Spider and Beetle Communities Across Urban Greenspaces in Cleveland, Ohio

Spider and Beetle Communities across Urban Greenspaces in Cleveland, Ohio:

Distributions, Patterns, and Processes

Dissertation

Presented in Partial Fulfillment of the Requirements for the Degree Doctor of Philosophy in the

Graduate School of The Ohio State University

Yvan A. Delgado de la flor, B.S.

Graduate Program in Entomology

The Ohio State University

2020

Dissertation Committee:

Mary M. Gardiner, Adviser

Luis Cañas

Robert J. Gates

Andrew P. Michel

William O.C. Symondson

Yvan A. Delgado de la flor

2020

Abstract

Urban areas are often considered adverse environments for wildlife communities given that the colonization and establishment of local species pools are disrupted by biotic and abiotic changes at different spatial scales such as the introduction of invasive species, periodic mowing, and changes in soil and air quality. Although the number of people residing in cities has increased in the last century, over 300 cities worldwide have shrunk due to prolonged economic decline and population loss, resulting in thousands of newly available greenspaces scattered throughout cities. Consequently, interest in urban greenspaces as sites for conservation has grown considerably, raising questions about the ability of these habitats to support wildlife. As novel ecosystems, urban areas represent a set of new challenges for local species pools, yet the mechanisms driving community assembly processes within cities is a major knowledge gap. My work focused on identifying species distributions, patterns, and processes leading to the successful establishment arthropods in cities. For this, I chose to work with beetle and spider assemblages as they are considered biological indicators of environmental changes at small and large spatial scales and are taxonomically and functionally diverse predatory groups.

In Chapter 1, my objective was to determine how urban greenspaces management and design impacts Carabidae and Staphylinidae beetles using taxonomic and life-history trait approaches. I found that ecological and morphological traits were good indicators of how beetles were responding to greenspace management strategies. Most species were negatively associated with building structures, while undisturbed habitats supported more hygrophilous and brachypterous beetle populations. In Chapter 2, I investigated the importance of local and landscape ii characteristics on spider communities using taxonomic and functional diversity approaches. I found that Pardosa milvina (Lycosidae) was the most abundant spider in our system. Plant height favored larger species, while mowing frequency benefited small spiders, leading to lower than expected functional alpha and beta diversity of spider communities. I also found that patch isolation and impervious surface increased the functional dissimilarity of spider communities. In

Chapter 3, I examined the dietary niche breadth and niche overlap of the generalist spider P. milvina using DNA gut-content analysis. I found that dietary niche breadth was not greater within urban pockets, and Diptera, Hemiptera, and Collembola were often preyed by this urban- adapted spider. I also found that P. milvina predation patterns were driven by foraging strategies, suggesting that taxonomic identify and nutritional content were more important than prey availability. In Chapter 4, my goal was to determine the habitat characteristics and landscape variables influencing spider functional groups directly and indirectly. Using structural equation modeling, I found that web spiders were negatively correlated with plant biomass, hunting spiders were positively associated with available prey breadth, and both spider functional groups were positively correlated with heavy-metal pollution in the soil.

In summary, my work showed that urban greenspace management and design play important roles shaping beetle and spider communities within cities. I also highlighted the importance of using multi-trophic and molecular approaches coupled with multivariate methodologies to disentangle local food web interactions and elucidate the mechanisms shaping community assembly processes of urban species pools. Investments in managed greenspaces that require infrequent mowing (i.e. urban prairies) will enhance the populations of beneficial arthropods, such as spiders and beetles, and beautify local neighborhoods. iii

Dedication

A mi papá Jesus y mi mamá Roxana, que de jóvenes arriesgaron mucho para que nada me faltara

y me enseñaron que podría alcanzar cualquier meta que me trazara.

A mi familia, en especial Angela, Beatriz y Alexis, y a mis amigos del Trian, que a pesar de la

distancia me brindaron su apoyo incondicional y me acompañaron durante estos años.

A mis mentores, quienes me enseñaron a navegar los mares de la educación, me guiaron por

muchos años y me ayudaron a crecer como estudiante y persona.

A mi asesora Mary por formarme como profesional, guiarme y apoyarme en todo momento, y a

mis compañeros de laboratorio, en especial Chris, con quienes compartí y aprendí tanto.

A mi prometida Melissa por ser la mejor compañera que pude haber anhelado, por apoyarme y

mantenerse a mi lado en todo momento.

A todos ustedes que de alguna u otra forma estuvieron conmigo durante esta travesía;

Gracias.

Acknowledgements

I want to acknowledge members of my graduate committee Luis Cañas, Andy Michel, Robert

Gates, and William Symondson for their support and guidance. My friends and colleagues in the

Gardiner Lab: Mary, Alex, Chelsea, Chris, Denisha, Emily, Frances, Kayla, Katie, Leo, Lydia,

MaLisa, Molly, Nicole, Sarah and Scott for their help and advice during my years in graduate school. My collaborators Caitlin Burkman, Christopher Riley, Denisha Parker, Kayla Perry,

Katherine Turo, Jennifer Thompson, Larry Phelan, Lyndsie Collins, Rodney Richardson, and

Taro Eldredge for their expertise. Faculty, staff, and graduate students in the Department of

Entomology for their willingness to help and positive attitude during my years as a graduate student. The Ohio State University and the National Science Foundation Graduate Research

Fellowship Program (DGE-1343012) for the support, resources and opportunities to succeed in graduate school. I also want to thank the hard work put by several summer fieldwork assistants.

Without this list of individuals, my research would have not been possible.

Vita

2003...... Precursores de la Independencia Nacional High School - Lima, Peru 2011...... A.A. Evergreen Valley College - San Jose, California 2014...... B.S. Humboldt State University - Arcata, California 2020...... Ph.D. The Ohio State University - Columbus, Ohio

Publications

Delgado de la flor, Y. A., K.I. Perry, K.J. Turo, D.M. Parker, J.L. Thompson, and M.M. Gardiner. 2020. Local and landscape features as drivers of the functional diversity and taxonomic of spiders across urban greenspaces. Journal of Applied Ecology. https://doi.org/10.1111/1365-2664.13636

Perry, K.I., N.C. Hoekstra, Y.A. Delgado de la flor, and M.M. Gardiner. 2020. Disentangling landscape and local drivers of ground-dwelling beetle community assembly in an urban ecosystem. Ecological Applications (In Press).

Parker D.M., K.J. Turo, Y.A. Delgado de la flor, and M.M. Gardiner. 2020. Landscape context influences the abundance and richness of native lady beetles occupying urban vacant land. Urban Ecosystems (In Press).

Delgado de la flor, Y.A., C.E. Burkman, T.K. Eldredge, and M.M. Gardiner. 2017. Patch and landscape-scale variables influence the taxonomic and functional composition of beetles in urban greenspaces. Ecosphere 8(11):e2007. https://doi.org/10.1002/ecs2.2007

Delgado de la flor, Y.A. and M.D. Johnson. 2015. Influence of invasive European beachgrass on mesopredator activity in the coastal dunes of northern California. Western Wildlife 2:29-34.

Fields of Study Major Field: Entomology

Table of Contents

Abstract ...... ii

Dedication ...... iv

Acknowledgments...... v

Vita ...... vi

List of Tables ...... ix

List of Figures ...... xi

Chapter 1: Patch and landscape-scale variables influence the taxonomic and functional composition of beetles in urban greenspaces ...... 1

1. Abstract ...... 1

2. Introduction ...... 2

3. Methods...... 6

4. Results ...... 10

5. Discussion ...... 22

6. Conclusions ...... 25

Chapter 2: Local and landscape-scale environmental filters drive the functional diversity and taxonomic composition of spiders across urban greenspaces ...... 28

1. Abstract ...... 28

2. Introduction ...... 29

3. Methods...... 32

4. Results ...... 39

5. Discussion ...... 45

6. Conclusions ...... 48 vii

Chapter 3: Landscape composition and configuration influence the dietary niche breadth and overlap of an urban-adapted generalist wolf spider ...... 50

1. Abstract ...... 50

2. Introduction ...... 51

3. Methods...... 55

4. Results ...... 61

5. Discussion ...... 65

6. Conclusions ...... 69

Chapter 4: Multi-trophic interactions reveal the biotic and abiotic components driving spiders across urban greenspaces at different spatial scales ...... 71

1. Abstract ...... 71

2. Introduction ...... 72

3. Methods...... 76

4. Results ...... 82

5. Discussion ...... 87

6. Conclusions ...... 89

References ...... 91

Appendix A: Plant species seeded in low-diversity and high-diversity prairies in Cleveland, Ohio in 2014 and 2016 ...... 125

Appendix B: Spider functional traits and mean body size (mm) ...... 127

Appendix C: Spiders sampled in Cleveland, Ohio in 2015 and 2016 ...... 131

Appendix D: Prey taxa that tested positive in the guts of Pardosa milvina ...... 136

viii

List of Tables

Table 1.1. Total number of Carabidae and Staphylinidae beetles collected in vacant lots (Vac), urban farms (Far), and urban prairies (Pra) in Cleveland, Ohio in June-August 2011 and 2012…

...... 11

Table 1.2. Canonical correspondence analysis (CCA) and partial CCA. All environmental variables that remained significant predictors following model selection are listed for each model examined (Variables). If no variable was significant for a given model, ‘None’ is listed and percent variation and p-values were not calculated. The percentage of variation (% Variation) in beetle assemblages explained by the environmental variables included in a given model is reported for all models with significant predictors. For partial CCA, ‘Constrained’ represents the effect exercised only of the desired variables of the selected model after removing the effect of the conditioned model. ‘Correlation scores’ refer to the correlation of the vectors with respect to

Axis 1 and Axis 2 of the CCA tri-plot. Data were collected from vacant lots, urban farms, and urban prairies in Cleveland, Ohio in June-August of 2012 ...... 17

Table 2.1. Functional alpha and beta diversity in 2015 and 2016 compared to null expectations using Wilcoxon signed-rank tests ...... 42

Table 3.1. Linear models among predator-prey interactions and environmental variables. Initial models included plant biomass, plant height, bloom abundance, patch size & landscape diversity at 200 m, and impervious surface & patch isolation at 1500 m radii. Final models were selected using Akaike’s Information Criterion and backward variable selection ...... 65

Table 4.1. Spiders sampled from eight vacant lots and seven pocket prairies in Cleveland, Ohio in July 2017 ...... 82

Table 4.2. Regression coefficients of the log-transformed dependent and independent variables from our final SEM model ...... 83

List of Figures

Figure 1.1. Mean (SE) Carabidae abundance (A) and richness (B), and mean (SE)

Staphylinidae abundance (C) and richness (D) in vacant lots, urban farms, and urban prairies within Cleveland, Ohio. Specimens collected in pitfall traps from 7 d sampling periods monthly in June-August 2011 and 2012. Letters indicate significant differences (p < 0.05) using Tukey’s range test ...... 13

Figure 1.2. Venn diagram of the beetle taxa found in vacant lots, community gardens, and urban prairies in Cleveland, Ohio from June-August 2011 and 2012. Numbers in brackets represent the percentage of beetle taxa that were found only in a given greenspace type ...... 14

Figure 1.3. CCA tri-plot for beetle taxa collected in at least five sites from June-August 2012.

Each square, circle, or triangle represents a given site. The proximity of sites to one another indicates the relative similarity of their beetle communities. The proximity of sites to beetle taxa indicates the relative abundance of a given taxonomic group within a site. Environmental variables are shown as vectors along two CCA axes. The proximity of sites and species to these vectors suggest a positive, neutral, or negative correlation with a given environmental variable ....

...... 16

Cleveland, Ohio. Specimens collected in pitfall traps from 7 d sampling periods monthly in June-

August 2011 and 2012. Letters indicate significant differences (p < 0.05) using Tukey’s test ......

...... 19

Figure 2.1. Map of Cleveland, Ohio showing our 32 experimental sites where spiders were sampled in 2015 & 2016 ...... 33

Figure 2.2. Four treatments were established in 2014 across 32 vacant lots in Cleveland, Ohio:

A) Control Vacant Lots (seeded fescue grass and weedy flowering plant species, mowed monthly in 2015 & 2016 reflecting the city’s management practices), B) Urban Meadows (seeded fescue grass and weedy flowering plans, mowed annually in October 2015 & 2016), C) Low-Diversity

Prairies and D) High-Diversity Prairies (mowed monthly in 2015 & annually in October 2016).

Prairies were seeded with a mixture of native grasses and flowing plants (Appendix A)...... 34

Figure 2.3. Functional diversity (standardized effect size, mean  standard error) versus null expectations (dash lines) in Vacant Lots [Vac], Urban Meadows [Mea], Low-Diversity Prairies

[Low], High-Diversity Prairies [High], and across all sites [ALL]: (A) Functional alpha diversity in 2015 and (B) 2016; and (C) Functional beta diversity in 2015 and (D) 2016. Each diversity index was compared to zero using Wilcoxon’s test and significance is denoted with asterisks when P < 0.05 ...... 41

Figure 2.4. Clustered image-map generated from cPLS. Spider response variables (Y-axis) and environmental features (X-axis) in A) 2015 and B) 2016. Positive pairwise correlations are shown in blue and negative in orange. Light color indicates correlation 0.5-0.6 and dark color

0.6-0.75. Highly correlated variables (threshold set at 0.5) are only shown here ...... 43

Figure 3.1. Frequency of taxa detected in the guts of Pardosa milvina in Control Vacant Lots

(right) and Pocket Prairies (left). Out of 171 analyzed spider guts, 113 specimens (66%) tested

xii positive for preying on Diptera (green), 68 spiders (40%) for Hemiptera (magenta), 65 (38%) for

Collembola (blue), and 51 (30%) tested positive for Coleoptera (red). Most detected taxa:

6.Leiodidae, 10.Cambaridae, 17.Chloropidae, 18.Culicidae, 19.Dolichopodidae, 30.Syrphidae,

31.Entomobryidae, 36.Cicadellidae, 41.Miridae, and 46.Formicidae. (Appendix D) ...... 62

Figure 3.2. Partial residuals plot of dietary niche overlap and mean patch size at 200 m radius.

Gray-shaded area indicates 95% confidence intervals...... 63

Figure 4.1. Hypothesized SEM model. Blue and red arrows indicate expected positive and negative correlations, respectively ...... 76

Figure 4.2. Treatments established in 2014 across 15 vacant lots in Cleveland, Ohio: A) Control

Vacant Lots (seeded fescue grass and weedy flowering plant species, mowed monthly reflecting the city’s management practices), and B) Pocket Prairies (seeded with a mixture of three native grasses and 22 flowering plants, mowed annually in October) ...... 77

Figure 4.3. Final SEM model. Blue represents a positive correlation and red indicates a negative correlation. The width of the arrows is an indicative of the strength of the relationship. Statistics available in Table 4.2 ...... 84

Figure 4.4. Partial residual plots of web spiders from our final SEM model. Grey area represents

95% confidence intervals; statistics are available in Table 4.2 ...... 85

Figure 4.5. Partial residual plots of hunting spiders from our final SEM model. Grey area represents 95% confidence intervals; statistics are available in Table 4.2 ...... 86

Figure 4.6. Partial residual plots of plant Biomass and percent of Clay in the soil, and plant biomass in Control Vacant Lots and Pocket Prairies. Grey area represents 95% confidence intervals; statistics are available in Table 4.2 ...... 86

xiii

CHAPTER 1

Patch and landscape-scale variables influence the taxonomic and functional composition of

beetles in urban greenspaces

Published in: Ecosphere 8(11):e02007 https://doi.org/10.1002/ecs2.2007

1. Abstract

Urbanization is a leading cause of species extinction, however, interest in urban greenspaces as sites for conservation has grown considerably in recent decades, raising questions about the ability of these habitats to support desired wildlife. Our goal was to determine how distinct forms of urban greenspace and their landscape context influenced the species composition of ground- dwelling beetle communities. We examined the taxonomic and functional composition of

Carabidae and Staphylinidae in urban vacant lots, urban farms, and planted urban prairies within the city of Cleveland, Ohio. Beetles were collected using pitfall traps across 23 sites. We found that the three habitats examined varied significantly in beetle composition, with several unique species found within each type of greenspace. Carabidae abundance and richness were greater in urban prairies and urban farms relative to vacant lots. The abundance and taxonomic richness of

Staphylinidae was highest within urban farms. Canonical correspondence analysis (CCA) and partial CCA revealed that both local features and landscape variables influenced beetle community assembly. Most beetle taxa were negatively associated with buildings within the surrounding 1 km landscape, whereas grass was the most important local habitat feature.

Additionally, we found variation in the distribution of species traits among habitats; ecological traits such as moisture tolerance and dispersal capacity differed significantly among urban greenspaces. Most interesting from a conservation perspective was a greater abundance of brachypterous carabids found in urban prairies, which suggests that these habitats provide overwintering and breeding habitat for some beetle species. Our findings demonstrate that urban greenspaces play important roles in shaping arthropod diversity in cities, and maintaining habitats that vary in design and management is important for conservation and the provisioning of desired ecosystem services.

2. Introduction

Urbanization is recognized as a leading cause of species extinction, known to create homogeneous communities dominated by non-native, generalist, and urban-adapted species

(Wilcove et al. 1998, Mckinney 2002, Shochat et al. 2010, Pickett et al. 2011, Knop 2016).

Nevertheless, evidence that some urban habitats can also offer resources important for conservation is mounting (Matteson et al. 2008, McKinney 2008, Bang and Faeth 2011, Beninde et al. 2015), due in part to the growing expanse of available greenspace in many cities (Gardiner et al. 2013). In the last 60 years, over 450 cities worldwide have experienced significant population decline including several cities in the US (Rieniets 2009). The legacy of these changes occurring within “shrinking cities” is an overabundance of infrastructure, which is demolished over time resulting in large tracts of vacant land (Blanco et al. 2009, Herrmann et al.

2016). Within many of these cities, greening efforts to convert vacant parcels to diverse habitats such as urban farms, rain gardens, and native wildflower plantings or “urban prairies” are underway (Gardiner et al. 2013, Park and Ciorici 2013, Chaffin et al. 2016). Vacant lots and 2 other forms of greenspace established on these formerly-developed patches have been highlighted as potentially important for conservation as they support a high richness and abundance of beneficial taxa that provide desired ecosystem services (Moorhead and Philpott

2013, Gardiner et al. 2014, Burkman and Gardiner 2015, Deák et al. 2016), as well as several rare and threatened species (Small et al. 2002, Eyre et al. 2003, Kamp et al. 2015). Given the potential of urban greenspaces to contribute to conservation, it is critical to understand if detected patterns of species occurrence shift with a changing urban habitat mosaic, and what impacts these changes might have on ecosystem functioning and desired ecosystem services.

The scale at which habitat change within inner-city landscapes influences distinct taxonomic groups varies (McIntyre 2000, Pickett et al. 2011). For example, at regional scales, urbanization gradients demonstrate that vertebrate species richness generally decreases with human activity as these species typically require larger and connected habitat patches than are available within a highly developed urban landscape (Mahan and O’Connell 2005, Parris 2006). Conversely, arthropod abundance and species richness has been shown to both decrease (Denys and Schmidt

1998, Deguines et al. 2016) and increase (Horváth et al. 2012, Lowenstein et al. 2014) with human population density and the proportion of impervious surface in the surrounding landscape. Following the beta diversity hypothesis, patchy greenspace fragments contribute to increased landscape heterogeneity which can lead to more dissimilar biological communities

(Tscharntke et al. 2002, 2012). The opposing responses to landscape fragmentation detected for arthropods is likely driven in part by the differential responses of species known as urban exploiters, which are highly dependent on human-subsidized resources such as some pollinators and many herbivorous pests, versus urban avoiders that are sensitive to human disturbance such 3 as many brachypterous species and aquatic insects (McIntyre 2000, Mckinney 2002, Kark et al.

2007). Therefore, any shift in species composition related to landscape-scale changes can influence the conservation value of focal patches and should be documented. Further, not all patches within a fragmented landscape are ecologically equivalent; local habitat features such as vegetation density and biomass have demonstrated to structure arthropod communities (Philpott et al. 2014, Eckert et al. 2017). Therefore, a multiscale approach that integrates both local and landscape heterogeneity is needed to elucidate mechanisms driving arthropod community assembly in cities.

Historically, the conservation implications of local habitats and landscape changes have been quantified taxonomically, yet a functional approach can aid in recognizing the drivers that impact community assembly in urban greenspaces (Díaz and Cabido 2001, Shochat et al. 2006, Flynn et al. 2011, Cadotte et al. 2011, Gagic et al. 2015). Consequently, several ecological frameworks have been proposed to quantify shifts in functional traits among species groups (Poff et al. 2006,

Violle et al. 2007, Suding et al. 2008), including local and landscape filters (Poff 1997). These spatial filters are demonstrated to be important in determining how functional groups influence ecological processes at the individual, population, community, and ecosystem level (Violle et al.

2007, Gagic et al. 2015). Importantly, species-based and trait-based approaches can sometimes show contradicting patterns (e.g. overall species richness increases, whereas a species group with a particular trait decreases along the same gradient), hence impacting future conservation and management decisions (Gobbi and Fontaneto 2008, Magura et al. 2010). For instance, although species richness often decreases in proximity to the city, some functional groups of butterflies, beetles, and bees have been shown to thrive in urban greenspaces (Lizée et al. 2012, Magura et 4 al. 2013, Normandin et al. 2017). With greater study, documenting the distribution of functional traits within a particular arthropod community can be used to predict the conservation value of focal habitats as well as potential desired insect-mediated ecosystem services such as biological control, pollination, and nutrient cycling (Mulder and Vonk 2011, Vonk et al. 2013, Rusch et al.

2015, Jonsson et al. 2017, Normandin et al. 2017).

Herein, we examined how small-scale shifts in urban habitat patch structure as well as landscape context influenced the taxonomic and functional trait distributions of ground-dwelling beetles.

We compared beetle assemblages among urban vacant lots, urban farms, and urban prairies within the shrinking city of Cleveland, Ohio. We collected ground beetles (Carabidae) and rove beetles (Staphylinidae) as they are considered biological indicators of habitat change (Bohac

1999, Rainio and Niemelä 2003, Beck et al. 2013), their distribution patterns vary across natural and urban gradients (Deichsel 2006, Tóthmérész et al. 2011, Work et al. 2013, Magura et al.

2013, Vergnes et al. 2014), and some function as biological control agents (Coaker and Williams

1963, Hatteland et al. 2010). Our research objective was to answer the following questions: (1)

Does the conversion of vacant land to alternative forms of greenspace influence beetle abundance and species richness? (2) Does beetle community composition vary across urban greenspaces? (3) What are the effects of local habitat vegetation characteristics and landscape composition on beetle communities? (4) Are beetle functional groups distributed equally across urban greenspaces? We hypothesized that the species composition of carabid and staphylinid communities would vary among greenspaces, driven by habitat affinity, and both local vegetation and landscape-scale variables.

3. Methods

3.1. Study area

Our study took place in Cleveland, Ohio. Once the seventh largest city in the United States,

Cleveland has experienced a 57% population decline in the last 60 years from 914,808 people in

1950 to 385,809 in 2016 (U.S. Census Bureau 2017). Protracted economic decline and population loss has resulted in over 27,000 vacant lots throughout the city (Western Reserve

Land Conservancy 2018). We collected data from eight vacant lots, eight urban farms, and seven planted urban native grass and wildflower habitats, referred to herein as urban prairies; see

Burkman and Gardiner (2015) for a map of study site locations. Habitat composition and management differed among urban greenspaces. Vacant lots are sites wherein foreclosed or abandoned residential properties were demolished, all housing debris was removed, and a fescue grass seed-mix was established. Vacant lot plant communities consisted of these grasses and a diversity of common early-successional weedy forb species and were bordered by trees and shrubs. Vacant lots were mown monthly during the growing season (April-October) by the City of Cleveland Land Bank. Urban farm sites, established on former vacant lots between 1996-

2010, were managed by the Cleveland Botanical Gardens Green Corps Program, and underwent frequent agricultural activity during the growing season. Urban prairies were managed by the

Cleveland Metroparks (six sites) and Case Western Reserve University (one site) and were only mowed or burned once a year early in the spring. Pesticides, including herbicides and insecticides, were not applied to any site during our study. Sites were separated by a minimum of

0.5 km and varied in size: vacant lots averaged 2,857 m2 (±767 m2), urban farms were 1,736 m2

(±491 m2), and urban prairies were 6,833 m2 (±1,615 m2). Site area was measured using Google

Earth Pro, and boundaries were defined by property parcels, fences, and tree lines. 6

3.2. Beetle sampling, identification, and classification

We collected Carabidae and Staphylinidae using pitfall traps that were made of 1 L plastic cups

(11.5 cm in diameter by 14.0 cm depth) filled halfway with water and a small amount of dish soap (Dawn Ultra, original scent). Each site was divided into four plots, and one pitfall trap was randomly placed within each plot while avoiding the site edges by a minimum of 10 m. Pitfall traps were set monthly in June 14-16, July 12-14, and August 16-18 in 2011; and June 11-13,

July 9-11, and August 7-9 in 2012. Pitfalls were deployed for seven consecutive days, after which their contents were stored in 75% ethanol until identification. Specimens were identified to the finest taxonomic resolution possible in North America: Carabidae were identified to species (Lindroth 1961) and Staphylinidae to genus (Arnett and Thomas 2000, Brunke et al.

2011). Analyses on species traits required a taxonomic group to be well described, and the habitat affinity information needed for our analyses were available for Carabidae species only

(Larochelle and Larivière 2003). Following Larochelle and Lariviere (2003), we assigned each

Carabidae specimen to the following trait categories: native (North America), non-native, xerophilous (found in dry environments), hygrophilous (found in moist/wet environments), open- habitat preference, shade-habitat preference, macropterous (large wings), and brachypterous

(reduced wings). We also recorded an average body size (mm) for each species (Lindroth 1961).

3.3. Local and landscape variables

We measured vegetation within a 1 m2 quadrat placed adjacent to each pitfall trap (a total of four per site) to investigate the effects of local vegetation on beetle assemblages in 2012. Using estimates of 5% increments, we recorded percent cover of grass, bare ground, litter, and forbs

(flowering and non-flowering). We also obtained plant average height measured from three 7 random points within each quadrat. To investigate the effects of landscape composition on beetle assemblages, we quantified the landscape surrounding each site using remotely-sensed images at

1 m resolution (Zhou and Troy 2008, Blaschke 2010, de Pinho et al. 2012). We used eCognition

Developer 8.7 and ArcGIS 10.1 for land cover identification and landscape composition analysis respectively. We classified land cover into four categories following Cadenasso et al. (2007): buildings, pavement/bare-soil, coarse vegetation (trees and woody shrubs), and fine vegetation

(grasses and low-growing forbs). Following previous studies that evaluated ground-dwelling beetles across the landscape (Sadler et al. 2006, Gardiner et al. 2010), circular buffers at 100 m,

250 m, 500 m, 750 m, and 1 km radius were created for each site.

3.4. Data analysis

Analyses were performed in R version 3.2.5 (R Core Team 2019). First, to compare the richness and abundance of carabids and staphylinids among urban greenspaces, we fit generalized linear mixed-effects models with Habitat (vacant lot, urban farm, urban prairie) as a fixed effect and

Year (2011 and 2012) as a random effect. Second, to compare the species composition of beetle communities across habitats, we performed a permutational multivariate analysis of variance using distance matrices using Bray-Curtis dissimilarity index and 999 permutations. This was carried out using the ‘adonis’ function in the ‘vegan’ package (Oksanen et al. 2018). Third, to better understand what environmental features influenced carabids and staphylinids among habitats, we followed a multi-scale approach using the ‘vegan’ package (Oksanen et al. 2018).

We performed canonical correspondence analysis (CCA) and partial CCA incorporating beetles along with local vegetation and landscape composition data collected in 2012 (ter Braak 1986,

Borcard et al. 1992). We checked the species distribution via detrended correspondence analysis 8 using the ‘decorana’ function in the ‘vegan’ package (Oksanen et al. 2018). The length of the gradient exceeded 3 in standard deviation units indicating that our species matrix followed a unimodal gradient. We included beetles that were trapped in more than five sites; consequently, two prairie sites were removed as they lacked all 12 most-trapped species. We fitted six CCA models: local vegetation and landscape composition at 100 m, 250 m, 500 m, 750 m, and 1 km.

Next, we followed a forward and backward selection comparing each model with all variables included (Beetles ~ .) against an intercept-model (Beetles ~ 1) to eliminate the variables that explained no significant variation. We then performed a model selection procedure based on the percent of the inertia. We also performed a partial CCA on the selected models to partial out the covariate effect of local features on landscape variables and vice-versa (Borcard et al. 1992). We tested for multicollinearity in each model and examined all models using the functions ‘vif.cca’ and ‘anova.cca’ in the ‘vegan’ package (Oksanen et al. 2018). Lastly, to assess what morphological and ecological characteristics were influencing community composition, we followed a functional trait approach. We compared the abundance and richness of xerophilous, hygrophilous, open-habitat preference, macropterous, and brachypterous species across urban greenspaces. Shade-habitat species richness was too low for proper statistical analysis; hence we only examined the abundance of this group. Body size was examined using community weighted

푆 mean 퐶푊푀 = ∑푖=1 푝푖푥푖 where pi is the relative abundance of the i-th species and xi is the trait value (body size) of the i-th species (Garnier et al. 2004). Species traits were examined by fitting generalized linear mixed-effects models with Habitat (vacant lot, urban farm, urban prairie) as a fixed effect and Year (2011 and 2012) as a random effect. Models were fitted using Poisson and negative binomial distributions based on their diagnostic plots and using the ‘fitdistr’ function in

9 the ‘fitdistrplus’ package (Delignette-Muller and Dutang 2015). Analysis of deviance tables were generated using the ‘Anova’ function in the ‘car’ package (Fox and Weisberg 2019), and pairwise comparisons were performed using Tukey’s range test using the ‘ghlt’ function in the

‘multcomp’ package (Hothorn et al. 2008).

4. Results

4.1. Abundance and richness

We collected 1,497 staphylinids (43 genera) and 316 carabids (44 species) for a total of 1,813 beetles (Table 1.1). Carabidae abundance (ANOVA, X2 = 31.36; df = 2,46; p < 0.01; Figure

1.1A) and richness (ANOVA, X2 = 20.18; df = 2,46; p < 0.01; Figure 1.1B) differed among urban greenspaces with prairies and farms supporting a greater abundance and richness of ground beetles than vacant lots. For staphylinids, abundance was significantly higher in urban farms than vacant lots and prairies (ANOVA, X2 = 30.76; df = 2,46; p < 0.01; Figure 1.1C). Staphylinidae richness was significantly higher in urban farms when compared with vacant lots, while prairies had an intermediate richness that was not significantly different from the other habitats

(ANOVA, X2 = 18.90; df = 2,46; p < 0.01; Figure 1.1D).

Table 1.1. Total number of Carabidae and Staphylinidae beetles collected in vacant lots (Vac), urban farms (Far), and urban prairies (Pra) in Cleveland, Ohio in June-August 2011 and 2012.

Carabidae Staphylinidae Species Vac Far Pra Origin Genus Vac Far Pra Agonoleptus conjunctus 0 2 0 us can Acrotona 1 14 6 Agonum muelleri 0 0 4 eurasia Aleochara 7 29 83 Agonum nutans 0 0 3 us can Amischa 0 1 1 Agonum punctiforme 0 12 17 us can Anotylus 235 641 75 Agonum rufipes 0 0 1 us can Apocellus 3 92 26 Amara aenea 1 17 1 eurasia Autalia 0 0 1 Amara convexa 0 1 0 us can Batriasymmodes 1 0 0 Amara cupreolata 0 4 0 us can Belonuchus 0 1 2 Amara familiaris 0 1 0 eurasia Coproporus 1 2 4 Amara impuncticollis 0 3 1 us can Cordalia 2 6 0 Amara littoralis 0 1 1 us can Datomicra 1 1 0 Anisodactylus harrisii 1 2 0 us can Dinaraea 2 7 2 Anisodactylus nigerrimus 0 1 0 us can Falagria 0 2 0 Anisodactylus ovularis 0 1 0 us can Gauropterus 0 1 0 Anisodactylus rusticus 4 1 0 us can Gyrohypnus 1 5 0 Anisodactylus sanctaecrucis 0 2 0 us can Heterothops 0 1 0 Bembidion quadrimaculatum 1 1 0 us can Hoplandria 0 2 1 Bembidion viridicolle 1 0 0 us can Hydrosmecta 0 1 0 Bradycellus rupestris 0 1 0 us can Hypnogyra 0 4 0 Bradycellus tantillus 0 0 1 us can Lithocharis 1 2 0 Calathus opaculus 0 0 2 us can Medon 0 3 0 Chlaenius nemoralis 4 2 4 us can Meronera 10 0 1 Chlaenius tricolor 4 7 2 us can Myrmedonota 7 0 0 Cicindela punctulata 2 0 2 us can Oxybleptes 0 1 0 Cyclotrachelus gravesi 0 0 3 us can Oxypoda 0 0 2 Dicaelus elongatus 0 0 2 us can Paederus 0 0 2 Diplocheila obtusa 2 0 3 us can Philonthus 6 12 23 Elaphropus anceps 0 13 0 us can Pinophilus 0 0 1 Elaphropus vernicatus 1 2 0 us can Platydracus 12 10 38 Elaphropus xanthopus 0 0 27 us can Platystethus 0 3 0 Harpalus affinis 0 3 0 eurasia Quedius 0 0 4 Harpalus erraticus 1 0 0 us can Sepedophilus 3 1 2 Harpalus paratus 1 0 0 us can Stamnoderus 0 0 2

Harpalus pensylvanicus 7 12 1 us can Stenus 0 0 2 Notiophilus semistriatus 0 0 2 us can Stethusa 4 6 21 Ophonus puncticeps 0 7 4 europe Strigota 8 15 4 Poecilus chalcites 0 0 8 us can Tachinus 0 0 2 Poecilus lucublandus 0 0 9 us can Tachyporus 0 1 2 Pterostichus atratus 0 0 17 us can Tasgius 1 0 0 Pterostichus permundus 0 0 25 us can Tinotus 0 5 1 Pterostichus stygicus 0 0 2 us can Trichiusa 0 6 0 Scarites subterraneus 1 22 3 us can Xantholinus 0 1 0 Stenolophus ochropezus 0 0 1 us can Xestolinus 0 0 1 Stenolophus spretus 0 2 0 us can Unknown staphs 1 3 2 Unknown carabids 6 4 9 Total 37 124 155 Total 307 879 311 # of species 14 24 26 # of genera 19 30 26

Figure 1.1. Mean (SE) Carabidae abundance (A) and richness (B), and mean (SE) Staphylinidae abundance (C) and richness (D) in vacant lots, urban farms, and urban prairies within Cleveland, Ohio. Specimens collected in pitfall traps from 7 d sampling periods monthly in June-August 2011 and 2012. Letters indicate significant differences (p < 0.05) using Tukey’s range test.

4.2. Species composition

Carabidae and Staphylinidae communities sampled in both years differed in species composition among habitats with 24 unique taxa (46%) in urban prairies (15 carabids and 9 staphylinids), followed by 21 taxa (39%) in urban farms (11 carabids and 10 staphylinids), and 6 (18%) in vacant lots (3 carabids and 3 staphylinids, Figure 1.2). We collected five non-native carabid species, which represented 11% of the overall richness and 12% of the total beetle catch (Table

1.1). Permutational multivariate analysis of variance using distance matrices for both sampling years showed that ground-dwelling beetle assemblages were significantly different among vacant lots, farms, and prairies (ADONIS, F = 3.59; p < 0.01).

4.3. Ordination analysis

Following a percent of inertia model procedure, we found that the two best-fit CCA models were local vegetation and landscape cover at 1 km explaining most of the variation in species distribution across greenspaces (Table 1.2). Our stepwise variable selection resulted in local vegetation retaining the variables grass percent cover and plant height, whereas landscape cover at 1 km retained the variables buildings and pavement (Table 1.2). The first axis of the local vegetation ordination explained 20.9% of the species variance (ANOVA, F = 4.46; p < 0.01), whereas the second axis explained an additional 15.4% (ANOVA, F = 1.45; p = 0.08). In our landscape ordination, the first and second axes explained 23.8% (ANOVA, F = 5.61; p < 0.01) and 14.1% (ANOVA, F = 1.30; p = 0.14) of the variance, respectively. Our partial CCA showed that the effect exercised only by local vegetation variables (after removing the landscape effect) explained 10.7% of the dispersion of beetles in our plot, whereas the effect exercised only by landscape variables (after removing local effects) explained 13.7% of the distribution of beetles

(Table 1.2). The CCA tri-plot shows the relationship among sites, beetles, and environmental variables (Table 1.2, Figure 1.3). The first axis explained 22.2% of the species variance

(ANOVA, F = 6.83; p < 0.01), whereas the second axis explained an additional 11.1%

(ANOVA, F = 2.07; p = 0.01). Beetles found predominantly in urban farm sites, including

Anotylus, Apocellus, Amara aenea, and Scarites subterraneus, were not strongly associated with local vegetation or landscape predictors. Philontus and Meronera species were found to be most common in urban prairies and vacant lots, respectively. Additionally, Aleochara and Stethusa species were positively associated with grass cover and were predominantly found in vacant lots and prairie habitats (Figure 1.3).

Figure 1.3. CCA tri-plot for beetle taxa collected in at least five sites from June-August 2012. Each square, circle, or triangle represents a given site. The proximity of sites to one another indicates the relative similarity of their beetle communities. The proximity of sites to beetle taxa indicates the relative abundance of a given taxonomic group within a site. Environmental variables are shown as vectors along two CCA axes. The proximity of sites and species to these vectors suggest a positive, neutral, or negative correlation with a given environmental variable.

CCA model Variables % Variation P-value Local vegetation Grass + Height 24.7 < 0.01 Landscape 1 km Buildings+Pavement 27.8 < 0.01 Landscape 750 m None N/A N/A Landscape 500 m None N/A N/A Landscape 250 m None N/A N/A Landscape 100 m Buildings 13.4 0.03

% Variation Partial CCA model Constrained Conditioned P-value Local vegetation 10.7 27.8 0.07 Landscape 1 km 13.7 24.7 < 0.05

Correlation scores CCA tri-plot Axis 1 Axis 2 P-value Buildings -0.8 13.5 < 0.01 Pavement 0.3 7.5 0.05 Grass 0.7 -0.6 0.08 Height 0.5 5.5 0.65

4.4. Species traits

For ecological traits, xerophilous ground beetle abundance (ANOVA, X2 = 11.21; df = 2,23; p <

0.01; Figure 1.4A) and richness (ANOVA, X2 = 11.43; df = 2,23; p < 0.01; Figure 1.4B) varied across habitats with urban farms supporting more than twice the abundance as well as a greater taxonomic richness than vacant lots, while urban prairies did not differ from the other habitats.

Hygrophilous ground beetle abundance also differed among habitats (ANOVA, X2 = 25.96; df =

2,23; p < 0.01; Figure 1.4C); the greatest abundance was collected from urban prairies followed by farms and vacant lots. Hygrophilous richness was significantly higher in urban prairies and farms compared to vacant lots (ANOVA, X2 = 10.77; df = 2,23; p < 0.01; Figure 1.4D). The abundance of ground beetle species with an affinity for open-habitats was higher in urban prairies and farms than vacant lots (ANOVA, X2 = 18.56; df = 2,23; p < 0.01; Figure 1.4E), whereas richness was higher in urban farms than vacant lots, yet it did not differ between urban prairies and the other two habitats (ANOVA, X2 = 16.99; df = 2,23; p < 0.01; Figure 1.4F). The abundance of shade-tolerant ground beetle species differed among greenspaces (ANOVA, X2 =

17.83; df = 2,23; p < 0.01) with urban prairies supporting a significantly higher abundance than the other two habitats (Figure 1.4G).

For morphological traits, we found that carabid body size did not differ among urban greenspaces (ANOVA, X2 = 0.68; df = 2,23; p = 0.71; Figure 1.4H). In terms of flight capabilities, macropterous abundance was significantly higher in urban farms and prairies

(ANOVA, X2 = 15.58; df = 2,23; p < 0.01; Figure 1.4I), while richness was higher in urban farms than the other habitats (ANOVA, X2 = 20.86; df = 2,23; p < 0.01; Figure 1.4J). Brachypterous

18 abundance was significantly higher in urban prairies supporting more than six times as many brachypterous carabids than urban farms or vacant lots (ANOVA, X2 = 17.61; df = 2,23; p <

0.01; Figure 1.4K). Brachypterous richness was also significantly greater in prairies compared with vacant lots; farms supported an intermediate richness of these species (ANOVA, X2 =

10.89; df = 2,23; p < 0.01; Figure 1.4L).

Figure 1.4. Carabidae traits: Mean (SE) of xerophilous abundance (A) and richness (B), hygrophilous abundance (C) and richness (D), open-habitat abundance (E) and richness (F), shade-habitat abundance (G), body size (H), macropterous abundance (I) and richness (J), and brachypterous abundance (K) and richness (L) in vacant lots, urban farms, and urban prairies in Cleveland, Ohio. Specimens collected in pitfall traps from 7 d sampling periods monthly in June- August 2011 and 2012. Letters indicate significant differences (p < 0.05) using Tukey’s test.

5. Discussion

Urbanization processes result in habitat loss, fragmentation, and degradation, which are often associated with reductions in biodiversity and significant shifts in species composition (Wilcove et al. 1998, Mckinney 2002, Kark et al. 2007, Knop 2016). These processes are known to favor urban adapters and exploiters, which are often non-native generalist species (Mckinney 2002,

McKinney 2006, Kark et al. 2007). Nevertheless, we found that urban greenspaces within a shrinking city can support diverse beetle communities dominated by native taxa. We collected 44 carabid species and 43 staphylinid genera within vacant lots, farms, and prairies within the city of Cleveland, Ohio. Only five carabid species were non-native, which represented 11% of the overall richness and 12% of the total abundance. We also found that these greenspaces varied widely in beetle community composition, with up to 24 unique taxa representing 46% of the community composition found within a particular habitat type.

Localized habitat quality has been shown to mitigate large-scale landscape changes (Rundlöf et al. 2008, Concepción et al. 2012). Within shrinking cities, diversification and management of vacant land patches created by deindustrialization increases landscape-scale heterogeneity, which is predicted to support beta diversity in fragmented landscapes (Tscharntke et al. 2002, 2012,

Gibb and Hochuli 2002). We found that urban farm and prairie habitats, managed to deliver a diversity of ecosystem services, supported a greater species richness and abundance of ground- dwelling beetles than vacant lots and high proportions of unique taxa. Differences in habitat quality for focal biotic communities can be highly influenced by the human-mediated environmental filters acting upon them (Swan et al. 2011). This process of “facilitated assembly”, wherein humans modify existing low-input urban greenspace, has often been found 22 to support a greater richness and abundance of several arthropod groups versus “self-assembled” vacant lots (Swan et al. 2011, Moorhead and Philpott 2013, Gardiner et al. 2014, Philpott et al.

2014, Burkman and Gardiner 2015).

We found that both local vegetation features and large-scale landscape composition influenced the quality of urban greenspaces for ground-dwelling beetles. At the landscape scale, patches surrounded by a high density of built infrastructure were negatively associated with the abundance of most beetle taxa. Connectivity among green elements within urban landscapes has shown to influence the dispersal and abundance of arthropods (Cranmer et al. 2012, Braaker et al. 2014, Carper et al. 2014). Thus, our results may indicate that when greenspace is scarcer within a landscape, source populations are scarce, and low emigration contributes to reduced population density. Among local vegetation characteristics, the taller vegetation and dense grass cover common within urban prairies were positively associated with several species. These vegetative features have shown to be important to other insects, including butterfly communities and natural enemies in semi-natural habitats (Bergman et al. 2008, Sarthou et al. 2014). Further, other unmeasured variables such as plant composition and soil characteristics may be responsible for the remaining unexplained variation in beetle richness and abundance among habitats, as these factors commonly influence beetle assemblages in urban, forest, and agricultural settings

(Jukes et al. 2001, Brose 2003, Dauber et al. 2005, Aviron et al. 2005, Woodcock et al. 2010,

Philpott et al. 2014).

Our examination of ground beetle species traits revealed how three distinct management regimes operating in vacant lots, urban farms, and urban prairies structured ground beetle communities. 23

Urban prairies received the least frequent management, with vegetation cut once annually

(versus monthly for vacant lots) allowing the establishment of a high biomass of native forbs and grasses during the growing season. This habitat supported a greater number of hygrophilous and shade-tolerant species; high plant biomass likely results in higher humidity at the soil surface and a greater proportion of shaded foraging micro-habitats within prairie sites. Urban prairies also supported a greater abundance of brachypterous beetles, which are generally considered poor dispersers (Den Boer 1970). This suggests that urban prairies may provide these beetles, and potentially other arthropods, with overwintering and breeding habitat (Thomas et al. 1992,

Eversham and Telfer 1994, Noordijk et al. 2011, Schaffers et al. 2012) and may act as a source for re-colonization of surrounding patches following disturbance (Clough et al. 2007, Brunke et al. 2014). An affinity towards patches with reduced disturbance intensity has been recorded for other brachypterous (Venn 2007, Gobbi and Fontaneto 2008, Pedley and Dolman 2014), hygrophilous (Magura et al. 2013), and habitat-specialist groups (Halme and Niemelä 1993,

Deichsel 2006, Magura et al. 2008) across urban gradients.

Urban farm sites were the most intensively managed habitat studied, with frequent weeding, irrigation, and harvesting activities occurring throughout the growing season. On average, these habitats supported a greater abundance and richness of xerophilous species, which may be able to best tolerate the variation in soil moisture present within irrigated vegetable crops and unirrigated mulched areas and bare soil present between crop rows. Urban farms also supported a higher richness of macropterous ground beetles, able to disperse from these sites during times of low prey availability or unfavorable environmental conditions. In response to high levels of human activity, dry-tolerant, heat-tolerant, and macropterous arthropods are often found in 24 highly disturbed urban environments. For example, thermophilous beetles (Magura et al. 2013,

Piano et al. 2017), xerophilous spiders (Horváth et al. 2012), and macropterous carabids (Venn

2007, Gobbi and Fontaneto 2008, Pedley and Dolman 2014) have been recorded in large numbers in city cores along urbanization gradients. Interestingly, we did not find differences in body-size among habitat types, despite studies reporting that beetle body-size decreased as urbanization, fragmentation, and management intensified (Halme and Niemelä 1993, Alaruikka et al. 2002, Magura et al. 2004). It is possible that the environmental filters influencing body size operate at larger spatial scales, detectable among gradients but not necessarily among patches within an urban area. Previous studies have found that large-size beetles commonly decrease along urbanization gradients and are more common in semi-natural forests and grassland patches outside cities (Halme and Niemelä 1993, Alaruikka et al. 2002, Magura et al. 2004, Gobbi and

Fontaneto 2008).

6. Conclusions

The expanse of vacant land is expected to increase globally over the next several decades as many cities continue to undergo protracted population loss (Rieniets 2009, Herrmann et al.

2016). Urban greenspaces have demonstrated a capacity to support greater richness and abundance, reproductive success, and even a unique species pool when compared to agricultural or natural habitats sampled outside cites (Sattler et al. 2011, Ives et al. 2016). Additionally, urban greenspaces reduce landscape-scale impervious surface, increase spatial heterogeneity, and provide key ecosystem services such as carbon stormwater infiltration, storage and sequestration, and atmospheric pollutant removal (Gardiner et al. 2013, Riley et al. 2018a). Given the vast number of vacant lots present within shrinking cities, and a low probability of redevelopment 25

(Rieniets 2009), these habitats offer a reliably-accessible resource for local citizen groups to initiate small-scale conservation efforts (Blanco et al. 2009, Green et al. 2016). Our results highlight that creation of unique urban greenspaces through these efforts, such as planted prairies and urban farms, can have measurable positive and distinct influences on species abundance and richness. This study adds to a body of literature documenting abundant and taxonomically rich communities found among these habitats (Moorhead and Philpott 2013, Gardiner et al. 2014,

Philpott et al. 2014, Burkman and Gardiner 2015). Further, our research demonstrates that habitat design can filter species based on ecological traits such as moisture tolerance and dispersal capacity. Gaining a mechanistic understanding of how management activities are likely to filter species assemblages though functional analyses such as this are key to advancing conservation planning in cities (Shochat et al. 2006, Faeth et al. 2011). For example, we found that urban prairies supported an increased abundance of dispersal-limited brachypterous species, which may mean that these patches can act as source habitats able to support overwintering and breeding habitat. Further, we found that when selecting vacant lot patches for conservation investment, the landscape surrounding the site can influence the community assemblage found.

The abundance of several beetle species within urban farm and prairie habitats was negatively associated with highly developed landscapes dominated by built infrastructure. As the ecological value of vacant land becomes realized, growth in urban conservation research and planning that incorporates habitat design, management, landscape context, and greenspace connectivity is needed (Harrison and Davies 2002, Muratet et al. 2007, Kattwinkel et al. 2011, Gardiner et al.

2013). Further, it is critical to keep in mind that such conservation efforts can only be successful when stakeholders including local municipalities, non-governmental organizations, and 26 neighborhood residents work together to define how to generate, maintain, and assess a diversity of desired services such as biodiversity conservation, habitat aesthetics, recreational opportunities, and food production from available vacant land resources (Su 2010, Dearborn and

Kark 2010). Our work aims to influence city planners to include diversity and heterogeneity in the planning, design, and development of urban greenspaces to increase biodiversity, benefit conservation, and enhance ecosystem functions.

Acknowledgements

I thank Caitlin Burkman for letting me use her samples for this project, Taro Eldredge for identifying rove beetles, Kayla Perry for teaching me Carabidae identification. This research was conducted as part of the Cleveland Urban Long-Term Research Area Exploratory (ULTRA-Ex)

Project. We also thank Kelsey Greathouse, Alec Norris, Andrea Kautz, Mike Shields, Shawn

Probst, Jared Power, Steve Ryan, and Emiko Waight who assisted with the collection and processing of samples for this project. The Cleveland Botanical Gardens and Cleveland Land

Bank provided access to urban farms and vacant lot sites. Access to prairie sites was through the

Cleveland Metroparks and Case Western Reserve University. This work was supported in part by

Ohio State University fellowships awarded to C. E. Burkman, the National Science Foundation

CAREER DEB Ecosystem Studies Program (CAREER-1253197) to Mary Gardiner, and the

National Science Foundation Graduate Research Fellowship Program (DGE-1343012) to Yvan

Delgado de la flor.

CHAPTER 2

Local and landscape-scale environmental filters drive the functional diversity and

taxonomic composition of spiders across urban greenspaces

Published in: Journal of Apply Ecology https://doi.org/10.1111/1365-2664.13636

1. Abstract

Urban patch colonization and species establishment within cities are restricted by the behavioral, life history, and physiological attributes of colonizing species, in conjunction with environmental filtering processes at small and large spatial scales. To enhance local biodiversity in urban greenspaces, these filtering processes need to be assessed so that greenspace design and management can guide establishment of local species pools. We investigated the relative importance of local and landscape-scale features on spider community assembly using a functional and taxonomic approach. Within the shrinking city of Cleveland Ohio, we established a field experiment wherein control vacant lots, urban meadows, and low- and high-diversity pocket prairies were established across eight neighborhoods (N = 32). Spiders were sampled monthly during June-August 2015 and 2016 using pitfall traps and vacuums. Spider functional diversity was assessed using null models, while local and landscape drivers were analyzed via canonical partial least squares and clustered image maps. Increased mowing strongly influenced spider communities leading to lower-than-expected spider functional alpha and beta diversity in

2015. Patch isolation and percentage impervious surface increased the functional dissimilarity

28 and taxonomic diversity of spiders in 2016, resulting in higher-than-expected overall functional alpha diversity. We also found that increasing plant height and biomass favored spiders with large body-size and decreased the abundance of small web-weavers. Our findings suggest that impervious surface is a strong environmental filter that influences the colonization and establishment of spider communities in cities. Additionally, while periodic mowing in vacant lots benefits some spider taxa, it has a negative impact on the establishment of several species, mainly larger spiders and those most sensitive to disturbance. To conserve spiders and their dependent biota , investment in managed greenspaces such as pocket prairies that require infrequent mowing is warranted. In so doing, cities can enhance urban biodiversity and beautify local neighborhoods.

2. Introduction

Wildlife conservation research in cities has received considerable attention in the last decade due to the number of studies reporting that urban greenspaces can support diverse local communities, including rare and endangered species (Mckinney 2002, Ives et al. 2016). Invertebrates have been a major focus of this research, and speciose arthropod assemblages have been reported in urban areas worldwide (McIntyre 2000, Jones and Leather 2012). In some cities, greenspaces are increasing due to local conservation efforts or newly available parcels resulting from prolonged economic and population decline (Blanco et al. 2009, Gardiner et al. 2013). However, managing this influx of green infrastructure to support biodiversity is challenging as species have different resource requirements. To achieve the conservation potential of urban greenspaces, a multi-scale spatial approach is needed to elucidate the hierarchical filters involved in the distribution and establishment of species within cities (Pickett et al. 2011, Aronson et al. 2016). 29

There is substantial evidence demonstrating that environmental filters (local habitat characteristics and landscape composition and configuration) influence the species composition and spatial distribution of arthropod communities in cities (Tscharntke et al. 2012, Aronson et al.

2016). At the local scale, insects have been associated with plant diversity and vegetation complexity (Randlkofer et al. 2010, Eckert et al. 2017), whereas landscape-scale features such as the amount of impervious surface and greenspace connectivity have been implicated as important environmental predictors (Davis et al. 2007, Braaker et al. 2017), especially for species with limited dispersal capabilities (Sivakoff et al. 2018). To date, many studies have described how environmental filters shape local species pools, yet it is still challenging to determine the spatial scale responsible for assembling species communities, as the relative importance of local and landscape drivers varies by taxa (McIntyre 2000, Jones and Leather 2012, Lepczyk et al. 2017).

This has led to the application of ecological theory to identify the mechanisms driving arthropod communities within cities (Tscharntke et al. 2012, Fattorini et al. 2018). Derived from the “single large or several small” dilemma in ecology, one hypothesis suggests that fragmented habitats increase biodiversity due to species turnover across several small patches (Tscharntke et al.

2012, Fahrig 2017). This hypothesis has been criticized due to the well-documented detrimental effects of fragmentation on local species (Pfeifer et al. 2017, Fletcher et al. 2018), yet to draw definitive conclusions, more research is needed examining the response of arthropods to habitat fragmentation (Gibb and Hochuli 2002, Fattorini et al. 2018), especially within cities.

To better understand how environmental filters affect species assemblages in urban ecosystems, we need to shift our focus from biodiversity assessments to the mechanistic processes behind these patterns (Cadotte et al. 2011, Schwarz et al. 2017). Despite the important contributions of 30 taxonomic diversity (e.g. abundance and richness) to conservation biology, functional diversity, wherein ecologists group species by traits and quantify the diversity of traits and functions, has shown to more accurately link biodiversity and ecosystem processes at different spatial scales

(Díaz and Cabido 2001, Cadotte et al. 2011, Gagic et al. 2015). Trait-based analyses have elucidated the mechanisms by which functional groups respond to habitat alteration, changes in temperature, and urban greenspace management (Venn 2007, Magura et al. 2013, Delgado de la flor et al. 2017). Additionally, trait-based approaches are particularly useful in monitoring arthropod communities because practitioners can target specific functional groups and achieve desired management or conservation goals. For example, studies have shown that urban greenspace management is a determinant of colonization and establishment of beetle and spider functional groups (Braaker et al. 2017, Delgado de la flor et al. 2017).

Consequently, we employed both a functional and taxonomic approach to investigate the local and landscape drivers of spider community assembly. To do so, we established the Cleveland

Pocket Prairie Project, a large-scale field experiment where four habitat treatments were replicated on individual vacant lots across eight inner-city neighborhoods (Figure 1). Cleveland,

Ohio has experienced steady human population decline and infrastructure loss over the last 60 years resulting in 27,000+ vacant lots spread across the city (Western Reserve Land Conservancy

2018). These reclaimed greenspaces represent an opportunity to examine the impact of urban greenspace management on biodiversity across a large landscape scale (Blanco et al. 2009,

Gardiner et al. 2013). We focused on spiders as they are among the most abundant and diverse predatory groups in urban ecosystems and are considered biological indicators of habitat alteration due to their sensitivity to environmental changes (Langellotto and Denno 2004, 31

Sarthou et al. 2014). Within an urban landscape, a single patch may not have the capacity to sustain spider communities long-term, nonetheless their spatial configuration of may enable greenspaces to act as stepping stones, facilitating interactions among distant populations

(Colding 2007, Lepczyk et al. 2017). Following the premise that heterogenous, fragmented landscapes determine biodiversity patterns (see Hypothesis 2 in Tscharntke et al. 2012), we hypothesized that landscape-scale features would drive the functional dissimilarity of spider communities and determine the functional diversity of spiders occupying a patch, overriding local habitat characteristics. We predicted that spider functional beta diversity (dissimilarity) would increase with impervious surface, as variation in dispersal capabilities and adaptation to disturbance are demonstrated to facilitate spider community colonization (Braaker et al. 2017,

Argañaraz et al. 2018). We also predicted that spider functional alpha diversity would be higher than expected from random, resulting from variation in successful colonization and establishment of spider functional groups within patches that vary in vegetation community composition, management and landscape context (Cattin et al. 2003, Lowe et al. 2018).

3. Methods

3.1. Study area

Our study was conducted in the city of Cleveland, Ohio, USA. Since the 1950s, Cleveland has experienced prolonged population decline and infrastructure loss, leading to the overabundance of vacant lots, that are seeded with turf grass and mown monthly by the city. In 2014, we established the Cleveland Pocket Prairie Project across eight inner-city neighborhoods and selected 32 vacant lots (15 x 30 m on average) wherein four experimental treatments were established (Figure 2.1 & 2.2): Control Vacant Lots, Urban Meadows, Low-Diversity Pocket 32

Prairies and High-Diversity Pocket Prairies (Figure 2.2, Appendix A). Pocket prairies were mowed to a height of 20 cm monthly in 2015 to reduce weed competition and facilitate the establishment of flowering species. All data collection occurred within a 7 x 15 m grid of 105 quadrats (1 m2 each), placed within the center of each site.

Figure 2.1. Map of Cleveland, Ohio showing our 32 experimental sites where spiders were sampled in 2015 and 2016. (©Denisha Parker)

Figure 2.2. Four treatments were established in 2014 across 32 vacant lots in Cleveland, Ohio: A) Control Vacant Lots (seeded fescue grass and weedy flowering plant species, mowed monthly in 2015 & 2016 reflecting the city’s management practices), B) Urban Meadows (seeded fescue grass and weedy flowering plans, mowed annually in October 2015 & 2016), C) Low-Diversity Prairies and D) High-Diversity Prairies (mowed monthly in 2015 & annually in October 2016). Prairies were seeded with a mixture of native grasses and flowing plants (Appendix A).

3.2. Spider sampling and identification

To sample both active and less-mobile spiders, specimens were collected in each site using four pitfall traps and four vacuum samples three times per year in 2015 (12-22 June, 8-20 July, 11-18

August), and in 2016 (1-9 June, 6-14 July, 3-11 August). Within each site, four quadrats were randomly selected, and pitfall traps were active for seven consecutive days. Pitfall traps consisted of 1 L plastic cups (12 cm diameter x 14 cm depth) filled halfway with water containing a small amount of dish soap (Dawn® Ultra, original scent). While pitfall traps were active, we vacuumed an area of 0.25 m2 (30-50 cm away from each trap in any direction) for 45 seconds using a modified leaf vacuum (12 cm diameter). Specimens were stored in 80% ethanol and transported to the laboratory for sorting and identification. Due to weather and issues with landscaping contractors, vacuum sampling did not occur in July 2015.

Lycosidae and Linyphiidae adult spiders were identified to species and other adult and sub-adult spiders were identified to genus. We used several identification resources including Spiders of

North America: An Identification Manual (Ubick et al. 2017), Guide d’identification des

Araignées (Araneae) du Québec (Paquin and Dupérré 2003), and taxon-specific keys available at the World Spider Catalog (2018). Spider functional traits and groups (Appendix B) were classified following Cardoso et al. (2011). Functional traits comprised foraging activity (web type or hunting method), prey range (stenophagous or euryphagous), vertical stratification

(ground or vegetation), circadian activity (diurnal or nocturnal), and mean body size measured as the community-weighted mean (Uetz et al. 1999, Cardoso et al. 2011). Voucher specimens were deposited in the Museum of Biological Diversity at The Ohio State University.

3.3. Local habitat variables and landscape cover

Local habitat variables were measured within 20 randomly selected quadrats. Using a 0.5 m2 sub-quadrat, vegetation was sampled twice in 2015 (16 June - 3 July and 22 July - 13 August) and three times in 2016 (13-24 June, 11-22 July, and 4-16 August). We recorded the three most dominant plant species per quadrat and diversity was calculated per site using the Shannon-

Wiener index (Shannon 1948). Plant biomass was estimated with a comparative-yield method and the dry-weight-rank method from the 20 selected quadrats (Mannetje and Haydock 1963,

Haydock and Shaw 1975). First, five quadrats were ranked (1 = lowest biomass density, 5 = highest biomass density, and 2-4 in between) and established as the ‘standard yields’ reflecting the range of biomass within each site. In each of our 20 quadrats, we estimated the biomass yield, on a scale of 1 to 5, in comparison to our five standard yields. Only the five standard yields were harvested, oven-dried at 75 C for 48 hours and weighted. Finally, we plotted our standard yields, obtained an equation from the trendline, and inserted our 20 ranked comparative yields from each site into this equation to estimate biomass in each quadrat. Average site-wide biomass was then calculated as the mean of these 20 comparative yield estimates.

Mean bloom abundance, bloom area, and plant height were also calculated at each site from an additional six randomly selected quadrats. Average plant height was derived from three height measurements (cm) taken in each quadrat. Likewise, bloom abundances were counted per each flowering species present in the six quadrats. For each flowering species present, we recorded five bloom area measurements (mm2) and then multiplied the average bloom size by the number of blooms present at a site to derive an average bloom area. Additionally, twenty soil cores were

36 randomly sampled and pooled per site in April 2014 to measure the concentration of heavy metals. The Contamination Factor (Loska et al. 2004) of aluminum, antimony, arsenic, barium, cadmium, chromium, cobalt, copper, iron, lead, manganese, nickel, vanadium, and zinc was calculated using regional background levels from eastern United States (US EPA 2007), and from these values Pollution Load Indices were calculated per site (Tomlinson et al. 1980,

Weissmannová and Pavlovský 2017).

Landscape information was obtained from the Cuyahoga County Planning Commission using remotely sensed images at 1-2 m resolution, captured in 2011. Following previous studies that reported spider communities patterns across the landscape (Gardiner et al. 2010, Philpott et al.

2014), we selected buffer zones at 200 m and 1500 m radii from each site. Landscape cover was classified into percentage: grass/shrubs, bare soil, water, buildings, roads/railroads, other paved surfaces, tree canopy (TC) over vegetation, TC over buildings, TC over roads/railroads, and TC over other paved surfaces. We limited the landscape covers to those directly affecting ground- dwelling spiders. To assess landscape composition, percentage grass/shrubs, percentage buildings, percentage flat impervious surface (roads/railroads & other paved), and Shannon landscape diversity were included in the analysis. For landscape configuration, we re-classified our categories into either ‘greenspace’ (grass/shrubs & TC over vegetation) or ‘other’ based on the importance of patch connectivity on our spider functional groups (Bonte et al. 2004, Braaker et al. 2017), and calculated the class-metrics patch size (m2) and patch isolation (m). Shannon landscape diversity, patch size, and patch isolation were computed at 200 m and 1500 m radii using Fragstats v4.2 (McGarigal et al. 2012).

3.4. Statistical analysis

To account for lost/stolen pitfall traps, catches were standardized to 84 trap days per site for each year (4 traps x 7 days x 3 months). Samples were pooled per site per year to obtain a comprehensive representation of the spider community and a robust dataset suitable for functional diversity calculations (van der Plas et al. 2017). Years were analyzed separately and statistical analyses were performed in R v3.5.1 and RStudio v1.1.456 (RStudio Team 2016, R

Core Team 2019).

We calculated six functional diversity indices: the dendrogram-based functional alpha diversity

(Petchey and Gaston 2002) and functional beta diversity (Swenson 2011) were computed using the ‘picante’ package (Kembel et al. 2010), while the distance-based functional divergence, functional richness, functional evenness, and functional dispersion were calculated using the

‘FD’ package (Villéger et al. 2008, Laliberté and Legendre 2010). Functional alpha and beta diversity were compared across treatments with Analysis of Deviance and Type-II Wald’s Chi- square tests using the ‘car’ package (Fox and Weisberg 2019). Randomized null model communities (999) were generated using the trait-swap approach (Mason et al. 2008, Swenson

2014) and standardized effect sizes (SES) were calculated for functional alpha and beta diversity.

Observed SES functional diversity values were each compared to a null expectation using

Wilcoxon signed-rank tests, where any significant deviation from zero represents a higher or lower value than would be expected by random chance (Chase et al. 2011).

To examine patterns among spiders and environmental variables, we performed canonical partial least squares analyses (cPLS) using the ‘mixOmics’ R package v6.8 (Rohart et al. 2017). cPLS is 38 a multivariate approach commonly used for analyzing large datasets and maximizing the correlation between two datasets via two sets of latent variables (Tenenhaus 1998, Krishnan et al.

2011). Our response dataset consisted of nine taxonomic metrics, ten spider traits, and six functional diversity indices. Taxonomic variables included abundance, richness, and diversity of

Lycosidae and Linyphiidae species and the abundance, richness and diversity of all spider genera. Spider functional traits comprised hunters, sheet/cob-web weavers, orb-web weavers, stenophagous, euryphagous, ground dwellers, plant dwellers, diurnal spiders, nocturnal spiders, and spider body size. Functional diversity indices consisted of functional alpha- and beta- diversity, and functional divergence, richness, evenness, and dispersion. Environmental predictor variables included seven local habitat variables and 12 landscape-scale variables (six variables at

200 m and six variables at 1500 m). Local features included plant biomass, plant height, Shannon plant diversity, number of blooms, bloom area, pollution load index, and mowing frequency. At each 200 m and 1500 m buffer zone, landscape composition variables consisted of percentage buildings, percentage impervious surface, and percentage grass/shrubs; while landscape configuration variables comprised landscape Shannon diversity, patch isolation, and patch size. cPLS was performed between all 25 spider response variables and 19 environmental predictor variables, and pairwise correlations and clustered image maps were generated using Ward’s hierarchical agglomerative clustering method (González et al. 2012).

4. Results

A total of 16,549 adult and sub-adult spiders were identified during our two-year study representing 21 spider families and 58 genera out of a possible 43 spider families and 254 genera 39 recorded in Ohio. We sampled 7,943 spiders representing 17 families, 47 genera, 22 Linyphiidae species, and seven Lycosidae species in 2015; and 8,606 spiders representing 20 families, 51 genera, 22 Linyphiidae, and eight Lycosidae species in 2016 (Appendix C). Lycosidae and

Linyphiidae represented over 64% and 22% of all spiders collected, respectively. Pardosa milvina (Hentz) was the most frequently trapped spider (45% of the total catch), followed by

Trochosa ruricola (De Geer, 15%), Xysticus spp. (8%), and Grammonota inornata (Emerton,

6%).

4.1. Functional diversity and null models

The overall functional alpha diversity was lower than expected by random chance in 2015 and higher than expected in 2016 (Figure 2.3A-B, Table 2.1). When functional alpha diversity was discriminated by treatment, we found that none of the treatments deviated significantly from null expectations during both years (Figure 2.3A-B). Functional beta diversity was significantly lower than expected by random chance in 2015 for each individual habitat treatment and when all sites were examined together (Figure 2.3C). However, functional beta diversity did not differ from null expectations in 2016 (Figure 2.3D, Table 2.1). We also compared functional alpha and beta diversity across treatments, yet no differences were found in 2015 or 2016.

Figure 2.3. Functional diversity (standardized effect size, mean  standard error) versus null expectations (dash lines) in Vacant Lots [Vac], Urban Meadows [Mea], Low-Diversity Prairies [Low], High-Diversity Prairies [High], and across all sites [ALL]: (A) Functional alpha diversity in 2015 and (B) 2016; and (C) Functional beta diversity in 2015 and (D) 2016. Each diversity index was compared to zero using Wilcoxon’s test and significance is denoted with asterisks when P < 0.05.

Table 2.1. Functional alpha and beta diversity in 2015 and 2016 compared to null expectations using Wilcoxon signed-rank tests.

Functional Diversity Treatment V Pseudo median 95% Conf. Interv. p-value

ALL sites 130 -0.400 -0.694 , -0.090 0.011 Vacant lots 5 -0.587 -1.285 , 0.248 0.078 Alpha 2015 Meadows 7 -0.393 -1.215 , 0.421 0.148 Low-Div Pra 8 -0.479 -0.918 , 0.288 0.195 High-Div Pra 13 -0.096 -0.717 , 0.541 0.547

ALL sites 6 -0.746 -0.910 , -0.565 <0.001 Vacant lots 0 -0.864 -1.246 , -0.428 0.008 Beta 2015 Meadows 0 -0.725 -1.100 , -0.315 0.008 Low-Div Pra 1 -0.793 -1.175 , -0.269 0.016 High-Div Pra 2 -0.557 -0.867 , -0.065 0.023

ALL sites 375 0.250 0.012 , 0.505 0.037 Vacant lots 17 -1.013 -0.437 , 0.406 0.945 Alpha 2016 Meadows 24 0.336 -0.480 , 0.990 0.461 Low-Div Pra 30 0.433 -0.091 , 1.013 0.109 High-Div Pra 27 0.387 -0.293 , 0.816 0.250

ALL sites 246 -0.039 -0.215 , 0.157 0.747 Vacant lots 7 -0.244 -0.540 , 0.181 0.148 Beta 2016 Meadows 19 0.054 -0.660 , 0.557 0.945 Low-Div Pra 28 0.197 -0.107 , 1.049 0.195 High-Div Pra 9 -0.146 -0.526 , 0.210 0.250

4.2. Multivariate pairwise correlation analysis

In 2015, functional beta diversity and taxonomic diversity of spiders were positively associated with plant height and biomass, and negatively correlated with mowing frequency. Conversely, total spider abundance, lycosid abundance, orb weavers, active hunters, ground dwellers, diurnal, nocturnal, and spiders with broad diets were positively correlated with mowing and negatively associated with plant height and biomass (Figure 2.4). Functional divergence was positively associated with landscape diversity and negatively correlated with percentage impervious surface, patch isolation, percentage buildings at 1500 m, whereas linyphiid abundance and plant- dwelling spiders showed the opposite pattern (Figure 2.4). Percentage grass/shrubs at 1500 m and patch size at 200 m were negatively correlated with plant-dwellers and spider functional divergence, respectively.

In 2016, patch isolation and percentage impervious surface at 1500 m were positively correlated with functional beta diversity, spider genera diversity, and lycosid diversity, but negatively correlated with the abundance of lycosids (Figure 2.4). On the other hand, total abundance of spiders, hunters, ground dwellers, diurnal, nocturnal, and spiders with generalist feeding habits were positively correlated with landscape diversity at 200 m, yet negatively associated with patch size and percentage grass/shrubs at 200 m, and patch isolation and percentage impervious surface at 1500 m. Linyphiidae richness also decreased with percentage grass/shrubs at 200 m.

Lastly, spider body size increased with plant height and biomass and decreased with mowing frequency, whereas the abundance and richness of linyphiids and the abundance of orb-web weavers showed the opposite pattern (Figure 2.4).

5. Discussion

To disentangle the environmental filters structuring spider communities, we examined the functional and taxonomic response of these predators to local greenspace management and landscape-scale features in four experimental vacant lot habitats in Cleveland, Ohio. Our results illustrated that a combination of local and landscape processes influenced the assembly of spider communities, and that the relative importance of these assembly processes changed over time.

Our study revealed that: 1) mowing frequency and vegetation structural characteristics influenced the functional diversity of spiders at the local patch scale; 2) at the landscape scale, patch isolation and the amount of surrounding impervious surface shaped patterns of spider functional diversity; 3) functional diversity was similar across vacant lot planting treatments; and

4) Pardosa milvina, a ground-dwelling wolf spider, was the dominant species in urban greenspaces. 45

At the local patch scale, arthropod communities are influenced by environmental conditions associated with habitat quality such as plant diversity, vegetation complexity, and disturbance frequency (McIntyre 2000, Randlkofer et al. 2010, Delgado de la flor et al. 2017). Our results indicated that monthly mowing was a strong local environmental filter for spiders. We predicted that spider functional alpha diversity would be higher than expected by chance, since patches embedded in a fragmented urban landscape would vary in their habitat quality and accessibility for the urban spider species pool. However, we observed that functional alpha and beta diversity were lower than expected in 2015 when Control Vacant Lots, Low- and High-Diversity Pocket

Prairies were mowed monthly to improve prairie establishment and growth. Additionally, we found that high functional similarity among greenspaces was driven by a few dominant disturbance-tolerant species. Spider communities were dominated (45% of sampled individuals) by the ground-hunter Pardosa milvina (Lycosidae), which is known to respond positively to resource availability and is commonly found within disturbed habitats (Marshall et al. 2000,

Burkman and Gardiner 2015). Orb-web weavers Glenognatha spp. (Tetragnathidae) were also abundant in Control Vacant Lot sites. These small (3 mm) species build their webs close to the ground in low-lying vegetation where periodic mowing may not affect their populations.

Although disturbance-tolerant arthropods may benefit, mowing regimes generally have negative impacts on arthropod communities (Garbuzov et al. 2015, Unterweger et al. 2017), and increased mowing likely impairs establishment of spiders with unique habitat requirements and limited dispersal capabilities (Cattin et al. 2003).

In addition to mowing, vegetation structural characteristics were important predictors of spider functional and taxonomic diversity. Plant height and biomass were positively correlated with 46 functional beta and taxonomic diversity of spiders, implying that greenspaces with complex vegetation enhanced the diversity and distinctiveness of spider communities. Studies have shown that managed habitats with complex vegetation structure create more microhabitats, enhancing natural enemy communities (Langellotto and Denno 2004, Sarthou et al. 2014). Moreover, despite studies reporting that spider body size was not adversely affected by urbanization or landscape features (Alaruikka et al. 2002, Kaltsas et al. 2014), we provide evidence that spider body size is driven by local environmental conditions. Not only were small (< 5 mm) linyphiids and tetragnathids more abundant in highly disturbed sites, but some large (> 15 mm) agelenids and lycosids were only found in Meadows and Pocket Prairies. This suggests that changes in vacant lot management, such as increasing vegetation structural complexity and reduced mowing frequency, facilitates the colonization of large and mobile spiders capable of consuming larger prey items.

At the landscape scale, the dominance of impervious surface and built structures, as well as the amount and connectivity of surrounding greenspaces can influence arthropod communities

(Davis et al. 2007, Braaker et al. 2017, Sivakoff et al. 2018). We predicted that spider functional beta diversity (dissimilarity) will increase with impervious surface, due to variation in dispersal capabilities and adaptation to disturbance among species. In 2016, functional dissimilarity increased with impervious surface and the overall spider functional alpha diversity was higher than expected; consequently, our hypothesis was supported in the second year. Null model results indicated that spider functional beta diversity was neither higher nor lower than expected by chance, suggesting stochastic mechanisms of assembly. In 2016, monthly mowing only occurred in the Control Vacant Lots, and once this strong local environmental pressure was 47 relaxed, landscape processes were stronger drivers of spider community assembly. Functional similarity of spider communities decreased as vacant lots became increasingly isolated and surrounded by impervious surface, suggesting that outlying urban green patches contribute to the landscape-wide spider diversity. These isolated vacant lots supported more tiny ballooning spiders rather than larger generalist ground-hunter species. This compositional change in spider communities with increased patch isolation from urbanization likely represents an ecological barrier that hinders colonization of species that lack the traits to overcome dispersal challenges

(Bonte et al. 2004, Braaker et al. 2017). Small web-weavers are effective colonizers via ballooning than larger ground-hunter spiders that disperse by walking or running (Bonte et al.

2004, Blandenier 2009). Moreover, the inherent stochasticity of successful colonization via ballooning could produce patterns of spider assembly in isolated vacant lots that appear random.

Therefore, our findings suggest that in the absence of strong local environmental filters, the permeability of the urban landscape or the extent to which the matrix facilitates or limits dispersal, shapes patterns of spider assembly based on key dispersal traits.

6. Conclusions

This study provides evidence that habitat management, principally mowing frequency, determined the establishment of spider functional groups within distinct vacant habitats, whereas landscape composition and configuration restricted the colonization potential of spiders across the urban mosaic. The relative importance of these assembly processes changed over the two- year study, with landscape processes becoming a major driver as mowing frequency was reduced within Low-Diversity and High-Diversity Pocket Prairie treatments. We also demonstrated that the mechanisms behind spider community assembly were better understood when both 48 functional and taxonomic approaches were investigated concurrently. Most importantly, our functional diversity approach revealed that current vacant lot management strategies favor the establishment of functionally redundant spider groups, suggesting that food-web interactions and ecological processes might be more complex and stable in habitats with reduced disturbance.

Therefore, to promote spiders and reliant conservation targets such as other predatory arthropods and insectivorous birds, shrinking city land use planners should incorporate vegetation plantings such as native grass and wildflower habitats that require reduced mowing. Importantly, to succeed in these endeavors it is imperative for the city to form a coalition with local community members and co-design these conservation themed greenspaces (Riley et al. 2018b, Turo and

Gardiner 2019). In summary, biodiversity and urban conservation depends on patch management and landscape design, as both drive species community assembly across the urban matrix, informing practitioners about the possible consequences of their management decisions.

Acknowledgements

I thank Katie Turo, Denisha Parker and MaLisa Spring for helping and processing vegetation samples for this project. Kayla Perry for guiding me in the conceptual framework of this project.

This work was supported by CAREER DEB Ecosystem Studies Program (CAREER-1253197) to

MMG, and the National Science Foundation Graduate Research Fellowship Program (DGE-

1343012) to YDA and KJT.

CHAPTER 3

Dietary niche breadth and niche overlap of the generalist spider Pardosa milvina

(Lycosidae) is not influenced by urban greenspace management

1. Abstract

Despite the challenges that local species pools experience in urban environments, studies suggest that cities support speciose biotic communities, especially within wildlife-friendly habitats such as urban prairies. Nonetheless, the effect of these greening efforts on the niche breadth of species at higher trophic levels is a major gap knowledge. We investigated the dietary niche breadth and overlap of the generalist spider Pardosa milvina, and whether consumed and available prey were influenced by habitat management and landscape features. Our study took place across eight vacant lots and seven pocket prairies established in Cleveland, Ohio, USA. We hand-collected

220 P. milvina specimens and dietary niche breadth and overlap were calculated per site from prey identified via DNA gut-content analysis. We also sampled P. milvina and available prey using pitfall traps and vacuums to determine if P. milvina abundance was influenced by prey availability, and whether available prey influence its niche breadth and overlap. We found that

Diptera, Collembola, and Hemiptera were frequently detected in the guts of P. milvina, yet the composition of consumed prey did not vary between treatments. We also found that P. milvina dietary niche overlap increased with patch size, yet niche breadth and overlap were not associated with prey availability. Our study showed that greenspace revitalization efforts within cities do not influence the dietary niche of P. milvina spiders. Spider predation patterns were 50 driven by foraging strategies, where taxonomic identify and nutritional content might be more important than prey availability.

2. Introduction

Evidence shows that biodiversity is associated with ecosystem functioning, and therefore conservation programs have focused on protecting and managing habitats to increase the diversity of local species pools (Cardinale et al. 2012). Similar conservation efforts are also underway across cities, where studies have demonstrated that urban greenspaces can harbor diverse arthropod communities (McIntyre 2000, Jones and Leather 2012, Ives et al. 2016).

However, how these diversity patterns affect the foraging activity of predatory groups within cities is poorly known. One of the most important evolutionary processes is competition for available resources (i.e. shelter and food), which can be evaluated by measuring the ecological niche of an organism. Niche breadth is defined as the variety of resources utilized by a species, and is critical to understanding how species will adapt to potential environmental changes (Krebs

1998, Sexton et al. 2017, Kennedy et al. 2019). As food is considered one of the most important dimensions of the niche, elucidating the diet of predominant species can inform us whether the dietary niche breadth is influenced by environmental stochasticity (Krebs 1998, Sexton et al.

2017). This is particularly important in cities where natural habitats have been abruptly altered and new management strategies implemented, revealing that the mechanisms driving local communities in urban areas remain a major knowledge gap (Vanoverbeke et al. 2016, Aronson et al. 2016, Cadotte et al. 2017).

Greening efforts within cities include the establishment of urban farms, rain gardens, and wildflower habitats to alleviate food insecurity and support several other important ecosystem services (Gardiner et al. 2013, Philpott et al. 2014, Green et al. 2016). For instance, Grewal and

Grewal (2012) estimated that urban farms in Cleveland, OH could generate up to 48% of its demand for vegetables and fruits. Empirical evidence, especially in the last decade, show that the revitalization of urban greenspaces supports diverse arthropod communities, such as ants (Uno et al. 2010, Philpott and Bichier 2017), bees (Normandin et al. 2017, Sivakoff et al. 2018), beetles

(Delgado de la flor et al. 2017, Philpott et al. 2019), and spiders (Burkman and Gardiner 2015,

Otoshi et al. 2015). However, whether the management of urban greenspaces broadens the supply available prey and attracts local predators is poorly known.

Due to their ecological relevance as top predators, spiders are cosmopolitan predators widely studied in urban ecosystems (Moorhead and Philpott 2013, Burkman and Gardiner 2015, Lövei et al. 2019). Along urban gradients, spider community composition and distribution patterns vary spatially and temporally (Magura et al. 2010, Varet et al. 2011, Lowe et al. 2018). This variation is attributed to the combination of biotic and abiotic factors interacting with spiders at different spatial scales including food availability (Marshall et al. 2000, Harwood et al. 2003), local habitat characteristics (Greenstone 1984, Horváth et al. 2005), and landscape-scale features

(Braaker et al. 2017, Argañaraz et al. 2018). For instance, the abundance and community composition of spiders are reported to be affected by the degree of greenspace fragmentation

(Marshall et al. 2006, Braaker et al. 2017, Gallé et al. 2019), while the abundance of prey (and its nutritional) value has been positively associated with the abundance and fitness of spiders (Toft and Wise 1999, Harwood et al. 2001, Pekár et al. 2010, Schmidt et al. 2013). 52

Given that niche breadth is a fundamental concept for understanding the adaptation of species to habitat change (Krebs 1998, Sexton et al. 2017, Kennedy et al. 2019), examining the dietary niche of a predominant species at different spatial scales will help us understand if urban greenspace management and design benefit local predators, such as spiders and other predatory groups. Due to the challenges of recording predator-prey interactions, the dietary niche of predatory arthropods has not been thoroughly characterized to this date. Fortunately, new molecular techniques offer efficient and practical means to estimate the dietary niche of species, especially cryptic organisms, such as spiders (Sheppard and Harwood 2005, Pompanon et al.

2012, Symondson and Harwood 2014, Roubinet et al. 2018).

Predator-prey interactions have historically been investigated with visual observations; however, this approach is challenging for nocturnal and fluid-feeding taxa such as spiders. For instance, most spiders are active at night and the lack of hard remains from their meals make it impossible to quantify their dietary niche (Pompanon et al. 2012, Symondson and Harwood 2014). New technological and molecular tools developed in the last decade, such as multiplex PCR and next generation sequencing (Sheppard and Harwood 2005, Symondson and Harwood 2014), have allowed researchers to explore the dietary niche breadth of minute, cryptic organisms, including spiders (King et al. 2008, Pompanon et al. 2012, Pearson et al. 2018). For example, DNA gut- content studies have confirmed that predatory arthropods serve as biological control agents by suppressing pest populations (Roubinet et al. 2017, Yang et al. 2017, Mabin et al. 2020).

Initially, DNA gut-content studies faced some challenges such as designing efficient general primers and avoiding the amplification of the predator’s DNA (Piñol et al. 2014, Krehenwinkel et al. 2018), however new molecular tools are overcoming these challenges. Recent 53 improvements in spider DNA extraction procedures (Krehenwinkel et al. 2017, Macías-

Hernández et al. 2018), universal arthropod primer design (Elbrecht and Leese 2017,

Krehenwinkel et al. 2018, Lafage et al. 2019), and bioinformatic pipelines for ecological metagenetics (Richardson et al. 2019, Wang et al. 2019) are promising means for unravelling spider trophic ecology.

Using DNA gut-content analysis, our objective was to examine if the establishment of urban prairies across the city of Cleveland, Ohio will influence the dietary niche of Pardosa milvina

Hentz (Lycosidae). Cleveland is one of the 350+ cities worldwide that has experienced protracted population loss (Rieniets 2009, Haase 2013), leading to the creation of thousands of greenspaces in the form of vacant lots throughout the city (Gardiner et al. 2013, Herrmann et al.

2016). To do this, we selected 16 vacant lots (former residential properties) where a control and pocket prairie treatment were established. We hypothesized that increased herbaceous vegetation complexity and landscape heterogeneity will enhance food resource partitioning among wolf spiders. Empirical evidence shows that diverse and structurally complex habitats support speciose arthropod assemblages (Siemann 1998, Randlkofer et al. 2010); therefore, we predicted that the dietary niche breadth of P. milvina populations will be greater in urban prairies and positively correlated with plant height, biomass, and bloom abundance (Prediction 1). Likewise, as ecological theory suggests that both the size and heterogeneity of green patches enhance biodiversity (Gibb and Hochuli 2002, Tscharntke et al. 2012), we predicted that niche breadth will be positively correlated with patch size and landscape diversity (Prediction 2).

3. Methods

3.1. Study sites

Prolonged economic decline throughout the Midwestern United States has led to the overabundance of neglected infrastructure in cities across the region. In Cleveland, Ohio, vacated residential properties are demolished and seeded with turf grass, which has resulted in the creation of more than 27,000 vacant lots scattered throughout the city (Western Reserve Land

Conservancy 2018). In 2014, we initiated the Cleveland Pocket Prairie Project, an urban ecology research program that spanned eight inner-city neighborhoods (Delgado de la flor et al. 2020).

For this research project specifically, we selected fifteen vacant lots distributed across the eight neighborhoods wherein two treatments were established. Control Vacant Lots (VL) (n = 8) were seeded with a mixed of non-native grasses, dominated by weedy flowering species (e.g.

Trifolium and Plantago spp.) and mowed monthly to reflect the city’s standard management practices. Pocket Prairies (PP) (n = 7) were seeded with three native grasses and 12-22 native

Ohio flowering plant species (Appendix A).

3.2. Arthropod sampling

P. milvina abundance and available prey were sampled using four pitfall traps and four large vacuums (heavy-duty modified leaf blowers) in all eight VL and (a subset of) four PP sites.

Pitfall traps consisted of 1 L plastic cups (12 cm diameter x 14 cm depth) filled halfway with soapy water (Dawn® Ultra) and were set up for seven consecutive days. While pitfall traps were active, we vacuumed a 0.25 m2 area for 45 seconds using the large vacuums (12 cm diameter).

Adult and sub-adult spiders were confirmed to be P. milvina (Vogel 2004), and potential prey

55 was identified to Order using the American Museum of Natural History online resources

(www.amnh.org). Prey collected with pitfall traps and vacuums is referred to as “available prey”.

3.3. Pardosa milvina sampling, DNA extraction and sequencing

At each site, 12-22 P. milvina (N = 220) were collected for DNA analysis using a small hand- held DC vacuum (BioQuip, Rancho Dominguez, CA) between 10:00 AM and 3:00 PM in 17-28

July 2017. To avoid contamination, spiders were sampled individually using mesh socks attached to the posterior end of the extension nozzle (3 cm diameter). Socks were replaced while nozzles were disinfected and rinsed with 80% ethanol. Specimens were stored individually in 1.5 mL microtubes in 100% ethanol, transported to the laboratory and stored at -70 C.

To maximize prey detection and minimize spider DNA amplification, we only dissected spiders’ opisthosomas (Krehenwinkel et al. 2017, Macías-Hernández et al. 2018). Using a new scalpel for each sample, each opisthosoma was placed in a 1.5 mL centrifuge tube and DNA was extracted following the Qiagen DNeasy Blood & Tissue Kit protocol (Supplementary protocol for insects,

QIAGEN 2019). We selected two universal primers: the forward primer NoSpi2 (5’to3’:

TTYCCHCGWATAAAYAAYATAAG, Lafage et al. 2019) and the reverse primer BR2 (5’to3’:

TCDGGRTGNCCRAARAAYCA, Elbrecht and Leese 2017) as they have shown to successfully amplify DNA of terrestrial arthropods while restricting DNA amplification of Pardosa

(Lycosidae) species. Prey identified in P. milvina guts via DNA gut-content analysis will be referred to as “consumed prey”.

A two-step PCR protocol was used to prepare sample libraries for Illumina sequencing and conditions were as follows: initial denaturation at 96 C for 5 minutes, followed by 38 cycles of

1 minute at 96 C, 30 seconds at 48 C, 50 seconds at 72 C and a final extension of 10 minutes at 72 C. In the initial PCR, the 5 prime ends of the NoSpi2 and BR2 universal COI primers contained Illumina-specific adapters for read priming during sequencing. For this reaction, 25 µL of purified sample DNA extract was combined with 0. 5 µL of 10 mM dNTPs, 2.5 µL of 2 µM forward and reverse primer, 0.75 µL of 100 percent DMSO, 0.2 µL of Phusion HF polymerase and 5 µL of 5x Phusion HF buffer. The total reaction volume was brought to 25 µL with 8.55 µL of molecular grade water. During the second PCR, primers which annealed to the Illumina adapters of the first PCR and contained sample-specific 5 prime oligos for dual-indexing and addition of the Illumina lane hybridization adapter were used. In this reaction, 3 µL of purified product from the initial product was combined with 0.5 µL of 10 mM dNTPs, 2.5 µL of 2 µM forward and reverse primer, 0.2 µL of Phusion HF enzyme and 5 µL of 5X Phusion HF buffer.

The total reaction volume was then brought to 25 µL with the addition of 11.3 µL of molecular grade water. Sequencing was carried out at the Molecular and Cellular Imaging Center located at

The Ohio State University in Wooster, Ohio on an Illumina MiSeq using 600 v3 cycle kit with

30% PhiX spike in.

3.4. Bioinformatics

Following Illumina sequencing, forward and reverse mate-paired reads were merged using Pear v.0.9.1 (Zhang et al. 2014), using a quality trimming Phred score threshold of 20, minimum assembly length of 250 and minimum trim length of 100. Merged fastq files were then converted

57 to fasta format using the ‘--fastq_filter’ function of VSEARCH v2.8.1 (Rognes et al. 2016).

Semi-global VSEARCH alignment was then used to search these fasta-formatted sequences against a global arthropod reference database representative of 87,120 species from 538,864 trimmed and dereplicated sequences. Sequences in this reference database were trimmed to the

BF2-BR2 amplicon region of the COI marker using MetaCurator v1.0.1 as described in

Richardson et al. (2019). Database curation with MetaCurator was conducted with VSEARCH,

MAFFT v7.270 (Katoh 2002) and HMMER v3.1 (Eddy 2011) as dependencies. During alignment, the following VSEARCH settings were used: --id 0.70 --maxaccepts 100 --maxrejects

50 --maxhits 1 --gapopen 0TE --gapext 0TE --query_cov 0.9. The percent identity of the top-hit alignment was then used for classification. A percent identity threshold was set for each rank, from species to order, and the top-hit alignment percent identity was used to determine the lowest classifiable rank according to the following thresholds: Species, 98 percent; genus, 93 percent; family, 82 percent; order, 70 percent. We then used the methods and code provided in

Richardson et al. (2018) to estimate the expected proportion of true positive, true negative, false negative and false positive classifications when using these thresholds. In this cross-validation analysis, 10 percent of the reference data was randomly sampled and set aside as testing data and classified using the remaining 90 percent according to the thresholds described above. All analysis was conducted on the Owens cluster of the Ohio Supercomputer Center.

3.5. Environmental variables

Vegetation sampling took place on July 17-19, 2017. In the middle of each site, a 7 x 15 m grid was placed generating 105 quadrats (1 m2 each) wherein data was collected. Plant biomass was estimated from 20 randomly pre-selected quadrats using the comparative yield method (Haydock 58 and Shaw 1975). Using this method, five quadrats were ranked from 1 to 5 (lowest to highest) reflecting the dry biomass within each site, and each of the 20 quadrats was compared to the five ranked quadrats to estimate the total biomass (Delgado de la flor et al. 2020). Additionally, three vegetation heights and the number of blooms were recorded from six randomly selected quadrats. Plant biomass, plant height and bloom abundance were averaged per site for further analysis. All land cover data were provided by the Cleveland City Planning Commission (CCPC) derived from their 2013 Urban Tree Canopy Cover Assessment (City of Cleveland 2011). For this assessment, the CCPC obtained hi-resolution ground cover data reflecting the physical environment in the City in 2011. Land cover data were classified into the following categories: grass/shrubs, bare soil, water, buildings, roads/railroads, other paved, tree canopy (TC) over vegetation/soil, TC over buildings, TC over roads/railroads, and TC over other paved. Buffers at

200 and 1500 m radii were created around each site and the composition of each landscape was calculated. For the landscape composition calculation, the variables percentage buildings, roads/railroads, and other paved were combined to form “percentage impervious surface”.

The variable Shannon landscape diversity was calculated using all aforementioned land cover classes in FRAGSTATS v4.2 (McGarigal et al. 2012). Taking into account the potential landscape-scale components that most likely influence P. milvina colonization patterns (Bonte et al. 2003, Braaker et al. 2017), we re-classified our landscape cover into either ‘greenspace’

(grass/shrubs & TC over vegetation) or ‘other’, and calculated the class-metrics patch isolation patch size at 200 m and 1500 m radii in FRAGSTATS v4.2 (McGarigal et al. 2012).

3.6. Statistical analysis

Analyses were performed in R v3.6.2 programming language (RStudio Team 2016, R Core

Team 2019). P. milvina’s with a minimum depth of 1000 Arthropoda prey reads were included in the analysis, detected prey DNA was identified to Family and transformed to incidence data.

First, the community composition of consumed prey families was compared between VL and PP using permutational multivariate analysis of variance with the ‘adonis’ function, while the contribution of each prey family to the overall dissimilarities of communities was performed with the ‘simper’ function in the vegan R package (Oksanen et al. 2018).

Secondly, we calculated the dietary niche breadth and overlap of spiders between VL and PP.

DNA reads from each spider were transformed to incidence data and the average frequency was obtained per site. The dietary niche breadth was computed per site using Levins’ measure in the spaa R package (Levins 1968, Zhang 2016). Then, the standardized Levins measure, hereafter niche breadth, was calculated following Hurlbert (1978), where 0 suggests that spiders consume some prey items more than others and 1 indicates equal prey consumption. Dietary niche overlap was calculated using Pianka’s index in the EcoSimR R package (Gotelli et al. 2015), where 0 suggest no overlap in proportion of prey consumption and 1 indicates complete overlap. We also computed Pianka’s niche overlap using 1000 null models and the reshuffling method, and standardized Pianka niche overlap indices, hereafter niche overlap, were calculated at each site.

Thirdly, to investigate the influence of local and landscape variables on predator-prey interactions, we fit linear models using niche breadth and niche overlap as response variables.

Following previous studies in our vacant lot system (Delgado de la flor et al. 2020), we included seven predictor variables: plant biomass, plant height, bloom abundance, patch size & landscape

60 diversity at 200 m, and impervious surface & patch isolation at 1500 m radii. Final models were chosen using backward variable selection and Akaike’s Information Criterion (AIC) procedures.

Fourth, we examined if the abundance of P. milvina in each site was influenced by consumed or available prey. For this, we fit four different negative binomial models using P. milvina abundance as dependent variable, and niche breadth, niche overlap, prey abundance, and prey

Simpson diversity index as independent variables. Finally, we tested if P. milvina dietary niche breadth and overlap were influenced by prey abundance or prey diversity. For this, we fit four different linear models with niche breadth and niche overlap as dependent variables and prey abundance and prey diversity as independent variables.

4. Results

4.1. Community composition of prey detected in P. milvina guts

Out of 220 P. milvina specimens, 171 spiders (92 from VL and 79 from PP) were sequenced to a minimum depth of 1000 Arthropoda prey reads and were included in the analysis. DNA metabarcoding identified 44 families (representing 13 orders) of prey within VL and 46 prey families (12 orders) from PP, for a total of 67 families (18 orders) of arthropods (Appendix D).

The three most frequent prey orders detected were Diptera (36.8% of spiders in VL & 29.2% in

PP), followed by Hemiptera (21.6% VL & 18.1% PP), and Entomobryomorpha (18.1% VL &

19.3% PP). The five most frequent prey family detected were Culicidae (19.3% VL & 19.3%

PP), Entomobryidae (18.1% VL & 19.3% PP), Dolichopodidae (14.6% VL & 10.5% PP),

Leiodidae (10.5% VL & 14.6% PP), and Cicadellidae (16.9% VL & 6.4% PP) (Appendix D).

The community composition of prey detected in the guts of P. milvina and the contribution of

61 each prey taxa to the overall dissimilarity did not differ between Vacant Lots and Pocket Prairies

(R2 = 0.06, F = 0.85, p = 0.67).

Figure 3.1. Frequency of taxa detected in the guts of Pardosa milvina in Control Vacant Lots (right) and Pocket Prairies (left). Out of 171 analyzed spider guts, 113 specimens (66%) tested positive for preying on Diptera (green), 68 spiders (40%) for Hemiptera (magenta), 65 (38%) for Collembola (blue), and 51 (30%) tested positive for Coleoptera (red). Most detected taxa: 6.Leiodidae, 10.Cambaridae, 17.Chloropidae, 18.Culicidae, 19.Dolichopodidae, 30.Syrphidae, 31.Entomobryidae, 36.Cicadellidae, 41.Miridae, and 46.Formicidae. All family names available in Appendix D.

6 10 17 18 19 30 31 36 41 46

4.2. Dietary niche breadth, niche overlap and environmental features

We found that dietary niche breadth (t = 0.66, df = 13, p = 0.52) and dietary niche overlap (t = -

1.44, df = 13, p = 0.17) did not vary between VL and PP. Similarly, when consumed prey was pooled for each treatment, we found that niche overlap between VL and PP was 0.92 (Figure 3.1) and was significantly higher than the simulated niche overlap of 0.18 (Pianka’s Standardized

Effect Size = 7.36, p < 0.001).

Following AIC model selection procedures, we found that the dietary niche breadth was best predicted by a model containing plant biomass, plant height, patch size and landscape diversity at

200 m, yet these variables were not statistically significant (Table 3.1). Dietary niche overlap was best predicted by a model containing plant biomass, patch size and landscape diversity at

200 m (Table 3.1). Of these variables, patch size had a significant positive correlation with niche overlap (t = 2.85, p = 0.02, Figure 3.2).

Figure 3.2. Partial residuals plot of dietary niche overlap and mean patch size at 200 m radius. Gray-shaded area indicates 95% confidence intervals.

4.3. Pardosa milvina abundance and available prey

We analyzed our specimens collected with pitfall traps and vacuums and found that prey abundance (z = -0.48, p = 0.63) and prey diversity (t = -0.08, p = 0.94) did not vary between treatments. We also analyzed the four most abundant orders and found that Diptera was more abundant in Pocket Prairies (z = -2.18, p = 0.03), Hemiptera was more abundant in Vacant Lots

(z = 1.99, p = 0.05), whereas Collembola (z = -1.30, p = 0.19) and Coleoptera abundance (z = -

0.94, p = 0.35) did not differ between treatments.

Additionally, P. milvina abundance did not vary between VL and PP (z = -0.28, p = 0.78), did not respond to prey abundance (z = 0.54, p = 0.59) or prey diversity (z = 0.05 , p = 0.96), and was not influenced by dietary niche breadth (z = -0.92, p = 0.36) or niche overlap (z = 1.18, p =

0.24). Similarly, P. milvina dietary niche breadth was not associated with prey abundance (t = -

1.00, p = 0.34) or prey diversity (t = 0.27, p = 0.79), while niche overlap was neither correlated with prey abundance (t = -0.14, p = 0.89) nor with prey diversity (t = 1.19, p = 0.26).

Dependent Variables Independent Variables Estimate t-value df p-value Levins Niche Breadth ~ Plant Biomass -0.01 -1.80 9 0.11 Plant Height 0.01 1.66 9 0.13 Patch Size at 200 m -2.31 -2.19 9 0.06 Landscape Diversity at 200 m 1.19 1.42 9 0.19

Pianka Niche Overlap ~ Plant Biomass 0.02 1.47 10 0.17 Patch Size at 200 m 360.45 2.85 10 0.02 Landscape Diversity at 200 m 29.85 2.00 10 0.07

5. Discussion

We evaluated several biotic and abiotic factors influencing the abundance of P. milvina, an urban-exploiter and generalist wolf spider. To do this, we compared its dietary niche breadth and overlap across two distinct types of urban greenspace, and examined the effects of environmental variables on niche partitioning and predator-prey interactions. Our results showed that 1) the community composition of prey detected in the guts P. milvina did not vary between treatments,

2) Pocket Prairie establishment did not enhance spider resource partitioning, 3) dietary niche overlap increased with patch size, and 4) P. milvina abundance was not influenced by available or consumed prey.

5.1. Prey community composition

The overall community composition of prey detected in the guts of P. milvina did not differ between VL and PP. The most detected prey taxa in P. milvina guts were Diptera, Hemiptera,

Collembola, Coleoptera. When we analyzed the abundance of these four taxa sampled with pitfall traps and vacuums, Diptera was more abundant in Pocket Prairies while Hemiptera was more abundant in Vacant Lots. However, Diptera and Hemiptera were equally detected in P. milvina guts in both treatments suggesting that a small number of prey items, regardless of their abundance, play a relatively more important role in P. milvina diets. In fact, studies in natural and agricultural systems have reported that Pardosa spp. prey largely on Diptera and Hemiptera

(Yeargan 1975, Nyffeler 1999, Eitzinger et al. 2019), and that prey taxonomic identity was more important than prey abundance and environmental conditions (Kuusk and Ekbom 2010, Eitzinger et al. 2019). This selective feeding behavior has been attributed to the nutritional value and limited anti-predator behavior of prey, especially flies (Mayntz et al. 2005, Rickers et al. 2006).

For example, Schmidt et al. (2012) found that P. milvina selected high-quality prey from a mixture of Drosophila flies. We also found that Culicidae was the most detected taxa in the guts of P. milvina suggesting that this spider often preys on mosquito populations within cities. In natural and open water habitats, culicid mosquitoes have also been reported as part of Pardosa spp. diets (Greenstone 1980, Futami et al. 2008, Wirta et al. 2015, Eitzinger et al. 2019); nevertheless, the extent to which urban areas are benefitted from this service has not been explored and hence warrant further studies.

5.2. Dietary niche breadth, niche overlap and environmental features

We found that P. milvina dietary niche breadth did not differ between VL and PP, suggesting than urban conservation efforts did not affect how spiders partitioned food resources. When the dietary niche breadth was analyzed with local and landscape variables, we found that no local or landscape variable influence the proportion of taxa preyed by this spider. In the Arctic, Eitzinger et al. (2019) examined Pardosa glacialis diet along an environmental gradient and found that neither vegetation nor elevation affected the spider’s diet. We also expected that spiders will be sharing more food resources across Vacant Lot sites, yet the dietary niche overlap of P. milvina did not differ between treatments. Additionally, we found that patch size was positively correlated with niche overlap indicating that spiders were sharing more prey resources when the greenspaces at 200 m radii were on average larger. P. milvina is a generalist predator and there is mounting evidence showing that species abundance and diversity increase with patch size

(Gaston 2000, Tscharntke et al. 2012). Therefore, a plausible explanation is that the type and proportion of prey consumed by P. milvina in each site was higher in larger patches (every member of the population within a site was eating more), leading to the positive correlation between dietary niche overlap and patch size.

5.3 Food-web interactions and environmental components

Urbanization often homogenizes biological communities leading to the overabundance of a small number of urban-adapted species (McKinney 2006, Groffman et al. 2014, Deguines et al. 2016,

Knop 2016). Studies along urbanization gradients have reported that Pardosa and Trochosa spiders are cosmopolitan species in suburban and urban areas (Magura et al. 2010, Varet et al.

2011, Horváth et al. 2012, Burkman and Gardiner 2015). The predominance of Pardosa spiders 67 within cities can be related to co-existence theory along with the concept of fundamental niche

(Krebs 1998, Kraft et al. 2015, Sexton et al. 2017). In other words, the decline of some species

(especially those susceptible disturbance) lead to an increasing number of unoccupied niches across urban greenspaces, which are eventually filled by generalists like Pardosa spiders. In an experimental setting, Marshall et al. (2006) found that Hogna (Tigrosa) helluo populations declined in fragmented and small patches while Pardosa milvina numbers increased in smaller areas. They concluded that P. milvina was a superior colonist, and it colonization strategies coupled with landscape patterns were more important than interspecific interactions or prey availability (Marshall et al. 2000, 2006).

Ecological theory suggests that environmental conditions have the potential to alter niche breadth

(Krebs 1998, Sexton et al. 2017); nonetheless, this is considered an evolutionary process that can take decades. For instance, studies in urban and natural ecosystems have found that the effect of plant communities on higher trophic levels was a slow and gradual process (Johnson et al. 2018,

Damschen et al. 2019, Kennedy et al. 2019). Although we could not identify what drives P. milvina niche breadth and niche overlap patterns, our models suggest that a combination of biotic and abiotic characteristics at different spatial scales influence P. milvina resource utilization.

Habitat structure and behavioral patterns such as cannibalism risk and life history traits explains foraging behavior of P. milvina in disturbed environments (Marshall et al. 2000, Schmidt and

Rypstra 2010, Schmidt et al. 2012, 2013). Moreover, patch size, fragmentation, impervious surface, and air temperature have also shown to influence spiders across urban environments

(Bonte et al. 2004, Magura et al. 2010, Meineke et al. 2017, Braaker et al. 2017).

5.4. Challenges and future directions

Despite the advancements in the field of molecular ecology, gut content analysis and DNA metabarcoding methods present some limitations. It is not uncommon for spiders to prey on other generalist predators, such as ground-dwelling beetles and spiders, and consequently it is challenging to distinguish direct and secondary predation (Sheppard et al. 2005, King et al. 2008,

Pompanon et al. 2012). Another challenge includes detecting prey items that were consumed several hours before the spider was killed (King et al. 2008, Welch et al. 2014). For example,

Kuusk and Ekbom (2010) found that Collembola eaten by Pardosa spp. were only detected if ingested in the last 24 hours, which raise questions about the number and frequency of prey that we could have missed in our study. Once secondary predation and DNA detection have been taken into account, another challenge is to decide how to convert sequence reads to dietary data

(Deagle et al. 2019). Current metabarcoding approaches cannot accurately detect the abundance of consumed prey, and making inferences from incidence data might fail to reveal relevant ecological patterns (Banašek-Richter et al. 2004, Tylianakis et al. 2007). However, we sampled

92 (54%) specimens from VL and 79 (46%) from PP, and dietary niche indices were standardized, showing that spiders were well-represented from both treatments and that errors, if any, were avoided.

6. Conclusions

In summary, our study showed that the dietary niche breadth and overlap of P. milvina was not influenced by the addition of wildlife-friendly habitats, such as urban prairies, to the urban matrix. This shows that although urban greenspace management strategies is important for some taxa such as bees and beetles (Delgado de la flor et al. 2017, Sivakoff et al. 2018), this may not 69 be a defining factor for other predators such as spiders or ants (Uno et al. 2010, Gardiner et al.

2014). This also highlights the importance of examining several taxonomic and functional groups in order to make mindful management decisions that benefit all local species (Politi

Bertoncini et al. 2012, Zhou et al. 2017, Aronson et al. 2017). Given the complexity of the system, future studies should examine species co-occurrence patterns along urban gradients

(Kraft et al. 2015, Cadotte and Tucker 2017), to better understand what biotic and abiotic components influence the niche of this generalist wolf spider.

Acknowledgements

I want to acknowledge Christopher Riley for helping with DNA extraction and Rodney

Richardson for performing the bioinformatics components of the study.

CHAPTER 4

Multi-trophic interactions reveal the biotic and abiotic components driving spiders across

urban greenspaces at different spatial scales

1. Abstract

Urbanization is often considered detrimental to biodiversity, yet urban greenspaces are known to support diverse arthropod communities. Spiders are one of the most abundant natural predators in urban areas, and their establishment is associated to prey availability, local habitat characteristic and landscape-scale components. However, the extent to which these biotic and abiotic features jointly influence spider assemblages is largely unknown. Using a Structural

Equation Modeling approach, we investigated three mechanisms influencing web and hunting spiders across urban greenspaces: prey availability, habitat characteristics, and landscape features. We established eight control vacant lots and seven urban prairies in Cleveland, Ohio where spiders, potential prey, soil properties, vegetation complexity, and landscape data were collected in July 2017. We found that web spiders were negatively correlated with plant biomass, hunting spiders were positively correlated with available prey breadth, and both spider functional groups were positively associated with heavy-metal pollution in the soil. Overall, we revealed that urban greenspace management strategies influence spider functional groups differently, and highlighted the advantages of using pathway analysis to elucidate the most relevant environmental variables shaping local spider assemblages.

2. Introduction

Cities are often considered hostile environments for biotic communities due to the biological and physical constraints that impact the behavior and life cycle, of local species pools (Pickett et al.

2001, Mckinney 2002, Miles et al. 2019). As novel ecosystems, urban areas represent distinct challenges to local wildlife, such as impervious surface and artificial light at night, which disrupt dispersal and feeding patterns (Schoeman 2016, Egerer et al. 2017, Johnson et al. 2018, Philpott et al. 2019). The effects of urbanization on biodiversity has revealed the loss of species in some cities (Deichsel 2006, Horváth et al. 2012, Magura et al. 2013), yet other studies have found that cities support species-rich communities (Dearborn and Kark 2010, Ives et al. 2016, Normandin et al. 2017), warranting urban conservation efforts. These discrepancies are attributed to varied responses among trophic levels and variability in the local management and landscape context of urban greenspaces across cities (McKinney 2006, Kark et al. 2007, Aronson et al. 2014). For instance, McKinney (2008) found that while plant richness increased with urbanization, higher trophic groups such as invertebrates and vertebrates decline along the same urban gradient.

Therefore, to disentangle and identify the main environmental filters influencing species within cities, we need to investigate how multitrophic interactions are influenced by urbanization processes, such as greenspace management and landscape fragmentation (Pickett et al. 2001,

Raupp et al. 2009).

Despite 55% of the world’s population residing in cities and this number projected to increase, over 350 cities worldwide are shrinking due to prolonged population loss and economic decline

(Rieniets 2009, Haase 2013). The demolition and eventual replacement of residential properties by vacant lots has led to an increasing number of newly available greenspaces in shrinking cities. 72

In the Midwest of the United States, local studies have documented that vacant lots support unique and diverse arthropod communities (Gardiner et al. 2013, Philpott et al. 2014, Burkman and Gardiner 2015, Delgado de la flor et al. 2017), providing desired ecosystem services such as pollination and the removal of atmospheric pollutants (Burkman and Gardiner 2014, Riley et al.

2018a, Sivakoff et al. 2018). Local organizations are transforming these vacant lots to community gardens and urban prairies, yet predicting how these management practices might influence the conservation of arthropod assemblages is still poorly known (Gardiner et al. 2013,

Herrmann et al. 2016). Consequently, these novel ecosystems provide a unique opportunity to evaluate how urban greenspace management and design influence multitrophic interactions, and whether urban conservation efforts are having the desired effects on local wildlife communities.

Due to their abundance and role as top predators, spiders have been largely studied in urban environments (Shochat et al. 2004, Magura et al. 2010, Lowe et al. 2018). Emerging trends suggest that although spider diversity patterns have not been greatly affected by urbanization, changes in community composition are attributed to the replacement of habitat-specialists by generalists that thrive in disturbed environments (Samu and Szinetár 2002, Fedoriak et al. 2012,

Burkman and Gardiner 2015, Egerer et al. 2017). In efforts to understand the environmental filters driving spider community assembly within cities, studies have focused on investigating the direct effects of local and landscape variables on spider assemblages (Alaruikka et al. 2002,

Moorhead and Philpott 2013, Meineke et al. 2017, Braaker et al. 2017, Argañaraz et al. 2018).

Spider-prey interaction studies suggest that spiders establish in microsites where more prey is available (Wise 1993, Marshall et al. 2000, Harwood et al. 2001, Lowe et al. 2016), implying a direct relationship between spiders and food resources. However, due to the difficulties of 73 separating the effects of habitat and prey, it has been argued that rather than being influenced by prey abundance, spiders are responding to localized habitat characteristics (Greenstone 1984,

Uetz 1991), as vegetation structural complexity and productivity are known to enhance arthropod assemblages (Langellotto and Denno 2004, Borer et al. 2012, Burkman and Gardiner 2014,

Sarthou et al. 2014). For instance, when the effects of both prey and environmental features were examined concurrently, investigators found that spider communities had a stronger response to local habitat characteristics than prey availability (Greenstone 1984, Bradley 1993).

Furthermore, biotic and abiotic components play an important role in colonization of local species. Below-ground communities and soil properties have been shown to enhance the establishment of diverse plant communities via nutrient availability (Wilder et al. 2011, Maathuis and Diatloff 2013), or in some cases by limiting the establishment of arthropods in sites with high concentrations of heavy metals (Jung et al. 2008, Gongalsky et al. 2010, Gardiner and

Harwood 2017). Additionally, recent studies have reported that species community assembly is driven by landscape-scale patterns (Tscharntke et al. 2012, Aronson et al. 2016), such us patch connectivity and impervious surfaces, which are strong environmental filters detrimental to the dispersal of beetles and spiders (Deichsel 2006, Braaker et al. 2017, Delgado de la flor et al.

2017, Lowe et al. 2018, Grez et al. 2019). The challenges of empirically testing multitrophic interactions and their indirect effects on spiders could lead to the misinterpretation of local food webs (Ings et al. 2009, Schmitz 2017), with implications for urban conservation and management. Therefore, to elucidate the environmental filters driving species community assembly in cities, we need to follow a multitrophic approach where the effects of local biotic and abiotic components are included in the estimation of parameters. 74

Our objective was to determine the relative importance of environmental characteristics influencing spiders directly and indirectly across greenspaces within a city. Using structural equation modeling, we hypothesized three mechanisms influencing spider food webs:

Prey availability: Evidence shows that spiders inhabit locations where prey is abundant

(Harwood et al. 2001, Afzal et al. 2013, Lowe et al. 2016), hence we predicted that web spiders and hunting spiders will be positively correlated with prey breadth and abundance (Figure 4.1).

Habitat characteristics: Spider community composition has been associated with vegetation structural complexity (Uetz 1991, Langellotto and Denno 2004, Burkman and Gardiner 2014).

Therefore, we predicted that plant biomass will be positively correlated with the abundance of hunting spiders, but negatively associated with web spiders that are mainly composed of small linyphiid and tetragnathid species with affinities for turf grass (Burkman and Gardiner 2015,

Delgado de la flor et al. 2020). We also proposed indirect pathways of soil properties and plant biomass on spiders via changes in prey availability (Wilder et al. 2011, Maathuis and Diatloff

2013, Philpott et al. 2014, Gardiner and Harwood 2017), as soil, vegetation, prey and predators are interconnected in ecological foodwebs (Figure 4.1).

Landscape-scale features: Due to their capacity to balloon throughout their life cycle, small web- weaver spiders are strong dispersers (Bonte et al. 2004, Blandenier 2009), whereas ground- dwelling spiders rely on patch connectivity (Bonte et al. 2003, Braaker et al. 2017); hence, we predicted the patch isolation will negatively affect hunting spiders but positively influence web spiders. Furthermore, wolf spiders (Lycosidae) are generalist predators that thrive in disturbed environments (Samu and Szinetár 2002, Burkman and Gardiner 2015). Given that hunting spiders were predominantly lycosid specimens, we predicted that landscape diversity will favor hunting spiders, and therefore restrict the abundance of web spiders (Figure 4.1). 75

Figure 4.1. Hypothesized SEM model. Blue and red arrows indicate predicted positive and negative correlations, respectively.

3. Methods

The study took place in July 2017 within the city of Cleveland, Ohio, USA. Since the 1950s,

Cleveland has experienced significant economic decline and prolonged population loss leading to the abandonment and demolition of residential properties resulting in over 27,000 vacant lots throughout the city (Herrmann et al. 2016, Western Reserve Land Conservancy 2018). Current urban greenspace management strategies include mowing vacant lots to a height of 20 cm. every month in spring and summer. In 2014, we established the Cleveland Pocket Prairie Project

76 wherein 15 vacant lots (former residential properties) across eight residential neighborhoods were seeded with mixes of grass and forb species. We selected eight Control Vacant Lot sites

(seeded with a mixed of non-native grasses and dominated by weedy species) to represent current city management strategies, and seven Pocket Prairie sites (seeded with three native grasses and 22 native Ohio forb species) representing urban conservation efforts (Figure 4.2).

Data collection took place in a 7 x 15 m grid of 105 quadrats (1 m2 each) placed in the middle of each site.

Figure 4.2. Treatments established in 2014 across 15 vacant lots in Cleveland, Ohio: A) Control Vacant Lots (seeded fescue grass and weedy flowering plant species, mowed monthly reflecting the city’s management practices), and B) Pocket Prairies (seeded with a mixture of three native grasses and 22 flowering plants, mowed annually in October).

3.1. Arthropod sampling

Arthropods were collected in each site using four pitfall traps and four vacuum samples (hand- held modified leaf blowers) during 5-14 July 2017. In each site, four quadrats were randomly selected, and pitfall traps were set up for seven consecutive days. Pitfall traps consisted of 1 L plastic cups (12 cm diameter x 14 cm depth) filled halfway with water and a small amount of dish soap (Dawn® Ultra). During non-rainy days and while pitfall traps were active, we vacuumed a 0.25 m2 area for 45 seconds using a modified leaf-blower vacuum (12 cm diameter).

Adult and sub-adult spiders were identified to genus (Paquin and Dupérré 2003, Ubick et al.

2017, World Spider Catalog 2018), web-weaving spiders were classified as web spiders and others were classified as hunting spiders (Table 4.1). Potential prey items were identified to order using the American Museum of Natural History online resources (www.amnh.org), and niche breadth was calculated using Levin’s equation and standardized across sites (Levins 1968,

Hurlbert 1978).

3.2. Local and landscape variables

Vegetation was sampled during 17-19 July 2017 using the comparative yield method (Haydock and Shaw 1975) where five quadrats were ranked from 1 to 5 reflecting the lowest and highest biomass in each site. Then, 20 quadrats were compared to the five ranked quadrats and total biomass per site was estimated (details in Delgado de la flor et al. 2020). Additionally, six random quadrats were selected per site wherein blooming flowers were counted and three vegetation heights were measured. Soils were sampled by collecting at each site three composited soil cores, 2.5 cm diam x 15 cm deep. Stones and debris were removed, and samples were oven-dried at 110 °C for 24 hours to measure soil moisture and passed through a 2 mm 78 sieve. Soil pH was measured with 1:1 deionized water:soil using an AccupHastTM pH meter

(Fisher Scientific, USA) and soil organic matter was determined as loss on ignition at 400 °C for

16 hr. Soil texture was analyzed using the micropipette method to measure particle size distribution (Miller and Miller 1987). Sieved soil samples were analyzed for total elemental content by X-ray fluorescence (XRF) on a NitonTM FXL-959 (Thermo Scientific, USA) according to US EPA Method 6200 (US EPA 2007). Plant-available nitrogen was measured as ammonium (NH4-N) and nitrate (NO3-N), by extraction with 2M KCl (Maynard and Kalra 1993) and quantified using a QuikChem 8500 automatic flow injection analyzer (Lachat, USA).

Soluble plant nutrients and heavy metals were measured by extraction with Mehlich 3 reagent

(Mehlich 1984), followed with quantitation by inductively coupled plasma–atomic emission spectroscopy (ICP–OES). We then estimated the accumulation of heavy metals in the soil. The

Contamination Factor (Loska et al. 2004) of aluminum, antimony, arsenic, barium, cadmium, chromium, cobalt, copper, iron, lead, manganese, nickel, vanadium, and zinc was estimated using regional background levels from eastern United States Environmental Protection Agency.

Subsequently, Pollution Load Indices were calculated per site (Tomlinson et al. 1980,

Weissmannová and Pavlovský 2017).

We also obtained land cover data from the Urban Tree Canopy Cover Assessment from the

Cleveland City Planning Commission captured in 2011 (City of Cleveland 2011). Following previous studies that found relevant patterns on arthropod assemblages across the landscape

(Gardiner et al. 2010, Philpott et al. 2014), we created buffer zones at 200 m and 1500 m radii around each site. We classified landscape cover types and calculated the percentage of: grass/shrubs, bare soil, water, buildings, roads/railroads, other paved, tree canopy (TC) over 79 vegetation/soil, TC over buildings, TC over roads/railroads, and TC over other paved for both

200 and 1500 m buffer zones. For landscape composition, percent buildings, percent roads/railroads, and percent other paved were combined to form “percent impervious surface”, whereas the Shannon landscape diversity was calculated in FRAGSTATS v4.2 (McGarigal et al.

2012). Considering the importance of patch connectivity for our spider functional groups (Bonte et al. 2004, Braaker et al. 2017), we re-classified our landscape cover into “greenspace”

(grass/shrubs and TC vegetation/soil) or “other”, and the class-metrics patch size and patch isolation were also computed in FRAGSTATS v4.2 (McGarigal et al. 2012).

3.3. Statistical analysis

Analysis was performed in R v3.6.2 programming language (RStudio Team 2016, R Core Team

2019). We compared web spider and hunting spider abundance between Vacant Lots and Pocket

Prairies. Our initial list of potential biotic and abiotic components influencing Spiders (web & hunting) included: Prey (abundance & breadth), Soil (% moisture, % organic matter, % clay, % sand, % silt, plant available nutrient, available mycorrhizal fungi & pollution load index),

Vegetation (bloom abundance, plant height & plant biomass), and Landscape (% impervious surface, landscape diversity, mean patch size & patch isolation; each at 200 m and 1500 m radii).

First, we selected the most potentially influential environmental variables on spider ecological networks using canonical partial least squares analyses (cPLS) and clustered image maps (CIM) using the ‘mixOmics’ R package (Rohart et al. 2017). cPLS is a multivariate approach that maximizes the correlation between large datasets via two sets of latent variables (Tenenhaus

1998, Krishnan et al. 2011), whereas CIM calculates Pearson’s pair-wise correlations among 80 each response and predictor variable (González et al. 2012). cPLS followed by CIM, hereafter cPLS-CIM, were performed among Spiders as response variables and Prey, Vegetation,

Landscape and pollution load index as predictor variables. We also computed cPLS-CIM among

Prey as response variables and Spiders, Soil, Vegetation and Landscape as predictor variables.

Likewise, we performed cPLS-CIM among Vegetation as response variables and Prey, Soil and

Landscape as predictor variables. During each cPLS-CIM, only pair-wise correlations higher than 0.5 were considered ecologically important and selected for further analysis.

Secondly, we used structural equation modelling (SEM) to examine potential causal pathways among our variables and identify mediating factors that influence spider composition and abundance in urban landscapes (Grace 2006). We used results from our cPLS-CIM analyses and from Delgado de la flor et al. (2020) to identify potentially important variables and relationships among variables. All variables were log-transformed to meet assumptions of normality and linearity (Ullman 2007) and Pearson correlation coefficients were examined to ensure no multicollinearity (r2 < 0.7) among variables (Grewal et al. 2004). We then constructed a hypothetical SEM that specified relationships among variables (Figure 4.1) and fit our SEM using the ‘lavaan’ package in R (Rosseel 2012). To obtain our final model, we sequentially removed non-significant paths (p > 0.10) until we obtained an adequate model fit, and maximum likelihood methods were used for parameter estimation (Fan et al. 2016). Model performance was evaluated using the chi-square-based goodness-of-fit test, with p > 0.05 indicating a model structure consistent with the data (Grace 2006). We also used the comparative fit index (CFI) and root mean square error of approximation (RMSEA) to assess model fit, with values of >0.97 and

<0.05 indicating acceptable model fit, respectively (Schermelleh-Engel et al. 2003). 81

4. Results

We collected 1625 spiders representing 17 families, of which 838 spiders (52%) represented

Lycosidae and 508 (31%) were Linyphiidae spiders (Table 4.1). A total of 1018 spiders (9 families) were classified as hunting spiders, and 607 (8 families) were classified as web spiders.

Table 4.1. Spiders sampled from eight vacant lots and seven pocket prairies in Cleveland, Ohio in July 2017.

Family Guild Vacant Lot Pocket Prairie Total Total (%) Agelenidae Web 1 1 0.06% Amaurobiidae Web 1 1 0.06% Clubionidae Web 10 11 21 1.29% Corinnidae Hunter 1 1 0.06% Dysderidae Hunter 3 5 8 0.49% Gnaphosidae Hunter 1 1 0.06% Hahniidae Web 1 1 0.06% Linyphiidae Web 327 181 508 31.26% Lycosidae Hunter 468 370 838 51.57% Phrurolithidae Hunter 1 4 5 0.31% Salticidae Hunter 3 11 14 0.86% Tetragnathidae Web 64 8 72 4.43% Theridiidae Web 3 3 0.18% Thomisidae Hunter 80 53 133 8.18% Trachelidae Hunter 2 2 0.12% Zodariidae Hunter 4 12 16 0.98% Grand Total 963 662 1625 100%

Our SEM analysis indicated that web spiders responded differently to environmental factors compared to hunting spiders (Figure 4.3 and Table 4.2). The final SEM model was consistent with the data (Χ2 test, p = 0.78, df = 10) and the CFI and RMSEA values were 1.00 and 0.00, respectively, indicating good model fit.

Table 4.2. Regression coefficients of the log-transformed dependent and independent variables from our final SEM model. Significance was set when p < 0.05

Pathway Response Variable Predictor Variable Estimate Std Error z-value p-value Std Estimate Direct Plant Biomass ~ Clay in Soil 1.02 0.30 3.47 0.00 0.67

Prey Breadth ~ Plant Biomass -0.20 0.14 -1.43 0.15 -0.35

Web Spiders ~ Soil Pollution 2.13 0.90 2.38 0.02 0.49 Plant Biomass -0.50 0.22 -2.32 0.02 -0.47 Lscp Diversity -10.80 5.64 -1.81 0.07 -0.37

Hunting Spiders~ Soil Pollution 1.45 0.62 2.35 0.02 0.41 Prey Breadth 0.72 0.26 2.84 0.01 0.49 Lscp Diversity 6.74 3.87 1.74 0.08 0.31 Indirect Clay x Biomass x Web Spiders -0.56 0.44 -1.26 0.21 -0.34 Biomass x Breadth x Hunting Spiders -0.12 0.17 -0.73 0.46 -0.15 Clay x Biomass x Breadth x Hunting Spiders -0.13 0.15 -0.81 0.42 -0.10

Web spiders were negatively associated with plant biomass (Estimate = -0.47, p = 0.02, Figure

4.3 & 4.4) and landscape diversity (Estimate = -0.37, p = 0.07, Figure 4.3 & 4.4), but positively influenced by soil pollution index (Estimate = 0.49, p = 0.02; Figure 4.3 & 4.4, Table 4.2).

Figure 4.4. Partial residual plots of web spiders from our final SEM model. Grey area represents

95% confidence intervals; statistics are available in Table 4.2.

Hunting spiders were positively associated with soil pollution index (Estimate = 0.41, p = 0.02,

Figure 4.3 & 4.5), prey breadth (Estimate = 0.49, p < 0.01, Figure 4.3 & 4.5), and landscape diversity (Estimate = 0.31, p = 0.08; Figure 4.3 & 4.5, Table 4.2). We also examined indirect pathways among spiders and environmental variables and found no significant associations

(Table 4.2). Our final model also revealed that the percentage of clay in the soil had a positive influence on vegetation biomass (Estimate = 0.67, p < 0.01; Figure 4.6, Table 4.2).

Figure 4.5. Partial residual plots of hunting spiders from our final SEM model. Grey area represents 95% confidence intervals; statistics are available in Table 4.2.

5. Discussion

We investigated the biotic and abiotic components influencing spiders directly and indirectly across urban greenspaces in Cleveland, Ohio. We highlighted the importance of utilizing a path analysis approach and classifying spiders in web-weaving and hunting predators as they responded differently to prey availability and environmental characteristics (Uetz et al. 1999,

Cardoso et al. 2011). We found partial support for each of the three mechanisms shaping spider assemblages, and that no single mechanism can explain the response of every spider hunting guild. Our study found that (1) hunting spider abundance increased with available prey breadth,

(2) hunting and web spider abundance increased with soil pollution, (3) web spiders declined with vegetation biomass, and (4) landscape diversity was associated with the abundance of web and hunting spiders.

Our first mechanism predicted that the abundance and dietary niche breadth of available prey will increase the abundance of hunting and web spiders. Web spiders were not associated with available prey, whereas hunting spiders were positively correlated with potential prey breadth.

Therefore, our Prey Availability hypothesis was supported for hunting spiders. This indicates that prey breadth was important in the establishment of ambush, ground-dwelling, and active- hunter spiders across urban greenspaces. In fact, the establishment of spiders in natural and disturbed environments (Harwood et al. 2003, Wirta et al. 2015, Lowe et al. 2016), and the fitness of Lycosidae spiders have been positively associated with prey availability (Uetz et al.

1992, Schmidt et al. 2013). However, we are aware that the diversity of available prey does not always reflect the proportion of prey that spiders consume in the field.

Our second mechanism predicted that plant biomass will be positively and negatively associated with hunting and web spiders, respectively. Hunting spiders were not correlated with plant biomass, yet we found a negative correlation between web spiders and plant biomass.

Consequently, our Habitat Characteristics hypothesis was supported for web spiders. Most web spiders were 3 mm linyphiid and tetragnathid species that prefer structurally simple vegetation as their webs are built close to the ground. These spiders have been found in large numbers in vacant lot sites, where their minute size and proximity to the ground allow them to overcome periodic mowing (Burkman and Gardiner 2015, Delgado de la flor et al. 2020). Additionally, we found that web and hunting spider abundance increased with the concentration of heavy metals in the soil, which suggest that spiders within cities have adapted to higher-than-expected heavy metal concentrations. Despite that heavy metal contamination can reduce arthropods’ overall fitness (Gardiner and Harwood 2017), the abundance of spiders in contaminated sites could also be in response to the emergence of available niches due to the loss of other predatory groups

(Jung et al. 2008, Żmudzki and Laskowski 2012).

Our third mechanism predicted that landscape diversity will be positively correlated with hunting spiders and negatively associated with web spiders. Our SEM model was best fit when landscape diversity was included in the final model, yet the predicted patterns were not statistically significant. Ecological studies have shown that landscape-scale characteristics shape local species pools (Tscharntke et al. 2012, Gámez-Virués et al. 2015, Fahrig 2017). Across urban ecosystems, studies have found that patch size, impervious surface, connectivity, and air temperature drive spider diversity and community composition (Bonte et al. 2004, Magura et al.

2010, Meineke et al. 2017, Braaker et al. 2017, Lowe et al. 2018). In this study, we observed that 88 urban landscape diversity was negatively associated with web spiders and positively associated with hunting spiders, suggesting heterogenous landscapes affect local spider guilds differently. It is possible that the increasing the number of hunting spiders in heterogenous urban landscapes leads to the decline of web spiders. Yet, given the complexity of ecological foodwebs, more research is needed investigating co-occurrence interactions of spider communities spatially and temporally (Kraft et al. 2015, Aronson et al. 2016, Cadotte and Tucker 2017).

6. Conclusions

We demonstrated that urban greenspace management and design affected web and hunting spiders differently. In fact, empirical evidence have shown that the extent to which structurally complex habitats and land cover types affect arthropod communities is taxa specific (McIntyre et al. 2001, Jones and Leather 2012, Gardiner et al. 2013, Philpott et al. 2014). Our work showed that spider communities and their ecological food webs can be better interpreted when biotic and abiotic variables are examined concurrently at different spatial scales. In fact, we found that both spider groups were positively associated with contaminated soils, web building spiders were affected by vegetation structural complexity, while hunting spiders increased with available prey diversity. Therefore, the mechanisms driving spider community assembly within cities can be elucidated if are spiders classified by foraging strategies. We are aware that intra-guild predation, cannibalism, and bird predation could be driving the colonization and establishment of spiders within cities (Wise 2006, Gunnarsson 2007, Davey et al. 2013), and hence should be examined.

Acknowledgements

I want to acknowledge Lyndsie Collins for analyzing the data using Structural Equation

Modeling, and Larry Phelan for sampling and analyzing the soil component of the study. I also thank members of the Gardiner’s lab and summer filed assistants for collecting and processing vegetation data.

References

Afzal, G., S. Mushtaq, S. A. Rana, and M. A. Sheikh. 2013. Trophic niche breadth and niche

overlap among different guilds of spider species in wheat Agro-ecosystem. Pakistan Journal

of Life and Social Sciences 11:107–111.

Alaruikka, D., D. J. Kotze, K. Matveinen, and J. Niemela. 2002. Carabid beetle and spider

assemblages along a forested urban–rural gradient in southern Finland. Journal of Insect

Conservation 6:195–206.

Argañaraz, C. I., G. D. Rubio, and R. M. Gleiser. 2018. Spider communities in urban green

patches and their relation to local and landscape traits. Biodiversity and Conservation

27:981–1009.

Arnett, J., and M. Thomas. 2000. American Beetles, Volume I. CRC Press, Boca Raton, FL.

Aronson, M. F. J., C. A. Lepczyk, K. L. Evans, M. A. Goddard, S. B. Lerman, J. S. MacIvor, C.

H. Nilon, and T. Vargo. 2017. Biodiversity in the city: key challenges for urban green space

management. Frontiers in Ecology and the Environment 15:189–196.

Aronson, M. F. J., C. H. Nilon, C. A. Lepczyk, T. S. Parker, P. S. Warren, S. S. Cilliers, M. A.

Goddard, A. K. Hahs, C. Herzog, M. Katti, F. A. La Sorte, N. S. G. Williams, and W.

Zipperer. 2016. Hierarchical filters determine community assembly of urban species pools.

Ecology 97:2952–2963.

Aronson, M. F. J., F. A. La Sorte, C. H. Nilon, M. Katti, M. A. Goddard, C. A. Lepczyk, P. S.

Warren, N. S. G. Williams, S. Cilliers, B. Clarkson, C. Dobbs, R. Dolan, M. Hedblom, S.

Klotz, J. L. Kooijmans, I. Kuhn, I. MacGregor-Fors, M. McDonnell, U. Mortberg, P. Pysek,

S. Siebert, J. Sushinsky, P. Werner, and M. Winter. 2014. A global analysis of the impacts

of urbanization on bird and plant diversity reveals key anthropogenic drivers. Proceedings 91

of the Royal Society B: Biological Sciences 281:1–8.

Aviron, S., F. Burel, J. Baudry, and N. Schermann. 2005. Carabid assemblages in agricultural

landscapes: impacts of habitat features, landscape context at different spatial scales and

farming intensity. Agriculture, Ecosystems & Environment 108:205–217.

Banašek-Richter, C., M. F. Cattin, and L. F. Bersier. 2004. Sampling effects and the robustness

of quantitative and qualitative food-web descriptors. Journal of Theoretical Biology

226:23–32.

Bang, C., and S. H. Faeth. 2011. Variation in arthropod communities in response to urbanization:

Seven years of arthropod monitoring in a desert city. Landscape and Urban Planning

103:383–399.

Beck, J., L. Pfiffner, L. Ballesteros-Mejia, T. Blick, and H. Luka. 2013. Revisiting the indicator

problem: Can three epigean arthropod taxa inform about each other’s biodiversity?

Diversity and Distributions 19:688–699.

Beninde, J., M. Veith, and A. Hochkirch. 2015. Biodiversity in cities needs space: A meta-

analysis of factors determining intra-urban biodiversity variation. Ecology Letters 18:581–

592.

Bergman, K. O., L. Ask, J. Askling, H. Ignell, H. Wahlman, and P. Milberg. 2008. Importance of

boreal grasslands in Sweden for butterfly diversity and effects of local and landscape habitat

factors. Biodiversity and Conservation 17:139–153.

Blanco, H., M. Alberti, R. Olshansky, S. Chang, S. M. Wheeler, J. Randolph, J. B. London, J. B.

Hollander, K. M. Pallagst, T. Schwarz, F. J. Popper, S. Parnell, E. Pieterse, and V. Watson.

2009. Shaken, shrinking, hot, impoverished and informal: Emerging research agendas in

planning. Progress in Planning 72:195–250. 92

Blandenier, G. 2009. Ballooning of spiders (Araneae) in Switzerland: general results from an

eleven-year survey. Arachnology 14:308–316.

Blaschke, T. 2010. Object based image analysis for remote sensing. ISPRS Journal of

Photogrammetry and Remote Sensing 65:2–16.

Den Boer, P. J. 1970. On the significance of dispersal power for populations of carabid-beetles

(Coleoptera, Carabidae). Oecologia 4:1–28.

Bohac, J. 1999. Staphylinid beetles as bioindicators. Agriculture, Ecosystems and Environment

74:357–372.

Bonte, D., L. Baert, L. Lens, and J. P. Maelfait. 2004. Effects of aerial dispersal, habitat

specialisation, and landscape structure on spider distribution across fragmented grey dunes.

Ecography 27:343–349.

Bonte, D., L. Lens, J. P. Maelfait, M. Hoffmann, and E. Kuijken. 2003. Patch quality and

connectivity influence spatial dynamics in a dune wolfspider. Oecologia 135:227–233.

Borcard, D., P. Legendre, and P. Drapeau. 1992. Partialling out the spatial component of

ecological variation. Ecology 73:1045–1055.

Borer, E. T., E. W. Seabloom, and D. Tilman. 2012. Plant diversity controls arthropod biomass

and temporal stability. Ecology Letters 15:1457–1464. ter Braak, C. J. F. 1986. Canonical Correspondence Analysis: A New Eigenvector Technique for

Multivariate Direct Gradient Analysis. Ecology 67:1167–1179.

Braaker, S., J. Ghazoul, M. K. Obrist, and M. Moretti. 2014. Habitat connectivity shapes urban

arthropod communities: the key role of green roofs. Ecology 95:1010–1021.

Braaker, S., M. K. Obrist, J. Ghazoul, and M. Moretti. 2017. Habitat connectivity and local

conditions shape taxonomic and functional diversity of arthropods on green roofs. Journal 93

of Animal Ecology 86:521–531.

Bradley, R. A. 1993. The influence of prey availability and habitat on activity patterns and

abundance of Argiope keyserlingi (Araneae: Araneidae). The Journal of Arachnology

21:91–106.

Brose, U. 2003. Bottom-up control of carabid beetle communities in early successional wetlands:

Mediated by vegetation structure or plant diversity? Oecologia 135:407–413.

Brunke, A. J., C. A. Bahlai, J. Klimaszewski, and R. H. Hallett. 2014. Rove beetles (Coleoptera:

Staphylinidae) in Ontario, Canada soybean agroecosystems: Assemblage diversity,

composition, seasonality, and habitat use. Canadian Entomologist 146:652–670.

Brunke, A., A. Newton, J. Klimaszewski, C. Majka, and S. Marshall. 2011. Staphylinidae of

Eastern Canada and Adjacent United States. Canadian Journal of Arthropod Identification

12:1–110.

Burkman, C. E., and M. M. Gardiner. 2014. Urban greenspace composition and landscape

context influence natural enemy community composition and function. Biological Control

75:58–67.

Burkman, C. E., and M. M. Gardiner. 2015. Spider assemblages within greenspaces of a

deindustrialized urban landscape. Urban Ecosystems 18:793–818.

Cadotte, M. W., K. Carscadden, and N. Mirotchnick. 2011. Beyond species: functional diversity

and the maintenance of ecological processes and services. Journal of Applied Ecology

48:1079–1087.

Cadotte, M. W., and C. M. Tucker. 2017. Should Environmental Filtering be Abandoned?

Trends in Ecology and Evolution 32:429–437.

Cadotte, M. W., S. L. E. Yasui, S. Livingstone, and J. S. MacIvor. 2017. Are urban systems 94

beneficial, detrimental, or indifferent for biological invasion? Biological Invasions

19:3489–3503.

Cardinale, B. J., J. E. Duffy, A. Gonzalez, D. U. Hooper, C. Perrings, P. Venail, A. Narwani, G.

M. MacE, D. Tilman, D. A. Wardle, A. P. Kinzig, G. C. Daily, M. Loreau, J. B. Grace, A.

Larigauderie, D. S. Srivastava, and S. Naeem. 2012. Biodiversity loss and its impact on

humanity. Nature 486:59–67.

Cardoso, P., S. Pekár, R. Jocqué, and J. A. Coddington. 2011. Global patterns of guild

composition and functional diversity of spiders. PLoS ONE 6.

Carper, A. L., L. S. Adler, P. S. Warren, and R. E. Irwin. 2014. Effects of suburbanization on

forest bee communities. Environmental Entomology 43:253–262.

Cattin, M. F., G. Blandenier, C. Banašek-Richter, and L. F. Bersier. 2003. The impact of mowing

as a management strategy for wet meadows on spider (Araneae) communities. Biological

Conservation 113:179–188.

Chaffin, B. C., W. D. Shuster, A. S. Garmestani, B. Furio, S. L. Albro, M. M. Gardiner, M.

Spring, and O. O. Green. 2016. A tale of two rain gardens: Barriers and bridges to adaptive

management of urban stormwater in Cleveland, Ohio. Journal of Environmental

Management 183:431–441.

Chase, J. M., N. J. B. Kraft, K. G. Smith, M. Vellend, and B. D. Inouye. 2011. Using null models

to disentangle variation in community dissimilarity from variation in α-diversity. Ecosphere

2:art24.

City of Cleveland. 2011. Cleveland City Planning Commission.

Clough, Y., A. Kruess, and T. Tscharntke. 2007. Organic versus conventional arable farming

systems: Functional grouping helps understand staphylinid response. Agriculture, 95

Ecosystems and Environment 118:285–290.

Coaker, T. H., and D. A. Williams. 1963. The importance of some Carabidae and Staphylinidae

as predators of the cabbage root fly, Erioischia brassicae (Bouché). Entomologia

Experimentalis et Applicata 6:156–164.

Colding, J. 2007. “Ecological land-use complementation” for building resilience in urban

ecosystems. Landscape and Urban Planning 81:46–55.

Concepción, E. D., M. Díaz, D. Kleijn, A. Báldi, P. Batáry, Y. Clough, D. Gabriel, F. Herzog, A.

Holzschuh, E. Knop, E. J. P. Marshall, T. Tscharntke, and J. Verhulst. 2012. Interactive

effects of landscape context constrain the effectiveness of local agri-environmental

management. Journal of Applied Ecology 49:695–705.

Cranmer, L., D. McCollin, and J. Ollerton. 2012. Landscape structure influences pollinator

movements and directly affects plant reproductive success. Oikos 121:562–568.

Damschen, E. I., L. A. Brudvig, M. A. Burt, R. J. Fletcher, N. M. Haddad, D. J. Levey, J. L.

Orrock, J. Resasco, and J. J. Tewksbury. 2019. Ongoing accumulation of plant diversity

through habitat connectivity in an 18-year experiment. Science 365:1478–1480.

Dauber, J., T. Purtauf, A. Allspach, J. Frisch, K. Voigtländer, and V. Wolters. 2005. Local vs.

landscape controls on diversity: A test using surface-dwelling soil macroinvertebrates of

differing mobility. Global Ecology and Biogeography 14:213–221.

Davey, J. S., I. P. Vaughan, R. Andrew King, J. R. Bell, D. A. Bohan, M. W. Bruford, J. M.

Holland, and W. O. C. Symondson. 2013. Intraguild predation in winter wheat: Prey choice

by a common epigeal carabid consuming spiders. Journal of Applied Ecology 50:271–279.

Davis, J. D., D. M. Debinski, and B. J. Danielson. 2007. Local and landscape effects on the

butterfly community in fragmented Midwest USA prairie habitats. Landscape Ecology 96

22:1341–1354.

Deagle, B. E., A. C. Thomas, J. C. McInnes, L. J. Clarke, E. J. Vesterinen, E. L. Clare, T. R.

Kartzinel, and J. P. Eveson. 2019. Counting with DNA in metabarcoding studies: How

should we convert sequence reads to dietary data? Molecular Ecology 28:391–406.

Deák, B., B. Hüse, and B. Tóthmérész. 2016. Grassland vegetation in urban habitats - Testing

ecological theories. Tuexenia 36:379–393.

Dearborn, D. C., and S. Kark. 2010. Motivations for Conserving Urban Biodiversity.

Conservation Biology 24:432–440.

Deguines, N., R. Julliard, M. de Flores, and C. Fontaine. 2016. Functional homogenization of

flower visitor communities with urbanization. Ecology and Evolution 6:1967–1976.

Deichsel, R. 2006. Species change in an urban setting-ground and rove beetles (Coleoptera:

Carabidae and Staphylinidae) in Berlin. Urban Ecosystems 9:161–178.

Delgado de la flor, Y. A., C. E. Burkman, T. K. Eldredge, and M. M. Gardiner. 2017. Patch and

landscape-scale variables influence the taxonomic and functional composition of beetles in

urban greenspaces. Ecosphere 8:e02007.

Delgado de la flor, Y. A., K. I. Perry, K. J. Turo, D. M. Parker, J. L. Thompson, and M. M.

Gardiner. 2020. Local and landscape-scale environmental filters drive the functional

diversity and taxonomic composition of spiders across urban greenspaces. Journal of

Applied Ecology.

Delignette-Muller, M. L., and C. Dutang. 2015. fitdistrplus: An R package for fitting

distributions. Journal of Statistical Software 64:1–34.

Denys, C., and H. Schmidt. 1998. Insect communities on experimental mugwort (Artemisia

vulgaris L.) plots along an urban gradient. Oecologia 113:269–277. 97

Díaz, S., and M. Cabido. 2001. Vive la différence: Plant functional diversity matters to

ecosystem processes. Trends in Ecology and Evolution 16:646–655.

Eckert, S., M. Möller, and S. Buchholz. 2017. Grasshopper diversity of urban wastelands is

primarily boosted by habitat factors. Insect Conservation and Diversity 10:248–257.

Eddy, S. R. 2011. Accelerated profile HMM searches. PLoS Computational Biology 7.

Egerer, M. H., C. Arel, M. D. Otoshi, R. D. Quistberg, P. Bichier, and S. M. Philpott. 2017.

Urban arthropods respond variably to changes in landscape context and spatial scale.

Journal of Urban Ecology 3:1–10.

Eitzinger, B., N. Abrego, D. Gravel, T. Huotari, E. J. Vesterinen, and T. Roslin. 2019. Assessing

changes in arthropod predator–prey interactions through DNA-based gut content analysis—

variable environment, stable diet. Molecular Ecology 28:266–280.

Elbrecht, V., and F. Leese. 2017. Validation and development of COI metabarcoding primers for

freshwater macroinvertebrate bioassessment. Frontiers in Environmental Science 5:1–11.

Eversham, B. C., and M. G. Telfer. 1994. Conservation value of roadside verges for stenotopic

heathland Carabidae: corridors or refugia? Biodiversity and Conservation 3:538–545.

Eyre, M. D., M. L. Luff, and J. C. Woodward. 2003. Beetles (Coleoptera) on brownfield sites in

England: An important conservation resource? Journal of Insect Conservation 7:223–231.

Faeth, S. H., C. Bang, and S. Saari. 2011. Urban biodiversity: patterns and mechanisms. Annals

of the New York Academy of Sciences 1223:69–81.

Fahrig, L. 2017. Ecological Responses to Habitat Fragmentation Per Se. Annual Review of

Ecology, Evolution, and Systematics 48:1–23.

Fan, Y., J. Chen, G. Shirkey, R. John, S. R. Wu, H. Park, and C. Shao. 2016. Applications of

structural equation modeling (SEM) in ecological studies: an updated review. Ecological 98

Processes 5.

Fattorini, S., C. Mantoni, L. De Simoni, and Di. M. P. Galassi. 2018. Island biogeography of

insect conservation in urban green spaces. Environmental Conservation 45:1–10.

Fedoriak, M., S. Rudenko, O. Iaroshynska, and E. Zhukovets. 2012. Spiders (Araneae) of

Chernivtsi City (Ukraine). Arachnologische Mitteilungen 43:37–50.

Fletcher, R. J., R. K. Didham, C. Banks-Leite, J. Barlow, R. M. Ewers, J. Rosindell, R. D. Holt,

A. Gonzalez, R. Pardini, E. I. Damschen, F. P. L. Melo, L. Ries, J. A. Prevedello, T.

Tscharntke, W. F. Laurance, T. Lovejoy, and N. M. Haddad. 2018. Is habitat fragmentation

good for biodiversity? Biological Conservation 226:9–15.

Flynn, D. F. B., N. Mirotchnick, M. Jain, M. I. Palmer, and S. Naeem. 2011. Functional and

phylogenetic diversity as predictors of biodiversity--ecosystem-function relationships.

Ecology 92:1573–1581.

Fox, J., and S. Weisberg. 2019. An R companion to applied regression. Third edition. Sage,

Thousand Oaks, CA.

Futami, K., G. Sonye, P. Akweywa, S. Kaneko, and N. Minakawa. 2008. Diving Behavior in

Anopheles gambiae (Diptera: Culicidae): Avoidance of a Predacious Wolf Spider (Araneae:

Lycosidae) in Relation to Life Stage and Water Depth. Journal of Medical Entomology

45:1050–1056.

Gagic, V., I. Bartomeus, T. Jonsson, A. Taylor, C. Winqvist, C. Fischer, E. M. Slade, I. Steffan-

Dewenter, M. Emmerson, S. G. Potts, T. Tscharntke, W. Weisser, and R. Bommarco. 2015.

Functional identity and diversity of animals predict ecosystem functioning better than

species-based indices. Proceedings of the Royal Society B: Biological Sciences

282:20142620. 99

Gallé, R., A.-K. Happe, A. B. Baillod, T. Tscharntke, and P. Batáry. 2019. Landscape

configuration, organic management, and within-field position drive functional diversity of

spiders and carabids. Journal of Applied Ecology 56:63–72.

Gámez-Virués, S., D. J. Perović, M. M. Gossner, C. Börschig, N. Blüthgen, H. de Jong, N. K.

Simons, A.-M. Klein, J. Krauss, G. Maier, C. Scherber, J. Steckel, C. Rothenwöhrer, I.

Steffan-Dewenter, C. N. Weiner, W. Weisser, M. Werner, T. Tscharntke, and C. Westphal.

2015. Landscape simplification filters species traits and drives biotic homogenization.

Nature Communications 6:8568.

Garbuzov, M., K. A. Fensome, and F. L. W. Ratnieks. 2015. Public approval plus more wildlife:

Twin benefits of reduced mowing of amenity grass in a suburban public park in Saltdean,

UK. Insect Conservation and Diversity 8:107–119.

Gardiner, M. M., C. E. Burkman, and S. P. Prajzner. 2013. The value of urban vacant land to

support arthropod biodiversity and ecosystem services. Environmental Entomology

42:1123–1136.

Gardiner, M. M., and J. D. Harwood. 2017. Influence of heavy metal contamination on urban

natural enemies and biological control. Current Opinion in Insect Science 20:45–53.

Gardiner, M. M., D. A. Landis, C. Gratton, N. Schmidt, M. O’Neal, E. Mueller, J. Chacon, and

G. E. Heimpel. 2010. Landscape composition influences the activity density of Carabidae

and Arachnida in soybean fields. Biological Control 55:11–19.

Gardiner, M. M., S. P. Prajzner, C. E. Burkman, S. Albro, and P. S. Grewal. 2014. Vacant land

conversion to community gardens: Influences on generalist arthropod predators and

biocontrol services in urban greenspaces. Urban Ecosystems 17:101–122.

Garnier, E., J. Cortez, G. Billès, M. L. Navas, C. Roumet, M. Debussche, G. Laurent, A. 100

Blanchard, D. Aubry, A. Bellmann, C. Neill, and J. P. Toussaint. 2004. Plant functional

markers capture ecosystem properties during secondary succession. Ecology 85:2630–2637.

Gaston, K. J. 2000. Global patterns in biodiversity. Nature 405:220–227.

Gibb, H., and D. F. Hochuli. 2002. Habitat fragmentation in an urban environment: large and

small fragments support different arthropod assemblages. Biological Conservation 106:91–

100.

Gobbi, M., and D. Fontaneto. 2008. Biodiversity of ground beetles (Coleoptera: Carabidae) in

different habitats of the Italian Po lowland. Agriculture, Ecosystems and Environment

127:273–276.

Gongalsky, K. B., Z. V. Filimonova, and A. S. Zaitsev. 2010. Relationship between soil

invertebrate abundance and soil heavy metal contents in the environs of the Kosogorsky

Metallurgical Plant, Tula oblast. Russian Journal of Ecology 41:67–70.

González, I., K. A. L. Cao, M. J. Davis, and S. Déjean. 2012. Visualising associations between

paired “omics” data sets. BioData Mining 5:1–23.

Gotelli, N. J., E. M. Hart, and A. M. Ellison. 2015. EcoSimR: Null model analysis for ecological

data.

Grace, J. B. 2006. Structural Equation Modeling and Natural Systems. Cambridge University

Press, Cambridge.

Green, O. O., A. S. Garmestani, S. Albro, N. C. Ban, A. Berland, C. E. Burkman, M. M.

Gardiner, L. Gunderson, M. E. Hopton, M. L. Schoon, and W. D. Shuster. 2016. Adaptive

governance to promote ecosystem services in urban green spaces. Urban Ecosystems 19:77–

93.

Greenstone, M. H. 1980. Contiguous allotopy of Pardosa ramulosa and Pardosa tuoba (Araneae: 101

Lycosidae) in the San Francisco Bay Region, and its implications for patterns of resource

partitioning in the genus. The American Midland Naturalist 104:305–311.

Greenstone, M. H. 1984. Determinants of web spider species diversity: Vegetation structural

diversity vs. prey availability. Oecologia 62:299–304.

Grewal, R., J. A. Cote, and H. Baumgartner. 2004. Multicollinearity and measurement error in

structural equation models: Implications for theory testing. Marketing Science 23:519–529.

Grewal, S. S., and P. S. Grewal. 2012. Can cities become self-reliant in food? Cities 29:1–11.

Grez, A. A., T. Zaviezo, M. M. Gardiner, and A. Alaniz. 2019. Urbanization filters coccinellids

composition and functional trait distributions in greenspaces across greater Santiago, Chile.

Urban Forestry & Urban Greening 38:337–345.

Groffman, P. M., J. Cavender-Bares, N. D. Bettez, J. M. Grove, S. J. Hall, J. B. Heffernan, S. E.

Hobbie, K. L. Larson, J. L. Morse, C. Neill, K. Nelson, J. O’Neil-Dunne, L. Ogden, D. E.

Pataki, C. Polsky, R. R. Chowdhury, and M. K. Steele. 2014. Ecological homogenization of

urban USA. Frontiers in Ecology and the Environment 12:74–81.

Gunnarsson, B. 2007. Bird Predation on Spiders: Ecological Mechanisms and. The Journal of

Arachnology 35:509–529.

Haase, D. 2013. Shrinking Cities, Biodiversity and Ecosystem Services. Pages 253–274

Urbanization, Biodiversity and Ecosystem Services: Challenges and Opportunities. Springer

Netherlands, Dordrecht.

Halme, E., and J. Niemelä. 1993. Carabid beetles in fragments of coniferpus forest. Ann. Zool.

Fennici:17–30.

Harrison, C., and G. Davies. 2002. Conserving biodiversity that matters: Practitioners’

perspectives on brownfield development and urban nature conservation in London. Journal 102

of Environmental Management 65:95–108.

Harwood, J. D., K. D. Sunderland, and W. O. C. Symondson. 2001. Living where the food is:

Web location by linyphiid spiders in relation to prey availability in winter wheat. Journal of

Applied Ecology 38:88–99.

Harwood, J. D., K. D. Sunderland, and W. O. C. Symondson. 2003. Web-location by linyphiid

spiders: Prey-specific aggregation and foraging strategies. Journal of Animal Ecology

72:745–756.

Hatteland, B. A., K. Grutle, C. E. Mong, J. Skartveit, W. O. C. Symondson, and T. Solhøy. 2010.

Predation by beetles (Carabidae, Staphylinidae) on eggs and juveniles of the Iberian slug

Arion lusitanicus in the laboratory. Bulletin of Entomological Research 100:559–567.

Haydock, K. P., and N. H. Shaw. 1975. The comparative yield method for estimating dry matter

yield of pasture. Australian Journal of Experimental Agriculture and Animal Husbandry

15:663–670.

Herrmann, D. L., K. Schwarz, W. D. Shuster, A. Berland, B. C. Chaffin, A. S. Garmestani, and

M. E. Hopton. 2016. Ecology for the Shrinking City. BioScience 66:965–973.

Horváth, R., S. Lengyel, C. Szinetár, and L. Jakab. 2005. The effect of prey availability on spider

assemblages on European black pine (Pinus nigra) bark: Spatial patterns and guild structure.

Canadian Journal of Zoology 83:324–335.

Horváth, R., T. Magura, and B. Tóthmérész. 2012. Ignoring ecological demands masks the real

effect of urbanization: a case study of ground-dwelling spiders along a rural–urban gradient

in a lowland forest in Hungary. Ecological Research 27:1069–1077.

Hothorn, T., F. Bretz, and P. Westfall. 2008. Simultaneous inference in general parametric

models. Biometrical Journal 50:346–363. 103

Hurlbert, S. H. 1978. The measurement of niche overlap and some relatives. Ecology 59:67–77.

Ings, T. C., J. M. Montoya, J. Bascompte, N. Blüthgen, L. Brown, C. F. Dormann, F. Edwards,

D. Figueroa, U. Jacob, J. I. Jones, R. B. Lauridsen, M. E. Ledger, H. M. Lewis, J. M.

Olesen, F. J. F. van Veen, P. H. Warren, and G. Woodward. 2009. Review: Ecological

networks - beyond food webs. Journal of Animal Ecology 78:253–269.

Ives, C. D., P. E. Lentini, C. G. Threlfall, K. Ikin, D. F. Shanahan, G. E. Garrard, S. A. Bekessy,

R. A. Fuller, L. Mumaw, L. Rayner, R. Rowe, L. E. Valentine, and D. Kendal. 2016. Cities

are hotspots for threatened species. Global Ecology and Biogeography 25:117–126.

Johnson, A. L., D. Borowy, and C. M. Swan. 2018. Land use history and seed dispersal drive

divergent plant community assembly patterns in urban vacant lots. Journal of Applied

Ecology 55:451–460.

Jones, E. L., and S. R. Leather. 2012. Invertebrates in urban areas: A review. European Journal

of Entomology 109:463–478.

Jonsson, M., R. Kaartinen, and C. S. Straub. 2017. Relationships between natural enemy

diversity and biological control. Current Opinion in Insect Science 20:1–6.

Jukes, M. R., A. J. Peace, and R. Ferris. 2001. Carabid beetle communities associated with

coniferous plantations in Britain: The influence of site, ground vegetation and stand

structure. Forest Ecology and Management 148:271–286.

Jung, M. P., S. T. Kim, H. Kim, and J. H. Lee. 2008. Species diversity and community structure

of ground-dwelling spiders in unpolluted and moderately heavy metal-polluted habitats.

Water, Air, and Soil Pollution 195:15–22.

Kaltsas, D., E. Panayiotou, M. Chatzaki, and M. Mylonas. 2014. Ground spider assemblages

(araneae: Gnaphosidae) along an urban-rural gradient in the city of heraklion, Greece. 104

European Journal of Entomology 111:59–67.

Kamp, J., A. Pelster, L. Gaedicke, J. Karthäuser, P. Dieker, and K. Mantel. 2015. High nest

survival and productivity of Northern Lapwings Vanellus vanellus breeding on urban

brownfield sites. Journal of Ornithology 156:179–190.

Kark, S., A. Iwaniuk, A. Schalimtzek, and E. Banker. 2007. Living in the city: Can anyone

become an “urban exploiter”? Journal of Biogeography 34:638–651.

Katoh, K. 2002. MAFFT: a novel method for rapid multiple sequence alignment based on fast

Fourier transform. Nucleic Acids Research 30:3059–3066.

Kattwinkel, M., R. Biedermann, and M. Kleyer. 2011. Temporary conservation for urban

biodiversity. Biological Conservation 144:2335–2343.

Kembel, S. W., P. D. Cowan, M. R. Helmus, W. K. Cornwell, H. Morlon, D. D. Ackerly, S. P.

Blomberg, and C. O. Webb. 2010. Picante: R tools for integrating phylogenies and ecology.

Bioinformatics 26:1463–1464.

Kennedy, S., J. Y. Lim, J. Clavel, H. Krehenwinkel, and R. G. Gillespie. 2019. Spider webs,

stable isotopes and molecular gut content analysis: Multiple lines of evidence support

trophic niche differentiation in a community of Hawaiian spiders. Functional

Ecology:1722–1733.

King, R. A., D. S. Read, M. Traugott, and W. O. C. Symondson. 2008. Molecular analysis of

predation: a review of best practice for DNA-based approaches. Molecular Ecology 17:947–

963.

Knop, E. 2016. Biotic homogenization of three insect groups due to urbanization. Global Change

Biology 22:228–236.

Kraft, N. J. B., P. B. Adler, O. Godoy, E. C. James, S. Fuller, and J. M. Levine. 2015. 105

Community assembly, coexistence and the environmental filtering metaphor. Functional

Ecology 29:592–599.

Krebs, C. J. 1998. Niche measures and resource preference. Page 624 Ecological methodology.

Second edition.

Krehenwinkel, H., S. Kennedy, S. Pekár, and R. G. Gillespie. 2017. A cost-efficient and simple

protocol to enrich prey DNA from extractions of predatory arthropods for large-scale gut

content analysis by Illumina sequencing. Methods in Ecology and Evolution 8:126–134.

Krehenwinkel, H., S. R. Kennedy, A. Rueda, A. Lam, and R. G. Gillespie. 2018. Scaling up

DNA barcoding – Primer sets for simple and cost efficient arthropod

systematics by multiplex PCR and Illumina amplicon sequencing. Methods in

Ecology and Evolution 9:2181–2193.

Krishnan, A., L. J. Williams, A. R. McIntosh, and H. Abdi. 2011. Partial Least Squares (PLS)

methods for neuroimaging: A tutorial and review. NeuroImage 56:455–475.

Kuusk, A.-K., and B. Ekbom. 2010. Lycosid spiders and alternative food: Feeding behavior and

implications for biological control. Biological Control 55:20–26.

Lafage, D., V. Elbrecht, J. P. Cuff, D. Steinke, P. A. Hambäck, and A. Erlandsson. 2019. A new

primer for metabarcoding of spider gut contents. Environmental DNA:edn3.62.

Laliberté, E., and P. Legendre. 2010. A distance-based framework for measuring functional

diversity from multiple traits. Ecology 91:299–305.

Langellotto, G. A., and R. F. Denno. 2004. Responses of invertebrate natural enemies to

complex-structured habitats: A meta-analytical synthesis. Oecologia 139:1–10.

Larochelle, A., and M.-C. Larivière. 2003. Natural history of the ground-beetles (Coleoptera:

Carabidae) of America north of Mexico. Pensoft, Sofia, Bulgaria. 106

Lepczyk, C. A., M. F. J. Aronson, K. L. Evans, M. A. Goddard, S. B. Lerman, and J. S. Macivor.

2017. Biodiversity in the city: fundamental questions for understanding the ecology of

urban green spaces for biodiversity conservation. BioScience 67:799–807.

Levins, R. 1968. Evolution in changing environments: some theoretical explorations. Princeton

University Press, New Jersey, USA.

Lindroth, C. H. 1961. The ground beetles (Carabidae, excl. Cicindelinae) of Canada and Alaska,

Part 2-6. Opuscula Entomologica:1–1192.

Lizée, M. H., S. Manel, J. F. Mauffrey, T. Tatoni, and M. Deschamps-Cottin. 2012. Matrix

configuration and patch isolation influences override the species-area relationship for urban

butterfly communities. Landscape Ecology 27:159–169.

Loska, K., D. Wiechuła, and I. Korus. 2004. Metal contamination of farming soils affected by

industry. Environment International 30:159–165.

Lövei, G. L., R. Horváth, Z. Elek, and T. Magura. 2019. Diversity and assemblage filtering in

ground-dwelling spiders (Araneae) along an urbanisation gradient in Denmark. Urban

Ecosystems 22:345–353.

Lowe, E. C., C. G. Threlfall, S. M. Wilder, and D. F. Hochuli. 2018. Environmental drivers of

spider community composition at multiple scales along an urban gradient. Biodiversity and

Conservation 27:829–852.

Lowe, E. C., S. M. Wilder, and D. F. Hochuli. 2016. Persistence and survival of the spider

Nephila plumipes in cities: do increased prey resources drive the success of an urban

exploiter? Urban Ecosystems 19:705–720.

Lowenstein, D. M., K. C. Matteson, I. Xiao, A. M. Silva, and E. S. Minor. 2014. Humans, bees,

and pollination services in the city: The case of Chicago, IL (USA). Biodiversity and 107

Conservation 23:2857–2874.

Maathuis, F. J. M., and E. Diatloff. 2013. Roles and Functions of Plant Mineral Nutrients. Pages

1–21 in F. J. M. Maathuis, editor. Plant Mineral Nutrients. Humana Press, Totowa, NJ.

Mabin, M. D., C. Welty, and M. M. Gardiner. 2020. Predator richness predicts pest suppression

within organic and conventional summer squash (Cucurbita pepo L. Cucurbitales:

Cucurbitaceae). Agriculture, Ecosystems and Environment 287:106689.

Macías-Hernández, N., K. Athey, V. Tonzo, O. S. Wangensteen, M. Arnedo, and J. D. Harwood.

2018. Molecular gut content analysis of different spider body parts. PLoS ONE 13:1–16.

Magura, T., R. Horváth, and B. Tóthmérész. 2010. Effects of urbanization on ground-dwelling

spiders in forest patches, in Hungary. Landscape Ecology 25:621–629.

Magura, T., D. Nagy, and B. Tóthmérész. 2013. Rove beetles respond heterogeneously to

urbanization. Journal of Insect Conservation 17:715–724.

Magura, T., B. Tóthmérész, and T. Molnár. 2004. Changes in carabid beetle assemblages along

an urbanisation gradient in the city of Debrecen, Hungary. Landscape Ecology 19:747–759.

Magura, T., B. Tóthmérész, and T. Molnár. 2008. A species-level comparison of occurrence

patterns in carabids along an urbanisation gradient. Landscape and Urban Planning 86:134–

140.

Mahan, C. G., and T. J. O’Connell. 2005. Small Mammal Use of Suburban and Urban Parks in

Central Pennsylvania. Northeastern Naturalist 12:307–314.

Mannetje, L. T., and K. P. Haydock. 1963. The dry-weight-rank method for the botanical analyis

of pasture. Journal of British Grassland Society 18:268–275.

Marshall, S. D., S. E. Walker, and A. L. Rypstra. 2000. A test for a differential colonization and

competitive ability in two generalist predators. Ecology 81:3341–3349. 108

Marshall, S. D., S. E. Walker, and A. L. Rypstra. 2006. Two ecologically-divergent generalist

predators have different responses to landscape fragmentation. Oikos 114:241–248.

Mason, N. W. H., P. Irz, C. Lanoiselée, D. Mouillot, and C. Argillier. 2008. Evidence that niche

specialization explains species-energy relationships in lake fish communities. Journal of

Animal Ecology 77:285–296.

Matteson, K. C., J. S. Ascher, and G. A. Langellotto. 2008. Bee richness and abundance in New

York City urban gardens. Annals of the Entomological Society of America 101:140–150.

Maynard, D. G., and Y. P. Kalra. 1993. Nitrogen and exchangeable ammonium nitrogen. Pages

25–26 in M. R. Carter, editor. Soil Sampling and Methods of Analysis. Lewis, Boca Raton,

FL.

Mayntz, D., D. Raubenheimer, M. Salomon, S. Toft, and S. J. Simpson. 2005. Nutrient-specific

foraging in invertebrate predators. Science 307:111–113.

McGarigal, K., S. A. Cushman, and E. Ene. 2012. FRAGSTATS v4: Spatial pattern analysis

program for categorical and continuous maps. University of Massachusetts Amherst.

McIntyre, N. E. 2000. Ecology of urban arthropods: A review and a call to action. Annals of the

Entomological Society of America 93:825–835.

McIntyre, N. E., J. Rango, W. F. Fagan, and S. H. Faeth. 2001. Ground arthropod community

structure in a heterogeneous urban environment. Landscape and Urban Planning 52:257–

274.

Mckinney, M. L. 2002. Urbanization, biodiversity, and conservation. BioScience 52:883–890.

McKinney, M. L. 2006. Urbanization as a major cause of biotic homogenization. Biological

Conservation 127:247–260.

McKinney, M. L. 2008. Effects of urbanization on species richness: A review of plants and 109

animals. Urban Ecosystems 11:161–176.

Mehlich, A. 1984. Mehlich 3 soil test extractant: A modification of Mehlich 2 extractant.

Communications in Soil Science and Plant Analysis 15:1409–1416.

Meineke, E. K., A. J. Holmquist, G. M. Wimp, and S. D. Frank. 2017. Changes in spider

community composition are associated with urban temperature, not herbivore abundance.

Journal of Urban Ecology 3:juw010.

Miles, L. S., S. T. Breitbart, H. H. Wagner, and M. T. J. Johnson. 2019. Urbanization Shapes the

Ecology and Evolution of Plant-Arthropod Herbivore Interactions. Frontiers in Ecology and

Evolution 7:1–14.

Miller, W. P., and D. M. Miller. 1987. A micro‐pipette method for soil mechanical analysis.

Communications in Soil Science and Plant Analysis 18:1–15.

Moorhead, L. C., and S. M. Philpott. 2013. Richness and composition of spiders in urban green

spaces in Toledo Ohio. The Journal of Arachnology 41:356–363.

Mulder, C., and J. A. Vonk. 2011. Nematode traits and environmental constraints in 200 soil

systems: Scaling within the 60–6000 µm body size range. Ecology 92:2004.

Muratet, A., N. Machon, F. Jiguet, J. Moret, and E. Porcher. 2007. The role of urban structures in

the distribution of wasteland flora in the Greater Paris area, France. Ecosystems 10:661–

671.

Noordijk, J., A. P. Schaffers, T. Heijerman, and K. V. Sýkora. 2011. Using movement and

habitat corridors to improve the connectivity for heathland carabid beetles. Journal for

Nature Conservation 19:276–284.

Normandin, É., N. J. Vereecken, C. M. Buddle, and V. Fournier. 2017. Taxonomic and

functional trait diversity of wild bees in different urban settings. PeerJ 5:e3051. 110

Nyffeler, M. 1999. Prey selection of spiders in the field. The Journal of Arachnology 27:317–

324.

Oksanen, J., F. G. Blanchet, M. Friendly, R. Kindt, P. Legendre, D. Mcglinn, P. R. Minchin, R.

B. O. Hara, G. L. Simpson, P. Solymos, M. H. H. Stevens, E. Szoecs, and H. Wagner. 2018.

vegan: Community Ecology Package.

Otoshi, M. D., P. Bichier, and S. M. Philpott. 2015. Local and Landscape Correlates of Spider

Activity Density and Species Richness in Urban Gardens. Environmental Entomology

44:1043–1051.

Paquin, P., and N. Dupérré. 2003. Guide d’identification des araignées (Araneae) du Québec.

Association des entomologistes amateurs du Québec, Varennes, Québec.

Park, I. K., and P. Ciorici. 2013. Determinants of vacant lot conversion into community gardens:

evidence from Philadelphia. International Journal of Urban Sciences 17:385–398.

Parris, K. M. 2006. Urban amphibian assemblages as metacommunities. Journal of Animal

Ecology 75:757–764.

Pearson, C. E., W. O. C. Symondson, E. L. Clare, S. J. Ormerod, E. Iparraguirre Bolaños, and I.

P. Vaughan. 2018. The effects of pastoral intensification on the feeding interactions of

generalist predators in streams. Molecular Ecology 27:590–602.

Pedley, S. M., and P. M. Dolman. 2014. Multi-taxa trait and functional responses to physical

disturbance. Journal of Animal Ecology 83:1542–1552.

Pekár, S., D. Mayntz, T. Ribeiro, and M. E. Herberstein. 2010. Specialist ant-eating spiders

selectively feed on different body parts to balance nutrient intake. Animal Behaviour

79:1301–1306.

Petchey, O. L., and K. J. Gaston. 2002. Functional diversity (FD), species richness and 111

community composition. Ecology Letters 5:402–411.

Pfeifer, M., V. Lefebvre, C. A. Peres, C. Banks-Leite, O. R. Wearn, C. J. Marsh, S. H. M.

Butchart, V. Arroyo-Rodríguez, J. Barlow, A. Cerezo, L. Cisneros, N. D’Cruze, D. Faria, A.

Hadley, S. M. Harris, B. T. Klingbeil, U. Kormann, L. Lens, G. F. Medina-Rangel, J. C.

Morante-Filho, P. Olivier, S. L. Peters, A. Pidgeon, D. B. Ribeiro, C. Scherber, L.

Schneider-Maunoury, M. Struebig, N. Urbina-Cardona, J. I. Watling, M. R. Willig, E. M.

Wood, and R. M. Ewers. 2017. Creation of forest edges has a global impact on forest

vertebrates. Nature 551:187–191.

Philpott, S. M., S. Albuquerque, P. Bichier, H. Cohen, M. H. Egerer, C. Kirk, and K. W. Will.

2019. Local and Landscape Drivers of Carabid Activity, Species Richness, and Traits in

Urban Gardens in Coastal California. Insects 10:112.

Philpott, S. M., and P. Bichier. 2017. Local and landscape drivers of predation services in urban

gardens. Ecological Applications 27:966–976.

Philpott, S. M., J. Cotton, P. Bichier, R. L. Friedrich, L. C. Moorhead, S. Uno, and M. Valdez.

2014. Local and landscape drivers of arthropod abundance, richness, and trophic

composition in urban habitats. Urban Ecosystems 17:513–532.

Piano, E., K. De Wolf, F. Bona, D. Bonte, D. E. Bowler, M. Isaia, L. Lens, T. Merckx, D.

Mertens, M. van Kerckvoorde, L. De Meester, and F. Hendrickx. 2017. Urbanization drives

community shifts towards thermophilic and dispersive species at local and landscape scales.

Global Change Biology 23:2554–2564.

Pickett, S. T. A., M. L. Cadenasso, J. M. Grove, C. G. Boone, P. M. Groffman, E. Irwin, S. S.

Kaushal, V. Marshall, B. P. McGrath, C. H. Nilon, R. V. Pouyat, K. Szlavecz, A. Troy, and

P. Warren. 2011. Urban ecological systems: Scientific foundations and a decade of 112

progress. Journal of Environmental Management 92:331–362.

Pickett, S. T. A., M. L. Cadenasso, J. M. Grove, C. H. Nilon, R. V Pouyat, W. C. Zipperer, and

R. Costanza. 2001. Urban ecological systems: Linking terrestrial ecological, physical, and

socioeconomic components of metropolitan areas. Annual Review of Ecology and

Systematics 32:127–157. de Pinho, C. M. D., L. M. G. Fonseca, T. S. Korting, C. M. de Almeida, and H. J. H. Kux. 2012.

Land-cover classification of an intra-urban environment using high-resolution images and

object-based image analysis. International Journal of Remote Sensing 33:5973–5995.

Piñol, J., V. San Andres, E. L. Clare, and W. O. C. Mir, G. Symondson. 2014. A pragmatic

approach to the analysis of diets of generalist predators : the use of next-generation

sequencing with no blocking probes. Molecular Ecology:18–26. van der Plas, F., R. van Klink, P. Manning, H. Olff, and M. Fischer. 2017. Sensitivity of

functional diversity metrics to sampling intensity. Methods in Ecology and Evolution

8:1072–1080.

Poff, N. L. 1997. Landscape Filters and Species Traits: Towards Mechanistic Understanding and

Prediction in Stream Ecology. Journal of the North American Benthological Society

16:391–409.

Poff, N. L., J. D. Olden, N. K. M. Vieira, D. S. Finn, M. P. Simmons, and B. C. Kondratieff.

2006. Functional trait niches of North American lotic insects: traits-based ecological

applications in light of phylogenetic relationships. Journal of the North American

Benthological Society 25:730–755.

Politi Bertoncini, A., N. Machon, S. Pavoine, and A. Muratet. 2012. Local gardening practices

shape urban lawn floristic communities. Landscape and Urban Planning 105:53–61. 113

Pompanon, F., B. E. Deagle, W. O. C. Symondson, D. S. Brown, S. N. Jarman, and P. Taberlet.

2012. Who is eating what: Diet assessment using next generation sequencing. Molecular

Ecology 21:1931–1950.

QIAGEN. 2019. DNeasy Blood & Tisue Kit.

https://www.qiagen.com/us/resources/resourcedetail?id=cabd47a4-cb5a-4327-b10d-

d90b8542421e&lang=en.

R Core Team. 2019. R: A language and environment for statistical Computing. R Foundation for

Statistical Computing, Vienna, Austria.

Rainio, J., and J. Niemelä. 2003. Ground beetles (Coleoptera: Carabidae) as bioindicators.

Biodiversity and Conservation 12:487–506.

Randlkofer, B., E. Obermaier, M. Hilker, and T. Meiners. 2010. Vegetation complexity—The

influence of plant species diversity and plant structures on plant chemical complexity and

arthropods. Basic and Applied Ecology 11:383–395.

Raupp, M. J., P. M. Shrewsbury, and D. A. Herms. 2009. Ecology of herbivorous arthropods in

urban landscapes. Annual Review of Entomology 55:19–38.

Richardson, R. T., J. Bengtsson-Palme, M. M. Gardiner, and R. M. Johnson. 2018. A reference

cytochrome c oxidase subunit I database curated for hierarchical classification of arthropod

metabarcoding data. PeerJ 2018:1–19.

Richardson, R. T., D. B. Sponsler, H. McMinn-Sauder, and R. M. Johnson. 2019. MetaCurator:

A hidden Markov model-based toolkit for extracting and curating sequences from

taxonomically-informative genetic markers. Methods in Ecology and Evolution 2019:1–6.

Rickers, S., R. Langel, and S. Scheu. 2006. Dietary routing of nutrients from prey to offspring in

a generalist predator: Effects of prey quality. Functional Ecology 20:124–131. 114

Rieniets, T. 2009. Shrinking cities: Causes and effects of urban population losses in the twentieth

century. Nature and Culture 4:231–254.

Riley, C. B., D. A. Herms, and M. M. Gardiner. 2018a. Exotic trees contribute to urban forest

diversity and ecosystem services in inner-city Cleveland, OH. Urban Forestry and Urban

Greening 29:367–376.

Riley, C. B., K. I. Perry, K. Ard, and M. M. Gardiner. 2018b. Asset or liability? Ecological and

sociological tradeoffs of urban spontaneous vegetation on vacant land in shrinking cities.

Sustainability (Switzerland) 10:1–19.

Rognes, T., T. Flouri, B. Nichols, C. Quince, and F. Mahé. 2016. VSEARCH: A versatile open

source tool for metagenomics. PeerJ 2016:1–22.

Rohart, F., B. Gautier, A. Singh, and K.-A. Lê Cao. 2017. mixOmics: An R package for ‘omics

feature selection and multiple data integration. PLOS Computational Biology 13:e1005752.

Rosseel, Y. 2012. lavaan : an R package for structural equation modeling. Journal of Statistical

Software 48:1–93.

Roubinet, E., K. Birkhofer, G. Malsher, K. Staudacher, B. Ekbom, M. Traugott, and M. Jonsson.

2017. Diet of generalist predators reflects effects of cropping period and farming system on

extra- and intraguild prey. Ecological Applications 27:1167–1177.

Roubinet, E., T. Jonsson, G. Malsher, K. Staudacher, M. Traugott, B. Ekbom, and M. Jonsson.

2018. High redundancy as well as complementary prey choice characterize generalist

predator food webs in agroecosystems. Scientific Reports 8:1–10.

RStudio Team. 2016. RStudio: Integrated development environment for R. RStudio, Inc.,

Boston, MA.

Rundlöf, M., H. Nilsson, and H. G. Smith. 2008. Interacting effects of farming practice and 115

landscape context on bumble bees. Biological Conservation 141:417–426.

Rusch, A., K. Birkhofer, R. Bommarco, H. G. Smith, and B. Ekbom. 2015. Predator body sizes

and habitat preferences predict predation rates in an agroecosystem. Basic and Applied

Ecology 16:250–259.

Sadler, J. P., E. C. Small, H. Fiszpan, M. G. Telfer, and J. Niemelä. 2006. Investigating

environmental variation and landscape characteristics of an urban-rural gradient using

woodland carabid assemblages. Journal of Biogeography 33:1126–1138.

Samu, F., and C. Szinetár. 2002. On the nature of agrobiont spiders. Journal of Arachnology

30:389–402.

Sarthou, J.-P., A. Badoz, B. Vaissière, A. Chevallier, and A. Rusch. 2014. Local more than

landscape parameters structure natural enemy communities during their overwintering in

semi-natural habitats. Agriculture, Ecosystems & Environment 194:17–28.

Sattler, T., M. K. Obrist, P. Duelli, and M. Moretti. 2011. Urban arthropod communities: Added

value or just a blend of surrounding biodiversity? Landscape and Urban Planning 103:347–

361.

Schaffers, A. P., I. P. Raemakers, and K. V. Sýkora. 2012. Successful overwintering of

arthropods in roadside verges. Journal of Insect Conservation 16:511–522.

Schermelleh-Engel, K., H. Moosbrugger, and H. Müller. 2003. Evaluating the fit of structural

equation models: Tests of significance and descriptive goodness-of-fit measures. MPR-

online 8:23–74.

Schmidt, J. M., J. D. Harwood, and A. L. Rypstra. 2013. Influence of prey availability on

seasonal fluctuation in body condition in the wolf spider, Pardosa milvina (Araneae:

Lycosidae). Journal of Arachnology 41:400–403. 116

Schmidt, J. M., and A. L. Rypstra. 2010. Opportunistic predator prefers habitat complexity that

exposes prey while reducing cannibalism and intraguild encounters. Oecologia 164:899–

910.

Schmidt, J. M., P. Sebastian, S. M. Wilder, and A. L. Rypstra. 2012. The nutritional content of

prey affects the foraging of a generalist arthropod predator. PLoS ONE 7.

Schmitz, O. 2017. Predator and prey functional traits: understanding the adaptive machinery

driving predator–prey interactions. F1000Research 6:1767.

Schoeman, M. C. 2016. Light pollution at stadiums favors urban exploiter bats. Animal

Conservation 19:120–130.

Schwarz, K., M. L. Cadenassol, and S. T. A. Pickett. 2007. Spatial heterogeneity in urban

ecosystems: reconceptualizing land cover ecosystems and a framework for classification.

Frontiers in Ecology and the Environment 5:80–88.

Schwarz, N., M. Moretti, M. N. Bugalho, Z. G. Davies, D. Haase, J. Hack, A. Hof, Y. Melero, T.

J. Pett, and S. Knapp. 2017. Understanding biodiversity-ecosystem service relationships in

urban areas: A comprehensive literature review. Ecosystem Services 27:161–171.

Sexton, J. P., J. Montiel, J. E. Shay, M. R. Stephens, and R. A. Slatyer. 2017. Evolution of

Ecological Niche Breadth. Annual Review of Ecology, Evolution, and Systematics 48:183–

206.

Shannon, C. E. 1948. A mathematical theory of communication. Bell system technical journal

27:379–423.

Sheppard, S. K., J. Bell, K. D. Sunderland, J. Fenlon, D. Skervin, and W. O. C. Symondson.

2005. Detection of secondary predation by PCR analyses of the gut contents of invertebrate

generalist predators. Molecular Ecology 14:4461–4468. 117

Sheppard, S. K., and J. D. Harwood. 2005. Advances in molecular ecology: Tracking trophic

links through predator-prey food-webs. Functional Ecology 19:751–762.

Shochat, E., S. B. Lerman, J. M. Anderies, P. S. Warren, S. H. Faeth, and C. H. Nilon. 2010.

Invasion, Competition, and Biodiversity Loss in Urban Ecosystems. BioScience 60:199–

208.

Shochat, E., W. L. Stefanov, M. E. A. Whitehouse, and S. H. Faeth. 2004. Urbanization and

spider diversity: Influences of human modification of habitat structure and productivity.

Ecological Applications 14:268–280.

Shochat, E., P. S. Warren, S. H. Faeth, N. E. McIntyre, and D. Hope. 2006. From patterns to

emerging processes in mechanistic urban ecology. Trends in Ecology and Evolution

21:186–191.

Siemann, E. 1998. Experimental tests of effects of plant productivity and diversity on grassland

arthropod diversity. Ecology 79:2057–2070.

Sivakoff, F. S., S. P. Prajzner, and M. M. Gardiner. 2018. Unique Bee Communities within

Vacant Lots and Urban Farms Result from Variation in Surrounding Urbanization Intensity.

Sustainability 10:1926.

Small, E. C., J. P. Sadler, and M. G. Telfer. 2002. Carabid beetle assemblages on urban derelict

sites in Birmingham, UK. Journal of Insect Conservation 6:233–246.

Su, X. 2010. Urban conservation in Lijiang, China: Power structure and funding systems. Cities

27:164–171.

Suding, K. N., S. Lavorel, F. S. Chapin, J. H. C. Cornelissen, S. Díaz, E. Garnier, D. Goldberg,

D. U. Hooper, S. T. Jackson, and M. L. Navas. 2008. Scaling environmental change through

the community-level: A trait-based response-and-effect framework for plants. Global 118

Change Biology 14:1125–1140.

Swan, C. M., S. T. A. Pickett, K. Szlavecz, P. Warren, and K. T. Willey. 2011. Biodiversity and

Community Composition in Urban Ecosystems: Coupled Human, Spatial, and

Metacommunity Processes. Pages 179–186 Urban Ecology. Oxford University Press,

Oxford.

Swenson, N. G. 2011. Phylogenetic beta diversity metrics, trait evolution and inferring the

functional beta diversity of communities. PLoS ONE 6:e21264.

Swenson, N. G. 2014. Functional and Phylogenetic Ecology in R. Springer New York, New

York, NY.

Symondson, W. O. C., and J. D. Harwood. 2014. Special issue on molecular detection of trophic

interactions: Unpicking the tangled bank. Molecular Ecology 23:3601–3604.

Tenenhaus, M. 1998. La Regression PLS: theorie et pratique. Editions Technip, Paris, France.

Thomas, M. B., H. J. Mitchell, and S. D. Wratten. 1992. Abiotic and biotic factors influencing

the winter distribution of predatory insects. Oecologia 89:78–84.

Toft, S., and D. H. Wise. 1999. Growth, development, and survival of a generalist predator fed

single- and mixed-species diets of different quality. Oecologia 119:191–197.

Tomlinson, D. L., J. G. Wilson, C. R. Harris, and D. W. Jeffrey. 1980. Problems in the

assessment of heavy-metal levels in estuaries and the formation of a pollution index.

Helgoländer Meeresuntersuchungen 33:566–575.

Tóthmérész, B., I. Máthé, E. Balázs, and T. Magura. 2011. Responses of carabid beetles to

urbanization in Transylvania (Romania). Landscape and Urban Planning 101:330–337.

Tscharntke, T., I. Steffan-Dewenter, A. Kruess, and C. Thies. 2002. Contribution of small habitat

fragments to conservation of insect communities of grassland-cropland landscapes. 119

Ecological Applications 12:354–363.

Tscharntke, T., J. M. Tylianakis, T. A. Rand, R. K. Didham, L. Fahrig, P. Batáry, J. Bengtsson,

Y. Clough, T. O. Crist, C. F. Dormann, R. M. Ewers, J. Fründ, R. D. Holt, A. Holzschuh, A.

M. Klein, D. Kleijn, C. Kremen, D. A. Landis, W. Laurance, D. Lindenmayer, C. Scherber,

N. Sodhi, I. Steffan-Dewenter, C. Thies, W. H. van der Putten, and C. Westphal. 2012.

Landscape moderation of biodiversity patterns and processes - eight hypotheses. Biological

Reviews 87:661–685.

Turo, K. J., and M. M. Gardiner. 2019. From potential to practical: conserving bees in urban

public green spaces. Frontiers in Ecology and the Environment 17:167–175.

Tylianakis, J. M., T. Tscharntke, and O. T. Lewis. 2007. Habitat modification alters the structure

of tropical host–parasitoid food webs. Nature 445:202–205.

Ubick, D., P. Paquin, and P. E. Cushing. 2017. Spiders of North America: an identification

manual. Page (D. Ubick, P. Paquin, P. E. Cushing, and V. Roth, Eds.) American

Arachnological Society. Second edition. American Arachnological Society, Keene, New

Hampshire.

Uetz, G. W. 1991. Habitat structure and spider foraging. Pages 325–348 Habitat Structure.

Springer Netherlands, Dordrecht.

Uetz, G. W., J. Bischoff, and J. Raver. 1992. Survivorship of wolf spiders (Lycosidae) reared on

different diets. The Journal of Arachnology 20:207–211.

Uetz, G. W., J. Halaj, and A. B. Cady. 1999. Guild structure of spiders in major crops. Journal of

Arachnology 27:270–280.

Ullman, J. 2007. Structural equation modeling. Pages 676–780 in B. G. Tabachnick and L. S.

Fidell, editors. Using Multivariate Statistics. 5th ed. Pearson/Allyn & Bacon, Boston, USA. 120

Uno, S., J. Cotton, and S. M. Philpott. 2010. Diversity, abundance, and species composition of

ants in urban green spaces. Urban Ecosystems 13:425–441.

Unterweger, P. A., C. Rieger, and O. Betz. 2017. The influence of urban lawn mowing regimes

on diversity of Heteroptera (Hemiptera). Heteropteron 48:7–21.

US EPA. 2007. Method 6200: Field portable X-ray Fluorescence Spectrometry for the

determination of elemental concentrations in soil and sediment. Unites States

Environmental Protection Agency, Washington D.C.

Vanoverbeke, J., M. C. Urban, and L. De Meester. 2016. Community assembly is a race between

immigration and adaptation: eco-evolutionary interactions across spatial scales. Ecography

39:858–870.

Varet, M., J. Pétillon, and F. Burel. 2011. Comparative responses of spider and carabid beetle

assemblages along an urban-rural boundary gradient. Journal of Arachnology 39:236–243.

Venn, S. 2007. Morphological responses to disturbance in wing-polymorphic carabid species

(Coleoptera: Carabidae) of managed urban grasslands. Baltic Journal of Coleopterology

7:51–59.

Vergnes, A., V. Pellissier, G. Lemperiere, C. Rollard, and P. Clergeau. 2014. Urban densification

causes the decline of ground-dwelling arthropods. Biodiversity and Conservation 23:1859–

1877.

Villéger, S., N. W. H. Mason, and D. Mouillot. 2008. New multidimensional functional diversity

indices for a multifaceted framework in functional ecology. Ecology 89:2290–2301.

Violle, C., M. L. Navas, D. Vile, E. Kazakou, C. Fortunel, I. Hummel, and E. Garnier. 2007. Let

the concept of trait be functional! Oikos 116:882–892.

Vogel, B. R. 2004. A review of the spider genera Pardosa and Acantholycosa (Araneae, 121

Lycosidae) of the 48 contiguous United States. Journal of Arachnology 32:55–108.

Vonk, J. A., A. M. Breure, and C. Mulder. 2013. Environmentally-driven dissimilarity of trait-

based indices of nematodes under different agricultural management and soil types.

Agriculture, Ecosystems and Environment 179:133–138.

Wang, X., F. Hua, L. Wang, D. S. Wilcove, and D. W. Yu. 2019. The biodiversity benefit of

native forests and mixed-species plantations over monoculture plantations. Diversity and

Distributions 25:1721–1735.

Weissmannová, H. D., and J. Pavlovský. 2017. Indices of soil contamination by heavy metals –

methodology of calculation for pollution assessment (minireview). Environmental

Monitoring and Assessment 189.

Welch, K. D., M. R. Schofield, E. G. Chapman, and J. D. Harwood. 2014. Comparing rates of

springtail predation by web-building spiders using Bayesian inference. Molecular Ecology

23:3814–3825.

Western Reserve Land Conservancy. 2018. Cleveland by the Numbers: 2018 Cleveland Property

Inventory. https://www.wrlandconservancy.org/publications/.

Wilcove, D. S., D. Rothstein, J. Dubow, A. Phillips, and E. Losos. 1998. Quantifying Threats to

Imperiled Species in the United States. BioScience 48:607–615.

Wilder, S. M., D. A. Holway, A. V. Suarez, and M. D. Eubanks. 2011. Macronutrient content of

plant-based food affects growth of a carnivorous arthropod. Ecology 92:325–332.

Wirta, H. K., E. Weingartner, P. A. Hambäck, and T. Roslin. 2015. Extensive niche overlap

among the dominant arthropod predators of the High Arctic. Basic and Applied Ecology

16:86–92.

Wise, D. H. 1993. Spiders in Ecological Webs. Cambridge University Press, Cambridge. 122

Wise, D. H. 2006. Cannibalism, Food Limitation, Intraspecific Competition, and the Regulation

of Spider Populations. Annual Review of Entomology 51:441–465.

Woodcock, B. A., J. Redhead, A. J. Vanbergen, L. Hulmes, S. Hulmes, J. Peyton, M.

Nowakowski, R. F. Pywell, and M. S. Heard. 2010. Impact of habitat type and landscape

structure on biomass, species richness and functional diversity of ground beetles.

Agriculture, Ecosystems and Environment 139:181–186.

Work, T. T., J. Klimaszewski, E. Thiffault, C. Bourdon, D. Paré, Y. Bousquet, L. Venier, and B.

Titus. 2013. Initial responses of rove and ground beetles (Coleoptera, Staphylinidae,

Carabidae) to removal of logging residues following clearcut harvesting in the boreal forest

of Quebec, Canada. ZooKeys 258:31–52.

World Spider Catalog. 2018. World Spider Catalog. https://wsc.nmbe.ch.

Yang, T., J. Liu, L. Yuan, Y. Zhang, D. Li, I. Agnarsson, and J. Chen. 2017. Molecular

identification of spiders preying on Empoasca vitis in a tea plantation. Scientific Reports

7:7784.

Yeargan, K. V. 1975. Prey and periodicity of Pardosa ramulosa (McCook) in Alfalfa.

Environmental Entomology 4:137–141.

Zhang, J. 2016. spaa: SPecies Association Analysis.

Zhang, J., K. Kobert, T. Flouri, and A. Stamatakis. 2014. PEAR: A fast and accurate Illumina

Paired-End reAd mergeR. Bioinformatics 30:614–620.

Zhou, W., S. T. A. Pickett, and M. L. Cadenasso. 2017. Shifting concepts of urban spatial

heterogeneity and their implications for sustainability. Landscape Ecology 32:15–30.

Zhou, W., and A. Troy. 2008. An object‐oriented approach for analysing and characterizing

urban landscape at the parcel level. International Journal of Remote Sensing 29:3119–3135. 123

Żmudzki, S., and R. Laskowski. 2012. Biodiversity and structure of spider communities along a

metal pollution gradient. Ecotoxicology 21:1523–1532.

124

Appendix A: Plant species seeded in low-diversity and high-diversity prairies in Cleveland,

Ohio in 2014 and 2016.

125

Low-Diversity Prairies were seeded with 3 native grasses and 4 native Ohio flowering species, whereas High-Diversity Prairies were seeded with 3 native grasses and 16 native Ohio flowering species. Prairie sites were seeded with a blend of 40% forbs and 60% grasses by Ohio Prairie Nursery (Hiram, Ohio) after 2 applications of glyphosate based non-selective herbicide in May and June 2014. All prairie sites were over-seeded in January of 2016 with a mixture of 2 or 6 additional forb species in our Low- and High-Diversity pocket prairie sites, respectively. Prairie Treatment Plant type Species Common Name Year seeded Low diversity Forb Eupatorium purpureum Sweet Joe pye 2014 Low diversity Forb Lobelia siphilitica Great blue lobelia 2014 Low diversity Forb Ratibida pinnata Grey headed coneflower 2014 Low diversity Forb Zizia aurea Golden Alexander 2014 Low diversity Grass Elymus canadensis Canada wildrye 2014 Low diversity Grass Schizachyrium scoparium Little bluestem 2014 Low diversity Grass Sorghastrum nutans Indian grass 2014 Low diversity Forb Monarda citriodora Lemon mint 2016 Low diversity Forb Rudbeckia hirta Black-eyed Susan 2016 High diversity Forb Aster novae angliae New England aster 2014 High diversity Forb Coreopsis lanceolata Lance leaf coreopsis 2014 High diversity Forb Eryngium yuccifolium Rattlesnake master 2014 High diversity Forb Eupatorium purpureum Sweet Joe pye 2014 High diversity Forb Liatris spicata Spiked blazing star 2014 High diversity Forb Lobelia siphilitica Great blue lobelia 2014 High diversity Forb Monarda fistulosa Wild bergamot 2014 High diversity Forb Penstemon digitalis Foxglove beardtongue 2014 High diversity Forb Ratibida pinnata Grey headed coneflower 2014 High diversity Forb Silphium perfoliatum Cup plant 2014 High diversity Forb Silphium terebinthinaceum Prairie dock 2014 High diversity Forb Solidago riddellii Ridell's goldenrod 2014 High diversity Forb Tradescantia ohiensis Ohio spiderwort 2014 High diversity Forb Verbena hastata Blue vervain 2014 High diversity Forb Vernonia fasciculata Common ironweed 2014 High diversity Forb Zizia aurea Golden Alexander 2014 High diversity Grass Elymus canadensis Canada wildrye 2014 High diversity Grass Schizachyrium scoparium Little bluestem 2014 High diversity Grass Sorghastrum nutans Indian grass 2014 High diversity Forb Bidens aristosa Tickseed sunflower 2016 High diversity Forb Chamaecrista fasciculata Partridge pea 2016 High diversity Forb Coreopsis tinctoria Plains coreopsis 2016 High diversity Forb Gaillardia pulchella Indian blanket 2016 High diversity Forb Monarda citriodora Lemon mint 2016 High diversity Forb Rudbeckia hirta Black-eyed Susan 2016

126

Appendix B: Spider functional traits and mean body size (mm)

127

Spider functional traits and mean body size (adapted from Table S1 in Cardoso et al. 2011; doi:10.1371/journal.pone.0021710).

Family (Subfamily) Genus Hunter SheetSpace OrbWeb Stenophagous Euryphagous Ground Vegetation Diurnal Nocturnal BodySize Agelenidae Agelenidae sp. 0 1 0 0 1 1 1 1 1 10.9 Agelenidae Agelenopsis 0 1 0 0 1 1 1 1 1 10.85 Agelenidae Tegenaria 0 1 0 0 1 1 1 1 1 11 Araneidae Araneidae sp. 0 0 1 0 1 0 1 1 1 1 Cheiracanthiidae Cheiracanthium 1 0 0 0 1 0 1 0 1 6.85 Clubionidae Clubiona 1 0 0 0 1 0 1 0 1 5.6 Clubionidae Clubionidae sp. 1 0 0 0 1 0 1 0 1 5.6 Corinnidae Castianeira 1 0 0 0 1 1 0 1 1 6 Corinnidae Phrurotimpus 1 0 0 0 1 1 0 1 1 2.85 Corinnidae Scotinella 1 0 0 0 1 1 0 1 1 2.3 Corinnidae Trachelas 1 0 0 0 1 1 0 1 1 6.55 Dictynidae Dictyna 0 1 0 0 1 1 1 1 1 3.05 Dysderidae Dysdera 1 0 0 1 0 1 0 0 1 12.5 Gnaphosidae Drassyllus 1 0 0 0 1 1 0 1 1 5.1 Gnaphosidae Sergiolus 1 0 0 0 1 1 0 1 1 5.9 Gnaphosidae Zelotes 1 0 0 0 1 1 0 1 1 5.9 Hahniidae Neoantistea 0 1 0 0 1 1 0 1 1 3.1 Linyphiidae Linyphiidae sp. 1 1 0 0 1 1 1 1 1 2.2 Linyphiidae (Erigoninae) Ceraticelus 1 0 0 0 1 1 0 1 1 1.4 Linyphiidae (Erigoninae) Ceratinella 1 0 0 0 1 1 1 1 1 1.75 Linyphiidae (Erigoninae) Ceratinops 1 0 0 0 1 1 1 1 1 1.85 Linyphiidae (Erigoninae) Ceratinopsis 1 0 0 0 1 1 1 1 1 1.6 Linyphiidae (Erigoninae) Eridantes 1 0 0 0 1 1 1 1 1 1.6

128

Linyphiidae (Erigoninae) Erigone 1 0 0 0 1 1 1 1 1 2.1 Linyphiidae (Erigoninae) Grammonota 1 0 0 0 1 1 1 1 1 2.85 Linyphiidae (Erigoninae) Islandiana 1 0 0 0 1 1 1 1 1 1.4 Linyphiidae (Erigoninae) Mermessus 1 0 0 0 1 1 1 1 1 1.7 Linyphiidae (Erigoninae) Walckenaeria 1 0 0 0 1 1 1 1 1 2.4 Linyphiidae (Linyphiinae) Agyneta 0 1 0 0 1 1 1 1 1 2.15 Linyphiidae (Linyphiinae) Bathyphantes 0 1 0 0 1 1 1 1 1 2.35 Linyphiidae (Linyphiinae) Diplostyla 0 1 0 0 1 1 1 1 1 2.4 Linyphiidae (Linyphiinae) Lepthyphantes 0 1 0 0 1 1 1 1 1 2.5 Linyphiidae (Linyphiinae) Neriene 0 1 0 0 1 1 1 1 1 4.4 Linyphiidae (Linyphiinae) Tennesseellum 0 1 0 0 1 1 1 1 1 2.1 Linyphiidae (Linyphiinae) Tenuiphantes 0 1 0 0 1 1 1 1 1 2.75 Lycosidae Allocosa 1 0 0 0 1 1 0 1 1 4.55 Lycosidae Lycosidae sp. 1 0 0 0 1 1 0 1 1 6.85 Lycosidae Pardosa 1 0 0 0 1 1 0 1 1 5.5 Lycosidae Piratula 1 0 0 0 1 1 0 1 1 3.1 Lycosidae Rabidosa 1 0 0 0 1 1 0 1 1 18 Lycosidae Schizocosa 1 0 0 0 1 1 0 1 1 9.75 Lycosidae Tigrosa 1 0 0 0 1 1 0 1 1 15 Lycosidae Trochosa 1 0 0 0 1 1 0 1 1 10.5 Oxyopidae Oxyopes 1 0 0 0 1 0 1 1 1 10.15 Philodromidae Philodromus 1 0 0 0 1 1 1 1 1 5.25 Salticidae Eris 1 0 0 0 1 1 1 1 0 6.35 Salticidae Habronattus 1 0 0 0 1 1 1 1 0 5.5 Salticidae Marpissa 1 0 0 0 1 1 1 1 0 7.75

129

Salticidae Myrmarachne 1 0 0 0 1 1 1 1 0 6 Salticidae Neon 1 0 0 0 1 1 1 1 0 2.5 Salticidae Salticidae sp. 1 0 0 0 1 1 1 1 0 4.75 Salticidae Sitticus 1 0 0 0 1 1 1 1 0 4.55 Salticidae Synageles 1 0 0 0 1 1 1 1 0 3 Salticidae Talavera 1 0 0 0 1 1 1 1 0 2.35 Tetragnathidae Glenognatha 0 0 1 0 1 0 1 1 1 1.6 Tetragnathidae Pachygnatha 0 0 1 0 1 0 1 1 1 6.5 Tetragnathidae Tetragnatha 0 0 1 0 1 0 1 1 1 8.4 Theridiidae Asagena 0 1 0 0 1 1 1 1 1 5.5 Theridiidae Parasteatoda 0 1 0 0 1 1 1 1 1 4.25 Theridiidae Theridiidae sp. 0 1 0 0 1 1 1 1 1 4.4 Theridiidae Theridion 0 1 0 0 1 1 1 1 1 3.4 Thomisidae Xysticus 1 0 0 0 1 1 1 1 1 5.9 Uloboridae Uloboridae sp. 0 0 1 0 1 0 1 1 1 3.5 Uloboridae Uloborus 0 0 1 0 1 0 1 1 1 3.5 Zodariidae Zodarion 1 0 0 1 0 1 0 1 1 3.5

130

Appendix C: Spiders sampled in Cleveland, Ohio in 2015 and 2016.

131

2015 2016 2015-2016 Vacant Urban Low-Div High-Div Vacant Urban Low-Div High-Div Grand Total Total Family Genus or Species Lots Meadows Prairies Prairies Lots Meadows Prairies Prairies Total

Agelenidae 5 2 7 4 4 5 13 20

Agelenidae sp. 1 1 2 2

Agelenopsis sp. 2 1 3 2 4 3 9 12

Tegenaria sp. 3 1 4 1 1 2 6

Araneidae 1 1 2 1 1 3

Araneidae sp. 1 1 2 1 1 3 Clubionidae 6 3 6 14 29 2 8 8 12 30 59 Clubiona sp. 6 3 6 12 27 2 8 8 12 30 57

Clubionidae sp. 2 2 2

Corinnidae 2 1 3 4 4 6 14 17

Castianeira sp. 2 1 3 4 4 6 14 17

Dictynidae 1 1 1

Dictyna sp. 1 1 1

Dysderidae 1 2 3 11 16 11 18 56 59

Dysdera crocata 1 2 3 11 16 11 18 56 59

Eutichuridae 1 2 1 4 4

Cheiracanthium sp. 1 2 1 4 4

Gnaphosidae 3 1 4 2 3 5 9

Drassyllus sp. 1 1 1 1 2

Sergiolus sp. 1 1 1 1 2

Zelotes sp. 1 1 2 2 1 3 5

Hahniidae 1 1 1

Neoantistea sp. 1 1 1 Linyphiidae 769 374 555 384 2082 566 325 407 299 1597 3679

Agyneta fabra 1 1 1 4 5 10 11

132

Agyneta micaria 3 7 9 5 24 4 2 6 30 Agyneta unimaculata 4 5 2 1 12 1 4 1 8 14 26

Agyneta sp. 3 3 3

Bathyphantes pallidus 2 2 2

Bathyphantes sp. 1 1 2 1 1 3 Ceraticelus similis 31 7 11 17 66 29 13 6 14 62 128

Ceratinella buna 1 1 1 Ceratinops crenatus 3 5 1 5 14 4 7 4 4 19 33 Ceratinopsis laticeps 38 20 12 28 98 57 21 27 42 147 245

Ceratinopsis sp. 1 1 2 2 1 1 4 6

Diplostyla concolor 3 6 2 11 4 32 8 5 49 60 Eridantes erigonoides 7 3 5 1 16 2 3 1 1 7 23

Erigone atra 6 1 10 2 19 5 1 6 25 Erigone autumnalis 69 40 72 46 227 40 30 33 27 130 357

Erigone blaesa 3 1 33 37 7 2 2 11 48

Erigone dentigera 4 2 6 2 1 1 4 10

Erigone dentosa 37 17 31 43 128 26 18 16 60 188

Erigone sp. 1 1 5 5 12 12 Grammonota gentilis 82 6 73 6 167 48 14 26 9 97 264 Grammonota inornata 298 75 144 120 637 161 70 111 51 393 1030

Islandiana flaveola 2 2 1 5 1 1 6

Lepthyphantes sp. 2 2 2 Linyphiidae sp. 34 32 23 20 109 2 7 2 4 15 124

Mermessus trilobatus 16 6 3 25 6 7 4 1 18 43

Neriene clathrata 1 11 3 15 3 10 2 8 23 38 Tennesseellum formica 30 2 68 54 154 121 6 48 17 192 346

Tenuiphantes sp. 5 9 14 14 Tenuiphantes tenuis 60 102 19 16 197 35 87 98 68 288 485 Walckenaeria spiralis 36 17 13 11 77 13 7 6 13 39 116 133

Lycosidae 1651 710 884 1745 4990 1141 1196 1383 1900 5620 10610 Allocosa funerea 74 22 56 136 288 71 48 28 65 212 500

Lycosidae sp. 1 3 4 1 1 5 Pardosa milvina 1195 365 686 1298 3544 759 800 997 1424 3980 7524

Pardosa pauxilla 5 5 5

Pardosa sp. 24 1 25 25

Piratula minuta 1 2 1 4 1 10 2 3 16 20

Rabidosa rabida 4 1 5 2 1 3 8

Schizocosa avida 7 6 13 13

Schizocosa bilineata 1 1 10 10 11

Schizocosa sp. 1 1 1 1 2

Tigrosa helluo 2 2 2 Trochosa ruricola 377 316 137 310 1140 277 313 347 404 1341 2481

Trochosa sp. 1 1 4 5 2 2 13 14

Oxyopidae 4 1 5 5

Oxyopes sp. 4 1 5 5

Philodromidae 1 1 1

Philodromus sp. 1 1 1

Phrurolithidae 1 1 4 2 2 5 13 14

Phrurotimpus sp. 1 1 1

Scotinella sp. 1 1 4 1 2 5 12 13 Salticidae 2 3 3 4 12 1 11 1 10 23 35

Eris sp. 2 2 2

Habronattus sp. 2 1 3 1 3 4 7

Marpissa sp. 1 1 1

Myrmarachne sp. 1 1 2 1 5 5 11 13

Neon sp. 2 2 2

Salticidae sp. 1 1 2 1 1 3

Sitticus sp. 1 3 4 4 134

Synageles sp. 1 1 1

Talavera sp. 1 1 1 1 2 Tetragnathidae 112 53 121 169 455 64 36 43 19 162 617 Glenognatha sp. 83 24 105 140 352 51 25 30 13 119 471 Pachygnatha sp. 29 26 14 26 95 12 11 13 6 42 137

Tetragnatha sp. 3 2 3 8 1 1 9

Theridiidae 1 2 5 5 13 1 3 6 10 23

Asagena sp. 1 1 1

Parasteatoda sp. 1 3 4 4 4 8

Parasteatoda tabulata 1 1 1

Theridiidae sp. 1 2 3 1 7 2 2 9

Theridion sp. 1 1 1 2 3 4 Thomisidae 88 82 84 76 330 115 324 290 287 1016 1346 Xysticus sp. 88 82 84 76 330 115 324 290 287 1016 1346

Trachelidae 1 1 2 1 1 3 5 7

Trachelas sp. 1 1 2 1 1 3 5 7

Uloboridae 1 1 1 1 2

Uloboridae sp. 1 1 1

Uloborus sp. 1 1 1

Zodariidae 1 5 2 8 4 15 10 29 37

Zodarion sp. 1 5 2 8 4 15 10 29 37

0 Family richness 11 10 13 12 17 12 15 18 13 20 21 Genus richness 30 30 33 29 47 28 39 33 35 51 58 Linyphiidae species richness 20 19 20 16 22 18 17 19 20 22 24 Lycosidae species richness 5 5 5 4 7 5 7 5 5 8 9

Total spider abundance 2635 1238 1666 2404 7943 1912 1950 2176 2568 8606 16549

135

Appendix D: Prey taxa that tested positive in the guts of Pardosa milvina using DNA

metabarcoding

136

Order Family Vacant Lot Pocket Prairie Coleoptera 23 28 (Beetles) 1 Cerambycidae 0 1 2 Chrysomelidae 4 2 3 Dermestidae 1 0 4 Dytiscidae 0 1 5 Latridiidae 0 1 6 Leiodidae 18 25 7 Phalacridae 2 0 8 Scarabaeidae 0 1 9 Staphylinidae 1 3 Decapoda 10 8 (Crayfishes) 10 Cambaridae 10 8 Diptera 63 50 (Flies & Moquitos) 11 Anthomyiidae 4 1 12 Asilidae 0 2 13 Calliphoridae 0 1 14 Cecidomyiidae 1 2 15 Ceratopogonidae 0 1 16 Chironomidae 1 1 17 Chloropidae 8 2 18 Culicidae 33 33 19 Dolichopodidae 25 18 20 Drosophilidae 0 3 21 Empididae 1 0 22 Fanniidae 0 1 23 Hybotidae 1 3 24 Lonchopteridae 2 0 25 Phoridae 3 2 137

26 Rhiniidae 0 1 27 Sarcophagidae 1 0 28 Sciaridae 1 3 29 Sepsidae 1 0 30 Syrphidae 12 4 Entomobryomorpha 31 33 (Springtails) 31 Entomobryidae 31 33 32 Isotomidae 1 0 33 Tomoceridae 1 0 Ephemeroptera 0 1 (Mayflies) 34 Baetidae 0 1 Hemiptera 37 31 (Bugs) 35 Anthocoridae 3 7 36 Cicadellidae 29 11 37 Cydnidae 1 0 38 Delphacidae 4 4 39 Flatidae 1 0 40 Halimococcidae 1 1 41 Miridae 8 9 42 Nabidae 0 1 43 Pentatomidae 0 2

7 4 Hymenoptera (Wasps & Ants) 44 Ceraphronidae 0 1 45 Eucharitidae 1 0 46 Formicidae 6 3 Isopoda 0 2 (Woodlice) 47 Armadillidiidae 0 2 Lepidoptera 4 2 (Butterflies & Moths) 48 Coleophoridae 1 0 138

49 Crambidae 0 1 50 Hesperiidae 2 0 51 Lycaenidae 1 0 52 Noctuidae 0 1 53 Nymphalidae 2 1 54 Saturniidae 0 1 55 Sphingidae 0 1 Mesostigmata 1 0 (Mites) 56 Phytoseiidae 1 0 Oribatida 0 3 (Mites) 57 Ceratozetidae 0 2 58 Scutoverticidae 0 1 Orthoptera 0 1 (Crickets) 59 Gryllidae 0 1 Protura 1 0 60 Fujientomidae 1 0 Psocoptera 1 0 (Bark lice) 61 Ectopsocidae 1 0 Symphypleona 1 0 (Springtails) 62 Bourletiellidae 1 0 63 Katiannidae 1 0 Thysanoptera 1 3 (Thrips) 64 Aeolothripidae 1 3 Trichoptera 0 1 (Caddisflies) 65 Anomalopsychidae 0 1 Trombidiformes 2 0 (Mites) 66 Bdellidae 1 0 67 Demodicidae 1 0 Total 231 208

139