Results and Discussion

Hierarchal Partitioning

Hierarchal partitioning was done for each spatial to determine which landcover types are explaining the most variation within the spatial scales (figure 8). Broadleaf forests were consistently found to be the most influential showing that no matter the distance to a site, broadleaf forests will be the most impactful on bee community composition. Coniferous forests are the only other land cover type that consistently showed any sort of influence on bee communities. The similarity between all four of these graphs further exemplify the lack of importance regarding the spatial scale of landcover impacting bee communites.

db-RDA

Per landcover spatial scale with adjustments for climatic conditions

All four distance based redundancy analyses show similar results in that broadleaf and coniferous forests as well as shrubland consistently have the longest vectors meaning they are the strongest drivers of community composition. Landcover types like developed, exposed land, grasslands, agriculture and rock and rubble tend to be clustered together and opposite to the top species. showing that the most common bee species are negatively impacted by these land cover types. Instead showing correlation with forest and shrubland landcover types. Bombus flavifrons is shows a strong relationship with broadleaf forests regardless of the spatial scale. Similarly, Bombus mixtus shows a tendency towards coniferous forests. Interestingly, the top five species shift slightly between spatial scales, at the smaller scales (figure 9 & 10) Bombus rucocincus is within the top five and shows a relationship with grasslands. As the spatial scale grows (figure 11 &12), this species become less relevant and none of the top five species show any liking towards grasslands. This suggests that grasslands nearby a nesting site could provide sources of food or habitat for certain species such as Bombus rucocincus but when looking the larger landscape scale grasslands no longer provide unique benefits. Another conclusion that can be drawn from these graphs is that developed land tends to have the most negative effect on Alberta's native bees.

Although some patterns can be interpreted from these graphs, the R squared values at all four scales are extremely low and even negative so no significant conclusions can be made. This is because of landcovers lack of explanation of bee community variance. Instead, climate is more explanatory. The 1000m scale is the only one with a positive R squared value (figure 11) suggesting that the intermediate spatial scale has the strongest influence on the bee community.

Comparing the Effects of Landcover With Climate

Table 3 shows that while land cover at 1000m accounts for less than 1% of the variation in the bee community, climate was found to explain almost 7% of variation. Climates influence on the community is enough for it to statistically significant. This result confirms that in Alberta, climate is more prevalent driver of bee communities than landcover type.

db-RDA per Climate Group

In order to answer my third objective —whether bees occurring in different climates across Alberta require different landcover types— I performed a cluster analysis to break the sites into 3 climatic groups. The groups that were found are outlined in table 4.

I did a hierchical partition and db-rda on landcover types at the 1000m buffer for each climate cluster. Both climate clusters one and three remain largely the same with broadleaf and coniferous forests showing the most impact on bee variability. However, cluster two representing hot and dry climates typically in Southern Alberta are highly impacted by agriculture, looking at the species vectors in figure 13 the top species in this cluster tend to have a positive correlation with agriculture. Surprisingly, these species seem to be negatively impacted by broadleaf forests but positvely by agriculture in these climatic conditions. The lack of native grasslands and forested areas in Southern Alberta seem to have forces the bees to become reliant on agricultural crops. This implies that conservation of bees in the grasslands ecoregion would look very different compared to in the rest of the province. Additionally, the top species within each climate cluster vary slightly, with cluster two being the most unique. This is to be expected and demonstrates that management of wild bees in Alberta is not a one size fits all solution due to wide climate gradient which favours different species. 

Revisiting Objectives

1.) Identify the spatial scale at which surrounding land cover most strongly shapes native bee communities in Alberta to determine whether conservation efforts should focus on local habitat enhancement or landscape-scale habitat protection.

No single spatial scale was found to be significantly more influential to wild bee communities than another. Suggesting that the size of conservation efforts is largely irrelevant. While large areas of habitat restoration and protection such as provincial parks would be greatly beneficial to bee communties, smaller sized efforts like habitat restoration on landowners property could be equally beneficial to these pollinators.

2.) Identify which land cover types are most strongly associated with differences in native bee community composition across Alberta to help prioritize habitats for bee conservation.

Broadleaf forests tend to be the most impactful on bee communities throughout Alberta. However, this is dependent on the climate and the specific species that are focused on. The bee community tended to be the most negatively affected by areas of developed land. 

3.) Evaluate how broad climatic gradients across Alberta shape native bee communities, and whether bee communities occurring under different climatic conditions may require different habitat priorities for conservation.

Bees whom are favoured by hotter, drier conditions seen in Southern Alberta are more positively impacted by agriculture rather than broadleaf forests. The shifts in influential landcover types across different climates highlight that a universal solution for bee conservation in Alberta is not possible. Instead, conservation needs will need to be assessed on a case by case basis throughout the province.

Conclusion

Because of Alberta's extreme diversity in climate, landcover and bee species it is difficult to pin point a specific driver of bee communities, in particular, which spatial scale is the most influential. I can conclude that climate plays a larger role in shaping communities rather than land cover and therefore conservation planning for Alberta's wild bees needs to be tailored to each particular ecoregion of the province. Additionally, climate change could be a more worrisome driver of bee diversity decline rather than habitat loss.

Potentially more conclusive results regarding which spatial scale of landcover is more influential to native bees could be drawn if the study area or number of species was significantly narrowed.