Characterizing insect communities within thin-soil environments

Natural thin-soil environments are those which have little to no soil accumulation atop hard substrates. Many of these natural thin-soil environments, such as alvars, rocky lakeshores or glades, cliffs and cliff bluffs, and barrens, are found in the Great Lakes Region of North America. Due to their ubiquity and ecosystem services they provide, characterizing insects in sensitive environments such as these is important. This study monitored insects in nine thin-soil sites, within three regions, on a 630 km latitudinal gradient in the Southeastern Great Lakes Region of North America from June - August 2019. Over 22,000 insect specimens collected were identified to order or family, and bee specimens were identified to genus or species. We found that overall insect community composition and biodiversity characteristics were similar between the three regions examined. However, the central region had higher taxonomic richness than the southern region. Although unique bee taxa were observed in each region, diversity metrics and community composition of bees were similar among sites. This study provides taxonomic information about the insect, particularly bees, and plant communities in thin-soil environments in this region, which could support conservation and management efforts.


23
Ecological communities are shaped by the physical attributes of their environments. In 24 the Great Lakes basin, numerous globally rare habitats occur, including those characterized by 25 their paucity of surface soil. The ecological class 'Primary' is characterized by having little to no 26 soil accumulation on top of bedrock, cobble, or exposed mineral soil. There are many natural 27 thin-soil community types, such as alvars, rocky (cobble or bedrock) lakeshores or glades, cliffs 28 and cliff bluffs, and barrens (Cohen et al. 2015). These habitats may experience intense wind and 29 solar radiation, as well as varying precipitation conditions, from heavy rain to drought, due to the 30 bedrock being at or near the surface (Stephenson andHerendeen 1986, Brunton 1988, Albert 31 2006). Together, these factors limit the primary producers that can survive there (Lundholm  (Comer et al. 1997, Reschke et al. 1999, Albert 2006, Neufeld et al. 2018 2019). 37 In the Great Lakes Region of North America, research in these rare thin-soil 38 environments, mostly alvars, has generally focused on identifying the taxa that occur in them or Surveys of insects in Great Lakes thin-soil environments have largely focused on smaller 68 geographical regions, such as the intense sampling of alvars of Ontario, especially the Saugeen 69 (Bruce) Peninsula, and in Michigan near the northern shorelines of Lakes Michigan and Huron 70 (Albert et al. 1994, 1995, Reschke et al. 1999, Bouchard et al. 2005, Albert 2006 2015). Few surveys have occurred over the north-south expanses of this ecoregion, and none 72 have explicitly examined the communities of insects across these thin-soil habitats spanning the 73 region. Additionally, to our knowledge bee communities in these thin-soil environments have not 74 been characterized before. To address these knowledge gaps, we conducted a study of the 75 community composition of insects in thin-soil environments in three regions on a latitudinal 76 gradient across the Southeastern Great Lakes Region of North America. In this study, we 77 examined biodiversity metrics of insect communities, including bees as our focal taxa, between 78 northern, central, and southern regions of this area. We predicted that even if habitats and plant 79 communities differed, insect communities would be similar across these three regions because 80 the community will be shaped by the abiotic traits of the landscape. (about 15 cm or less) and typically lacked direct canopy cover, leaving them relatively open and exposed to solar radiation, winds, and ranging precipitation conditions. Permits to collect insects 89 and plants were obtained for all sites in accordance with landholder policy. shrubby cinquefoil dwarf shrub alvar, with a slightly rocky landscape. The Cape Hurd Alvar was 96 similar to Beaton in vegetation and landscape, but closer to the lakeshore making it subject to 97 frequent low floods. All sites occurred in a rural landscape, primarily consisting of undeveloped 98 land, and protected areas, but the area had previously been subjected to heavy logging pressure.

99
Central sites were located in Cuyahoga County in Northeast Ohio, USA, specifically in 100 the municipalities of Walton Hills, Parma, and Highland Heights. All sites were embedded in a 101 greater landscape of mixed use, urban, industrial, and semi-urban land use histories. In Walton 102 Hills, we sampled a thin-soil mossy barren near a hiking trail, between a meadow and forest, 103 near a cliff edge over a creek in the Cleveland Metroparks Bedford Reservation. Cleveland surveyed using three types of passive sampling traps: yellow sticky cards (Pherocon,Zoecon,126 Palo Alto, CA, USA), bee bowls (also known as pan traps, inspired by Leong & Thorp, 1999),  (Table S1). Each  (Table S2).

143
In the lab, the liquid samples were strained, and specimens were identified and placed in 144 vials with 70% ethanol for storage. The sticky cards were frozen, and specimens were identified 145 while remaining in the bag, afterwards they remained stored in the freezer. Specimens were 146 identified to order, suborder, superfamily, group ("wingless parasitoid wasps"), or family (Table   147 1). We identified bees (Hymenoptera: Apoidea: Anthophila) to the highest precision possible 148 (genus or species) using expert identification and various interactive Discover Life keys (det.

182
Three sampling periods at our nine sites yielded 252 samples: 72 from the north, 107 183 from the central, and 73 from the south. From these samples, we identified 22,459 specimens: 184 9,304 from the north sites, 11,107 from central, and 2,048 from the south. In the central and 185 southern regions, Diptera was the most abundant order. In the northern region, Hymenoptera was 186 the most abundant order, due to a single emergence event where 6,032 chalcid wasps were 187 collected, followed by Diptera (Table 1).

188
Traps at the southern sites produced the fewest unique taxa, followed by northern sites, southern. When compared with first order jackknife richness estimates, capture efficiency in the 206 northern region was 74%, 71% in the central region, and 63% in the southern region ( Figure 6).

207
Region was not found to contribute significantly to bee community variation as there was no 208 difference in bee richness or Shannon diversity between regions ( Figure S2).

209
The single most abundant species in the north was Augochlorella aurata (Smith), but the  (Table S2). Red maple was 225 found at all three sites in the central region and two sites in the southern region. America. Our findings suggest that there is strong similarity among the insect communities 230 within these habitats, at least when using typical sampling and community classification metrics. 231 We saw no differences in Shannon diversity between any regions when compared over a three were also observed when data were examined at the site level.

239
Bee analyses followed similar trends to the overall insect community analysis, as we  (14), compared to the central (87) and southern (49) regions, and between all the regions 244 specimen numbers were relatively low, which should be considered when interpreting these 245 results. However, there is strength in the higher resolution of identification of the bee specimens.

246
With these data, we can identify some apparent associations of bee taxa with regions and sites. is a primitively eusocial, generalist, ground-nesting species (Michener 1966). This species was 265 captured at all three southern sites, on sandstone cliff bluffs. However, it was not unique to this 266 region as specimens were captured in the central region as well, mostly at the meadow site.

267
During the three months of our study, we captured and identified 151 bee specimens 268 belonging to 26 taxa overall. The northern region had the lowest bee richness and raw abundance 269 and the central region had the highest bee richness and raw abundance. Though the bee bowls 270 were the only trap targeting bees, we used the specimens collected in the ramp traps, targeting 271 ground-dwelling insects, as well. A bee monitoring study that took place on non-thin soil  Overall, we captured few large-bodied bees, and notably no bumble bees, during the 285 study, but that is most likely a result of the bias of bee bowls to collect smaller bodied genera  (Albert et al. 1995, 373 Comer et al. 1997, Reschke et al. 1999, Neufeld et al. 2018, McMullin 2019. While previous 374 work has examined the composition of these habitats (Albert et al. 1994, 1995, Comer et al. 375 1997, Bouchard et al. 1998, 2005, Reschke et al. 1999, Albert 2006, Cohen et al. 2015, Neufeld 376 et al. 2018, McMullin 2019, in this study, we examined how these environments shape 377 similarities in insect communities. This study leverages a unique design that encompasses a large