Environmental Variance of Community Structure
Redundancy analysis (RDA) modelling was applied to Hellinger transformed observational data with respect to five environmental explanatory variables (wind speed, solar radiation, temperature, humidity, and field edge effects). While the model was statistically significant (F = 3.94, p< 0.001), it only explained 14% of the variance. All environmental variables are well represented by the axes but did not match the spread of communities in the model. The effect of sites on community composition appeared to be low since there was no distinct clustering with this variable. However, the model including temporal variables explained 65.2 % (R2 adj = 53.9%) of the variance (Figure 4a) was significant (F = 5.79, p< 0.001); and so were first 7 axes (ANOVA, p< 0.001). Most of the temporal variance was explained by sampling date (ANOVA, p< 0.001). Models that included date improved the explained variance by more than 45%. Time of day was also significant (ANOVA, p< 0.03), with ellipses depicting a small gradient across the communities (Figure 4a). General trends in insect community composition across the season showed consistent presence of solitary bees and Hemiptera, while the abundance of Syrphids and other Diptera, as well as observed Apis melliferashowed more seasonal variation (Figure 4b). While Formicidae (ants) were generally rare or absent, there was a large surge in their abundance on flowers for one week (May 18th). Coleoptera and Lepidoptera consistently showed low levels of occurrence on strawberry flowers (Figure 4b).