:No Harvest:years since 0.1385 0.0675 NS NS Chert Protected Backslope Forest:Single-Tree Selection -0.9509 0.4689 NS NS Chert Protected Backslope Forest:Single-Tree Selection:years since 0.1277 0.0548 NS NS Dry Footslope Forest:No Harvest:years since 0.1079 0.0539 NS NS Low-base Chert Protected Backslope Woodland:Intermediate Thin -2.6060 1.1810 NS NS

Table 10 All harvest types, ESDs, years since harvest and interaction terms thereof with significant relationships with percent cover or frequency of Apocynum cannabinum. Apocynum cannabinum Response cover frequency Predictor Est. St. Error Est. St. Error (Intercept) NS NS 0.6400 0.0944 Chert Protected Backslope Forest NS NS 0.2561 0.0999 Chert Protected Backslope Forest:Intermediate Thin NS NS -0.2699 0.1005 Chert Protected Backslope Forest:No Harvest NS NS -0.2561 0.0929 Chert Protected Backslope Forest:Single-Tree Selection NS NS -0.1827 0.0616 Chert Protected Backslope Forest:years since:Intermediate Thin NS NS 0.0539 0.0266 Chert Protected Backslope Forest:years since:No Harvest NS NS 0.0499 0.0253 Chert Protected Backslope Forest:years since:Single-Tree Selection NS NS 0.0175 0.0054 Low-base Chert Protected Backslope Woodland NS NS 0.2167 0.0969 Low-base Chert Protected Backslope Woodland:No Harvest NS NS -0.2188 0.0901 Low-base Chert Protected Backslope Woodland:Single-Tree Selection NS NS -0.1434 0.0619 Low-base Chert Protected Backslope Woodland:years since:Single-Tree Selection NS NS 0.0158 0.0040 Wet Footslope Forest NS NS 0.0997 0.0468 Wet Footslope Forest:years since -0.4574 0.2178 NS NS Sandy/Gravelly Floodplain Forest:years since NS NS 0.0078 0.0031 years since:Single-Tree Selection NS NS -0.0090 0.0029

47 Table 11 All harvest types, ESDs, years since harvest and interaction terms thereof with significant relationships with percent cover or frequency of Aristolochia serpentaria. Aristolochia serpentaria Response cover frequency Predictor Est. St. Error Est. St. Error Calcareous Dolomite Protected Backslope Forest:years since NS NS 0.0335 0.0143 Chert Dolomite Exposed Backslope Woodland NS NS -0.7727 0.3240 Chert Dolomite Exposed Backslope Woodland:No Harvest NS NS 0.7080 0.2921 Chert Dolomite Exposed Backslope Woodland:years since NS NS 0.0495 0.0193 Chert Dolomite Protected Backslope Forest NS NS -0.7159 0.3204 Chert Dolomite Protected Backslope Forest:No Harvest NS NS 0.6122 0.2883 Chert Dolomite Upland Woodland NS NS -0.5003 0.2350 Chert Dolomite Upland Woodland:years since NS NS 0.0341 0.0160 Chert Exposed Backslope Woodland NS NS -0.7290 0.2529 Chert Exposed Backslope Woodland:No Harvest NS NS 0.4478 0.2066 Chert Protected Backslope Forest NS NS -0.6854 0.2521 Chert Protected Backslope Forest:No Harvest NS NS 0.4587 0.2055 Dry Footslope Forest NS NS -0.6852 0.3134 Dry Footslope Forest:years since:Intermediate Thin NS NS 0.0451 0.0214 Fragipan Upland Woodland NS NS -0.7758 0.3439 Fragipan Upland Woodland:years since NS NS 0.0518 0.0207 Fragipan Upland Woodland:years since:No Harvest NS NS -0.0368 0.0183 Loamy Footslope Forest:years since NS NS 0.0273 0.0136 Loamy Terrace Forest 1.3910 0.6050 NS NS Low-base Chert Exposed Backslope Woodland NS NS -0.8030 0.2519 Low-base Chert Exposed Backslope Woodland:No Harvest NS NS 0.4187 0.2055 Low-base Chert Protected Backslope Woodland NS NS -0.7395 0.2563 Low-base Chert Protected Backslope Woodland:No Harvest NS NS 0.4732 0.2108 Low-base Chert Upland Woodland NS NS -0.6353 0.2715 Low-base Loamy Upland Woodland NS NS -0.3925 0.2000 Low-base Loamy Upland Woodland:years since NS NS 0.0306 0.0153 Low-base Sandstone Exposed Backslope Woodland NS NS -0.8550 0.3007 Low-base Sandstone Exposed Backslope Woodland:No Harvest NS NS 0.5417 0.2639 Low-base Sandstone Exposed Backslope Woodland:years since NS NS 0.0417 0.0178 Low-base Sandstone Protected Backslope Woodland NS NS -0.7800 0.2793 Sandy/Gravelly Floodplain Forest:years since NS NS 0.0258 0.0131 Shallow Dolomite Upland Glade/Woodland:years since NS NS 0.0287 0.0129 Wet Footslope Forest:years since NS NS 0.0362 0.0144

48

Table 12 All harvest types, ESDs, years since harvest and interaction terms thereof with significant relationships with percent cover or frequency of Dioscorea quaternata. Dioscorea quaternata Response cover frequency Predictor Est. St. Error Est. St. Error years since 0.3420 0.1390 NS NS Group Selection 0.8209 0.3856 NS NS Chert Exposed Backslope Woodland:years since -0.3395 0.1393 NS NS Chert Exposed Backslope Woodland:years since:Group Selection 0.3468 0.1400 NS NS Chert Exposed Backslope Woodland:years since:Intermediate Thin 0.3592 0.1413 NS NS Chert Exposed Backslope Woodland:years since:No Harvest 0.3413 0.1392 NS NS Chert Protected Backslope Forest:Group Selection -0.8486 0.3952 NS NS Chert Protected Backslope Forest:years since -0.3415 0.1390 NS NS Chert Protected Backslope Forest:years since:Group Selection 0.3623 0.1398 NS NS Chert Protected Backslope Forest:years since:Intermediate Thin 0.3559 0.1409 NS NS Chert Protected Backslope Forest:years since:No Harvest 0.3473 0.1389 NS NS Low-base Chert Exposed Backslope Woodland:years since -0.3320 0.1399 NS NS Low-base Chert Exposed Backslope Woodland:years since:No Harvest 0.3364 0.1399 NS NS Low-base Chert Protected Backslope Woodland:Group Selection -0.7836 0.3920 NS NS Low-base Chert Protected Backslope Woodland:years since -0.3389 0.1390 NS NS Low-base Chert Protected Backslope Woodland:years since:Group Selection 0.3584 0.1397 NS NS Low-base Chert Protected Backslope Woodland:years since:Intermediate Thin 0.3561 0.1410 NS NS Low-base Chert Protected Backslope Woodland:years since:No Harvest 0.3476 0.1389 NS NS Wet Footslope Forest:years since 0.0192 0.0078 NS NS years since:Group Selection -0.3603 0.1395 NS NS years since:Intermediate Thin -0.3570 0.1408 NS NS years since:No Harvest -0.3461 0.1389 NS NS

Table 13 All harvest types, ESDs, years since harvest and interaction terms thereof with significant relationships with percent cover or frequency of Echinacea simulata. Echinacea simulata Response cover frequency Predictor Est. St. Error Est. St. Error Single-Tree Selection 0.5089 0.2536 NS NS

Table 14 All harvest types, ESDs, years since harvest and interaction terms thereof with significant relationships with percent cover or frequency of Geranium maculatum. Geranium maculatum Response cover frequency Predictor Est. St. Error Est. St. Error (Intercept) NS NS -1.4371 0.6474 Chert Exposed Backslope Woodland:years.since:Group Selection NS NS 0.0559 0.0233 Low-base Chert Protected Backslope Woodland:Single-Tree Selection 1.9550 0.9908 NS NS

49

Table 15 All harvest types, ESDs, years since harvest and interaction terms thereof with significant relationships with percent cover or frequency of Hydrastis canadensis. Hydrastis canadensis Response cover frequency Predictor Est. St. Error Est. St. Error Chert Dolomite Protected Backslope Forest 1.6978 0.7860 NS NS Chert Dolomite Protected Backslope Forest:years since -0.1524 0.0681 NS NS Chert Protected Backslope Forest 1.7856 0.7732 NS NS Chert Protected Backslope Forest:years since -0.1425 0.0678 NS NS Dry Footslope Forest 1.6336 0.7796 NS NS Dry Footslope Forest:years since -0.1020 0.0455 NS NS Low-base Chert Protected Backslope Woodland 1.9355 0.7588 NS NS Low-base Chert Protected Backslope Woodland:years since -0.1405 0.0650 NS NS Sandy/Gravelly Floodplain Forest 1.8166 0.7534 NS NS Shallow Dolomite Upland Glade/Woodland 1.8249 0.7574 NS NS Sandy/Gravelly Floodplain Forest:years since -0.1351 0.0468 NS NS Shallow Dolomite Upland Glade/Woodland:years since -0.1192 0.0467 NS NS

Table 16 All harvest types, ESDs, years since harvest and interaction terms thereof with significant relationships with percent cover or frequency of Parthenium integrifolium. Parthenium integrifolium Response cover frequency Predictor Est. St. Error Est. St. Error No Harvest NS NS 0.3706 0.1725 (Intercept) NS NS -1.2420 0.2106 Chert Dolomite Exposed Backslope Woodland:years since NS NS 0.0173 0.0083 Chert Exposed Backslope Woodland:No Harvest NS NS -0.4681 0.1800 Chert Upland Woodland:years since NS NS 0.0233 0.0095 Low-base Chert Exposed Backslope Woodland:Group Selection -6.0090 2.7670 NS NS Low-base Chert Exposed Backslope Woodland:No Harvest NS NS -0.4002 0.1869 Low-base Chert Protected Backslope Woodland:Group Selection -8.7640 3.8280 NS NS Low-base Sandstone Protected Backslope Woodland:years since NS NS 0.0206 0.0088 Sandy/Gravelly Floodplain Forest NS NS -0.3494 0.1655 Shallow Dolomite Upland Glade/Woodland:years since NS NS 0.0187 0.0078

Table 17 All harvest types, ESDs, years since harvest and interaction terms thereof with significant relationships with percent cover or frequency of Sanguinaria canadensis. Sanguinaria canadensis Response cover frequency Predictor Est. St. Error Est. St. Error (Intercept) -2.0493 0.7730 -0.8829 0.2771 Chert Upland Woodland 2.7404 1.3307 NS NS

50 Table 18 All harvest types, ESDs, years since harvest and interaction terms thereof with significant relationships with percent cover or frequency of Tephrosia virginiana. Tephrosia virginiana Response cover frequency Predictor Est. St. Error Est. St. Error (Intercept) NS NS -1.1590 0.3371 Chert Protected Backslope Forest:years since 0.9148 0.4338 NS NS Chert Protected Backslope Forest:years since:No Harvest -0.8960 0.4374 NS NS Dry Igneous Exposed Backslope Woodland NS NS 0.6438 0.1275 Dry Igneous Exposed Backslope Woodland:years since 0.0667 0.0298 NS NS Fragipan Upland Woodland 3.4990 1.1170 NS NS Low-base Sandstone Exposed Backslope Woodland:Group Selection -11.8300 5.0010 NS NS Shallow Sandstone Upland Glade/Woodland:years since NS NS -0.0171 0.0059

Significant positive relationships between ESDs, harvest types and years since harvest with percent cover and frequency of the study species are summarized. (Table 9, Table 10, Table 11, Table 12,

Table 13, Table 14, Table 15, Table 16,Table 17 and Table 18). These relationships point to locations and time periods which have passively accumulated these species, without intentional cultivation. If the combination of these factors occurs on the Ozark Highland Plateau within a forest that is under silvicultural management, forest farmers could use these relationships as a guide to integrating forest farming into the management plan. For every species, the harvest types, years since harvest, ESDs and interaction terms between these that had solely positive relationships with percent cover and frequency are examined.

A. racemosa percent cover was positively related to Single-Tree Selection (P<0.0001) and

Intermediate Thin (P=0.0179) (Table 9). Percent cover was positively related to the three way interactions of Chert Protected Backslope Forest, years since harvest and both Single-Tree Selection (P=0.0199) and No

Harvest (P=0.0402). The interaction of ESD Dry Footslope Forest and years since harvest was positively related (P=0.0453) to A. racemosa cover.

Apocynum cannabinum frequency was positively related to the three-way interactions of ESD

Chert Protected Backslope Forest with years since harvest and No Harvest (P=0.0492), Single-Tree

Selection (P=0.0012) and Intermediate Thin (P=0.0436) (Table 10). Frequency was positively related to the interaction of Sandy/Gravelly Floodplain Forest and years since harvest (P=0.0134). Frequency was

51 positively related to the three way interaction of ESD Low-Base Chert Protected Backslope Woodland,

Single-Tree Selection and years since harvest (P < 0.0001).

Aristolochia serpentaria cover was positively related to the ESD Loamy Terrace Forest

(P=0.0216) (Table 11). Frequency was positively related to the three way interaction of ESD Dry Footslope

Forest, Intermediate Thin and years since harvest (P=0.0348). Frequency was also positively related to interactions of years since harvest with the following ESDs: Sandy/Gravelly Floodplain Forest (P=0.0497),

Shallow Dolomite Upland Glade/Woodland (P=0.0256), and Wet Footslope Forest (P=0.0121).

Dioscorea quaternata cover is positively related to Group Selection (P=0.0333), as well as the three way interactions of Group Selection, years since harvest and the following ESDs: Chert Exposed

Backslope Woodland (P=0.0311), and Chert Protected Backslope Forest (P=0.0096) and Low-Base Chert

Protected Backslope Woodland (P=0.0103) (Table 12). Also noteworthy is the positive relationship between cover and ESD Wet Footslope Forest with years since (P=0.0231).

Echinacea simulata had only one significant relationship with any predictor variable, that was a positive relationship between cover and Single-Tree Selection (P=0.0468) (Table 13).

Geranium maculatum cover was positvely related to the interaction of ESD Low-Base Chert

Protected Backslope Woodland and Single-Tree Selection (P=0.0487) (Table 14). G. maculatum frequency was positively related to the interaction of ESD Chert Exposed Backslope Woodland, Group Selection and years since (P=0.0167).

Hydrastis canadensis cover was positively related to six ESD: Chert Dolomite Protected

Backslope Forest (P=0.0334), Chert Protected Backslope Forest (P=0.0231), Dry Footslope Forest

(P=0.0388), Low-Base Chert Protected Backslope Woodland (P=0.0124), Sandy/Gravelly Floodplain

Forest (P=0.0179) and Shallow Dolomite Upland Glade/Woodland (P=0.0179) but negatively related to the interactions of the same ESDs with years since harvest (Table 15). Percent cover of H. canadensis was highest in plots receiving No Harvest and the next highest was Group Selection (Figure 37).

Parthenium integrifolium frequency was positively related to No Harvest (P=0.0320) (Table 16).

Frequency was positively related to the interactions of years since harvest with the ESDs Shallow Dolomite

Upland Glade/Woodland (P=0.0611) and Low-base Sandstone Protected Backslope Woodland (P=0.0179).

52

Mean percent cover of P. integrifolium largely followed the gradient of harvest intensity, with the highest cover corresponding to harvest type with the lowest mean residual stocking (Figure 38).

Neither percent cover or frequency of P. peltatum had significant positive relationships with ESDs or harvest types. Mean percent cover was highest in plots receiving the Single-Tree Selection harvest type, while mean frequency of P. peltatum was highest in No Treatment plots (Figure 39).

Sanguinaria canadensis cover was positively related to Chert Upland Woodland (P=0.0424)

(Table 17). Mean percent cover and frequency of S. canadensis were highest in plots receiving no harvest

(Figure 40).

Significant positive relationships were found between T. virginiana cover and the ESD Fragipan

Upland Woodland (P=0.0021), as well as the interaction of years since harvest with ESD Dry Igneous

Exposed Backslope Woodland (P=0.0255) (Table 18). Frequency was positively related to the ESD Dry

Igneous Exposed Backslope Woodland (P < 0.0001). Mean untransformed cover of T. virginiana was highest in Clearcut and Group Selection plots, while mean untransformed frequency was highest in plots receiving the Clearcut or Single-Tree Selection (Figure 41)

DISCUSSION A previous study conducted at MOFEP revealed a general trend of increasing percent cover of forb species, measured as an aggregate, with increasing harvest intensity (Zenner et al., 2006). The results presented here contradict that general trend in some cases, and match it in others. The following species matched that trend: A. cannabinum, A. serpentaria, D. quaternata, P. integrifolium and T. virginiana.

Conversely, some species contradicted that trend and supported the hypothesis of this study. Uneven-aged management harvest types were significantly positively related to percent cover and/or frequency of certain species. In this case, those species were A. racemosa, E. simulata, G. maculatum, H. canadensis, P. peltatum and S. canadensis. Positive relationships between Single-Tree Selection and Group Selection with

NTFP responses were common.

The percent cover and/or plot frequency of some study species were positively related to mesic sites. H. canadensis occurred almost exclusively on Dry Footslope Forest and several protected backslope

ESDs, and S. canadensis grew mostly on Dry Footslope Forest and Sandy/Gravelly Floodplain Forest

ESDs, which are all relatively mesic sites- dry here being a relative term, compared with fens that 53

otherwise occupy this landscape position. Similarly, there were positive relationships between A. serpentaria’s cover with protected backslope ESDs, and negative relationships with exposed backslope

ESDs. Compare these findings with those of a southeast Ohio study, in which A. serpentaria had the highest coverage on intermediate aspects (Small and McCarthy, 2010). The Ozark Highland Plateau constitutes the far western part of the range of A. serpentaria and many of this study’s other species, and is more xeric than Appalachian Ohio. In a drier region, a species might be more limited in cover except on the most mesic sites. Many understory perennial herbs enter dormancy in response to physiological stress from drought, and this response is particularly common for herbs growing in dry, calcareous soils (Reintal et al.,

2010), conditions typical of soils on exposed aspects in the study area (Shifley and Brookshire, 2000a). If droughts become more common due to climate change, this may impose additional stress on NTFP species, and the effects of drought on NTFPs may need to be explored (McGraw et al., 2013; Souther and McGraw,

2014). However, mesic ESDs did not support the highest percent cover or frequency of all study species.

For instance, E. simulata occurred exclusively on exposed backslope woodland and upland glade ESDs, which are particularly xeric.

Multispecies assemblages of forbs, such as the NTFP species in this study, can have shifting competitive hierarchies along gradients of moisture and light levels (Fortner and Weltzin, 2007). This means that at sites with low soil moisture levels, certain species may be able to compete for space effectively, while at higher soil moisture levels, different species may be more competitive. For example,

H. canadensis and S. canadensis were more abundant on mesic, low light conditions, while T. virginiana had significantly higher percent cover and frequency on more xeric sites, and had the highest mean cover value in clearcuts (Figure 41), which concurrently are associated with low percent stocking of overstory species and therefore higher PAR available at the ground level. This shifting competitive hierarchy along the hypothesized resource gradient might also explain significant relationships between A. serpentaria frequency and ESDs and harvest types (Table 11). A. serpentaria frequency was positively related to ESD

Fragipan Upland Woodland, but negatively related to the interaction of this ESD with No Harvest. Some amount of harvest was good, but not necessarily a Clearcut. This can be seen again with A. serpentaria frequency being negatively related to Dry Footslope Forest, but positively related to this ESD’s interaction

54 with years since and Intermediate Thin. Increased competition from other herbs in higher light areas could become less pronounced over time after a canopy thinning, assuming adequate moisture stripping the most xeric-adapted species of their competitive advantage. One Iowa woodland study found that competition for light with pre-existing ground flora had no effect on the establishment of re-introduced herbs under an undisturbed overstory canopy (Brudvig et al., 2011), but without the gradient of light and other effects from timber harvests as treatments this may not be an applicable finding.

Many slow growing perennial forbs, which include most of the NTFPs in this study, have limited seed dispersal capabilities (Glasgow and Matlack, 2007). However, slow growing perennial forbs in the

Ozarks have been observed to increase their cover in response to increases in available light from timber harvest (Zenner et al., 2006). This seed dispersal limitation might translate into a faster localized percent cover response, rather than a frequency response, as leaf area increases without the germination of new seeds. Moderately increased light can translate into both higher seed production but also poorer seed germination conditions (Wagner and McGraw, 2013). The lack of significant increases in frequency of H. canadensis in this study could be related to this species’ reliance on clonal reproduction (Sinclair et al.,

2005), which would likely show up in this data as an increase in cover rather than frequency. H. canadensis has been shown in other studies to increase size and seed production but not germination in response to canopy gap formation (Sinclair and Catling, 2004). This may reflect the regeneration niche concept, stating that adult individuals can endure a wider range of environmental conditions than seedlings can recruit into

(Albrecht and McCarthy, 2009). If conditions for NTFP seed production are improved by administering a harvest type that adversely affects conditions required for germination, this could afford enhanced opportunities for artificial regeneration of NTFPs in a forest farming system on similar sites that have not had the canopy disturbed. After NTFP seedlings are well-established, timber may be harvested from the site in a way that enhances NTFP seed production without killing parent plants. Such temporary gains in seed production could be leveraged to expand the spatial extent of NTFP populations of a forest to all potentially productive sites.

Some NTFP responses had negative relationships with harvest types but positive relationships with interactions of the same harvest types with years since harvest. This may be accounted for by soil

55

disturbance-related mortality by logging equipment (Chandler and McGraw, 2015; Demir et al., 2007). It was shown in a previous study at MOFEP that uneven-age harvest types resulted in significantly (P < 0.05) more damage to saleable trees than even age harvest types, when special considerations were not taken to minimize the impact of logging equipment from selection cuts (Dwyer et al., 2004b); and the same could be true of NTFP populations. Skidder damage may also explain why interactions of ESD with years since treatment were positively related to NTFP responses, while the relationships with the same ESD by itself were sometimes negative. In the years after disturbance, populations may have been recovering, and in plots receiving no harvest, the understory may still be in recovery from the effects of pre-MOFEP timber harvests due to high rates of erosion from historical exploitative timber harvest(Jacobson and Primm,

1997). Disturbance from timber harvest that extirpates a species locally puts rare species at greater risk of extirpation from a region (Belote et al., 2009). In a forest farming scenario, timber harvest may be conducted to minimize damage to NTFPs as well as to timber species.

Despite transformations of the data to correct for normality, some distributions of frequency and cover were kurtotic to the left. The frequency data at the low end of the range may be suffering in quality from being rounded together as one out of 16. Measuring stems per quadrat would change the frequency data into a more precise density or abundance metric, which might be more informative and correct the frequency measure’s skewness to the left. Percent cover data at the low end of the range are probably suffering in quality from being rounded together in the categories of “0.1%” and “1”. More precise measurements of cover at the low end of the range, particularly 0-1.5%, might allow a normal distribution curve for these herbs, many of which are quite small in stature. For instance, the mean cover of A. serpentaria was 0.42%, or 42 cm2 (Table 6). Recognizing this limitation, linear mixed-effect models still convey general trends in size and abundance of these species as they relate to site and harvest types.

In conclusion, significant positive relationships were found between percent cover and frequency with ESDs and uneven-aged management harvest types. ESD-based models were useful for predicting frequency and percent cover of these species. The Clearcut harvest type had no significant relationships with either percent cover or frequency. Together, the ESD framework combined with sound silvicultural management can be used for co-management of NTFPs and timber.

56

Prospective forest farmers may wish to consider the following lessons. The combination of site type and species identity can determine which timber harvest types are compatible with the goal of expanding populations of that herb. For instance, A. racemosa had significantly higher percent cover values in plots receiving the Group Selection harvest type than with other harvest types, and this positive relationship held true with the interaction of Chert Protected Backslope Forest, years since harvest and

Group Selection as well. If one were to manage wild A. racemosa populations in southeastern Missouri, the combination of this management and site type should be conducive to that population expanding. Forest farming practitioners are advised to focus efforts only on such site and treatment combinations that naturally support their chosen crop species, as recorded in Table 9,

Table 10, Table 11, Table 12, Table 13, Table 14, Table 15, Table 16, Table 17 and Table 18.

With a specific subset of locations selected for NTFP cultivation, and compatible harvest type(s) in mind, the forest farmer may increase NTFP success through cooperating with the forester, logger and skidder operator (Dwyer et al., 2004b). Planning as a team can result in limiting the number and spatial extent of roads and skid trails required to remove timber from the site, proactively protecting advance regeneration and NTFPs alike. Additionally, loggers can use directional felling to avoid destroying NTFP patches, and the NTFP specialist can help by actively monitoring sale administration on those days which logging operations are conducted at the forest farming site.

ACKNOWLEDGEMENTS

We thank the University of Missouri Center for Agroforestry and the Agricultural Research

Service for funding this project, Elizabeth Olson at the Missouri Department of Conservation for their collaboration, and the public for funding holistic natural resource management research.

57 CHAPTER 4: EFFECTS OF PRESCRIBED FIRE ON MEDICINAL HERBS IN THE MISSOURI OZARKS

Abstract: Prescribed fire may lead to increases in size and abundance of woodland perennial herbs. This study evaluated relationships between landscape-scale prescribed fire with percent cover and frequency of eight herbaceous, non-timber forest product species: Actaea racemosa L., Apocynum cannabinum L., Aristolochia serpentaria L., Dioscorea quaternata J. F. Gmel., Geranium maculatum L.,

Parthenium integrifolium L., Sanguinaria canadensis L. and Tephrosia virginiana (L.) Pers. Relationships were examined between plants and ecological site descriptions, which account for landscape position, soil series and depth, slope, aspect and the overstory vegetation community of a site. Additional relationships were explored between these species with the number of fires and years since the experiment’s beginning.

Percent cover and frequency of some herb species was significantly (P < 0.05) related to the number of fires and/or time since beginning of experiment, as well as to particular ecological site types. For instance, a generalized positive trend in Parthenium integrifolium cover was observed, a trend that was significantly pronounced on the “Chert Dolomite Exposed Backslope Woodland” site type- despite there being a negative relationship between individuals of this species’ size and the number of fires. These results may be used to identify forest farming sites and to tailor burn plans to stimulate growth of these species.

Keywords: ecological site descriptions; forest farming; medicinal herbs; non-timber forest product; Ozark Highland Plateau, USA; prescribed fire.

INTRODUCTION Regular, low-severity prescribed fire has been shown to benefit perennial herbs in multiple ways.

Depending on the frequency of prescribed burning, significant reduction in overstory and/or midstory woody species canopy can be expected (Knapp et al., 2015). A recent study (Hanberry et al., 2014c) suggests that Missouri forestland canopies have densified (trees/hectare) by a factor of 2.3 and become more uniform in the time between American settlement in the early to mid-1800’s, and the early 2000’s; this increase in canopy density coincided with a period of fire suppression starting in the 1940’s (Shang et al., 2007). Canopy structure can be a dominant factor affecting ground level flora; the relationship between understory herb cover with basal area and density of trees is significantly (P < 0.05) negative (Gilliam and 58

Turrill, 1993). Light available to ground-level forest herbs is primarily controlled by canopy closure

(Blizzard et al., 2013). Thinning the canopy from below increases available light, which can in turn, be positively related with NTFP herb cover (Clason et al., 2008b) and flowering rates (Mooney et al., 2015).

Additionally, maintaining a consistent, low-severity prescribed fire regime in Midwest oak woodlands may increase available soil moisture and enhance available soil nutrients stores; these soil changes are positively related with native herbaceous species richness (Scharenbroch et al., 2012). Due to increased soil moisture and nutrients, and reduced overstory and midstory cover, fire-related decreases in competition for light, growing space, nutrients and water may favor perennial herbs. Additionally, after the combustion of the litter layer, opportunities for seed-to-ground contact can increase and therefore enhance seed germination

(Maret and Wilson, 2005). Destruction of the aerial portion of the ground flora through disturbances such as prescribed fire or grazing can support increased prominence of perennial herbs in community assemblages, due to an ability to re-sprout from undamaged below-ground organs and efficiently re- colonize and compete for available growing space (Klimešová and Klimeš, 2003). One study on two perennial herbs in a southeastern US pine savannah showed that prescribed fire was significantly (P < 0.05) positively related to seedling establishment, possibly by reducing competition for growing space (Barker and Williamson, 1988). Contemporary efforts at ecological restoration in the Ozarks often employ prescribed fire (Ladd, 2014), and understory forbs have responded positively to prescribed fire treatments by an increase in cover, abundance and diversity (Lettow et al., 2014; Maginel et al., 2016b; McMurry et al., 2007).

Forest farming is an agroforestry system in which crops are cultivated beneath the canopy of a forest or woodland (Chamberlain, 2013; Naud et al., 2010); understory crops are classified as non-timber forest products (NTFPs). Wild populations of certain NTFPs in the eastern United States are shrinking to levels indicative of being threatened, as a result of overharvesting (Farrington et al., 2009b; Souther and

McGraw, 2014). Deliberate cultivation for the market has been suggested as a conservation measure for these species (Burkhart, 2011b), but the cost of labor to actively tend NTFP plantings can be prohibitively expensive (Burkhart and Jacobson, 2009b). However, forest farming can be conducted across a scale of management intensity. Multi-purpose management that deliberately supports NTFP populations for harvest

59

is a kind of extensive forest farming. One example of this can be intentional timber stand improvement activities in order to passively enhance existing NTFP populations (Chamberlain et al., 2009). Another example of this extensive approach to forest farming is to deliberately encourage the proliferation of

NTFPs for commercial harvest through prescribed fire management (Vander Kloet and Pither, 2000;

Vance, 2000; Lake and Long, 2014). In some places, NTFP proliferation has been a welcome, yet unintended consequence of prescribed fire-related disturbance (Abrahamson, 1999; Berkes and Davidson-

Hunt, 2006, 2006; Elliot and Vose, 2010; Hutchinson et al., 2005; Pilz et al., 2004; Sinclair and Catling,

2004; Storm and Shebitz, 2006). If prescribed fire treatments in the Ozarks benefit wild populations of

NTFP species, perhaps forest farming profits could incentivize this ecological restoration activity. As a precedent, First Nations people in the area are known to have used landscape-scale prescribed fire to modify and maintain specific community assemblages for harvesting NTFPs (Jurney, 2012).

The Natural Resource Conservation Service has developed a typology for classifying terrestrial ecosystems called the Ecological Site Description (ESD) (Nauman et al., 2015). ESDs factor in a site’s landscape position, soil series, steepness class, aspect class and dominant overstory and understory vegetation communities. In this way, the classification can incorporate interactions between soil, climate, flora, fauna, and humans into a decision-making framework for land management (Nauman et al., 2015).

Overstory composition combined with slope position was significantly related to herb species richness at the stand level (Ellum et al., 2010). Aspect can be a dominant factor in determining species composition of ground level flora species assemblages (Jenkins and Parker, 1999), as it is with overstory species composition in the Ozarks (Villwock et al., 2011). Timber harvest-related canopy gaps and soil disturbance can interact with soil type and slope position to have significant impacts on diversity of ground level flora

(Duguid et al., 2013). In a study conducted in the mixed oak forests of southern Ohio, no significant relationships were found between ground level flora species richness and the interaction of burning with the integrated moisture index of a site (Hutchinson et al., 2005). Interactive effects of site type, treatment and time can be different when examined at the species level (Elliot and Vose, 2010). Soil moisture may act as a habitat filter in fire-adapted plant communities (Kirkman et al., 2001) such as the those in the Ozarks.

NTFPs may respond differently to fire depending on habitat characteristics.

60 The Chilton Creek Management Area (CCMA), owned and managed by The Nature Conservancy, is the site of a longitudinal experiment for monitoring the effects of landscape-scale prescribed fire on plant communities in the Missouri Ozarks (Fan et al., 2015). Increasing ground flora species richness and increasing the Pinus echinata Mill. component of the overstory are organizational objectives for this burning (Maginel et al., 2016b). These management objectives are supported by fire-induced tree mortality, correlated with a subsequent increase of PAR in canopy gaps and as well as Pinus echinata recruitment to canopy dominance (King and Muzika, 2014b). This chapter examines trends over time of percent cover and frequency for eight NTFPs. If relationships exist between time since beginning of experiment and number of fires, this means there is a relationship between NTFP size and abundance with the length of time an area has received periodic prescribed fires and the number of fires during that time; if these relationships are positive, forest farming and prescribed fire plans could be coordinated to stimulate NTFP production while meeting the ecological restoration objectives of the fires.

It is hypothesized that percent cover and frequency of each of the study species (Actaea racemosa

L., Apocynum cannabinum L., Aristolochia serpentaria L., Dioscorea quaternata J. F. Gmel., Geranium maculatum L., Parthenium integrifolium L., Sanguinaria canadensis L. and

Tephrosia virginiana (L.) Pers.) will be significantly positively related to particular ESDs, years since beginning of the experiment and the number of fires.

MATERIALS AND METHODS

CCMA’s experimental and plot design was modeled after the Missouri Ozark Forest Ecosystem

Project (MOFEP) (Shifley and Brookshire, 2000b; Fan et al., 2012; Knapp et al., 2014). CCMA is divided into five management units: Chilton East, Chilton North, Chilton South, Kelly North and Kelly South

(Figure 18). Each burn unit received between four and 15 burns, primarily springtime, dormant season burns, out of the 16 year study period. Each plot contains 16 1-m2 quadrats (Figure 19). The species and percent cover of all herbaceous plants growing in quadrats were determined and recorded over the summers of 1997, 2001, 2009 and 2013. These units have been burned at randomly assigned intervals of once every one to four years. Within the CCMA experimental design, randomization occurred at the management unit

61 and plot levels, but not the quadrat level, as the location of quadrats is fixed in the plot design (Dey and

Hartman, 2005).

Cover and frequency data were collected by Justin Thomas and the field staff of the Institute for

Botanical Training. Plants in quadrats were identified to the species level. The estimated cover of each plant in each taxon being studied, across the soil’s surface area in each 1-m2 quadrat was visually estimated to the nearest 1%, or 100 cm2. If a plant’s surface area was estimated to be more than 0.5 percent cover of the quadrat, it was rounded up to 1%, and if its presence was estimated at <0.5%, it was assigned a value of

0.1% on the data collection sheet (10 cm2) by the field crew. Precision at this 10-100 cm2 scale of percent cover estimation was not a priority for The Nature Conservancy’s original goal for CCMA, which is to observe generalized responses of the ground level flora to fire. Eight NTFPs were chosen on the basis of being purchased by Missouri-based American Botanicals: Actaea racemosa L. (Ranunculaceae),

Apocynum cannabinum L. (Apocynaceae), Aristolochia serpentaria L. (Aristolochiaceae),

Dioscorea quaternata J.F. Gmel. (Dioscoreaceae), Geranium maculatum L. (Geraniaceae),

Parthenium integrifolium L. (Asteraceae), Sanguinaria canadensis L. (Papaveraceae) or

Tephrosia virginiana (L.) Pers. (Fabaceae).

62 Figure 37

Figure 18 Map of management unit at CCMA.

Figure 38

Figure 19 Plot design at CCMA.

The percent cover (0.1-100%) within individual quadrats or “cover”, and the proportion of quadrats/plot with a given NTFP occurring inside it (1/16-16/16) or “frequency”, served as two response variables. See Figure 19 for plot design at CCMA. These variables were investigated via linear mixed- effects models. An example of the basic model is the frequency formula for A. racemosa 63 ACTRACfreq=lmer(arcsinFREQ~ESD*fires+(1|Unit/Plot),data=cohoshCCMAfreq)

Where:

• “lmer” is a function for fitting linear mixed-effects models using the “lme4” package for the R

software program (Bates et al., 2014);

• “arcsinFREQ” is an arcsin transformation of the frequency for the sake of normality;

• “~” is used to separate the left- and right-hand sides in a model formula ;

• “ESD*fires” is the interaction term between ESD assigned at the plot level and the number of

fires to date at that plot (interaction terms also call up each component variables for analysis as

well);

• “(1|Unit/Plot)” signifies the random-effects terms of management unit and plot;

• “data=cohoshCCMAfreq” refers to the name of the file from which the data were used.

To determine whether fires and ESDs were positively related to the study species, linear mixed- effects modeling was conducted in RStudio. Models were constructed using the lme4 package (Bates et al.,

2014) and lmerTest package (Schaalje et al., 2002). Visualization of the data was conducted with the ggplot2 package (Teutonico, 2015) and in Excel. Random effects were found to have significant impact (P=

0.1 >|t|) when cover or frequency were examined.

Though the numbers of individuals of each medicinal herb species (taxon) in each quadrat were not tallied, the percent cover provided an approximation of the biomass of an individual or a small group of individuals. The frequency approximated the extent of that population at the plot level. Higher frequency could be the result of already existing individuals sexually reproducing across the landscape, with seedlings appearing in quadrats that had previously not been colonized, while decreases could be the result of mortality or dormancy. Significant positive relationships between percent cover or frequency particular

ESDs, number of fires or time since beginning of experiment, indicate especially productive NTFP growing conditions. Percent cover and frequency were usually transformed with a lot, gamma or arcsin transformation to satisfy the assumption of normality. For all analyses in this study, alpha=0.05.

64 RESULTS Summary statistics for explanatory variables are presented (Table 19). Significant relationships were found between the percent cover and frequency of all NTFP species with number of fires, ESDs and/or years since beginning of experiment (Table 20). The names and associated numbers of the ESDs is summarized (Table 21) The mean cover and frequency of each species in each data collection year and

ESD, is summarized (Table 22 and Table 23), which shows

Table 19 Summary statistics (mean, standard deviation, range) and description of explanatory variables in the CCMA study. SD=standard deviation. Units of measurement indicated in brackets. Variable Name Description Mean SD Min. Max. Time since beginning of experiment at time of Time since beginning of experiment sampling (years). 8.30 6.30 0 16 Number of times management units in which plots were nested received fire treatments over course of Fires study (fire treatments) 4.85 4.28 0 15 Frequency (quadrats/plot) of Actaea racemosa in Actaea racemosa frequency plots containing this species. 3.61 2.59 1 13 Frequency (quadrats/plot) of Apocynum Apocynum cannabinum frequency cannabinum in plots containing this species. 1.64 1.35 1 8 Frequency (quadrats/plot) of Aristolochia Aristolochia serpentaria frequency serpentaria in plots containing this species. 1.88 1.12 1 7 Frequency (quadrats/plot) of Dioscorea Dioscorea quaternata frequency quaternata in plots containing this species. 3.23 2.29 1 15 Frequency (quadrats/plot) of Geranium Geranium maculatum frequency maculatum in plots containing this species. 2.60 2.12 1 13 Frequency (quadrats/plot) of Parthenium Parthenium integrifolium frequency integrifolium in plots containing this species. 1.45 0.96 1 6 Frequency (quadrats/plot) of Sanguinaria Sanguinaria canadensis frequency canadensis in plots containing this species. 4.90 3.51 1 11 Frequency (quadrats/plot) of Tephrosia Tephrosia virginiana frequency virginiana in plots containing this species. 1.58 0.88 1 5 Percent cover (0.01-100%) of Actaea racemosa in Actaea racemosa cover plots containing this species. 0.09 0.11 0.10 75.00 Percent cover (0.01-100%) of Apocynum Apocynum cannabinum cover cannabinum in plots containing this species. 0.06 0.07 0.10 41.00 Percent cover (0.01-100%) of Aristolochia Aristolochia serpentaria cover serpentaria in plots containing this species. 0.00 0.01 0.10 5.00 Percent cover (0.01-100%) of Dioscorea Dioscorea quaternata cover quaternata in plots containing this species. 0.02 0.02 0.10 30.00 Percent cover (0.01-100%) of Geranium Geranium maculatum cover maculatum in plots containing this species. 0.01 0.01 0.10 7.00 Percent cover (0.01-100%) of Parthenium Parthenium integrifolium cover integrifoliumin plots containing this species. 0.01 0.02 0.10 20.00 Percent cover (0.01-100%) of Sanguinaria Sanguinaria canadensis cover canadensis in plots containing this species. 0.01 0.01 0.10 8.00 Percent cover (0.01-100%) of Tephrosia virginiana Tephrosia virginiana cover in plots containing this species. 0.03 0.04 0.10 23.00 65

Significant relationships were found between the number of fires and the percent cover of two taxa: Dioscorea quaternata percent cover (P=0.0903) and Sanguinaria canadensis (P=0.0054) (Table 20).

Parthenium integrifolium percent cover was inversely related to the number of fires (P=0.0271). Actaea racemosa cover was positively related to years since beginning of experiment (P < 0.0001). Significant positively relationships were found between years since beginning of experiment with Apocynum cannabinum percent cover (P=0.0002) and frequency (P=0.0065). A significant positive relationship was found between years since beginning of experiment and Aristolochia serpentaria percent cover (P=0.0002), but there was a significant negative relationship between years since beginning of experiment and the frequency of A. serpentaria (P=0.0065). A significant positive relationship was found between years since beginning of experiment and Dioscorea quaternata percent cover (P < 0.0001), but there was a significant negative relationship between years since beginning of experiment and the frequency of D. quaternata

(P=0.0081). There was a significant positive relationship was found between years since beginning of experiment and Geraneum maculatum percent cover (P < 0.0001), but there was a significant negative relationship between years since beginning of experiment and the frequency of G. maculatum (P=0.0469).

A significant positive relationship existed between years since beginning of experiment and Parthenium integrifolium percent cover (P < 0.0001). A significant negative relationship existed between years since beginning of experiment and Sanguinaria canadensis percent cover (P=0.0249). Tephrosia virginiana percent cover was significantly positively related (P < 0.0001) to years since beginning of experiment.

Fire-related changes in percent cover and frequency of these NTFPs may reflect fire-altered competitive relationships with other understory species. Increases in PAR at the ground level (such as from fire-induced overstory mortality) are known to shift the relative competitive hierarchy of multispecies assemblages of forbs (Fortner and Weltzin, 2007).

66

Table 20 Guide to significant relationships between NTFPs and ESDs at CCMA. Significant relationships between percent cover and frequency with ESDs, number of years and number of fires. "NS" stands for "not significant", and represents a relationship with a P-value= >0.1. "-" indicates there were no individuals of that species found growing in plots assigned to the corresponding ESD

(Intercept) years fires ESD (1) ESD (2) ESD (3) ESD (4) ESD (5) ESD (6) ESD (7) ESD (8) ESD (9) ESD (10) ESD (11) ESD (12) ESD (13) species response Est Pr(>|t|) Est Pr(>|t|) Est Pr(>|t|) Est Pr(>|t|) Est Pr(>|t|) Est Pr(>|t|) Est Pr(>|t|) Est Pr(>|t|) Est Pr(>|t|) Est Pr(>|t|) Est Pr(>|t|) Est Pr(>|t|) Est Pr(>|t|) Est Pr(>|t|) Est Pr(>|t|) Est Pr(>|t|) Actaea cover -1.7963 <0.0001 0.0313 <0.0001 NS NS NS NS - - - - NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS racemosa frequency NS NS NS NS NS NS NS NS - - - - NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS Apocynum cover -3.7572 <0.0001 0.0491 0.0002 NS NS - - 1.9228 0.0022 1.7886 0.0035 - - 1.5444 0.0180 1.5398 0.0112 1.6369 0.0070 2.0822 0.0009 - - 2.5470 0.0025 1.5117 0.0178 1.7315 0.0118 - - cannabinum frequency NS NS 0.0109 0.0065 NS NS - - - - -0.6945 0.0039 - - -0.9613 0.0003 -0.8211 0.0006 -0.7953 0.0008 -0.9382 0.0003 - - -1.0280 0.0024 -0.9040 0.0006 -0.8996 0.0014 - - Aristolochia cover NS NS 0.0551 0.0045 NS NS NS NS NS NS NS NS NS NS NS NS 0.4127 0.0620 0.4165 0.0573 0.3991 0.0829 NS NS NS NS NS NS NS NS NS NS serpentaria frequency 0.0982 0.0606 -0.0039 0.0001 NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS Dioscorea cover NS NS 0.0457 <0.0001 0.0127 0.0903 NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS quaternata frequency 0.3461 0.0068 -0.0036 0.0081 NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS Geranium cover -2.7235 <0.0001 0.0512 <0.0001 NS NS NS NS - - NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS maculatum frequency -0.5883 0.0322 -0.0130 0.0469 NS NS NS NS - - NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS Parthenium cover -2.9914 <0.0001 0.0925 <0.0001 -0.0476 0.0271 - - NS NS 1.0735 0.0163 - - NS NS NS NS NS NS NS NS ------NS NS - - integrifolium frequency NS NS NS NS NS NS - - 0.2365 0.0001 NS NS - - 0.1028 0.0391 NS NS NS NS NS NS ------0.1068 0.0571 - - Sanguinaria cover NS NS -0.0473 0.0249 0.1082 0.0054 ------NS NS ------NS NS - - - - NS NS NS NS canadensis frequency NS NS NS NS NS NS ------NS NS ------NS NS - - - - NS NS NS NS Tephrosia cover -3.0851 <0.0001 0.0575 <0.0001 NS NS - - 1.2469 0.0636 NS NS - - 1.1015 0.0642 NS NS 0.9297 0.0966 NS NS ------67 virginiana frequency NS NS NS NS NS NS - - NS NS NS NS - - NS NS NS NS NS NS NS NS ------

ESD (14) ESD (15) ESD (16) ESD (17) ESD (18) ESD (19) ESD (20) ESD (21) ESD (22) ESD (23) ESD (24) ESD (25) ESD (26) ESD (27) ESD (28) ESD (29) species response Est Pr(>|t|) Est Pr(>|t|) Est Pr(>|t|) Est Pr(>|t|) Est Pr(>|t|) Est Pr(>|t|) Est Pr(>|t|) Est Pr(>|t|) Est Pr(>|t|) Est Pr(>|t|) Est Pr(>|t|) Est Pr(>|t|) Est Pr(>|t|) Est Pr(>|t|) Est Pr(>|t|) Est Pr(>|t|) Actaea cover NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS -0.8176 0.0814 - - NS NS -0.8523 0.0760 NS NS 0.6742 0.0992 racemosa frequency NS NS NS NS 0.5715 0.0071 NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS - - 0.6546 0.0023 NS NS NS NS NS NS Apocynum cover 1.8080 0.0044 1.7875 0.0063 - - 1.7841 0.0035 1.4829 0.0162 1.6257 0.0076 1.9463 0.0022 1.6670 0.0078 1.6674 0.0178 NS NS 1.7888 0.0040 1.5993 0.0185 - - 1.5024 0.0333 - - - - cannabinum frequency -0.7527 0.0059 -0.6712 0.0318 - - -0.8467 0.0005 -0.9255 0.0002 -0.8727 0.0003 -0.8741 0.0009 -0.8042 0.0018 -1.0180 0.0004 NS NS -0.7276 0.0039 -0.7729 0.0136 - - -1.0100 0.0021 - - - - Aristolochia cover NS NS NS NS 0.5546 0.0692 0.3779 0.0857 NS NS NS NS 0.3930 0.0888 0.3794 0.0947 0.4509 0.0506 NS NS NS NS NS NS NS NS 0.5758 0.0238 NS NS 0.7694 0.0097 serpentaria frequency NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS 0.1140 0.0943 NS NS Dioscorea cover NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS -0.6368 0.0396 NS NS quaternata frequency NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS Geranium cover NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS - - NS NS NS NS - - NS NS maculatum frequency NS NS NS NS 0.6714 0.0766 NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS - - NS NS NS NS - - NS NS Parthenium cover - - NS NS - - NS NS NS NS NS NS NS NS NS NS - - NS NS NS NS ------integrifolium frequency - - NS NS - - NS NS NS NS NS NS NS NS NS NS - - NS NS NS NS ------Sanguinaria cover NS NS ------NS NS NS NS NS NS NS NS ------NS NS canadensis frequency NS NS ------NS NS NS NS NS NS NS NS ------NS NS Tephrosia cover NS NS - - - - NS NS NS NS NS NS NS NS NS NS - - - - NS NS ------virginiana frequency NS NS - - - - NS NS NS NS NS NS NS NS NS NS - - - - NS NS ------

Table 21 Name of all ESDs at CCMA, with number of plots per ESD.

Number ESD Plots 1 Calcareous Dolomite Exposed Backslope Woodland 1 2 Calcareous Dolomite Protected Backslope Forest 2 3 Calcareous Dolomite Upland Woodland 2 4 Chert Dolomite Exposed Backslope Woodland 9 5 Chert Dolomite Protected Backslope Forest 2 6 Chert Dolomite Upland Woodland 5 7 Chert Exposed Backslope Woodland 24 8 Chert Protected Backslope Forest 30 9 Chert Upland Woodland 6 10 Dolomite Protected Cliff 1 11 Dry Footslope Forest 1 12 Fragipan Upland Woodland 7 13 Gravelly/Loamy Upland Drainageway Forest 9 14 Loamy Dolomite Protected Backslope Forest 1 15 Loamy Protected Backslope Forest 4 16 Loamy Terrace Forest 4 17 Loamy Upland Woodland 1 18 Low-base Chert Exposed Backslope Woodland 41 19 Low-base Chert Protected Backslope Woodland 31 20 Low-base Chert Upland Woodland 26 21 Low-base Sandstone Exposed Backslope Woodland 11 22 Low-base Sandstone Protected Backslope Woodland 10 23 Low-base Sandstone Upland Woodland 6 24 Sandy/Gravelly Floodplain Forest 3 25 Shallow Dolomite Upland Glade/Woodland 6 26 Shallow Sandstone Upland Glade/Woodland 1 27 Wet Floodplain Step Forest 1 28 Wet Footslope Forest 2 29 Wet Terrace Forest 1 30 Wet Upland Drainageway Forest 2

68 that species found growing " there indicates that no individuals of were - . " . 20 Table frequency by frequency ESD at by CCMA year. This table is useful for interpreting the significant relationships

Mean Mean percent cover NTFPsof and by between NTFP percentand coverESDs as summarized in plotsin assigned tocorresponding the ESD.

22 Table

69 uency and uency " indicates that " there indicates that were no individualsthat of species found growing in plots assigned to the - . " . 20 Table

ized in Mean Mean frequency NTFPsof by ESD at CCMA. This table useful is for interpreting the significant relationships between NTFP freq ESDs as summar corresponding ESD.

23 Table

70 DISCUSSION Because of the prevalent usage prescribed fire in Ozark woodlands, foresters managing wild populations of NTFP plant species can benefit from knowing how these taxons performed under conditions related to repeated, low-intensity, dormant season prescribed fires at CCMA. Since all management units received fire treatments, no data were available from a fire-excluded control treatment. However, significant positive relationships between years since beginning of experiment and NTFP percent cover and frequency still show population trends. Significant positive relationships between NTFP percent cover and frequency with specific ecological site types, all of which were under fire management, reveal specific habitats in which populations of these species proliferated. These positive relationships suggest individuals of these species competed well for space with other ground flora species in those site types, by taking up more space within quadrats (percent cover) and across plots (frequency).

Sites that stand out in their ability to sustain populations of NTFP under fire management in the

Ozarks may be easily identified from the results of the regression analyses, by looking at site and species relationships that had positive estimate of the coefficient, and p-values of less than 0.05 (Table 20).

Percent cover of the following species was significantly positively related to years since the beginning of the study: Actaea racmosa, Aristolochia serpentaria, Apocynum cannabinum, Dioscorea quaternata and Tephrosia virginiana. In addition, for A. cannabinum and D. quaternata, frequency was positively related to years since beginning of relationship. Relationships with particular ESDs aside, these species became larger and in the latter subset, more abundant over the course of prescribed fire- management at the study site. Contrast this with the negative relationships between times since beginning of experiment and frequency of Geranium maculatum (P=0.0469) and Parthenium integrifolium

(P=0.0271), even as positive relationships were found between years since beginning of experiment with the percent cover of each species. Finally, Sanguinaria canadensis percent cover was negatively related to years since beginning of experiment (P=0.0249). These significant but seemingly contrasting relationships, a mix of positive and negative, illustrate how some NTFP species did better than others during the 16 years of fire treatment at CCMA. Select significant relationships between percent cover of Actaea racemosa and

Tephrosia virginiana with ESDs are reported (Figure 20 and Figure 21).

71

Actaea racemosa is a large herbaceous perennial that grows across a wide range of habitats in association with woody vegetation. Frequency of A. racemosa was significantly positively related to ESD

Loamy Upland Woodland (P=0.0071) and Floodplain Step Forest (P=0.0028): (Table 20) It might be that soils in these mesic ESDs were acting as a fire-mediated habitat filter, by retaining more moisture and cooling the peak temperatures of the O and A soil horizons during fires (Kirkman et al., 2001). A. racemosa rhizomes tend to grow several inches below the soil surface (Kaur et al., 2013), which may be deep enough to prevent fire-induced mortality.

Apocynum cannabinum is a tall, erect perennial herb that is widely distributed across the US. A. cannabinum percent cover was related to 20 of the 29 ESDs in this study (Table 20). A. cannabinum reproduces asexually in expanding patches through creeping, fibrous roots, so unless the root network was killed by a fire it could be well-positioned to produce new aerial growth following spring burns, increasing in cover over time. A. cannabinum produces short-lived, non-dormant seeds that are generated in greater numbers, especially when there is little competition from other vegetation (Webster and Cardina, 1999).

Because of the ability to quickly sexually reproduce, the positive relationship between frequency and time is not surprising.

Aristolochia serpentaria is a small perennial woodland and forest herb. Percent cover was significantly positively related to ESDs Wet Footslope Forest (P=0.0238) and Wet Upland Drainageway

Forest (P=0.0097) and frequency was positively related to one ESD (Table 20). Cover was positively related to years since beginning of experiment. A. serpentaria produces flowers in late spring at or near the soil surface level which are often hidden underneath leaf litter (Dávalos et al., 2014); burning in early spring may aid in seed dispersal.

Dioscorea quaternata is a small vine growing from a tuber. Cover was significantly negatively related to ESD Wet Terrace Forest (P=0.0396) (Table 20). Years since beginning of the experiment was positively related to cover (P < 0.0001) and frequency (P < 0.0081).

Geranium maculatum is a small perennial woodland herb. There was a positive significant relationship between cover of this species with time since beginning of experiment (P < 0.0001), but the relationship between G. maculatum frequency and time since beginning of experiment was negative

72

(P=0.0469) (Table 20). The aerial portion of G. maculatum forms from a rhizome sitting on or just below the soil surface (Ågren and Willson, 1992), which may leave root crowns vulnerable to fire-induced mortality.

Figure 20

Figure 20 Figure 9a Graph showing mean percent of Actaea racemosa, by year since beginning of experiment, on significantly (P < 0.05) related ESDs. A. racemosa cover had a positive relationship with two ESDs over time, and a negative relationship with one ESD.

73 Figure 40

Figure 21 Figure 9b. Graph showing mean percent of Tephrosia virginiana, by year since beginning of experiment, on significantly (P < 0.05) related ESDs. T. virginiana cover had a positive relationship with two ESDs over time and a negative relationship with one ESD. Parthenium integrifolium is a small upright perennial forb that grows in prairies, savannas and woodlands. P. integrifolium percent cover was significantly positively related to ESD Chert Dolomite

Exposed Backslope Woodland (P=0.0163), and frequency was positively related to ESD Chert Dolomite

Upland Woodland (P=0.0391) (Table 20). Percent cover was significantly positively (P=0.0391) related to years since beginning of experiment, but negatively related to the number of fires (P=0.0271)

Sanguinaria canadensis is a short, herbaceous perennial species that is known to grow in rich, well-drained forest soils (Furgurson et al., 2012). This species displayed no significant relationships between any ESDs and percent cover or frequency (Table 20). The S. canadensis cover model showed a positive relationship with number of fires, but a negative relationship with years since the beginning of the experiment. These seemingly contradictory results make it difficult to discern how representatives of this taxon are responding to fire treatments. Hypotheses for the recorded/observed responses of plants of this taxon to time since beginning of experiment and fire may be formulated by examining the species’ morphological traits and reproductive history, as well as details of the fire prescription at CCMA. S. canadensis reproduces asexually and expands its population through shoots emerging from stout, shallow 74 rhizomes (Pudlo et al., 1980; Small et al., 2011). Flowering occurs in early spring, between March and May

(Schemske et al., 1978). Fires at CCMA are typically performed in March or April and may have destroyed both rametes and flowers during a critical phase of the growing season. S. canadensis reproduces sexually and produces "two year seeds” that require a period of warm, then cold, then warm again to germinate

(Kaur et al., 2013). Increasing the number of years between fires, and changing the timing of the burn to earlier in the season when S. canadensis has not yet leafed out might yield different results. Avoiding the destruction of the flowering aerial tissues of this spring ephemeral might require burning only before

March, or not at all. The ESDs in which S. canadensis was recorded were the relatively uncommon

Gravelly/Loamy Upland Drainageway Forest (ESD 13) and Loamy Dolomite Protected Backslope Forest

(ESD 14), which typified the habitat of nine and one plots in this study, respectively. Restricting late-spring burning in these limited locations might prove helpful for this species.

Tephrosia virginiana is an upright, perennial, leguminous forb. The cover of this species was positively related to two ESDs and negatively related to one (Figure 21). Neither T. virginiana cover or frequency had significant relationships with any ESDs at alpha=0.05. Like Parthenium integrifolium, , T. virginiana is considered an indicator species for open-canopy woodland structure in Missouri (Kinkead,

2013). Other studies have shown T. virginiana to be fire-adapted, exhibiting increased aboveground bio- mass and flowering in response to dormant season burns (Dudley and Lajtha, 1993), and significant (P <

0.05) increases in germination rates following heat treatments (Pitman, 2009).

Despite transformations of the data to correct for normality, some frequency and percent cover distributions still had kurtosis to the left (Figure 22).

75 Figure 41

Figure 22 Distribution of Geranium maculatum frequency at CCMA. For most species, arcsin, gamma or Log10 transformations were applied to cover and frequency values in an effort to satisfy the assumption of normality, but this was not always successful.

If there was further study made of specific herb species from this data set, an alteration to the sampling protocol for future years could be made to enhance the resolution of the data at the lower end of the percent cover and frequency range. To improve the population density metric that frequency is meant to represent, measuring stems per quadrat or plot could change the frequency data into a more precise measure. This would likely reveal more detailed stem count values that were all lumped together as having a frequency of (1/16), and would thus offer a correction to the frequency measure’s skewedness to the left in this study. Percent cover data at the low end of the range are probably suffering in quality from being rounded together in the categories of “0.1%” and “1”. More precise measurements of percent cover at the low end of the range, particularly 0-1.5% cover, might allow a normal distribution curve to be seen for these herbs, many of which are small in stature.

There are additional caveats about using these data to make definitive statements about effects of prescribed fire on NTFPs in the Missouri Ozarks. This study’s mean fire interval was relatively low, usually between one and three years, compared to 7.7 years which is estimated to be the historic mean fire interval for the Ozark mountains (Stambaugh and Guyette, 2006). The mean fire interval on this project is 76 generally even shorter than the three to five years typically recommended for maintaining woodland structure in Missouri (Nelson, 2002). Seedlings that germinated on bare soil the year after a fire may have been destroyed by a burn the next year. Historically, the next fire would usually not have come for a few more years. The difference between the fire frequency at CCMA and the historical fire frequency could mean responses to the historic burn regime would be different. This would provide a complementary explanation for why distributions of the frequency and percent cover data were often skewed to the left.

Sexual or vegetative reproduction in response to fire would have been included in percent cover, but often within the same quadrat as the parent plant, therefore not impacting frequency.

In conclusion, significant relationships between percent cover and frequency of NTFPs were found with ESDs, number of fires and years since the beginning of the experiment. The ESD framework may be helpful to forest farmers in identifying potential sites for NTFP cultivation. Most NTFP species had higher percent cover and frequency over time, co-occurring with the re-introduction of prescribed fire.

Only D. quaternata, S. canadensis and T. virginiana showed significant relationships with number of fires.

Managers who cannot burn at the landscape level but want a positive response in size and abundance of

NTFPs could tailor burning to match the boundaries of ESDs that support robust populations of fire- adapted NTFP species

ACKNOWLEDGEMENTS We would like to thank the University of Missouri Center for Agroforestry and the Agricultural

Research Service for funding this project, and TNC Missouri for collaborating.

77 CHAPTER 5: CONCLUSIONS Prescribed fire is a common landscape management tool on wooded, public lands in Missouri, logging is a very common way for landowners to make money from the forest, and consequently canopy gaps of anthropogenic origin are also common. Forest farming will most likely be practiced in the presence of these disturbances, so the relationships with NTFPs and this kind of management are important to determine when designing productive forest farming systems. Forest farming is becoming more commonly practiced in Missouri. Because forests receive prescribed fire management and timber harvest throughout

Missouri and other parts of the world, land managers benefit from understood how prescribed fire, timber harvest and canopy gaps impact NTFP crops.

The common garden experiment in chapter two was established to examine the relationship of growth, mortality and reproduction of four NTFPs with canopy gap treatments. The test was replicated in two central Missouri forest fragments. Both planting sites had a steepness of approximately 30°, and were placed on North-facing backslopes to take advantage of increased soil moisture on this aspect. It was predicted that the height, mortality and reproduction of all study species would be positively related to the canopy gap treatment. The results of the studies indicated that Actaea racemosa, Allium tricoccum and

Collinsonia canadensis grew significantly (P < 0.05) taller in canopy gaps, and A. racemosa flowered at significantly (P < 0.05) higher rates when grown in canopy gaps. However, mortality was significantly (P

< 0.05) higher for all species in canopy gaps. Transplant shock may have been more severe in canopy gaps due to higher soil temperatures and lower soil moisture. While PAR under closed canopy was found to be inadequate for these species to reach their light saturation points, gains in growth rate from planting in canopy gaps may be offset by decreased survival rates. Site selection for medicinal herb cultivation may make opportunistic use of canopy gaps to benefit herb growth, but management to ameliorate against deleterious impacts of the canopy gap treatments on survival may be needed to ensure net gains in productivity of NTFP plantings.

The objective of the study in chapter three was determining how ecosystem site type and timber harvest type are related to percent cover and frequency of NTFPs. It was expected that thinning, single-tree selection and group selection harvest types would be positively related to percent cover and frequency of

NTFPs, and that No Harvest and Clearcut would either have no relationships, or negative relationships,with 78

these species. Protected ESDs were predicted to be positively related to percent cover and frequency of

NTFPs. The results supported these hypotheses: percent cover and frequency of many species studied had significant (P < 0.05) positive relationships with mesic ESDs; percent cover and frequency of Actaea racemosa appears to have been stimulated by the group selection harvest type. However, some species were positively related with no harvest and even-aged harvest types. For instance, Sanguinaria canadensis did significantly (P < 0.05) better in areas receiving no harvest at all, and Parthenium integrifolium response variables were positively and significantly (P < 0.05) related to clearcuts. Foresters with a grasp of these species-specific dynamics could write forest farming recommendations into their management plans.

Which species would be emphasized would be dependent on which ESDs were available for cultivation, current market demand for each, and which harvest types that met forest farming objectives were compatible with other management goals.

The objective of the study in chapter four was to ascertain the relationships between NTFPs and prescribed fire and ecosystem site descriptions over time. The percent cover of all but one study species was positively related to time since beginning of experiment, the exception being S. canadensis percent cover which was positively related to the number of fires since the study’s inception. D. quaternata percent cover was similarly positively related to the number of fires, though P. integrifolium percent cover had a negative relationship with the number of fires. Meanwhile, the frequency values of three species- A. cannabinum, A. serpentaria and D. quaternata- were positively related to age (years since the beginning of the experiment). Frequency of G. maculatum and percent cover of T. virginiana were negatively related to years since the beginning of the experiment. This suggests that prescribed fire is not incompatible with the proliferation of these species. Since prescribed fires in Missouri are typically conducted for purposes of ecological restoration, the results of this study increasing populations of wild populations of NTFPs can occur on land being managed for ecological restoration. Forest farmers planting NTFPs should consider what site types are available, the current market demand for potential NTFP crops, and the harvest types and fire regime that the site will see for the foreseeable future.

79

Chapters three and four show how Ecological Site Description typology successfully predicted plots which supported greater percent cover and frequency of NTFP species. Knowing the recent land use history, managers may tailor forest farming to their management plan using ESD maps.

80 LITERATURE CITED Abrahamson, W. G. (1999). Episodic Reproduction in Two Fire-Prone Palms, Serenoa repens and Sabal

etonia (Palmae). Ecology, 80(1), 100. https://doi.org/10.2307/176982

AFTA. (2015). What is agroforestry? Retrieved May 1, 2015, from http://www.aftaweb.org/about/what-is-

agroforestry.html

Ågren, J., & Willson, M. F. (1992). Determinants of seed production in Geranium maculatum. Oecologia,

92(2), 177–182.

Aikens, M. L., Ellum, D., McKenna, J. J., Kelty, M. J., & Ashton, M. S. (2007). The effects of disturbance

intensity on temporal and spatial patterns of herb colonization in a southern New England mixed-

oak forest. Forest Ecology and Management, 252(1–3), 144–158.

https://doi.org/10.1016/j.foreco.2007.06.031

Albrecht, M. A., & McCarthy, B. C. (2009). Seedling establishment shapes the distribution of shade-

adapted forest herbs across a topographical moisture gradient. Journal of Ecology, 97(5), 1037–

1049. https://doi.org/10.1111/j.1365-2745.2009.01527.x

Alexander, S., Oswalt, S. N., & Emery, M. R. (2011). Nontimber forest products in the United States:

Montreal process indicators as measures of current conditions and sustainability. Retrieved from

http://www.arlis.org/docs/vol1/C/748283785.pdf

Barker, N. G., & Williamson, G. B. (1988). Effects of a Winter Fire on Sarracenia alata and S. psittacina.

American Journal of Botany, 75(1), 138. https://doi.org/10.2307/2443912

Bates, D., Mächler, M., Bolker, B., & Walker, S. (2014). Fitting linear mixed-effects models using lme4.

arXiv Preprint arXiv:1406.5823. Retrieved from http://arxiv.org/abs/1406.5823

Belote, R. T., Sanders, N. J., & Jones, R. H. (2009). Disturbance alters local–regional richness relationships

in Appalachian forests. Ecology, 90(10), 2940–2947.

Berkes, F., & Davidson-Hunt, I. J. (2006). Biodiversity, traditional management systems, and cultural

landscapes: examples from the boreal forest of Canada. International Social Science Journal,

58(187), 35–47.

81 Bernatchez, A., Bussières, J., & Lapointe, L. (2013). Testing fertilizer, gypsum, planting season and

varieties of wild leek (Allium tricoccum) in forest farming system. Agroforestry Systems, 87(5),

977–991. https://doi.org/10.1007/s10457-013-9613-1

Blizzard, E. M., Kabrick, J. M., Dey, D. C., Larsen, D. R., Pallardy, S. G., & Gwaze, D. P. (2013). Light,

canopy closure, and overstory retention in upland Ozark forests. Retrieved from

http://www.treesearch.fs.fed.us/pubs/43478

Brudvig, L. A., Mabry, C. M., & Mottl, L. M. (2011). Dispersal, not Understory Light Competition, Limits

Restoration of Iowa Woodland Understory Herbs. Restoration Ecology, 19(101), 24–31.

https://doi.org/10.1111/j.1526-100X.2010.00675.x

Bruhn, J. N. (2008). Growing shiitake mushrooms in an agroforestry practice. Extension Publications

(MU). Retrieved from https://mospace.umsystem.edu/xmlui/handle/10355/8383

Burkhart, E. P. (2011). “Conservation Through Cultivation”: Economic, Socio-Political and Ecological

Considerations Regarding the Adoption of Ginseng Forest Farming in Pennsylvania. The

Pennsylvania State University. Retrieved from https://etda.libraries.psu.edu/paper/11456/7455

Burkhart, E. P., & Jacobson, M. G. (2009). Transitioning from wild collection to forest cultivation of

indigenous medicinal forest plants in eastern North America is constrained by lack of profitability.

Agroforestry Systems, 76(2), 437–453. https://doi.org/10.1007/s10457-008-9173-y

Chamberlain, J. (2013, April 30). Presentation on Ramps (Allium tricoccum) in Appalachia - eXtension.

Retrieved June 12, 2014, from http://www.extension.org/pages/67942/presentation-on-ramps-

allium-tricoccum-in-appalachia

Chamberlain, J. L., Bush, R. J., Hammett, A. L., & Araman, P. A. (2002). Managing for nontimber

products. Journal of Forestry, 100(1), 8–14.

Chamberlain, J. L., Mitchell, D., Brigham, T., Hobby, T., Zabek, L., & Davis, J. (2009). Forest Farming

Practices. In North American Agroforestry: an integrated science and practice (2nd Edition, p.

50). Madison, WI: American Society of Agronomy.

Chamberlain, J. L., Prisley, S., & McGuffin, M. (2013). Understanding the Relationships Between

American Ginseng Harvest and Hardwood Forests Inventory and Timber Harvest to Improve Co-

82

Management of the Forests of Eastern United States. Journal of Sustainable Forestry, 32(6), 605–

624. https://doi.org/10.1080/10549811.2013.798828

Chandler, J. L., & McGraw, J. B. (2015). Variable effects of timber harvest on the survival, growth, and

reproduction of American ginseng (Panax quinquefolius L.). Forest Ecology and Management,

344, 1–9. https://doi.org/10.1016/j.foreco.2015.02.007

Chazdon, R. L., & Pearcy, R. W. (1991). The importance of sunflecks for forest understory plants.

BioScience, 41(11), 760–766.

Clason, A. J., Lindgren, P. M. F., & Sullivan, T. P. (2008a). Comparison of potential non-timber forest

products in intensively managed young stands and mature/old-growth forests in south-central

British Columbia. Forest Ecology and Management, 256(11), 1897–1909.

https://doi.org/10.1016/j.foreco.2008.07.013

Clason, A. J., Lindgren, P. M. F., & Sullivan, T. P. (2008b). Comparison of potential non-timber forest

products in intensively managed young stands and mature/old-growth forests in south-central

British Columbia. Forest Ecology and Management, 256(11), 1897–1909.

https://doi.org/10.1016/j.foreco.2008.07.013

Collins, B. S., & Pickett, S. T. A. (1988). Response of Herb Layer Cover to Experimental Canopy Gaps.

American Midland Naturalist, 119(2), 282. https://doi.org/10.2307/2425811

Collins, E. P. (2014). Agroforestry : a practical guide to profitable farms and woodlands. New York :

Novinka : Nova Science. Retrieved from http://laurel.lso.missouri.edu/record=b10802106~S1

D’Amato, A. W., Bradford, J. B., Fraver, S., & Palik, B. J. (2011a). Forest management for mitigation and

adaptation to climate change: Insights from long-term silviculture experiments. Forest Ecology

and Management, 262(5), 803–816. https://doi.org/10.1016/j.foreco.2011.05.014

D’Amato, A. W., Bradford, J. B., Fraver, S., & Palik, B. J. (2011b). Forest management for mitigation and

adaptation to climate change: Insights from long-term silviculture experiments. Forest Ecology

and Management, 262(5), 803–816. https://doi.org/10.1016/j.foreco.2011.05.014

83

Dávalos, A., Nuzzo, V., & Blossey, B. (2014). Demographic responses of rare forest plants to multiple

stressors: the role of deer, invasive species and nutrients. Journal of Ecology, 102(5), 1222–1233.

https://doi.org/10.1111/1365-2745.12279

De Beenhouwer, M., Aerts, R., & Honnay, O. (2013). A global meta-analysis of the biodiversity and

ecosystem service benefits of coffee and cacao agroforestry. Agriculture, Ecosystems &

Environment, 175, 1–7. https://doi.org/10.1016/j.agee.2013.05.003

Demir, M., Makineci, E., & Yilmaz, E. (2007). Investigation of timber harvesting impacts on herbaceous

cover, forest floor and surface soil properties on skid road in an oak (Quercus petrea L.) stand.

Building and Environment, 42(3), 1194–1199. https://doi.org/10.1016/j.buildenv.2005.11.008

Dey, D. C., & Hartman, G. (2005). Returning fire to Ozark Highland forest ecosystems: Effects on advance

regeneration. Forest Ecology and Management, 217(1), 37–53.

https://doi.org/10.1016/j.foreco.2005.05.002

Diamond, A. K., & Emery, M. R. (2011). Black Ash (Fraxinus nigra Marsh.): Local Ecological Knowledge

of Site Characteristics and Morphology Associated with Basket–Grade Specimens in New

England (USA) 1. Economic Botany, 65(4), 422–426.

Dress, W. J., & Boerner, R. E. (2004). Patterns of microarthropod abundance in oak-hickory forest

ecosystems in relation to prescribed fire and landscape position. Pedobiologia, 48(1), 1–8.

Dudley, J. L., & Lajtha, K. (1993). The Effects of Prescribed Burning on Nutrient Availability and Primary

Production in Sandplain Grasslands. American Midland Naturalist, 130(2), 286.

https://doi.org/10.2307/2426128

Duguid, M. C., & Ashton, M. S. (2013). A meta-analysis of the effect of forest management for timber on

understory plant species diversity in temperate forests. Forest Ecology and Management, 303, 81–

90. https://doi.org/10.1016/j.foreco.2013.04.009

Duguid, M. C., Frey, B. R., Ellum, D. S., Kelty, M., & Ashton, M. S. (2013). The influence of ground

disturbance and gap position on understory plant diversity in upland forests of southern New

England. Forest Ecology and Management, 303, 148–159.

https://doi.org/10.1016/j.foreco.2013.04.018

84

Dwyer, J. P., Dey, D. C., Walter, W. D., & Jensen, R. G. (2004a). Harvest Impacts in Uneven-Aged and

Even-Aged Missouri Ozark Forests. Northern Journal of Applied Forestry, 21(4), 197–193.

Dwyer, J. P., Dey, D. C., Walter, W. D., & Jensen, R. G. (2004b). Harvest Impacts in Uneven-Aged and

Even-Aged Missouri Ozark Forests. Northern Journal of Applied Forestry, 4(21), 187–193.

Elliot, K. J., & Vose, J. M. (2010). Short-term effects of prescribed fire on mixed oak forests in the

southern Appalachians: vegetation response. Journal of the Torrey Botanical Society, 137(1), 49–

66.

Ellum, D. S., Ashton, M. S., & Siccama, T. G. (2010). Spatial pattern in herb diversity and abundance of

second growth mixed deciduous-evergreen forest of southern New England, USA. Forest Ecology

and Management, 259(8), 1416–1426. https://doi.org/10.1016/j.foreco.2010.01.011

Fan, Z., Ma, Z., Dey, D. C., & Roberts, S. D. (2012). Response of advance reproduction of oaks and

associated species to repeated prescribed fires in upland oak-hickory forests, Missouri. Forest

Ecology and Management, 266, 160–169. https://doi.org/10.1016/j.foreco.2011.08.034

Fan, Z., Yao, Q., Dey, D., Spetich, M., Ezell, A., Shifley, S., … Jensen, R. (2015). Efficacy and Associated

Factors of Even- and Uneven-Aged Management to Promote Oak Regeneration in the Missouri

Ozarks. Forest Science, 61(2), 397–408. https://doi.org/10.5849/forsci.13-069

Farnsworth, E. J., & Ogurcak, D. E. (2006). Biogeography and decline of rare plants in New England:

historical evidence and contemporary monitoring. Ecological Applications, 16(4), 1327–1337.

Farrington, S. J., Muzika, R.-M., Drees, D., & Knight, T. M. (2009a). Interactive Effects of Harvest and

Deer Herbivory on the Population Dynamics of American Ginseng. Conservation Biology, 23(3),

719–728. https://doi.org/10.1111/j.1523-1739.2008.01136.x

Farrington, S. J., Muzika, R.-M., Drees, D., & Knight, T. M. (2009b). Interactive Effects of Harvest and

Deer Herbivory on the Population Dynamics of American Ginseng. Conservation Biology, 23(3),

719–728. https://doi.org/10.1111/j.1523-1739.2008.01136.x

Fortner, A. M., & Weltzin, J. F. (2007). Competitive hierarchy for four common old-field plant species

depends on resource identity and availability. The Journal of the Torrey Botanical Society, 134(2),

166–176.

85

Furgurson, J., Cubbage, F., Sills, E., & Bates, P. (2012). Bloodroot (Sanguinaria canadensis L.) Extent and

Sustainability in Western North Carolina. Open Journal of Forestry, 02(04), 213–218.

https://doi.org/10.4236/ojf.2012.24026

Gillespie, A. R., Yu, C. K., Rathfon, R., & Seifert, J. R. (2006a). Microenvironments for Forest-Farming of

Medicinal Herbs in Varying Forest Structures of the Midwestern USA. Journal of Sustainable

Agriculture, 27(4), 71–96. https://doi.org/10.1300/J064v27n04_06

Gillespie, A. R., Yu, C. K., Rathfon, R., & Seifert, J. R. (2006b). Microenvironments for Forest-Farming of

Medicinal Herbs in Varying Forest Structures of the Midwestern USA. Journal of Sustainable

Agriculture, 27(4), 71–96. https://doi.org/10.1300/J064v27n04_06

Gilliam, F. S., & Turrill, N. L. (1993). Herbaceous Layer Cover and Biomass in a Young Versus a Mature

Stand of a Central Appalachian Hardwood Forest. Bulletin of the Torrey Botanical Club, 120(4),

445. https://doi.org/10.2307/2996749

Gingrich, S. F. (1967). Measuring and Evaluating Stocking and Stand Density in Upland Hardwood Forests

In the Central States. Forest Science, 13(1`), 38–53.

Glasgow, L. S., & Matlack, G. R. (2007). Prescribed burning and understory composition in a temperate

deciduous forest, Ohio, USA. Forest Ecology and Management, 238(1–3), 54–64.

https://doi.org/10.1016/j.foreco.2006.08.344

Goebel, P. C., & Hix, D. M. (1996). Development of mixed-oak forests in southeastern Ohio: a comparison

of second-growth and old-growth forests. Forest Ecology and Management, 84, 1–21.

Graves, A. R., Burgess, P. J., Palma, J., Keesman, K. J., van der Werf, W., Dupraz, C., … Mayus, M.

(2010). Implementation and calibration of the parameter-sparse Yield-SAFE model to predict

production and land equivalent ratio in mixed tree and crop systems under two contrasting

production situations in Europe. Ecological Modelling, 221(13), 1744–1756.

Grubb, P. J. (1998). A reassessment of the strategies of plants which cope with shortages of resources.

Perspectives in Plant Ecology, Evolution and Systematics, 1(1), 3–31.

86

Guldin, J. M., Iffrig, G. F., & Flader, S. L. (2008). Pioneer forest-a half century of sustainable uneven-aged

forest management in the Missouri Ozarks. Retrieved from

http://www.treesearch.fs.fed.us/pubs/30744

Hanberry, B. B., Dey, D. C., & He, H. S. (2014). The history of widespread decrease in oak dominance

exemplified in a grassland–forest landscape. Science of The Total Environment, 476–477, 591–

600. https://doi.org/10.1016/j.scitotenv.2014.01.048

Hanberry, B. B., Jones-Farrand, D. T., & Kabrick, J. M. (2014). Historical open forest ecosystems in the

Missouri Ozarks: reconstruction and restoration targets. Ecological Restoration, 32(4), 407–416.

Hanberry, B. B., Kabrick, J. M., & He, H. S. (2014). Densification and State Transition Across the

Missouri Ozarks Landscape. Ecosystems, 17(1), 66–81. https://doi.org/10.1007/s10021-013-9707-

7

Hernández-Barrios, J. C., Anten, N. P. R., & Martínez-Ramos, M. (2015). Sustainable harvesting of non-

timber forest products based on ecological and economic criteria. Journal of Applied Ecology,

52(2), 389–401. https://doi.org/10.1111/1365-2664.12384

Hutchinson, T. F., Boerner, R. E., Sutherland, S., Sutherland, E. K., Ortt, M., & Iverson, L. R. (2005).

Prescribed fire effects on the herbaceous layer of mixed-oak forests. Canadian Journal of Forest

Research, 35(4), 877–890. https://doi.org/10.1139/x04-189

Jacobson, R. B., & Primm, A. T. (1997). Historical land-use changes and potential effects on stream

disturbance in the Ozark Plateaus, Missouri. U.S. Geological Survey Water-Supply Paper, (2484),

1–85.

Jenkins, M. A., & Parker, G. R. (1999). Composition and Diversity of Ground-Layer Vegetation in

Silvicultural Openings of Southern Indiana Forests. American Midland Naturalist, 142(1), 1–16.

Jenkins, M. A., & Parker, G. R. (2000). The response of herbaceous-layer vegetation to anthropogenic

disturbance in intermittent stream bottomland forests of southern Indiana, USA. Plant Ecology,

151(2), 223–237.

87

Jensen, R. G., & Kabrick, J. M. (2008). Comparing single-tree selection, group selection, and clearcutting

for regenerating oaks and pines in the Missouri Ozarks. Retrieved from

http://www.treesearch.fs.fed.us/pubs/download/13911.pdf

Jose, S., Gillespie, A. R., & Pallardy, S. (2004a). Interspecific interactions in temperate agroforestry.

Agroforestry Systems, 61(1), 237–255.

Jose, S., Gillespie, A. R., & Pallardy, S. G. (2004b). Interspecific interactions in temperate agroforestry.

Agroforestry Systems, 61, 237–255.

Jurney, D. H. (2012). Anthropology of fire in the Ozark Highland region. USDA Forest Service GTR NRS-

P-102, 12–33.

Kaur, B., McCoy, J.-A., & Eisenstein, E. (2013). Efficient, Season-Independent Seed Germination in Black

Cohosh (&lt;i&gt;Actaea racemosa&lt;/i&gt; L.). American Journal of Plant

Sciences, 04(05), 77–83. https://doi.org/10.4236/ajps.2013.45A012

Kern, C. C., Montgomery, R. A., Reich, P. B., & Strong, T. F. (2014). Harvest-Created Canopy Gaps

Increase Species and Functional Trait Diversity of the Forest Ground-Layer Community. Forest

Science, 60(2), 335–344. https://doi.org/10.5849/forsci.13-015

Kerns, B. K., Alexander, S. J., & Bailey, J. D. (2004). Huckleberry abundance, stand conditions, and use in

Western Oregon: evaluating the role of forest management. Economic Botany, 58(4), 668–678.

Kilchling, P., Hansmann, R., & Seeland, K. (2009). Demand for non-timber forest products: Surveys of

urban consumers and sellers in Switzerland. Forest Policy and Economics, 11(4), 294–300.

https://doi.org/10.1016/j.forpol.2009.05.003

King, C. B., & Muzika, R.-M. (2014a). Historic Fire and Canopy Disturbance Dynamics in an Oak-Pine (

Quercus-Pinus ) Forest of the Missouri Ozarks (1624–2010). Castanea, 79(2), 78–87.

https://doi.org/10.2179/13-046

King, C. B., & Muzika, R.-M. (2014b). Historic Fire and Canopy Disturbance Dynamics in an Oak-Pine

Quercus-Pinus Forest of the Missouri Ozarks (1624–2010). Castanea, 79(2), 78–87.

https://doi.org/10.2179/13-046

88

Kinkead, C. (2013). Thinning and burning in oak woodlands. University of Missouri–Columbia. Retrieved

from https://mospace.umsystem.edu/xmlui/handle/10355/40188

Kirkman, L. K., Mitchell, R. J., Helton, R. C., & Drew, M. B. (2001). Productivity and Species Richness

across an Environmental Gradient in a Fire-Dependent Ecosystem. American Journal of Botany,

88(11), 2119. https://doi.org/10.2307/3558437

Klimešová, J., & Klimeš, L. (2003). Resprouting of herbs in disturbed habitats: is it adequately described

by Bellingham–Sparrow’s model? Oikos, 103(1), 225–229.

Knapp, B. O., Olson, M. G., Larsen, D. R., Kabrick, J. M., & Jensen, R. G. (2014). Missouri Ozark Forest

Ecosystem Project: A Long-Term, Landscape-Scale, Collaborative Forest Management Research

Project. Journal of Forestry, 112(5), 513–524. https://doi.org/10.5849/jof.13-102

Knapp, B. O., Stephan, K., & Hubbart, J. A. (2015). Structure and composition of an oak-hickory forest

after over 60 years of repeated prescribed burning in Missouri, U.S.A. Forest Ecology and

Management, 344, 95–109. https://doi.org/10.1016/j.foreco.2015.02.009

Knowledge, I. T. E. (2014). AGROFORESTRY NOTES. Retrieved from

http://www.fs.fed.us/psw/publications/lake/psw_2014_lake001_rossier.pdf

Ladd, D. (2014). Ecologically Appropriate Fire in the Missouri Landscape: A 35 Year Reflection. Missouri

Natural Areas Newsletter, 14(1), 31–35.

Lake, F. K., & Long, J. W. (2014). Fire and Tribal Cultural Resources. In PSW-GTR-247 Science Synthesis

to Support Socioecological Resilience in the Sierra Nevada and Southern Cascade Range. U.S.

Department of Agriculture, Forest Service, Pacific Southwest Research Station. Retrieved from

http://www.fs.fed.us/psw/publications/documents/psw_gtr247/chapters/psw_gtr247_chapter4_2.p

df?

Lettow, M. C., Brudvig, L. A., Bahlai, C. A., & Landis, D. A. (2014). Oak savanna management strategies

and their differential effects on vegetative structure, understory light, and flowering forbs. Forest

Ecology and Management, 329, 89–98. https://doi.org/10.1016/j.foreco.2014.06.019

89

Li, Q., Chen, J., Moorhead, D. L., DeForest, J. L., Jensen, R., & Henderson, R. (2007). Effects of timber

harvest on carbon pools in Ozark forests. Canadian Journal of Forest Research, 37(11), 2337–

2348. https://doi.org/10.1139/X07-086

Lupardus, J. L., Muller, L. I., & Kindall, J. L. (2011). Seasonal forage availability and diet for reintroduced

elk in the Cumberland Mountains, Tennessee. Southeastern Naturalist, 10(1), 53–74.

Maginel, C. J., Knapp, B. O., Kabrick, J. M., Olson, E. K., & Muzika, R.-M. (2016a). Floristic Quality

Index for woodland ground flora restoration: Utility and effectiveness in a fire-managed

landscape. Ecological Indicators, 67, 58–67. https://doi.org/10.1016/j.ecolind.2016.02.035

Maginel, C. J., Knapp, B. O., Kabrick, J. M., Olson, E. K., & Muzika, R.-M. (2016b). Floristic Quality

Index for woodland ground flora restoration: Utility and effectiveness in a fire-managed

landscape. Ecological Indicators, 67, 58–67. https://doi.org/10.1016/j.ecolind.2016.02.035

Maret, M. P., & Wilson, M. V. (2005). Fire and litter effects on seedling establishment in western Oregon

upland prairies. Restoration Ecology, 13(3), 562–568.

Marino, P. C., Eisenberg, R. M., & Cornell, H. V. (1997). Influence of Sunlight and Soil Nutrients on

Clonal Growth and Sexual Reproduction of the Understory Perennial Herb Sanguinaria canadensis

L. Journal of the Torrey Botanical Society, 124(3), 219. https://doi.org/10.2307/2996609

McGraw, J. B., Lubbers, A. E., Van der Voort, M., Mooney, E. H., Furedi, M. A., Souther, S., … Chandler,

J. (2013). Ecology and conservation of ginseng ( Panax quinquefolius ) in a changing world:

Ecology and conservation of ginseng. Annals of the New York Academy of Sciences, 1286(1), 62–

91. https://doi.org/10.1111/nyas.12032

McLain, R. J., & Pacific Northwest Research Station (Portland, Or.). (2005). Nontimber forest products

management on national forests in the United States. Portland, OR: USDA, Forest Service, Pacific

Northwest Research Station. Retrieved from http://www.fs.fed.us/pnw/pubs/pnw_gtr655.pdf

McMurry, E. R., Muzika, R.-M., Loewenstein, E. F., Grabner, K. W., Hartman, G., & others. (2007). Initial

effects of prescribed burning and thinning on plant communities in the Southeast Missouri ozarks.

In 15th Central Hardwood Forest Conference (pp. 241–249). Retrieved from

https://firescience.gov/projects/00-2-04/project/McMurryandothers200715thCHFC.pdf

90

Miller, D. A., & Chamberlain, M. J. (2008). Plant community response to burning and herbicide site

preparation in eastern Louisiana, USA. Forest Ecology and Management, 255(3–4), 774–780.

https://doi.org/10.1016/j.foreco.2007.09.064

Missouri Department of Conservation. (2016a). Chert Protected Backslope Forest (Ecological Site

Description (Provisional)) (p. 11). Missouri. Retrieved from

https://efotg.sc.egov.usda.gov/references/Public/MO/F116BY009MO_STMfb_2016.pdf

Missouri Department of Conservation. (2016b). Deep Loess Protected Backslope Forest (Ecological Site

Description (Abstract)) (p. 2). Missouri. Retrieved from

https://efotg.sc.egov.usda.gov/references/Public/MO/F115BY003MO_summary.pdf

Mooney, E. H., Martin, A. A., & Blessin, R. P. (2015). Effects of Light Environment on Recovery from

Harvest and Antibacterial Properties of Oshá Ligusticum porteri (Apiaceae)1. Economic Botany,

69(1), 72–82. https://doi.org/10.1007/s12231-015-9302-2

Nadeau, I., Olivier, A., Simard, R. R., Coulombe, J., & Yelle, S. (1998). Growing American ginseng in

maple forests as an alternative land-use system in Québec, Canada. Agroforestry Systems, 44(2–3),

345–353.

Naud, J., Olivier, A., Bélanger, A., & Lapointe, L. (2010). Medicinal understory herbaceous species

cultivated under different light and soil conditions in maple forests in southern Québec, Canada.

Agroforestry Systems, 79(3), 303–326. https://doi.org/10.1007/s10457-009-9262-6

Nauman, T. W., Thompson, J. A., Teets, J., Dilliplane, T., Bell, J. W., Connolly, S. J., … Yoast, K. (2015).

Pedoecological Modeling to Guide Forest Restoration using Ecological Site Descriptions. Soil

Science Society of America Journal, 79(5), 1406. https://doi.org/10.2136/sssaj2015.02.0062

Nelson, P. W. (2002). CLASSIFICATION AND CHARACTERIZATION OF SAVANNAS AND

WOODLANDS IN MISSOURI. Savanna/Woodland Symposium Proceedings, 13–29.

Ni, X., Yang, W., Tan, B., Li, H., He, J., Xu, L., & Wu, F. (2016). Forest gaps slow the sequestration of

soil organic matter: a humification experiment with six foliar litters in an alpine forest. Scientific

Reports, 6, 19744. https://doi.org/10.1038/srep19744

91

O’Hara, K. L. (2002). The historical development of uneven-aged silviculture in North America. Forestry:

An International Journal of Forest Research, 75(4), 339–347.

Oláh, R., & Masarovičová, E. (1997). Response of CO2 uptake, chlorophyll content, and some

productional features of forest herb Smyrnium perfoliatum L. (Apiaceae) to different light

conditions. Acta Physiologiae Plantarum, 19(3), 285–293.

Pilz, D., Weber, N. S., Carol Carter, M., Parks, C. G., & Molina, R. (2004). Productivity and diversity of

morel mushrooms in healthy, burned, and insect-damaged forests of northeastern Oregon. Forest

Ecology and Management, 198(1–3), 367–386. https://doi.org/10.1016/j.foreco.2004.05.028

Pitman, W. D. (2009). Establishment and survival of native legumes on upland sites in Louisiana. Native

Plants Journal, 10(3), 240–250.

Price, D. M., & Kindscher, K. (2007). One Hundred Years of Echinacea angustifolia Harvest in the Smoky

Hills sof Kansas, USA. Economic Botany, 61(1), 86–95. https://doi.org/10.1663/0013-

0001(2007)61[86:OHYOEA]2.0.CO;2

Proctor, J. T. A. (1980). Some aspects of the Canadian culture of ginseng (P. quinquefolius L.), particularly

the growing environment. In Proceedings of the 3rd International Ginseng Symposium (pp. 39–

47). Seoul: Korea Ginseng Research Institute.

Pudlo, R. J., Beattie, A. J., & Culver, D. C. (1980). Population consequences of changes in an ant-seed

mutualism in Sanguinaria canadensis. Oecologia, 46(1), 32–37.

Pukkala, T., Lähde, E., Laiho, O., Salo, K., & Hotanen, J.-P. (2011). A multifunctional comparison of

even-aged and uneven-aged forest management in a boreal region. Canadian Journal of Forest

Research, 41(4), 851–862. https://doi.org/10.1139/x11-009

Rao, M. R., Palada, M. C., & Becker, B. N. (2004). Medicinal and aromatic plants in agroforestry systems.

In New Vistas in Agroforestry (pp. 107–122). Springer. Retrieved from

http://link.springer.com/chapter/10.1007/978-94-017-2424-1_8

Reintal, M., Tali, K., Haldna, M., & Kull, T. (2010). Habitat preferences as related to the prolonged

dormancy of perennial herbs and ferns. Plant Ecology, 210(1), 111–123.

https://doi.org/10.1007/s11258-010-9742-9

92

Renaud, E. (2001, June 1). The National Center for the Preservation of Medicinal Herbs Research Update.

Retrieved from

https://outlook.office365.com/owa/service.svc/s/GetFileAttachment?id=AAMkADNiMDBmZDBj

LTQzZDEtNDE1MC04MGQxLTBlNTViYmZlZTM4YQBGAAAAAABAdMEfcMOQTq2Ryhj

ma5QDBwAvcSi8v9pPTLQLyN4jYWMYAAAAAAENAAAvcSi8v9pPTLQLyN4jYWMYAA

AN91vNAAABEgAQADFWRoaYF4pAr4XTKk%2BSAqs%3D&X-OWA-

CANARY=IKT2hyOhMEWZaa_4GEm0ofAmVG9OqtIY1xmlfznQVc1uZOnvJhxsudcab2Bz4kp

M4F3pAMeKDKA.

Roberts, M. R., & Gilliam, F. S. (1995). Disturbance effects on herbaceous layer vegetation and soil

nutrients in Populus forests of northern lower Michigan. Journal of Vegetation Science, 6(6), 903–

912. https://doi.org/10.2307/3236405

Sakai, T., Saigusa, N., Yamamoto, S., & Akiyama, T. (2005). Microsite variation in light availability and

photosynthesis in a cool-temperate deciduous broadleaf forest in central Japan. Ecological

Research, 20(5), 537–545. https://doi.org/10.1007/s11284-005-0067-4

Schaalje, G. B., McBride, J. B., & Fellingham, G. W. (2002). Adequacy of approximations to distributions

of test statistics in complex mixed linear models. Journal of Agricultural, Biological, and

Environmental Statistics, 7(4), 512–524. https://doi.org/10.1198/108571102726

Scharenbroch, B. C., Nix, B., Jacobs, K. A., & Bowles, M. L. (2012). Two decades of low-severity

prescribed fire increases soil nutrient availability in a Midwestern, USA oak (Quercus) forest.

Geoderma, 183–184, 80–91. https://doi.org/10.1016/j.geoderma.2012.03.010

Schemske, D. W., Willson, M. F., Melampy, M. N., Miller, L. J., Verner, L., Schemske, K. M., & Best, L.

B. (1978). Flowering Ecology of Some Spring Woodland Herbs. Ecology, 59(2), 351.

https://doi.org/10.2307/1936379

Shackleton, C. M., & Pandey, A. K. (2014). Positioning non-timber forest products on the development

agenda. Forest Policy and Economics, 38, 1–7. https://doi.org/10.1016/j.forpol.2013.07.004

93

Shang, Z., He, H. S., Lytle, D. E., Shifley, S. R., & Crow, T. R. (2007). Modeling the long-term effects of

fire suppression on central hardwood forests in Missouri Ozarks, using LANDIS. Forest Ecology

and Management, 242(2–3), 776–790. https://doi.org/10.1016/j.foreco.2007.02.026

Shifley, S. R., & Brookshire, B. L. (2000a). Missouri Ozark Forest Ecosystem Project: site history, soils,

landforms, woody and herbaceous vegetation, down wood, and inventory methods for the

landscape experiment. U.S. Dept. of Agriculture, Forest Service, North Central Forest Experiment

Station. Retrieved from http://www.fs.fed.us/nrs/pubs/gtr/gtr_nc208.pdf

Shifley, S. R., & Brookshire, B. L. (2000b). Missouri Ozark Forest Ecosystem Project: Site History, Soils,

Landforms, Woody and Herbaceous Vegetation, Down Wood, and Inventory Methods for the

Landscape Experiment (General Technical Report No. NC-208) (p. 314). St. Paul, MN: U.S. Dept.

of Agriculture, Forest Service, North Central Forest Experiment Station. Retrieved from

http://www.fs.fed.us/nrs/pubs/gtr/gtr_nc208.pdf

Siegert, C. M., Levia, D. F., Hudson, S. A., Dowtin, A. L., Zhang, F., & Mitchell, M. J. (2016). Small-scale

topographic variability influences tree species distribution and canopy throughfall partitioning in a

temperate deciduous forest. Forest Ecology and Management, 359, 109–117.

https://doi.org/10.1016/j.foreco.2015.09.028

Sinclair, A., & Catling, P. M. (2004). Restoration of Hydrastis canadensis: experimental test of a

disturbance hypothesis after two growing seasons. Restoration Ecology, 12(2), 184–189.

Sinclair, A., Nantel, P., & Catling, P. (2005). Dynamics of threatened goldenseal populations and

implications for recovery. Biological Conservation, 123(3), 355–360.

https://doi.org/10.1016/j.biocon.2004.12.004

Small, C. J., Chamberlain, J. L., & Mathews, D. S. (2011). Recovery of Black Cohosh (Actaea racemosa

L.) Following Experimental Harvests. American Midland Naturalist, 166(2), 339–348.

Small, C. J., & McCarthy, B. C. (2010). Seed Bank Variation under Contrasting Site Quality Conditions in

Mixed Oak Forests of Southeastern Ohio, USA. International Journal of Forestry Research, 2010,

1–13. https://doi.org/10.1155/2010/419482

94

Souther, S., & McGraw, J. B. (2014). Synergistic effects of climate change and harvest on extinction risk of

American ginseng. Ecological Applications, 24(6), 1463–1477.

Stambaugh, M. C., & Guyette, R. P. (2006). Fire Regime of an Ozark Wilderness Area, Arkansas. The

American Midland Naturalist, 156(2), 237–251. https://doi.org/10.1674/0003-

0031(2006)156[237:FROAOW]2.0.CO;2

Starr, A. (2010). Local Food: A Social Movement? Cultural Studies <=> Critical Methodologies, 10(6),

479–490. https://doi.org/10.1177/1532708610372769

Storm, L., & Shebitz, D. (2006). Evaluating the Purpose , Extent , and Ecological Restoration

Applications of Indigenous Burning Practice s in Southwestern Washington. Ecological

Restoration, 24(4), 256–268.

Teutonico, D. (2015). ggplot2 essentials: explore the full range of ggplot2 plotting capabilities to create

meaningful and spectacular graphs. Retrieved from

http://proquest.safaribooksonline.com/?fpi=9781785283529

Thomas, A. L., Applequist, W. L., Rottinghaus, G. E., & Miller, J. S. (2010). Black Cohosh Rhizome and

Phytochemical Production in Response to Shading, Spacing, and Age. In XXVIII International

Horticultural Congress on Science and Horticulture for People (IHC2010): A New Look at

Medicinal and 925 (pp. 175–183). Retrieved from http://www.actahort.org/books/925/925_26.htm

Thomas, A. L., Craword Jr., R. J., Havermann, L. J., Applequist, W. L., Schweitzer, B. E., Woodbury, S.

F., & Miller, James S. (2006). Effect of Planting Depth, Planting Season, and Fungicide Treatment

on Establishment of Black Cohosh in a Poorly Drained Soil. HortScience, 41(2), 374–376.

Ticktin, T. (2004). The ecological implications of harvesting non-timber forest products. Journal of Applied

Ecology, 41(1), 11–21.

Valdivia, C., & Poulos, C. (2009). Factors affecting farm operators’ interest in incorporating riparian

buffers and forest farming practices in northeast and southeast Missouri. Agroforestry Systems,

75(1), 61–71.

Van Sambeek, J. W., Kobe, J. M., & Fralish, J. S. (1997). Interference with shoot growth and flowering of

dittany (Cunila origanoides (L.) Britton) by hardwood leachates. In Regeneration 1 (p. 294).

95

Columbia, MO: Northern Research Station. Retrieved from

http://www.nrs.fs.fed.us/pubs/ch/ch11/CHvolume11page294.pdf

Vance, E. D. (2000). Agricultural site productivity: principles derived from long-term experiments and

their implications for intensively managed forests. Forest Ecology and Management, 138(1), 369–

396.

Vance, N. C., Borsting, M., Pilz, D., & Freed, J. (2001). Special forests products: species information guide

for the Pacific Northwest (General Technical Report No. PNW-GTR-513) (p. 169). Portland, OR:

U.S. Department of Agriculture, Forest Service, Pacific Northwest Research Station. Retrieved

from http://www.fs.fed.us/pnw/pubs/gtr513/gtr513e.pdf

Vander Kloet, S. P., & Pither, J. (2000). Is fall burning preferable to spring burning for promoting growth

characteristics favorable for mechanical harvesting in Vaccinium myrtilloides Michaux?

HortScience, 35(4), 608–610.

Vasseur, L., & Gagnon, D. (1994). Survival and growth of Allium tricoccum AIT. Transplants in different

habitats. Biological Conservation, 68(2), 107–114. https://doi.org/10.1016/0006-3207(94)90340-9

Vaughan, R. C., Munsell, J. F., & Chamberlain, J. L. (2013). Opportunities for Enhancing Nontimber

Forest Products Management in the United States. Journal of Forestry, 111(1), 26–33.

https://doi.org/10.5849/jof.10-106

Vickers, L. A., Larsen, D. R., Knapp, B. O., Kabrick, J. M., & Dey, D. C. (2014). The impact of overstory

density on sapling height growth in the Missouri Ozarks: implications for interspecific

differentiation during canopy recruitment. Canadian Journal of Forest Research, 44(11), 1320–

1330. https://doi.org/10.1139/cjfr-2014-0237

Villwock, J. L., Kabrick, J. M., McNab, W. H., Dey, D. C., & others. (2011). Landform and terrain shape

indices are related to oak site index in the Missouri Ozarks. In Proceeding: The 17 th Central

Hardwood Forest Conference (pp. 197–207). Retrieved from

http://www.nrs.fs.fed.us/pubs/gtr/gtr-p-78papers/21villwockp78.pdf

Vockenhuber, E. A., Scherber, C., Langenbruch, C., Meißner, M., Seidel, D., & Tscharntke, T. (2011).

Tree diversity and environmental context predict herb species richness and cover in Germany’s

96

largest connected deciduous forest. Perspectives in Plant Ecology, Evolution and Systematics,

13(2), 111–119. https://doi.org/10.1016/j.ppees.2011.02.004

Wagner, A., & McGraw, J. B. (2013). Sunfleck effects on physiology, growth, and local demography of

American ginseng (Panax quinquefolius L.). Forest Ecology and Management, 291, 220–227.

https://doi.org/10.1016/j.foreco.2012.11.038

Webster, T. M., & Cardina, J. (1999). Apocynum cannabinum seed germination and vegetative shoot

emergence. Weed Science, 524–528.

Williams, A., Hunter, M. C., Kammerer, M., Kane, D. A., Jordan, N. R., Mortensen, D. A., … Davis, A. S.

(2016). Soil Water Holding Capacity Mitigates Downside Risk and Volatility in US Rainfed

Maize: Time to Invest in Soil Organic Matter? PLOS ONE, 11(8), e0160974.

https://doi.org/10.1371/journal.pone.0160974

Winkler, J. A., Andresen, J. A., Hatfield, J. L., Bidwell, D., & Brown, D. (2014). Climate change

vulnerabilities within the forestry sector for the Midwestern United States. Retrieved from

http://www.fs.fed.us/nrs/pubs/jrnl/2014/nrs_2014_handler_001.pdf

Woo, S.-Y. (2004). Growth and eco-physiological characteristics of Panax ginseng grown under three

different forest types. Journal of Plant Biology, 47(3), 230–235.

Young, J. A., van Manen, F. T., & Thatcher, C. A. (2011). Geographic profiling to assess the risk of rare

plant poaching in natural areas. Environmental Management, 48(3), 577–587.

Yuen, E., Anda, M., Mathew, K., & Ho, G. (2001). Water harvesting techniques for small communities in

arid areas. Water Science and Technology, 44(6), 189–194.

Zenner, E. K., Kabrick, J. M., Jensen, R. G., Peck, J. E., & Grabner, J. K. (2006). Responses of ground

flora to a gradient of harvest intensity in the Missouri Ozarks. Forest Ecology and Management,

222(1–3), 326–334. https://doi.org/10.1016/j.foreco.2005.10.027

97 Url on TROPCIOS on Url

tropicos.org/Specimen/100851683 tropicos.org/Specimen/100851685 tropicos.org/Specimen/100851690 tropicos.org/Specimen/100851696 tropicos.org/Specimen/100851694 tropicos.org/Specimen/100851681 tropicos.org/Specimen/100851688 tropicos.org/Specimen/100851689 tropicos.org/Specimen/100851693 tropicos.org/Specimen/100851680 tropicos.org/Specimen/100851697 tropicos.org/Specimen/100851682 UMO UMO UMO UMO UMO UMO UMO UMO UMO UMO UMO UMO Herbarium 8/4/2015 8/4/2015 8/4/2015 8/4/2015 8/4/2015 8/4/2015 8/4/2015 8/4/2015 8/4/2015 8/4/2015 8/4/2015 8/4/2015 Date Collected Date 9 8 5 2 3 7 6 4 1 11 12 10

Collector’s Number Collector’s APPENDIX 1 Palmer Herbarium, relocatedsince toMissouri the Botanical Garden. - Johnson,Badger Johnson,Badger Johnson,Badger Johnson,Badger Johnson,Badger Johnson,Badger Johnson,Badger Johnson,Badger Johnson,Badger Johnson,Badger Johnson,Badger Johnson,Badger Collector’s Name Collector’s Collection Location Collection Peck Ranch,Shannon Peck County, Missouri Ranch,Shannon Peck County, Missouri Ranch,Shannon Peck County, Missouri Ranch,Shannon Peck County, Missouri Ranch,Shannon Peck County, Missouri Ranch,Shannon Peck County, Missouri Ranch,Shannon Peck County, Missouri Ranch,Shannon Peck County, Missouri Ranch,Shannon Peck County, Missouri Ranch,Shannon Peck County, Missouri Ranch,Shannon Peck County, Missouri Ranch,Shannon Peck County, Missouri L. L. L. L. L. J.F. Gmel. J.F. L. (L.) Pers. (L.) L. List of type of type List specimens assessed Dunn the into L. L. Muhl. Species

24 Partheniumintegrifolium Sanguinariacanadensis Tephrosiavirginiana rubra Ulmus Actaearacemosa Apocynumcannabinum Aristolochiaserpentaria Dioscoreaquaternata Geraniummaculatum Hydrastiscanadensis Partheniumintegrifolium Actaearacemosa Table

98 APPENDIX 2

Figure 16

Figur’e 23 Difference (P < 0.05) in soil pH by site.

Figure 17

Figure 24 Bar graph showing significant difference (P < 0.05) in neutralizable acidity (meq/100 g) by site.

Figure 18

Figure 25 Bar graph showing significant difference (P < 0.05) in Ca (kg/hectare) by site.

99 Figure 19

Figure 26 Bar graph showing significant difference (P < 0.05) in Mg (kg/hectare) by site.

Figure 20

Figure 27 Bar graph showing significant difference (P < 0.05) in Fe (ppm) by site.

Figure 21

Figure 28 Bar graph showing significant difference (P < 0.05) in Mn (ppm) by site.

Figure 22

100 Figure 29 Bar graph showing significant difference (P < 0.05) in NH4-N (ppm) by site.

Figure 23

Figure 30 Bar graph showing significant difference (P < 0.05) in B (ppm) by site.

101 APPENDIX 3 Figure 26

Figure 31 Bar graph showing the mean percent cover and frequency values for Actaea racemosa by harvest type at MOFEP.

Figure 27

102

Figure 32 Bar graph showing the mean percent cover and frequency values for Apocynum cannabinum by harvest type at MOFEP. Figure 28

Figure 33 Bar graph showing the mean percent cover and frequency values for Aristolochia serpentaria by harvest type at MOFEP.

103

Figure 29

Figure 34 Bar graph showing the mean percent cover and frequency values for Dioscorea quaternata by harvest type at MOFEP. Figure 30

Figure 35 Bar graph showing the mean percent cover and frequency values for Echinacea simulata by harvest type at MOFEP.

104

Figure 31

Figure 36 Bar graph showing the mean percent cover and frequency values for Geranium maculatum by harvest type at MOFEP. Figure 32

Figure 37 Bar graph showing the mean percent cover and frequency values for Hydrastis canadensis by harvest type at MOFEP.

105

Figure 33

Figure 38 Bar graph showing the mean percent cover and frequency values for Parthenium integrifolium by harvest type at MOFEP.

106

Figure 34

Figure 39 Bar graph showing the mean percent cover and frequency values for Podophyllum peltatum by harvest type at MOFEP. Figure 35

Figure 40 Bar graph showing the mean percent cover and frequency values for Sanguinaria canadensis by harvest type at MOFEP.

107

Figure 36

Figure 41 Bar graph showing the mean percent cover and frequency values for Tephrosia virginiana by harvest type at MOFEP.

108