FACTORS INFLUENCING BEE COMMUNITIES AND POLLINATION SERVICES ACROSS AN URBAN ENVIRONMENT
Justin D. Burdine
A Dissertation
Submitted to the Graduate College of Bowling Green State University in partial fulfillment of the requirements for the degree of
DOCTOR OF PHILOSOPHY
May 2019
Committee:
Kevin McCluney, Advisor
Mary-Jon Ludy Graduate Faculty Representative
Andrew Gregory
Helen Michaels
Shannon Pelini ii ABSTRACT
Kevin McCluney, Advisor
Current declines in the abundance and diversity of bees and other pollinators has created uncertainty in their ability to reliably deliver pollination services. Recent studies examining urban bee communities show that bees respond to urbanization-mediated changes in land-use and environmental conditions. This includes increases in thermal and desiccation threats via urban heat island (UHI) effects that have not been well explored in bees. But it is unclear whether or how urbanization-related changes to pollinators influence pollination services.
In this dissertation, I surveyed urban gardens and city parks across the metropolitan region of Toledo, Ohio (USA). First, I examined thermal and desiccation tolerances and safety margins for three bee species: silky striped sweat bees (Agapostemon sericeus), western honey bees (Apis mellifera), and common eastern bumble bees (Bombus impatiens). Second, I examined how urbanization and local habitat characteristics (herbaceous cover, floral abundance and color, tree abundance, canopy cover, soil moisture, gardens size) influenced bee communities
(abundance, diversity, composition) and pollination services (visitation frequency). Third, I examined how bee species with specific functional traits and combinations of traits (functional guilds) were influenced by urbanization.
The findings from this dissertation suggest that bees have differential sensitivities to urbanization, and managing for diverse bee communities in urban environments may require mitigating changes in temperature and water and increasing floral resource availability.
iii
To my wife.
iv ACKNOWLEDGMENTS
The first people I want to thank are the members of committee: Kevin McCluney,
Shannon Pelini, Helen Michaels, Andrew Gregory, and Mary-Jon Ludy. They had a significant impact on my research design and methodology, and they shaped me into the scientist that I am today. My committee members also demonstrated how to find success in academia, and provided me with invaluable advice on work-life balance. My advisor, Kevin McCluney, was a great mentor and motivator that influenced my approach to teaching, research, and mentoring.
Many individuals in the McCluney Lab made significant contributions to this dissertation:
Jamie Becker, Margaret Duffy, Edward Lagucki, Melanie Marshall, Gabriella Metzner, Kaleigh
Obrock, Rachel Paul, Melissa Seidel, Erin Plummer, and John Woloschuk. Many of you assisted me on important research projects, and spent hours sitting in front of a microscope looking at bees. Others listened to practice talks, proofread manuscripts and grant proposals, and helped me design conference posters. We shared meals and coffee breaks, and you all reassured me that I was not the only person who had no idea what they were doing. Thank you for doing life beside me as a graduate (or undergraduate) student.
There were also many organizations and land owners that gave me permission to conduct research on their properties in Lucas and Wood County. I want to thank the Olander Park
System, Wood County Parks System, The Nature Conservancy, and the Toledo Zoo. Multifaith
Grows and Toledo Grows helped connect me with urban gardeners throughout the region, and I enjoyed getting to know each of the gardeners I worked with. I also want to thank all the gardeners for giving me your extra fruits, vegetables, and eggs. They were delicious.
v Lastly, I want to think my wife and daughter for being supportive and loving. You both helped me enjoy each step in the dissertation process, and gave me a life outside of academia. So much life happened during these years and I am glad I had you two by my side.
vi
TABLE OF CONTENTS
Page
CHAPTER I: DIFFERENTIAL SENSITIVITY OF BEES TO URBANIZATION-DRIVEN
CHANGES IN BODY TEMPERATURE AND WATER CONTENT ...... 1
Introduction ...... 1
Methods...... 5
Study Area ...... 5
Study Organisms ...... 6
Field Body Temperature and Water Content ...... 6
Thermal and Water Content Limits Sampling ...... 7
Thermal Tolerance Experiment ...... 7
Desiccation Tolerance Experiment ...... 8
Thermal and Hygric Safety Margins...... 9
Statistical Methods ...... 9
Results ...... 10
Critical Thermal Maximum (CTmax) ...... 10
Critical Water Content (CWC) ...... 10
Thermal Safety Margin ...... 11
Hygric Safety Margin ...... 12
Discussion ...... 12
Thermal Tolerance ...... 13
Desiccation Tolerance ...... 15
Thermal and Hygric Safety Margins...... 15 vii
CHAPTER II: INTERACTIVE EFFECTS OF URBANIZATION AND LOCAL HABITAT
CHARACTERISTICS INFLUENCE BEE COMMUNITIES AND FLOWER
VISITATION RATES ...... 19
Introduction ...... 19
Methods...... 22
Sampling Locations ...... 22
Sampling Methods ...... 22
Local Habitat Characteristics ...... 23
Visitation Rates ...... 23
Statistical Methods ...... 24
Results ...... 25
Summary Statistics...... 25
Community Composition ...... 25
Overall Abundance and Diversity ...... 26
Visitation Rates ...... 26
Discussion ...... 26
Impervious Surface ...... 27
Canopy Cover ...... 28
Flower Abundance ...... 28
Caveats ...... 29
Conclusions ...... 29
CHAPTER III: BEE GUILDS SHOW DISSIMILAR RESPONSES TO URBANIZATION IN A
MEDIUM-SIZED CITY, ALTERING POLLINATOR VISITATION ...... 31 viii
Introduction ...... 31
Methods...... 33
Sampling Locations ...... 33
Sampling Methods ...... 33
Habitat Characteristics ...... 34
Visitation Rates ...... 35
Functional Traits ...... 35
Functional Guilds ...... 36
Statistical Methods ...... 36
Results ...... 37
Functional Trait Diversity and Guild Richness ...... 37
Guild Composition ...... 38
Guild Specific Results...... 38
Visitation Rates ...... 39
Discussion ...... 39
Conclusion ...... 42
REFERENCES ...... 43
APPENDIX A: TABLES ...... 61
APPENDIX B: FIGURES ...... 78
APPENDIX C: CHAPTER I SUPPLEMENTARY INFORMATION ...... 90
APPENDIX D: COAUTHOR CONSENT FORM ...... 105
APPENDIX E: COAUTHOR CONSENT FORM ...... 106
APPENDIX F: COAUTHOR CONSENT FORM ...... 107 ix
APPENDIX G: URBANIZATION ALTERS COMMUNITIES OF FLYING ARTHROPODS IN
PARKS AND GARDENS OF A MEDIUM-SIZED CITY ...... 108
1
CHAPTER I: DIFFERENTIAL SENSITIVITY OF BEES TO URBANIZATION-DRIVEN
CHANGES IN BODY TEMPERATURE AND WATER CONTENT1
Introduction
Climate change and land use modification can have negative consequences for many species, leading to local population declines (Walther et al. 2002, Foley et al. 2005, Both et al.
2006) and extinctions (Thomas et al. 2004). When certain taxa decline, the services they provide
(e.g. pollination) can be disturbed or degraded (Kremen et al. 2007, Schweiger et al. 2010).
While examples of population declines with climate change and land-use modification are accumulating (Walther et al. 2002, Foley et al. 2005, Biesmeijer 2006, Both et al. 2006), proximate mechanisms mediating these declines are often unclear because very species-specific patterns occur. Understanding these mechanisms is vital to identifying actions that mitigate the potential losses of ecosystem services. Multiple studies have identified the importance of physiological tolerances in predicting species responses to global change (Walther et al. 2002,
Parmesan 2006, Williams et al. 2008, Scaven and Rafferty 2013, Sunday et al. 2014), as these tolerances demarcate the environmental conditions necessary for survival. However, the majority of studies investigating physiological tolerances focus on thermal tolerances (Pereboom and
Biesmeijer 2003, Verble-Pearson et al. 2015, Oyen et al. 2016), while desiccation tolerances may be just as important.
Thermal tolerances are an important tool for investigating species responses to changes in temperature (Nguyen et al. 2014, Diamond et al. 2017, Hamblin et al. 2017). Critical thermal maximum (CTmax) and minimum (CTmin), and thermal safety margins are the most common metrics of physiological vulnerability to climate change for a variety of organisms (Deutsch et al.
1 This chapter has been published as: Burdine, J.D. and K.E. McCluney. 2019. Differential sensitivity of bees to urbanization-driven changes in body temperature and water content. Scientific Reports 9: 1643. 2
2008, Kellermann et al. 2012, Oyen et al. 2016). CTmax is an organism’s upper sub-lethal temperature and CTmin the lower sub-lethal temperature (Terblanche et al. 2011), and these are the temperatures at which an organism loses muscular control and suffers an ecological death.
The difference between CTmax and CTmin is defined as the thermal range (Huey and Stevenson
1979). Thermal safety margin is defined as the differences between CTmax and either optimal body temperature, field body temperature, or air temperature, and offers a metric for understanding vulnerabilities to warming (Sunday et al. 2014). In general, thermal tolerance has been found to vary with natural temperature gradients. Insects from high and low latitudes tend to have similar CTmax values, while CTmin declines with latitude (Deutsch et al. 2008, Araújo et al. 2013) and thermal safety margin increases with latitude (Deutsch et al. 2008). CTmax has been shown to decreases with altitude (Gaston and Chown 1999, Oyen et al. 2016). There is also evidence that insect thermal tolerance varies across smaller climatic gradients (Baudier et al.
2018). Body size (surface-volume ratios) may also influence thermal tolerance, because smaller animals dissipate heat better but may be more prone to desiccation (Willmer and Stone 1997).
Desiccation tolerances may also be an important mediator of climate effects on species
(Sinclair et al. 2003), but few studies have explored desiccation in this context (Addo-Bediako et al. 2001, Hoffmann et al. 2003, Chown et al. 2011, Hoffmann and Sgró 2011), and little information exists about desiccation tolerance of animals in general (Hoffmann and Parsons
1989, Hadley 1994, Gibbs et al. 1997, Block 2003, Chen et al. 2014). Hadley (1994) reported that the majority of arthropods maintain a water content between 65 and 75%, while some arthropods may survive with a water content as low as 40% (Block 2003). We know little about how desiccation tolerance varies with environmental gradients, but Hoffman et al (2003) suggest that fruit fly desiccation tolerance declines with increasing precipitation in Australia. Weldon et 3 al. (2018) provide evidence that desiccation tolerance varies geographically among populations of Mediterranean fruit flies. Moreover, McCluney et al. (2017) found that arthropod hydration decreased with urbanization and consequent warming in moist/mild cities (Raleigh, NC), but sometimes increased with urbanization in warmer cities (Phoenix, AZ; Orlando, FL). The observed variations in desiccation tolerance may be due to local adaptations, or plasticity
(Terblanche and Kleynhans 2009). These studies suggest that desiccation tolerance might be an important predictor of the effects of land use and climate change on animals. Desiccation may be particularly likely for smaller animals, like insects, due to their greater surface area to volume ratios and higher water loss relative to metabolic rate (McCluney 2017).
Multiple metrics of desiccation tolerance have been employed by others. Many studies have used the time until 50% of the animals die (LT50)(Hoffmann et al. 2003, Markow et al.
2007) for both desiccation and thermal tolerance measurements. However, another measure, body water content at death (critical water content, CWC), provides values comparable across studies. CWC is an experimental measurement of desiccation tolerance, or the lethal water content, calculated gravimetrically as the difference between wet and dry mass, divided by wet mass (Schilman et al. 2007). Unlike thermal safety margins, none have calculated a hygric safety margin (hygric = relating to moisture), which we have used in this study and define here as the difference between CWC and field body water content. This metric could be highly informative in predicting potential responses to climate change, complementing thermal analogs.
Urban environments are unique systems for examining how land use modification and climate change influence thermal and desiccation tolerances (Youngsteadt et al. 2015). Summer urban heat islands (UHIs) in the USA typically generate a mosaic of hotter and cooler locations that differ, on average, by 1 – 4°C from each other and form cooler, rural temperatures (Gaffin et 4 al. 2008). The intensity of UHIs is often influenced by the biome in which the city is located.
Urban areas within regions dominated by temperate and mixed forest can experience temperature increases of 8°C, while urban areas within desert regions experience less pronounced changes or temperature decreases in urban centers (Imhoff et al. 2010). UHIs provides a gradient of temperatures (Schueler 1994, Yuan and Bauer 2007) that can replicate projected climate change
(Youngsteadt et al. 2015). Urban areas can also experience altered soil moisture (Shuster et al.
2005), and soil moisture is found to vary among habitat types within and among cities (Groffman et al. 2014). These changes in temperature and moisture availability can impact the field body water content of arthropods (McCluney et al. 2017). A recent study by Hamblin et al. (2017) found CTmax to predict population change for 15 bee species across an urbanization gradient. We know relatively little about bee thermal tolerances (Atmowidjojo et al. 1997, Kovac et al. 2014,
Oyen et al. 2016, Hamblin et al. 2017), and even less about the desiccation tolerances of bees
(Atmowidjojo et al. 1997), even though these factors may provide strong predictive power in explaining population declines or changes in distributions. Elevated temperatures due to UHI effects are likely to increase desiccation threats. Others have found that changes in insect water balance can have consequences on growth, reproduction, and survival (Dale and Frank 2018).
Here we examine how a gradient of urbanization (impervious surface, e.g. areas of pavement), in a medium-sized city, alters both the CTmax and CWC of three bee species: the silky striped sweat bee (Agapostemon sericeus), the western honeybee (Apis mellifera), and the common eastern bumblebee (Bombus impatiens). We combine measurements of thermal and water content limits with measurements of field body temperature and field water content, to quantify thermal and hygric safety margins. We test the relative importance of two competing hypotheses about limits and safety margins. First, thermal and water content limits may change 5 rapidly within taxa due to changes in temperature and moisture conditions, with sensitivity decreasing with urbanization-induced warming and drying. In this case, thermal and hygric safety margins would not vary with urbanization because limits change in concert with field body temperature and water content (“shifting limits hypothesis,” i.e. limits are plastic). Second, thermal and water content limits may change slowly within taxa, being insensitive to urbanization-induced warming and drying. In this case, limits would not change with urbanization, even as average field body temperature or water content did change, and we would expect safety margins to be related to urbanization-induced warming or drying (“stable limits hypothesis,” i.e. limits are non-plastic).
Methods
Study Area. Field body temperature and water content were measured in bees collected from 19 sites, and CTmax and CWC were measured on bees collected from a subset of sites, across the metropolitan region of Toledo, Ohio, USA (Figure 1). Toledo has been shown to display urban heat island effects with 2°C warmer mean annual temperatures (Schmidlin 1989).
Toledo has a 622.6 square-kilometer metropolitan area with a population of 507,643 residents and a large network of greenspaces (125 city parks and 150 urban gardens)(Burdine and Taylor
2017) that were used as sampling locations. Sites were selected by overlaying a grid over a map of Toledo in ArcGIS and numbering each grid cell (2 × 2 km). A random number generator was used to identify grid cells for sampling, and we chose a park or garden within each selected cell.
We quantified imperviousness (land surfaces that prevent water infiltration) for each site using the percent developed imperviousness layer in the 2011 National Land Cover Dataset (Homer et al. 2015) in ArcMap 10.3. We calculated local imperviousness (300 m radius) and landscape 6 imperviousness (2000 m radius), as scale can be important in appropriately measuring the urban landscape for bees (Glaum et al. 2017).
Study Organisms. We focused on the silky striped sweat bee (Agapostemon sericeus), western honeybee (Apis mellifera), and common eastern bumblebee (Bombus impatiens) due to the relative commonness of these species in parks and gardens across Toledo (Pardee and
Philpott 2014). In general, sweat bees begin their activity in early to mid-Spring when temperatures reach 14°C (Richards and Packer 1995), while bumble bees and honey bees are known to forage at much lower temperatures (Goulson 2003). Most research on thermal tolerance in bees is focused on bumble bees and honey bees, and we included sweat bees because little is known of their physiological tolerances, they are abundant and widely distributed, and they are smaller than the other species. These three species differ in size, foraging preference, sociality, and nest specificity (Michener 2007), increasing the likelihood of detecting differential responses among species.
Field Body Temperature and Water Content. We hand collected sweat bees, honey bees, and bumble bees from a series of 19 sites from June-August 2016 using 120 ml specimen containers (QuickMedical Model#4256). Bees were sampled in the mornings on days with clear skies and temperatures above 22°C. Upon capture, bees were anesthetized with CO2 to estimate body temperature using a thermocouple (Atkins 351/352) held between the thorax and abdomen for approximately 30 s. Surface temperatures of bees and other insects, at rest, have been found to be similar to internal body temperatures (within 1°C)(Casey 1976, Stabentheiner and
Schmaranzer 1987). We then transported bees back to the lab in airtight vials (Pelco® Mini
Vials) to prevent water loss, and placed the vials in a freezer for 24 hours prior to water content determination. Bee water content was calculated gravimetrically using an analytical balance 7
(Mettler Toledo XPE56) with precision to one microgram. We calculated the water content of each bee as the difference in wet and dry mass divided by wet mass. To calculate dry mass, bees were placed into a drying oven (Thermo Fisher Scientific #151030520) set at 55°C for at least 48 hours, and then reweighed (McCluney et al. 2017). We obtained estimates of water content from all 19 sites, but our temperature measurements are from only 13 sites. We were unable to take body temperature measurements at all 19 sites due to equipment availability.
Thermal and Water Content Limits Sampling. To measure thermal and water content limits, we focused on 6 sites (3 urban, 3 rural) in the metropolitan region of Toledo (Figure 1), that were selected based on the percent impervious surface within a 300 m radius of the site center (see Figure 1). Urban sites had an average local impervious surface (300m) of 57.79% ±
9.6, while rural sites had an average local impervious surface of 23.24% ± 4.5. For each experiment, we collected 15 bees from each site of the same species identified above (five sweat bees, five honey bees, five bumble bees). At each site, we walked a linear transect starting at the site’s center and collected the first five bees of each species we encountered using 120 ml specimen containers. Upon collection, bees were placed in a cooler box without ice (25°C) to standardize temperatures prior to each experiment, and immediately taken to Bowling Green
State University. Bees were given a 1 M sucrose solution upon capture to ensure bee survival during transportation prior to thermal ramping and desiccation trials, which was especially important to ensure survival of honey bees (Kovac et al. 2014). All experimental trials began within two hours of collection. The collection period for bees in our thermal and hygric experiments occurred during September 6-22, 2016.
Thermal Tolerance Experiment. We measured critical thermal maximum (CTmax) for each sampled bee with a temperature ramping experiment using an environmental chamber (Memmert 8
HPP 750). The temperature ramp began at 25°C and increased at a rate of 0.5°C min-1, following standard methods (Terblanche et al. 2011). Humidity was held constant at 30%. We used iButton® temperature and humidity data loggers (DS1923) to verify the conditions of the environmental chamber. Bees were placed individually into a 120 ml specimen cup
(QuickMedical Model#4256) with a mesh cover that allowed air temperatures within specimen cups to increase with the temperature ramp. Additionally, the mesh top allowed us to easily test the righting response using a puff of air (Giottos AA1900). The loss of righting response indicates an endpoint when muscle functions begin to fail, and is commonly used to estimate
CTmax (Verble-Pearson et al. 2015, Oyen et al. 2016). Bees unable to flip themselves upright within 15 seconds after receiving a puff of air were considered to have lost the righting response.
The temperature at which the righting response was lost was recorded as the CTmax, and bees were removed from the environmental chamber after this point. The temperature ramp ended when all 90 bees had reached their CTmax (~2 hrs). All samples were immediately weighed and stored in air-tight vials after the temperature ramp ended. Bees were not given food or water during the temperature ramping period.
Desiccation Tolerance Experiment. We measured critical water content (CWC) by placing a different set of 90 bees in a desiccation cabinet at 0% humidity and a constant temperature of 25°C (well below CTmax). Humidity was maintained at 0% using Drierite
(#60011T), and was verified using an iButton® temperature and humidity logger (DS1923).
Bees were placed individually into the same 120 ml specimen cups used in the CTmax experiment and checked on an incremental timescale as follows: (1) every 15 minutes, (2) every 30 minutes,
(3) every 1 hour, (4) every 3 hours, (5) every 6 hours, and (6) every 12 hours. We performed three checks of each time period. Each check lasted approximately five minutes, and we tested 9 the bees’ righting response and removed unresponsive bees. Immediately after removal from the desiccation cabinet, bees were placed into pre-weighed Pelco® Minivials for water content determination.
Thermal and Hygric Safety Margins. We calculated thermal and hygric safety margins as measures of the proximity of bees to limits (Sunday et al. 2014). Thermal safety margins were
th calculated as the difference in the measured field body temperature and the 90 percentile CTmax.
Hygric safety margins were calculated similarly as the difference in the measured field water content and the 10th percentile CWC. Using 90th and 10th percentiles, respectively, prevented us from using abnormally tolerant individuals to estimate safety margins, while simultaneously preventing negative safety margins (see Mccluney et al. 2017) that might have arisen had we used mean CTmax or CWC values.
Statistical Methods. All statistical tests were conducted in the R statistical environment
(version 3.1.3). We used the nlme package (Pinheiro et al. 2014) to fit linear mixed effects models (lme), comparing CTmax and CWC among bee species (A. sericeus, A. mellifera, B. impatiens) and site class (rural, urban), and the interaction effect between species and site class.
We included site ID as a random effect to account for inclusion of multiple bees per site. Post- hoc Tukey multiple comparison of means (glht) tests, in the multcomp package (Hothorn et al.
2009), were used following significant main effects.
Linear mixed effects models were also used to examine relationships between thermal and hygric safety margins and our urbanization metrics (local imperviousness, landscape imperviousness), with site as a random effect to account for inclusion of multiple bees per site.
We also used linear mixed effects models to test for relationships between safety margins and bee mass, including site and bee species as random effects. We tested for the significance of 10 fixed effects via likelihood ratio tests with removal of the fixed effect from the model (Bolker et al. 2009). Significance was indicated and discussed at α = 0.05 for a test of the core hypotheses, and at α = 0.1 as a threshold for patterns that could be further explored in the future.
Assumptions of normality were checked via examination of plots of residuals, and measures of hygric safety margins were logit transformed.
Results
Critical Thermal Maximum (CTmax). We measured CTmax for 88 bees (30 sweat bees, 30 honey bees, 28 bumble bees; S5). Our measurements varied significantly between species (df =
2, χ2 = 6.56, p = 0.038; Figure 2A). A post-hoc test revealed that bumble bees had significantly higher CTmax than honey bees and sweat bees, but there was no difference between honey bees and sweat bees. Although we did not detect a significant difference in CTmax between urban and rural sites at α = 0.05 (df = 1, χ2 = 3.84, p = 0.05; Figure 3; S1), a strong trend was apparent in mean CTmax between urban (52.0°C ± 0.76) and rural sites (49.6°C ± 1.1). We found no difference in the slope of the relationship between urbanization and CTmax between species
(Urbanization x Species: df = 2, χ2 = 0.745, p = 0.69). We also found no significant relationship
2 2 between CTmax and bee mass (df = 1, χ = 2.357, p = 0.125, R = 0.02). Bee mass did not differ between urban and rural sites (df = 1, χ2 = 0.488, p = 0.485), but we did find significant differences in mass between species as expected (df = 2, χ2 = 139.8, p < 0.001). Bumble bees had the largest mass (44.97 mg ± 2.81), followed by honey bees (31.39 mg ± 0.46), and sweat bees
(4.00 mg ± 0.44).
Critical Water Content (CWC). We measured CWC for a different set of 89 bees (30 sweat bees, 30 honey bees, 29 bumble bees; S6), and our measurements varied significantly between species (df = 2, χ2 = 77.81, p < 0.001; Figure 2B). A post-hoc multiple comparisons test 11 showed significant differences in CWC between all bee species. Sweat bees had the lowest CWC
(highest tolerance), and honey bees the highest CWC (lowest tolerance). However, we did not detect a significant difference in CWC between bees from urban and rural sites (df = 1, χ2 = 3.26, p = 0.071), and we found no interaction between species and urbanization (df = 2, χ2 = 2.44, p =
0.29). We also found no relationship between CWC and bee mass (df = 1, χ2 = 1.94, p = 0.16; R2
< 0.001).
We also found significant differences between species in the length of time they survived in the desiccation cabinet (df = 2, χ2 = 9.87, p = 0.007). A post-hoc multiple comparisons test found sweat bees (39.5 hours ± 8.03) to have longer survival times than honey bees (12.3 hours ±
6.85) (p = 0.005), and statistically similar survival times to bumble bees (19.5 hours ± 2.65) (p =
0.057). We found no differences in survival times between honey bees and bumble bees (p =
0.70). We also found no differences in survival times between bees collected from urban and rural sites (df = 1, χ2 = 0.71, p = 0.40).
Thermal Safety Margin. We calculated thermal safety margins for 52 bees (22 bumble bees, 21 honey bees, 9 sweat bees; S7) taken from 13 sites that varied in the amount of impervious surface present around our collection sites (our metric of urbanization) at the local
(300m radius) and landscape (2 km radius) scale. Thermal safety margins declined with increasing impervious surface, but bee species differed in the strength of this response (Table 1;
S2). Overall, bumble bees (29.38°C ± 0.38) had the largest thermal safety margin followed by honey bees (26.95°C ± 0.37) and sweat bees (24.91°C ± 0.47). We found thermal safety margin declined with local (300m) imperviousness for both bumble bees (Figure 4D) and sweat bees
(Figure 4C), but not for landscape imperviousness. For honey bees, thermal safety margin was 12 not associated with imperviousness at either scale. Overall, we found no relationship between thermal safety margin and bee mass (df = 1, χ2 = 1.93, p = 0.165, R2 = 0.02).
Hygric Safety Margin. We calculated hygric safety margins for 88 bees (42 bumble bee,
24 sweat bees, 22 honey bees; S8) taken from 19 sites that varied in the amount of impervious surface present around our collection sites (our metric of urbanization). Hygric safety margin varied interactively with imperviousness and bee species at the landscape level (df = 2, χ2 =
9.014, p=0.01; Table 1). Sweat bees had the largest hygric safety margin (26.15% ± 1.22), followed by bumble bees (5.25% ± 0.46) and honey bees (2.50% ± 0.45). For honey bees, hygric safety margin declined with landscape imperviousness at α = 0.05 (df = 1, χ2 = 8.847, p = 0.003;
Figure 4A; S4), with a similar trend for local imperviousness at α = 0.1 (df = 1, χ2 = 2.81, p =
0.094; Figure 4B; S3). For sweat bees and bumble bees, we found no associations between imperviousness and hygric safety margin. Overall, we found no significant relationship between hygric safety margin and bee mass (df = 1, χ2 = 0.06, p = 0.807; R2 < 0.001). Field water content did not differ between urban and rural sites (df = 1, χ2 = -311.34, p = 0.91), but we did find significant differences in field water content between species (df = 2, χ2 = -253.04, p < 0.001).
Sweat bees had the lowest field water content (59.3% ± 1.2), followed by bumble bees (65.6% ±
0.4), and honey bees (70.1 % ± 0.46).
Discussion
Overall, we found that bee species differ in their thermal limits and water content limits, and their proximity to those limits, with certain species (honey bees) being more susceptible to desiccation than to extreme temperatures. This points out the need to better investigate desiccation, along with temperature, as potential mechanisms underlying the biological effects of global change. We also found that all three species were closer to their physiological limits 13
(either thermal or hygric) with increasing urbanization (local or landscape imperviousness), and that thermal and hygric limits did not vary significantly with urbanization, supporting the “stable limits hypothesis”.
The proximity of honey bees to their water content limits in highly urbanized areas likely represents a high risk for population declines during a drought or heat wave (due to increased water loss rates). However, given that bumble bees and sweat bees were not close to their thermal limits, and that physiological performance often increases as an animal approaches its thermal limits (Deutsch et al. 2008, Angilletta 2009), the increasing proximity to thermal limits for these two species may actually represent a benefit rather than an increased risk. Essentially, warming related to urbanization may benefit bumble bees and sweat bees in cool cities like
Toledo, Ohio (<15 days per year above 32°C), compared to many other US cities. However, we suggest that if increased proximity to thermal limits with urbanization also occurs in warmer cities, urbanization may represent substantial increased risk of population declines in those cities.
This hypothesis needs further testing, but is supported by research from a warmer, southeastern
US city, Raleigh, NC, USA, where critical thermal maximum (CTmax) was a strong predictor of species presence with increased urbanization (Hamblin et al. 2017).
Thermal Tolerance. We found no significant differences in CTmax between urban and rural sites. Others have found CTmax to vary with natural temperature gradients, such as elevation
(Oyen et al. 2016) and latitude (Kerr et al. 2015), and multiple studies have identified patterns between CTmax and urbanization (Angilletta et al. 2007, Brans et al. 2017, Diamond et al. 2017,
Hamblin et al. 2017). For example, Diamond et al. (2017) found differences in CTmax between urban and rural acorn ants, and attributed these differences to evolved plasticity to local environmental conditions. Angilletta et al. (2007) found urban ants more heat tolerant than rural 14 ants, but were unable to determine whether the differences were attributed to environmental or genetic effects. These studies seem to support the “shifting limits hypothesis”. We hypothesize that bee movement distances may help explain why we found stronger effects on safety margins than limits. Research suggests that bee body size predicts foraging distance (Greenleaf et al.
2007), and the species used in our study are relatively large bees. Greater mobility may prevent local adaptation or even local acclimation, as mobile species encounter both very urbanized and poorly urbanized landscapes, and have greater ability to evade lethal conditions. In addition, any small amount of local adaptation could be erased by genetic mixing. Results might be different for small bees (i.e. Lasioglossum) with short foraging ranges (sensu local adaptation was observed for acorn ants in Diamond et al. (2017)). In addition, results may be different with a larger sample size, or at sites with more extreme levels of impervious surface as others have found when using rural sites with 0% impervious surface (Diamond et al. 2017).
We note that bumble bees had higher CTmax than both honey bees and sweat bees, independent of urbanization. Others have found CTmax to increase with bee size (Oyen et al.
2016), although the exact mechanism is unclear. We found no significant relationship between bee mass and CTmax, when controlling for species. Furthermore, sweat bees were the smallest bees measured and their CTmax was not significantly different from honey bees. Hamblin et al.
(2017)found sociality to be an important factor in heat tolerance, and our results offer some support for the importance of sociality in heat tolerance. Honey bees are eusocial and thermoregulation has been commonly observed in honeybee colonies (Heinrich 1980, Cooper et al. 1985). While thermoregulation has also been observed for some bumble bee species (Heinrich
1972, Heinrich and Heinrich 1983), we suggest that highly social honey bees may experience less selection pressure for increased CTmax than bumble bees, due to hive thermoregulation. 15
Desiccation Tolerance. Although little research has been done on critical water content
(CWC), the values we collected for bee water content are generally consistent with the reporting of other studies on arthropods (Hadley 1994, Gibbs et al. 1997, Schilman et al. 2007, Chen et al.
2014). While CWC values for sweat bees are consistent with the other arthropod taxa mentioned above, honey bees and bumble bees have much higher CWC. Other research has documented particularly high water loss rates in honey bees (Atmowidjojo et al. 1997), and here we find that they are relatively intolerant of desiccation. It is also interesting to note that honey bees and bumble bees have a tight range of tolerance, suggesting relatively little variation in water content limits in these populations.
Our research suggests key unanswered questions about CWC that future research should investigate. For instance, a better understanding of mechanistic drivers of variation in CWC is needed in general, including for bees. Following earlier work on other arthropods (Hadley 1994), we suggest CWC may be associated with functional traits like cuticular hydrocarbon content, physiological stress response pathways, coloration, or other traits, with a variable degree of genetic influence. Moreover, our work suggests there may be a tradeoff between CWC and
CTmax (bees were either more sensitive to heat or desiccation). But these ideas are in need of further, more explicit testing.
Thermal and Hygric Safety Margins. We found evidence that all species of bees have decreased safety margins with increasing urbanization. At high impervious sites, bumble bees and sweat bees were both closer to their thermal limit, and honey bees were closer to their hygric limit. These results have several significant implications.
First, many studies have shown that urban areas are able to support diverse bee communities (Matteson et al. 2008, Hall et al. 2017), particularly bumble bees (McFrederick and 16
LeBuhn 2006, Winfree et al. 2007, Glaum et al. 2017). The observed positive effects are typically thought to be due to increased floral resources or nest site availability in urbanized regions (Bates et al. 2011, Hennig and Ghazoul 2012, Pardee and Philpott 2014). However, our research suggests that associations between urbanization and bumblebee thermal physiology could be beneficial or harmful depending on the typical regional temperatures, altering thermal safety margins. Although bees function as facultative endotherms in flight, the body temperatures of flying bees increase with environmental warming (Willmer and Stone 1997).
Populations of both bumble bees and sweat bees may benefit from urbanization in cool regions, because higher ambient temperatures bring bees towards optimal operative temperatures. But these same species may be harmed by urbanization in warmer cities where bees are already near optimal operative temperatures. Likewise, in warm cities, these species may have reduced capacity to tolerate further increases in temperature associated with climate change or heat waves. Efforts to increase green spaces and shade within warm cities could help bees thermoregulate (Pereboom and Biesmeijer 2003) and reduce risk. Although we did not investigate CTmin (e.g. cold resistance) in this study, urban warming may lead to higher CTmin values and an overall decrease in thermal tolerance range, as has been shown in ants (Diamond et al. 2017).
Second, we found that honey bees were closer to their hygric limits with increasing urbanization, causing them to be on the edge of their desiccation tolerance. Although bees drink a nectar diet and produce metabolic water, they can regularly become dehydrated in the field
(Willmer and Stone 1997). Water drinking behavior has been observed in honey bees (Visscher et al. 1996) and this is consistent with their low desiccation tolerance. Here we also show that honey bees are unable to maintain their water content with increasing urbanization, even in this 17 relatively cool, mesic city. This suggests that metabolic water production may be insufficient to prevent dehydration, in honey bees. Thus, urban gardeners in any city may want to consider providing water sources for honey bees.
Finally, we note that although bumblebee hygric safety margins were not associated with urbanization, bumble bees had narrow hygric safety margins overall. This suggests that bumble bees are able to maintain their water content with increasing urbanization in this cool, mesic city, but urban bumble bees may be at greater risk in drier or hotter cities or during heatwaves or droughts.
Here, we provide the first measurements of bee CWC and one of only a small number examining bee CTmax (only one other in an urban context). Interesting insights emerge from these investigations. First, we found that species differ in their thermal and desiccation tolerance and in their safety margins. Honey bees may be particularly sensitive to desiccation associated with urbanization and climate change, while sweat bees may be more sensitive to temperature changes. More research is needed to determine whether bumble bees are more sensitive to changes in temperature or water balance. This is because, for bumble bees, we found a decline in thermal safety margins, but not hygric safety margins, with urbanization, but overall hygric safety margins were much narrower than their thermal safety margins, indicating potential susceptibility to changes in water balance.
Overall, we argue that a physiological approach provides important information about how bees may respond to changes in climate and land use. Additionally, our research highlights that water balance should not be ignored. Moreover, if changes in bee thermal and water physiology associated with land use or climate change alters bee communities, there could be significant economic consequences related to altered pollination services. Thus, physiological 18 approaches could prove useful in developing strategies to maintain key ecosystem services under climate and land use change.
19
CHAPTER II: INTERACTIVE EFFECTS OF URBANIZATION AND LOCAL
HABITAT CHARACTERISTICS INFLUENCE BEE COMMUNITIES AND FLOWER
VISITATION RATES
Introduction
The global significance of pollinators has been well established — bees and other pollinating animals provide important pollination services that benefit ~87% of flowering plants
(angiosperms) worldwide (Ollerton et al. 2011), including 1500 agricultural crops (Klein et al.
2007). Pollination can increase the quality, quantity, and stability of agricultural yields (Allen-
Wardell et al. 1998, Ricketts et al. 2008), and the estimated global value of pollination services is
$117 billion annually (Costanza et al. 1997). However, there is strong evidence that wild pollinators are in decline globally (Potts et al. 2010), and anthropogenic modification of natural landscapes via habitat loss, fragmentation, and land use intensification is a primary driver behind these declines (Ricketts et al. 2008, Potts et al. 2010). But evaluating how anthropogenic landscape modification affects pollination can be difficult to assess, because pollination is influenced by a myriad of environmental conditions that vary across spatial scales.
In general, pollination services are strongly associated with the availability of floral and nesting resources. Multiple studies have found positive effects of floral resource availability on pollination (Kells et al. 2001, Blaauw and Isaacs 2014). For instance, Blaauw and Isaacs (2014) found that increased wildflower abundances near crop fields improved pollination and crop yields, and even lead to profit gains. Nesting resource availability (e.g. bare ground, pre-existing cavities) also influences pollination through changes in bee community structure (Potts et al.
2005). Others have found declines in pollination with increased distance from natural areas
(Ricketts et al. 2008). Land use intensification (e.g. urbanization) may also influence pollination, 20 but few have investigated pollination across intensification gradients outside of traditional agricultural systems. Much of our understanding of pollination comes from traditional agriculture, but urban agriculture is an increasingly important sector of the global food supply
(Hodgson et al. 2011).
Currently, over half the global human population lives in urban regions (Pickett et al.
2011), and over 80% of the United States population is considered urban (United Nations 2018).
The amount of terrestrial land classified as urban is expected to triple by 2030 (Seto et al. 2012), transforming rural regions into residences and infrastructure for an increasingly urban human population. The overall impact urbanization has on species and ecosystem services is difficult to assess because there is a great deal of variation in the spatial heterogeneity and development intensity within and among cities (Lin and Fuller 2013), and between shrinking and growing cities (Haase 2008). But a number of studies over the past decade have shown that bee community responses to urbanization are often mediated by local and landscape habitat conditions (Ahrne et al. 2009, Hernandez et al. 2009, Fortel et al. 2014, Quistberg et al. 2016,
Glaum et al. 2017, Hall et al. 2017). Floral resource availability is consistently found to be a strong predictor of bee abundance (Lowenstein et al. 2014, Pardee and Philpott 2014) and diversity (Lowenstein et al. 2014, Pardee and Philpott 2014) in cities. Quistberg et al. (2016) found that larger urban greenspaces harbor more bee individuals and species, and that ground cover (e.g. mulch, leaflitter) influenced the types of bees present (e.g. cavity nesting taxa). But studies have yet to simultaneously consider the effects of urbanization and local habitat features on both diversity and pollination services (but see Potter and LeBuhn 2015).
Despite the importance of pollination in urban greenspaces (e.g. city parks and urban gardens), surprisingly few studies have explicitly measured pollination services in cities 21
(Lowenstein et al. 2014, 2015, Theodorou et al. 2016, Hou et al. 2019). Lowenstein et al. (2014) measured pollination services in residential yards in Chicago (USA), and found a positive correlation of bee abundance and diversity on visitation frequency, but they did not examine drivers of these patterns and did not examine highly urbanized parts of the city. Others have also identified the positive effect bee diversity has on fruit set in pollinator-dependent plants (Kremen et al. 2002), suggesting a direct relationship between bee diversity and pollination. However, the factors that drive bee diversity are not always directly associated with pollination. For instance,
Lowenstein et al. (2014) found a positive association of floral richness with bee diversity, but not pollination (e.g. visitation frequency). Thus, additional studies are needed to investigate how local habitat characteristics influence both bee diversity and pollination across urbanization gradients in cities.
In this study, we investigated how habitat characteristics (herbaceous cover, floral abundance and color, tree abundance, canopy cover, soil moisture, garden size) in city parks and urban gardens influenced the abundance, diversity, and community composition of bees, and the visitation frequency of insect pollinators. We divided this overarching question into two parts: 1)
How does urbanization influence bee communities (abundance, diversity, composition) and pollination services (visitation frequency)? 2) How do local habitat features within urban gardens and city parks modify the effects of urbanization on bee communities and visitation frequencies?
We expected to see changes in bee community composition with urbanization, with concomitant declines in abundance and diversity, likely due to changes in the availability and quality of habitat (e.g. highly urban areas have less greenspace and are hotter). We also predicted positive correlations between bee abundance and flower abundance, due to increased resource availability. But we were uncertain whether positive effects of floral resources would be 22 sufficient to counteract the negative effects of impervious surface or of the relative importance of other local habitat factors on pollination services.
Methods
Sampling Location. We sampled bees from a total of 30 sites (parks and gardens) in the metropolitan region of Toledo, OH, USA (Figure 5). This 620 km2 region is home to a half- million people, and its network of over 150 community gardens and 125 city parks was utilized for sampling locations (Burdine and Taylor 2017). We selected our 30 sites by overlaying a grid across a map of metropolitan Toledo in ArcGIS, and each grid cell (2 km x 2 km) was numbered.
We then used a random number generator to select which grid cells to include in the study, and within each of those selected grid cells, we identified a single park or garden to sample using a random number generator. Parks or gardens ranged in size from 0.001 to 0.46 km2.
Sampling Methods. We collected bees using elevated pan traps once per month between
June and August in 2016. Sampling was restricted to sunny days with temperatures above 22°C.
We constructed the elevated pan traps by placing a 175-ml plastic bowl (yellow, blue, or white) atop 1 m PVC pipe (Tuell and Isaacs 2009). Bowls were painted with Krylon ColorMaster® spray paint to enhance visibility, and each bowl was filled with a 150 ml mixture of water and dish soap. On sampling days, we placed 9 elevated pan traps (3 yellow, 3 blue, 3 white) along a transect at the center of each site and left them in the field for 24 hours. During collection, the contents of all 9 pan traps were combined into a single container for transport to the lab, and in the lab bees were separated from the other bycatch insects. Once sorted, bees were preserved in ethanol prior to pinning, and identified to species or morphospecies using a synoptic collection from Pardee and Philpott (2014), and the Discover Life bee species guides (Ascher and Pickering
2016). 23
Local Habitat Characteristics. We measured local habitat characteristics at each site during sampling events. All characteristics were measured with the center of each site as the focal point, corresponding to the pan trap locations. We calculated vegetation canopy cover when facing each cardinal direction away from the site’s center using a densiometer. We also counted the number of trees within 25 m of the site’s center. Additionally, we walked a 10 m transect out from the center of each site and counted the total number of flowers in bloom within
1 m of the transect line, and recorded their color. We measured floral color as a predictor as others have done (Pardee and Philpott 2014; Quistberg et al. 2016) since bees can have preferences for specific flower colors (Campbell et al. 2010). We calculated groundcover by randomly placing four quadrats (1m x 1m) along each transect, and estimated the percentage of herbaceous vegetation, woody vegetation, leaf litter, and bare ground cover (similar to Lagucki,
Burdine, & McCluney, 2017). We also took four measurements of volumetric soil moisture along the transect using a soil moisture meter (Delta-T Devices SM150), as a proxy for water availability. Water drinking behavior has been commonly observed in honeybees, and there is evidence that honeybees may be vulnerable to desiccation in cities (Burdine & McCluney 2019).
Thus, we focused on highly localized factors (within 25 m) that could influence bee abundance, diversity, community composition, and pollination given variation in the degree of urbanization surrounding each garden or park.
To assess the urbanization of the surrounding landscape, we estimated percent impervious surface within a 300 m radius of each site’s center, using the National Land Cover
Databases’ dataset for 2011 Percent Development Imperviousness (Homer et al. 2015).
Visitation Rates. We estimated pollinator visitation rates at the center of each site, during each sampling event, by placing five flowering plants: 1) tomato (early girl variety), 2) purple 24 headed cone flower (Echinacea purpurea), (3) brown-eyed susans (Rudbeckia triloba), 4) bergamot (Monarda fistulosa), and 5) foxglove (Penstemon digitalis). We selected these five plants because they are attractive to pollinators and are commonly found in Toledo parks and gardens (Pardee and Philpott 2014; Burdine and Taylor 2017). We counted the total number of individual insect pollinators that visited the plants over three 20-minute timespans in the earlier afternoon (similar to Lowenstein et al., 2014). We used these measures to calculate a visitation rate for each site (visits/hour), which others found to be well correlated with fruit production
(Garibaldi et al. 2013) in an agriculatural system.
Statistical Methods. We conducted all statistical tests using the program R. The cor function in R was used to examine collinearity between our environmental factors; for co- correlated factors (R > 0.5) we dropped one of the factors from statistical analyses (Table 2). We examined the effects of environmental factors on community composition with a Type II
PERMANOVA (adonis.II) using the “RVAideMemoire” package. We also utilized non-metric multidimensional scaling (metaMDS) within the “vegan” package to display differences in community composition, and associations with environmental factors. We used Bray-Curtis distances for all community composition analyses.
We used generalized linear models (glm) to examine relationships among environmental factors and dependent variables (abundance, diversity, visitation frequency). Models were developed by first establishing a list of candidate models that contained each potential predictor independently (Table 3). Then, we took the model(s) with the lowest AICc values (models within
2 AICc units were considered equivalent) and combined the models to test whether the combined model was a better fit (2 AICc units lower). We chose this process instead of model averaging approaches because they can be problematic with interactive models (Cade 2015; Harrison et al. 25
2018). There are also issues with model averaging because AIC weights are relative to the model, and models and are not independent. We also tested for interactions between various site- level environmental factors and impervious surface to examine potential modifiers of an urbanization effect. Assumptions of normality and equal variance were assessed by examining plots of residuals and data transformations were used when necessary. We tested for spatial autocorrelation using the “ape” package in R, and our results showed no spatial autocorrelation in the dependent variables (Table 4). We also tested whether site type (urban garden vs. city park) had an impact on the dependent variables (abundance, diversity, visitation frequency), and found no significant differences (Table 5).
Results
Summary Statistics. We collected a total of 727 bees representing 19 genera and 58 species from 30 sites. The majority of bees sampled were females (84.2%). The most common genera in order of abundance were the sweat bee Lasioglossum (48.4%), the long-horned bee
Melissodes (8.8%), the striped sweat bee Agapostemon (7.57%), the mining bee Halictus
(7.02%), and the sweat bee Augochlora (6.88%). The most diverse genera in order of number of species were the sweat bee Lasioglossum (13 species), the long-horned bee Melissodes (7 species), the bumblebee Bombus (5 species), and the leafcutter bee Megachile (5 species). Across sampling periods, we collected between 2 and 19 species per site.
Community Composition. We found impervious surface to be the only environmental variable significantly associated with bee community composition (PERMANOVA F1,21 = 1.99, p = 0.01; Table 6). Nonmetric dimensional scaling plots (Figure 6B) indicated that Lasioglossum imitatum was positively associated with urbanization, and 4 species were negatively associated 26 with urbanization: Bombus bimaculatus, Lasioglossum fatiggi, Hylaeus annulatus, and Hylaeus illinoisensis.
Overall Abundance and Diversity. We found several local factors to be strongly associated with the overall abundance and diversity of bees. The most parsimonious model for bee abundance was an additive model with abundance declining with increased canopy cover and impervious surface (AIC = 247.11, R2 = 0.27, Table 7, Figure 7 A-B). We identified two additional models with similar AIC values: the interaction of canopy cover and impervious surface (AIC = 248.31, R2 = 0.30, Table 7, Figure 8A) and canopy cover, but not impervious surface (AIC = 248.37, R2 = 0.18, Table 7, Figure 7A). For bee diversity, the most parsimonious model included the interaction of impervious surface and purple flower abundance (AIC = 31.96,
R2 = 0.5, Table 7, Figure 8B), with diversity declining with impervious surface, but only when purple flowers were not abundant. We identified an additional model with similar AIC that included the interaction of total flower abundance and impervious surface (AIC = 33.48, R2 =
0.48, Table 7), showing a similar pattern.
Visitation Rates. The most parsimonious model for visitation rates was an additive model with visitation declining with impervious surface, but increasing with flower abundance (AIC =
255.59, R2 = 0.67, Table 7, Figure 7 C-D), explaining 67% of the variation in visitation. In addition, we found a positive correlation between visitation rates and bee abundance (R = 0.49) and diversity (R = 0.44).
Discussion
Overall, our results indicate that bee diversity and pollination services decline with increased urbanization, but local habitat features can modify the effects of urbanization. More specifically, abundant flowers (all or purple) can help prevent urbanization-related declines in 27 diversity and pollination services. Although these results might have been expected from other research, mostly outside cities, showing positive effects of flowers (Kells et al. 2001, Blaauw and
Isaacs 2014) on bee abundance, diversity, and pollination, they are in contrast with another recent study that does not indicate that flowers can rescue urban bees (Hamblin et al. 2018).
Although there are many potential mechanisms underlying the differences observed between that study and ours, background climate may be one important factor. Hamblin et al. (2017) found that differences in thermal tolerance between species strongly drove abundance of bees along a gradient of urban-related warming in Raleigh, NC, an already warm southeastern city. This contrasts with a recent study finding that three species of bees in urban parts of Toledo, OH, a cooler city, are not near their thermal limits, and thus are unlikely to be influenced by urban- warming (Burdine and Mccluney 2019). Thus, flowers may be unable to rescue bees from urban- warming in already warm climates, but may be sufficient to reduce declines in cooler cities.
Other explanations are possible; in general, more work is needed to better identify regional differences in both the effects of urbanization and the potential for mitigation of urbanization via floral resources, or other factors. But here we show that improved floral resources can mitigate urban-related declines in pollinators and pollination services in Toledo, OH.
Impervious Surface. Impervious surface was the only habitat characteristic associated with community composition. In particular, we found a positive association of impervious surface on Lasioglossum imitatum, a solitary and ground nesting species. Normandi et al. (2017) provide evidence that this species can be abundant in certain urban habitats (e.g. cemeteries). On the other hand, we identified 11 species that were exclusively present at low impervious sites (<
25%), and 1 cavity-nesting species (Hylaeus illinoisensis) present only at high impervious sites
(> 50%). Multiple studies have identified changes in bee community composition across 28 urbanization gradients (Bates et al. 2011, Fortel et al. 2014), and degraded nest site availability with increasing impervious surface may help explain changes in composition (Cane et al. 2006).
Canopy Cover. We found a negative association between canopy cover and bee abundance, and the association was stronger at low impervious sites. Declines in floral abundance with increasing canopy cover may mediate declines in abundance. However, others have also identified canopy cover as a significant predictor of bee abundance in non-urban systems (Jha and Vandermeer 2010), and particularly for solitary species. There are examples of other arthropod taxa responding negatively to canopy cover in cities (Philpott et al. 2014,
Lagucki et al. 2017). Matteson and Langellotto (2010) found that shading from buildings in New
York City reduced sunlight availability in urban greenspaces, negatively impacting species richness of bees. Increased shade may prevent bees from maintaining optimal body temperatures by passive basking in greenspaces (Matteson & Langellotto, 2010), and this may explain why canopy cover associations were stronger at low impervious sites (reduced heat island effects).
Increased shade in urban regions can also reduce floral abundance, and Matteson et al. (2013) show that this can indirectly impact insect pollinators.
Flower Abundance. Others have identified the importance of floral availability in maintaining diverse bee assemblages in cities (Matteson Langellotto, 2010; Pardee & Philpott,
2014; Quistberg et al., 2016), but here we find this pattern occurs across a wide range of impervious surface, with flowers restricting declines in bee diversity and pollination that would otherwise be seen in cities. In particular, we found a strong influence of purple flowers on bees and this may due to the types of purple flowers that were present (e.g. purple clover, purple- headed coneflower, bergamot). There is also evidence that bees may be more attracted to blue and ultraviolet flowers due to visual signals that are visible to bees in these flowers (Mayround et 29 al. 2017). Lowenstein et al. (2014) shows that increased floral diversity can mitigate any potential negative effects of urbanization, even in densely-population regions, and visitation frequencies may even increase with urbanization. However, increasing floral resources is less effective in warmer cities like Raleigh, NC (Hamblin et al. 2018). We expected to find a positive relationship between flower abundance and visitation frequencies, and the strength of the relationship (R2 = 0.57) suggests that increased flower availability might strongly help prevent declines in pollination services at high impervious sites.
Caveats. Our research has several methodological limitations. First, by only using pan traps to sample bees we may have under-sampled certain taxa. Others have shown that pan traps can underrepresent larger bees (Roulston et al. 2007), but we still collected many large bees (e.g. honeybees, bumblebees) and this method of capture was constant across all sampling sites, providing robust metrics of relative differences between sites. Second, visitation rates may not always reflect pollination. Visitors are not necessarily pollinating flowers, and others have suggested combining measures of visitation with an estimate of pollinator effectiveness (King et al. 2013). However, there are instances within agricultural systems, such as those studied here, where visitation rates have been shown to be a good metric of pollination (Garibaldi et al. 2013), but more work is needed in urban agricultural systems. Third, we included both urban gardens and city parks as samping sites because both are greenspaces embedded within an urban landscapes. Even though these site types are different types of greenspaces, we did not detect different effects based on site type.
Conclusions. We show that negative effects of urbanization on bee communities and pollination services can be altered by local habitat characteristics (flower abundance, canopy cover). More specifically, increasing the total number of flowers could be an important strategy 30 for improving pollination services, independent of whether the garden is embedded within a highly impervious habitat.
31
CHAPTER III: BEE GUILDS SHOW DISSIMILAR RESPONSES TO URBANIZATION
IN A MEDIUM-SIZED CITY, ALTERING POLLINATOR VISITATION
Introduction
Recent pollinator declines may disrupt or impair pollination services (Kremen et al. 2002,
Klein et al. 2003a, Potts et al. 2010). The economic value of animal-mediated pollination is $117 billion annually (Costanza et al. 1997), and over a third of the global food supply requires pollination (Klein et al. 2007). Arguably, the most important and well-studied of the animal pollinators are European honey bees (Apis mellifera), and honey bees have been shown to increase yields in 96% of pollinator-dependent crops (Klein et al. 2007, Potts et al. 2010).
However, many regions have become increasingly dependent on wild bees for crop pollination due to regional declines in honeybee colonies (van Engelsdorp et al. 2008), and there is growing evidence that wild bees deliver satisfactory pollination services to many crops in the absence of honey bees (Ricketts 2004, Shuler et al. 2005, Klein et al. 2007). Pollination services tend to increase with bee abundance (Klein et al. 2003b, Winston and Morandin 2005), and when the abundance of different bee species is more even (Garibaldi et al. 2013). However, the relationship between bee diversity and pollination is often inconsistent and difficult to predict.
One reason relating bee diversity to pollination is problematic is because increased species diversity doesn’t always mean increased trait diversity. The relationship between species and functional diversity is generally positive because diverse communities often display a larger number of taxonomic, morphological, and physiological differences between species (Biswas and Mallik 2011), and this may also be true for bees. However, environmental conditions (e.g. habitat features, disturbance) do influence the strength of the species-functional diversity relationship, which suggests that certain traits are more sensitive than others to changing 32 environmental conditions (see Williams et al. 2010). Others have found declines in the functional diversity of plant (Flynn et al. 2009, Laliberté et al. 2010) and pollinator communities (Rader et al. 2014) with land-use intensification in cities and agricultural areas, but implications for pollination services have been less explored (but see van der Plas 2019).
Another reason it is difficult to relate bee diversity to pollination is because the relationship between trait diversity and pollination is often inconsistent (see Bartomeus et al.
2018). While certain traits (e.g. body size, nesting location, sociality) may be strongly predictive of species responses to environmental change (Williams et al. 2010), they are often less predictive of pollination services (Bartomeus et al. 2018). Few studies have attempted to investigate whether combinations of traits could be used to predict pollination services (but see
(Williams et al. 2010). Some have separated bees by functional traits into functional guilds for experimental studies about pollination (Fontaine et al. 2006, Hoehn et al. 2008). Functional guilds are established by assigning bee species to groups based on shared traits (i.e. body size, tongue length) or combinations of traits (i.e. large bees with long tongues). But the effects of functional guilds on pollination services have received less investigation.
To assess how functional traits and guilds respond to urbanization and how this relates to pollination services, we sampled bee communities in parks and gardens across a medium-sized city, and measured habitat features (herbaceous cover, floral abundance and color, tree abundance, canopy cover, soil moisture, park/garden size) and pollination services (visitation frequency) at each site. We also compiled a database of functional traits for the sampled species from the literature (Fortel et al. 2014, Pardee & Philpott 2014, Forrest et al. 2015) and used these traits to classify bee species into functional guilds. These data were then used to investigate how urbanization impacts functional diversity and guild richness and composition, exploring whether 33 habitat features explain differences in the abundance of individuals within and among guilds. We expected to see declines in functional diversity and guild richness with increasing urbanization
(impervious surface), likely due to changes in the types and quality of habitat. We then expected declines in functional diversity and guild richness to be strongly correlated with declines in pollination services. We expected these relationships to be more greatly driven by certain traits and guilds. However, we also assessed the alternative prediction that pollination is more associated with total bee abundance or species diversity than functional traits.
Methods
Sampling Locations. We sampled bees from a total of 30 sites (parks and gardens) in the metropolitan region of Toledo, OH, USA (Figure 5). This 620 km2 region is home to a half- million people, and its network of over 150 community gardens and 125 city parks was utilized for sampling locations (Burdine and Taylor 2017). We selected our 30 sites by overlaying a grid across a map of metropolitan Toledo in ArcGIS, and each grid cell (2 km x 2 km) was numbered.
We then used a random number generator to select which grid cells to include in the study, and within each of those selected grid cells, we identified a single park or garden to sample using a random number generator. Parks or gardens ranged in size from 0.001 to 0.46 km2.
Sampling Methods. We collected bees using elevated pan traps once per month between
June and August in 2016. Sampling was restricted to sunny days with temperatures above 22°C.
We constructed the elevated pan traps by placing a 175-ml plastic bowl (yellow, blue, or white) atop 1 m PVC pipe (Tuell and Isaacs 2009). Bowls were painted with Krylon ColorMaster® spray paint to enhance visibility, and each bowl was filled with a 150 ml mixture of water and dish soap. On sampling days, we placed 9 elevated pan traps (3 yellow, 3 blue, 3 white) along a transect at the center of each site and left them in the field for 24 hours. During collection, the 34 contents of all 9 pan traps were combined into a single container for transport to the lab, and in the lab bees were separated from the other bycatch insects (bycatch insects were analyzed separately, see Lagucki et al. 2017). Once sorted, bees were preserved in ethanol prior to pinning, and identified to species or morphospecies using a synoptic collection from Pardee and
Philpott (2014), and the Discover Life bee species guides (Ascher and Pickering 2016).
Habitat Characteristics. We measured local habitat characteristics at each site during sampling events. All characteristics were measured with the center of each site as the focal point, corresponding to the pan trap locations. We calculated canopy cover when facing each cardinal direction away from the site’s center using a densiometer. We also counted the number of trees within 25 m of the site’s center. Additionally, we walked a 10 m transect out from the center of each site and counted the total number of flowers in bloom within 1 m of the transect line, and recorded their color. We measured floral color as a predictor as others have done (Pardee and
Philpott 2014; Quistberg et al. 2016) since bees can have preferences for specific flower colors
(Campbell et al. 2010). We calculated groundcover by randomly placing four quadrats (1m x
1m) along each transect, and estimated the percentage of herbaceous vegetation, woody vegetation, leaf litter, and bare ground cover (similar to Lagucki, Burdine, & McCluney, 2017).
We also took four measurements of volumetric soil moisture along the transect using a soil moisture meter (Delta-T Devices SM150), as a proxy for water availability. Water drinking behavior has been commonly observed in honeybees, and there is evidence that honeybees may be vulnerable to desiccation in cities (Burdine & McCluney 2019). Thus, we focused on highly localized factors (within 25 m) that could influence bee abundance, diversity, community composition, and pollination given variation in the degree of urbanization surrounding each garden or park. 35
To assess the urbanization of the surrounding landscape, we estimated percent impervious surface within a 300 m radius of each site’s center, using the National Land Cover
Databases’ dataset for 2011 Percent Development Imperviousness (Homer et al. 2015).
Visitation Rates. We estimated pollinator visitation rates at the center of each site, during each sampling event, by placing five flowering plants: 1) tomato (early girl variety), 2) purple headed cone flower (Echinacea purpurea), (3) brown-eyed susans (Rudbeckia triloba), 4) bergamot (Monarda fistulosa), and 5) foxglove (Penstemon digitalis). We selected these five plants because they are attractive to pollinators and are commonly found in Toledo parks and gardens (Pardee and Philpott 2014; Burdine and Taylor 2017). We counted the total number of individual insect pollinators that visited the plants over a 20-minute timespan (similar to
Lowenstein et al., 2014). We used these measures to calculate a visitation rate for each site
(visits/hour), which others found to be well correlated with fruit production in agricultural environments (Garibaldi et al. 2013).
Functional Traits. For each species we collected (see Table 8), we assembled information on seven life-history traits from similar studies (Fortel et al. 2014, Pardee and Philpott 2014,
Forrest et al. 2015) and from personal observations and measurements. These traits included body size (small, medium, large), nesting location (above or belowground), dietary specialization
(polylectic = pollen generalist; oligolectic = pollen specialist), sociality (solitary, eusocial, cleptoparasitic), pollen deposition (legs and body, abdomen underside, corbiculae, crop, legs only, accidental), seasonality (year around, spring only, spring and summer, summer only), and tongue length (short, medium, long). Body size was estimated with a measurement of intertegular distance (ITD, length between the 2 wing tegulae on a bee’s thorax) using digital calipers, and bees were classified as either small (ITD < 2mm), medium (ITD between 2.1-3 36 mm), or large (ITD >3 mm), as others have done in similar studies (Hoehn et al. 2008). Tongue lengths were classified as either small (< 3mm), medium (3 – 7 mm), or long (> 7 mm).
We used the metric of functional dispersion (FDis) to represent functional trait diversity using the FD package in R (Laliberte and Legendre 2010). This metric uses a list of functional traits to construct a multivariate space, and then measures the average distance between each species and the community centroid of the multivariate space. FDis was calculated for each site using the Cailliez correction since we included categorical traits (Laliberte and Legendre 2010).
We then used generalized linear models (glm) to compare FDis to habitat characteristics and estimates of pollination services.
Functional Guilds. We classified bee species into 12 functional guilds based on combinations of three traits: body size, tongue length, nesting location (see Table 9). These traits were selected because they have been shown to vary with urbanization, and because others have utilized these traits when classifying bee species to guild (Fontaine et al. 2006, Hoehn et al.
2008).
Statistical Methods. We conducted all statistical tests using the program R. The cor function in R was used to examine collinearity between our environmental factors; for co- correlated factors (R > 0.5) we dropped one of the factors from statistical analyses (Table 2). We used this cut-off for the level of correlation in order to constrain the potential model set. We examined the effects of environmental factors on guild community composition with a Type II
PERMANOVA (adonis.II) using the “RVAideMemoire” package. We also utilized non-metric multidimensional scaling (metaMDS) within the “vegan” package to display differences in guild community composition, and associations with environmental factors. We used Bray-Curtis distances for all community composition analyses. 37
We used generalized linear models (glm) to examine relationships among environmental factors and dependent variables (FDis, guild richness). Models were developed by first establishing a list of candidate models that contained each potential predictor independently (see
Tables 7 and 8). Then, we took the model(s) with the lowest AIC values (models within 2 AIC units were considered equivalent) and combined the models to test whether the combined model was more parsimonious (2 AIC units lower, see Table 7). We also tested for interactions between various site-level environmental factors and impervious surface to examine potential modifiers of an urbanization effect. We used glm to compare which variables (abundance, species diversity, functional trait diversity, guild richness) were stronger predictors of visitation frequencies (pollination). We also used glm to examine guild-specific relationships with environmental conditions for guilds that had greater than 15 individuals. We used logistic regression (presence/absence) as a follow-up to guild richness results, and Gaussian glm as a follow-up to PERMANOVA results. Assumptions of normality and equal variance were assessed by examining plots of residuals and data transformations were used when necessary. We tested for spatial autocorrelation using the “ape” package in R, and our results showed no spatial autocorrelation in the dependent variables (Table 4). We also tested whether site type (urban garden vs. city park) had an impact on the dependent variables (abundance, diversity, visitation frequency), and found no significant differences (Table 5).
Results
Functional Trait Diversity and Guild Richness. We did not detect any associations between habitat characteristics and functional dispersion (FDis, our metric for functional trait diversity). The most parsimonious models for FDis were within 2 AIC units of the null model
(Table 10). For functional guild richness, the most parsimonious model included the interaction 38 of the total number of flowers and impervious surface (AICc = 110.46, R2 = 0.51, Table 12,
Figure 9A), with guild richness declining with impervious surface, but only when floral availability was low.
Guild Composition. Impervious surface was the only habitat characteristic significantly associated with the community composition of bee guilds (PERMANOVA F1,21 = 2.095, p =
0.047; Table 13). The NMDS plots (Figure 10 A-C) showed that small bees with medium tongues that nested aboveground (Lasioglossum platyparium), and large bees with long tongues that nested belowground (long-horned bees; Melissodes spp) were positively associated with impervious surface. The NMDS plots also showed that some guilds were negatively associated with impervious surface: large bees with long tongues that nest aboveground (LAL, honey bees, bumble bees), medium-sized bees with medium-length tongues that nest aboveground (mason bees and wood-nesting bees), and small-bodied bees with medium tongues that nest aboveground
(Sphecodes and Hylaeus spp). We also found that visitation rates were associated with guild community composition in multivariate space (F1,29 = 4.21, P < 0.01, Figure 10C). The NMDS plots showed that honey bees and bumble bees (LAL) were positively associated with visitation rates, while L. platyparium (MAS) was negatively associated with visitation rates.
Guild Specific Results. We identified additional habitat characteristics that were strongly associated with the abundance of bees within specific guilds. Large, long-tongued bees (Guild
LAL: bumble bees, honey bees) were more abundant with more purple flowers (AICc = 96.89,
R2 = 0.18, Table 12, Figure 11C). A second model with similar AIC included the interaction of purple flowers and impervious surface (AICc = 97.40, R2 = 0.29, Table 12, Figure 9B), with fewer bumble bees and honey bees found at high impervious sites, but only when purple flower abundance was low. Medium-sized bees with medium-length tongues that nested aboveground 39
(guild MAM, mason and wood-nesting bees) declined with increasing impervious surface and canopy cover (AICc = 172.84, R2 = 0.39, Table 12, Figure 11A-B). For the remaining guilds, the most parsimonious models were within 2 AIC units of the null model (Table 11).
Visitation Rates. We compared four candidate models that used bee community metrics
(abundance, species diversity, functional diversity, functional guild richness) to predict visitation rates. The most parsimonious model indicated a positive association between visitation rates and guild richness (AICc =273.57, R2 = 0.36; Table 14; Figure 12A), and this model was distinctly better than the models using bee abundance (ΔAICc = 5.75, R2 = 0.23, Table 14, Figure 12C), bee diversity (ΔAICc = 7.17, R2 = 0.19, Table 14, Figure 12D), and functional trait diversity
(ΔAICc = 13.5, R2 = 0.01, Table 14, Figure 12D). Then, we used logistic regression
(presence/absence) post-hoc tests to investigate which functional guilds most strongly contributed to visitation rates, and which environmental factors influenced presence/absence in these guilds. We found that that presence of mason and wood-nesting bees (guild MAM) was the strongest predictor of visitation rates (AICc = 40.65), and the presence of these bees was influenced by floral availability (AICc = 26.80). In addition, the abundance of mason and wood- nesting bees (guild MAM) was positively associated with visitation rates (AICc = 177.9, R2 =
0.13).
Discussion
Overall, our results indicate that functional guild richness is a strong predictor of pollination services, and that bee guilds are differentially impacted by urbanization (impervious surface). Others have found that urbanization can alter the types of functional traits present in bee communities (Cane et al. 2006, Bates et al. 2011, Banaszak-Cibicka and Zmihorski 2012,
Fortel et al. 2014). Our results suggest that combinations of multiple functional traits may 40 explain why certain bees were abundant with increasing urbanization. For instance, large bees with long tongues that nest belowground (Guild LBL; long-horned bees; Melissodes spp.) performed well with increasing urbanization, but Guild LAL (honey bees, aboveground bumble bees) performed poorly. While both guilds contain species with long-tongues and large body sizes, they diverge in nesting behavior (aboveground, belowground). Thus, including this third trait helped explain why long-horned bees performed well and honey bees and bumble bees performed poorly—above ground nesting species may be more adversely affected by urbanization. In addition, guild richness was positively associated with flower abundance (all or purple), and within guilds we identified multiple habitat features that had positive (total flowers, purple flowers) and negative (canopy cover, impervious surface) impacts on bee abundance within guilds.
One of the guilds that was negatively impacted by urbanization contained mason and wood-nesting bees (Guild MAM). This was also the only functional guild associated with visitation rates, and others have found mason bees to be as effective as honeybees when investigating pollination at the level of the individual flower (Jauker et al. 2012). Our ability to detect the presence of these bees at a site depended on floral resource availability, which suggests that floral resources mediate declines in mason and wood-nesting bees with increasing urbanization. Urbanization also negatively impacted honey bees (A. mellifera) and bumble bees
(Bombus spp.), which are known to be important generalist pollinators (Winfree et al. 2007,
Rader et al. 2009, Cameron et al. 2011), and their large body size allows for increased foraging distance and heavier pollen loads. Additionally, our results indicate that increasing the number of purple flowers may counteract the negative effects of urbanization, and the positive effects of 41 floral resources on bumble bees and honey bees is supported in the literature (Pardee and
Philpott 2014, Williams et al. 2015, Quistberg et al. 2016).
One of the guilds that appeared to do well with increasing urbanization contained the small, parasitic Lasioglossum platyparium. This species is cleptoparasitic and lays its eggs in the nests of other taxa, including some Lasioglossum and long-horned bee species. Fortel et al.
(2014) also found that parasitic bees do well at sites with intermediate levels of impervious surface, and the presence of parasitic bee species may be a good indicator of bee community stability since parasitic bees are generally the first to respond to disturbances (Sheffield et al.
2013). However, the presence of L. platyparium was negatively associated with visitation rates.
Parasitic bees are generally poor pollinators because they lack pollen-collecting hairs, and so their increased presence in urban habitats does not increase pollination. In fact, their parasitic behavior may negatively impact their hosts and indirectly effect pollination. Thus, in urban systems, they may not be a good indicator of bee community stability. The other guild that appeared to do well with urbanization (LBL) contained 7 Melissodes species (which can be the parasitic hosts of L. platyparium). These species nest in belowground tunnels and are often oligolectic (plant specialists), which suggests that urban parks and gardens are quality sources of floral and nesting resources. Others have found that Melissodes species were the primary pollinators of sunflowers and cucumbers in urban greenspaces (Fitch 2017). But here we find that overall pollination may decline in those same locations, due to the loss of other guilds.
Others have identified changes in bee communities with urbanization due to changes in floral and nesting resources (Ahrne et al. 2009, Banaszak-Cibicka and Zmihorski 2012, Fortel et al. 2014). While functional guild composition was not found to be strongly influenced by floral and nesting resource availability, we did find that functional guild richness was higher at sites 42 with increased floral resource availability. Multiple studies have shown that floral resources mediate the relationship between functional diversity and pollination (Fontaine et al. 2006,
Garibaldi et al. 2013). For instance, Fontaine et al. (2006) found that pollination (e.g. fruit production) increased with pollinator trait diversity, but only when plant communities were also functionally diverse. Therefore, increasing floral resource availability and diversity may be an important strategy for maintaining functionally diverse bee communities.
Conclusion. We show that bee guilds have dissimilar responses to urbanization, and these responses are influenced by habitat characteristics. In particular, the guild containing mason and wood-nesting bees increased pollination services, and the availability of floral resources influenced the presence of these bees. Our results suggest that increasing floral resources could be an important strategy for increasing the abundances of honey bees, bumble bees, and mason bees, and may indirectly enhance pollination services.
43
REFERENCES
Addo-Bediako, A., S.L. Chown, and K.J. Gaston. 2001. Revisiting water loss in insects: A large
scale view. Journal of Insect Physiology 47:1377–1388.
Ahrne, K., J. Bengtsson, and T. Elmqvist. 2009. Bumble bees (Bombus spp) along a gradient of
increasing urbanization. Plos One 4:e5574.
Allen-Wardell, G., P. Bernhardt, R. Bitner, A. Burquez, S. Buchmann, J. Cane, P.A. Cox, V.
Dalton, P. Feinsinger, M. Ingram, D. Inouye, C.E. Jones, K. Kennedy, P. Kevan, H.
Koopowitz, S. Medellin-Morales, G.P. Nabhan, B. Pavlik, V. Tepedino, and S. Walker.
1998. The potential consequences of pollinator declines on the conservatoin of biodiversity
and stability of food crop yields. Conservation Biology 12:8–17.
Angilletta, M.J. 2009. Thermal adaptation: a theoretical and empirical synthesis. Oxford
University Press, Oxford.
Angilletta, M.J., R.S. Wilson, A.C. Niehaus, M.W. Sears, C.A. Navas, and P.L. Ribeiro. 2007.
Urban physiology: City ants possess high heat tolerance. PLoS One 2:e258.
Araújo, M.B., F. Ferri-Yáñez, F. Bozinovic, P.A. Marquet, F. Valladares, and S.L. Chown. 2013.
Heat freezes niche evolution. Ecology Letters 16:1206–1219.
Ascher, J.S., and J. Pickering. 2016. Discover Life bee species guide and world checklist
(Hymenoptera: Apoida: Anthopila). Available at: http://www.discoverlife.org/
Atmowidjojo, A.H., D.E. Wheeler, E.H. Erickson, and A.C. Cohen. 1997. Temperature tolerance
and water balance in feral and domestic honey bees, Apis mellifera L. Comparative
Biochemistry and Physiology 118A:1399–1403.
Banaszak-Cibicka, W., and M. Zmihorski. 2012. Wild bees along an urban gradient: Winners
and losers. Journal of Insect Conservation 16:331–343. 44
Bartomeus, I., D.P. Cariveau, T. Harrison, and R. Winfree. 2018. On the inconsistency of
pollinator species traits for predicting either response to land-use change or functional
contribution. Oikos 127:306–315.
Bates, A.J., J.P. Sadler, A.J. Fairbrass, S.J. Falk, J.D. Hale, and T.J. Matthews. 2011. Changing
bee and hoverfly pollinator assemblages along an urban-rural gradient. PLoS ONE
6:e23459.
Baudier, K.M., C.L.D. Amelio, R. Malhotra, M.P.O. Connor, and S.O. Donnell. 2018. Extreme
insolation : Climatic variation shapes the evolution of thermal tolerance at multiple scales.
The American Naturalist 192:347–359.
Biesmeijer, J.C. 2006. Parallel declines in pollinators and insect-pollinated plants in Britain and
the Netherlands. Science 313:351–354.
Biswas, S.R., and A.U. Mallik. 2011. Species diversity and functional diversity relationship
varies with disturbance intensity. Ecosphere 2:1–10.
Blaauw, B.R., and R. Isaacs. 2014. Flower plantings increase wild bee abundance and the
pollination services provided to a pollination-dependent crop. Journal of Applied Ecology
51:890–898.
Block, W. 2003. Water or ice? The challenge for invertebrate cold survival. Science progress
86:77–101.
Bolker, B.M., M.E. Brooks, C.J. Clark, S.W. Geange, J.R. Poulsen, M.H.H. Stevens, and J.S.S.
White. 2009. Generalized linear mixed models: a practical guide for ecology and evolution.
Trends in Ecology and Evolution 24:127–135.
Both, C., S. Bouwhuis, C.M. Lessells, and M.E. Visser. 2006. Climate change and population
declines in a long-distance migratory bird. Nature 441:81–83. 45
Brans, K.I., M. Jansen, J. Vanoverbeke, N. Tuzun, R. Stoks, and L. De Meester. 2017. The heat
is on: genetic adaptation to urbanization mediated by thermal tolerance and body size.
Global Change Biology:1–10.
Burdine, J.D., and K.E. Mccluney. 2019. Differential sensitivity of bees to urbanization-driven
changes in body temperature and water content. Scientific Reports:1–10.
Burdine, J.D., and D.E. Taylor. 2017. Neighbourhood characteristics and urban gardens in the
Toledo metropolitan area: staffing and voluntarism, food production, infrastructure, and
sustainability practices. Local Environment 23:198–219.
Cade, B.S. 2015. Model averaging and muddled multimodel inferences. Ecology 96:9:2370-
2382.
Cameron, S.A., J.D. Lozier, J.P. Strange, J.B. Koch, N. Cordes, L.F. Solter, and T.L. Griswold.
2011. Patterns of widespread decline in North American bumble bees. Proceedings of the
National Academy of Sciences 108:662–667.
Campbell, D.R., M. Bischoff, J.M. Lord, A.W. Robertson. 2010. Flower color influences insect
visitaion in alpine New Zealand. Ecology 91:9:2638-2649.
Cane, J.H., R.L. Minckley, L.J. Kervin, T. Roulston, and N.M. Williams. 2006. Complex
responses within a desert bee guild (Hymenoptera: Apiformes) to urban habitat
fragmentation. Ecological Applications 16:632–644.
Casey, T.M. 1976. Activity patterns, body temperature and thermal ecology in two desert
caterpillars (Lepidoptera : Sphingidae). Ecology 57:485–497.
Chen, J., T. Rashid, and G. Feng. 2014. A comparative study between Solenopsis invicta and
Solenopsis richteri on tolerance to heat and desiccation stresses. PLoS ONE 9:1–6.
Chown, S.L., J.G. Sørensen, and J.S. Terblanche. 2011. Water loss in insects: An environmental 46
change perspective. Journal of Insect Physiology 57:1070–1084.
Cooper, P.D., W.M. Schaffer, and S.L. Buchmann. 1985. Temperature regulation of honey bees
(Apis mellifera) foraging in the Sonoran Desert. Journal of Experimental Biology 114:1–15.
Costanza, R., R. Arge, R. De Groot, S. Farberk, M. Grasso, B. Hannon, K. Limburg, S. Naeem,
R.V. O’Neill, J. Paruelo, R.G. Raskin, P. Suttonkk, and M. van den Belt. 1997. The value of
the world’s ecosystem services and natural capital. Nature 387:253–260.
Dale, A.G., and S.D. Frank. 2018. Urban plants and climate drive unique arthropod interactions
with unpredictable consequences. Current Opinion in Insect Science 29:27–33.
Deutsch, C.A., J.J. Tewksbury, R.B. Huey, K.S. Sheldon, C.K. Ghalambor, D.C. Haak, and P.R.
Martin. 2008. Impacts of climate warming on terrestrial ectotherms across latitude Thermal
Safety margin. Proceedings of the National Academy of Sciences 105:6668–6672.
Diamond, S.E., L. Chick, A. Perez, S.A. Strickler, and R.A. Martin. 2017. Rapid evolution of ant
thermal tolerance across an urban-rural temperature cline. Biological Journal of the Linnean
Society 121:248–257. van Engelsdorp, D., J. Hayes, R.M. Underwood, and J. Pettis. 2008. A survey of honey bee
colony losses in the U.S., Fall 2007 to Spring 2008. PLoS ONE 3:8–13.
Fitch, G.M. 2017. Urbanization-mediated context dependence in the effect of floral
neighborhood on pollinator visitation. Oecologia 185:1–11.
Flynn, D.F.B., M. Gogol-Prokurat, T. Nogeire, N. Molinari, B.T. Richers, B.B. Lin, N. Simpson,
M.M. Mayfield, and F. DeClerck. 2009. Loss of functional diversity under land use
intensification across multiple taxa. Ecology Letters 12:22–33.
Foley, J.A., R. Defries, G.P. Asner, C. Barford, G. Bonan, S.R. Carpenter, F.S. Chapin, M.T.
Coe, G.C. Daily, H.K. Gibbs, J.H. Helkowski, T. Holloway, E.A. Howard, C.J. Kucharik, 47
C. Monfreda, J.A. Patz, I.C. Prentice, N. Ramankutty, and P.K. Snyder. 2005. Global
consequences of land use. Science 309:570–574.
Fontaine, C., I. Dajoz, J. Meriguet, and M. Loreau. 2006. Functional diversity of plant-pollinator
interaction webs enhances the persistence of plant communities. PLoS Biology 4:0129–
0135.
Forrest, J.R.K., R.W. Thorp, C. Kremen, and N.M. Williams. 2015. Contrasting patterns in
species and functional-trait diversity of bees in an agricultural landscape. Journal of Applied
Ecology 52:706–715.
Fortel, L., M. Henry, L. Guilbaud, A.L. Guirao, M. Kuhlmann, H. Mouret, O. Rollin, B.E.
Vaissière, L. Fortel, H. Mouret, O. Rollin, and B.E. Vaissie. 2014. Decreasing abundance,
increasing diversity and changing structure of the wild bee community (Hymenoptera:
Anthophila) along an urbanization gradient. PLoS One 9:e104679.
Gaffin, S.R., C. Rosenzweig, R. Khanbilvardi, L. Parshall, S. Mahani, H. Glickman, R.
Goldberg, R. Blake, R.B. Slosberg, and D. Hillel. 2008. Variations in New York City’s
urban heat island strength over time and space. Theoretical and Applied Climatology 94:1–
11.
Garibaldi, L.A., I. Steffan-Dewenter, R. Winfree, M.A. Aizen, R. Bommarco, S.A. Cunningham,
C. Kremen, L.G. Carvalheiro, L.D. Harder, O. Afik, I. Bartomeus, F. Benjamin, V. Boreux,
D. Cariveau, N.P. Chacoff, J.H. Dudenhöffer, B.M. Freitas, J. Ghazoul, S. Greenleaf, J.
Hipólito, A. Holzschuh, B. Howlett, R. Isaacs, S.K. Javorek, C.M. Kennedy, K.M.
Krewenka, S. Krishnan, Y. Mandelik, M.M. Mayfield, I. Motzke, T. Munyuli, B.A. Nault,
M. Otieno, J. Petersen, G. Pisanty, S.G. Potts, R. Rader, T.H. Ricketts, M. Rundlöf, C.L.
Seymour, C. Schüepp, H. Szentgyörgyi, H. Taki, T. Tscharntke, C.H. Vergara, B.F. Viana, 48
T.C. Wanger, C. Westphal, N. Williams, and A.M. Klein. 2013. Wild pollinators enhance
fruit set of crops regardless of honey bee abundance. Science 339:1608–1611.
Gaston, K.J., and S.L. Chown. 1999. Elevation and climatic tolerance : A test using dung beetles.
Oikos 86:584–590.
Gibbs, A.G., A.K. Chippindale, and M.R. Rose. 1997. Physiological mechanisms of evolved
desiccation resistance in Drosophila melanogaster. The Journal of Experimental Biology
200:1821–1832.
Glaum, P., M. Simao, C. Vaidya, G. Fitch, and B. Iulinao. 2017. Big city Bombus: Using natural
history and land-use history to find significant environmental drivers in bumblebee declines
in urban development. Royal Society Open Science 4:170156.
Goulson, D. 2003. Effects of introduced bees on native ecosystems. Annual Review of Ecology,
Evolution, and Systematics 34:1–26.
Greenleaf, S.S., N.M. Williams, R. Winfree, and C. Kremen. 2007. Bee foraging ranges and their
relationship to body size. Oecologia 153:589–596.
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.
Haase, D. 2008. Urban ecology of shrinking cities: an unrecognized opportunity? Nature and
Culture 3:1–8.
Hadley, N.F. 1994. Water relations of terrestrial arthropods. Academic Press, San Diego, CA.
Hall, D.M., G. Camilo, R.K. Tonietto, J. Ollerton, K. Ahrne, M. Arduser, J. Ascher, K. Baldock,
R. Fowler, G. Frankie, D. Goulson, B. Gunnarsson, M. Hanley, J. Jackson, G. Langellotto, 49
D. Lowenstein, E. Minor, S.M. Philpott, S.G. Potts, M. Sirohi, E. Spevak, G. Stone, and C.
Threlfall. 2017. The city as a refuge for insect pollinators. Conservation Biology 31:24–29.
Hamblin, A.L., E. Youngsteadt, and S.D. Frank. 2018. Wild bee abundance declines with urban
warming, regardless of floral density. Urban Ecosystems 21:419–428.
Hamblin, A.L., E. Youngsteadt, M.M. Lopez-Uribe, and S.D. Frank. 2017. Physiological thermal
limits predict differential responses of bees to urban heat-island effects. Biological Letters
13:20170125.
Harrison X.A., L. Donaldson, M.E. Correa-Cano, J. Evans, D.N. Fisher, C.E.D. Goodwin, B.S.
Robinson, D.J. Hodgson, and R. Inger. 2018. A brief introduction to mixed effects
modelling and multi-model infernece in ecology. PeerJ 6:e4794.
Heinrich, B. 1972. Temperature regulation in the bumblebee Bombus vagans: A field study.
Science 4018:185–187.
Heinrich, B. 1980. Mechanisms of body-temperature regulation in honeybees, Apis mellifera. II.
Regulation of thoracic temperature at high air temperatures. Journal of Experimental
Biology 85:73–87.
Heinrich, B., and M.J.E. Heinrich. 1983. Size and caste in temperature regulation by
Bumblebees. Physiological Zoology 56:552–562.
Hennig, E.I., and J. Ghazoul. 2012. Pollinating animals in the urban environment. Urban
Ecosystems 15:149–166.
Hernandez, J.L., G.W. Frankie, and R.W. Thorp. 2009. Ecology of urban bees: a review of
current knowledge and directions for future study. Cities and the Environment 2:1–15.
Hodgson, K., M.C. Campbell, and M. Bailkey. 2011. Urban agriculture: growing healthy,
sustainable places. Page American Planning Association. Washington DC. 50
Hoehn, P., T. Tscharntke, J.M. Tylianakis, and I. Steffan-Dewenter. 2008. Functional group
diversity of bee pollinators increases crop yield. Proceedings of the Royal Society B:
Biological Sciences 275:2283–2291.
Hoffmann, A.A., and P.A. Parsons. 1989. Selection for increased desiccation resistance in
Drosophila melanogaster: additive genetic control and correlated responses for other
stresses. Genetics 122:837–845.
Hoffmann, A.A., and C.M. Sgró. 2011. Climate change and evolutionary adaptation. Nature
470:479–485.
Hoffmann, A., R. Hallas, J. Dean, and M. Schiffer. 2003. Low potential for climate stress
adaptation in a rainforest Drosophila species. Science 301:100–103.
Homer, C.G., J.A. Dewitz, L. Yang, S. Jin, P. Danielson, G. Xian, J. Coulston, N.D. Herold, J.D.
Wickham, and K. Megown. 2015. Completion of the 2011 National Land Cover Database
for the conterminous United States - Representing a decade of land cover change
information. Photogrammetric Engineering and Remote Sensing 81:345–354.
Hothorn, T., F. Bretz, and M. Horthorn. 2009. The multcomp package. The R Project for
Statistical Computing.
Hou, Q., X. Pang, Y. Wang, K. Sun, L. Jia, S. Zhang, and Q. Li. 2019. Urbanization threaten the
pollination of Gentiana dahurica. Scientific Reports 9:583.
Huey, R.B., and R.D. Stevenson. 1979. Integrating thermal physiology and ecology of
ectotherms: A discussion of approaches. American Zoologist 19:357–366.
Imhoff, M.L., P. Zhang, R.E. Wolfe, and L. Bounoua. 2010. Remote sensing of the urban heat
island effect across biomes in the continental USA. Remote Sensing of Environment
114:504–513. 51
Jauker, F., B. Bondarenko, H.C. Becker, and I. Steffan-Dewenter. 2012. Pollination efficiency of
wild bees and hoverflies provided to oilseed rape. Agricultural and Forest Entomology
14:81–87.
Jha, S., and J.H. Vandermeer. 2010. Impacts of coffee agroforestry management on tropical bee
communities. Biological Conservation 143:1423–1431.
Kellermann, V., J. Overgaard, A.A. Hoffmann, C. Flojgaard, J.C. Svenning, and V. Loeschcke.
2012. Upper thermal limits of Drosophila are linked to species distributions and strongly
constrained phylogenetically. Proceedings of the National Academy of Sciences
109:16228–16233.
Kells, A.R., J.M. Holland, and D. Goulson. 2001. The value of uncropped field margins for
foraging bumblebees. Journal of Insect Conservation 5:283–291.
Kerr, J.T., A. Pindar, P. Galpern, L. Packer, S.G. Potts, S.M. Roberts, P. Rasmont, O. Schweiger,
S.R. Colla, L.L. Richardson, D.L. Wagner, L.F. Gall, D.S. Sikes, and A. Pantoja. 2015.
Climate change impacts on bumblebees converge across continents. Science 349:177–180.
King, C., G. Ballantyne, P.G. Willmer. 2013. Why flower visitation is a poor proxy for
pollination: measuring single-visit pollen deposition, with implications for pollination
networks and conservation. Methods in Ecology and Evolution 4:811-818.
Klein, A.M., I. Steffan-Dewenter, and T. Tscharntke. 2003a. Fruit set of highland coffee
increases with the diversity of pollinating bees. Proceedings of the Royal Society B:
Biological Sciences 270:955–961.
Klein, A.M., I. Steffan-Dewenter, and T. Tscharntke. 2003b. Pollination of Coffea canephora in
relation to local and regional agroforestry management. Journal of Applied Ecology
40:837–845. 52
Klein, A.M., B.E. Vaissière, J.H. Cane, I. Steffan-Dewenter, S.A. Cunningham, C. Kremen, and
T. Tscharntke. 2007. Importance of pollinators in changing landscapes for world crops.
Proceedings of the The Royal Society B: Biological Sciences 274:303-313.
Kovac, H., H. Kafer, A. Stabentheiner, and C. Costa. 2014. Metabolism and upper thermal limits
of Apis mellifera carnica and A. m. ligustica. Apidologie 45:664–677.
Kremen, C., N.M. Williams, M.A. Aizen, B. Gemmill-Herren, G. LeBuhn, R. Minckley, L.
Packer, S.G. Potts, T. Roulston, I. Steffan-Dewenter, D.P. Vázquez, R. Winfree, L. Adams,
E.E. Crone, S.S. Greenleaf, T.H. Keitt, A.M. Klein, J. Regetz, and T.H. Ricketts. 2007.
Pollination and other ecosystem services produced by mobile organisms: A conceptual
framework for the effects of land-use change. Ecology Letters 10:299–314.
Kremen, C., N.M. Williams, and R.W. Thorp. 2002. Crop pollination from native bees at risk
from agricultural intensification. Proceedings of the National Academy of Sciences
99:16812–16816.
Lagucki, E., J.D. Burdine, and K.E. McCluney. 2017. Urbanization alters communities of flying
arthropods in parks and gardens of a medium-sized city. PeerJ 5:e3620.
Laliberte, E., and P. Legendre. 2010. A distance-based framework for measuring functional
diversity from multiple traits. Ecology 91:299–305.
Laliberté, E., J.A. Wells, F. Declerck, D.J. Metcalfe, C.P. Catterall, C. Queiroz, I. Aubin, S.P.
Bonser, Y. Ding, J.M. Fraterrigo, S. McNamara, J.W. Morgan, D.S. Merlos, P.A. Vesk, and
M.M. Mayfield. 2010. Land-use intensification reduces functional redundancy and response
diversity in plant communities. Ecology Letters 13:76–86.
Lin, B.B., and R.A. Fuller. 2013. Sharing or sparing? How should we grow the world’s cities?
Journal of Applied Ecology 50:1161–1168. 53
Lowenstein, D.M., K.C. Matteson, and E.S. Minor. 2015. Diversity of wild bees supports
pollination services in an urbanized landscape. Oecologia 179:811–821.
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
Conservation 23:2857–2874.
Markow, T.A., L.M. Matzkin, and T.D. Watts. 2007. Desiccation resistance in four Drosophila
species: Sex and population effects. Fly 1:268–273.
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.
Matteson, K.C., J.B. Grace, and E.S. Minor. 2013. Direct and indirect effects of land use on
floral resources and flower-visiting insects across an urban landscape. Oikos 122:682–694.
Matteson, K., and G.A. Langellotto. 2010. Determinates of inner city butterfly and bee species
richness. Urban Ecosystems 13:333–347.
McCluney, K.E. 2017. Implications of animal water balance for terrestrial food webs. Current
Opinion in Insect Science 23:13–21.
McCluney, K.E., J.D. Burdine, and S.D. Frank. 2017. Variation in arthropod hydration across US
cities with distinct climate. Journal of Urban Ecology 3:1–9.
McFrederick, Q.S., and G. LeBuhn. 2006. Are urban parks refuges for bumble bees Bombus spp.
(Hymenoptera: Apidae)? Biological Conservation 129:372–382.
Michener, C.D. 2007. The bees of the world. Johns Hopkins University Press.
Moyround, E., T. Wenzel, R. Middleton, P.J. Rudall, H. Banks, A. Reed, G. Mellers, P. Killoran,
M.M. Westwood, U. Steiner, S. Vignolini, and B.J. Glover. 2017. Disorder in convergent
floral nanostructures enhances signalling to bees. Nature 550: 469-474. 54
Nguyen, C., H. Bahar, G. Baker, and N.R. Andrew. 2014. Thermal tolerance limits of
diamondback moth in ramping and plunging assays. PLoS ONE 9:e87535.
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.
Ollerton, J., R. Winfree, and S. Tarrant. 2011. How many flowering plants are pollinated by
animals? Oikos 120:321–326.
Oyen, K.J., S. Giri, and M.E. Dillon. 2016. Altitudinal variation in bumble bee (Bombus) critical
thermal limits. Journal of Thermal Biology 59:52–57.
Pardee, G.L., and S.M. Philpott. 2014. Native plants are the bee’s knees: local and landscape
predictors of bee richness and abundance in backyard gardens. Urban Ecosystems 17:641–
659.
Parmesan, C. 2006. Ecological and evolutionary responses to recent climate change. Annual
Review of Ecology, Evolution, and Systematics 37:637–669.
Pereboom, J.J.M., and J.C. Biesmeijer. 2003. Thermal constraints for stingless bee foragers: The
importance of body size and coloration. Oecologia 137:42–50.
Philpott, S.M., J. Cotton, P. Bichier, 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.
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 progress.
Journal of Environmental Management 92:331–362.
Pinheiro, J., D. Bates, S. DebRoy, and D. Sarkar. 2014. R Core Team. 2014. nlme: linear and 55
nonlinear mixed effects models. R Core Team. van der Plas, F. 2019. Biodiversity and ecosystem functioning in naturally assembled
communities. Biological Reviews.
Potter, A., and G. Lebuhn. 2015. Pollination service to urban agriculture in San Francisco, CA.
Urban Ecosystems 18:885-893.
Potts, S.G., J.C. Biesmeijer, C. Kremen, P. Neumann, O. Schweiger, and W.E. Kunin. 2010.
Global pollinator declines: Trends, impacts and drivers. Trends in Ecology and Evolution2
25:345–353.
Potts, S.G., B. Vulliamy, S. Roberts, C. O’Toole, A. Dafni, G. Ne’eman, and P. Willmer. 2005.
Role of nesting resources in organising diverse bee communities in a Mediterranean
landscape. Ecological Entomology 30:78–85.
Quistberg, R.D., P. Bichier, and S.M. Philpott. 2016. Landscape and local correlates of bee
abundance and species richness in urban gardens. Environmental Entomology 45:592–601.
Rader, R., I. Bartomeus, J.M. Tylianakis, and E. Laliberté. 2014. The winners and losers of land
use intensification: Pollinator community disassembly is non-random and alters functional
diversity. Diversity and Distributions 20:908–917.
Rader, R., B.G. Howlett, S.A. Cunningham, D.A. Westcott, L.E. Newstrom-Lloyd, M.K.
Walker, D.A.J. Teulon, and W. Edwards. 2009. Alternative pollinator taxa are equally
efficient but not as effective as the honeybee in a mass flowering crop. Journal of Applied
Ecology 46:1080–1087.
Richards, M.H., and L. Packer. 1995. Annual variation in survival and reproduction of the
primitively eusocial sweat bee Halictus ligatus (Hymenoptera: Halictidae). Canadian
Journal of Zoology 73:933–941. 56
Ricketts, T.H. 2004. Tropical forest fragments enhance pollinator activity in nearby coffee crops.
Conservation Biology 18:1262–1271.
Ricketts, T., J. Regetz, S.A. Steffan-Dewenter Cunningham, C. Kremen, A. Bogdanski, B.
Gemmill-herren, S.S. Greenleaf, A.M. Klein, M.M. Mayfield, L.A. Morandin, A. Ochieng,
and B.F. Viana. 2008. Landscape effects on crop pollination services: are there general
patterns? Ecology Letters 11:499–515.
Roulston, T.A., S.A. Smith, A.L. Brewster. 2007. A comparison of pan trap and intensive net
sampling technique for documenting a bee (Hymenotpera: Apiformes) fauna. Journal of the
Kansas Entomological Society 80: 171-181.
Scaven, V.L., and N.E. Rafferty. 2013. Physiological effects of climate warming on flowering
plants and insect pollinators and potential consequences for their interactions. Current
Zoology 59:418–426.
Schilman, P.E., J.R.B. Lighton, and D.A. Holway. 2007. Water balance in the Argentine ant
(Linepithema humile) compared with five common native ant species from southern
California. Physiological Entomology 32:1–7.
Schmidlin, T.W. 1989. The urban heat island at Toledo , Ohio. Ohio Journal of Science 89:38–
41.
Schueler, T.R. 1994. The importance of imperviousness. Watershed Protection Techniques
1:100–111.
Schweiger, O., J.C. Biesmeijer, R. Bommarco, T. Hickler, P.E. Hulme, S. Klotz, I. Kühn, M.
Moora, A. Nielsen, R. Ohlemüller, T. Petanidou, S.G. Potts, P. Pyšek, J.C. Stout, M.T.
Sykes, T. Tscheulin, M. Vilà, G.R. Walther, C. Westphal, M. Winter, M. Zobel, and J.
Settele. 2010. Multiple stressors on biotic interactions: How climate change and alien 57
species interact to affect pollination. Biological Reviews 85:777–795.
Seto, K.C., B. Güneralp, and L.R. Hutyra. 2012. Global forecasts of urban expansion to 2030 and
direct impacts on biodiversity and carbon pools. Proceedings of the National Academy of
Sciences 109:16083–16088.
Sheffield, C.S., A. Pindar, L. Packer, and P.G. Kevan. 2013. The potential of cleptoparasitic bees
as indicator taxa for assessing bee communities. Apidologie 44:501–510.
Shuler, R.E., T.H. Roulston, and G.E. Farris. 2005. Farming Practices Influence Wild pollinator
populations on squash and pumpkin. Journal of Economic Entomology 98:790–795.
Shuster, W.D., J. Bonta, H. Thurston, E. Warnemuende, and D.R. Smith. 2005. Impacts of
impervious surface on watershed hydrology: A review. Urban Water Journal 2:263–275.
Sinclair, B.J., A. Addo-Bediako, and S.L. Chown. 2003. Climatic variability and the evolution of
insect freeze tolerance. Biological Reviews 78:181–195.
Stabentheiner, A., and S. Schmaranzer. 1987. Thermographic determination of body
temperatures in honey bees and hornets, calibration and applications. Thermology 2:563–
572.
Sunday, J.M., A.E. Bates, M.R. Kearney, R.K. Colwell, N.K. Dulvy, J.T. Longino, and R.B.
Huey. 2014. Thermal-safety margins and the necessity of thermoregulatory behavior across
latitude and elevation. Proceedings of the National Academy of Sciences 111:5610–5615.
Terblanche, J.S., A.A. Hoffmann, K.A. Mitchell, L. Rako, P.C. le Roux, and S.L. Chown. 2011.
Ecologically relevant measures of tolerance to potentially lethal temperatures. Journal of
Experimental Biology 214:3713–3725.
Terblanche, J.S., and E. Kleynhans. 2009. Phenotypic plasticity of desiccation resistance in
Glossina puparia: Are there ecotype constraints on acclimation responses? Journal of 58
Evolutionary Biology 22:1636–1648.
Theodorou, P., R. Radzevičiūtė, J. Settele, O. Schweiger, T.E. Murray, and R.J. Paxton. 2016.
Pollination services enhanced with urbanization despite increasing pollinator parasitism.
Proceedings of the Royal Society B: Biological Sciences 283:20160561.
Thomas, C.D., A. Cameron, R. Green, M. Bakkenes, L. Beaurmon, Y. Collingham, B. Erasmus,
M. de Siqueira, A. Grainger, L. Hannah, L. Hughes, B. Huntley, A. van Jaarsveld, G.
Midgley, L. Miles, M. Ortega-Huerta, A. Peterson, O. Phillips, and S. Williams. 2004.
Extinction risk from climate change. Nature 427:145–148.
Tuell, J.K., and R. Isaacs. 2009. Elevated pan traps to monitor bees in flowering crop canopies.
Entomologia Experimentalis et Applicata 131:93–98.
United Nations (2018) 2018 Revision of World Urbanization Prospects (United Nations: New
York). Available at: https://population.un.org/wup/
Verble-Pearson, R.M., M.E. Gifford, and S.P. Yanoviak. 2015. Variation in thermal tolerance of
North American ants. Journal of Thermal Biology 48:65–68.
Visscher, P.K., K. Crailsheim, and G. Sherman. 1996. How do honey bees (Apis mellifera) fuel
their water foraging flights? Journal of Insect Physiology 42:1089–1094.
Walther, G.R., E. Post, P. Convey, A. Menzel, C. Parmesan, T.J. C. Beebee, J.M. Fromentin, O.
Hoegh-Guldberg, and F. Bairlein. 2002. Ecological responses to recent climate change.
Nature 416:389–395.
Weldon, C.W., C. Nyamukondiwa, M. Karsten, S.L. Chown, and J.S. Terblanche. 2018.
Geographic variation and plasticity in climate stress resistance among southern African
populations of Ceratitis capitata (Wiedemann) (Diptera: Tephritidae). Scientific Reports
8:1–13. 59
Williams, N.M., E.E. Crone, T.H. Roulston, R.L. Minckley, L. Packer, and S.G. Potts. 2010.
Ecological and life-history traits predict bee species responses to environmental
disturbances. Biological Conservation 143:2280–2291.
Williams, N.M., K.L. Ward, N. Pope, R. Isaacs, J. Wilson, E.A. May, J. Ellis, J. Daniels, A.
Pence, K. Ullmann, and J. Peters. 2015. Native wildflower plantings support wild bee
abundance and diversity in agricultural landscapes across the United States. Ecological
Applications 25:2119–2131.
Williams, S.E., L.P. Shoo, J.L. Isaac, A.A. Hoffmann, and G. Langham. 2008. Towards an
integrated framework for assessing the vulnerability of species to climate change. PLoS
Biology 6:e325.
Willmer, P., and G. Stone. 1997. Temperature and water relations in desert bees. Journal of
Thermal Biology 22:453–465.
Winfree, R., N.M. Williams, J. Dushoff, and C. Kremen. 2007. Native bees provide insurance
against ongoing honey bee losses. Ecology Letters 10:1105–1113.
Winston, M.L., and L.A. Morandin. 2005. Wild bee abundance and seed production in
conventional, organic, and genetically modified canola. Ecological Applications 15:871–
881.
Xian, G., C. Homer, J. Dewitz, J. Fry, N. Hossain, and J. Wickham. 2011. Change of impervious
surface area between 2001 and 2006 in the conterminous United States. Photogrammetric
Engineering and Remote Sensing 77:754–762.
Youngsteadt, E., A.G. Dale, A.J. Terando, R.R. Dunn, and S.D. Frank. 2015. Do cities simulate
climate change? A comparison of herbivore response to urban and global warming. Global
Change Biology 21:97–105. 60
Yuan, F., and M.E. Bauer. 2007. Comparison of impervious surface area and normalized
difference vegetation index as indicators of surface urban heat island effects in Landsat
imagery. Remote Sensing of Environment 106:375–386. 61
APPENDIX A: TABLES
Table 1. Results of linear mixed effects models comparing thermal and hygric safety margins to local (300m) and landscape (2km) imperviousness. Each model begins with interaction terms, followed by the effect of each component removed individually. Significant results with a p- value < 0.05 are in bold.
Model Component Removed df ΔAIC LRT(χ2) P-Value Hygric Safety Margin ~ Local Species * Local 2 204.51 3.361 0.186 Species 2 296.09 95.584 <0.001 Local 1 205.47 2.962 0.085 Hygric Safety Margin ~ Landscape Species * Landscape 2 200.85 9.424 0.01 Species 2 294.28 97.433 <0.001 Landscape 1 205.47 6.623 0.01 Thermal Safety Margin ~ Local Species * Local 2 183.47 0.457 0.796 Species 2 235.09 55.623 <0.001 Local 1 186.42 3.951 0.047 Thermal Safety Margin ~ Landscape Species * Landscape 2 186.67 0.570 0.752 Species 2 235.51 52.842 <0.001 Landscape 1 185.42 0.749 0.387
62
Table 2. Results of our variable selection based on correlation coefficients that includes the eight variables included in statistical models, and not-selected correlated variables that were removed from statistical models (indicated by *). Correlations that were above
0.5 are bolded.
Bare No. No. No. No. Ground Impervious Soil No. Purple White Yellow No. Red Total Canopy Herbaceous Cover Area Surface Moisture Trees Flowers Flowers* Flowers* Flowers* Flowers Cover Cover * Area 1 -0.278 -0.024 -0.161 0.467 0.432 0.077 -0.217 0.277 -0.082 0.036 -0.239 Impervious Surface -0.278 1 0.157 0.075 -0.286 -0.437 -0.109 0.081 -0.339 0.195 -0.142 0.461 Soil Moisture -0.024 0.157 1 -0.123 -0.022 0.066 0.284 0.504 0.272 -0.352 0.036 0.017 No. Trees -0.161 0.075 -0.123 1 -0.255 0.009 -0.367 -0.280 -0.349 0.284 -0.192 0.104 No. Purple Flowers 0.467 -0.286 -0.022 -0.255 1 0.199 0.119 -0.023 0.441 -0.007 0.349 -0.460 No. White Flowers* 0.432 -0.437 0.066 0.009 0.199 1 0.213 0.114 0.637 -0.316 0.084 -0.214 No. Yellow Flowers* 0.077 -0.109 0.284 -0.367 0.119 0.213 1 0.306 0.750 -0.518 0.134 -0.101 No. Red Flowers* -0.217 0.081 0.504 -0.280 -0.023 0.114 0.306 1 0.442 -0.277 0.061 0.043 No. Total Flowers 0.277 -0.339 0.272 -0.349 0.441 0.637 0.750 0.442 1 -0.434 0.273 -0.314 Canopy Cover -0.082 0.195 -0.352 0.284 -0.007 -0.316 -0.518 -0.277 -0.434 1 -0.094 0.065 Herbaceous Cover 0.036 -0.142 0.036 -0.192 0.349 0.084 0.134 0.061 0.273 -0.094 1 -0.829 Bare Ground Cover* -0.239 0.461 0.017 0.104 -0.460 -0.214 -0.101 0.043 -0.314 0.065 -0.829 1
63
Table 3. All candidate models initially used in the analysis. The most parsimonious model for each response metric (abundance, diversity, visitation rates) was determined by the lowest AICc value. Models within 2 AICc units were considered equivalent, and both models were selected.
Model K AICc ΔAICc Weight LL R2 Bee Abundance ~ Canopy Cover 3 248.37 0.00 0. 34 - 120.72 0.18 ~ Impervious Surface 3 249.28 0.91 0.22 -121.18 0.15 ~ No. Total Flowers 3 250.08 1.72 0.14 -121.58 0.13 ~ Area 3 250.51 2.15 0.12 -121.80 0.12 ~ Null Model 2 251.86 3.50 0.06 -123.71 ----- ~ No. Purple Flowers 3 252.12 3.75 0.05 -122.60 0.07 ~ Soil Moisture 3 253.04 4.68 0.03 -123.06 0.07 ~ No. Trees 3 253.90 5.53 0.02 -123.49 0.01 ~ Herbaceous Cover 3 254.32 5.95 0.02 -123.70 <0.01 ~ Full Model 10 264.86 16.50 0.00 -116.64 0.30 Bee Diversity ~ Impervious Surface 3 38.76 0.00 0.81 -15.92 0.28 ~ No. Purple Flowers 3 42.74 3.97 0.11 -17.91 0.18 ~ Null Model 2 46.41 7.65 0.02 -20.98 ----- ~ Area 3 46.46 7.70 0.02 -19.77 0.08 ~ No. Total Flowers 3 46.91 8.14 0.01 -19.99 0.06 ~ Canopy Cover 3 47.41 8.64 0.01 -20.24 0.05 ~ Herbaceous Cover 3 47.61 8.84 0.01 -20.34 0.04 ~ Soil Moisture 3 48.74 9.97 0.01 -20.91 <0.01 ~ No. Trees 3 48.89 10.12 0.01 -20.98 <0.01 ~ Full Model 10 56.38 17.62 0.00 -12.40 0.36 Visitation Frequency ~ No. Total Flowers 3 261.21 0.00 0.99 -127.15 0.57 ~ Full Model 10 270.90 9.68 0.01 -119.66 0.68 ~ Impervious Surface 3 276.15 14.93 0.00 -134.61 0.30 ~ No. Purple Flowers 3 281.56 20.34 0.00 -137.32 0.17 ~ No. Trees 3 282.80 21.58 0.00 -137.94 0.14 ~ Canopy Cover 3 283.43 22.22 0.00 -138.25 0.12 ~ Null Model 2 284.86 23.64 0.00 -140.21 ----- ~ Herbaceous Cover 3 284.87 23.66 0.00 -138.97 0.08 ~ Soil Moisture 3 285.79 24.57 0.00 -139.43 0.05 ~ Area 3 286.47 25.25 0.00 -139.77 0.03
64
Table 4. Results from Moran's I test for spatial autocorrelation.
Dependent Variable Moran’s I Statistic Moran’s I Standard Dev Expected P Value
Abundance -0.0184 0.1185 -0.0344 0.892
Diversity -0.1203 0.1125 -0.0344 0.445
Visitation -0.0142 0.1178 -0.0344 0.863
65
Table 5. Results from GLMs to test whether site type (urban parks, city gardens) influenced the effects we observed of flowers on bee abundance, diversity, and visitation rates. We found no significant differences with site type.
Dependent Factor Estimate SE T P Abundance Site Type -0.749 15.03 -0.05 0.961 Flowers 0.191 0.246 0.776 0.445 Site Type * Flowers 0.175 0.293 0.599 0.554
Diversity Site Type 0.200 0.501 0.399 0.693 Flowers 0.006 0.008 0.769 0.449 Site Type * Flowers 0.002 0.010 0.253 0.802
Visitation Rates Site Type 13.258 18.65 0.711 0.483 Flowers 1.168 0.306 3.820 <0.001 Site Type * Flowers -0.286 0.363 -0.787 0.438
66
Table 6. Results comparing bee community composition with environmental variables from our
PERMANOVA analysis. Bold indicates significance at α = 0.05.
Source Df SS MS F P Impervious Surface 1 0.58 0.58 1.99 0.01 Area 1 0.16 0.16 0.55 0.95 Soil Moisture 1 0.30 0.30 1.03 0.43 No. Trees 1 0.27 0.27 0.92 0.55 No. Purple Flowers 1 0.26 0.26 0.88 0.59 No. Total Flowers 1 0.11 0.11 0.38 0.99 Canopy Cover 1 0.23 0.23 0.77 0.74 Herbaceous Cover 1 0.18 0.18 0.63 0.88 Residuals 21 6.15 0.29 Total 29 8.63
67
Table 7. Results displaying the most parsimonious models, and the null model, for each response metric. For each response metric, we considered models within 2 AIC units to be equivalent.
Model K AICc ΔAICc Weight LL R2 Bee Abundance ~ Canopy Cover + Imperviousness 4 247.11 0.00 0.46 - 118.75 0.267 ~ Canopy Cover * Imperviousness 5 248.31 1.20 0.25 -117.91 0.298 ~ Canopy Cover 3 248.37 1.26 0.25 -120.72 0.176 ~ Null Model 2 251.86 4.76 0.04 -123.71 ----- Bee Diversity ~ No. Purple Flowers * Imperviousness 5 31.96 0.00 0.68 - 9.73 0.500 ~ No. Flowers * Imperviousness 5 33.48 1.52 0.32 -10.49 0.476 ~ Null Model 2 46.41 14.45 0.00 -20.98 ----- Visitation Frequency ~ No. Flowers + Imperviousness 4 255.59 0.00 1.00 - 122.99 0.667 ~ Null Model 2 284.86 29.27 0.00 -140.21 -----
68
Table 8. Compiled list of the 57 bee species we collected. We used information from the literature on species functional traits to aid in bee species classification (Fortel et al. 2014,
Pardee and Philpott 2014, Forrest et al. 2015).
Species Name Tongue Nesting Body Size Diet Season Agapostemon sericeus Medium Ground Medium Polylectic Spring/Summer Agapostemon splendens Medium Ground Medium Polylectic Spring/Summer Agapostemon texanus Medium Ground Large Polylectic Spring/Summer Agapostemon virescens Medium Ground Large Polylectic Spring/Summer Anthidium oblongatum Medium Cavity Medium Polylectic Summer Apis mellifera Long Cavity Large Polylectic Yearlong Augochlora pura Medium Wood Medium Polylectic Spring/Summer Augochlorella aurata Medium Ground Medium Polylectic Spring/Summer Augochlorella persimilis Short Ground Medium Polylectic Spring Augoloropsis metallica Medium Ground Medium Polylectic Spring/Summer Bombus bimaculatus Long Cavity Large Polylectic Yearlong Bombus citrinus Long Cuckoo Large Polylectic Summer Bombus fervidus Long Cavity Large Polylectic Yearlong Bombus griseocollis Long Cavity Large Polylectic Yearlong Bombus impatiens Long Cavity Large Polylectic Yearlong Colletes nudus Short Ground Large Polylectic Summer Dieunomia heteropoda Medium Ground Medium Oligolectic Spring/Summer Halictus confusus Medium Ground Medium Polylectic Spring/Summer Halictus ligatus Medium Ground Medium Polylectic Spring/Summer Halictus parallelus Medium Ground Medium Polylectic Spring/Summer Halictus rubicundus Medium Ground Medium Polylectic Spring/Summer Hoplitis producta Medium Cavity Medium Polylectic Summer Hoplitis truncata Medium Cavity Medium Polylectic Summer Hylaeus affinis Short Cavity Small Polylectic Summer Hylaeus annulatus Short Cavity Medium Polylectic Summer Hylaeus illinoisensis Short Cavity Medium Polylectic Summer Hylaeus saniculae Short Cavity Small Polylectic Summer Lasioglossum anomalum Medium Ground Medium Polylectic Spring/Summer Lasioglossum birkmanni Medium Ground Medium Polylectic Spring/Summer Lasioglossum coeruleum Medium Wood Medium Polylectic Spring Lasioglossum divergens Medium Ground Medium Polylectic Spring/Summer Lasioglossum fattigi Medium Ground Medium Polylectic Spring/Summer Lasioglossum foveolatum Medium Ground Medium Polylectic Spring/Summer Lasioglossum gotham Medium Ground Medium Polylectic Spring/Summer Lasioglossum imitatum Medium Ground Small Polylectic Spring/Summer Lasioglossum leucocomum Medium Ground Medium Polylectic Spring/Summer Lasioglossum platyparium Medium Cuckoo Small Polylectic Spring/Summer 69
Lasioglossum rufitarse Medium Ground Small Polylectic Spring/Summer Lasioglossum tegulare Medium Ground Medium Polylectic Spring/Summer Lasioglossum zephyrum Medium Ground Medium Polylectic Spring/Summer Megachile brevis Medium Cavity Large Polylectic Summer Megachile communis Medium Cavity Large Polylectic Summer Megachile concinna Medium Cavity Large Polylectic Summer Megachile pugnata Medium Cavity Large Oligolectic Summer Megachile trinodis Medium Cavity Large Polylectic Summer Melissodes agilis Long Ground Large Oligolectic Summer Melissodes bimaculata Long Ground Large Polylectic Summer Melissodes boltoniae Long Ground Large Oligolectic Summer Melissodes communis Long Ground Large Polylectic Summer Melissodes denticulatus Long Ground Large Oligolectic Summer Melissodes desponsus Long Ground Large Oligolectic Summer Melissodes trinodis Long Ground Large Oligolectic Summer Osmia simillima Medium Cavity Medium Polylectic Spring/Summer Perdita georgica Short Ground Medium Polylectic Summer Perdita octomaculata Short Ground Medium Oligolectic Summer Sphecodes banksii Short Cuckoo Medium Polylectic Spring Sphecodes illinoensis Short Cuckoo Medium Polylectic Spring Xylocopa virginica Medium Wood Large Polylectic Summer
70
Table 9. Description of the 12 bee guilds used in the analysis. Each species was added to a single guild based on 3 traits (tongue length, nesting location, body size), and the total number of collected bees from each guild is listed.
No. Guild ID Traits Species Individuals 1 LAL long tongue, aboveground, large mass 5 34 2 LBL long tongue, belowground, large mass 8 64 3 MAL medium tongue, aboveground, large mass 6 7 4 MAM medium tongue, aboveground, medium mass 6 130 5 MAS medium tongue, aboveground, small mass 1 122 6 MBL medium, tongue, belowground, large mass 2 44 7 MBM medium tongue, belowground, medium mass 19 272 8 MBS medium tongue, belowground, small mass 2 14 9 SAM short tongue, aboveground, medium mass 4 15 10 SAS short tongue, aboveground, small mass 2 13 11 SBL short tongue, belowground, large mass 1 7 12 SBM short tongue. belowground, medium mass 2 5
71
Table 10. All candidate models initially used in the analysis. The most parsimonious model for each response metric (functional dispersion, functional guild richness) was determined by the lowest AICc value. Models within 2 AICc units were considered equivalent, and both models were selected.
Model K AICc ΔAICc Weight LL R2 Functional Dispersion ~ Herbaceous Cover 2 -73.16 0.00 0.31 40.04 0.11 ~ Null Model 3 -72.17 1.00 0.19 38.3 ----- ~ No. Purple Flowers 3 -71.03 2.12 0.11 38.98 0.03 ~ Canopy Cover 3 -70.78 2.38 0.09 38.85 0.02 ~ Soil Moisture 3 -70.20 2.96 0.07 38.56 0.01 ~ No. Total Flowers 3 -70.14 3.02 0.07 38.53 <0.01 ~ Area 3 -69.77 3.39 0.06 38.35 <0.01 ~ No. Trees 3 -69.71 3.44 0.06 38.32 <0.01 ~ Impervious Surface 3 -69.70 3.45 0.05 38.31 <0.01 ~ Full Model 10 -56.62 16.54 0.00 44.10 0.25 Functional Guild Richness ~ No. Total Flowers 3 117.66 0.00 0.58 -55.37 0.28 ~ No. Purple Flowers 3 119.31 1.64 0.26 -56.19 0.24 ~ Canopy Cover 3 122.15 4.48 0.06 -57.61 0.17 ~ Impervious Surface 3 122.16 4.50 0.06 -57.62 0.17 ~ Area 3 125.27 7.61 0.01 -59.17 0.08 ~ Null Model 2 125.51 7.84 0.01 -60.53 ----- ~ No. Trees 3 126.96 9.30 0.01 -60.02 0.03 ~ Herbaceous Cover 3 127.34 9.67 0.00 -60.21 0.02 ~ Soil Moisture 3 127.96 10.29 0.00 -60.52 <0.01 ~ Full Model 10 131.98 14.31 0.00 -50.20 0.42
72
Table 11. Table displaying all candidate models for post-hoc tests that investigated guild- specific responses. Candidate models were calculated for guilds with greater than 15 individuals.
Model K AICc ΔAICc Weight LL R2 Guild LAL ~ No. Purple Flowers 3 96.89 0.00 0.61 -44.99 0.18 ~ Null Model 2 100.70 3.81 0.09 -48.13 ----- ~ Area 3 101.06 4.16 0.08 -47.07 0.07 ~ Soil Moisture 3 102.13 5.23 0.04 -47.60 0.03 ~ No. Total Flowers 3 102.16 5.27 0.04 -47.62 0.03 ~ No. Trees 3 102.43 5.53 0.04 -47.75 0.02 ~ Impervious Surface 3 102.72 5.83 0.03 -47.90 0.01 ~ Canopy Cover 3 102.90 6.00 0.03 -47.99 <0.01 ~ Herbaceous Cover 3 103.13 6.23 0.03 -48.10 <0.01 ~ Full Model 10 118.81 21.92 0.00 -43.62 0.20 Guild LBL ~ Impervious Surface 3 131.87 0.00 0.30 -62.47 0.10 ~ Null Model 2 132.51 0.64 0.22 -64.03 ----- ~ Canopy Cover 3 134.37 2.51 0.09 -63.73 0.02 ~ Herbaceous Cover 3 134.60 2.73 0.08 -63.84 0.01 ~ No. Total Flowers 3 134.90 3.03 0.07 -63.99 <0.01 ~ Soil Moisture 3 134.94 3.07 0.06 -64.01 <0.01 ~ No. Purple Flowers 3 134.95 3.08 0.06 -64.01 <0.01 ~ No. Trees 3 134.95 3.08 0.06 -64.01 <0.01 ~ Area 3 134.95 3.09 0.06 -64.02 <0.01 ~ Full Model 10 152.95 21.08 0.00 -60.68 0.15 Guild MAM ~ Canopy Cover 3 176.41 0.00 0.39 -84.74 0.20 ~ Impervious Surface 3 176.94 0.53 0.30 -85.01 0.18 ~ No. Total Flowers 3 179.07 2.66 0.10 -86.07 0.13 ~ Area 3 180.25 3.85 0.06 -86.67 0.09 ~ Soil Moisture 3 180.57 4.16 0.05 -86.82 0.08 ~ Null Model 2 180.76 4.35 0.04 -88.16 ----- ~ No. Purple Flowers 3 181.08 4.67 0.04 -87.08 0.07 73
~ No. Trees 3 182.77 6.37 0.02 -87.93 0.01 ~ Herbacous Cover 3 183.24 6.83 0.01 -88.16 <0.01 ~ Full Model 10 189.94 13.53 0.00 -79.18 0.37 Guild MAS ~ Canopy Cover 3 169.76 0.00 0.24 -81.42 0.11 ~ No. Total Flowers 3 170.10 0.34 0.20 -81.59 0.10 ~ Null Model 2 170.81 1.06 0.14 -83.18 ----- ~ Soil Moisture 3 171.36 1.60 0.11 -82.22 0.06 ~ Area 3 171.36 1.61 0.11 -82.22 0.06 ~ No. Trees 3 172.34 2.58 0.07 -82.71 0.03 ~ No. Purple Flowers 3 173.19 3.44 0.04 -83.13 <0.01 ~ Impervious Surface 3 173.23 3.47 0.04 -83.15 <0.01 ~ Herbaceous Cover 3 173.24 3.49 0.04 -83.16 <0.01 ~ Full Model 10 190.86 21.11 0.00 -79.64 0.16 Guild MBL ~ Canopy Cover 3 124.98 0.00 0.30 -59.03 0.11 ~ Area 3 126.08 1.10 0.17 -59.58 0.08 ~ Null Model 2 126.14 1.16 0.17 -60.85 ----- ~ Soil Moisture 3 127.67 2.70 0.08 -60.37 0.03 ~ Herbaceous Cover 3 127.92 2.94 0.07 -60.50 0.02 ~ No. Trees 3 128.21 3.23 0.06 -60.64 0.01 ~ No. Total Flowers 3 128.21 3.24 0.06 -60.64 0.01 ~ No. Purple Flowers 3 128.35 3.38 0.05 -60.72 <0.01 ~ Impervious Surface 3 128.61 3.64 0.05 -60.84 <0.01 ~ Full Model 10 142.30 17.32 0.00 -55.36 0.24 Guild MBM ~ Impervious Surface 3 202.67 0.00 0.28 -97.87 0.12 ~ No. Total Flowers 3 203.82 1.15 0.16 -98.45 0.08 ~ Null Model 2 204.09 1.42 0.14 -99.82 ----- ~ Area 3 204.65 1.98 0.11 -98.86 0.06 ~ Canopy Cover 3 204.79 2.12 0.10 -98.93 0.06 ~ No. Purple Flowers 3 205.33 2.67 0.07 -99.20 0.04 ~ Herbaceous Cover 3 206.02 3.35 0.05 -99.55 0.02 ~ No. Trees 3 206.29 3.63 0.05 -99.69 <0.01 ~ Soil Moisture 3 206.53 3.86 0.04 -99.80 <0.01 74
~ Full Model 10 225.02 22.35 0.00 -96.72 0.14
75
Table 12. Results displaying the best fit model for each response metric. For each response metric, we considered models within 2 AIC units to be equivalent.
Model K AICc ΔAICc Weight LL R2 Functional Guild Richness ~ Flowers * Imperviousness 5 110.46 0.00 1.00 - 48.98 0.51 ~ Null Model 2 125.51 15.05 0.00 -60.53 ----- LAL ~ Purple Flowers 3 96.89 0.00 0.52 -44.99 0.18 ~ Purple Flowers * Imperviousness 5 97.40 0.50 0.40 -42.45 0.29 ~ Null Model 2 100.70 3.81 0.08 -48.13 ----- MAM ~ Canopy Cover * Imperviousness 5 172.84 0.00 0.76 -80.17 0.39 ~ Null Model 2 180.76 7.91 0.01 -88.16 -----
76
Table 13. Results comparing functional guild community composition with environmental variables from our PERMANOVA analysis. Bold indicates significance at α = 0.05.
Environmental Factors SS MS Df F P-Value Impervious Surface 0.341 0.341 1 2.095 0.047 Area 0.061 0.061 1 0.374 0.928 Soil Moisture 0.092 0.092 1 0.565 0.756 No. Trees 0.037 0.037 1 0.229 0.965 No. Purple Flowers 0.179 0.179 1 1.105 0.348 No Total Flowers 0.074 0.074 1 0.454 0.873 Canopy Cover 0.106 0.106 1 0.649 0.719 Herbaceous Cover 0.130 0.130 1 0.801 0.571 Residuals 3.418 0.163 21 Total 4.836 29
77
Table 14. Results from glm model comparisons showing which bee community metrics
(functional guild richness, bee abundance, bee diversity, functional trait diversity) were the stronger predictor of visitation frequency.
Model K AICc ΔAICc Weight LL R2 Visitation Frequency ~ Functional Guild Richness 3 273.57 0.00 0.92 -133.32 0.36 ~ Bee Abundance 3 279.32 5.75 0.05 -136.20 0.23 ~ Bee Diversity 3 280.74 7.17 0.03 -136.07 0.19 ~ Functional Trait Diversity (FDis) 3 287.07 13.50 0.00 -140.07 0.01
78
APPENDIX B: FIGURES
Figure 1. Map displaying site locations in Toledo, Ohio, with reference to the state of Ohio.
Base layer coloration indicates percent impervious surface. Darker shades of red represent high impervious surface, and light shades represent low impervious surface. CTmax and CWC were determined for square sites (3 urban, 3 rural). Field measurements body temperature and water content were measured at all 19 sites. This map was created using the 2011 National Landcover
Dataset percent developed imperviousness layer (Xian et al. 2011, Homer et al. 2015) in ArcGIS
10.3.1.
79
Figure 2. Differences across bee species for thermal and hygric limits. (A) CTmax was significantly different between species (χ2 = 6.56; p = 0.038), and bumble bees had a higher
CTmax than honey bees and sweat bees; (B) CWC was significantly different between all species (χ2 = 77.81; p < 0.001) and sweat bees had the lowest CWC; (C) thermal safety margins were significantly different across all species (χ2 = 55.62; p < 0.001), and honey bees had the largest thermal safety margin; and (D) hygric safety margin was significantly different across all species (χ2 = 97.433; p < 0.001), and sweat bees had the largest hygric safety.
80
2 Figure 3. CTmax between site class (χ = 3.84, p = 0.05). The three investigated bee species at urban sites (52.39°C ± 0.77) and rural sites (49.57°C ± 8 1.13). Each data point represents an individual bee.
81
Figure 4. Significant relationships between safety margins (thermal and hygric) and imperviousness (local and landscape) for each bee species. (A) Honeybee hygric safety margin is significantly associated with landscape imperviousness at α = 0.05 (χ2 = 8.847, p = 0.003, R2 =
0.19); (B) Honeybee hygric safety margin is also associated with local imperviousness at the α =
0.1 level (χ2 = 2.81, p = 0.094, R2 = 0.04); (C) Sweat bee thermal safety margin is significantly associated with local imperviousness (χ2 = 6.28, p = 0.012, R2 = 0.49); and (D) Bumblebee thermal safety margin is significantly associated with local imperviousness (χ2 = 4.905, p =
0.027).
82
Figure 5. Map displaying site locations and percent impervious surface data in Toledo, Ohio.
Darker colors indicate high impervious surface and light colors indicate low impervious surface.
This map was created using the 2011 National Landcover Dataset percent developed imperviousness layer (Xian et al. 2011, Homer et al. 2015) in ArcGIS 10.3.1.
83
Figure 6. Nonmetric multidimensional scaling (NMDS) analysis for bee species sampled in
Toledo, Ohio (USA). (A) Each bee species is represented by single point, and environmental factors are represented by arrows. (B) Impervious surface was significantly associated with community composition.
84
Figure 7. Panel figure displaying associations between environmental factors and bee abundance
(A-B), and visitation rate (C-D).
85
Figure 8. Panel figure displaying interaction plots. (A) Relationship between impervious surface and bee abundance when canopy cover is at a high (+ 1 SD), low (-1 SD), or medium level
(mean). (B) Relationship between impervious surface and bee diversity when purple flower abundance is at a high (+ 1 SD), low (-1 SD), or medium level (mean).
86
Figure 9. Panel figure displaying interaction plots. (A) Relationship between impervious surface and functional guild richness when the total number of flowers is high (+ 1 SD), low (-1 SD), or medium level (mean). (B) Relationship between impervious surface and functional guilds LAL abundance when the number of purple flowers high (+ 1 SD), low (-1 SD), or medium (mean).
87
Figure 10. Non-metric multidimensional scaling (NMDS) analysis for bee guilds sampled in
Toledo, OH. (A) All habitat characteristics were plotted with arrows, and bee guild were plotted with points. (B) Impervious surface was the only habitat characteristic that influenced guild composition. (C) Visitation frequencies were plotted to show how guilds influenced pollination.
88
Figure 11. Panel figure displaying associations between environmental factors and within-guild bee abundance. (A) Guild MAM negatively associated with percent canopy cover and (B) percent impervious surface. (C) Guild LAL (bumble bees, honey bees) positively associated with the number of purple flowers.
89
Figure 12. Panel figure displaying associations between visitation rates and bee community metrics: (A) guild richness, (B) functional dispersion, (C) bee abundance, and (D) bee diversity.
90
APPENDIX C: CHAPTER I SUPPLEMENTARY INFORMATION
2 S1. CTmax between site class (χ = 3.84, p = 0.05). The three investigated bee species at urban sites (52.39°C ± 0.77) and rural sites (49.57°C ± 8 1.13). Each data point represents an individual bee.
91
S2. Panel figure displaying the relationship between temperature (CTmax and Field Body
Temperature) and percent imperviousness surface for the three investigated bee species. We show these relationships at the scale of (A) local (300 m) percent impervious surface, and (B) landscape (2000 m) percent impervious surface.
92
S3. Panel figure displaying the relationship for the three investigated bee species between local
(300 m) percent impervious surface and (A) field body water content, and (B) critical water content.
93
S4. Panel figure displaying the relationship for the three investigated bee species between landscape (2000 m) percent impervious surface and (A) field body water content, and (B) critical water content.
94
S5. Table displaying all data used for the calculation of critical thermal maximum (CTmax). The table displays site names (Site), site geographical location (Lat = latitude; Long=longitude), and classification (Class). The table also includes the mass (Mass) of each bee (in mg), species name
(Species), and the recorded CTmax (in °C).
Record Site Lat Long Class Mass Species CTmax 1 MCD 41.641 -83.513 Urban 5.357 Agapostemon sericeus 54.5 2 DOY 41.681 -83.524 Urban 29.554 Bombus impatiens 52.5 3 CHE 41.657 -83.549 Urban 34.161 Bombus impatiens 49 4 SYL 41.696 -83.752 Rural 45.836 Bombus impatiens 48.5 5 BU 41.548 -83.671 Rural 18.905 Apis mellifera 40.5 6 BU 41.548 -83.671 Rural 19.342 Apis mellifera 44.5 7 CHE 41.657 -83.549 Urban 21.593 Apis mellifera 55.5 8 SV 41.682 -83.717 Rural 19.382 Apis mellifera 40.5 9 MCD 41.641 -83.513 Urban 23.622 Apis mellifera 56.5 10 DOY 41.681 -83.524 Urban 55.572 Bombus impatiens 57.5 11 MCD 41.641 -83.513 Urban 23.371 Apis mellifera 47 12 SV 41.682 -83.717 Rural 18.749 Apis mellifera 56 13 SV 41.682 -83.717 Rural 36.974 Bombus impatiens 56 14 SYL 41.696 -83.752 Rural 18.858 Apis mellifera 53.5 15 MCD 41.641 -83.513 Urban 39.483 Bombus impatiens 53.5 16 DOY 41.681 -83.524 Urban 32.692 Bombus impatiens 55 17 MCD 41.641 -83.513 Urban 25.104 Apis mellifera 56 18 SYL 41.696 -83.752 Rural 22.023 Apis mellifera 56.5 19 DOY 41.681 -83.524 Urban 50.401 Bombus impatiens 59.5 20 BU 41.548 -83.671 Rural 24.836 Apis mellifera 56 21 CHE 41.657 -83.549 Urban 23.619 Apis mellifera 44 22 SV 41.682 -83.717 Rural 38.181 Bombus impatiens 57.5 23 BU 41.548 -83.671 Rural 21.565 Apis mellifera 54 24 SYL 41.696 -83.752 Rural 20.276 Apis mellifera 55 25 SYL 41.696 -83.752 Rural 52.24 Bombus impatiens 53 26 BU 41.548 -83.671 Rural 35.834 Bombus impatiens 50 27 SV 41.682 -83.717 Rural 17.35 Apis mellifera 30.5 28 MCD 41.641 -83.513 Urban 24.865 Apis mellifera 55 29 BU 41.548 -83.671 Rural 44.017 Bombus impatiens 50.5 30 BU 41.548 -83.671 Rural 40.947 Bombus impatiens 52 31 MCD 41.641 -83.513 Urban 7.891 Agapostemon sericeus 49 32 SYL 41.696 -83.752 Rural 60.597 Bombus impatiens 30.5 33 SV 41.682 -83.717 Rural 44.457 Bombus impatiens 53.5 34 SV 41.682 -83.717 Rural 23.211 Apis mellifera 55.5 95
35 SYL 41.696 -83.752 Rural 17.328 Apis mellifera 30.5 36 SV 41.682 -83.717 Rural 17.907 Apis mellifera 30.5 37 DOY 41.681 -83.524 Urban 38.585 Bombus impatiens 51 38 SYL 41.696 -83.752 Rural 54.37 Bombus impatiens 58.5 39 SV 41.682 -83.717 Rural 33.609 Bombus impatiens 40.5 40 SYL 41.696 -83.752 Rural 19.036 Apis mellifera 51 41 BU 41.548 -83.671 Rural 49.561 Bombus impatiens 55 42 CHE 41.657 -83.549 Urban 18.874 Apis mellifera 56 43 BU 41.548 -83.671 Rural 25.751 Bombus impatiens 57 44 MCD 41.641 -83.513 Urban 25.367 Apis mellifera 47 45 MCD 41.641 -83.513 Urban 24.608 Bombus impatiens 56.5 46 BU 41.548 -83.671 Rural 21.604 Apis mellifera 55.5 47 MCD 41.641 -83.513 Urban 99.73 Bombus impatiens 51 48 SYL 41.696 -83.752 Rural 65.474 Bombus impatiens 53.5 49 MCD 41.641 -83.513 Urban 35.3 Bombus impatiens 57.5 50 CHE 41.657 -83.549 Urban 47.678 Bombus impatiens 52 51 CHE 41.657 -83.549 Urban 22.892 Apis mellifera 55.5 52 CHE 41.657 -83.549 Urban 56.026 Bombus impatiens 58 53 SV 41.682 -83.717 Rural 38.147 Bombus impatiens 58 54 CHE 41.657 -83.549 Urban 19.78 Apis mellifera 54 55 MCD 41.641 -83.513 Urban 1.057 Agapostemon sericeus 49 56 SYL 41.696 -83.752 Rural 2.388 Agapostemon sericeus 47 57 SYL 41.696 -83.752 Rural 4.081 Agapostemon sericeus 47.5 58 SV 41.682 -83.717 Rural 4.863 Agapostemon sericeus 50 59 CHE 41.657 -83.549 Urban 4.2 Agapostemon sericeus 48.5 60 SYL 41.696 -83.752 Rural 2.552 Agapostemon sericeus 47.5 61 BU 41.548 -83.671 Rural 1.384 Agapostemon sericeus 46.5 62 SV 41.682 -83.717 Rural 2 Agapostemon sericeus 49.5 63 SYL 41.696 -83.752 Rural 2.053 Agapostemon sericeus 47.5 64 CHE 41.657 -83.549 Urban 7.078 Agapostemon sericeus 52 65 BU 41.548 -83.671 Rural 2.599 Agapostemon sericeus 54 66 SV 41.682 -83.717 Rural 8.731 Agapostemon sericeus 56 67 CHE 41.657 -83.549 Urban 6.705 Agapostemon sericeus 52.5 68 SV 41.682 -83.717 Rural 3.819 Agapostemon sericeus 50.5 69 CHE 41.657 -83.549 Urban 4.637 Agapostemon sericeus 50 70 BU 41.548 -83.671 Rural 6.89 Agapostemon sericeus 47.5 71 BU 41.548 -83.671 Rural 5.281 Agapostemon sericeus 52 72 DOY 41.681 -83.524 Urban 1.141 Agapostemon sericeus 53.5 73 DOY 41.681 -83.524 Urban 1.91 Agapostemon sericeus 49 74 DOY 41.681 -83.524 Urban 1.896 Agapostemon sericeus 51 75 SYL 41.696 -83.752 Rural 2.69 Agapostemon sericeus 52.5 76 SV 41.682 -83.717 Rural 7.976 Agapostemon sericeus 52 77 MCD 41.641 -83.513 Urban 8.384 Agapostemon sericeus 47 96
78 BU 41.548 -83.671 Rural 3.915 Agapostemon sericeus 46.5 79 MCD 41.641 -83.513 Urban 1.48 Agapostemon sericeus 49 80 DOY 41.681 -83.524 Urban 22.852 Apis mellifera 49.5 81 DOY 41.681 -83.524 Urban 25.399 Apis mellifera 43.5 82 DOY 41.681 -83.524 Urban 20.537 Apis mellifera 53 83 DOY 41.681 -83.524 Urban 19.981 Apis mellifera 58 84 DOY 41.681 -83.524 Urban 23.449 Apis mellifera 32.5 85 DOY 41.681 -83.524 Urban 2.749 Agapostemon sericeus 52.5 86 DOY 41.681 -83.524 Urban 1.865 Agapostemon sericeus 51 87 MCD 41.641 -83.513 Urban 49.251 Bombus impatiens 60 88 CHE 41.657 -83.549 Urban 2.355 Agapostemon sericeus 53.5
97
S6. Table displaying all data used for the calculation of critical water content (CWC). The table displays site names (Site), geographical locations (Lat = latitude; Long=longitude), and classification (Class). The table also includes the mass (Mass) of each bee (in mg), species name
(Species), desiccation survival time in hours (Time), and CWC.
Record Site Lat Long Class Mass Species Time CWC 1 DOY 41.681 -83.524 Urban 39.756 Bombus impatiens 23 65.348 2 MCD 41.641 -83.513 Urban 45.306 Bombus impatiens 32 58.953 3 SV 41.682 -83.717 Rural 34.277 Bombus impatiens 41 63.718 4 SV 41.682 -83.717 Rural 37.028 Bombus impatiens 8 60.647 5 BU 41.548 -83.671 Rural 21.421 Apis mellifera 8 71.314 6 BU 41.548 -83.671 Rural 11.555 Agapostemon sericeus 8 35.262 7 MCD 41.641 -83.513 Urban 20.046 Apis mellifera 9 73.636 8 SV 41.682 -83.717 Rural 6.028 Agapostemon sericeus 8 33.119 9 BU 41.548 -83.671 Rural 21.456 Apis mellifera 3 70.375 10 CHE 41.657 -83.549 Urban 1.731 Agapostemon sericeus 17 48.065 11 CHE 41.657 -83.549 Urban 21.161 Apis mellifera 0.5 68.021 12 BU 41.548 -83.671 Rural 2.862 Agapostemon sericeus 8 33.750 13 BU 41.548 -83.671 Rural 31.701 Bombus impatiens 23 61.214 14 CHE 41.657 -83.549 Urban 6.439 Agapostemon sericeus 8 53.003 15 BU 41.548 -83.671 Rural 20.475 Apis mellifera 8 67.889 16 DOY 41.681 -83.524 Urban 32.048 Bombus impatiens 23 60.384 17 MCD 41.641 -83.513 Urban 0.364 Agapostemon sericeus 23 72.937 18 BU 41.548 -83.671 Rural 38.442 Bombus impatiens 23 57.384 19 CHE 41.657 -83.549 Urban 0.962 Agapostemon sericeus 17 21.277 20 SV 41.682 -83.717 Rural 41.514 Bombus impatiens 17 60.319 21 MCD 41.641 -83.513 Urban 22.729 Apis mellifera 9 72.759 22 CHE 41.657 -83.549 Urban 18.207 Apis mellifera 8 76.906 23 SYL 41.696 -83.752 Rural 51.277 Bombus impatiens 8 64.622 24 DOY 41.681 -83.524 Urban 2.319 Agapostemon sericeus 17 64.943 25 SV 41.682 -83.717 Rural 31.339 Bombus impatiens 3 60.699 26 SV 41.682 -83.717 Rural 20.531 Apis mellifera 210 82.654 27 BU 41.548 -83.671 Rural 20.062 Apis mellifera 2 77.912 28 SYL 41.696 -83.752 Rural 51.618 Bombus impatiens 17 61.031 29 SYL 41.696 -83.752 Rural 7.329 Agapostemon sericeus 8 52.111 30 DOY 41.681 -83.524 Urban 34.945 Bombus impatiens 9 63.174 31 SYL 41.696 -83.752 Rural 20.092 Apis mellifera 3 81.695 32 DOY 41.681 -83.524 Urban 20.495 Apis mellifera 9 74.436 33 MCD 41.641 -83.513 Urban 21.79 Apis mellifera 1 70.007 34 BU 41.548 -83.671 Rural 4.398 Agapostemon sericeus 8 29.778 98
35 DOY 41.681 -83.524 Urban 6.815 Agapostemon sericeus 9 48.017 36 BU 41.548 -83.671 Rural 30.782 Bombus impatiens 17 63.770 37 MCD 41.641 -83.513 Urban 35.172 Bombus impatiens 9 63.651 38 DOY 41.681 -83.524 Urban 30.05 Bombus impatiens 32 65.476 39 DOY 41.681 -83.524 Urban 20.527 Apis mellifera 9 81.231 40 BU 41.548 -83.671 Rural 21.377 Apis mellifera 8 69.389 41 SYL 41.696 -83.752 Rural 31.193 Bombus impatiens 8 65.085 42 SYL 41.696 -83.752 Rural 31.582 Bombus impatiens 17 64.932 43 CHE 41.657 -83.549 Urban 7.452 Agapostemon sericeus 17 46.218 44 DOY 41.681 -83.524 Urban 21.905 Apis mellifera 9 76.695 45 DOY 41.681 -83.524 Urban 20.492 Apis mellifera 9 79.572 46 BU 41.548 -83.671 Rural 4.488 Agapostemon sericeus 17 37.240 47 BU 41.548 -83.671 Rural 41.181 Bombus impatiens 23 64.920 48 DOY 41.681 -83.524 Urban 18.175 Apis mellifera 9 82.232 49 SYL 41.696 -83.752 Rural 22.404 Apis mellifera 2 66.118 50 SV 41.682 -83.717 Rural 18.49 Apis mellifera 0.25 67.969 51 SV 41.682 -83.717 Rural 17.414 Apis mellifera 0.25 71.124 52 SV 41.682 -83.717 Rural 16.271 Apis mellifera 0.25 67.600 53 SYL 41.696 -83.752 Rural 44.559 Bombus impatiens 0.5 64.309 54 SYL 41.696 -83.752 Rural 21.254 Apis mellifera 8 82.038 55 SYL 41.696 -83.752 Rural 21.05 Apis mellifera 1.5 76.566 56 DOY 41.681 -83.524 Urban 32.866 Bombus impatiens 41 60.817 57 SV 41.682 -83.717 Rural 19.046 Apis mellifera 2 71.350 58 SYL 41.696 -83.752 Rural 19.264 Apis mellifera 3 62.137 59 SYL 41.696 -83.752 Rural 5.78 Agapostemon sericeus 2 48.259 60 SV 41.682 -83.717 Rural 5.297 Agapostemon sericeus 8 48.473 61 SV 41.682 -83.717 Rural 0.846 Agapostemon sericeus 3 77.824 62 SV 41.682 -83.717 Rural 5.332 Agapostemon sericeus 1.5 53.199 63 DOY 41.681 -83.524 Urban 6.577 Agapostemon sericeus 95 69.513 64 MCD 41.641 -83.513 Urban 3.746 Agapostemon sericeus 95 53.338 65 MCD 41.641 -83.513 Urban 2.137 Agapostemon sericeus 95 57.825 66 MCD 41.641 -83.513 Urban 1.068 Agapostemon sericeus 95 51.322 67 SV 41.682 -83.717 Rural 4.541 Agapostemon sericeus 168 50.690 68 BU 41.548 -83.671 Rural 8.958 Agapostemon sericeus 9 56.426 69 CHE 41.657 -83.549 Urban 6.055 Agapostemon sericeus 48 64.384 70 MCD 41.641 -83.513 Urban 7.271 Agapostemon sericeus 48 60.220 71 DOY 41.681 -83.524 Urban 5.101 Agapostemon sericeus 72 63.866 72 DOY 41.681 -83.524 Urban 10.894 Agapostemon sericeus 96 57.260 73 SYL 41.696 -83.752 Rural 2.41 Agapostemon sericeus 17 57.186 74 SYL 41.696 -83.752 Rural 5.944 Agapostemon sericeus 48 50.343 75 SYL 41.696 -83.752 Rural 5.065 Agapostemon sericeus 120 53.841 76 CHE 41.657 -83.549 Urban 18.76 Apis mellifera 2 68.596 77 CHE 41.657 -83.549 Urban 18.779 Apis mellifera 9 79.758 99
78 CHE 41.657 -83.549 Urban 19.355 Apis mellifera 9 77.230 79 MCD 41.641 -83.513 Urban 21.001 Apis mellifera 9 78.046 80 MCD 41.641 -83.513 Urban 20.246 Apis mellifera 9 80.826 81 CHE 41.657 -83.549 Urban 31.175 Bombus impatiens 9 64.186 82 CHE 41.657 -83.549 Urban 34.536 Bombus impatiens 9 66.482 83 CHE 41.657 -83.549 Urban 29.883 Bombus impatiens 17 67.037 84 CHE 41.657 -83.549 Urban 36.577 Bombus impatiens 72 64.723 85 CHE 41.657 -83.549 Urban 38.48 Bombus impatiens 23 67.939 86 MCD 41.641 -83.513 Urban 58.373 Bombus impatiens 17 66.656 87 MCD 41.641 -83.513 Urban 42.144 Bombus impatiens 17 64.755 88 MCD 41.641 -83.513 Urban 55.384 Bombus impatiens 9 65.408 89 SV 41.682 -83.717 Rural 31.558 Bombus impatiens 17 68.329
100
S7. Table displaying all data used for the calculation of thermal safety margin (TSM). The table displays site names (Site), geographical locations (Lat = latitude; Long=longitude), and measurements of local (Local: 300m) and landscape (Landscape: 2000m) percent impervious surface. The table also includes the mass (Mass) of each bee (in mg), species name (Species), field body temperatures (Temp: in °C), and TSM (in °C).
Record Site Lat Long Local Landscape Mass Species Temp TSM 1 TV 41.607 -83.657 36.93 40.05 32.793 Bombus impatiens 27.3 31.3 2 TV 41.607 -83.657 36.93 40.05 32.455 Bombus impatiens 29.7 28.9 3 TV 41.607 -83.657 36.93 40.05 26.512 Bombus impatiens 30 28.6 4 TV 41.607 -83.657 36.93 40.05 44.711 Bombus impatiens 30.2 28.4 5 TUC 41.637 -83.654 33.14 40.85 20.823 Apis mellifera 31.2 25.25 6 TV 41.607 -83.657 36.93 40.05 59.409 Bombus impatiens 33 25.6 7 TUC 41.637 -83.654 33.14 40.85 59.379 Bombus impatiens 29 29.6 8 MNR 41.670 -83.584 40.10 45.92 22.457 Apis mellifera 29.9 26.55 9 TV 41.607 -83.657 36.93 40.05 54.005 Bombus impatiens 31.2 27.4 10 TV 41.607 -83.657 36.93 40.05 4.448 Agapostemon sericeus 30.7 23.25 11 ZL 41.621 -83.585 47.74 44.26 66.098 Bombus impatiens 31.6 27 12 TV 41.607 -83.657 36.93 40.05 4.427 Agapostemon sericeus 29 24.95 13 TV 41.607 -83.657 36.93 40.05 5.347 Agapostemon sericeus 30.6 23.35 14 TV 41.607 -83.657 36.93 40.05 46.064 Bombus impatiens 30.4 28.2 15 TV 41.607 -83.657 36.93 40.05 34.796 Apis mellifera 30.6 25.85 16 OWN 41.584 -83.541 53.19 52.55 25.635 Apis mellifera 31.3 25.15 17 INT 41.647 -83.527 59.28 61.33 31.604 Apis mellifera 30.9 25.55 18 FX 41.699 -83.628 28.28 43.28 40.832 Bombus impatiens 28.3 30.3 19 INT 41.647 -83.527 59.28 61.33 53.571 Bombus impatiens 29.44 29.16 20 MNR 41.670 -83.584 40.10 45.92 25.44 Apis mellifera 31.7 24.75 21 FX 41.699 -83.628 28.28 43.28 25.552 Apis mellifera 26.2 30.25 22 UT 41.659 -83.621 42.99 39.05 27.337 Bombus impatiens 30.8 27.8 23 FX 41.699 -83.628 28.28 43.28 28.687 Bombus impatiens 26.9 31.7 24 FX 41.699 -83.628 28.28 43.28 27.506 Bombus impatiens 27.4 31.2 25 INT 41.647 -83.527 59.28 61.33 24.155 Apis mellifera 29 27.45 26 INT 41.647 -83.527 59.28 61.33 20.729 Apis mellifera 30.1 26.35 27 CC 41.591 -83.441 16.24 32.16 31.871 Apis mellifera 28.1 28.35 28 UT 41.659 -83.621 42.99 39.05 54.764 Bombus impatiens 29.3 29.3 29 INT 41.647 -83.527 59.28 61.33 34.325 Bombus impatiens 31 27.6 30 CC 41.591 -83.441 16.24 32.16 77.09 Bombus impatiens 27.8 30.8 31 FX 41.699 -83.628 28.28 43.28 24.895 Apis mellifera 27.9 28.55 32 FX 41.699 -83.628 28.28 43.28 22.077 Apis mellifera 27.3 29.15 33 FX 41.699 -83.628 28.28 43.28 27 Bombus impatiens 27.7 30.9 34 FX 41.699 -83.628 28.28 43.28 22.132 Bombus impatiens 26.9 31.7 35 FX 41.699 -83.628 28.28 43.28 22.589 Apis mellifera 29.2 27.25 36 FX 41.699 -83.628 28.28 43.28 34.476 Apis mellifera 26.6 29.85 101
37 TUC 41.637 -83.654 33.14 40.85 27.069 Apis mellifera 30.3 26.15 38 TUC 41.637 -83.654 33.14 40.85 26.134 Apis mellifera 29.7 26.75 39 TUC 41.637 -83.654 33.14 40.85 28.299 Apis mellifera 31.4 25.05 40 TUC 41.637 -83.654 33.14 40.85 24.915 Apis mellifera 31.4 25.05 41 CC 41.591 -83.441 16.24 32.16 27.523 Bombus impatiens 26.1 32.5 42 CC 41.591 -83.441 16.24 32.16 34.814 Bombus impatiens 29.1 29.5 43 CHS 41.684 -83.496 37.03 48.87 21.592 Apis mellifera 27.3 29.15 44 TUC 41.637 -83.654 33.14 40.85 23.84 Apis mellifera 30.7 25.75 45 COL 41.663 -83.483 20.24 44.06 46.352 Bombus impatiens 29.7 28.9 46 CC 41.591 -83.441 16.24 32.16 5.962 Agapostemon sericeus 29 24.95 47 ZP 41.618 -83.578 53.06 43.30 17.094 Apis mellifera 28.7 27.75 48 MNR 41.670 -83.584 40.10 45.92 6.73 Agapostemon sericeus 30.1 23.85 49 COL 41.663 -83.483 20.24 44.06 5.247 Agapostemon sericeus 29.3 24.65 50 COL 41.663 -83.483 20.24 44.06 10.287 Agapostemon sericeus 28.1 25.85 51 CC 41.591 -83.441 16.24 32.16 2.789 Agapostemon sericeus 26.1 27.85 52 CC 41.591 -83.441 16.24 32.16 6.249 Agapostemon sericeus 28.5 25.45
102
S8. Table displaying all data used for the calculation of hygric safety margin (HSM). The table displays site names (Site), geographical locations (Lat = latitude; Long=longitude), and measurements of local (Local: 300m) and landscape (Landscape: 2000m) percent impervious surface. The table also includes the mass (Mass) of each bee (in mg), species name (Species), and HSM (%).
Record Site Lat Long Local Landscape Mass Species HSM 1 TV 41.607 -83.657 36.935 40.055 27.586 Bombus impatiens 5.69 2 SV 41.682 -83.717 20.878 35.425 19.13 Apis mellifera 2.55 3 TV 41.607 -83.657 36.935 40.055 32.793 Bombus impatiens 7.18 4 TV 41.607 -83.657 36.935 40.055 32.455 Bombus impatiens 8.24 5 FTU 41.611 -83.619 47.690 39.347 40.281 Bombus impatiens 4.91 6 FTU 41.611 -83.619 47.690 39.347 45.233 Bombus impatiens 7.15 7 CHE 41.657 -83.549 73.297 58.649 37.813 Bombus impatiens 6.81 8 FTU 41.611 -83.619 47.690 39.347 40.262 Bombus impatiens 5.82 9 FTU 41.611 -83.619 47.690 39.347 43.961 Bombus impatiens 5.94 10 SYL 41.696 -83.752 32.000 31.829 5.754 Agapostemon sericeus 30.04 11 SYL 41.696 -83.752 32.000 31.829 90.22 Bombus impatiens 6.54 12 SYL 41.696 -83.752 32.000 31.829 26.119 Apis mellifera 1.90 13 SV 41.682 -83.717 20.878 35.425 21.034 Apis mellifera 3.92 14 DOY 41.681 -83.524 48.658 56.269 70.387 Bombus impatiens 5.32 15 DOY 41.681 -83.524 48.658 56.269 62.51 Bombus impatiens 14.10 16 CHS 41.684 -83.496 37.025 48.875 28.153 Bombus impatiens 1.89 17 TV 41.607 -83.657 36.935 40.055 26.512 Bombus impatiens 3.51 18 TV 41.607 -83.657 36.935 40.055 44.711 Bombus impatiens 2.44 19 ZL 41.621 -83.585 47.741 44.260 74.237 Bombus impatiens 4.69 20 MNR 41.670 -83.584 40.104 45.924 66.921 Bombus impatiens 2.74 21 COL 41.663 -83.483 20.238 44.064 20.003 Apis mellifera 0.58 22 ZL 41.621 -83.585 47.741 44.260 48.979 Bombus impatiens 1.70 23 TV 41.607 -83.657 36.935 40.055 59.409 Bombus impatiens 5.37 24 TUC 41.637 -83.654 33.143 40.848 59.379 Bombus impatiens 3.32 25 MNR 41.670 -83.584 40.104 45.924 31.452 Bombus impatiens 0.45 26 ZL 41.621 -83.585 47.741 44.260 14.045 Agapostemon sericeus 25.88 27 FTU 41.611 -83.619 47.690 39.347 31.978 Apis mellifera 4.79 28 TV 41.607 -83.657 36.935 40.055 4.448 Agapostemon sericeus 28.37 29 ZL 41.621 -83.585 47.741 44.260 66.098 Bombus impatiens 0.93 30 TV 41.607 -83.657 36.935 40.055 4.427 Agapostemon sericeus 5.16 31 TV 41.607 -83.657 36.935 40.055 5.347 Agapostemon sericeus 21.45 32 TV 41.607 -83.657 36.935 40.055 46.064 Bombus impatiens 4.26 103
33 FTU 41.611 -83.619 47.690 39.347 25.814 Apis mellifera 8.57 34 FX 41.699 -83.628 28.283 43.279 40.832 Bombus impatiens 2.40 35 ZL 41.621 -83.585 47.741 44.260 56.855 Bombus impatiens 4.08 36 FX 41.699 -83.628 28.283 43.279 25.552 Apis mellifera 2.14 37 UT 41.659 -83.621 42.991 39.053 27.337 Bombus impatiens 6.74 38 FX 41.699 -83.628 28.283 43.279 28.687 Bombus impatiens 8.38 39 FX 41.699 -83.628 28.283 43.279 27.506 Bombus impatiens 7.86 40 INT 41.647 -83.527 59.284 61.328 24.155 Apis mellifera 0.61 41 INT 41.647 -83.527 59.284 61.328 20.729 Apis mellifera 1.33 42 FTU 41.611 -83.619 47.690 39.347 4.535 Agapostemon sericeus 19.49 43 CC 41.591 -83.441 16.245 32.162 31.871 Apis mellifera 4.83 44 UT 41.659 -83.621 42.991 39.053 54.764 Bombus impatiens 3.97 45 INT 41.647 -83.527 59.284 61.328 34.325 Bombus impatiens 1.62 46 CC 41.591 -83.441 16.245 32.162 77.09 Bombus impatiens 6.30 47 FX 41.699 -83.628 28.283 43.279 24.895 Apis mellifera 2.67 48 FX 41.699 -83.628 28.283 43.279 22.077 Apis mellifera 2.36 49 FX 41.699 -83.628 28.283 43.279 27 Bombus impatiens 6.09 50 FX 41.699 -83.628 28.283 43.279 22.132 Bombus impatiens 3.30 51 FX 41.699 -83.628 28.283 43.279 22.589 Apis mellifera 4.42 52 INT 41.647 -83.527 59.284 61.328 21.943 Apis mellifera 0.09 53 MNR 41.670 -83.584 40.104 45.924 19.058 Apis mellifera 4.85 54 MNR 41.670 -83.584 40.104 45.924 36.722 Bombus impatiens 3.09 55 MNR 41.670 -83.584 40.104 45.924 18.725 Apis mellifera 0.37 56 TUC 41.637 -83.654 33.143 40.848 24.915 Apis mellifera 0.87 57 CC 41.591 -83.441 16.245 32.162 27.523 Bombus impatiens 4.73 58 CC 41.591 -83.441 16.245 32.162 34.814 Bombus impatiens 6.61 59 BWN 41.699 -83.589 41.200 46.614 42.465 Bombus impatiens 4.92 60 CHE 41.657 -83.549 73.297 58.649 95.481 Bombus impatiens 5.21 61 UT 41.659 -83.621 42.991 39.053 23.102 Apis mellifera 0.80 62 BWN 41.699 -83.589 41.200 46.614 61.655 Bombus impatiens 2.86 63 MNR 41.670 -83.584 40.104 45.924 31.541 Bombus impatiens 16.11 64 TUC 41.637 -83.654 33.143 40.848 23.84 Apis mellifera 0.81 65 UT 41.659 -83.621 42.991 39.053 23.946 Apis mellifera 1.97 66 CHE 41.657 -83.549 73.297 58.649 36.159 Bombus impatiens 5.76 67 COL 41.663 -83.483 20.238 44.064 46.352 Bombus impatiens 6.30 68 UT 41.659 -83.621 42.991 39.053 21.7 Apis mellifera 3.96 69 CC 41.591 -83.441 16.245 32.162 5.791 Agapostemon sericeus 25.25 70 CC 41.591 -83.441 16.245 32.162 5.962 Agapostemon sericeus 23.45 71 BU 41.548 -83.671 16.840 29.589 16.911 Bombus impatiens 5.35 72 WW 41.584 -83.590 19.448 27.120 5.041 Agapostemon sericeus 22.88 73 ZP 41.618 -83.578 53.060 43.301 13.093 Agapostemon sericeus 26.64 74 MNR 41.670 -83.584 40.104 45.924 6.73 Agapostemon sericeus 23.48 75 CHE 41.657 -83.549 73.297 58.649 24.328 Apis mellifera 0.55 104
76 COL 41.663 -83.483 20.238 44.064 6.51 Agapostemon sericeus 25.65 77 COL 41.663 -83.483 20.238 44.064 6.073 Agapostemon sericeus 20.98 78 COL 41.663 -83.483 20.238 44.064 5.521 Agapostemon sericeus 28.82 79 CC 41.591 -83.441 16.245 32.162 7.685 Agapostemon sericeus 32.59 80 COL 41.663 -83.483 20.238 44.064 5.247 Agapostemon sericeus 28.16 81 FTU 41.611 -83.619 47.690 39.347 10.618 Agapostemon sericeus 27.65 82 CC 41.591 -83.441 16.245 32.162 3.649 Agapostemon sericeus 30.50 83 COL 41.663 -83.483 20.238 44.064 10.287 Agapostemon sericeus 28.02 84 CC 41.591 -83.441 16.245 32.162 2.789 Agapostemon sericeus 37.53 85 CC 41.591 -83.441 16.245 32.162 6.249 Agapostemon sericeus 31.02 86 COL 41.663 -83.483 20.238 44.064 10.393 Agapostemon sericeus 27.17 87 TUC 41.637 -83.654 33.143 40.848 8.481 Agapostemon sericeus 28.25 88 BWN 41.699 -83.589 41.200 46.614 2.105 Agapostemon sericeus 29.13
105
APPENDIX D: COAUTHOR CONSENT FORM
I, Kevin McCluney, give Justin Burdine permission to submit Differential sensitivity of bees to urbanization-driven changes in body temperature and water content as a dissertation chapter.
Signature:
Date: 4/4/2019
106
APPENDIX E: COAUTHOR CONSENT FORM
I, Kevin McCluney, give Justin Burdine permission to submit Urbanization alters communities of flying arthropods in parks and gardens of a medium-sized city as a dissertation appendix.
Name:
Date: 4/4/2019
107
APPENDIX F: COAUTHOR CONSENT FORM
I, Edward Lagucki, give Justin Burdine permission to submit Urbanization alters communities of flying arthropods in parks and gardens of a medium-sized city as a dissertation appendix.
Signature:
Date: 4/2/2019
108
APPENDIX G: URBANIZATION ALTERS COMMUNITIES OF FLYING
ARTHROPODS IN PARKS AND GARDENS OF A MEDIUM-SIZED CITY1
Introduction
For the past two centuries, the global population has migrated from rural landscapes into densely-populated urban environments. Currently more than half of the world’s population resides in urban regions (United Nations 2014), and this number is growing. As more people move into urban regions, habitats are transformed into built environments and this impacts biodiversity and ecosystem processes (McKinney 2008). The process of urbanization fragments landscapes and creates a mosaic of habitat patches of different size, use, and quality.
Urbanization has been found to be a contributor to species endangerment (Czech et al. 2000), and often leads to the homogenization of biotic communities (McKinney 2006; Groffman et al.
2014). In addition, habitat loss and fragmentation in cities can alter important species interactions, such as plant-pollinator interactions (Harrison and Winfree 2015). These changes in community structure and species interactions may affect important abiotic and biotic processes, like pollination, nutrient cycling, and decomposition (McIntyre et al. 2001a), in the locations where most people now live. Thus, it is important to understand how urbanization influences organisms in order to maintain the services these organisms provide.
Urbanization can have strong positive and negative effects on a variety of organisms, making patterns of change unclear. Bird diversity, in general, is negatively affected (Blair 2004), but total bird abundance and that of introduced species can be positively affected (Clergeau et al.
1998). Arthropod pests can have higher abundances in urban habitats, possibly due to reduced predation and parasitism (Kahn 1998; Kahn & Cornell 1989; McIntyre 2000; Meineke et al.
1 This paper has been published as: Lagucki, E., J.D. Burdine, and K.E. McCluney. 2017. Urbanization alters communities of flying arthropods in parks and gardens of a medium-sized city. PeerJ 109
2017), or due to direct environmental effects (Meineke et al. 2013; Dale & Frank 2014), facilitating their proliferation and the likelihood of outbreaks. Others have argued that urbanization homogenizes biological communities because certain taxa are able to take advantage of urban environments worldwide (McKinney 2006). But much remains to be understood about how urbanization influences biota.
Flying arthropods are abundant and diverse, and perform numerous ecosystem functions within urban environments. Many studies have shown that arthropod diversity along urbanization gradients is lowest near urban centers (Centeno et al. 2004; Venn et al. 2003; Blair & Launer
1997). However, one study found that ant richness can be higher with urbanization (Uno et al.
2010). Differences in abundance and richness across urban environments can result in shifts in the composition of ant assemblages (Uno et al. 2010), and bee communities (McIntyre &
Hostetler 2001b; Pardee & Philpott 2014). Some influential drivers of Hymenoptera (ants, bees, wasps) population declines in urbanized areas include habitat fragmentation and pollution (Potts et al. 2010). Studies have shown that impervious surface cover has a negative effect on specialist cavity and ground nesting bees (Geslin et al. 2016; Threlfall et al. 2015), but a positive effect on generalist honey bees (Threlfall et al. 2015). Lepidoptera (butterflies, moths) have also been shown to have reduced species richness in heavily urbanized areas (McGeoch et al. 1997). Much of the reduction in Lepidoptera species richness is caused by a loss of vegetation or the replacement of native with introduced plants (Majer 1997). Furthermore, the plants many adult butterflies depend on for nectar can be more sensitive to heavy metal pollutants (Mulder et al.
2005), and this further explains the negative effects of urbanization on butterflies. Diptera abundance and community composition have also been found to vary along urbanization gradients (Avondet et al. 2003). Hemiptera abundance has been shown to increase with building 110 cover, and to decrease with proximity to natural habitat cover in an urban environment (Philpott et al. 2014). Thus, flying arthropods may respond strongly to urbanization, but additional work is needed to help us gain a better understanding of the mechanisms behind these patterns and the potential effects on ecosystem functions and services.
Two important habitat types within urban environments are urban gardens and city parks.
Urban gardens are an important source of local, healthy food (Taylor and Ard 2015), and are increasingly used in the remediation of vacant lots in post-industrial cities like Detroit and
Toledo (Our City in a Garden, 2010). Additionally, urban food production accounts for 15-20% of the global food supply (Hodgson et al. 2011). City parks provide many social and psychological benefits to urban residents, along with environmental services like air purification and noise reduction (Chiesura 2004). Furthermore, both urban gardens and city parks increase property values and can lead to tax revenues for cities (Luttik 2000: Bremer et al. 2003). Flying arthropods play important roles in urban gardens and city parks as pollinators, predators, and decomposers. Therefore, understanding how urbanization impacts flying arthropods is necessary to maintain the delivery of ecosystem services to urban gardens and city parks.
Here we examine how the abundance, diversity, and composition of flying (and floating) arthropod communities change with urbanization (percent impervious surface and distance to city center) in urban gardens and city parks. We predicted that flying arthropod abundance and diversity would be strongly correlated with percent impervious surface and distance to city center. In addition, we explored associations with other environmental variables and local habitat characteristics, in the hopes of identifying factors that might be influencing these communities across changes in urbanization, for future investigation.
Methods 111
Site Location. This study was conducted in Toledo, OH, USA. We sampled flying arthropods in a total of 30 parks and garden across the metro Toledo region (Figure 1). Sites were chosen by overlaying a grid (2km x 2km) across a Northwest Ohio map and assigning each grid cell a number value. A random number generator was used to select which grid cells we used in our study. Within each selected grid cell, a park or garden was chosen. Garden sites were managed by the Toledo Botanical Gardens outreach program and the MultiFaith Grows organization. Park sites were managed by the following entities: Toledo City Parks, Olander
Parks Systems, Toledo Zoo, Wood County Parks, and the City of Holland.
Sampling Methods. Flying (and floating) arthropods were sampled using elevated pan traps in July and August 2016 (Permit: Ohio Division of Wildlife 17-204). Elevated pan traps were constructed by placing a 175ml bowl atop a 1m pvc pipe (Tuell and Isaacs 2009). Bowls were painted white (#137990), blue (#51910), or yellow (#51806) using Krylon ColorMaster® spray paint. Each site contained three of each color type, for a total of nine elevated pan traps per site. Traps were left in the field for 24 hours. Each pan trap contained a water and soap mixture.
Sites were sampled once per month on days with weather conditions that were sunny with a temperature of at least 70°F. Upon collection, insects were rinsed with water and placed into vials containing ethanol to preserve specimens. Specimens were stored and identified to order.
Habitat Characteristics. Local habitat characteristics of each site were recorded during each sampling event (Table 1). We calculated the canopy cover at the center of each site in four cardinal directions using a densiometer. We counted the total number of trees within 25m of the site’s center. We walked a 10m transect starting at the site’s center and counted the number of flowers and floral colors for all vegetation within 1m on each side of the transect. Ground cover was measured by randomly placing four 1m quadrats along each transect, calculated as a 112 percentage in the following categories: bare ground, debris, herbaceous vegetation, leaf litter, or woody vegetation. Volumetric soil moisture was measured using a soil moisture meter (Delta-T
Devices SM150) at four random points along each transect. Unshaded air temperature and relative humidity were taken with a handheld weather station (Ambient Weather WS-HT-350).
Percent impervious surface was measured within a 300m radius circle around the center of each site using the NLCD 2011 Percent Developed Imperviousness dataset from the National Land
Cover Database. The distance of each site to the city center of Toledo (i.e. City Hall) was measured using Google Earth.
Multivariate Responses. We tested for associations between environmental factors and flying arthropod community composition with nonparametric permutational anova (adonis) using the “vegan” package of R. Also within this package, we used non-metric multidimensional scaling (metaMDS) to show differences in community composition between sites, and used the
“envfit” function to show associations with environmental factors. Bray-Curtis distances were used for all community composition techniques. For these multivariate analyses, we analyzed data combined from the two months, removing the need for repeated measures statistical approaches. We used the correlation function (cor) in R to test for collinearity between environmental variables, and environmental variables were considered highly correlated at a correlation coefficient of r = +/- 0.7. When this occurred, one of the two highly correlated variables was dropped from the analysis.
Univariate Responses. All statistical analyses utilized the program R (v. 3.1.3). The
“vegan” package in R was used to calculate the Shannon Diversity Index and Pielou’s Evenness.
We tested for associations of abundance (total flying arthropod and within order), diversity, or evenness of flying arthropods with our environmental factors and metrics of urbanization 113
(percent impervious surface and distance to city center) using linear regression analysis.
Abundance data were log-transformed, and evenness data were squared to better meet the normality and equal variance assumptions (assessed via plots of residuals). We consider alpha values below 0.1 to point towards potential patterns in need of further exploration and specify our exact p-values explicitly throughout. The purpose of this research is to identify patterns rather than test hypotheses. Future research will be needed to test hypotheses and infer mechanisms.
Results
Collection Summary Statistics. We sampled and identified 2369 individual arthropods representing nine orders (Araneae, Coleoptera, Diptera, Hemiptera, Hymenoptera, Lepidoptera,
Odonata, Orthoptera, and Thysanoptera). The three most common orders in terms of relative abundance were Diptera (~30% of all sampled insects), Hymenoptera (~29% of all sampled insects), and Coleoptera (~15% of all sampled insects). Diptera varied from 0 to 63 individuals per site, Coleoptera varied from 0 to 40, and Hymenoptera varied from 0 to 45.
At each site, we measured a wide range of values for our environmental variables of canopy cover (0-82.3%), number of trees (0-10 individuals), soil moisture (6.0-62.3%), impervious surface (5.6%-73.3%), humidity (28.4-75%), temperature (70.3-100.4°F), herbaceous cover (45-97.5%), bare ground (0-37.5%), distance to city center (638-12,884m), and flower abundance (6-202). We tested for collinearity between environmental variables, and removed bare ground from further analyses due to its high collinearity with herbaceous cover (r = -0.78).
Community Composition Results. Our PERMANOVA (Table 2) test showed two environmental variables that were associated with flying arthropod community composition: 114
impervious surface (F1,20 = 4.39, p = 0.004 at α = 0.05; Figure 2) and canopy cover (F1,20 = 2.31, p = 0.057 at α = 0.1; Figure 2).
Distance to City Center Results: The total abundance of flying arthropods was positively associated with distance to city center (Figure 3). For order-specific responses, we found positive associations with distance to city center for abundances of Lepidoptera (F1,27 = 10.523, p =
0.003), Hymenoptera (F1,26 = 4.686, p = 0.0398), and Araneae (F1,26 = 3.742, p = 0.064 at α =
0.1). Distance to city center was not associated with the diversity or evenness of flying arthropod communities.
Percent Impervious Surface Results: The total abundance of flying arthropods was negatively associated with impervious surface (Figure 3). For order-specific responses, we found negative associations with percent impervious surface for abundances of Araneae (F1,26 = 4.682, p = 0.040), Diptera ((F1,28 = 6.739, p = 0.0149), and Hemiptera (F1,25 = 3.228, p = 0.084 at α =
0.1). Percent impervious surface was not associated with the diversity or evenness of flying arthropod communities.
Vegetation Results. The total abundance of all flying arthropods combined was negatively associated with canopy cover and herbaceous cover, and positively associated with the number of flowering plants (Table 3). For order-specific response, we found negative associations with canopy cover for the abundances of Hemiptera (F1,25 =4.385, p = 0.047) and Hymenoptera (F1,26
= 3.865, p = 0.0601 at α = 0.1). Additionally, Lepidoptera abundance was negatively associated with the number of trees (F1,27 = 4.472, p = 0.0438), and Hemiptera abundance was negatively associated with herbaceous cover (F1,25 = 9.664, p = 0.005). Vegetation factors were not associated with the diversity or evenness of flying arthropod communities. 115
Soil Moisture Results. The total abundance of all flying arthropods was positively associated with soil moisture. Arthropod diversity was also positively associated with soil moisture. For order-specific responses, we found positive associations with soil moisture and the abundances of Hemiptera (F1,25 = 4.762, p = 0.039), Hymenoptera (F1,26 = 3.001, p = 0.0951 at α
= 0.1), and Araneae (F1,26 = 3.312, p = 0.080 at α = 0.1).
Discussion
Understanding how flying arthropod communities are impacted within community gardens and city parks in urban areas is important for maintaining the many ecosystem functions flying arthropods provide. We found evidence that pollinator-containing orders of insects (i.e.
Hymenoptera, Lepidoptera) are less abundant with more impervious surface and more abundant farther from the city center (i.e. Diptera). These patterns are supported across the literature for butterflies (Clark et al. 2007; Mauro et al. 2007), bees (Hernandez et al. 2009), and parasitoids
(Bennett & Gratton 2012). In addition, we found evidence that orders containing both pests and predators (i.e. Araneae and Hemiptera) are less abundant with more impervious surface. These results are interesting because many of these taxa are important in providing pollination and pest control services for urban gardens and city parks. More research targeting the mechanisms of effect upon these taxa are needed.
Associations with Distance. We found more flying arthropods in general, and more
Araneae, Hymenoptera, and Lepidoptera farther from the city center. Hulsmann et al. (2015) found a similar pattern with distance to city center for bumblebee abundance and diversity, while
Pacheco and Vasconcelos (2007) found no effect on ant abundance in an urban region. Others have found that butterfly abundance peaks at intermediate distances (Blair et al. 1997). Peaks in abundance at intermediate distances may be explained by additional food and water resources 116 made available in suburban regions, while peaks at distances further from the urban core are often explained by plant community composition and density (Hulsmann et al. 2015).
Associations with Impervious Surface. We found that impervious surface was associated with shifts in flying arthropod community composition, with fewer flying arthropods overall with higher impervious surface. In addition, Hemiptera, Araneae, and Diptera showed lower abundances with more impervious surface. Studies have shown similar patterns for bumble bees
(Ahrne et al. 2009), ground spiders (Kaltsas et al. 2014), and tree spiders (Meineke et al. 2017).
However, scale insects (Hemiptera) are positively affected by impervious surface (Dale et al.
2016; Speight et al. 1998). Other studies have found percent impervious surface to have no effect on the abundance of arthropods (Pacheco and Vasconcelos 2007). One mechanism to explain why impervious surface reduced arthropod abundance is a species-area effect, since impervious surfaces can lead to a loss in habitat area (McKinney 2008). Another mechanism is a physiological effect of impervious surface on arthropods. Diamond et al. (2017) found difference in physiological limits for ants sampled at sites with high and low impervious surface. However, many other possibilities exist (e.g. increased soil contaminates, reduced nesting sites).
It is interesting to note that except for spiders, the orders influenced by distance to city center were different than those influenced by percent impervious surface. This suggests that
Hymenoptera and Lepidoptera may be more influenced by habitat fragmentation and a loss of connectivity, while Hemiptera and Diptera may be more influenced by local habitat characteristics associated with impervious surface (e.g. increased temperatures). This hypothesis warrants further testing.
Association with Vegetation Factors. We found negative associations with canopy and herbaceous cover on Hymenoptera and Hemiptera abundance, as well as overall flying arthropod 117 abundance. Additionally, canopy cover was associated with the composition of flying arthropod communities. Previous studies in urban systems support our findings on canopy cover, but not herbaceous cover. Studies show that canopy cover reduces herbivorous ground arthropod abundance (Philpott et al. 2014), and has a significant impact on ant community composition
(Uno et al. 2010). But these studies found herbaceous cover to have positive or no effects on arthropods, and others have found similar positive effects of herbaceous cover on arthropods
(Pinna et al. 2008). The differences between our findings (negative associations with herbaceous cover) and those of others (positive or no associations) may be due to the herbaceous cover structure or composition (i.e. height, diversity, or type). Studies have found that vegetation height is an important predictor of community composition for leafhoppers and grasshoppers
(Strauss and Bierdermann 2006). Our findings that the total arthropod abundance was negatively associated with herbaceous cover, but positively associated with flowing plants, could be explained by aspects of herbaceous cover for which we did not account. One might expect flowering plants to be associated with herbaceous vegetation in undeveloped areas, but we suggest that this relationship may not hold in managed urban landscapes, where turf grass is part of the herbaceous cover. Additionally, a previous study showed that Hymenoptera were more attracted to specific plant species, and not necessarily diverse gardens (Barbir et al. 2015).
Combined this suggests that the relative abundance of herbaceous vegetation should not necessarily be expected to be positively associated with arthropod abundance in urban areas.
Associations with Soil Moisture. Soil moisture also had strong associations with flying arthropod abundance. Soil moisture was the only factor to have positive associations on arthropod diversity, and it was positively associated with the abundance of Araneae, Hemiptera, and Hymenoptera. Studies have found positive effects of soil moisture on arthropod movement 118
(Green et al. 2005), arthropod water content (McCluney et al. 2017), and arthropod abundance
(Allen et al. 2014), but research is lacking on the role of soil moisture in altering community composition and diversity. However, a study found that the absolute number of insect species increased with increasing soil moisture levels and suggests that soil moisture plays a key role in overall ecosystem health (Janzen et al. 1968). Our finding that soil moisture is associated with flying arthropod abundance and diversity is interesting because urban gardens (and many city parks) are irrigated and receive water inputs. Studies have shown that irrigation can positively impact arthropod abundance (Cook and Faeth 2006), and these inputs could be important in maintaining abundant and diverse flying arthropod communities in urbanized sites.
Conclusions. Understanding drivers of flying arthropod declines is necessary in maintaining the important services they provide in urban gardens and city parks. Upwards of 150 agricultural crops in the US require pollination services, and flying arthropods are the primary pollinator of these crops. Additionally, pest control services are important in reducing crop loss.
With estimates that 15-20% of the world’s food supply comes from urban agriculture (Maxwell et al. 2000), conservation of flying arthropods with urban environments should be an issue of global concern. The patterns we observed indicate that urbanization plays an important role in shaping arthropod communities, and particularly may reduce the abundance of Lepidoptera and
Hymenoptera.