Pollinator interactions with yellow starthistle (Centaurea solstitialis) across urban, agricultural, and natural landscapes

Abstract

Pollinator-plant relationships are found to be particularly vulnerable to land use change. Yet despite extensive research in agricultural and natural systems, less attention has focused on these interactions in neighboring urban areas and its impact on pollination services. We investigated pollinator-plant interactions in a peri-urban landscape on the outskirts of the San Francisco Bay Area, California, where urban, agricultural, and natural land use types interface. We made standardized observations of floral visitation and measured seed set of yellow starthistle (Centaurea solstitialis), a common grassland invasive, to test the hypotheses that increasing urbanization decreases 1) rates of bee visitation, 2) viable seed set, and 3) the efficiency of pollination (relationship between bee visitation and seed set). We unexpectedly found that bee visitation was highest in urban and agricultural land use contexts, but in contrast, seed set rates in these human-altered landscapes were lower than in natural sites. An explanation for the discrepancy between floral visitation and seed set is that higher plant diversity in urban and agricultural areas, as a result of more introduced species, decreases pollinator efficiency. If these patterns are consistent across other plant species, the novel plant communities created in these managed landscapes and the generalist bee species that are favored by human-altered environments will reduce pollination services.

Figure 1. Map of study area and locations of plots in East Contra Costa County, California.

Light blue dots represent a 500(green), urban (red), and natural (yellow) land use types.

Figure 2. Box plots of bee visitation response variables in natural, agricultural, and urban sites.

Bee morphotype visitation data and calculated community metrics were collected in East Contra Costa County, California.

Figure 3. Bee visitation response variables as a function of surrounding anthropogenic land use.

Bee morphotype visitation data and calculated community metrics were collected in East Contra Costa County, California. To examine in more detail the effect of anthropogenic land use on bee visitation, we created a continuous variable for land use with an index ranging from agricultural to urban land use based on proportional area of each type within a 500-axis moves from left to right, sites go from being more agricultural to more urban.

Figure 4. Box plot demonstrating the effect of land use type on percentage of viable seeds.

Yellow starthistle seed heads were collected in East Contra Costa County, California and dissected in the lab after maturity. We calculated significance using a generalized linear mixed model fit with a binomial distribution, with land use type as a fixed effect and site as a random effect. With natural sites as the baseline, urban areas had significantly lower rates of seed set (effect size±SE = −0.756±0.371, p = 0.042).

Figure 5. Correlation between the percentage of viable seeds in each yellow starthistle seed head and the average number of site visits by morphotype.

Bee morphotype visitation data, calculated community metrics, and yellow starthistle seed heads were collected in East Contra Costa County, California. Bee morphotypes that averaged at least one visit per 30 minute observation window were included as fixed effects in a linear mixed model fit with a binomial distribution, with site as a random effect and the ratio of viable to total seeds as the response variable. Medium hairy leg bees (effect size±SE = 0.284±0.069, p<0.001) and shield-tipped small dark bees (effect size±SE = −0.155±0.051, p = 0.002) had significant effect sizes in the model. Regression lines were added to illustrate relationships.