WHITE ASH SURVIVAL IN THE CORE OF THE EMERALD ASH BORER
INVASION AREA
By
Molly A. Robinett

A DISSERTATION
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
Forestry - Doctor of Philosophy
2017

ABSTRACT
WHITE ASH SURVIVAL IN THE CORE OF THE EMERALD ASH BORER
INVASION AREA
By
Molly A. Robinett
The invasive emerald ash borer (EAB); (Agrilus planipennis Fairmaire) (Coleoptera:
Buprestidae), a secondary pest of stressed or declining ash (Fraxinus) trees in its native Asia, is
the most destructive forest insect to ever invade North America. Since its arrival in the Detroit
metropolitan area in the early 1990’s, EAB has established in 31 states and three Canadian
provinces. To date, hundreds of millions of ash trees have been killed and nearly 8 billion trees
are threatened in U.S. forests.
While all North American ash species are susceptible to EAB, white ash (Fraxinus
americana L.) appears to be an intermediate host that is not highly preferred, but still becomes
infested and succumbs. Catastrophic levels of ash mortality, ranging from 80% to 99% of stems,
were recorded in forested sites located in southeast Michigan and Ohio. Despite these reports, we
have observed an unexpectedly high proportion of overstory white ash trees alive in southeast
and south central Michigan with the longest history of EAB infestation. To quantify this survival,
I intensively surveyed 28 sites in this region. I recorded 821 white ash trees ≥ 10 cm (DBH) in
fixed radius plots (18 m) in a 1 ha site. Most (75%) white ash stems are alive and in good
condition. More than half (68%) of the live white ash trees had external bark cracks left from
previous EAB larval feeding, however, 91% of these trees recovered. In contrast, 95% of the 373
green ash (F. pennsylvanica Marsh.) trees recorded in the 28 sites are dead. This supports
findings from previous plantation and urban ash tree studies reporting the preference of green ash
trees over white ash trees where they co-occur.

Stand-level variables that could predict the percentage of white ash stems and basal area
alive were evaluated. White ash stem density and distance from sites to the EAB origin were
positively related to white ash survival, while total basal area of all overstory species and white
ash were negatively related to white ash survival. Density of white ash stems and distance from
sites to the EAB origin were positively related to the proportion of white ash basal area alive.
White ash basal area averaged 1.5 ± 0.2 m2·ha–1, and ranged from 0.3 to 3.5 m2·ha–1 within sites,
totaling 42.7 m2·ha–1 across sites. I also tallied regeneration and found white ash accounted for
64%, 79%, and 74% of the total recruits, saplings, and seedlings recorded across the 28 sites.
Given the impact and ongoing spread of this highly destructive pest, effective monitoring
techniques are critical. I established two double decker (DD) trap types in 30 post-invasion sites
located in the EAB origin in 2014, 2015, and 2016; one DD comprised of a dark green upper
prism and a light purple lower prism (PG) baited with cis-3-hexanol, and one DD comprised of a
dark purple prism on top and bottom (PP) baited with cis-3-hexanol and Manuka oil. A total of
580, 585, and 932 EAB adults were captured in 2014, 2015, and 2016, respectively. The PG trap
captured significantly more EAB adults than the PP trap. Despite ample live white ash phloem
available for EAB development, EAB density has not reached levels high enough to kill trees.
In additional to stand-level variables, site variables could also have an influence on white
ash survival rates. I delineated a 1.5 km radial buffer from the center point of the 28 white ash
sites. Using a combination of downloaded geographic data and data collected during surveys, I
fit models at five linear distances (0 to 400 m, 400 to 800 m, 800 to 1200 m, 1200 to 1500 m, and
cross-scale). Green ash presence, specifically those growing along roads, had negative effects on
white ash survival. The proportion of woody wetlands and the presence of white ash trees,
specifically those growing along roads, were also important predictors across multiple distances.

ACKNOWLEDGMENTS

I thank my advisor, Dr. Deborah G. McCullough for all of her assistance, and most
importantly, for believing in me. Deb has greatly influenced the person I have grown to be; there
are no words to express my gratitude. I also thank my guidance committee: Dr. Robert A.
Montgomery, Dr. Therese M. Poland, and Dr. David E. Rothstein for all of their support and
assistance through this journey.
These projects would not have been possible without the help and support of my lab
mates in the Forest Entomology Lab. A special thank you to Brian Martell and Alyssa
Wethington for the countless hours spent in the field assisting with data collection.
My parents have stood by me and supported me in every decision I have made. Thank
you Dad, Kristy, Mom, and Ron for allowing me to find my place while loving me
unconditionally - I’ll always be your Bug. Thank you to my sister Cora for being not only my
sister, but my best friend.
Thanks to my Aunt Kara and Uncle Dave for being another set of parents to me, and my
Grandma Nancy for never letting me forget how proud you are of me. A special thanks to my
Grandma Lou Lou - you are my light. I love you all with my whole heart.
I thank Dr. Sara Tanis – you have been at my side through this journey, providing endless
support and comfort. There will never be enough words to express my gratitude for you. My
closest friends - Steven Murphy, Jaimie Manfre, Alyssa Wethington, Emily Eudy, and Glen
Messer. You all have helped me become the person I am today – thank you. Lastly, I want to
thank Kyle Redilla – you are everything that has been missing. Thank you for believing in me,
and pushing me to be my best self. I love you, always.

iv

TABLE OF CONTENTS

LIST OF TABLES........................................................................................................................vii
LIST OF FIGURES........................................................................................................................ix
INTRODUCTION...........................................................................................................................1
CHAPTER ONE..............................................................................................................................4
White ash (Fraxinus americana) survival in the core emerald ash borer (Agrilus planipennis)
invasion area.................................................................................................................................................4
ABSTRACT.................................................................................................................................................4
INTRODUCTION...........................................................................................................................5
METHODS....................................................................................................................................10
Site selection......................................................................................................................10
Macroplots.........................................................................................................................10
Regeneration......................................................................................................................11
Statistical analysis..............................................................................................................12
Stand-level variables and white ash survival.....................................................................12
Regeneration......................................................................................................................13
RESULTS......................................................................................................................................14
Macroplots.........................................................................................................................14
White ash stems.................................................................................................................17
White ash basal area..........................................................................................................17
Regeneration......................................................................................................................18
Recruits......................................................................................................................18
Saplings......................................................................................................................19
Seedlings.....................................................................................................................19
DISCUSSION................................................................................................................................21
APPENDIX....................................................................................................................................28
CHAPTER TWO...........................................................................................................................40
Captures of emerald ash borer (Agrilus planipennis) adults in sites with varying amounts of white
ash phloem..................................................................................................................................... 40
ABSTRACT...................................................................................................................................40
INTRODUCTION.........................................................................................................................41
METHODS....................................................................................................................................46
Site selection.....................................................................................................................46
Trapping............................................................................................................................46
Fixed radius plots..............................................................................................................47
Statistical analysis.............................................................................................................48
RESULTS......................................................................................................................................50
Trapping............................................................................................................................50
EAB adult captures and growing degree days..................................................................52
v

White ash phloem.............................................................................................................52
DISCUSSION................................................................................................................................56
APPENDIX....................................................................................................................................61
CHAPTER THREE.......................................................................................................................71
Variables related to white ash survival rates in forested sites in the core emerald ash borer invasion
area.................................................................................................................................................71
ABSTRACT...................................................................................................................................71
INTRODUCTION.........................................................................................................................72
METHODS....................................................................................................................................76
Site selection.....................................................................................................................76
White ash condition..........................................................................................................76
Variable radius plots.........................................................................................................76
Driving surveys.................................................................................................................77
Geographic data................................................................................................................77
Statistical analysis.............................................................................................................78
RESULTS......................................................................................................................................80
White ash condition..........................................................................................................80
Variable radius plots.........................................................................................................80
Driving surveys.................................................................................................................80
Geographic data................................................................................................................81
Cross-scale model.............................................................................................................81
400 m model.....................................................................................................................82
800 m model.....................................................................................................................82
1200 m model...................................................................................................................83
1500 m model...................................................................................................................84
DISCUSSION................................................................................................................................86
APPENDIX....................................................................................................................................91

LITERATURE CITED..........................................................................................................................104

vi

LIST OF TABLES
Table 1.1. Total number of trees (live and dead) (≥ 10 cm DBH), mean (± SE), minimum and
maximum diameter at breast height (DBH) (cm), total basal area (m2·ha-1) (live and
dead) for white ash trees and the five tree species with the highest relative importance
values (RIVs)1 recorded in 28 sites located in southeast and south central Michigan in
2015.. ............................................................................................................................ 28
Table 1.2. Number of live white ash and green ash trees (≥ 10 cm DBH) with signs of EAB,
including holes left by woodpeckers feeding on EAB larvae, cracks in the bark over
EAB larval galleries, and bark cracks categorized as healing by DBH class, and across
all classes...................................................................................................................... 29
Table 1.3. Parameter estimates (± SE), t values and P values of variables in a multiple linear
regression to predict percentage of white ash stems alive in 28 sites in Michigan in
2015 (R2 = 0.49; F = 4.72; DF = 7,20; P = 0.003).. ..................................................... 30
Table 1.4. Estimated (± SE) parameters, t values and P values (α < 0.05) of variables in a
multiple linear regression to predict the proportion of white ash basal area alive in 28
sites in Michigan (R2 = 0.61; F = 7.09; DF = 7,20; P < 0.001) (N = 28 sites). ............ 30
Table 1.5. Total number and mean (± SE) number of stems per hectare among sites, relative
frequency and density1 of the five most abundant recruits, saplings, and seedlings
recorded in 28 sites located in the core region of the EAB invasion in southeast and
south central Michigan. ................................................................................................ 31
Table 2.1. Mean (± SE) number of adult emerald ash borer (EAB) males and females captured
per prism in 2014, 2015, and 2016 in 30 sites in Michigan. ........................................ 60
Table 2.2. Number of male and female emerald ash borer adults captured by month and year.
Accumulated growing degree days (base 50 ºF /10 ºC)¹ from the nearest MSU Enviro
Weather station were compiled for each date. ............................................................. 61
Table 2.3. Number of white ash (F. americana) trees, estimated total and adjusted area of
phloem (m2) recorded in fixed radius plots (18 m radius) in 30 sites in Michigan in
2014 and 2016. Phloem area was calculated following methods of McCullough and
Siegert (2007), then adjusted to account for percent canopy dieback class. ................ 62
Table 3.1. Average (± SE) proportion of area classified as forested, open area, or woody
wetlands within the 1.5 km radius (706.8 ha) of the 28 sites located in southeast and
south central Michigan separated into high survival (≥ 85%), moderate survival (>
60.0% to ≤ 84.0%), or low survival (≤ 60%) survival sites based on proportion of
white ash stems live at the center of the sites in 2016. ................................................ 89

vii

Table 3.2. Competing models with a delta AICc < 1.0 for variables recorded across all four areas
bounded by 400 m, 800 m, 1200 m, 1500 m, and cross-scale models. 1Full-model
averaging results from the cross-scale model.. ............................................................ 90
Table 3.3. Competing models and parameters calculated for variables and used to predict percent
white ash survival (delta AICc < 2.0) in the center of areas bounded by a 400 m radius
in 28 sites in southeast and south central Michigan. 1Full-model averaging results from
the 0 to 400 m model.................................................................................................... 93
Table 3.4. Competing models and parameters calculated for variables and used to predict percent
white ash survival (delta AICc < 2.0) in the center of areas bounded by an 800 m
radius in 28 sites in southeast and south central Michigan. 1Full-model averaging
results from the 400 to 800 m model. .......................................................................... 94
Table 3.5. Competing models and parameters calculated for variables and used to predict percent
white ash survival (delta AICc < 2.0) in the center of areas bounded by a 1200 m
radius in 28 sites in southeast and south central Michigan. 1Full-model averaging
results from the 800 to 1200 m model. ........................................................................ 95
Table 3.6. Competing models and parameters calculated for variables and used to predict percent
white ash survival (delta AICc < 2.0) in the center of areas bounded by a 1500 m
radius in 28 sites in southeast and south central Michigan. 1Full-model averaging
results from my 1200 to 1500 m model. ...................................................................... 97

viii

LIST OF FIGURES

Figure 1.1. Location of 28 sites in 12 counties in Michigan dominated by white ash trees across
southeast and south central Michigan. ......................................................................... 32
Figure 1.2. Within each site, white ash trees (≥ 10 cm DBH) were surveyed in four macroplots
(18 m radius) in a 1 ha area. Seedlings were recorded in microplots (1.7 m radius), and
saplings and recruits were recorded in subplots (7 m radius) centered in each
macroplot...................................................................................................................... 33
Figure 1.3. Number of white ash trees (≥ 10 cm DBH) by canopy dieback class, recorded in 28
sites in southeast and south central Michigan in 2015. ............................................... 34
Figure 1.4. Percentage of white ash stems alive in 28 sites surveyed in southeast Michigan in
2015. Site order represents high to low percent white ash stem survival. ................... 35
Figure 1.5. Percentage of white ash trees alive by DBH class recorded in 28 sites. A total of 685,
107 and 29 white ash trees were recorded in the pole-sized, medium, and large DBH
classes, respectively. .................................................................................................... 36
Figure 1.6. Average DBH (cm) of white ash trees calculated for 28 forested sites located in
southeast and south central Michigan and proportion of live white ash trees. ............ 37
Figure 1.7. Linear distance (km) between the 28 white ash sites located in southeast and south
central Michigan, and the nearest site where introduced EAB parasitoids were (A)
released and (B) recovered. .......................................................................................... 38
Figure 2.1. Distribution of 30 white ash sites in southeast and south central Lower Michigan....63
Figure 2.2. Average (± SE) adult emerald ash borers captured in 2014, 2015, and 2016 in 30
white ash sites in southeast and south central lower MI on green-purple double decker
traps (PG) and purple-purple double decker traps (PP). Bars topped by the same letter
are not significantly different, P > 0.05. ...................................................................... 64
Figure 2.3. Linear relationship between total emerald ash borer adult captured in 2014-2016 and
estimated area live white ash phloem (m2) surrounding traps (18 m radius) in 30 sites
in Michigan. Area of live white ash phloem (m2) calculated from 2014 plot data is
compared to adult emerald ash borers captures in (A) 2014, (B) 2015 and (C) 2016.
Area of live white ash phloem (m2) calculated from 2016 plot data is compared to
2016 EAB captures in 2016 (D). Live white ash phloem area (m 2) was estimated
following the methods of McCullough and Siegert (2007) in 2014 and again in
2016 .............................................................................................................................. 65
Figure 2.4. Total adult emerald ash borer captures compared to total live white ash phloem area
(m2) adjusted for percent canopy dieback in 30 sites located in southeast and south
ix

central Michigan. Area of live white ash phloem (m 2) calculated from 2014 plot data
adjusted for percent canopy dieback is compared to total adult emerald ash borer
captured in 2014 (A), 2015 (B), and 2016 (C). Area of live white ash phloem (m2)
calculated from 2016 plot data adjusted for percent canopy dieback is compared to
total EAB captures in 2016 (D). Live white ash phloem area (m 2) was estimated using
diameter at breast height (DBH, cm) of white ash trees in two plots (18 m radius)
surrounding each trap in each site, and was adjusted to incorporate percent canopy
dieback by multiplying the percentage of the canopy not showing dieback. Phloem
area was calculated following the methods of McCullough and Siegert (2007) in 2014
and again in 2016. ........................................................................................................ 67
Figure 3.1. Location of 28 white ash sites in southeast and south central Michigan.................... 98
Figure 3.2. Percentage of live and dead white ash stems alive in 28 sites surveyed in southeast
Michigan in 2016 ......................................................................................................... 99
Figure 3.3. The relative variable importance for models developed to predict white ash survival
rates in areas bounded by 400 m, 800 m, 1200 m, 1500 m, and cross-scale models.
Radii and the cross-scale model for the entire 7.1 km2 area bounded by the 1.5 km
radius. Variables with a relative variable importance of 1.00 are of high importance in
predicting white ash survival rate. ....................... ......................................................100

x

INTRODUCTION
Since its arrival in the metropolitan area of southeast Detroit in the early 1990’s (Siegert
et al. 2014), emerald ash borer (EAB); (Agrilus planipennis Fairmaire) (Coleoptera: Buprestidae)
has become the most destructive and costly forest insect to invade North America (Aukema et al.
2011, Herms and McCullough 2014). To date, populations of this Asian-native beetle are known
to occur in 31 states and three Canadian provinces (EAB.info 2017). Hundreds of millions of ash
(Fraxinus spp.) trees in forests and landscapes have been killed by EAB, and more than 8 billion
trees are threatened in U.S. forests (Poland and McCullough 2006, Pugh et al. 2011, Herms and
McCullough 2014, Morin et al. 2016).
In North America, EAB preferentially colonize stressed ash trees, but will attack healthy
trees (Cappaert et al. 2005, Poland and McCullough 2006, McCullough et al. 2009b, Siegert et
al. 2010, Flower et al. 2013b, Herms and McCullough 2014). Feeding and development of EAB
can likely occur on all ash species in the U.S., including the four prominent species occurring in
Michigan (Anulewicz et al. 2007, Tanis and McCullough 2012, 2015). Though all ash species
are susceptible to EAB, there are interspecific differences in ash tree resistance or EAB host
preference among species (Cappaert et al. 2005, Anulewicz et al. 2007, 2008, Rebek et al. 2008,
Pureswaran and Poland 2009, Chen and Poland 2010, Tanis and McCullough 2012, 2015). Green
ash (Fraxinus pennsylvanica Marsh.) and black ash (F. nigra Marsh.) are consistently highly
preferred and vulnerable hosts of EAB (Cappaert et al 2005, Anulewicz et al. 2007, 2008, Rebek
et al. 2008, Chen and Poland 2010), while blue ash (F. quadrangulata Michx.) is relatively
resistant to EAB (Anulewicz et al. 2007, Tanis and McCullough, 2012).
White ash (F. americana L.) is considered an intermediate host for EAB (Anulewicz et
al. 2007, Tanis and McCullough 2012, 2015). Previous studies showed green ash trees were

1

preferentially colonized in multi-species plantations and in landscapes where green ash and white
ash trees co-occurred (Anulewicz et al. 2007, Limback 2010, Tanis and McCullough 2015).
While substantial white ash mortality has occurred in many areas in southeast Michigan (Gandhi
and Herms 2010, Knight et al. 2013, 2014, Klooster et al. 2014, Smith et al. 2015), evidence
suggests interactions between white ash trees and EAB may be relatively complex. We have
observed an unexpectedly high proportion of overstory white ash trees alive in forested sites in
southeastern and south central Michigan (McCullough, unpublished data), known as the EAB
“core” of the invasion (Burr and McCullough 2014). In chapter one, I intensively surveyed 28
sites to quantify the abundance and condition of white ash trees and white ash regeneration, and
evaluated stand-level variables that could influence white ash survival rates.
Given the impact and ongoing spread of EAB, effective monitoring techniques to detect
newly infested areas are a critical aspect of managing this highly destructive pest. Previous
studies have shown EAB adults do not produce long-range sex or aggregation pheromones
(Lelito et al. 2007, Rodriguez-Saona et al. 2007, Crook and Mastro 2010, Ryall 2015). Beetles
rely on olfactory and visual cues, including volatiles produced by stressed ash trees (Poland et al
2004, 2005), and visual attractants such as color (Francese et al. 2005c) to locate potential host
trees and encounter mates. Live phloem must be available for larval development to sustain EAB
populations. In chapter two, I compared EAB adult captures on two baited double decker trap
types in 30 post-invasion white ash sites in southeastern and south central Michigan in 2014,
2015, and 2016. I also tallied white ash trees by diameter class in each site in 2014 and 2016. I
related EAB captures to phloem area (m2) of live white ash to determine whether phloem
available for larval development limits EAB populations.

2

Characteristics of trees and sites can play a role in tree mortality caused by native or
exotic pests. In a previous study, Knight et al. (2013) reported site-level variables (hydrology)
and tree-level variables (ash density, crown class) were related to ash tree mortality once a stand
was infested with EAB. In contrast, Smith et al. (2015) reported no relationship between site and
stand variables and mortality rates in white, green, or black ash sites. A study by Jenkins et al.
(2016) quantified ash mortality risks using “detection tree” data collected in Ohio from 2005 to
2007, while also incorporating distance to existing ash stands, campgrounds, roads, and railways
that presumably provide corridors to facilitate EAB dispersal. They reported ash along riparian
corridors are at high relative risk of EAB-induced mortality. In heterogeneous sites with urban,
residential, and wooded areas, EAB preferentially colonized open-grown ash trees rather than
those shaded in woodlots, which supports previous studies (McCullough et al. 2009b, 2009c). In
chapter three, I evaluated an array of site-level variables in the 28 sites where I surveyed white
ash survival. I fit models to five linear distances to identify if any variables were associated with
rates of white ash survival.

3

CHAPTER ONE
White ash (Fraxinus americana) survival in the core of the emerald ash borer
(Agrilus planipennis) invasion

ABSTRACT
Previous studies have reported catastrophic levels of white ash (Fraxinus americana L.),
green ash (F. pennsylvanica Marsh.) and black ash (F. nigra Marsh.) mortality caused by
emerald ash borer (EAB) (Agrilus planipennis Fairmaire) in several forested areas in southeast
Michigan and Ohio. We observed, however, a relatively high proportion of white ash trees
remain alive in some southeast Michigan forests, despite the long history of EAB invasion across
the region. To assess white ash survival, I inventoried overstory trees in fixed radius plots in 28
forested sites in southeast and south central Michigan in 2015. Overall, 75% of the 821 white ash
stems (≥ 10 cm DBH) I recorded were alive. Most were in good condition; 95% of the live trees
had ≤ 30% canopy dieback. Similarly, 66% of the white ash basal area was alive, averaging 1.0 ±
0.2 m2·ha–1 across sites. Most live white ash trees were previously colonized by EAB. Bark
cracks over old larval galleries were observed on 68% of the live white ash, but callus tissue (i.e.,
wound periderm) had formed over or around 91% of those cracks. White ash regeneration was
abundant, accounting for 64%, 79%, and 74% of the recruits, saplings, and seedlings recorded.
Distance to locations where Asian parasitoids of EAB were released or recovered did not
influence white ash survival rates. Although the long term survival of these trees is unknown, the
persistence of white ash to date in some forested areas is encouraging.

4

INTRODUCTION
Since its discovery in 2002, emerald ash borer (EAB); (Agrilus planipennis Fairmaire)
(Coleoptera: Buprestidae) has become the most destructive and costly forest insect to invade
North America (Aukema et al. 2011, Herms and McCullough 2014). To date, EAB populations
are known to occur in 31 states and three Canadian provinces (EAB.info 2017). Although EAB
preferentially colonizes stressed ash (Fraxinus spp.) trees, it can also colonize and kill healthy
ash trees in North America (Cappaert et al. 2005, Poland and McCullough 2006, McCullough et
al. 2009a, 2009b, Flower et al. 2013b). An extensive dendrochronological study showed the
North American EAB invasion likely originated in the Detroit suburbs of Westland and Canton,
in southeast Michigan by the early 1990’s (Siegert et al. 2014). Since then, this beetle, which is
native to Asia, has killed hundreds of millions of ash trees in forests and landscapes, and
threatens more than 8 billion trees in U.S. forests (Poland and McCullough 2006, Pugh et al.
2011, Herms and McCullough 2014, Morin et al. 2016). Recent projections suggested an
estimated 10.7 billion USD will be spent by 2019 for removal or treatment of roughly half of the
municipal ash landscape trees likely to be affected (Kovacs et al. 2010). Impacts associated with
the EAB invasion are not only economic but also ecological. Widespread ash mortality will
likely have cascading effects in many forested ecosystems, altering species composition,
understory light, temperature and moisture availability, increasing coarse woody debris and
affecting carbon and nitrogen cycling (Gandhi and Herms 2010, Flower et al. 2013a, Burr and
McCullough 2014, Wagner and Todd 2016).
Catastrophic levels of ash mortality, ranging from 80% to 99% of stems, have been
recorded in forested sites dominated by white ash (Fraxinus americana L.), green ash (F.
pennsylvanica Marsh.), and black ash (F. nigra Marsh.) in southeast Michigan and Ohio (Gandhi
5

and Herms 2010, Flower et al. 2013b, Knight et al. 2013, Burr and McCullough 2014, Klooster
et al. 2014, Smith et al. 2015). Smith et al. (2015) evaluated an array of site and stand-level
variables from 2004-2007 and found distance from the individual sites in southeast Michigan to
the core of the EAB invasion in the greater Detroit area was the only variable significantly
related to the rate of ash mortality. In sites where either green, black, or white ash trees were
abundant, Klooster et al. (2014) reported ash mortality rates from 2005 to 2009 ranged from 40
to 99.7 %. Knight et al. (2013) reported ash trees died more rapidly when ash densities were low
compared to stands where ash trees were more abundant, but overall, 25% of the ash trees were
killed three years after the infestation became apparent and 99% of ash were killed after six
years.
Given the extent of ash mortality in forests, three EAB parasitoids native to China have
been imported into the U.S. for classical biological control of EAB. Releases began in sites in
southeast Michigan in 2007 (Federal Register 2007, Bauer et al. 2008, 2014, 2015) and are
ongoing. Only a small proportion of parasitoid release sites were subsequently monitored to
assess establishment (Duan et al. 2011, 2012b, 2013b, U.S. Department of Agriculture 2013,
Abell et al. 2014, Bauer et al. 2014). Two species, the larval parasitoid Tetrastichus planipennisi
Yang (Eulophidae) and the egg parasitoid Oobius agrili Zhang and Huang (Encyrtidae) (Yang et
al. 2005), have become established in some release sites in Michigan (Duan et al. 2013a, Bauer
et al. 2015), while establishment of the larval parasitoid Spathius agrili Yang (Braconidae)
(Yang et al. 2006) in Michigan has not been verified (Bauer et al. 2008, Slager, in press). To
date, however, there is no evidence that introduced parasitoids have slowed ash mortality rates
(Bauer et al. 2008, 2015, Abell et al. 2012, Duan et al. 2015).

6

While EAB can likely develop on all North American ash species (Herms and
McCullough 2014), consistent interspecific differences in EAB host preference or host resistance
have been observed (Cappaert et al. 2005, Anulewicz et al. 2007, 2008, Rebek et al. 2008,
Pureswaran and Poland 2009, Chen and Poland 2010, Tanis and McCullough 2012, 2015). Green
ash and black ash are consistently highly preferred and vulnerable hosts of EAB (Cappaert et al
2005, Anulewicz et al. 2007, 2008, Rebek et al. 2008, Chen and Poland 2010), while healthy
blue ash trees (F. quadrangulata Michx.) are less preferred hosts of EAB (Anulewicz et al. 2007,
Tanis and McCullough 2012, 2015).
While there is clearly substantial white ash mortality in many areas in this region (Gandhi
and Herms 2010, Knight et al. 2013, 2014, Klooster et al. 2014, Smith et al. 2015), interactions
between white ash trees and EAB may be less consistent or perhaps more complex compared
with green ash or black ash. For example, in a plantation study where five ash species were
arranged in randomized blocks, larval densities averaged (± SE) 41 ± 12 larvae per m2 of phloem
area on white ash trees compared to 220 ± 40 and 236 ± 36 larvae per m2 on green ash and black
ash trees, respectively (Tanis and McCullough 2015). Larval mortality on the green ash and
black ash trees was very high, however, because of intraspecific competition for phloem. In
another plantation with alternating green ash and white ash trees, green ash trees had nearly
twice as many larval galleries as similarly-sized white ash trees (Limback 2010). Similarly,
Anulewicz et al. (2007) reported green ash landscape trees were preferentially colonized over
white ash trees in four residential areas where these species co-occurred in southeast Michigan.
We observed over the years that white ash trees in at least a few forested areas in the
“core” of the EAB invasion in southeast and south central Michigan continued to survive, despite
the presence of EAB in the sites for at least 10-12 years (McCullough, unpublished data). It was

7

not clear, however, whether high rates of white ash survival were fairly common in areas with a
long EAB history or were limited to a few sites that we simply happened to notice. I
hypothesized that abundance of green ash, a preferred and highly vulnerable host of EAB, within
or in close proximity to white ash-dominated sites could affect white ash survival, either
negatively or positively. For example, high numbers of beetles emerging from heavily infested
green ash trees could subsequently colonize and kill white ash trees. Conversely, green ash trees
could function as sinks by attracting most of the ovipositing EAB females, perhaps reducing
pressure on less preferred white ash trees, particularly if interspecific competition resulted in
high mortality of larvae on the green ash. Distance to parasitoid release and recovery sites or
distance to the epicenter of the EAB invasion in the greater Detroit metropolitan area could also
influence white ash survival.
To address these questions, I identified and surveyed white ash growing in forested sites
across much of the core region of the EAB invasion in southeastern and south central Michigan.
My goals included documenting white ash survival rates, evaluating whether green ash or the
presence of other overstory species were related to white ash survival, and determining whether
other stand-related factors were consistently related to white ash survival.
I also quantified regeneration in these sites to assess potential long-term species
composition and to evaluate whether species composition of regeneration reflected overstory
composition. Previous studies showed that in sites where nearly 100% of the overstory ash trees
were killed by EAB, ash seed banks were rapidly depleted, and ash regeneration sharply declined
(Burr and McCullough 2014, Klooster et al. 2014, Smith et al. 2015). Orphan cohorts, consisting
of previously established seedlings and saplings too small to be colonized by EAB remained, but

8

competition with co-occurring species could limit the chances that young ash reach the overstory
(Burr and McCullough 2014, Herms and McCullough 2014).

9

METHODS
Site selection
To avoid biasing my site selection, I used a county atlas to identify 70 areas of public
forestland (state, county, or municipal property). These areas were distributed across 11 counties
in southeast and south central Michigan that were invaded by EAB at least 10-12 years earlier
(Pugh et al. 2011, Siegert et al. 2014). I queried managers of each area about the presence and
distribution of white ash and excluded 30 areas where white ash was scarce or absent. In spring
2014, I scouted the remaining areas and selected 28 sites where white ash trees, live or dead,
were relatively abundant (Fig 1.1). Most areas were comprised of second growth forests that had
regenerated on abandoned agricultural land, near floodplains, or following other disturbances.
I recorded coordinates of a center point established in the midst of the white ash trees in
each site, then delineated a rectangular 1 ha area around each center point. Four non-overlapping
macroplots, each with a radius of 18 m, were established in each cardinal direction from the
center point within the 1 ha area using ArcMap 10.3.1. At the center of each macroplot, I
established a subplot (radius 7 m) to quantify recruits and saplings by species and a microplot
(radius 1.7 m) to survey seedling species (Fig 1.2).
Macroplots
Between late May and mid-August 2015, I recorded DBH of all live and dead trees ≥ 10
cm DBH by species within each macroplot. Canopy dieback of live trees was visually assessed in
increments of 10% in early to mid-summer after trees were fully flushed, where 10% indicated a
nearly full, healthy canopy and 90% indicated a nearly dead canopy (Zarnoch et al. 2004).
Although standing dead ash were encountered in nearly all sites, fallen ash trees and other forms
of coarse woody debris were scarce or non-existent within and beyond my plots. Counts of live

10

and dead white ash trees overall and by DBH class were summed for the four macroplots in each
site and standardized per ha. Total (live and dead) and live basal area (m2·ha-1) were calculated
for all overstory species, for all white ash trees, and for all green ash trees, when present.
Relative importance values (RIV) for the five most abundant overstory species across sites and
within each site were calculated following methods of Kent and Coker (1992).
Evidence of EAB colonization, including D-shaped exit holes left by emerging adults and larger
holes left by woodpeckers preying on EAB larvae, were recorded as absent, scarce (≤ 4 holes
visible), medium (5-8 holes visible), or abundant (≥ 9 holes visible) on white ash and green ash
trees. Number of bark cracks above larval feeding galleries on the trunk were also tallied. If
callus tissue (i.e., wound periderm) and bark were growing along the perimeter or over the
surface of old larval galleries, I categorized the galleries as “healed”.
Asian parasitoid release and recovery data were obtained from the MapBioControl
database, a geospatial framework used to monitor parasitoid release, recovery, and spread
(MapBioControl 2016). A total of 565 release sites and 107 recovery sites were identified in
Lower Michigan, where 103 release and 77 recovery sites were located in southeast and south
central Michigan. The type and quality of data reported in MapBioControl varied considerably.
Frequently, no information on parasitoid recovery was available for release sites and conversely,
recovery sites were not necessarily locations where parasitoids were released. Despite these
limitations, I identified the parasitoid release and/or recovery sites closest each of the 28 sites I
surveyed.
Regeneration
Number and species of recruits (stems ≥ 3.0-9.9 cm DBH) and saplings (DBH < 3.0 cm;
height ≥ 45.0 cm) were tallied from May to August 2015 in subplots and number of seedlings (<

11

45.0 cm height) were recorded by species in microplots (Fig 1.2). Ash and elm seedlings were
recorded only to genus due to difficulty in distinguishing species. Recruit, sapling, and seedling
counts from subplots and microplots were summed by species (or genus) and standardized per ha
for each site.
Statistical analysis
Severely declining overstory trees with ≥ 80% canopy dieback appeared unlikely to
recover (Fig 1.3), and were classified as “dead” for analyses. To evaluate survival rates among
size classes, white ash trees were grouped into three DBH classes: pole-sized (10.0-20.0 cm),
medium (20.1-30.0 cm), and large (> 30.1 cm) trees. Variables were tested using the ShapiroWilk test. If variables were not normal and could not be normalized by transformations, the
nonparametric Kruskal-Wallis test was used to compare ranks of percent white ash survival
among DBH classes. Pairwise multiple comparisons with Tukey’s adjustment were used to
identify differences among DBH classes when the Kruskal-Wallis test was significant. Tests
were conducted using R statistical software (R Core Development Team, http://www.Rproject.org) at α of P < 0.05.
Stand-level variables and white ash survival
I used multiple linear regression to identify variables that could predict the percentage of
the white ash stems and basal area alive in the sites in 2015. Potential predictor variables
included basal area and stems per ha of white ash, green ash, and all species combined (live and
dead), and the distance from each site to the EAB origin near Westland, Michigan. Pearson’s
correlation coefficients were calculated for potential predictors and only variables with a
correlation coefficient < 0.7 were included in the initial models to avoid multicollinearity (Elith
et al., 2006, Wisz and Guisan, 2009). Generalized Variance Inflation Factors (VIFs) were

12

considered for the remaining predictor variables (Neter et al. 1996, Brunsdon et al. 2012) and
competing variables with the highest VIF values were eliminated until all VIFs were < 3.
Multicollinearity diagnoses were conducted using R statistical software (R Core Development
Team, http://www.R-project.org) at α of P < 0.05.
To assess whether varying levels of white ash survival could be related to effects of
Asian parasitoids, I used ArcMap 10.3.1 to measure the Euclidean distance between each of the
28 sites and the nearest parasitoid release or recovery site. Simple linear regression was used to
determine if distance between my sites and parasitoid release or recovery locations predicted
white ash survival rates.
Regeneration
I calculated stems per ha by species, and determined species richness within and across
sites. I identified the five most frequently recorded species in each regeneration stratum (recruits,
saplings, seedlings), then calculated total and mean (± SE) density and relative frequency of
these species. Counts of recruit, sapling, and seedling species within sites were compared to
overstory white ash survival rates and live white ash basal area using linear regression.

13

RESULTS
Macroplots
I recorded 3,587 live and dead trees (DBH ≥ 10 cm) representing 44 different species
across the 28 forested sites, including 821 white ash, 373 green ash and 2,393 stems of other
species. The most abundant species co-occurring with white ash included green ash, American
elm (Ulmus americana L.), black cherry (Prunus serotina Ehrh.), sugar maple (Acer saccharum
Marsh.), and red oak (Quercus rubra L.) (Table 1.1). Green ash, which was tallied in 27 sites and
sugar maple, tallied in 22 sites, each accounted for approximately 10% of the total stems
recorded but had relative importance values (RIVs) of 88.7 and 58.4, respectively. Sugar maple
was not frequent but abundant when it occurred in the sites, while green ash was both frequent
and abundant in the sites. Black cherry, American elm, and red oak accounted for 9%, 8%, and
6% (333, 296, and 215 stems) of the total stems and were recorded in 26, 26 and 18 sites,
respectively (Table 1.1). Trees that co-occurred with white ash averaged 20.7 ± 2.0 cm DBH and
most (74%) were ≤ 25 cm in DBH.
Overall, dead trees co-occurring with white ash were minimal, representing 18% of the
stems recorded. I tallied 344, 76, and 30 dead green ash, American elm, and black cherry trees,
respectively. Green ash accounted for 68% of all dead trees (mean DBH 17.9 ± 0.46 cm) and
were recorded in 27 sites (Table 1.1). Live green ash trees were recorded in ten sites,
representing 8% of the total green ash recorded. American elm (mean DBH 19.0 ± 1.1 cm) and
black cherry trees (mean DBH 16.5 ± 1.3 cm) were recorded in 18 and 9 sites, respectively, and
accounted for 15% and 6% of the dead trees.
White ash survival rates varied substantially among sites, ranging from 0% to 100% of
the stems recorded (Fig 1.4), but across all sites, 75% of the white ash trees were alive. In 14

14

sites, ≥ 80% of the total white ash stems were alive, and at least 50% of the white ash stems were
alive in 22 sites (Fig 4.1). Most (83%) of the live white ash stems had ≤ 10% dieback and 95%
had ≤ 30% canopy dieback (Fig 1.3). Only 27 trees, representing < 2% of the total white ash
stems, were severely declining with canopy dieback estimates of ≥ 80% (Fig 1.3). These trees
were heavily colonized by EAB, as evidenced by abundant adult beetle exits and bark cracks
along the trunk.
White ash survival varied with DBH, which ranged from 10.0 to 44.0 cm across the 28
sites. I recorded 685, 107, and 29 white ash trees in the pole-sized, medium, and large diameter
classes, respectively, which averaged (± SE) 13.1 ± 0.1 cm, 24.1 ± 0.2 cm, and 35.1 ± 0.6 cm
DBH, respectively. Overall, 77% of pole-sized trees and 63% of medium trees were alive, but
only 45% of large trees were alive, which was significantly lower than survival rates for trees in
the smaller DBH classes (F = 6.42; DF= 3,98; P < 0.001) (Fig 1.5). Mean DBH of dead white
ash trees averaged 17.2 ± 0.5 cm and was significantly greater than mean DBH of live white ash
trees (t=4.61; DF=277.6; P < 0.0001). Within sites, mean DBH of the white ash trees was
significantly and negatively related to the proportion of those trees alive in 2015 (Fig 1.6).
Total basal area of species co-occurring with white ash averaged 10.7 ± 1.3 m2·ha–1.
White ash basal area (live and dead) averaged 1.5 ± 0.2 m2·ha–1, and ranged from 0.3 to 3.5
m2·ha–1 within sites. Live white ash basal area averaged 1.0 ± 0.2 m2·ha–1 and comprised 66% of
the total white ash basal area (live and dead) summed across all sites (42.7 m2·ha–1). Green ash
basal area (live and dead) averaged 1.0 ± 0.8 m2·ha–1 and ranged from 0 to 3.8 m2·ha–1 within
sites. Dead green ash basal area averaged 0.9 ± 0.2 m 2·ha–1 and comprised 95% of the total green
ash basal area recorded across all sites (27.4 m2·ha–1), while live green ash basal area averaged
0.05 ± 0.02 m2·ha–1 across sites.

15

Evidence of EAB colonization was apparent on all dead overstory white ash trees.
Woodpecker holes were observed on the trunk of 99% of the dead white ash trees, while exit
holes left by EAB adults and bark cracks above larval galleries were observed on 90% and 68%
of the dead white ash trees, respectively. The dead green ash (344 trees) were also clearly killed
by EAB. I could see woodpecker holes and EAB adult exit holes on 96% and 88% of the dead
green ash, respectively, and when bark was intact, cracks over old galleries on the trunk were
apparent on 59% of these trees.
Evidence of EAB colonization was also common on nearly all live white ash trees. Exit
holes and woodpecker holes could be seen on 37% and 53% of the live white ash trees,
respectively. Bark cracks were observed on 68% of live white ash trees, but these trees had
formed callus tissue (i.e., wound periderm) over or around the edge of 91% of these cracks,
which I considered to be “healing”. Proportion of live white ash trees with woodpecker holes
increased as DBH class increased. Overall, 52%, 58%, and 69% of the white ash trees in the
pole-sized, medium, and large DBH classes had woodpecker holes, respectively (Table 1.2). The
percentage of live white ash trees with bark cracks varied among DBH classes, where 69%, 64%,
and 54% of the live stems in the pole, medium, and large DBH classes, respectively, had one or
more bark cracks over old galleries. Overall, 54% of the live white ash trees in the large DBH
class and 62% of trees in the pole-sized and medium DBH classes were forming wound periderm
around and over old larval galleries.
I tallied only 29 live green ash trees across the 28 sites. Nearly all (93%) of those trees
had at least five woodpecker holes visible. Bark cracks were present on 22 of the live green ash
trees, but cracks on only two trees appeared to be healing (Table 1.2).

16

White ash stems
I considered 11 potential variables as predictors of survival rates for white ash stems.
Four were dropped because of multicollinearity, including basal area of dead green ash, density
of stems (live and dead) of all species, density of live stems of all species, and density of live
green ash stems. Four variables were significant predictors of white ash survival rate. Total basal
area of all overstory species and total basal area of white ash were significantly and negatively
related to white ash survival, while white ash stem density and the distance from sites to the EAB
origin near Detroit were positively related to survival (Table 1.3). Overall, the model with the
four predictors explained 49% of the variability in the proportion of white ash stems alive in
2015 (Table 1.3).
White ash basal area
I considered 13 variables as potential predictors of the proportion of white ash basal area
alive in our 2015 surveys. Six variables were dropped because of multicollinearity, including
basal area of dead green ash, density of live green ash stems, density of total stems (live and
dead), density of total live stems, density of live white ash stems, and total (live and dead) white
ash basal area. Only two of the remaining variables were significant predictors; density of total
white ash stems and distance from the sites to the EAB origin were positively related to the
proportion of white ash basal area that was alive in 2015 (Table 1.4). Overall, the model with the
two predictors explained 61% of the variability in the proportion of white ash basal area that was
alive (Table 1.4).
I used simple linear regression to examine the relationship between white ash survival
(percent of stems alive) and distance between the sites and the closest locations where Asian
parasitoids of EAB were released and/or recovered. Distances between the white ash sites and

17

the nearest parasitoid release or recovery site recorded in the MapBioControl database ranged
from 0.09 to 47.7 km and 0.04 to 47.9 km, respectively. There was little influence of distance to
parasitoid release sites (R2 = 0.002) or recovery sites (R2 = 0.009) on white ash survival rates (Fig
1.7).
Regeneration
Recruits
Not surprisingly, white ash dominated all three regeneration strata (Table 1.5). Overall,
1,601 recruits (DBH ≥ 3.0 and < 10.0 cm) representing 25 species were recorded across the 28
sites. White ash comprised 64% of the 1,028 recruits tallied and were present in all 28 sites.
Number of white ash recruits recorded in individual sites ranged from 3 recruits in two sites to
171 recruits in one site. I recorded < 10 white ash recruits in five of the 28 sites and > 30 in 11
sites. Overall, density of white ash recruits averaged 596 ± 118.9 stems per ha. In 18 sites, white
ash recruits comprised ≥ 50% of all recruits, while in one site, white ash comprised < 20% of the
recruits. This site also had the fewest white ash overstory trees and the lowest white ash basal
area (live or dead). The linear relationship between the percent of white ash trees (≥ 10 cm DBH)
alive in our surveys and the density of white ash recruits was weak (y = 0.0001x + 0.677, R² =
0.013, P = 0.56), as was the linear relationship between total white ash basal area (live and dead)
and the density of white ash recruits (y = -0.0003x + 1.686, R² = 0.038, P = 0.32).
Species composition of recruits was similar to composition of overstory trees recorded in
macroplots, but American elm, hawthorn (Crataegus spp.), sugar maple, and black cherry, which
were important overstory species, collectively accounted for ≤ 8% of the total recruits recorded.
Green ash, an abundant, albeit mostly dead overstory species, was nearly absent in the recruit
stratum (Table 1.5). A total of 23 live green ash recruits were recorded in nine of the 28 sites.

18

Seven sites had a single green ash recruit, while two other sites had 14 recruits. No dead green
ash recruits were encountered. Other species of recruits recorded in our plots included box elder
(Acer negundo L.), black walnut (Juglans nigra L.), white oak (Q. alba L.), and American beech
(Fagus grandifolia Ehrh.), which collectively accounted for 12% of the recruits tallied.
American elm recruits were recorded in 19 of the 28 sites, while none of the other species were
recorded in more than 13 sites.
Saplings
A total of 2,849 saplings (DBH < 3.0 cm; height ≥ 45.0 cm) representing 25 species were
recorded. White ash accounted for 79% (2,256) of all saplings and were present in all 28 sites,
averaging 1,683 ± 168.6 stems per ha. In 24 sites, white ash accounted for ≥ 50% of all saplings,
and in two sites, white ash accounted for 100% of the saplings. Percent survival of overstory
white ash trees was not significantly related to the density of white ash saplings within sites (y =
0.001x + 0.591, R² = 0.110, P = 0.08), nor was the linear relationship between total white ash
basal area (live and dead) and the density of white ash saplings significant (y = 0.0002x + 1.236,
R2 = 0.07, P = 0.189).
Other abundant sapling species included American elm, black cherry, sugar maple, and
hawthorn (Table 1.5). As with recruits, green ash saplings were rare. A total of ten green ash
saplings were alive in four sites; no dead green ash saplings were recorded. American elm was
encountered in 18 sites and was again the most abundant species, representing 6% of saplings
recorded. No other species accounted for more than 3% of the total saplings.
Seedlings
A total of 5,973 seedlings (height < 45.0 cm) representing 24 species were tallied in the
28 microplots. Ash and elm seedlings were recorded only by genus due to the difficulty in
19

differentiating between species. Ash seedlings dominated all sites, accounting for 30% to 100%
of seedlings within sites and 74% (4,397) of all seedlings. Percent survival of overstory white
ash trees was not related to the density of white ash seedlings within sites (y = 0.003 x + 0.717,
R² = 0.001, P = 0.858), but total basal area of white ash trees (live and dead) was positively
related to the density of ash seedlings (y = 0.004x + 0.963, R2 = 0.296, P = 0.003). Sugar maple,
black cherry, and elm were relatively common and abundant (Table 1.5). Sugar maple
represented 8% of the total seedlings, and was present in 17 of the 28 sites. Elm represented ~
2% of the seedlings and was present in 16 sites while black cherry and sugar maple seedlings
were present in 17 sites.

20

DISCUSSION
A substantial number of white ash trees remain alive and healthy in the core of the EAB
invasion in southeast and south central Michigan, despite the relatively long history of EAB
presence in this region. More than 65% of the white ash basal area and 75% of the white ash
stems tallied across the 28 sites I surveyed were alive and nearly all of those trees (95%) had less
than 30% canopy dieback. I did survey sites where white ash mortality was similar to the
catastrophic rates reported in previous studies (Gandhi and Herms 2010, Knight et al. 2013,
Klooster et al. 2014, Smith et al. 2015). In one of my sites, every white ash tree was dead and
more than half of the white ash trees in four other sites were dead. Nevertheless, the continued
survival of white ash in sites spanning 11 counties is encouraging.
I originally hypothesized that the presence of green ash, a highly preferred and vulnerable
EAB host (Cappaert et al 2005, Anulewicz et al. 2007, 2008, Rebek et al. 2008, Chen and Poland
2010, Limback 2010), could affect white ash survival, either positively or negatively. If green
ash trees acted as a sink for EAB oviposition, the number of eggs laid on less attractive white ash
might be reduced, particularly if intraspecific competition for phloem reduced larval survival on
green ash trees. Conversely, if high numbers of EAB adults emerged from green ash trees,
nearby white ash trees could be heavily colonized as the green ash declined and succumbed, a
pattern previously observed with green ash and white ash landscape trees (Anulewicz et al.
2007). In addition, an abundance of green ash, which can grow in heavier soils or in ephemerally
flooded areas, could indicate site conditions were less than optimal for white ash, which is more
often found on upland mesic sites (Schlesinger 1990, NRCS 2017). In my survey, green ash trees
were present in all but one site, although 90% of those trees had been killed by EAB. I included
the density of green ash stems (live and dead), total green ash basal area and live green ash basal

21

area as potential predictors of the proportion of white ash stems or basal area alive in the
multiple regressions. None of the green ash-related variables, however, significantly entered
either regression.
Differential survival of green ash and white ash trees in my sites is clearly consistent with
previously observed interspecific variation in EAB host preference (Cappaert et al. 2005,
Anulewicz et al. 2008, Rebek et al. 2008, Pureswaran and Poland 2009, Chen and Poland 2010,
Tanis and McCullough 2012, 2015). Although both ash species appear to be highly suitable hosts
for leaf-feeding adult beetles and larvae developing in phloem and cambium (Anulewicz et al.
2008, Pureswaran and Poland 2009, Chen and Poland 2010), green ash is consistently colonized
sooner and at higher densities than white ash. For example, in a recent plantation study, 21
randomized blocks of four ash species were exposed to the local EAB population for a single
season, then debarked in autumn. All 21 green ash trees were heavily infested, with an average
of 220 ± 40 larvae per m2. In comparison, three of the 21 white ash trees were heavily infested,
while five white ash had no galleries and overall larval density on white ash averaged 41 ± 12
larvae per m2 (Tanis and McCullough 2015). Although both green ash and white ash appear to be
highly suitable hosts for leaf-feeding adult beetles and larvae developing in phloem and
cambium (Anulewicz et al. 2007, 2008, Rebek et al. 2008, Pureswaran and Poland 2009, Chen
and Poland 2010), host preference of ovipositing females appears to determine EAB attack
densities and mortality rates.
Variation in white ash survival in my sites was at least partially attributable to the size
and density of the white ash trees along with the basal area or density of co-occuring overstory
species. Most white ash trees in the second-growth sites I surveyed were ≤ 30 cm in DBH and
75% of these trees were alive and healthy. Larger trees, however, did not fare as well; less than

22

half of the trees > 30 cm DBH were alive and there was a negative relationship between DBH
and survival. The underlying mechanisms driving this pattern are not clear. Landing rates of
EAB adults on sticky bands affixed to tree trunks were higher on larger ash trees than on smaller
trees in one study (Marshall et al. 2009), but tree size was not related to either EAB adult
captures or larval densities in several other studies (Anulewicz et al. 2007, McCullough et al.
2009a, 2009b, Siegert et al. 2010, Jennings et al. 2014, Burr et al. 2018). Larger and presumably
older white ash trees may have been less vigorous than smaller trees; numerous studies have
shown EAB adults are attracted to and preferentially oviposit on stressed ash trees (McCullough
et al. 2009a, 2009b, Tluczek et al. 2011, Mercader et al. 2013). Additionally, basal area of all
species combined was negatively related to white ash survival rates, which further suggests
competition with other hardwood species or less than ideal growing conditions for white ash
corresponded to higher mortality.
The positive relationship between density of white ash stems and white ash survival
combined with the negative relationship between the diameter of white ash trees and the
proportion of white ash basal area that was alive could suggest that many of the white ash were
too small to be colonized during the peak of the EAB invasion in the early to mid-2000’s. This
seems unlikely, however, given that EAB will oviposit and develop on trees ≥ 2.5 cm DBH
(Cappert et al. 2005, Poland et al. 2006). Further, more than 50-60% of the live white ash trees in
the pole-sized or medium DBH classes were colonized by EAB in previous years, as evidenced
by the bark cracks over old galleries and woodpecker holes. While EAB attraction to stressed
trees is well-accepted (McCullough et al. 2009a, 2009b, 2015, Mercader et al. 2013, 2015),
further research, particularly in recently infested sites, would be useful to more fully understand

23

how the size and relative abundance of white ash trees can affect EAB host selection and
subsequent tree survival.
Distance between sites and the epicenter of the EAB invasion in North America was also
a significant predictor of white ash survival in both multiple regression models. Three of the six
sites where ≥ 90% of the white ash stems and basal area were alive are located more than 100 km
west of the southeast Michigan suburbs originally invaded by EAB (Siegert et al. 2014). In a
previous study, green ash mortality recorded in 2010-2011 generally decreased from the EAB
core in southeast Michigan, where nearly all overstory trees were dead, to southwest Michigan
where 9% to 17% of the trees had been killed in 2010 and 2011, respectively (Burr and
McCullough 2014). Smith et al. (2015), who monitored three ash species and measured an array
of stand-related variables in forested plots in southeast Michigan from 2004 to 2007, reported
that distance from the EAB origin was the only variable significantly related to ash decline and
mortality. Although EAB populations generally advanced outward from southeast Michigan,
human transport of infested ash nursery trees, logs, and firewood established numerous satellite
infestations in other regions of the state (Siegert et al. 2010, 2014). For example, one site I
surveyed where more than 30% of the white ash were dead, was >150 km west of the EAB
origin. Nevertheless, four of the five sites with less than 50% survival of white ash stems (and
basal area) were within 60 km of the EAB origin.
I did not attempt to directly assess parasitism by sampling EAB eggs or debarking trees
to expose EAB larvae, but there was no relationship between white ash survival and the distance
to locations where Asian parasitoids were released or have been recovered. Native parasitoids
such as Atanycolus spp. and the introduced Tetrastichus planipennisi can cause substantial larval
mortality, particularly in heavily infested trees (Duan et al. 2011, Duan et al. 2013, Cappaert and

24

McCullough 2009, Bauer et al. 2015), I am aware of no evidence that parasitoids have slowed
ash mortality rates (Abell et al. 2012, Bauer et al. 2015, Duan et al. 2015) in North America.
Although T. planipennisi appears capable of dispersing a considerable distance, its small
ovipositor (2-2.5 mm length) reportedly limits its ability to parasitize EAB larvae in branches
and boles > 12 cm DBH (Liu et al. 2003, Abell 2012, Duan et al. 2012a, 2013a). Variation in the
type and quality of records in the MapBioControl also varied considerably. Information on
parasitoid recovery was lacking for many release sites and many recovery sites did not
correspond to locations where parasitoids were released. Although I found no evidence that EAB
parasitism affected white ash survival, a limited amount of sampling could be useful in assessing
parasitoid presence and parasitism rates.
In previous studies, individual living ash trees in the midst of dead ash trees were
occasionally observed in forested areas following EAB invasion (Knight et al. 2013). These
“lingering” ash trees were assumed to exhibit at least some level of resistance to EAB (Knight et
al. 2013, 2014, Koch et al. 2012, 2015). Efforts were undertaken to collect seeds or propagate
tissue from these trees (Koch et al. 2012), although some of the lingering ash trees eventually
succumbed (Knight et al. 2014). I did not encounter individual lingering white ash trees, which
were defined by Knight et al. (2014) as an individual healthy ash tree (> 10 cm DBH) in a site
where > 95% of the ash were killed by EAB. In the four sites with high ash mortality (e.g., 55 to
80% of the white ash trees were dead), I still tallied four to seven live trees in my plots.
Moreover, on most of the live white ash, as well as the dead ash, I could readily observe
evidence of past EAB colonization on the tree trunks, including holes left by woodpeckers
preying on late instar EAB larvae, exit holes left by emerged EAB adults and bark cracks over

25

larval galleries. This suggests that while most trees are suitable EAB hosts, larval densities have
not been high enough to cause mortality.
The ability of white ash trees to tolerate some level of EAB colonization may be
facilitated by the highly sectored xylem tissue that characterizes the genus (Zwieniecki et al.
2001, Tanis et al. 2012b, 2016). In previous studies, ash trees produced callus tissue around the
perimeter of longitudinal wounds on the trunk (McCullough et al. 2009a, Tanis et al. 2016) or
laid new wood over plastic plugs or wounds following trunk injections of insecticide (Herms et
al. 2014, Tanis et al. 2016). In this study, approximately 90% of the live white ash trees had
formed wound periderm around the perimeter of old EAB larval galleries, re-establishing
phellogen and cambial integrity (Mullick and Jensen 1973, Mullick 1977, Biggs et al. 1984).
Whether white ash trees will continue to persist in areas of southeast and south central
Michigan over the long term remains unknown. Nearly complete mortality of green ash trees in
and near the sites were surveyed suggest current EAB populations are substantially lower than
during the peak of the invasion when beetles would have been emerging from those trees, as well
as the white ash trees. Populations of EAB remain present in all of the sites (Robinett et al. in
prep.) and densities could presumably build to damaging levels, particularly if trees are severely
stressed by drought or other problems.
White ash regeneration, however, was abundant in all sites I surveyed, suggesting these
sites will retain a white ash component for some time. Mature white ash trees (≤ 20 years old)
typically produce abundant seed one out of every three years (Schlesinger 1990) and while
seedlings are fairly shade tolerant, saplings and recruits require substantial sunlight (Wright
1965, Schlesinger 1990, Griffith 1991, NRCS 2017). In previous surveys in southeast Michigan
and Ohio, nearly complete mortality of white ash, green ash or black ash overstory trees

26

corresponded to depletion of the relatively short-lived ash seeds in seed banks (Knight et al.
2013, Klooster et al. 2014, Smith et al. 2015). An “orphan cohort” of previously established ash
seedlings and saplings remained in these sites, presumably because they were too small to be
colonized by EAB. In contrast to the abundant white ash regeneration in my sites, green ash was
not well-represented in any of the regeneration strata, despite sometimes abundant numbers of
dead overstory green ash. While this partially reflects my plot locations, which were established
in the midst of abundant white ash, the scarcity of live overstory or regenerating green ash
suggests this species has been effectively lost in these areas. White ash dominated not only the
seedling and sapling strata, but also the recruits (stems ≥ 3.0 and ≤ 9.9 cm DBH), which could
potentially replace overstory trees that decline or die. The general health of most overstory white
ash in most sites, along with the abundant white ash regeneration, is encouraging.

27

APPENDIX

28

APPENDIX
Table 1.1: Total number of trees (live and dead) (≥ 10 cm DBH), mean (± SE), minimum and
maximum diameter at breast height (DBH) (cm), total basal area (m2·ha-1) (live and dead) for
white ash trees and the five tree species with the highest relative importance values (RIVs)1
recorded in 28 sites located in southeast and south central Michigan in 2015.
Species
White ash
Green ash
Black cherry
American elm
Sugar maple
Red oak

No. of
stems

No. stems
per ha-1

DBH (cm) ± SE;
min, max

Total basal
area (m2·ha-1)

Relative
importance value

821
373
333
296
354
215

2016.4
916.1
817.9
727.0
869.5
528.1

15.3 ± 0.2; 10, 44
17.8 ± 0.4; 10, 53
18.6 ± 0.5; 10, 65
17.1 ± 0.5; 10, 77
18.3 ± 0.5; 10, 90
26.1 ± 0.4; 10, 92

42.7
27.9
26.9
20.4
28.4
35.8

132.9
88.7
79.5
75.4
58.4
52.7

1

Relative importance represents the sum of frequency, dominance, and density of a species relative
to all other species, following Kent and Coker (1992).

29

Table 1.2: Number of live white ash and green ash trees (≥ 10 cm DBH) with signs of EAB,
including holes left by woodpeckers feeding on EAB larvae, cracks in the bark over EAB larval
galleries, and bark cracks categorized as healing by DBH class, and across all classes.

Live white ash trees
Pole (10.0-20.0 cm)
Medium (20.1-30.0 cm)
Large (> 30.1 cm)
Total

No. live
trees
534
69
13
616

Woodpecker
holes
279
40
9
328

Bark
cracks
367
44
7
418

Healing bark
cracks
333
42
7
382

24
1
4
29

22
1
4
27

18
1
3
22

1
0
1
2

Live green ash trees
Pole (10.0-20.0 cm)
Medium (20.1-30.0 cm)
Large (> 30.1 cm)
Total

30

Table 1.3: Parameter estimates (± SE), t values and P values of variables in a multiple linear
regression to predict percentage of white ash stems alive in 28 sites in Michigan in 2015 (R 2 =
0.49; F = 4.72; DF = 7,20; P = 0.003).
Stand-level predictor variables

Estimate ± SE

t value

P value

Basal area of all species (m2·ha-1)

-0.020 ± 0.006

-3.35

0.003

Distance to EAB core (km)

0.004 ± 0.001

3.13

0.005

White ash stem ha-1

0.004 ± 0.001

2.65

0.015

White ash basal area (m2·ha-1)

-0.150 ± 0.066

-2.26

0.035

Green ash live basal area (m2·ha-1)

-0.869 ± 0.511

-1.70

0.104

Green ash basal area (m2·ha-1)

0.095 ± 0.065

1.46

0.159

-0.0004 ± 0.002

-2.22

0.832

Green ash stems (ha-1)

Table 1.4. Estimated (± SE) parameters, t values and P values (α < 0.05) of variables in a
multiple linear regression to predict the proportion of white ash basal area alive in 28 sites in
Michigan (R2 = 0.61; F = 7.09; DF = 7,20; P < 0.001) (N = 28 sites).
Stand-level predictor variables

Estimate ± SE

t value

P value

White ash stems (ha-1)

0.015 ± 0.003

4.916

< 0.001

Distance to EAB core (km)

0.010 ± 0.004

2.754

0.012

Green ash basal area (m2·ha-1)

0.186 ± 0.183

1.018

0.321

Basal area of all species (m2·ha-1)

-0.018 ± 0.020

-0.912

0.373

White ash dead stems (ha-1)

-0.009 ± 0.010

-0.906

0.376

Green ash stems (ha-1)

-0.003 ± 0.005

-0.510

0.616

Green ash live basal area (m2·ha-1)

-0.385 ± 1.403

-1.274

0.787

31

Table 1.5: Total number and mean (± SE) number of stems per hectare among sites, relative
frequency and density1 of the five most abundant recruits, saplings, and seedlings recorded in 28
sites located in the core region of the EAB invasion in southeast and south central Michigan.
RECRUITS
No. of stems·ha

−1
−1

Mean (± SE) stems·ha
Mean relative frequency ± SE
Mean relative density ± SE

White ash
16,695

American elm
2,192

Hawthorn
1,462

Sugar maple
1,397

Black cherry
731

596 ± 118.9
79 ± 4.6
59 ± 4.8

115 ± 38.2
29 ± 5.5
8 ± 2.0

112 ± 37.7
19 ± 5.1
5 ± 2.2

155 ± 45.1
13 ± 4.5
6 ± 2.8

60 ± 11.3
15 ± 4.5
5 ± 1.6

White ash
47,129

American elm
3,898

Black cherry
1,851

Sugar maple
1,462

Hawthorn
1,186

458 ± 34.3
90 ± 3.5
74 ± 4.4

97 ± 22.5
35 ± 6.5
6 ± 2.4

74 ± 22.2
22 ± 5.4
3 ± 1.2

81 ± 22.4
16 ± 5.6
2 ± 0.8

66 ± 16.3
16 ± 5.0
1 ± 0.7

Ash
1,079,813

Sugar maple
118,861

Black cherry
92,584

Hawthorn
38,310

Elm
35,609

10,186 ± 1262.9
93 ± 3.0
69 ± 4.4

3,495 ± 1484.6
30 ± 5.9
6 ± 2.5

2,153 ± 329.4
37 ± 7.0
9 ± 2.7

2,253 ± 434.8
15 ± 4.9
2 ± 1.2

1,112 ± 279.8
27 ± 5.0
2 ± 1.0

SAPLINGS
No. of stems·ha

−1

Mean (± SE) stems·ha−1
Mean relative frequency ± SE
Mean relative density ± SE
SEEDLINGS
No. of stems·ha

−1
−1

Mean (± SE) stems·ha
Mean relative frequency ± SE
Mean relative density ± SE
1

Relative importance represents the sum of frequency, dominance, and density of a species relative to
all other species, following Kent and Coker (1992).

32

Figure 1.1: Location of 28 sites in 11 counties in Michigan dominated by white ash trees across
southeast and south central Michigan.

33

18.0 m

1 ha

7.0 m
1.7 m

Figure 1.2: Within each site, white ash trees (≥ 10 cm DBH) were surveyed in four
macroplots (18 m radius) in a 1 ha area. Seedlings were recorded in microplots (1.7
m radius), and saplings and recruits were recorded in subplots (7 m radius) centered
in each macroplot.

34

Number of White Ash Stems

500
400
300
200
100
0
≤ 10

11-30

31-50

51-70

71-90

Percent Canopy Dieback
Figure 1.3. Number of white ash trees (≥ 10 cm DBH) by canopy dieback class,
recorded in 28 sites in southeast and south central Michigan in 2015.

35

Percent of White Ash Stems Alive

100%
80%
60%
40%
20%
0%
1 2 3 4 5 6 7 8 9 10111213141516171819202122232425262728
White Ash Field Sites

Figure 1.4: Percentage of white ash stems alive in 28 sites surveyed in southeast Michigan
in 2015. Site order represents high to low percent white ash stem survival.

36

Percentage of Live White Ash Stems

100%
80%
60%
40%
20%
0%
10.0-20.0 cm

20.1-30.0 cm
DBH Class (cm)

> 30.1 cm

Figure 1.5: Percentage of white ash trees alive by DBH class recorded in 28 sites. A
total of 685, 107 and 29 white ash trees were recorded in the pole-sized, medium,
and large DBH classes, respectively.

37

Percent of White Ash Stems Alive

100%
80%
60%
40%
20%

y = -0.048x + 1.4521
R² = 0.4199

0%
10

15

20
Average DBH (cm)

25

30

Figure 1.6. Average DBH (cm) of white ash trees calculated for 28 forested sites located
in southeast and south central Michigan and proportion of live white ash trees.

38

Proportion of Live White Ash Stems

A

1.0
0.9
0.8
0.7
0.6
0.5
0.4

0.3
0.2

y = -0.0008x + 0.7182
R² = 0.0018

0.1
0.0
0

10

20

30

40

50

60

Proportion of Live White Ash Stems

Distance (km)

B

1.0
0.9

B

0.8
0.7
0.6
0.5
0.4
0.3
0.2

y = 0.0016x + 0.6795
R² = 0.0087

0.1

0.0
0

10

20

30

40

50

60

Distance (km)
Figure 1.7: Linear distance (km) between the 28 white ash sites located in southeast and
south central Michigan, and the nearest site where introduced EAB parasitoids were (A)
released and (B) recovered.

39

CHAPTER TWO
Captures of emerald ash borer (Agrilus planipennis) adults
in sites with varying amounts of white ash phloem
ABSTRACT
Emerald ash borer (EAB), (Agrilus planipennis Fairmaire), first identified in 2002 in
southeast Michigan, is now established in 31 states and three Canadian provinces. Catastrophic
rates of ash (Fraxinus spp.) mortality have been recorded in the core of the EAB invasion in
southeast Michigan. While numerous studies have evaluated trap designs and lures for EAB
detection, little is known about the persistence and densities of EAB populations in post-invasion
sites after most ash trees have died. I monitored two baited double decker (DD) traps in the midst
of white ash (F. americana) trees in 30 post-invasion sites in southeast and south central
Michigan from 2014 to 2016. Traps consisted of (1) a dark green prism above a light purple
prism, each baited with cis-3-hexenol, or (2) two dark purple prisms baited with cis-3-hexenol on
the upper prism and Manuka oil on the lower prism. Size and condition of ash trees in an 18 m
radius around each trap were surveyed in 2014 and again in 2016. Area of live white ash phloem
was calculated and related to EAB adult captures. A total of 580, 585, and 932 EAB were
captured in 2014, 2015, and 2016, respectively. Green-purple DD traps captured more EAB
adults than traps with two dark purple prisms. Male beetles accounted for 69% to 77% of the
captures each year and consistently preferred the green panels. Linear relationships between ash
phloem area and EAB captures were not significant, indicating local phloem availability had
little effect on beetle captures.

40

INTRODUCTION
Emerald ash borer (EAB) (Agrilus planipennis Fairmaire) (Coleoptera: Buprestidae), an
invasive phloem-feeding beetle native to Asia, has been found in at least 31 U.S. states and three
Canadian provinces to date (EAB.info 2017), along with western Russia (Baranchikov et al.
2008, Straw et al. 2013, Orlova-Bienkowskaja 2014). Hundreds of millions of ash (Fraxinus
spp.) have been killed and more than 8 billion ash trees in U.S. forests are threatened (Poland and
McCullough 2006, Pugh et al. 2011, Herms and McCullough 2014). In its native range in Asia,
EAB is a secondary pest, colonizing severely stressed or declining ash trees (Baranchikov et al.
2008, Liu et al. 2003, Wei et al. 2004). In North America, EAB adults are attracted to and
preferentially oviposit on stressed ash trees, but can also colonize and kill healthy ash trees
(Cappaert et al. 2005, Poland and McCullough 2006, McCullough et al. 2009a, 2009b, Siegert et
al. 2010, Tluczek et al. 2011, Herms and McCullough 2014).
Although EAB can likely develop on all North American ash species, interspecific
variation in host preference has been documented (Cappaert et al. 2005, Anulewicz et al. 2007,
Rebek et al. 2008, Chen and Poland 2010, Tanis and McCullough 2012, 2015). Of the four
prominent ash species that occur in Michigan and much of the Great Lakes Region, green ash (F.
pennsylvanica Marsh.) and black ash (F. nigra Marsh.) are highly preferred and vulnerable hosts,
while blue ash (F. quadrangulata Michx.) is less preferred and/or more resistant (Anulewicz et
al. 2007, Tanis and McCullough 2012, 2015). White ash (F. americana L.) appears to be an
intermediate host (Limbeck 2010, Tanis et al. 2012, 2015). In some forested sites in southeast
Michigan and northwest Ohio, i.e., the original “core” of the EAB invasion, all or nearly all
white ash trees have been killed by EAB (Anuelwicz et al. 2007, Smith et al. 2010, Tanis and
McCullough 2012, Klooster et al. 2014, Knight et al. 2014). In other forested areas in this region,

41

however, a high proportion of white ash trees are alive and remain relatively healthy, despite the
presence of EAB for at least ten years (McCullough, unpublished data).
Adult EAB generally emerge in late May or June in southern Michigan and feed on the
margins of ash leaves throughout their 3-6 week life span. Larvae feed on phloem and cambium
in serpentine galleries that often score the outer sapwood, disrupting translocation of water and
nutrients. As larval densities build, canopies thin, branches die and eventually the entire tree
succumbs (Cappaert et al. 2005, Poland and McCullough 2006, McCullough et al. 2009a, Herms
and McCullough 2014). The number of EAB larvae that can successfully develop on a given ash
tree, or within a specific area where ash trees occur, depends primarily on the availability of live
phloem for larval development. McCullough and Siegert (2007) used DBH to estimate phloem
area in 148 green ash and white ash trees ranging from 5.2 to 65.0 cm DBH, then determined that
on average, 89 ± 4.6 EAB adults could potentially develop per m 2 of phloem. Empirical data
collected from debarked trees indicated an individual EAB larva requires approximately 10.13
cm2 of phloem for development (McCullough and Siegert 2007, Mercader et al. 2011). Thus, a
large tree can produce more EAB adults than a small tree and the number of EAB adults that
could potentially be produced in a specific area depends on the abundance and size of the ash
trees.
Numerous studies in North America have evaluated survey methods, trap designs and
lures for detection and delineation of recently established EAB infestations. Although a short
range, female-produced pheromone appears to facilitate mating (Lelito et al. 2009, Pureswaran
and Poland 2009, Silk et al. 2009, 2011, Ryall et al. 2012), EAB adults do not produce long
range sex or aggregation pheromones (Otis et al. 2005, Pureswaran and Poland 2009). Instead,
they rely on olfactory and visual cues to locate suitable hosts where they will presumably

42

encounter potential mates (Rodriguez-Saona et al. 2007, Crook et al. 2009, Crook and Mastro
2010).
Methods to detect low-density EAB populations include visual surveys of ash trees to
identify external signs of EAB infestation, girdled ash trees and artificial traps. External signs of
EAB infestation such as holes left by woodpeckers preying on late instar EAB larvae, small Dshaped holes left by emerging EAB adults, declining canopies and epicormic sprouts are
generally not apparent on recently infested trees with low densities of EAB. Girdling ash trees by
removing a band of outer bark and phloem around the tree trunk in spring stresses the tree, which
alters volatile profiles and increases EAB attraction (Crook et al. 2008) and attracting EAB
adults (McCullough et al. 2009a, 2009b, 2011, Poland and McCullough 2010, Rauscher 2006,
Tluczek et al. 2011, Mercader et al. 2013, 2015). Debarking girdled ash trees in fall to locate
EAB larval galleries is the most effective method of detecting or monitoring low density
populations (McCullough et al. 2009a, 2009b, 2011, 2015, Mercader et al. 2013, 2015), but
locating suitable trees for girdling can be challenging, especially for multi-year surveys.
National EAB detection programs in the U.S. and Canada primarily employ artificial
baited traps that incorporate visual and olfactory cues to attract EAB adults (USDA APHIS
2017, CFIA 2016). Most EAB detection programs currently use prism traps or funnel traps in
specific shades of green or purple that are coated with clear insect glue and hung from a branch
in an ash tree (Crook et al. 2009, 2012, 2014, Cook and Mastro 2010, Francese et al. 2010,
2013a, 2013b, 2014, Crook et al. 2014). A green leaf volatile, cis-3-hexenol, is most commonly
used in the U.S. to bait EAB traps (de Groot et al. 2008, Grant et al. 2010, 2011, Poland et al.
2011). In past years, EAB traps were also baited with lures containing Manuka oil, an extract of
the New Zealand tea tree (Leptospermum scoparium J.R. Forst and G. Forst), which contains

43

sesquiterpenes associated with ash wood and bark that are attractive to EAB beetles (Crook et al.
2008, de Groot et al. 2008, Crook and Mastro 2010, Grant et al. 2010, Poland et al. 2011).
An alternate trap design, known as the double decker trap (DD), is comprised of two
sticky prisms attached to a 2.4-m polyvinyl chloride (PVC) pipe that slides over a t post for
support (McCullough and Poland 2009, 2017, Poland et al. 2011). These traps, which resemble
the silhouette of a small tree, are intended to be placed within 5-10 m of ash trees but in an open
area to exploit the preference of EAB for sunny conditions (McCullough et al. 2009, Poland et
al. 2011) and to ensure visual and olfactory cues are not overwhelmed by live ash trees
(McCullough et al. 2009a, 2009b, Poland et al. 2011, 2014, McCullough and Poland 2017).
Multiple studies across dozens of sites have shown detection rates or EAB adult captures
(standardized by trapping surface area) are higher on DD traps baited with cis-3-hexenol or a
combination of cis-3-hexenol and Manuka oil than on baited purple or green prisms or funnel
traps hung from ash branches (Marshall 2010b, McCullough et al. 2011, Poland et al. 2011,
2014, Poland and McCullough 2017, Wieferich et al. 2017)
Although nearly all trapping studies to date have focused on EAB detection, I was
interested in evaluating the relationship between EAB captures and the availability of live ash
phloem in post-invasion sites. I hypothesized that live ash phloem represents a critical but
limited resource for EAB populations in post-invasion areas and that EAB adult captures would
be higher in areas where phloem remains abundant, while few adults would be captured in sites
where nearly all ash trees are dead. Conversely, the lack of a relationship between EAB captures
and phloem area could suggest that traps attract EAB adults from beyond the vicinity of the traps
or that local EAB populations are not limited by phloem availability and that other factors may

44

be influencing densities. I monitored two double decker trap designs in 30 post-invasion sites
over three seasons and evaluated relationships between EAB captures and live ash phloem.

45

METHODS
Site selection
I used a Michigan county atlas in April 2014 to identify 70 large, forested areas on state,
county, or municipal property in southeast and south central Lower Michigan, then queried
managers of each area to assess white ash tree presence and abundance. I visited 40 sites where
white ash occurred and identified 30 areas where white ash trees, whether dead or alive,
appeared to be relatively abundant (Fig 2.1). I recorded GPS coordinates in the midst of the
white ash trees in each site.
Trapping
At each of the 30 sites, I set up two double decker (DD) traps, 36 to 50 m apart, between
24 May and 12 June in 2014 and recorded GPS coordinates. Traps were placed in the same
locations in 2015 and 2016. Traps remained in place through mid-September each year. Traps
were placed in full sunlight, 5-10 m from the edge of wooded areas or in open areas among
scattered ash trees. Each DD trap consisted of two coroplast panels (36 × 60 cm panels) (Great
Lakes IPM, Vestaburg, MI) folded into three-sided prism (36 × 60 cm on each side). Prisms were
zip-tied to a 3 m tall PVC pipe (10 cm diam); the upper and lower prisms were 3.0 m and 1.8 m
from the ground, respectively. The PVC pipe was supported by sliding the pipe over a t-post
(McCullough and Poland 2017). Prism surfaces were coated with clear Pestick™ (Hummert
International, Earth City, MO) when the traps were set up. One DD trap was comprised of an
upper dark green prism and a lower light purple prism. Two cis-3-hexenol bubble caps (release
rate of 3.7 mg/d per bubble cap for a combined release rate of 7.4 mg/d; Contech Enterprises
Inc., Delta, BC) were attached to the lower edge of both prisms. The second DD trap was
comprised of two dark purple prisms. The top prism was baited with two cis-3-hexenol bubble

46

caps while the lower prism was baited with a Manuka oil pouch (release rate of 50 mg/d,
Synergy Semiochemicals, Corp.). Lures were replaced mid-season to ensure consistent volatile
release and Pestick™ was re-applied mid-season. The same trap designs were placed in the same
locations in all three years.
Captured EAB beetles were collected at two to three week intervals from early June to
mid-September each year. Beetles were typically collected from multiple sites on a given day but
it required approximately a week to check traps at all 30 sites. Beetles on each prism of the two
traps were bagged separately and returned to the Michigan State University (MSU) Forest
Entomology Lab. Beetles were soaked in Histoclear II (National Diagnostics, Atlanta, GA) to
remove Pestick™ and examined under a microscope to confirm species and sex.
I acquired cumulative growing degree days (GDD) (base 50 ºF/10 ºC) for each site from the
nearest MSU Enviro-weather station (Enviroweather.msu.edu). The Euclidean distance between
each site and nearby weather stations were determined using ArcMap 10.3 and ranged from 5 to
43 km. I recorded the cumulative GDDs (base 50 ºF /10 ºC) for each collection period.
Fixed radius plots
I centered a circular plot with an 18 m radius around each of the two DD traps and
recorded DBH of white ash trees (≥ 6 cm DBH). I visually estimated canopy dieback in
increments of 10%, where < 10% indicated a full, healthy canopy and ≥ 80% indicated severe
decline (Zarnoch et al. 2004). Ash trees were assigned to dieback classes (0-20%, 21-40%, 4160%, 61-80%, 81-100%) for some analyses. Each ash tree was examined for signs of EAB
infestation, including holes left by woodpeckers preying on EAB larvae and D-shaped adult EAB
emergence holes. Trees were re-surveyed from early July to mid-August in 2016.

47

Density of white ash stems and white ash basal area (m 2·ha-1) were calculated for each
plot and summed for each site in 2014 and 2016. White ash trees were assigned to four DBH
classes (6.0-13.0 cm, 13.1-25.0 cm, 25.1-42.0 cm, 42.1 to 60.0 cm). Area of phloem (m2) in live
and dead white ash trees was calculated using DBH for trees recorded in 2014 and 2016
following methods of McCullough and Siegert (2007). I also adjusted the area of live phloem to
account for canopy dieback in individual trees. When dieback estimates were > 10%, I multiplied
the estimated phloem area by the percentage of the canopy that remained healthy, e.g., the
estimated live phloem area was reduced by the proportion of the canopy that was dead. I also
estimated live phloem area for trees assigned to the five canopy dieback classes, using the
midpoint of each dieback class.
Statistical analysis
Adult male and female beetles captured per trap (both prisms) and per prism were tallied
for each collection period and summed by year. Differences in numbers of EAB captured per
trap or per prism were compared among years and trap types, and among prism colors or lures
within trap types using separate generalized linear mixed models (PROC GLIMMIX). In
separate models, the main effects of trap type, year, prism color, position (upper or lower), and
lure were tested as fixed effects. Site was a random effect in all models. To account for
collinearity among years, an autoregressive order one covariance structure was used for all
models, with the Kenward-Rogers modification of the denominator degrees of freedom
(Alnosaier 2007). Because the response variable was count data, the negative binomial
distribution was used with the log link function (Hilbe 2011). Differences among years, trap
types, prism color within trap types, and lure within trap types for the separate models were

48

tested with the Tukey-Kramer means comparison procedure. Residuals were tested for
homogeneity of variance assumption using Levene’s test and models were adjusted if needed.
Simple linear regression was used to assess whether the area of live ash phloem (m2)
significantly predicted EAB adult captures within sites (PROC REG). Captures of EAB in 2014,
2015, and 2016 were compared to total area of live phloem measured in 2014 and EAB captures
in 2016 were regressed on phloem area measured in 2016. Cumulative GDD for collection dates
each year were plotted to identify when EAB adult captures peaked. All analyses were conducted
using the SAS 9.4 for Windows statistical package (SAS Institute, 2012) with an α-level of 0.05.

49

RESULTS
Trapping
Overall, 580, 585, and 932 EAB adults were captured on the 60 DD traps in 2014, 2015,
and 2016, respectively. Significantly more beetles were captured in 2016 than in either 2014 or
2015 (F = 6.08; df = 2, 42; P = 0.0048) (Fig 2.2). In 2014, no EAB adults were captured in two
sites, while one to 61 beetles were captured in the other 28 sites. On average, 19 ± 2.8 beetles
were captured per site in my first year of trapping. In 2015, EAB beetles were captured in 29
sites; the site with zero EAB captures in 2015 also had no EAB captures in 2014. Total beetles
captured per site ranged from one to 64, with an average of 19 ± 3.0 per site. Between 2014 and
2015, EAB captures increased by ≥ 50% in 12 sites, remained consistent in four sites and
decreased by ≥ 50% in five sites. In 2016, EAB adults were captured in all sites and number of
captured beetles increased by approximately 60% in 14 of the 30 sites from 2014 to 2016 (Fig
2.2). I captured 191 and 194 more EAB adults in 2016 than in 2014 and 2015, respectively. Total
captures ranged from one to 127 EAB adults per site and averaged 31 ± 5.3 beetles per site.
The cumulative number of EAB adults captured throughout the season differed between
trap types (F = 24.91; df = 1, 178; P < 0.0001) and among years (F = 8.86; df = 1,14; P =
0.0034), but the interaction was not significant (P = 0.636). Overall, the PG trap caught
significantly more EAB adults than the PP trap, accounting for 73%, 68%, and 67% of total
captures in 2014, 2015, and 2016, respectively. The green prisms on the PG traps accounted for
40% and 35% of the total EAB captures in 2014 and 2015, respectively. On average, 7.7 ± 1.3
and 6.9 ± 1.3 EAB adults were captured on the green prisms in 2014 and 2015, respectively,
while captures on the lower light purple prisms averaged 6.4 ± 1.1 in both years. In 2016,
however, the light purple prisms on the PG traps caught an average of 11.2 ± 2.2 EAB adults,

50

which accounted for 36% of the total beetle captures (Table 2.1). On the PP traps, the upper
prisms accounted for 14%, 19%, and 20% of total EAB captures, while the lower prisms
captured for 12%, 13%, and 12% of the beetles in 2014, 2015, and 2016, respectively. On
average, 2.8 ± 0.7, 3.8 ± 0.9, and 6.3 ± 1.2 EAB adults were captured on the upper prisms, while
2.4 ± 0.8, 2.5 ± 0.6, and 3.9 ± 0.9 EAB adults were captured on the lower prisms in 2014, 2015,
and 2016, respectively.
Overall, individual prisms baited with cis-3-hexenol lures captured more EAB adults than
prisms baited with Manuka oil (F = 15.50; df = 3, 87; P < 0.001) and green prisms and light
purple prisms baited with cis-3-hexenol captured more beetles than the dark purple prisms baited
with either cis-3-hexenol or with Manuka oil. The trap designs and lures represented operational
trapping regimes (USDA APHIS 2017), but prism colors, position and lures were confounded,
making it difficult to fully determine which factor was most relevant to beetles. Results do
suggest the PG trap was more attractive than the PP trap given that the cis-3-hexenol was more
attractive than the Manuka oil lure on the PG trap, whereas cis-3-hexenol and Manuka oil
captured similar numbers of beetles on the PP traps.
Male beetles dominated captures on both trap designs in all three years. Overall, 70%,
69%, and 77% of the EAB adults captured in 2014, 2015, and 2016, respectively were males.
Male captures increased in 21 of the 30 sites from 2014 to 2016. Number of males captured
increased by ≥ 50% in 13 sites and decreased by ≥ 50% in four sites, while two sites stayed the
same. Two sites increased from zero males captured in 2014 to one captured in 2016, while one
site increased from zero males captured in 2014 to eight sites in 2016. Females captured
increased in 18 sites and remained stable in 2 sites from 2014 to 2016 (Fig 2.2). Number of
females captured increased by ≥ 50% in 16 sites and decreased by ≥ 50% in three sites. One site

51

increased from zero females captured in 2014 to one captured in 2016, while one site increased
from zero females captured in 2014 to five captured in 2016. Green prisms caught significantly
more males than the light purple prisms or either of the dark purple prisms (F = 15.50; df = 3,87;
P < 0.0001). Significantly more female EAB adults were captured in 2015 and 2016 than in 2014
(F = 4.57; df = 2,58 P = 0.014). Total number of females captured differed among years and
prisms, but were not affected by the interaction (P = 0.837). Light purple prisms on the PG traps
caught significantly more EAB females than either of the dark purple prisms on the PP traps or
the upper green prisms on the PG traps (F = 11.7; df = 3,87; P < 0.0001) (Table 2.1).
EAB adult captures and growing degree days
I began collecting EAB from the DD traps on 10 June each year. Sites were visited to
collect captured beetles following the same order used when I set up the traps. Cumulative
growing degree days (GDD) (base 50ºF/10ºC) associated with the first trap checks in each site
ranged from a minimum of 654 to 941 DD50 (Table 2.2). Peak EAB activity occurred in late June
to early July each year, with EAB captures accounting for 54%, 52%, and 62% of the total EAB
captured in 2014, 2015, and 2016, respectively. Following this peak in 2016, EAB adults
captured during my third check occurring in August decreased by 85%, whereas in 2014 and
2015, EAB adults captured decreased by 48% and 39%, respectively. Across all three years, I
caught ≤ 21 EAB during my final check in September (ending on September 17, 11, 23 in 2014,
2015, and 2016, respectively).
White ash phloem
Abundance and condition of white ash trees in the plots centered on the two traps in each
site varied substantially. In 2014, a total of 1,109 white ash trees (≥ 6 cm DBH) were measured,
including 477 trees with ≤ 10% canopy dieback and 657 trees with ≤ 50% canopy dieback.

52

Conversely, 44 white ash trees were severely declining with ≥ 80% canopy dieback and an
additional 373 trees are dead. External signs of EAB infestation were apparent on 58% of the
736 live white ash trees recorded, while 98% of the dead white ash trees were killed by EAB. Of
the 304 white ash trees showing 10% to 50% canopy dieback, external signs of EAB infestation
were apparent on 78%, while 98% of the 44 white ash trees showing 80% to 90% canopy
dieback had external signs of EAB infestation (Table 2.3).
In 2016, number and condition of white ash trees were similar to those in 2014. There
was a slight increase in the number of white ash trees recorded from 2014 to 2016 because of
young, healthy trees that were > 6 cm DBH in 2016 but not in 2014. In 2016, I measured a total
of 1,187 white ash trees, including 662 white ash trees with ≤ 10% canopy dieback, and 726 trees
with ≤ 50% canopy dieback. There were seven white ash trees that were heavily infested and
severely declining with ≥ 80% canopy dieback, and an additional 430 that are dead (Table 2.3).
External signs of EAB infestation were apparent on 71% of the 757 live white ash trees, and 99%
of the dead white ash trees. Of the 118 white ash trees showing 10% to 50% canopy dieback,
external signs of infestation were apparent on 94%, while all 15 white ash trees in severe decline
showing 80% to 90% canopy dieback were infested with EAB.
Estimated area of live white ash phloem within sites was similar in 2014 and 2016. In
2014, I recorded a total of 3,627.2 m2 of phloem, of which, 1,832.5 m2 was alive. A total of 684.8
m2 of phloem recorded was alive in the small DBH class (6.0 to 13.0 cm), while 279.1 m 2 of
phloem recorded was alive in the two large DBH classes combined (25.1 to 42.0 and 42.1 to
60.0). Based on the trees recorded in 2014 and phloem area, potentially 10,761 ± 1,558.4 EAB
could have emerged across the sites, ranging from a minimum of 1,673 to a maximum of 37,466
within sites. When adjusting phloem to account for canopy dieback, total phloem decreased to

53

1,566.0 m2, and the total live phloem decreased to 1,382.8 m 2. In the small DBH class, the
amount of live adjusted phloem recorded was 557.3 m2, and 195.4 m2 in the two large DBH
classes combined.
In 2016, I recorded a total of 3,569.1 m2 of phloem, of which 1,732.1 m2 was alive. The
small DBH class (6.0 to 13.0 cm) accounted for 728.6 m2 of the live phloem recorded, while the
two large DBH classes combined (25.1 to 42.0 and 42.1 to 60.0 cm) accounted for 259.2 m 2 of
the live phloem. Based on the trees recorded in 2016 and phloem area, potentially 10,881 ±
1,434.9 EAB could have emerged across the sites, ranging from a minimum of 2,095 to a
maximum of 36,567 within sites. When adjusting phloem to account for canopy dieback, total
phloem decreased to 1,641.1 m2, while the total live phloem recorded decreased to 1,457.4 m 2.
Live phloem adjusted for canopy dieback decreased to 635.9 m 2 for the small DBH class, and
201.4 m2 for the two large DBH classes combined.
The level of infestation and condition of white ash trees varied from nearly all trees alive
and in relatively good condition, to ≥ 90% mortality. In my third year of trapping, I captured
EAB in all sites using baited DD traps, even where nearly all trees were dead. Relationships
between the area of live white ash phloem derived from the 2014 surveys and the number of
EAB adults captured were weak in 2014 and 2015 (Fig 2.3). Despite the notable increase in EAB
captures in 2016, the area of live white ash phloem available for EAB larval development
explained relatively little variation in EAB adult captures in 2016 (P = 0.166), or when
comparing 2016 EAB captures to 2014 phloem area (P = 0.189) (Fig 2.3). I re-ran each
regression excluding the one outlier site with the highest amount of live phloem. This resulted in
a significant relationship between 2014 white ash phloem and 2015 EAB captures (y = 0.695x +
27.003; R2 = 0.16; P = 0.032).

54

When I adjusted the area of live white ash phloem (m2) measured in 2014 to account for
canopy dieback, the relationship with EAB captures remained weak in 2014, as well as 2015 and
2016. Similarly, EAB captures in 2016 were not strongly related to the adjusted area of live
phloem calculated from the 2016 survey (P = 0.241) (Fig 2.4). After re-running the regression
excluding the outlier site, the relationship between adjusted 2014 live phloem and 2015 EAB
captured was significant (y = 0.707x + 27.765; R2 = 0.16; P = 0.031).

55

DISCUSSION
Populations of both EAB and white ash trees continue to persist in these post-invasion
sites. I found that white ash phloem was not a limiting factor for EAB larval development. While
EAB are likely to colonize ash trees within 200 m of their point of emergence when phloem is
abundant (Siegert et al. 2010), it is likely captured EAB dispersed from parent trees outside my
surveyed plots. Across the 30 sites, the level of infestation and condition of white ash trees
varied from nearly all trees alive and in relatively good condition, to ≥ 90% mortality. However,
the condition of ash trees recorded in my plots was similar in 2014 and 2016, despite capturing
significantly more EAB adults in 2016 compared to the previous two years. This may be a result
of larvae developing in two years rather than one, or from beetles dispersing from ash trees
surrounding these sites. I found that green-light purple (PG) traps lured with cis-3-hexenol
captured significantly more EAB adults than the purple-purple (PP) traps.
While the availability of ash phloem for larval development would be a limiting resource
if it was related to EAB captures, I found this not to be the case. In both 2014 and 2016, one of
the 30 sites had low EAB captures with roughly 175 m2 of live white ash phloem area available.
When excluding this site to compare total EAB adults captured and live white ash phloem, there
was a significant relationship between phloem and adjusted phloem calculated from 2014 trees
and EAB adult captures in 2015. However, when comparing total EAB adult captures in 2016 to
white ash phloem available in both 2014 and 2016, the relationship was weak. Despite the
abundance and availability of phloem in many of the sites, trees continue to persist and remain in
relatively good condition.
Most EAB females will select ash tree hosts close to their point of emergence to oviposit
their eggs (Mercader et al. 2009). In a simulation study conducted by Mercader et al. (2011),

56

mortality rates and the relationship between the proportion of EAB larvae that develop in one or
two years and larval density for a site were found to likely be affected by the diffuse distribution
of ash phloem. Siegert et al. (2010) reported a significant bias in the dispersal of EAB towards
areas where ash phloem was relatively abundant within 200 m of their emergence point,
however, larvae were also reported up to 650 m from the origin. Mercader at al. (2009) reported
70% of larval galleries were on trees within 100 m of the point of emergence, while a single
larvae was reported 750 m from the point of emergence. I focused only on white ash trees
recorded within radius plots extending out 18 m from each trap. Since live white ash trees were
abundant in many of the sites I surveyed, it is possible EAB adults captured on my traps emerged
from live white ash trees present beyond my 18 m plots.
Determining the amount of total and live phloem (m 2) for an individual tree, or for
multiple trees, allows us to estimate the area of phloem available for EAB development on ash
trees of varying diameter classes. McCullough and Siegert (2007) found that, on average,
roughly 89 EAB could successfully develop and emerge per m2 of phloem surface area, and that
EAB production averaged 68.8 ± 5.9 beetles per m2 on small trees (2.5-13 cm DBH) and 105.3 ±
5.7 beetles per m2 on larger trees (DBH > 13 cm). Based on the trees recorded and phloem area
calculated, I found on average 10,761 ± 1,558.4 and 10,881 ± 1,434.9 EAB could have
potentially emerged across the sites in 2014 and 2016, respectively. The DBH of live white ash
trees recorded in sites ranged from 6.0 to 40.6 cm, with 80% having a DBH of ≤ 13 cm. Young
white ash trees generally have relatively smooth bark on the branches and much of the trunk.
When young trees are attacked by EAB, females are likely to lay most eggs on the trunk of the
tree, both because branches are too slender and because females usually prefer rough-textured
bark near the base or around branch crotches (Anulewicz et al. 2006. Larval feeding on tree

57

trunks and its effects on water and nutrient translocation can result in rapid mortality. Despite
this, many of the white ash trees recorded in the sites remain in good condition from my initial
plots conducted in 2014 to when I repeated these plots in 2016. Additionally, many show signs
of previous infestation, but are still persisting.
Development of an efficient and effective artificial trap for capturing and monitoring
EAB adults has proven to be a challenging task. Poland et al. (2011) found significantly more
EAB were captured on DD traps that are designed to imitate the silhouette of a tree compared to
canopy traps hung in infested ash trees. Poland and McCullough (2014) found both green and
purple double decker traps were the most successful trap design in capturing EAB adults when
compared to canopy traps. Multiple studies have shown males to be consistently attracted to dark
green panels, while females are attracted to the purple (light and dark) panels (Crook et al. 2009,
Grant et al. 2009, Francese et al. 2010). In my study, the PG trap which combines both green and
purple panels captured significantly more beetles than the PP trap. I found the green panels of the
PG traps captured significantly more males, while the light purple panels captured significantly
more females. Males tend to hover near the canopy of ash trees (Rodriguez-Saona et al. 2007),
resulting in higher proportions captured on green traps, while females are captured in higher
proportions on purple prisms that imitate a tree trunk or branches where they oviposit their eggs
(Crook and Mastro 2010, Francese 2010, Grant et al. 2011, Poland and McCullough 2014).
Additionally, EAB adult activity is greatest in full sun (Yu 1992). Field studies showed higher
numbers of EAB adults are captured on trap trees that are exposed to sunlight compared those
growing in a shaded canopy (McCullough et al. 2009a, 2009b). Chen and Poland (2009) found
EAB adults preferentially feed on ash leaves grown in sun over leaves grown in shade (Chen and
Poland 2009). In contrast to traps hung on a branch in the shaded canopy of ash trees, double

58

decker traps are placed in full sun, making them visually apparent to EAB adults (McCullough
and Poland 2009, McCullough et al. 2011, Poland et al. 2011), giving this design an advantage.
In addition to color, lure type is also important in the success of the trap (Poland and
McCullough 2014). Both the top and bottom panels of the PG trap and the top panel of the PP
trap were baited with cis-3-hexenol, which is a lure that imitates volatiles released by stressed
ash trees. Males are specifically attracted to cis-3-hexenol, and are often hovering the canopy of
ash trees feeding on foliage (Lelito et al. 2008, Rodriguez-Saona et al. 2007), which may explain
the abundance in male captures on panels baited with this lure. Poland et al. (2014) compared
DD traps of different colors and found dark green-light purple DD traps lured with cis-3 hexenol
to be the most effective DD trap, with a detection rate of 90%, while dark purple-dark purple DD
traps lured with cis-3 hexenol and Manuka oil, or cis-3-hexenol on both the top and bottom
prisms were the least attractive of the DD trap color combinations, with detection rates of 70%
and 60%, respectively. I also found the use of both dark green and light purple panels, both lured
with cis-3-hexenol, to be significantly more attractive to EAB adults than DD traps with dark
purple top and bottom panels lured with both cis-3-hexenol and Manuka oil.
Emergence of EAB typically begins in mid-May when accumulated growing degree days
(GDD) (base 50 ºF/10 ºC) reach 450 in southern Michigan. As in previous studies, EAB adult
activity peaked in late June and early July in all three years (McCullough et al. 2009a, Poland et
al. 2011). In 2016, however, the proportion of EAB adults captured during my first check
beginning June 10 was double that of the previous two years, despite similar accumulation of
GDD during this time across all three years. While this is notable, it could be a reflection of
differences in sun or daily temperatures during this time that resulted in greater EAB adult
emergence. Though EAB generally has a one-year life cycle, development in healthy ash has

59

been found to take two years (Cappaert et al. 2005, Siegert et al. 2010, Tluczek et al. 2011).
Based on conclusions from these previous studies, abundance of EAB emerging earlier in 2016
may be explained by the completion of a two-year life cycle in these sites due to the vigor of the
surviving white ash trees reflecting the low population densities (Cappaert et al. 2005, Wei et al.
2007, Tluczek et al. 2011). Destructive sampling or bark windows of ash trees within these sites
would have to be conducted to verify life stages to identify and quantify density of two-year
galleries. Based on the difference in the population trend of EAB when comparing captures from
2014 and 2015 to 2016, it is difficult to determine whether populations will continue to build in
these sites. Additional years of trapping would be beneficial to project the future condition of
these stands.
While phloem is readily available for EAB development in these sites located in the core
EAB invasion area, white ash trees seem to be coexisting with EAB populations. Natural
enemies of EAB larvae, including woodpeckers (Cappaert et al. 2005, Duan et al. 2010 Lindell et
al. 2008, ), native parasitoids (Cappaert and McCullough 2009, Duan et al. 2012) and introduced
parasitoids native to Asia (Yang et al. 2005, Yang et al. 2006, Bauer et al. 2008) could be
keeping EAB populations below carrying capacity within these sites. Destructive sampling
would have to occur to know for sure. Monitoring of these trees should continue in the future.

60

APPENDIX

61

APPENDIX
Table 2.1: Mean (± SE) number of adult emerald ash borer (EAB) males and
females captured per prism in 2014, 2015, and 2016 in 30 sites in Michigan.

Green
Upper

Prisms
Light Purple
Lower

Dark Purple
Upper

Dark Purple
Lower

Males

7.1 ± 1.2

4.5 ± 0.9

2.2 ± 0.5

1.6 ± 0.6

Females

0.7 ± 0.2

1.9 ± 0.4

0.6 ± 0.2

0.7 ± 0.3

Males

5.5 ± 1.1

4.1 ± 0.8

2.3 ± 0.6

1.3 ± 0.4

Females

1.4 ± 0.3

2.3 ± 0.4

1.4 ± 0.3

1.1 ± 0.3

Males

8.7 ± 2.0

7.9 ± 1.6

4.6 ± 0.9

2.8 ± 0.8

Females

1.0 ± 0.2

3.3 ± 0.7

3.9 ± 0.8

1.0 ± 0.3

2014

2015

2016

62

Table 2.2: Number of male and female emerald ash borer adults captured by month and year.
Accumulated growing degree days (base 50 ºF/10 ºC)¹ from the nearest MSU Enviro Weather
station were compiled for each date.
N = 30 sites
Trap Check
2014
June
July
August
September

Min; Max GDD

Male

Female

Total

625.3; 812.6
938.0; 1114.8
1256.0; 1483.9
1861.0; 2335.6

62
263
123
13

17
53
42
7

79
316
165
20

2015
June
July
August
September

669.5; 941.5
968.2; 1392.7
1475.9; 1964.3
2006.5; 2489.0

61
212
119
7

15
96
69
6

76
308
188
13

2016
June
July
August
September

654.3; 856.6
1045.7; 1302.8
1710.7; 2036.2
2308.7; 2891.3

232
141
37
15

38
438
25
6

270
579
62
21

¹Accumulted growing degree days base (50ºF) is the heat accumulation above the minimum
threshold of 50ºF, beginning on January 1.

63

Table 2.3: Number of white ash (F. americana) trees, estimated total and adjusted area of
phloem (m2) recorded in fixed radius plots (18 m radius) in 30 sites in Michigan in 2014 and
2016. Phloem area was calculated following methods of McCullough and Siegert (2007),
then adjusted to account for percent canopy dieback class.
2014
Percent Canopy
Dieback Class

No. Trees

Total Phloem

Adjusted Live
Phloem

0.0-20.0%

555

1243.1

1118.8

20.1-40.0%

86

254.4

178.1

40.1-60.0%

31

91.7

45.9

60.1-80.0%

39

97.1

29.1

80.1-100.0%

398

1940.9

194.1

0.0-20.0%

693

1503.7

1353.3

20.1-40.0%

22

76.9

53.8

40.1-60.0%

20

65.0

32.5

60.1-80.0%

15

45.7

13.7

80.1-100.0%

437

1877.8

187.8

2016

64

Figure 2.1: Distribution of 30 white ash sites in southeast and south central Lower Michigan.

65

Average EAB Captures

30
a

25
20
a

a

15

ab

10

b

b
5
0
PG

PP

2014

PG

PP

2015

PG

PP

2016

Trap Type by Year
Figure 2.2: Total adult emerald ash borer captured in 2014, 2015, and 2016 on
purple-green double decker traps (PG) and purple-purple double decker traps (PP).
Bars topped by the same letter are not significantly different, P > 0.05.

66

Total EAB Adults Captured

A)
y = 0.10x + 14.64
R2 = 0.07

70
60
50
40
30
20
10
0
0

50

100

150

200

B)
y = 0.16x + 12.34
R2 = 0.13

Live Phloem (m2)
Figure 2.3: Linear relationship between total emerald ash borer adult captured in 2014-2016 and
estimated area live white ash phloem (m 2) surrounding traps (18 m radius) in 30 sites in Michigan.
Area of live white ash phloem (m2) calculated from 2014 plot data is compared to adult emerald
ash borers captures in (A) 2014, (B) 2015 and (C) 2016. Area of live white ash phloem (m 2)
calculated from 2016 plot data is compared to 2016 EAB captures in 2016 (D). Live white ash
phloem area (m2) was estimated following the methods of McCullough and Siegert (2007) in 2014
and again in 2016.

67

Figure 2.3 (cont’d)

C)

Total EAB Adults Captured

y = 0.19x + 22.46
R2 = 0.06

D)
140
120
100
80
60
40
20
0

y = 0.14x + 24.38
R2 = 0.04

0

50

100

Live Phloem (m2)

68

150

200

70

A

60

y = 0.09x + 14.48
R² = 0.06

50
40

Total EAB Adults Captured

30
20
10
0
0

50

100

150

70

200

B

60

y = 0.15x + 11.70
R² = 0.13

50
40
30
20
10
0

0

50

100

150

200

Live Phloem (m2)
Figure 2.4: Total adult emerald ash borer captures compared to total live white ash phloem area (m 2)
adjusted for percent canopy dieback in 30 sites located in southeast and south central Michigan.
Area of live white ash phloem (m2) calculated from 2014 plot data adjusted for percent canopy
dieback is compared to total adult emerald ash borer captured in 2014 (A), 2015 (B), and 2016 (C).
Area of live white ash phloem (m2) calculated from 2016 plot data adjusted for percent canopy
dieback is compared to total EAB captures in 2016 (D). Live white ash phloem area (m2) was
estimated using diameter at breast height (DBH, cm) of white ash trees in two plots (18 m radius)
surrounding each trap in each site, and was adjusted to incorporate percent canopy dieback by
multiplying the percentage of the canopy not showing dieback. Phloem area was calculated
following the methods of McCullough and Siegert (2007) in 2014 and again in 2016.

69

Figure 2.4 (cont’d)
140

C

120
100

y = 0.17x + 22.25
R² = 0.05

80

Total EAB Adults Captured

60
40
20
0
0

50

100

150

140

200

D

120
100

y = 0.15x + 23.05
R² = 0.05

80
60
40
20
0
0

50

100

Live Phloem (m2)

70

150

200

CHAPTER THREE
Variables related to white ash survival rates in forested sites
in the core emerald ash borer invasion area
ABSTRACT
The invasive emerald ash borer (EAB) (Agrilus planipennis Fairmaire) (Coleoptera:
Buprestidae) has caused catastrophic mortality of ash (Fraxinus spp.) trees since its arrival in the
Detroit-metro area of southeast Michigan in the 1990’s, and currently threatens over 8 billion ash
trees in U.S. forests. Despite previous studies reporting nearly 100% mortality of prominent ash
species in Michigan following the initial wave of EAB infestation, an abundance of surviving
white ash trees have been recorded in sites located in this area. Determining site-level variables
associated with relatively high rates of white ash survival could be valuable, particularly for
forest managers. I collected data in intensive plots, variable radius plots, and driving surveys in
2014 and 2015 in 28 sites in southeast and south central Michigan. I downloaded land type
satellite data to fit models at five scales to determine whether any recorded variables consistently
predict white ash stem survival rates. Results showed that as distance increased from intensive
plots with abundant white ash, green ash presence, specifically those growing along roads,
negatively affected white ash survival. The proportion of area classed as woody wetlands and the
presence of white ash trees, specifically those along roads, are also important variables. These
variables should be considered in decisions related to white ash management.

71

INTRODUCTION
The emerald ash borer (EAB) (Agrilus planipennis Fairmaire) (Coleoptera: Buprestidae)
has caused catastrophic damage in U.S. forests (Aukema et al. 2011, Herms and McCullough
2014) following its arrival in southeast Michigan during the mid-1990’s (Cappaert et al. 2005,
Siegert et al. 2014). Currently, EAB populations are found in 31 states and three Canadian
provinces (EAB.info 2017). Hundreds of millions of ash (Fraxinus spp.) trees have died, and
more than 8 billion ash trees across U.S. forests are threatened by this Asian, phloem-feeding
beetle native to Asia (Poland and McCullough 2006). Projections suggest an estimated 10.7
billion USD will be spent by 2019 for removal or treatment of roughly half of the ash landscape
trees likely to be affected (Kovacs et al. 2010). This widespread ash mortality caused by this
invasive beetle has resulted in cascading effects through forested ecosystems, altering species
composition, understory light, temperature, and moisture availability, increasing coarse woody
debris, and affecting carbon and nitrogen cycling (Knight et al. 2013, Flower et al. 2013a, Burr
and McCullough 2014, Klooster et al. 2014).
Spread of EAB is characterized by stratified dispersal, represented by short-distance adult
flights (i.e., natural dispersal), while human-transport of infested ash material can result in longdistance spread (Mercader et al. 2011a, 2011b, Mercader et al. 2012, Siegert et al. 2014). Field
studies based on systematically felling and debarking ash trees to locate larval progeny from
beetles that emerged from a known point of origin have helped with the understanding of EAB
dispersal (Mercader et al. 2009, Siegert et al. 2010). While some portion of female EAB adults
likely have long range dispersed flights and colonize trees more than 800 m from their point of
emergence (Mercader et al. 2012), multiple field studies have shown 90% of eggs laid by

72

females are within 800 m of their emergence point, with most ovipositing on trees within 100 m
(Mercader et al. 2009, 2012, 2015, Siegert et al. 2010).
In North America, EAB preferentially colonize stressed ash trees, but will attack healthy
trees (Cappaert et al. 2005, Poland and McCullough 2006, McCullough et al. 2009a, 2009b,
Siegert et al. 2010, Flower et al. 2013, Herms and McCullough 2014). Feeding and development
of EAB can likely occur on all ash species in the U.S., including the four prominent species
occurring in Michigan (Anulewicz et al. 2007, Tanis and McCullough 2012, 2015). Interspecific
differences in ash resistance or EAB host preference among species have been observed
(Cappaert et al. 2005, Anulewicz et al. 2007, 2008, Rebek et al. 2008, Pureswaran and Poland
2009, Chen and Poland 2010, Tanis and McCullough 2012, 2015). Green ash (Fraxinus
pennsylvanica Marsh.) and black ash (F. nigra Marsh.) are consistently highly preferred and
vulnerable hosts of EAB (Cappaert et al 2005, Anulewicz et al. 2007, 2008, Rebek et al. 2008,
Chen and Poland 2010), while blue ash (F. quadrangulata Michx.) is relatively resistant to EAB
(Anulewicz et al. 2007, Tanis and McCullough, 2012).
White ash (F. americana L.) is considered an intermediate host for EAB (Anulewicz et
al. 2007, Tanis and McCullough 2012, 2015). Previous studies showed green ash trees are
preferentially colonized in multi-species plantations and in landscapes where green ash and white
ash trees co-occur (Anulewicz et al. 2007, Limback 2010, Tanis and McCullough 2015). While
substantial white ash mortality has occurred in many areas in southeast Michigan (Gandhi and
Herms 2010, Knight et al. 2013, 2014, Klooster et al. 2014, Smith et al. 2015), evidence to date
suggests interactions between white ash trees and EAB may be relatively complex. We have
observed live white ash trees in some forested sites in areas of southeast Michigan despite the
presence of EAB in these sites for at least 10-12 years (McCullough, unpublished data).

73

Identification of variables associated with relatively high or low survival could have important
management implications for the future of white ash trees as a functional component of forest
systems.
Both tree and site characteristics play a role in tree mortality caused by native or exotic
pests. Clark et al. (2015) developed decision models to predict the probability of ash tree
mortality using tree health assessments and signs of infestation, specifically for management of
hazardous trees in residential areas. This study focused on tree-level data, rather than data
collected across a forested stand, and did not incorporate variables such as land use. Knight et al.
(2013) evaluated relationships between site and tree-level factors and ash tree mortality in EABinfested stands. They reported crown class, hydrology, ash canopy health class, and the number
of ash trees per hectare were significant variables in models to predict ash survival rates. In
contrast, Smith et al. (2015) reported no relationship between site and stand attributes including
overstory species composition, ash density, basal area, and importance, and mortality rates of
white, green, or black ash.
Factors that promote EAB dispersal should also be considered. A study by Jenkins et al.
(2016) quantified ash mortality risks using “detection tree” data collected in Ohio from 2005 to
2007, while also incorporating distance to existing ash stands, campgrounds, roads, and railways
that are corridors for EAB dispersal. They found ash trees growing along riparian corridors are at
high relative risk of EAB-induced mortality. In heterogeneous sites with urban, residential, and
wooded areas, EAB preferentially colonized open-grown ash trees rather than shaded trees in
wooded areas, similar to findings from other studies (Siegert et al. 2010, McCullough et al.
2009a, 2009b).

74

I investigated site-level variables associated with white ash survival in southeast and
south central Michigan forests where EAB has had the longest history of invasion in North
America. I surveyed an array of variables in 28 sites located in this region to quantify site
characteristics potentially associated with rates of white ash survival. My primary objectives
were to document overstory species composition and condition in forested stands with a
substantial component of white ash varying in survival and at least 10-12 years of EAB presence.
I evaluated site-level variables within a 1.5 km radius from the center of each site, and
hypothesized EAB colonization or population growth rates could be related to abundant white
ash, abundant green ash, or land usage type. Since green ash is a highly preferred, vulnerable
host of EAB (Anulewicz et al. 2007, 2008, Rebek et al. 2008), I hypothesize its presence could
attract EAB, having a negative impact on white ash survival. However, if this preferred host is
absent, some white ash trees may be alive and continue to persist. If the presence of open-grown
ash trees growing in sunny-conditions, such as along roads around sites consisting of living
white ash are scarce or absent, white ash trees may have a higher rate of survival. I developed
models using survey results and spatial datasets to determine if any variables were consistent
predictors of white ash survival. These variables could then be used to develop white ash
survival prediction models where resource managers could prioritize harvest of white ash trees in
forested areas projected to be susceptible to the infestation of EAB, while white ash trees
projected to survive EAB infestation could be left to persist as a functional component within our
forests.

75

METHODS
Site selection
I identified 70 large areas of public forestland (state, county, or municipal property)
distributed across 11 counties in southeast and south central Michigan using a county atlas in
2014. Emerald ash borer is known to have invaded this area at least 10 to 12 years earlier (Pugh
et al. 2011, Marshall et al. 2013, Burr and McCullough 2014, Klooster et al. 2014, Siegert et al.
2014). I contacted managers of each area to acquire information about the presence of white ash
trees on the property. Based on forest inventory records and my visual observations, I eliminated
areas with few or no white ash trees from further consideration. Remaining areas were further
scouted to assess distribution and condition of white ash trees. I established a center point and
recorded GPS coordinates in the midst of abundant white ash trees in 28 sites (Figure 3.1).
White ash condition
In August 2015, I delineated a 1 ha area around the center point of each site, divided the
area into four equal quadrants, and established a fixed radius plot (18 m) in each quadrant. I
recorded DBH (cm) of all tree species (≥ 10 cm DBH) and visually estimated percent canopy
dieback of white ash trees and green ash trees ≥ 10 cm DBH when present. Sites were then
classed as high survival if (≥ 85% of the white ash trees were alive, low survival if ≤ 60% of the
white ash trees were alive, or moderate survival if > 60% but ≤ 84% of the white ash trees were
alive.
Variable radius plots
I delineated a 1.5 km radius around the center point of each site using satellite imagery
and ArcMap 10.3, then delineated forested areas within the 1.5 km radius. A 200 x 200 m grid
was overlaid on each map and grid cells falling within each forested polygon were numbered.

76

Grid cells were selected for field surveys using systematic random sampling based on the
polygon area. I surveyed a minimum of two grid cells in polygons < 8 ha and up to 20 cells in
polygons > 81 ha.
I established a variable radius plot (10 baf prism) in the center of each selected grid cell
and recorded species and DBH of all dead and live trees ≥ 10 cm. I visually assessed canopy
dieback of live trees in increments of 10%, where 10% indicated a nearly full, healthy canopy,
and 90% indicated a nearly dead canopy (Zarnoch et al. 2004). The total number of live and dead
white ash trees recorded in variable radius plots were summed for each site. When ash trees were
present, I recorded signs of EAB infestation, including holes left by woodpeckers feeding on
EAB larvae, adult EAB exit holes, and bark cracks above larval galleries.
Driving surveys
I conducted additional surveys within the area defined by the 1.5 km radius in August
2015 to assess presence and abundance of white ash trees and green ash trees along roads, in
open fields, hedgerows, and residential areas. A waypoint was taken when ash trees were
present. I recorded the ash species and visually estimated DBH and canopy dieback. In locations
where clusters of > 5 white ash trees or green ash trees were present, I counted or visually
estimated the number of ash trees, and estimated the average DBH and canopy dieback for trees
by species.
Geographic data
Land cover data were downloaded from the National Land Cover Database (NLCD 2011)
and clipped to each 1.5 km radius. I calculated total area (km2) of each land cover type with the
area bounded by each radius in each site, and the proportion of area represented by deciduous
forest, mixed forest, woody wetland, shrubland, cropland, grassland, and pasture. I summed the

77

total area of deciduous forest and mixed forest within the four areas defined by the radii, then
classified this variable as “forest” for analyses. I similarly summed the area of cropland,
grassland, shrubland, and pasture for each radius and classified this variable as “open” for
analyses.
Road shapefiles were downloaded from the Michigan Geographic Data Library (MiGDL)
and clipped each to the areas as above. Using ArcMap 10.3, I calculated the length (km) of roads
within each 1.5 km radius, and calculated the density of white ash and green ash trees (≥ 10 cm
DBH) by summing the count of each species and dividing by the determined area of roads within
the areas defined by the radii.
Statistical analysis
I summed the number of variable radius plots where white ash and green ash trees were
tallied for each site. Average DBH (cm) and average percent canopy dieback were calculated for
white ash and green ash trees (≥ 10 cm DBH) in the variables radius plots and driving surveys.
I fit a general linear model to variables within the areas defined by the 0 to 400, 400 to
800, 800 to 1200, and 1200 to 1500 m radii, and to a cross-scale model defined by the outer 1500
m radius. The response variable was the percentage of white ash trees (≥ 10 cm DBH) that were
alive in the fixed radius plots. I began my analyses with 16 potential predictor variables I
hypothesized could influence white ash tree survival, including counts, average DBH (cm) and
canopy dieback estimates for white ash and green ash trees recorded in driving surveys and
variable radius plots, number of plots with white ash or green ash trees present, proportion of
woody wetlands, open area (area of crops, grass, shrubs, and pasture summed), forested area
(summed mixed and deciduous forest), and densities of white ash and green ash trees along
roads.

78

To address multicollinearity among potential predictor variables, I calculated Pearson’s
correlation coefficients, and excluded variables with a correlation coefficient > 0.40 for a
conservative model (Elith et al. 2006, Wisz and Guisan 2009). Generalized Variance Inflation
Factors (VIFs) were calculated and considered for the remaining predictor variables (Brunsdon et
al. 2012, Neter et al. 1996) and competing variables with the highest VIF values were eliminated
until all VIF values were < 2. Multicollinearity diagnoses were conducted using R statistical
software (R Core Development Team, http://www.R-project.org)
I fit a set of competing models to my data and applied multi-model inference (Burnham
and Anderson 2002) to assess the fit. I used the Akaike information criterion (AICc) corrected
for small sample sizes (AICc; Shono 2000), as recommended when the ratio between n (the
number of observations used to fit the model) and K (the number of parameters in the largest
model) is < 40 (Shono 2000, Burnham and Anderson 2002). To optimize model strength, I
applied a multi-model dredging function in R’s MuMin package (dredge – MuMIn R package;
available at http://r-forge.r-project.org) to assess valid combinations of variables to predict the
percentage of white ash trees alive in the sites. All competing models with a delta AICc of < 2.0
were selected, as recommended by Burnham and Anderson (2002). Model averaging was used
for the competing parameter coefficient estimates in the model following methods of Buckland
et al. (1997), and the relative importance of variables were calculated (Burnham and Anderson
2002). Diagnostics for the competing models were assessed for normality of residuals and
homoscedasticity by evaluating plots of fitted values vs. residuals, Q-Q plots and Cook's
Distance plots. Model selection and residual diagnostics were conducted using R statistical
software (R Core Development Team, http://www.R-project.org) at α of P < 0.05.

79

RESULTS
White ash condition
Overall, 74% of the 821 white ash trees (≥ 10 cm DBH) recorded in the fixed radius plots
in the 28 sites were alive. There were 33 white ash trees with 70-90% canopy dieback that were
classed as “dead”. White ash survival rates, i.e., the percent of stems ≥ 10 cm DBH alive, varied
considerably among sites, ranging from 0% in one site up to 100% in two sites (Fig 3.2). In 14 of
the 28 sites, ≥ 80% of the white ash stems were alive, and at least 50% of the white ash stems
were alive in 22 sites. Nine sites were classified as high survival (≥ 85% of the white ash trees
alive), seven sites were classified as low survival (≤ 60% of the white ash trees alive), and twelve
sites were classified as moderate survival (> 60 % to 84% of the white ash trees alive).
Variable radius plots
White ash trees were present in a total of 98 of the variable radius plots established in
forested areas. I recorded 240 white ash trees total and survival rates within plots ranged from
0% in three sites to 100% in nine sites. Overall, 75% of the white ash trees were alive and all had
< 50% canopy dieback. In 15 sites, ≥ 80% of the white ash trees were alive and at least 60% of
the white ash trees were alive in 23 of the 28 sites. In contrast, of the 214 green ash trees
encountered in 91 variable radius plots, only 3% were alive in four of the 28 sites.
Driving surveys
Both white ash and green ash trees were abundant along roads running through all 28
sites. Density of ash trees along roads ranged from 1.6 to 37.7 trees per km, averaging 12.3 ± 4.4,
14.8 ± 3.1, and 17.1 ± 3.8 trees for the sites with high, moderate, and low white ash survival,
respectively. Density of ash along roads ranged from 73 per km to 472 per km, averaging 210 ±
23.3 trees per site. On average, there were 223 ± 52.0, 227 ± 38.4, and 178 ± 34.9 ash street trees

80

in the high, moderate, and low survival sites, respectively. Survival of the ash street trees in the
high, moderate, and low survival site classes was 46%, 46% and 26%, respectively. White ash
trees accounted for 50%, 54%, and 26% of the street trees, while green ash accounted for 50%,
46%, and 74% of the total recorded in high, moderate, and low survival site classes, respectively.
Geographic data
Land use data was analyzed across the 1.5 km radius. For the high, moderate, and low
white ash survival sites, forested area accounted for 34% to 40% of the total area. Open land
comprised 30% of the total area in sites with high white ash survival. Woody wetlands were the
least abundant land class, comprising 12% of the area in the low and moderate survival sites, to
19% in the high survival sites (Table 3.1).
Cross-scale model
A total of 29 competing models with a delta AICc < 2.0 (Table 3.2) had a ≤ 4% chance of
being the best model to predict white ash survival recorded in the four fixed radius plots in the
center of the sites. No single strongly weighted (wi > 0.9) model could be identified. Variables in
the top model (delta AICc = 0.00) with an AICc value of -8.79 included average DBH of green
ash recorded in driving surveys, green ash density along roads, proportion of area classed as
open, and white ash road density. Other variables included in competing models (delta AICc <
2.0) were average DBH of green ash and white ash recorded in variable radius plots, average
DBH of white ash recorded in driving surveys, and the proportion of area classed as open (Table
3.2).
Based on the cross-scale full-averaged model, there were no variables that significantly
affected white ash survival (Table 3.2). Density of white ash trees along roads had the highest
value of relative importance (0.91) compared to all other variables (Figure 3.3). Average DBH of

81

green ash trees along roads had a relative importance of 0.77, while green ash road density, the
proportion of area classed as open, and the proportion of area classed as woody wetlands had
importance values of 0.73, 0.67, and 0.53, respectively (Figure 3.3).
400 m model
Five competing models were identified for areas bounded by the 400 m radius. No single
strongly weighted (wi > 0.9) model could be identified. The intercept had an AICc value of 3.12
and was considered the top model (delta AICc = 0.00), with a weight of 0.30. Average DBH of
white ash trees along roads had the second lowest AICc value of 4.40, but only had a 16%
chance of explaining white ash survival.
The full-averaged model for the area within the 400 m radius showed the proportion of
area classed as woody wetlands, green ash density along roads, and the proportion of area
classed as forest were negatively related to white ash survival, but the parameters were not
significant (P = 0.83, 0.85, and 0.95, respectively) (Table 3.3). Average DBH of white ash
recorded during driving surveys, proportion of area classed as woody wetlands, and average
DBH of green ash recorded in variables radius plots had the highest relative importance (0.16,
0.12, and 0.12, respectively) (Figure 3.3).
800 m model
Two competing models had a delta AICc < 2.0 (Table 3.4), but none were identified as
the best model (wi > 0.9). Variables in the top model (delta AICc = 0.00) with an AICc value of
-1.16 included average DBH of green ash trees from the driving surveys and white ash tree
density along roads. This model, however, only had a 36% chance of being the best model to
explain white ash survival. The remaining model included average DBH of green ash recorded in

82

driving surveys, green ash tree density along roads, and white ash tree density along roads, with
a 19% chance of being the best model.
Based on the full-averaged model, average DBH of green ash trees recorded in driving
surveys was the only significant variable negatively related to white ash survival (P = 0.02). The
proportion of area classed as forest was also negatively related to white ash survival, but was not
significant (P = 0.90) (Table 3.4). Average DBH of green ash trees recorded in driving surveys
had the highest relative importance of 1.00 followed by white ash tree density along roads,
having a relative importance of 0.90 (Figure 3.3). Green ash density along roads which had a
relative importance of 0.19, while the remaining variables had a relative importance of ≤ 0.09
(Figure 3.3).
1200 m model
There were 23 competing models with a delta AICc < 2.0 in the 1200 m model, all of
which had < 5% chance of being the best model to explain white ash survival (Table 3.5). The
top model (delta AICc = 0.00) with an AICc value of 2.67 included average DBH of green ash
trees recorded in variable radius plots, green ash tree density along roads, and proportion of area
classed as forest. The second model with an AICc value of 2.72 included average DBH of green
ash recorded in variable radius plots and proportion of area classed as open. Other variables
included in the competing models were average DBH of green ash trees recorded in driving
surveys, average DBH of white ash trees recorded in driving surveys and variable radius plots,
and the proportion of area classed as woody wetlands (Table 3.5).
The full-averaged model for the 1200 m radius showed average DBH of green ash in
variable radius plots, green ash density along roads, the proportion area classed as forest, the
proportion of area classed as woody wetlands, and average DBH of green ash recorded in driving

83

surveys negatively affected white ash survival, but none were significant (P = 0.36, 0.47, 0.66,
0.72 and 0.83, respectively) (Table 3.5). Average DBH of green ash recorded in variables radius
plots had the highest relative importance of 0.61 while green ash tree density along roads had a
relative importance of 0.48. The proportion of area classed as open and forest, and average DBH
of white ash trees recorded in driving surveys had a relative importance of 0.34, 0.25, and 0.24,
respectively. The remaining variables had a relative importance of ≤ 0.20 (Figure 3.3).
1500 m model
There were only two competing models with a delta AICc < 2.0 within the 1200 to 1500
m radius, but none were clearly the best (wi > 0.9) (Table 3.6). The top model (delta AICc =
0.00) included green ash tree density along roads and the proportion of area classed as woody
wetlands, which had a 36% chance of being the best model of the candidate set. The next model
included green ash tree density along roads, average DBH of white ash in variable radius plots,
and the proportion of area classed as woody wetlands with a 16% chance of being the best model
(Table 3.6).
Green ash tree density along roads and the proportion of area classed as woody wetlands
had a significant negative affect on white ash tree survival (P = 0.02) (Table 3.6). Average DBH
of white ash trees recorded in driving surveys and average DBH of green ash trees recorded in
variables radius plots both negatively affected white ash survival, however, this effect was not
significant (P = 0.83 and P = 0.99, respectively). Green ash density along roads and the
proportion of area classed as woody wetlands had the highest importance (1.00) of all variables
in the 1200 to 1500 m model. Average DBH of white ash trees recorded in variables radius plots
and driving surveys, and the proportion of area classed as open had a relative importance of 0.16,

84

0.12, and 0.11, respectively, while the remaining variables had a relative importance of < 10
(Figure 3.3).
Competing models with a delta AICc value of < 2.0 are considered to be essentially as
good as the best model. When considering many models, the delta AICc value can become
problematic because models strongly competing with the best model (delta AICc < 2.0) may
differ very little structurally (Burnham and Anderson 2002, Symonds and Moussalli 2010).
Although many site-level variables were tested in the models, the sample was limited to 28 sites.
This resulted in model uncertainty, with many competing models at each scale. However, the
variables that were consistently of high importance across the models are notable and deserve
future exploration with a more abundant data set.
The presence of green ash was consistently important in predicting white ash survival at
four different scales (400 to 800 m, 800 to 1200 m, 1200 to 1500 m, and cross-scale), having a
negative effect on white ash survival. White ash tree presence along roads was also important in
multiple models (400 to 800 m and cross-scale), while the proportion area classed as woody
wetlands was one of two variables with the highest importance (1.00) in the 800 to 1200 m
model, and had a negative effect on white ash survival in all models but the 800 to 1200 m
model.

85

DISCUSSION
White ash trees are native forest species in 38 states and two Canadian provinces, are
landscape trees in all 48 states, and are a component of 26 unique forest cover types in North
America (NLCD 2017). Determining site-level variables associated with white ash tree survival
in areas with the longest history of EAB invasion could provide direction for the management
and information about the future of this species as a functional component of forest systems.
I hypothesized the presence of green ash, a highly preferred and vulnerable EAB host
(Anulewicz et al. 2007, 2008, Rebek et al. 2008) could impact white ash survival. Previous
studies have shown that EAB preferentially colonizes green ash trees over white ash trees where
these species co-occur (Anulewicz et al. 2007, Limback 2010, Tanis and McCullough 2012,
2015). I found green ash trees present along roads in the areas surrounding the concentrations of
white ash was negatively related to white ash survival. Additionally, the proportion of the area
bounded by the 1.5 km radius that was classed as woody wetlands had a negative effect on white
ash survival. Green ash is commonly found along rivers and streams, and can be abundant on
land subject to flooding (NRCS 2017). Area of woody wetlands may, therefore, reflect
conditions favorable and likely to support green ash trees. In this study, nearly all (97%) green
ash trees recorded in variable radius plots were dead while the majority (75%) of white ash trees
were alive and persisting. Continued monitoring to evaluate white ash persistence in these sites
would be useful to assess whether EAB populations may eventually kill most of these trees, or if
an equilibrium will eventually be reached with EAB functioning as a secondary pest, affecting
primarily stressed and low vigor trees.
Green ash presence along roads was an important variable in most of the models. Clark et
al. (2015) reported ash trees growing along riparian zones are at high relative risk of EAB

86

induced mortality. In this study, my focus was on ash trees growing within a forested stand or
along roads. While the area of railroads and streams within the area bounded by the 1.5 km
radius was low in all sites, these linear corridors may be of importance if their presence was
more abundant in the areas I studied. While we did not test for these variables in the models, they
should not be excluded as potential variables in future models. Assessing ash abundance and
condition along different types of corridors could yield more information about potential effects
of streams, roads, or other linear corridors on ash survival.
Past studies in sites with a mix of urban, residential, and wooded areas showed EAB
preferentially colonized open-grown ash trees exposed to sun compared to shaded trees (Siegert
et al. 2010). The proportion of open area that was present in three of the five models was not of
high importance. In the future, plots should be established in open areas based on the acreage of
each, as I selected variable radius plot locations. Ash trees in open-grown areas could assist in
the dispersal of EAB between forested areas where white ash are present. While the proportion
of area classed as open was an important variable in multiple models, surveying ash trees
growing in these open areas would give a better representation of the influence open-grown ash
trees potentially have on white ash survival.
Ash trees along roads were abundant in all 28 sites. Most of these trees were growing in
full sunlight, which is preferred by EAB (McCullough et al. 2009a, 2009b). Both white ash and
green ash trees are prolific seed producers, contributing to their abundance along roads (NRCS
2017, Gucker 2005, Griffith 1991). This data was valuable in determining whether the density,
size, and condition of white ash and green ash trees growing in the near proximity of abundant
white ash trees influence their survival. The abundance of ash trees along roads resulted in a
representation of the influence this corridor, and the density of roadside ash has on white ash

87

survival. Like ash trees present in open areas, roadside ash could also promote EAB dispersal
between forested areas.
Variables related to site nutrition were strongly related to white ash health in a study
conducted by Knight et al. (2012). Soil moisture is an important factor in the distribution of
white ash (NRCS 2017, Griffith 1991). White ash trees grow best on moderately well drained
soils where roots can penetrate to a depth of 40 cm or more (NRCS 2017, Hosie 1969). Green
ash trees grow best on fertile, moist, well-drained soils, and are most commonly found along
rivers and streams and on land subject to flooding for as long as 40% of the growing season
(NRCS 2017, Boerner and Brinkman 1996). Stress can alter volatiles emitted by ash trees,
increasing their attraction to EAB adults (Rodriguez-Saona et al. 2006, McCullough et al. 2009a,
2009b, Chen and Poland 2009). If green ash trees experienced poorer growing conditions than
white ash trees, survival rates could reflect EAB attraction to stressed versus healthier trees. This
should be tested in future modeling approaches. While the majority (97%) of green ash trees
were dead across the sites, I cannot state whether that is due to site-level variables, genetics, or a
combination of the two based on the results from this study.
Although many site-level variables were tested in the models, model uncertainty was
high as evidenced by the many competing models at each scale (Burnham and Anderson 2002).
By averaging the covariate data collected and extracted at the different scales, I condensed the
available information to enable these variables to be tested in models of white ash survival.
Because of the high model uncertainty, however, whether or not the mean representations of
these variables accurately reflect components of the true process of white ash survival requires
further inquiry. The spatial distribution of a measured environmental variable will certainly
affect whether a mean value adequately represents the factor in a model. For example, taking the

88

average of data points collected across space may be a defensible solution if the variable is
homogeneously distributed. But it is difficult to determine whether this is true in all cases;
further investigation of the variables found in competing models of this study is necessary to
determine which variables may actually be affecting white survival and at what scale. Rather
than averaging data across areas bounded by different radii, extracting and analyzing data at the
plot-level could influence, and potentially lead to, greater confidence in my results. This data
would be more fine-scale, focusing on where they are both measured and where white ash trees
are present. This might provide a larger sample size to analyze, and remove some of the
observational uncertainty that comes with representing variables by the mean value, or other
transformations.
High rates of white ash survival could result from infrequently occurring phenotypes that
cause the trees to be less preferred or more resilient to EAB attack (Knight et al. 2012, Marshall
et al. 2013). Knight et al. (2014) defined “lingering ash” as an individual tree that is alive in
areas where more than 98% of the ash trees have been killed by EAB (Knight et al. 2014,
Marshall et al. 2010). Bioassays suggest multiple aspects of the host-insect interaction are
important for increased resistance to EAB (Koch et al. 2015). In my study, white ash tree
populations were abundant within sites, therefore, they are not “lingering ash”. Including these
abundant white ash tree populations in selection and breeding programs could potentially assist
in determining what drives resistance to EAB to eventually yield resistant ash hybrids or
cultivars.
Results from this study are useful in identifying variables that could be considered in
future models to understand the driving factors of this unexpected white ash survival. Based on
my findings, it is important to consider land use type and the presence of green ash in models to

89

predict white ash survival. Understanding why white ash survives EAB in some sites but not in
other sites could have management implications. Ideally, resource managers could prioritize
harvest of white ash trees in areas that are projected to be vulnerable to EAB infestation, and
leave white ash that are unlikely to be killed, allowing white ash to persist as a functional
component within forest systems.

90

APPENDIX

91

APPENDIX
Table 3.1. Average (± SE) proportion of area classified as forested, open area, or woody
wetlands within the 1.5 km radius (706.8 ha) of the 28 sites located in southeast and south
central Michigan separated into high survival (≥ 85%), moderate survival (> 60.0% to ≤
84.0%), or low survival (≤ 60%) survival sites based on proportion of white ash stems live
at the center of the sites in 2016.
Forested area

Open area

Woody wetland

High survival sites (N = 7)
Average proportion ha ± SE
Min, max

0.34 ± 0.11
0.08, 0.93

0.30 ± 0.04
0.14, 0.44

0.12 ± 0.03
0.01, 0.27

Moderate survival sites (N = 12)
Average proportion ha ± SE
Min, max

0.37 ± 0.05
0.05, 0.62

0.25 ± 0.03
0.01, 0.56

0.12 ± 0.02
0.03, 0.23

Low Survival sites (N = 9)
Average proportion ha ± SE
Min, max

0.39 ± 0.07
0.03, 0.67

0.24 ± 0.02
0.02, 0.34

0.19 ± 0.03
0.04, 0.30

92

Table 3.2. Competing models with a delta AICc < 1.0 for variables recorded across all four
areas bounded by 400 m, 800 m, 1200 m, 1500 m, and cross-scale models. 1Full-model
averaging results from the cross-scale model.

Variables

Competing models
df logLik

Average DBH of green ash along roads + green ash
road density + proportion open land + white ash
road density
Average DBH of green ash along roads + green ash
road density + proportion open land + white ash
road density + proportion woody wetlands
Average DBH of green ash along roads + green ash
road density + average DBH of white ash + white
ash road density + proportion woody wetlands
Average DBH of green ash along roads + proportion
open land + white ash road density
Average DBH of green ash along roads + green ash
road density + proportion open land + average DBH
of white ash + white ash road density + proportion
woody wetlands
Average DBH of green ash along roads + green ash
road density + proportion open land + average DBH
of white ash + white ash road density
Average DBH of green ash along roads + green ash
road density + proportion of forested land + average
DBH of white ash + white ash road density +
proportion woody wetlands
Average DBH of green ash along roads + green ash
road density + proportion of forested land + white
ash road density + proportion woody wetlands
Average DBH of green ash along roads + green ash
road density + white ash road density + proportion
woody wetlands
Average DBH of green ash along roads + proportion
of open land + white ash road density + proportion
woody wetlands
Average DBH of green ash along roads + green ash
road density + proportion of open land + proportion
of forested land + white ash road density

93

AICc

delta

weight

6

10.79

-8.79

0.00

0.04

7

11.84

-8.61

0.18

0.03

7

11.75

-8.43

0.36

0.03

5

9.38

-8.19

0.60

0.03

8

12.79

-8.19

0.60

0.03

7

11.62

-8.16

0.62

0.03

8

12.75

-8.11

0.68

0.03

7

11.56

-8.04

0.75

0.02

6

10.37

-7.94

0.85

0.02

6

10.35

-7.91

0.88

0.02

7

11.47

-7.87

0.92

0.02

Table 3.2 (cont’d)

Average DBH of green ash along roads + proportion
of forested land + average DBH of white ash +
white ash road density
Average DBH of green ash along roads + green ash
road density + proportion of open land + proportion
of forested land + white ash road density +
proportion of woody wetlands
Average DBH of green ash along roads + proportion
of forested land + average DBH of white ash +
white ash road density + proportion of woody
wetlands
Green ash road density + proportion of open land +
white ash road density
Green ash road density + proportion of forested land
+ white ash road density + proportion of woody
wetlands
Green ash road density + proportion of open land +
proportion of forested land + white ash road density
Green ash road density + proportion of open land +
proportion of forested land + white ash road density
+ proportion of woody wetlands
Green ash road density + proportion of open land +
white ash road density + proportion of woody
wetlands
Green ash road density + proportion of forested land
+ average DBH of white ash + white ash road
density + proportion of woody wetlands
Average DBH of green ash along roads + average
DBH of white ash + white ash road density +
proportion of woody wetlands
Average DBH of green ash along roads + green ash
road density + average DBH of white ash + white
ash road density
Average DBH of green ash along roads + green ash
road density + proportion of open land + proportion
of forested land + average DBH of white ash +
white ash road density + proportion of woody
wetlands
Average DBH of green ash along roads + green ash
road density + proportion of forested land + white
ash road density

94

6

10.32

-7.84

0.92

0.02

8

12.59

-7.78

1.00

0.02

7

11.43

-7.78

1.01

0.02

5

9.15

-7.73

1.06

0.02

6

10.22

-7.64

1.15

0.02

6

10.22

-7.64

1.15

0.02

7

11.31

-7.55

1.24

0.02

6

10.15

-7.51

1.28

0.02

7

11.24

-7.40

1.38

0.02

6

10.10

-7.39

1.40

0.02

6

10.09

-7.39

1.40

0.02

9

13.47

-7.17

1.61

0.02

6

9.98

-7.15

1.63

0.02

Table 3.2 (cont’d)
Average DBH of green ash along roads + green ash
road density + proportion of forested land + average
DBH of white ash + white ash road density
Average DBH of green ash along roads + green ash
road density + proportion of open land + proportion
of forested land + average DBH of white ash +
white ash road density
Average DBH of green ash along roads + green ash
road density + white ash road density
Green ash road density + proportion of open land +
average DBH of white ash + white ash road density
Average DBH of green ash along roads + proportion
of open land + average DBH of white ash

7

11.09

-7.11

1.68

0.02

8

12.23

-7.07

1.72

0.02

5

8.80

-7.03

1.75

0.02

6

9.82

-6.85

1.94

0.02

5

8.69

-6.82

1.97

0.01

Parameters of predictors1
Variables
Estimate Std. Error
Intercept
0.745
0.080
Average DBH of green ash along roads
-0.003
0.003
Green ash road density
-0.7e-03
0.7e-03
Proportion of open land
0.173
0.174
White ash road density
0.002
0.9e-03
Proportion of woody wetlands
-0.153
0.199
Average DBH of white ash
0.001
0.002
Proportion of forested land
-0.049
0.095
Average DBH of green ash
-0.3e-04
0.3e-03
Average DBH of white ash along roads
-0.2e-04
0.5e-03
1

Adjusted SE
0.080
0.003
0.7e-03
0.176
0.9e-03
0.200
0.002
0.096
0.3e-03
0.6e-03

z value
9.324
1.200
1.132
0.987
1.668
0.764
0.685
0.511
0.098
0.039

Parameters were derived from full-model averaging. This approach is recommended by Burnham
and Anderson (2002) when high model uncertainty exists, where each averaged parameter estimate
is weighted to reduce the influence of variables with low Akaike parameter weights (w i).

95

Table 3.3. Competing models and parameters calculated for variables and used to predict percent
white ash survival (delta AICc < 2.0) in the center of areas bounded by a 400 m radius in 28 sites
in southeast and south central Michigan. 1Full-model averaging results from the 0 to 400 m model.

Competing Models
Predictors
df logLik
AICc
Intercept
2
0.68
3.12
Average DBH of white ash along roads 3
1.30
4.40
Proportion of woody wetlands
3
1.02
4.96

delta
0.00
1.28
1.84

weight
0.30
0.16
0.12

Average DBH of green ash

3

1.01

4.98

1.86

0.12

Green ash road density

3

0.95

5.10

1.98

0.11

Adjusted SE
0.075
0.002
0.127
0.001
0.001
0.001
0.062

z value
9.361
0.297
0.213
0.211
0.191
0.183
0.062

Parameters of Predictors1
Variables
Estimate
Std. Error
Intercept
0.703
0.072
Average DBH of white ash along roads
0.9e-03
0.003
Proportion of woody wetlands
-0.027
0.123
Average DBH of green ash
0.3e-03
0.001
Green ash road density
-0.2e-02
0.001
Average DBH of white ash
0.3e-02
0.001
Proportion of forested land
-0.004
0.059
1

Parameters were derived from full-model averaging. This approach is recommended by Burnham
and Anderson (2002) when high model uncertainty exists, where each averaged parameter estimate
is weighted to reduce the influence of variables with low Akaike parameter weights (w i).

96

Table 3.4. Competing models and parameters calculated for variables and used to predict percent
white ash survival (delta AICc < 2.0) in the center of areas bounded by an 800 m radius in 28
sites in southeast and south central Michigan. 1Full-model averaging results from the 400 to 800
m model.

Predictors

Competing models
df
logLik

AICc

delta

weight

Average DBH of green ash along roads
+ white ash road density

4

5.45

-1.16

0.00

0.36

Average DBH of green ash along roads
+ green ash road density + white ash
road density

5

6.33

0.07

1.23

0.19

Parameters of predictors1
Variables
Estimate Std. Error Adjusted SE z value
Intercept
9.396e-01 1.323e-01
1.390e-01
6.758
Average DBH of green ash along roads
-1.603e-02 6.650e-03
6.974e-03
2.299
White ash road density
7.597e-03 4.470e-03
4.626e-03
1.642
Green ash road density
1.279e-03 3.516e-03
3.600e-03
0.355
Average DBH of white ash along roads
1.442e-04 1.085e-03
1.133e-03
0.127
Proportion of forested land
-8.631e-03 6.911e-02
7.221e-02
0.120
Average DBH of green ash
9.139e-05 9.147e-04
9.582e-04
0.095
Average DBH of white ash along roads
4.197e-05 1.554e-03
1.636e-03
0.026
1

Parameters were derived from full-model averaging. This approach is recommended by Burnham
and Anderson (2002) when high model uncertainty exists, where each averaged parameter estimate
is weighted to reduce the influence of variables with low Akaike parameter weights (w i).

97

Table 3.5. Competing models and parameters calculated for variables and used to predict percent
white ash survival (delta AICc < 2.0) in the center of areas bounded by a 1200 m radius in 28
sites in southeast and south central Michigan. 1Full-model averaging results from the 800 to 1200
m model.
Competing models
Variables
df
Average DBH of green ash + green ash road
density + proportion of forested land
5
Average DBH of green ash + proportion of
4
open land
Average DBH of green ash + proportion of
open land + average DBH of white ash along
5
roads
Green ash road density + proportion of woody
4
wetlands
Average DBH of green ash
3
Intercept
2
Average DBH of green ash + green ash road
density + proportion of forested land +
6
average DBH of white ash along roads
Average DBH of green ash + green ash road
4
Green ash road density
3
Proportion of woody wetlands
3
Proportion of open land
3
Green ash road density + proportion of
4
forested land
Average DBH of white ash
3
Average DBH of green ash + green ash road
density + proportion of open land + average
6
DBH of white ash along roads
Average DBH of green ash + green ash road
5
density + proportion of open land
Average DBH of green ash + green ash road +
5
average DBH of white ash along roads
Average DBH of green ash + proportion of
4
forested land
Average DBH of green ash + Average DBH
4
of green ash along roads
Average DBH of green ash + green ash road
5
density + proportion of woody wetlands

98

logLik

AICc

delta weight

5.03

2.67

0.00

0.05

3.51

2.72

0.04

0.05

4.93

2.86

0.19

0.04

3.33

3.08

0.40

0.04

1.94
0.68

3.11
3.12

0.44
0.45

0.04
0.04

6.39

3.22

0.55

0.04

3.25
1.82
1.59
1.58

3.23
3.36
3.81
3.84

0.56
0.69
1.14
1.16

0.04
0.03
0.03
0.03

2.87

4.00

1.32

0.02

1.37

4.26

1.59

0.02

5.87

4.26

1.59

0.02

4.23

4.27

1.59

0.02

4.21

4.31

1.73

0.02

2.69

4.35

1.68

0.02

2.67

4.41

1.73

0.02

4.13

4.46

1.79

0.02

Table 3.5 (cont’d)
Average DBH of green ash along roads
Average DBH of green ash + proportion of
open land, average DBH of white ash +
average DBH of white ash along roads
Average DBH of green ash + average DBH
of white ash
Average DBH of green ash + average DBH
of white ash along roads

3

1.26

4.47

1.8

0.02

6

5.68

4.63

1.96

0.02

4

2.55

4.64

1.97

0.02

4

2.55

4.65

1.97

0.02

Parameters of predictors1
Variable
Estimate Std. Error
Intercept
0.755
0.151
Average DBH of green ash
-0.002
0.003
Green ash road density
-0.5e-03
0.7e-03
Proportion of forested land
-0.096
0.217
Proportion of open land
0.161
0.294
Average DBH of white ash along roads
0.002
0.004
Proportion of woody wetlands
-0.164
0.441
Average DBH of white ash
0.7e-03
0.003
Average DBH of green ash along roads
-0.4e-03
0.002
1

Adjusted SE
0.154
0.003
0.7e-03
0.221
0.300
0.004
0.450
0.003
0.002

z value
4.885
0.916
0.723
0.434
0.536
0.423
0.364
0.284
0.211

Parameters were derived from full-model averaging. This approach is recommended by Burnham
and Anderson (2002) when high model uncertainty exists, where each averaged parameter estimate
is weighted to reduce the influence of variables with low Akaike parameter weights (w i).

99

Table 3.6. Competing models and parameters calculated for variables and used to predict percent
white ash survival (delta AICc < 2.0) in the center of areas bounded by a 1500 m radius in 28
sites in southeast and south central Michigan. 1Full-model averaging results from my 1200 to
1500 m model.
Competing models
Variables
df
logLik AICc delta weight
Green ash road density + proportion
4
6.54
-3.34
0
0.36
of woody wetlands
Green ash road density + average
DBH of white ash + proportion of
5
7.23
-1.73 1.61
0.16
woody wetlands

Variables
Intercept
Green ash road density
Proportion of woody wetlands
Average DBH of white ash
Average DBH of white ash along
roads
Proportion of open land
White ash road density
Average DBH of green ash along
roads
Average DBH of green ash

Parameters of predictors1
Estimate
Std. Error
9.013e-01
8.674e-02
-5.417e-03
2.285e-03
-1.241e+00
5.011e-03
5.408e-04
1.737e-03

Adjusted SE
9.095e-02
2.404e-03
5.273e-01
1.783e-03

z value
9.911
2.253
2.353
0.303

-4.050e-04
1.922e-02
1.864e-04

1.810e-03
1.012e-01
1.956e-03

1.871e-03
1.049e-01
2.050e-03

0.217
0.183
0.091

4.208e-05
-5.174e-06

1.079e-03
9.253e-04

1.135e-03
9.743e-04

0.037
0.005

1

Parameters were derived from full-model averaging. This approach is recommended by Burnham
and Anderson (2002) when high model uncertainty exists, where each averaged parameter
estimate is weighted to reduce the influence of variables with low Akaike parameter weights (wi).

100

Figure 3.1. Location of 28 white ash sites in southeast and south central Michigan.

101

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28

Percent White Ash Survival

100%
90%
80%
70%
60%
50%
40%
30%
20%
10%
0%

Sites
Figure 3.2. Percentage of live and dead white ash stems alive in 28 sites surveyed in southeast
Michigan in 2016.

102

Scale

Importance Value

Potential Predictor Variables
Figure 3.3: The relative variable importance for models developed to predict white ash survival rates in areas bounded by 400 m, 800
m, 1200 m, 1500 m, and cross-scale models. Radii and the cross-scale model for the entire 7.1 km2 area bounded by the 1.5 km radius.
Variables with a relative variable importance of 1.00 are of high importance in predicting white ash survival rate.

103

LITERATURE CITED

104

LITERATURE CITED

Abell, K. J., Bauer, L. S., Duan, J. J., & Van Driesche, R. (2014). Long-term monitoring of the
introduced emerald ash borer (Coleoptera: Buprestidae) egg parasitoid, Oobius agrili
(Hymenoptera: Encyrtidae), in Michigan, USA and evaluation of a newly developed
monitoring technique. Biological Control, 79, 36–42.
Abell, K. J., Duan, J. J., Bauer, L., Lelito, J. P., & Van Driesche, R. G. (2012). The effect of bark
thickness on host partitioning between Tetrastichus planipennisi (Hymenoptera:
Eulophidae) and Atanycolus spp. (Hymenoptera: Braconidae), two parasitoids of emerald
ash borer (Coleoptera: Buprestidae). Biological Control, 63, 320-325.
Alnosaier, W. S. (2007). Kenward-Roger approximate F test for fixed effects in mixed linear
models (Doctoral dissertation).
Anulewicz, A.C., McCullough, D.G., and Miller, D.L. (2006). Oviposition and development of
emerald ash borer (Agrilus planipennis)(Coleoptera: Buprestidae) on hosts and potential
hosts in no-choice bioassays.
Anulewicz, A. C., McCullough, D. G., & Cappaert, D. L. (2007). Emerald ash borer (Agrilus
planipennis) density and canopy dieback in three North American ash species.
Arboriculture and Urban Forestry, 33, 338-349.
Anulewicz, A. C., McCullough, D. G., Cappaert, D. L., & Poland, T. M. (2008). Host range of
the emerald ash borer (Agrilus planipennis Fairmaire) (Coleoptera:Buprestidae) in North
America: results of multiple-choice field experiments. Environmental Entomology, 37,
230-241.
Aukema, J. E., Leung, B., Kovacs, K., Chivers, C., Britton, K. O., Englin, J., Frankel, S. J.,
Haight, R. G., Holmes, T. P., Liebhold, A., McCullough, D. G., & Von Holle, B. (2011).
Economic impacts of non-native forest insects in the continental United States. PLoS
ONE, 6, e24587.
Baranchikov, Y., Mozolevskaya, E., Yurchenko, G., & Kenis, M. (2008). Occurrence of the
emerald ash borer, Agrilus planipennis, in Russia and its potential impact on European
forestry. EPPO Bulletin, 38, 233-38.
Bauer, L. S., Duan, J. J., Gould, J. R., Abell, K. J., Lelito, J. P., & Van Driesche, R. (2014).
Classical Biocontrol of Emerald Ash Borer in North America: Monitoring Demonstrates
Successful Establishment. In USDA InterAgency reSeArch ForUm on InvASIve
SpecIeS (p. 15).
Bauer, L. S., Duan, J. J., Gould, J. R., & Van Driesche, R. (2015). Progress in the classical
biological control of Agrilus planipennis Fairmaire (Coleoptera: Buprestidae) in North
America. Canadian Entomologist, 147, 300-317.
105

Bauer, L. S., Liu, H., Miller, D., & Gould, J. (2008). Developing a classical biological control
program for Agrilus planipennis (Coleoptera: Buprestidae), an invasive ash pest in North
America. Newsletter of the Michigan Entomological Society, 53, 38-39.
Biggs, A. R., Merrill, W., & Davis, D. D. (1984). Discussion: response of bark tissues to injury
and infection. Canadian Journal of Forest Research, 14, 351-356.
Boerner, R. E. J., & Brinkman, J. A. (1996). Ten years of tree seedling establishment and
mortality in an Ohio deciduous forest complex. Bulletin of the Torrey Botanical Club,
123, 309-317. [48378]
Brown-Rytlewski, D. E., & Wilson, M. A. (2005). Tracking the emergence of emerald ash borer
adults. Mastro and Reardon, 13-14.
Buckland, S., Burnham, K., & Augustin, N. (1997). Model Selection: An Integral Part of
Inference. Biometrics, I, 603-618.
Burnham, K. P., & Anderson, D. R. (2002). Model Selection and Multimodel Inference. A
Practical Information—Theoretic Approach, 2nd edn, New York Springer-Verlag.
Burr, S. J., & McCullough, D. G. (2014). Condition of green ash (Fraxinus pennsylvanica)
overstory and regeneration at three stages of the emerald ash borer invasion wave.
Canadian Journal of Forest Research, 44, 768-776.
Burr, S. J., & McCullough, D.G. (201X). Density of Emerald Ash Borer (Agrilus planipennis)
(Coleoptera:Buprestida) Adults and Larvae at Three Stages of the Invasion Wave.
In press.
Brunsdon, C., Charlton, M., & Harris, P. (2012). Living with collinearity in local regression
Models. In proceedings of the 10th International Symposium on Spatial Accuracy
Assessment in Natural Resources and Environmental Science, Florianopolis SC, Brazil.
Canadian Food Inspection Agency. (2016). Emerald ash borer—Agrilus planipennis.
http://www.inspection.gc.ca/plants/plant-pests-invasive-species/insects/emerald-ashborer/fact-sheet/eng/1337368130250/1337368224454
Cappaert, D., McCullough, D. G., Poland, T. M., & Siegert, N. W. (2005). Emerald ash borer in
North America: a research and regulatory challenge. American Entomologist, 51, 152165.
Cappaert, D., & McCullough, D.G. (2009). Occurrence and seasonal abundance of Atanycolus
cappaerti (Hymenoptera: Braconidae) a native parasitoid of emerald ash borer, Agrilus
planipennis (Coleoptera: Buprestidae). Great Lakes Entomologist. 42(1/2): 16-29.

106

Chen, Y., & Poland, T. M. (2010). Nutritional and defensive chemistry of three North American
ash species: possible roles in host performance by emerald ash borer adults. Great Lakes
Entomologist, 43, 20-33.
Clark, R. E., Boyes, K. N., Morgan, L. E., Storer, A. J., & Marshall, J. M. (2015). Development
and assessment of ash mortality models in relation to emerald ash borer infestation.
Arboriculture & Urban Forestry, 41, 270-278.
Crook, D. J., & Mastro, V. C. (2010). Chemical ecology of the emerald ash borer Agrilus
planipennis. Journal of Chemical Ecology, 36, 101-112.
Crook, D. J., Khrimian, A., Cossé, A. A., Fraser, I., & Mastro, V. C. (2012). Influence of trap
color and host volatiles on capture of the emerald ash borer (Coleoptera: Buprestidae).
Economic Entomology, 105, 429-437.
Crook, D. J., Khrimian, A., Francese, J. A., Fraser, I., Poland, T. M., Sawyer, A. J., & Mastro,
V. C. (2008). Development of a host-based semiochemical lure for trapping emerald ash
borer Agrilus planipennis (Coleoptera: Buprestidae). Environmental Entomology, 37,
356-365.
Crook, D. J., Francese, J. A., Rietz, M. L., Lance, D. R., Hull-Sanders, H. M., Mastro, V. C.,
Silk, P. J., & Ryall, K. L. (2014). Improving detection tools for emerald ash borer
(Coleoptera: Buprestidae): Comparison of multifunnel traps, prism traps, and lure types
at varying population densities. Journal of Economic Entomology, 107, 1496-1501.
Crook, D. J., Francese, J. A., Zylstra, K. E., Fraser, I., Sawyer, A. J., Bartels, D. W., Lance, D.
R., & Mastro, V. C. (2009). Laboratory and field response of the emerald ash borer
(Coleoptera: Buprestidae), to selected regions of the electromagnetic spectrum. Journal
of Economic Entomology, 102, 2160-2169.
de Groot, P., Gary, G. G., Poland, T. M., Scharbach, R., Buchan, L., Nott, R. W.,
Macdonald, L., & Pitt, D. (2008). Electrophysiological response and attraction of emerald
ash borer to green leaf volatiles (GLVs) emitted by host foliage. Journal of Chemical
Ecology, 34, 1170-1179.
Domingue, M. J., Lelito, J. P., Fraser, I., Mastro, V. C., Tumlinson, J.H. & Baker, T. C. (2013).
Visual and chemical cues affecting the detection rate of the emerald ash borer in sticky
traps. Journal of Applied Entomology, 137, 77-87.
Duan, J. J., Bauer, L. S., Abell, K. J., Lelito, J. P., & Van Driesche, R. (2013a). Establishment
and abundance of Tetrastichus planipennisi (Hymenoptera: Eulophidae) in Michigan:
potential for success in classical biocontrol of the invasive emerald ash borer (Coleoptera:
Buprestidae). Journal of Economic Entomology, 106, 1145-1154.

107

Duan, J. J., Bauer, L. S., Abell, K. J., Ulyshen, M. D., & Van Driesche, R. G. (2015). Population
dynamics of an invasive forest insect and associated natural enemies in the aftermath of
invasion: implications for biological control. Journal of Applied Ecology, 52, 1246-1254.
Duan, J. J., Bauer, L. S., Hansen, J. A., Abell, K. J., & Van Driesche, R. (2012a). An improved
method for monitoring parasitism and establishment of Oobius agrili (Hymenoptera:
Encyrtidae), an egg parasitoid introduced for biological control of the emerald ash borer
(Coleoptera: Buprestidae) in North America. Biological Control, 60, 255-261.
Duan, J. J., Bauer, L. S., Ulyshen, M. D., Gould, J. R., & Van Driesche, R. (2011). Development
of methods for the field evaluation of Oobius agrili (Hymenoptera: Encyrtidae) in North
America, a newly introduced egg parasitoid of emerald ash borer (Coleoptera:
Buprestidae). Biological Control, 56, 170-174.
Duan, J. J., Taylor, P. H., Fuester, R. W., Kula, R. R., & Marsh, P. M. (2013b). Hymenopteran
parasitoids attacking the invasive emerald ash borer (Coleoptera: Buprestidae) in western
and central Pennsylvania. Florida Entomologist, 96, 166-172.
Duan J.J., Ulyshen, M.D., Bauer L.S., Gould J., and van Driesche R. (2010). Measuring the
impact of biotic factors on populations of immature emerald ash borer (Coleoptera:
Buprestidae). Environ. Entomol. 39:1513–22
Duan, J. J., Yurchenko, G., & Fuester, R. W. (2012b). Occurrence of emerald ash borer
(Coleoptera: Buprestidae) and biotic factors affecting its immature stages in the
Russian Far East. Environmental Entomology, 41, 245-254.
Elith, J., Graham, C. H., Anderson, P. R., Dudík, M., Ferrier, S., Guisan, A., Hijmans, R. J.,
Huettmann, F., Leathwick, J. R., Lehmann, A., Li, J., Lohmann, L. G., & Loiselle, B. A.,
(2006). Novel methods improve prediction of species’ distributions from occurrence data.
Ecography, 29, 129-151.
EAB.info (2017). Emerald Ash Borer. http://www.emeraldashborer.info/index.cfm. Accessed
June 2017.
Federal Register. (2007). Availability of an environmental assessment for the proposed release of
three parasitoids for the biological control of the emerald ash borer (Agrilus planipennis)
in the continental United States. Federal Register, 72: 28947–28948. [docket number
APHIS-2007-006]. Available from http://www.regulations.gov/#!documentDetail;
D=APHIS-2007-0060-0043 [accessed 21 January 2017].
Fierke, M. K., Kinney, D. L., Salisbury, V. B., Crook, D. J., & Stephen, F. M. (2005).
Development and comparison of intensive and extensive sampling methods and
preliminary within-tree population estimates of red oak borer (Coleoptera:
Cerambycidae) in the Ozark Mountains of Arkansas. Environmental Entomology, 34,
184-192.

108

Fierke, M. K., & Stephen, F. M. (2007). Red oak borer (Coleoptera: Cerambycidae) flight
trapping in the Ozark National Forest, Arkansas. Florida Entomologist, 90, 488-494.
Flower, C. E., Knight, K. S., & Gonzalez-Meler, M. A. (2013a). Impacts of the emerald ash
borer (Agrilus planipennis Fairmaire) induced ash (Fraxinus spp.) mortality on forest
carbon cycling and successional dynamics in the eastern United States. Biological
Invasions, 15, 931-944.
Flower, C. E., Knight, K. S., Rebbeck, J., & Gonzalez-Meler, M. A. (2013b). The relationship
between the emerald ash borer (Agrilus planipennis) and ash (Fraxinus spp.) tree decline:
Using visual canopy condition assessments and leaf isotope measurements to assess pest
damage. Forest Ecology and Management, 303, 143-147.
Francese, J. A., Mastro, V. C., Oliver, J. B., Lance, D. R., Youssef, N., & Lavallee, S. G. (2005).
Evaluation of colors for trapping Agrilus planipennis (Coleoptera: Buprestidae). Journal
of Entomological Science, 40, 93-95.
Francese, J. A., Oliver, J. B., Fraser, I., Youssef, N., Lance, D. R., Crook, D. J., & Mastro, V. C.
(2006). Effects of trap design and placement on capture of emerald ash borer, Agrilus
planipennis. In V. Mastro, D. Lance, R. Reardon, and G. Parra (compilers), Emerald ash
borer research and technology development meeting, 26-27 September 2005, Pittsburgh,
PA. U.S. Department of Agriculture, Morgantown, WV. (p. 57-58).
Francese, J. A., Oliver, J. B., Fraser, I., Lance, D. R., Youssef, N., Sawyer, A. J., & Mastro,
V. C. (2008). Influence of trap placement and design on capture of the emerald ash borer
(Coleoptera: Buprestidae). Economic Entomology, 101, 1831-1837.
Francese, J. A., Crook, D. J., Fraser, I., Lance, D. R., Sawyer, A. J., & Mastro, V. C. (2010).
Optimization of trap color for emerald ash borer (Coleoptera: Buprestidae). Economic
Entomology, 103, 1235-1241.
Francese, J. A., Rietz, M. L., Crook, D. J., Fraser, I., Lance, D. R., & Mastro, V. C. (2013a).
Improving detection tools for the emerald ash borer (Coleoptera: Buprestidae):
comparison of prism and multifunnel traps at varying population densities. Economic
Entomology, 106, 2407-2414.
Francese, J. A., Rietz, M. L., & Mastro, V. C. (2013b). Optimization of multifunnel traps for
emerald ash borer (Coleoptera: Buprestidae): Influence of size, trap coating, and color.
Economic Entomology, 106, 2415-2423.
Fraser, I., Francese, J. A., Lance, D. R., Crook, D. J., Oliver, J. B., Youssef, N., & Mastro, V. C.
(2006). Effects of tree wounding and banding on emerald ash borer capture, pp. 59-60. In
V. Mastro, D. Lance, R. Reardon, and G. Parra (compilers), Emerald ash borer research
and technology development meeting, 26-27 September 2005, Pittsburgh, PA. U.S.
Department of Agriculture, Morgantown, WV.

109

Gandhi, K. J. K., & Herms, D. A. (2010). Potential biodiversity loss due to impending
devastation of the North American genus Fraxinus by the exotic emerald ash borer.
Biological Invasions, 12, 1839-1846.
Grant, G. G., Ryall, K. L., Lyons, D. B., & Abou‐Zaid, M. M. (2010). Differential response of
male and female emerald ash borers (Coleoptera: Buprestidae) to (Z)‐3‐hexenol and
manuka oil. Applied Entomology, 134, 26-33.
Grant, G. G., Poland, T. M., Ciaramitaro, T., Lyons, D. B., & Jones, G. C. (2011). Comparison
of male and female emerald ash borer (Coleoptera: Buprestidae) responses to phoebe oil
and (Z)-3- hexenol lures in light green prism traps. Economic Entomology, 104, 173-179.
Griffith, R. S. (1991). Fraxinus americana. In: Fire effects information system (online). USDA
Forest Service, Rocky Mountain Research Station, Fire Sciences Laboratory, p 32.
Gucker, C. L. (2005). Fraxinus pennsylvanica. In: Fire Effects Information System, (Online).
U.S. Department of Agriculture, Forest Service, Rocky Mountain Research Station, Fire
Sciences Laboratory.
Herms, D. A., & McCullough, D. G. (2014). Emerald ash borer invasion of North America:
history, biology, ecology, impacts, and management. Annual Review of Entomology, 59,
13-30.
Hilbe, J. M. (2011). Negative binomial regression. Cambridge University Press.
Homer, C. G., Dewitz, J. A., Yang, L., Jin, S., Danielson, P., Xian, G., Coulston, J., Herold, N.
D., Wickham, J. D., and Megown, K. (2015). Completion of the 2011 National Land
Cover Database for the conterminous United States-Representing a decade of land cover
change information. Photogrammetric Engineering and Remote Sensing, 81, p. 345-354.
Hosie, R. C. (1969). Native trees of Canada. 7th ed. Ottawa, ON: Canadian Forestry Service,
Department of Fisheries and Forestry. 380 p. [3375]
Jennings, D. E., Taylor, P. B., and Duan, J. J. (2014). The mating and oviposition behavior of the
invasive emerald ash borer (Agrilus planipennis), with reference to the influence of host
tree condition. Journal of Pest Science, 87, 71-78.
Jenkins, S. G., Oduor, P. G., Kotchman, L., & Kangas, M. (2016). A cohort model for ash
mortality risk due to potential emerald ash borer infestation. Open Journal of
Biophysics, 6, 38-51.
Kent, M., & Coker, P. (1992). Vegetation description and analysis: a practical approach. John
Wiley and Sons, New York.
Klooster, W. S., Herms, D. A., Knight, K. S., Herms, C. P., McCullough, D. G., Smith, A.,
Gandhi, K. J. K., & Cardina, J. (2014). Ash (Fraxinus spp.) mortality, regeneration, and
110

seed bank dynamics in mixed hardwood forests following invasion by emerald ash borer
(Agrilus planipennis). Biological Invasions, 16, 859-873.
Knight, K. S., Brown, J. P., & Long, R. P. (2013). Factors affecting the survival of ash (Fraxinus
spp.) trees infested by emerald ash borer (Agrilus planipennis). Biological Invasions, 15,
371-383.
Knight, K. S., Flash, B. P., Kappler, R. H., Throckmorton, J. A., Grafton, B., & Flower, C. E.
(2014). Monitoring ash (Fraxinus spp.) decline and emerald ash borer (Agrilus
planipennis) symptoms in infested areas. Gen Tech Rep NRS-139. US Department of
Agriculture, Forest Service, Northern Research Station, Newtown Square, PA.
https://www.fs.fed.us/nrs/pubs/gtr/gtr_nrs139.pdf.
Knight, K. S., Herms, D., Plumb, R., Sawyer, E., Spalink, D., Pisarczyk, E., Wiggin, B., Kappler,
R., & Menard, K. (2012). Dynamics of surviving ash (Fraxinus spp.) populations in areas
long infested by emerald ash borer (Agrilus planipennis). In proceedings of the 4th
international workshop on the genetics of host-parasite interactions in forestry: disease
and insect resistance in forest trees; 2011 July 31-August 5; Eugene, OR. Gen. Tech.
Rep. PSW- 240. Albany, CA: U.S. Department of Agriculture, Forest Service, Pacific
Southwest Research Station: 143-152.
Koch, J. L., Carey, D. W., Knight, K. S., Poland, T., Herms, D. A., & Mason, M. E.
(2012). Breeding strategies for the development of emerald ash borer resistant - North
American ash. In proceedings of the 4th international workshop on the genetics of hostparasite interactions in forestry: disease and insect resistance in forest trees, 2011 July 31August 5; Eugene, OR. Gen. Tech. Rep. PSW-240. Albany, CA: U.S. Department of
Agriculture, Forest Service, Pacific Southwest Research Station: 235-239.
Koch, J. L., Carey, D. W., Mason, M. E., Poland, T. M., & Knight, K. S. (2015). Intraspecific
variation in Fraxinus pennsylvanica responses to emerald ash borer (Agrilus
planipennis). New Forests, 46, 995-1011.
Kovacs, K. F., Haight, R. G., McCullough, D. G., Mercader, R. J., Siegert, N. W., & Liebhold.,
A. M. (2010). Cost of potential emerald ash borer damage in U.S. communities, 20092019. Ecological Economics, 69, 569-578.
Kovacs, K. F., Mercader, R. J., Haight, R. G., Siegert, N. W., McCullough, D. G., &
Liebhold, A. M. (2011). The influence of satellite populations of emerald ash borer on
projected economic costs in US communities, 2010-2020. Journal of Environmental
Management, 92, 2170-2181.
Lelito, J. P., Fraser, I., Mastro, V. C., Tumlinson, J. H., Böröczky, K., & Baker, T. C. (2007).
Visually mediated ‘paratrooper copulations’ in the mating behavior of Agrilus
planipennis (Coleoptera: Buprestidae), a highly destructive invasive pest of North
American ash trees. Journal of Insect Behavior, 20, 537-552.

111

Limback, C. K. (2010). Tree vigor and its relation to emerald ash borer (Agrilus planipennis
Fairmarie) adult host preference and larval development on green and white ash trees, p.
98. M.S. Thesis, Department of Entomology, Michigan State University. East Lansing,
MI.
Liu, H. P., Bauer, L. S., Gao, R. T., Zhao, T. H., Petrice, T. R., & Haack, R. A. (2003).
Exploratory survey for the emerald ash borer, Agrilus planipennis (Coleoptera:
Buprestidae), and its natural enemies in China. Great Lakes Entomologist, 36, 191-204.
Lovett, G. M., Canham, C. D., Arthur, M. A., Weathers, K. C., & Fitzhugh, R. D. (2006). Forest
ecosystem responses to exotic pests and pathogens in eastern North America. Bioscience,
56, 395-405.
MapBioControl. (2016). Agent release tracking and data management for federal, state, and
researchers releasing three biocontrol agents released against emerald ash borer.
(http://www.mapbiocontrol.org/).
Marshall, J. M., Smith, E. L., Mech, R., & Storer, A. J. (2013). Estimates of Agrilus planipennis
infestation rates and potential survival of ash. The American Midland Naturalist, 169,
179-193.
Marshall, J. M., Storer, A. J., Fraser, I., Beachy, J. A., & Mastro, V. C. (2009). Effectiveness of
differing trap types for the detection of emerald ash borer (Coleoptera:
Buprestidae). Environmental Entomology, 38, 1226-1234.
Marshall, J. M., Storer, A. J., Fraser, I., & Mastro, V. C. (2010a). Efficacy of trap and lure types
for detection of Agrilus planipennis (Coleoptera: Buprestidae) at low density. Journal of
Applied Entomology, 134, 296-302.
Marshall J. M., Storer A. J., Fraser I., and Mastro V. C. (2010b). Multi-state comparison of
detection tools at low emerald ash borer densities. In. Mastro V. Lance D. Reardon R.
Parra G. Proceedings of the Emerald Ash Borer and Asian Longhorned Beetle Research
and Technology Development MeetingPittsburgh, PA19-21 OctU.S. Department of
Agriculture, Forest Service Publication, Morgantown, WV.
Marshall, J. M., Smith, E. L., Mech, R., and Storer, A. J. (2013). Estimates of Agrilus
planipennis infestation rates and potential survival of ash. The American Midland
Naturalist. 169, 179-193.
McCullough, D. G., Mercader, R. J., & Siegert, N. W. (2015). Developing and integrating tactics
to slow ash (Oleaceae) mortality caused by emerald ash borer (Coleoptera:
Buprestidae). The Canadian Entomologist, 147, 349-358.
McCullough, D. G., & Poland, T. M. (2009c). Using double-decker traps to detect emerald ash
borer. Emerald Ash Borer Info Research Reports: Survey Research, 9 pp.
http://www.emeraldashborer.info/files/ double_decker_eab_trap_guide.pdf

112

McCullough, D. G., Poland, T. M., Anulewicz, A. C., & Cappaert, D. L. (2008). Doubledeckers and towers: Emerald ash borer traps in 2007.
McCullough, D. G., Poland, T. M., Anulewicz, A. C., & Cappaert, D. (2009a). Emerald ash
borer (Coleoptera: Buprestidae) attraction to stressed or baited ash trees. Environmental
Entomology, 38, 1668-1679.
McCullough, D. G., Poland, T. M., & Cappaert, D. (2009c). Attraction of the emerald ash borer
to ash trees stressed by girdling, herbicide treatment, or wounding. Canadian Journal of
Forest Research, 39, 1331-1345.
McCullough, D. G., Schneeberger, N. F., & Katovich, S. A. (2008). Emerald ash borer pest alert.
NA-PR-02-04. USDA Forest Service.
https://www.na.fs.fed.us/spfo/pubs/pest_al/eab/eab.pdf.
McCullough, D. G., & Siegert, N. W. (2007). Estimating potential emerald ash borer
(Coleoptera: Buprestidae) populations using ash inventory data. Journal of Economic
Entomology, 100, 1577-1586.
McCullough, D. G., Siegert, N. W., Poland, T. M., Pierce, S. J., & Ahn, S. Z. (2011). Effects of
trap type, placement and ash distribution on emerald ash borer captures in a low density
site. Environmental Entomology, 40, 1239-1252.
McCullough, D. G., Work, T. T., Cavey, J. F., Liebhold, A. T., & Marshall, D. (2006).
Interceptions of nonindigenous plant pests at U.S. ports of entry and border crossings
over a 17 year period. Biological Invasions, 8, 611-630.
Meiners, S. J., Handel, S. N., & Pickett, S. T. A. (2000). Tree seedling establishment under
insect herbivory: edge effects and inter-annual variation. Plant Ecology, 151, 161-170.
Mercader, R. J., McCullough, D. G., & Bedford, J. M. (2013). A comparison of girdled ash
detection trees and baited artificial traps for emerald ash borer (Agrilus planipennis
Fairmaire) detection. Environmental Entomology, 42, 1027-1039.
Mercader, R. J., McCullough, D. G., Storer, A. J., Bedford, J. M., Heyd, R., Poland, T. M., &
Katovich, S. (2015). Evaluation of the potential use of a systemic insecticide and girdled
trees in area wide management of the emerald ash borer. Forest Ecology and
Management, 350, 70-80.
Mercader, R. J., Siegert, N. W., Liebhold, A. M., & McCullough, D. G. (2009). Dispersal of the
emerald ash borer, Agrilus planipennis, in newly-colonized sites. Agricultural and Forest
Entomology, 11, 421-424.
Mercader, R. J., Siegert, N. W., Liebhold, A. M., & McCullough, D. G. (2011a). Estimating the

113

effectiveness of three potential management options to slow the spread of emerald ash
borer populations in localized outlier sites. Canadian Journal of Forest Research, 41,
254–64.
Mercader, R. J., Siegert, N. W., Liebhold, A. M., & McCullough, D. G. (2011b). Influence of
foraging behavior and host spatial distribution on the localized spread of the emerald ash
borer, Agrilus planipennis. Population Ecology, 53, 271-285.
Mercader, R. J., Siegert, N. W., & McCullough, D. G. (2012). Estimating the influence of
population density and dispersal behavior on the ability to detect and monitor Agrilus
planipennis (Coleoptera: Buprestidae) populations. Economic Entomology, 105, 272-281.
Miller, R. O., Bloese, P. D., Hanover, J. W., & Haack, R. A. (1991). Paper birch and European
white birch vary in growth and resistance to bronze birch borer. Journal of the American
Society for Horticultural Science, 116, 580-584.
Morin, R. S., Liebhold, A. M., Pugh, S. A., & Crocker, S. J. (2016). Regional assessment of
emerald ash borer, Agrilus planipennis, impacts in forests of the Eastern United
States. Biological Invasions, 19, 703-711.
Mullick, D. B. (1977). The non-specific nature of defense in bark and wood during wounding,
insect and pathogen attack. In The Structure, Biosynthesis, and Degradation of Wood (pp.
395-441). Springer US.
Mullick, D. B., & G. D., Jensen. (1973). New concepts and terminology of coniferous periderms:
necrophylactic and exophylactic periderms. Canadian Journal of Botany, 51, 1459-1470.
Neter, J., Kutner, M. H., Nachtsheim, C. J., & Wasserman, W. (1996). Applied linear statistical
Models, vol. 4, p. 318. Chicago: Irwin.
Nielsen, D. G., Muilenburg, V. L., & Herms, D. A. (2011). Interspecific variation in resistance of
Asian, European, and North American birches (Betula spp.) to bronze birch borer
(Coleoptera:Buprestidae). Environmental Entomology, 40, 648-653.
Nisbet, D., Kreutzweiser, D., Sibley, P., & Scarr, T. (2015). Ecological risks posed by emerald
ash borer to riparian forest habitats: A review and problem formulation with management
implications. Forest Ecology and Management, 358, 165-173.
NRCS (USDA Natural Resource Conservation Service). (2017). Plants Database. Fraxinus
Americana L. http://plants.usda.gove/core/profile?symbol=fram2. Accessed September
2017.
Orlova-Bienkowskaja, M. J. (2014). European range of the emerald ash borer Agrilus
planipennis (Coleoptera: Buprestidae) is expanding: The pest destroys ashes in the
northwest of Moscow Oblast and in part of Tver Oblast. Russian journal of biological
invasions, 5(1), 32-37.
114

Poland, T. M. (2007). Twenty million ash trees later: current status of emerald ash borer in
Michigan. Newsletter of the Michigan entomological society, 52, 10-14.
Poland, T. M., de Groot, P., Grant, G., MacDonald, L., & McCullough, D. G. (2004).
Developing attractants and trapping techniques for the emerald ash borer, pp. 15-16. In
V. Mastro and R. Reardon (compilers), Emerald ash borer research and technology
development meeting, 30 September to 1 October 2003, Port Huron, MI. U.S.
Department of Agriculture, Morgantown, WV.
Poland, T. M., & McCullough, D. G. (2006). Emerald ash borer: Invasion of the urban forest and
the threat to North America’s ash resource. Journal of Forestry, 104, 118-124.
Poland, T. M., & McCullough, D. G. (2014). Comparison of trap types and colors for capturing
emerald ash borer adults at different population densities. Environmental Entomology,
43, 157-179.
Poland, T. M., McCullough, D. G., & Anulewicz, A. C. (2011). Evaluation of double-decker
traps for emerald ash borer (Coleoptera: Buprestidae). Economic Entomology, 104, 517531.
Poland, T. M., McCullough, D. G., de Groot, P., Grant, G. G., & Cappaert, D. (2005). Progress
toward developing trapping techniques for the emerald ash borer, pp. 53-54. In V. Mastro
and R. Reardon (compilers), Emerald ash borer research and technology development
meeting, 5- 6 October 2004, Romulus, MI. U.S. Department of Agriculture, Morgantown,
WV.
Poland, T. M., Petrice, T. R., & Ciaramitaro, T. M. (2014). Effects of trap design, color, and lure
on emerald ash borer attraction. In Proceedings of the Emerald Ash Borer Research and
Technology Development Meeting. Wooster, Ohio. 19-20 October 2014 (p. 100).
Pugh, S. A., Liebhold, A. M., & Morin, R. S. (2011). Changes in ash tree demography associated
with emerald ash borer invasion, indicated by regional forest inventory data from the
Great Lakes States. Canadian Journal of Forest Research, 41, 2165-2175.
Pureswaran, D. S., & Poland, T. M. (2009). Host selection and feeding preference of Agrilus
planipennis (Coleoptera: Buprestidae) on ash (Fraxinus spp.). Environmental
Entomology, 38, 757-765.
R Core Development Team. 2016. R: a language and environment for statistical computing.
(http://www.R-project.org).
Rebek, E. J., Herms, D. A., & Smitley, D. R. (2008). Interspecific variation in resistance to
emerald ash borer (Coleoptera: Buprestidae) among North American and Asian ash
(Fraxinus spp.). Environmental Entomology, 37, 242-246.
Robinett, M. A., McCullough, D. G., & Poland, T. M. (2014). Evaluation of two trap designs in

115

stands with varying levels of white ash mortality. In Proceedings of the Emerald Ash
Borer Research and Technology Development Meeting. Wooster, Ohio. 19-20 October
2014 (p. 98).
Rodriguez-Saona, C. R., Miller, J. R., Poland, T. M., Kuhn, T. M., Otis, G. W., Turk, T., &
Ward, D. L. (2007). Behaviors of adult Agrilus planipennis (Coleoptera:
Buprestidae). Great Lakes Entomologist, 40, 1-16.
Rodriguez-Saona, C. R., Poland, T. M., Miller, J. R., Stelinski, L. L., Grant, G. G., De Groot, P.,
& MacDonald, L. (2006). Behavioral and electrophysiological responses of the emerald
ash borer, Agrilus planipennis, to induced volatiles of Manchurian ash, Fraxinus
mandshurica. Chemoecology, 16, 75-86.
Ryall, K. (2015). Detection and sampling of emerald ash borer (Coleoptera: Buprestidae)
infestations. The Canadian Entomologist, 147, 290-299.
Ryall, K. L., Silk, P. J., Mayo, P., Crook, D., Khrimian, A., Cossé, A. A., Sweeney, J. and Scarr,
T. (2012). Attraction of Agrilus planipennis (Coleoptera: Buprestidae) to a volatile
pheromone: effects of release rate, host volatile, and trap placement. Environmental
entomology, 41(3), pp.648-656.
Ryall, K. L., Fidgen, J. G., Silk, P. J. and Scarr, T. A. (2013). Efficacy of the pheromone (3Z)‐
lactone and the host kairomone (3Z)‐hexenol at detecting early infestation of the emerald
ash borer, Agrilus planipennis. Entomologia Experimentalis et Applicata, 147(2), pp.126131.
SAS Institute. (2003). PROC user’s manual, version 9.1. SAS Institute, Cary, NC.
Schlesinger, R. C. (1990). Fraxinus americana L. white ash. 333–338. In: Silvics of North
American hardwoods U.S.D.A. Forest Service Agric. Handbook 654.
Shono, H., Peng, C. K., Goldberger, A. L., Shono M., & Sugimori, H. (2000). A new method to
determine a fractal dimension of non-stationary biological time-serial data. Computers in
Biology and Medicine, 30, 237-245.
Siegert, N. W., McCullough, D. G., Williams, D. W., Fraser, I., Poland, T. M., & Pierce, S. J.
(2010). Dispersal of Agrilus planipennis (Coleoptera: Buprestidae) from Discrete
Epicenters in Two Outlier Sites. Environmental Entomology, 39, 253-265.
Siegert, N. W., McCullough, D. G., Liebhold, A. M., & Telewski, F. W. (2014).
Dendrochronological reconstruction of the epicentre and early spread of emerald ash
borer in North America. Diversity and Distributions, 20, 847-858.
Siegert, N. W., McCullough, D. G., Williams, D. W., Fraser, I., & Poland, T. M. (2010).
Dispersal of Agrilus planipennis (Coleoptera: Buprestidae) from discrete epicenters in
two outlier sites. Environmental Entomology, 39, 253-265.
116

Slager, B. Parasitoids of the emerald ash borer: An update on rearing, release, and
recovery. In Proceedings of the Emerald Ash Borer Research and Technology
Development Meeting. Wooster, Ohio. 19-20 October 2016. USDA APHIS PPQ,
Brighton, MI. In press.
Silk, P., and Ryall, K. (2015). Semiochemistry and chemical ecology of the emerald ash borer
Agrilus planipennis (Coleoptera: Buprestidae). Canadian Entomology, 147, 277-289.
Silk, P. J., Ryall, K., Mayo, P., Lemay, M. A., Grant, G., Crook, D., Cossé, A., Fraser, I.,
Sweeney, J.D., Lyons, D.B. and Pitt, D. (2011). Evidence for a volatile pheromone in
Agrilus planipennis Fairmaire (Coleoptera: Buprestidae) that increases attraction to a host
foliar volatile. Environmental Entomology, 40, 904-916.
Straw, N. A., Williams, D. T., Kulinich, O., and Gninenko, Y. I. (2013). Distribution, impact and
rate of spread of emerald ash borer Agrilus planipennis (Coleoptera: Buprestidae) in the
Moscow region of Russia. Forestry, 86, 515-522.
Symonds, M. R., & Moussalli, A. (2011). A brief guide to model selection, multimodel inference
and model averaging in behavioural ecology using Akaike’s information
criterion. Behavioral Ecology and Sociobiology, 65, 13-21.
Smith, A., Herms, D. A., Long, R. P., Gandhi, K. J. K. (2015). Community composition and
structure had no effect on forest susceptibility to invasion by the emerald ash borer
(Coleoptera: Buprestidae). Canadian Entomologist, 147, 318-328.
Tanis, S. R., & McCullough, D. G. (2012). Differential persistence of blue ash and
white ash following emerald ash borer invasion. Canadian Journal of Forest Research,
42, 1542-1550.
Tanis, S. R., & McCullough, D. G. (2015). Host resistance of five Fraxinus species to Agrilus
planipennis (Coleoptera: Buprestidae) and effects of paclobutrazol and fertilization.
Environmental Entomology, 44, 287-299.
Tanis, S. R., & McCullough, D. G. (2016). Evaluation of Xylem Discoloration in Ash Trees
Associated with Macroinjections of a Systemic Insecticide. Arboriculture and Urban
Forestry, 42(6).
Tanis, S. R., Cregg, B. M., Mota‐Sanchez, D., McCullough, D. G., and Poland, T. M. (2012b).
Spatial and temporal distribution of trunk‐injected 14C‐imidacloprid in Fraxinus
trees. Pest Management Science, 68,529-536.
Tluczek, A. R., McCullough, D. G., & Poland, T. M. (2011). Influence of host stress on emerald
ash borer (Coleoptera: Buprestidae) adult density, development, and distribution in
Fraxinus pennsylvanica trees. Environmental Entomology, 40, 357-366.

117

United States Department of Agriculture. (2013). Emerald ash borer biological control release
and recovery guidelines. United States Department of Agriculture, Animal and Plant
Health Inspection Service, Forest Service, Agricultural Research Service, Riverdale,
Maryland, United States of America. Available from
http://www.aphis.usda.gov/plant_health/plant_pest_info/emerald_ash_b/downloads/EAB
-FieldRelease-Guidelines.pdf [accessed 11 August 2014].
[USDA–APHIS] U.S. Department of Agriculture–Animal and Plant Health Inspection Service.
(2017). Plant health: emerald ash borer. Emerald ash borer survey guidelines.
http://www.aphis.usda.gov/plant_health/plant_ pest_info/emerald_ash_b/index.shtml).
U.S. Geological Survey (2014) NLCD 2011 Land Cover (2011 Edition, amended 2014) –
National Geospatial Data Asset (NGDA) Land Use Land Cover: U.S. Geological Survey,
Sioux Falls, South Dakota.
Wagner, D. L., & Todd, K. J. (2016). New ecological assessment for the emerald ash borer.
American Entomologist, 62, 26-35.
Wei, X., Reardon, D., Wu, Y., & Sun, J. H. (2004). Emerald ash borer, Agrilus planipennis
Fairmaire (Coleoptera:Buprestidae), in China: a review and distribution survey. Acta
Entomological Sinica, 47, 679-685.
Wei, X., Wu, Y., Reardon, D., Sun, T. H., Lu, M., & Sun, J. H. (2007). Biology and damage
traits of emerald ash borer (Agrilus planipennis Farmaire). Journal of Insect Science, 14,
367-373.
Wisz, M. S., & Guisan, A. (2009). Do pseudo-absence selection strategies influence species
distribution models and their predictions? An information-theoretic approach based on
simulated data. BMC Ecology, 9, 8.
Wright, Jonathan W. (1965). White ash (Fraxinus americana L.), revised. In Silvics of forest
trees of the United States. p. 191-196. H. A. Fowells, comp. U.S. Department of
Agriculture, Agriculture Handbook 271. Washington, DC.
Yang, Z. Q., Achterberg, C. V., Choi, W. Y., & Marsh, P. M. (2005). First recorded parasitoid
from China of Agrilus planipennis: a new species of Spathius (Hymenoptera: Braconidae:
Doryctinae). Annals of the Entomological Society of America, 98, 636-642.
Yang, Z. Q., Yao, Y. X., & Wang, X. Y. (2006). A new species of emerald ash borer parasitoid
from China belonging to the genus Tetrastichus (Hymneoptera: Eulophidae). In
Proceedings of the Entomological Society of Washington, 108, 550-558.
Yu, C. (1992). Agrilus marcopoli Obenberger (Coleoptera: Buprestidae). In Forest Insects of
China, 2nd Ed., Xiao, G. (ed.). China Forestry Publishing House, Beijing, China, p. 400 –
401.

118

Zarnoch, S. J., Bechtold, W. A., & Stolte, K. W. (2004). Using crown condition variables as
indicators of forest health. Canadian Journal of Forest Research, 34, 1057-1070.
Zwieniecki M. A., Melcher P. J. and Holbrook N. M. (2001). Hydraulic properties of individual
xylem vessels of Fraxinus americana. Journal of Experimental Botany, 52, 257–264.

119