where the cumulative sum of Lijk was calculated across all time periods since bees first were observed visiting flowers (including the present time period) and was run both including and excluding the Tijk term. Models for H3b were only run for bee genera that might produce multiple broods per season (Augochlorella , Bombus , Halictus ,Lasioglossum ; Packer et al. 2007).
Model selection with the MuMIn package (Bartoń 2019) was used to determine which models best predicted bee visits, considering all models with ΔAICc < 2. Model selection was run in two iterations, with unknown land area assigned either the median FR value calculated from all land types within a specific radius during a given time period (‘median models’), or the minimum FR value calculated from all land types within a specific radius during a given time period, which was always zero (‘minimum models’). This was done to test the sensitivity of our conclusions to the presence of unknown areas; we found no differences between the identities of models with ΔAICc < 2 when unknown areas were assigned either minimum or median FR values, and we have presented the results from both iterations.
Results
Floral resources
Floral volume was the best predictor of both daily nectar sugar mass (µg/day; R2 = 0.39, p = 3.6×10-6, n = 46, Fig. 2a), and pollen volume (µl/flower; R2 = 0.40, p = 6.0×10-5, n = 33, Fig. 2b), and therefore was used to represent FR in all subsequent analyses. Flower length, width, height, and surface area were also significantly correlated with nectar and pollen volume, but to a lesser degree (R2 ≤ 0.34). Floral measurements, pollen volumes, and nectar sugar mass for individual species can be found in Appendix, Table A2–A4.
Across many of the landscapes sampled in this study, there was a high degree of correlation between the FR at each spatial scale within a given time period, particularly between the 500 m and 750 m spatial scales (Fig. 3). In most landscapes, the fourth (last) time period had the lowest FR, and the highest FR abundance generally was seen in either in the second or third time period, with a few landscapes having the highest FR of the season in the first time period (Fig. 3).
Most common bee genera
Over the season, 8422 bee visits were observed across all sites, with 1647 visits observed in the first time period, 1946 in the second, 2211 in the third, and 2618 in the fourth. Bees in the genera Andrena(mining bees), Apis (A. mellifera ; Western honey bee),Bombus (bumble bees), Halictus (furrow bees), andLasioglossum (sweat bees) were observed visiting flowers in at least four of the 27 sites surveyed during all time periods. Bees in the genus Augochlorella (A. aurata ; golden green sweat bee) were observed during just the first and second time periods in at least four sites, while Megachile (leafcutter bees) were observed in the second and third, and Peponapis (P. pruinosa ; hoary squash bee) in the third and fourth. All other bee genera found in at least four sites were only observed during one time period.
H1: Bee visits are only influenced by local FR
No models for any genera supported the hypothesis that bee visits were influenced only by local FR (Table 1). Several taxa did, however, show a positive relationship with the abundance of FR within transects based on the best models which also included landscape FR abundance as a predictor (Table 1; Fig, 4 for Halictus and Peponapis ).
H2: Bee visits are influenced by the present abundance of landscape FR
The number of visits observed from Andrena , Augochlorella ,Megachile , and Peponapis supported hypothesis H2b, that bee visits were negatively associated with the present abundance of landscape FR (Fig. 4; Table 1). The number of bee visits was best predicted by models of present landscape FR at a 250 m scale forMegachile (ΔAICc = 0), and at a 750 m scale forAugochlorella (ΔAICc = 0) and Peponapis (ΔAICc = 0).Andrena visits were best predicted by models of the present landscape FR at a 250 m scale (ΔAICc ≤ 1.00) but showed some support for hypothesis H3a (Table 1), described in the following section.
H3: Bee visits are influenced by the previous abundance of landscape FR
For Halictus and Lasioglossum , hypothesis H3a—that bee visits were influenced by the previous abundance of landscape FR within a season—was best supported (Table 1; Fig. 4). Halictus andLasioglossum visits were best predicted from models including landscape FR abundance during the first time period, and the non-positive change in the abundance of landscape FR since the first time period within 750 m of sampling locations (both ΔAICc = 0). For both genera, bee visits were negatively associated with landscape FR during the first time period and were also lower in landscapes that experienced greater decreases in landscape FR over the season (Table 1; Fig. 4). Andrena visits also showed some support for hypothesis H3a, with bee visits predicted by the previous abundance of landscape FR at the 250 m scale (ΔAICc ≤ 1.81), but in this genus, greater decreases in FR over the season were associated with a higher number of bee visits. However, hypothesis H2b had slightly better support inAndrena (Table 1), as described in the previous section.
Bombus visitation rates best supported hypothesis H3b, that bee visits are influenced by the cumulative abundance of landscape FR since the first time period (Table 1; Fig 4). Landscape FR was a predictor ofBombus visits at all spatial scales over which it was measured, but the 250 m scale was the strongest predictor (ΔAICc = 0). Visits fromBombus were positively associated with the cumulative abundance of landscape FR at all spatial scales.
Discussion
We found that FR in the landscape over preceding months was an important predictor of bee activity for most bees that were foraging over the entire season. High FR abundance in previous months was positively related to the number of visits for bees that were more social and produced multiple broods per season (Bombus ) , while stable FR abundance over previous months positively influenced the number of visits for genera that were less social and produced fewer broods per season (Halictus and Lasioglossum ). For all other bees, including solitary bees (Andrena , Megachile , and Peponapis ) and bees with shorter observed foraging periods (Augochlorella , Megachile , and Peponapis ), the present abundance of FR in the landscape was the best predictor of bee visits. However, the present abundance of FR was negatively related to local bee visits, possibly because of a dilution of pollinators across high-FR landscapes. Individual bee taxa exhibit unique combinations of foraging distances, foraging periods, and numbers of brood produced per season, all of which can influence how bees respond to changes in landscape FR abundance over a season. Our research acknowledged these taxonomic differences by assuming each bee genus observed would respond at different spatial and temporal scales to the abundance of FR in a landscape. By doing this, we were able to document important differences in the ways in which each genus was influenced by the spatial and temporal availability of FR.
Much of the previous research on wild bees responding to FR in agricultural landscapes has found that a higher abundance of FR is associated with larger populations, higher densities, or greater numbers of visits (Mandelik et al. 2012, Kovács-Hostyánszki et al. 2013, Mallinger et al. 2016, Martins et al. 2018). In our study, the number of visits by Bombus was positively related to cumulative landscape FR, in line with the hypothesis that bees producing multiple broods in a season can increase population sizes within one season with access to more FR through time. This finding agrees with previous research examining B. vosnesenskii colony responses to FR in agricultural landscapes, which found that the production of males and workers was more positively related to high early-season FR abundance in the surrounding landscape than to late-season FR (Williams et al. 2012). Bumble bees represent some of the most common and important pollinators for both wildflowers and crops in the Northern Hemisphere, but many species are facing declines (Goulson et al. 2008). Our results suggest that early-season FR in agricultural landscapes could promote high bumble bee visitation rates later in the season, and potentially maintain or increase colony sizes over a season.
For Halictus and Lasioglossum , our findings best supported the hypothesis that these bees were producing a single brood per season, and therefore were able to maintain (but not increase) their population sizes when FR was consistent or increased over time. However, in the region this study was conducted, both Halictus andLasioglossum include eusocial species that produce multiple broods per season and solitary species that only produce a single brood (Mitchell 1960, Packer et al. 2007). Given the stronger support for the hypothesis that these genera responded as expected for single-brood-producing bees, our results indicate either that more solitary species were present in our study locations, or that the more social species in this region were producing too few brood per season to exhibit a strong response to the cumulative abundance of FR. Geographic variations can influence both the degree of sociality and the number of broods produced within single species of Halictus andLasioglossum , with a general pattern of more solitary bees and fewer broods being produced at higher elevations and latitudes (Richards and Packer 1995, Wcislo 1997, Davison and Field 2016). In our study locations, most Halictus and Lasioglossum species are closer to their northern range limit (Mitchell 1960), which should increase the prevalence of populations in this region that produce fewer broods per season and exhibit less social behaviours.
The remaining bee genera (Andrena , Augochlorella ,Megachile , and Peponapis ) were all most influenced by the present abundance of FR in the landscape. These genera include both eusocial (Augochlorella ) and solitary bees (Andrena ,Megachile , and Peponapis ). Although Augochlorellavisits were primarily observed during the first and second sampling periods, the only species of Augochlorella in eastern Canada,A . aurata (Packer et al. 2007), is in flight for the entire growing season (Mitchell 1960). Similar to Halictus andLasioglossum , some A. aurata populations are solitary, while others produce just two broods (one worker and one reproductive brood) per season (Packer et al. 1989). If only two broods were produced over the entire season by A. aurata populations, we might have been unable to observe any influence of cumulative FR abundance by the second time period (only the first and second time periods were analyzed for this genus). The fact that visits from the solitary bee generaAndrena , Megachile , and Peponapis were apparently unaffected by past FR makes sense in light of the short flight periods of individual species within these genera. Although Andrena were found foraging over the entire season, this genus comprises 75 species in eastern Canada (Packer et al. 2007), many of which are active as adults for just one or two months (LaBerge 1986, Larkin et al. 2008).Megachile were observed foraging during the second, third, and fourth time periods, but a number of the species in this region have more restricted foraging periods (Mitchell 1962, Sheffield et al. 2011). The single local species of Peponapis , P. pruinosa , is a specialist on pollen in the Cucurbitaceae family and has a foraging period synchronized with its flowering in southern Ontario (Willis and Kevan 1995). Given the prevalence of species with short flight periods in our study area, the foraging periods of many individual species were likely too short to respond to fluctuations in FR at the roughly monthly scale we considered. Future research should focus on examining how fluctuations in FR over shorter temporal scales (e.g., weekly) influence bee activity, which would allow for development of agricultural landscapes that specifically benefit those species with short flight periods.
Despite the variety of life-history traits represented byAndrena , Augochlorella , Megachile , andPeponapis , all four genera responded similarly to the present abundance of FR; visits were generally fewer in landscapes that had a high abundance of FR. Several studies have previously found that abundant FR can decrease bee density on crops, either through dilution of pollinators across a landscape (Kovács-Hostyánszki et al. 2013, Holzschuh et al. 2016), or through distraction of pollinators from crops to other resource-rich areas (Lander et al. 2011, Nicholson et al. 2019). This may indicate that the landscapes used in our study generally provided a high amount of FR outside the local sampling areas, perhaps because we selected sampling locations based on their proximity to farms growing fruit or vegetables. Population sizes for solitary bees that produce a single brood per season (Andrena , Megachile , andPeponapis ) should also be strongly limited by the amount of FR available in the previous year, which would be used to produce the generation foraging in the current year. Since we did not know the relative abundance of FR in the previous year, this may have been the primary factor limiting bee populations, especially for species that only forage for a few weeks in a season.
Differences in body size can contribute to differences in the maximum foraging ranges of bee taxa (Greenleaf et al. 2007) and thus to availability of FR to bees within the landscape surrounding sampling locations. For most genera, we found that the number of bee visits observed within a transect was best predicted by landscape FR at a consistent spatial scale. Within genera like Andrena ,Halictus , and Lasioglossum , individual species can vary greatly in body size (Mitchell 1960), but the average body size across a genus did not seem to correlate with the spatial scale at which landscape FR was most relevant. Visits for Andrena , a relatively small-bodied bee (body length ranges from 4–15 mm; Mitchell 1960) were best predicted by FR at 250 m, the smallest spatial scale we measured, while visits from other small-bodied bees like Augochlorella(average body length of A. aurata is 5.25 mm; Mitchell 1960),Halictus (body length 7–13 mm; Mitchell 1960) andLasioglossum (body length 3.5–10 mm; Mitchell 1960) were best predicted by FR at 750 m, the largest spatial scale measured. On the other hand, visits from the relatively large-bodied Bombus (body length 5–28 mm; Laverty and Harder 1988) and Megachile (body length 6–25 mm; Sheffield et al. 2011) were best predicted by the abundance of FR within just 250 m around a sampling location. The spatial arrangement of both FR and nesting habitat in a landscape are likely the more relevant predictors of how far most bees are actually foraging (Zurbuchen et al. 2010). The spatial scales we have determined as the best predictor of bee visits may therefore represent the upper end of foraging areas used by bees in the majority of landscapes that we examined, with many bees actually foraging in smaller areas of landscapes with more densely packed resources.
By teasing apart the responses of individual genera, we discovered a diversity of responses among taxonomic groups, highlighting potential problems with lumping all non-Bombus bees into a single functional group, or with examining responses of the entire bee community to FR. Our research highlights the importance—particularly for social taxa—of not only the current floral resource landscape but also the FR present earlier in the season. This information can help determine how to configure agricultural landscapes in a way that promotes bee population persistence and growth, and thus, increases the pollination services crops receive. Our findings suggest that bees with longer flight periods likely benefit from continuous, consistent provision of FR throughout a single season, and high FR abundance early in the season. However, the number of bee visits observed may not be a good proxy for bee population sizes. An important next step will be to determine how across-year patterns in visitation rates change with spatiotemporal fluctuations in FR, particularly for single-brood species which can only respond positively to FR availability over this timescale. Though we observed fewer visits by bees with short foraging periods in landscapes with a high present abundance of FR, this pattern should not hold across years: if other factors are not limiting, more abundant floral resources in landscapes should yield higher bee abundances in subsequent years.
Acknowledgements
Thank you to Lenore Fahrig for providing the GPS unit; to Joseph Bennett, Lenore Fahrig, Jeremy Kerr, and Risa Sargent for their input on early drafts; and to Katie Baillie-David, Alexander Hare, and Megan McAulay for assistance in the field and lab. This work was funded by the University of Ottawa and an Ontario Graduate Scholarship to JG.