Conservation in post-industrial cities: how does vacant land management and landscape configuration influence urban bees?
Turo, Katherine et al. (2020), Conservation in post-industrial cities: how does vacant land management and landscape configuration influence urban bees?, Dryad, Dataset, https://doi.org/10.5061/dryad.8cz8w9gn5
1. Rich pollinator assemblages are documented in some cities despite habitat fragmentation and degradation, suggesting that urban areas have potential as pollinator refuges. To inform urban bee conservation, we assessed local and landscape scale drivers of bee community composition and foraging within vacant lots of Cleveland, Ohio, USA. Cleveland is a shrinking city, a type of urban area that has an over-abundance of vacated greenspaces as a result of population loss and subsequent demolition of abandoned infrastructure. As such, Cleveland represents over 350 post-industrial cities worldwide, that are all promising locations for bee conservation.
2. Across a network of 56 residential vacant lots (each ~30m x 12m), we established seven unique habitats, including seeded native prairies, to investigate how vegetation management and landscape context at a 1500m radius influenced urban bee communities. We assessed the distribution of several bee functional traits, diversity, and abundance with pan and malaise traps. Foraging frequency was determined with plant-pollinator interaction networks derived from vacuum collections of bees at flowers.
3. We observed higher bee richness and increased abundance of smaller sized bees as the size of surrounding greenspace patches increased within a 1500m radius landscape buffer. Within habitats, greater plant biomass positively influenced bee community richness and abundance whereas taller vegetation was a negative influence. Plant-pollinator interaction networks were dominated by spontaneous non-native vegetation, illustrating that this forage supports urban bees.
4. Synthesis and applications: Our study indicates that proximity to larger greenspaces within an urban landscape promotes overall bee richness and increased occurrence of smaller bee species within residential vacant lots. While we did not observe our seeded native plants enhancing the bee community, native wildflowers were still establishing during the study and may have a greater influence when blooming at higher densities. Importantly, spontaneous non-native vegetation provided the majority of urban bee’s forage. Thus, vacant land that is minimally managed and vegetated with what many consider undesirable “weeds”, provides valuable habitat for bee conservation in cities.
From June 2015 – August 2016, we studied 56 residential vacant lots across eight neighborhoods in Cleveland, Ohio, USA. Within each neighborhood, seven vacant lots were randomly assigned to habitat treatments. Control and Meadow treatments contained pre-existing vegetation; all other treatments were seeded. Treatments were as follows: 1) Control: non-native turf grasses and spontaneous vegetation, mown monthly, 2) Meadow: non-native turf grasses and spontaneous vegetation, mown annually, 3) Fine-fescue Lawn: low-growing, non-native fine-fescue grasses, 4) Flowering Lawn: low growing, non-native fine-fescue grasses with four non-native forbs, 5) Grass Prairie: three tall, native grass species, 6) Low Diversity Prairie: a tall, flowering habitat with three native grasses and six forbs, and 7) High Diversity Prairie: a tall, flowering habitat with three native grasses and 22 species of forbs. During 2015, all treatments, except Meadow, were mown monthly to facilitate establishment of seeded species. In 2016, Control and Lawn habitats were mown monthly while Meadow and Prairie sites were mown annually in October.
Vegetative Community Assessment
Plant assessments occurred in early and late season 2015 (early: 16 June–17 July, late: 22 July–13 August), and early, middle, and late season 2016 (early: 13–24 June, middle: 11–22 July, late: 4–16 August). In the center of each vacant lot, we created a 7x15m grid and randomly sampled this grid to monitor vegetation height, bloom area, dominant species diversity, and plant biomass.
To assess plant height, bloom abundance, and bloom area, we selected six 1 m2 grid quadrats and placed a 0.5 m2 PVC pipe square in each quadrat’s center, wherein data was collected. We measured plant height (cm) at the center and at two opposite corners of the quadrat. Then, we calculated a mean plant height for each site by averaging all height measurements. Bloom abundance and area were estimated from the same six quadrats. For each flowering species, we counted all floral units to determine bloom abundance (Table S5). We then took five random measurements (mm2) of individual floral units per species. Average bloom size per species was calculated and then multiplied by bloom abundance to quantify total bloom area at a site.
Plant biomass and diversity were estimated from 20 separate quadrats in our plot grid. We recorded the three most dominant plant species. Grasses were pooled into one general category. Dominant species diversity was calculated from a total of 60 plant records per site with a Shannon-Wiener Index. To estimate vegetation biomass, we used a comparative yield method (Haydock & Shaw, 1975). We identified five quadrats within the grid that represented a subjective scale of least (1.0) to most (5.0) plant biomass. We then compared the biomass from our 20 random quadrats to this scale (1.0 – 5.0) and estimated, in quarter increments, the relative density of plants within. We harvested vegetation from a representative 0.5 m2 portion of each standard, dried it, and measured dry plant weight. Linear regression equations were created for each site from these five dry weights, and plant biomass per site was averaged from the 20 dry weight estimates.
Landscape data were obtained through the Cleveland City Planning Commission and described landscape cover at 1m2 resolution from 2011 aerial imagery for a 1500m radius surrounding each site (Galvin & O’Neil-Dunne, 2013). We classified landscapes in a binary land cover system as either “greenspace” (shrubs and grass, tree canopy over shrubs and grass) or “other” (impervious surface, buildings, tree canopy over impervious surface and buildings). We used FRAGSTATs software (McGarigal, Cushman, & Ene, 2012) to measure four class level indices: (1) total area (m2) of greenspace within a landscape (total.gs), (2) mean patch size (m2) of greenspace (gs.size), (3) percentage of a landscape composed by the largest patch of greenspace (LPI.gs), and (4) mean isolation (m) between greenspaces (ENN).
Urban Bee Communities and Foraging Preferences
We conducted a bee community assessment with pan traps and SLAM Malaise traps (Bugdorm© MegaView Science Co., Taiwan) in a subset of treatments (Control, Fine-fescue Lawn, Flowering Lawn, Grass Prairie, Low Diversity Prairie, n=40) in 2015 and all treatments (n=55) in 2016. Sampling took place once per sampling period from 10am to 2pm on non-rainy days in early, middle, and late season in 2015 (early: 16–26 June, middle: 16–27 July, late: 11–21 August) and 2016 (early: 9–21 June, middle: 5–20 July, late: 2–12 August). Bees were sampled with one malaise trap and seven bright yellow plastic bowls (12 oz) (Solo© Dart Container Corporation, Mason, MI) filled 2/3 with a 1% dish soap solution (Blue Dawn© Proctor and Gamble, Cincinnati, OH). Pan traps were placed on the ground in randomly selected quadrats while malaise traps were deployed at the lot’s center.
Foraging bees were actively sampled once per month on non-rainy days within all flowering treatments (Control, Meadow, Flowering Lawn, Low Diversity Prairie, High Diversity Prairie) in early and late summer in 2015 (early: 8–15 July, late: 12–19 August, 13–25 September) and in 2016 (early: 9–21 June, 5–8 July, late: 2–8 August). Bees were collected with hand vacuums (© Bioquip, Rancho Dominguez, CA) at each lot for 4.5 minutes per floral species between the times of 10am and 4pm.
All collected bees were identified to species where possible. We categorized bees by functional traits, including nesting guild, status in North America (native or exotic), known foraging specialization (generalist or specialist), and the community-weighted mean of bee body size (CWM) (Ascher & Pickering, 2018; Sivakoff et al., 2018; Sam Droege pers. comm. 2 April 2019). CWM was calculated following Garnier et al. (2004). Nesting guilds included: human cultivated (Apis mellifera L.), ground nesting, cavity nesting (non-pith), pith nesting, and parasitic bees with no-nests.
Ascher, J. S., & Pickering, J. (2018). Discover Life bee species guide and world checklist (Hymenoptera: Apoidea: Anthophila). Retrieved from http://www.discoverlife.org/mp/20q?guide=Apoidea_species
Galvin, M., & O’Neil-Dunne, J. (2013). Cuyahoga Country urban tree canopy and land cover mapping. Retrieved from https://www.countyplanning.us/projects/urban-tree-canopy-assessment/background/
Garnier, E., Cortez, J., Billès, G., Navas, M. L., Roumet, C., Debussche, M., … Toussaint, J. P. (2004). Plant functional markers capture ecosystem properties during secondary succession. Ecology, 85(9), 2630–2637. doi: 10.1890/03-0799
Haydock, K. P., & Shaw, N. H. (1975). The comparitive yield method for estimating dry matter yield of pasture. Australian Journal of Experimental Agriculture and Animal Husbandry, 15, 663–670.
McGarigal, K., Cushman, S. A., & Ene, E. (2012). FRAGSTATS: Spatial pattern analysis program for categorical and continuous maps. Amherst, Massachusetts: University of Massachusetts.
Sivakoff, F. S., Prajzner, S. P., & Gardiner, M. M. (2018). Unique bee community assembly within vacant lots and urban farms results from variation in surrounding landscape urbanization intensity. Sustainability, 10(6), 1926.
Data is organized to support five analyses associated with the publication. A README file is provided to explain all data sets including which files are associated with what analysis and descriptions for column headers. NAs are resulting from errors in site management wherein the sites were erroneously mowed, disrupting sampling.
National Science Foundation, Award: CAREER 1253197
National Science Foundation, Award: DGE-1343012
North Central SARE, Award: GNC16-233