Every plant in the Arboretum was mapped with geographic coordinates as part of an earlier Arboretum study .Monthly bloom times were generalized to the plant genus level by recording the longest demonstrated bloom duration recorded for the plant genus or for any species within it. If research needed to be done to determine a bloom time, often, only the genus phylogenetic level of identification was possible. Though generalizations are not ideal from both phylogenetic and temporal perspectives, we believe these are reasonable estimates of responses based on the information currently available. Moreover, the same genus identification level of plant was possible in the field and is consistent throughout the study. Though originally recorded in Microsoft Excel , the bloom time table was subsequently joined to the Arboretum’s plant records table in the geodatabase , assigning a bloom duration by month to every plant genus. Observational bee-to-plant data collection was done on weekly intervals and compiled into monthly aggregates minus redundant association data per month . Each of the Arboretum’s gardens was sampled with random ordering, with some exceptions due to construction access issues and the fact that the Arboretum path is linear, composed of a loop, thus, cut flower bucket garden sampling order followed the provided path, but the direction and starting point were randomized throughout the study to reduce time-of-day bias.
Sampling was conducted for one calendar year, starting in January 2017 and ending in November 2017. Importantly, bee-toplant data were collected by garden, each of which represents a unique novel ecosystem, with distinct geographic boundaries.WHR predicted and observed habitat were assessed by comparing their polygonal area and overlap or lack thereof. First, potential foraging habitat was mapped by querying and buffering suitable plant locations for each bee genus in Chapter 1. The buffer distance used was based on the foraging radii of the respective bee genera being examined. Table 1 lists all bee genera and their estimated foraging distances . All mapping methods were implemented entirely with ArcGIS . The entire plant map methods section is diagrammed in Part A of Figure 1. To map potential habitat for each bee genus WHR model, plant genus names were queried in ArcMap using the ‘select by attributes’ tool. Then a new layer of the selected plants was made by exporting the Arboretum’s plant point geodatabase layer . The plant point data were then buffered with the respective bee genus’ foraging radius. The buffer distance was based on the bee genera’s body size . All plants were selected and buffered at the relevant radii before further analysis was done . Next, the buffered individual plant point data were combined using the ‘merge’ tool , and then followed by the ‘dissolve’ tool to eliminate interior boundaries and create a single polygon of contiguous predicted habitat . See Table 2 for an example of one bee genus’ predicted potential habitat with bloom times. The area for each contiguous habitat polygon was then calculated. It should be noted that plants not found in the Arboretum geodatabase were excluded from the mapping analysis.Garden design had a major effect on ecological value for bees. Overall, the number of gardens hosting bee foraging and the number of bee genera were quite consistent over time .
The number of floral resources, available to bees showed a marked increase in availability around April through August . While the number of foraging bee genera remained fairly constant, at approximately 13 bee genera on average. There was a clear turnover of bee genera throughout the year . Most arboretum bees seem to experience habitat continuity while only a few seem to experience gaps between utilized habitat areas. Area measurements were also made of each polygon and shown in Table 6 and Appendix 2. Both of which help to demonstrate the area calculations from each bee genera’s monthly and annual habitat trends. The complete tabular mapping data for all bees is shown in Appendix 5. See Table 6 for an example of the tabular monthly analysis of predicted versus utilized forage area and annual combination maps’ data, denoted as “annual” for bee genus Agapostemon.Table 4 shows bloom times for all plants suitable for bees per the literature and the partial WHR models from Chapter 1. The partial model contains 134 plant genera. The lowest predicted bloom month was December with only 28 possibly blooming plants. June was the month predicted to have the highest amount of plant genera blooming, with 120. Additionally, Table 5 demonstrates bloom times all plants observed for forage full WHR forage by bees in the Arboretum. The lowest predicted bloom month was December with 70 plant genera and the highest was June with 245 potential plant genus blooms. The totals for Table 4 and Table 5 have been summarized in Figure 4. While both tables help to visualize the potential floral resource in each month throughout the year with detail, Figure 4 compares the totals per month directly between the full and partial WHR model. Figure 4 demonstrates that bees are using nearly double the number of flowering plants per month than the partial WHR model suggests.The partial WHR has less potential foraging habitat than the full WHR . Area measurements and comparisons for both models are shown in Appendix 5.
Thirteen result maps were produced for each bee genus for both the partial and full WHR models. One map was made for each month as well as an annual composite map . Each bee genera map page has 13 maps per page for both the partial and full WHR models. Overall, 21 bee genera were expected to be seen in the Arboretum, being commonly found in California, however, one genus, Colletes, was never observed. I found 27 bee genera foraging total in the Arboretum, five of which were not commonly found throughout California. Foraging data sources varied in their ability to accurately predict if bees were common at the study site. For example, all but one of the bee genera described as common from the Frankie et al. 2014 source was confirmed for forage within the Arboretum. The Xerces list contained five bee genera , none of which were observed in the Arboretum. Overall, a great diversity of bee genera and their various foraging plants demonstrated the enormous ecological value of the Arboretum to native bees, which the full model demonstrated more effectively.WHR modeling revealed how rich the Arboretum is as a resource for native California bees, representing a near best case scenario in terms of habitat continuity in space and time. While all gardens do not function equally, there are plenty of bee foraging plants. Most bees appear to have connected habitat for the annual compilations . Even the partial WHR plant selection shown in Appendix 1, demonstrated abundant habitat throughout the entire length of the Arboretum for most bee genera. The full WHR habitat map set demonstrates spatially how strategic planting design can have key positive effects on native bees, flower display buckets even in human-dominated and artificially created environments. We hope that the compiled bloomtime database will serve as a possible source to aid in further plant phenological studies in the future. The full WHR results enforce the notion that bees do not experience habitat fragmentation in the Arboretum at their foraging scales.Importantly, garden design plant palettes were shown to have a significant ecological impact on the site’s bees. There appears to be several explanations for these differences. First, certain plants are more attractive to some bee genera and not to others. Another explanation may be that plant genera with different bloom times affect when and where bees will forage. Furthermore, garden peak blooms were offset between gardens, even the top two, for example, the MWB native plant garden experiences peak bee-to-flower associations in May, nearly identical association values are shown in the ornamental STOR garden in August . Thus, while both the MWB and STOR gardens demonstrate the best floral resource provisions and foraging in the entire Arboretum, their plant palettes create a significant variation in timing. These data show the spatial implications of the suitability of foraging plants, to ensure that basic bee foraging needs would be met in a strategic manner. California offers a high degree of both plant and bee diversity . High plant and bee diversity values contribute to novel foraging association re-combinations of bees and plants . With the demonstrated opportunistic foraging demonstrated in Chapter 1, there appears to be great potential for further integration of these bee genera into human-dominated environments. While some bees are being supported in these systems, the response currently varies. Importantly, most urban landscapes do not consistently possess the characteristics of plantings within the Arboretum .
Strategic planting design work could be done to maximize creation of more suitable anthroscape habitats for bees, in both spatial and temporal dimensions. Novel urban ecosystems have been known to provide habitat value for both bees and other animal species . Bee genera experience urban landscapes differently from each other due to variation in bee foraging preferences, annual timing, and foraging distance. For example, Bombus, an early spring-emerging generalist, forages on many plants in bloom in the Mary Wattis Brown California native garden. However, a mid-summer-emerging bee, such as Peponapis, which specializes on particular flowers, would be limited in finding forage plants throughout much of the year. Consequently, Peponapis is confined to the mid-summer peaking gardens with appropriate flowers, found in such places as the Southwest U.S. and Desert gardens in the Arboretum. Based on these examples, the timing and types of both bees and flowers creates dynamic habitat availability and usage throughout the year in time and space. For bees, plants in a site will be assessed as either attractive for foraging or not. It is known that unique horticultural planting designs represent novel ecosystems to pollinators . This research assists in gaining spatial clarity and understanding about which native bees use novel ecosystems in horticultural plantings. This information should help aid landscape designers best modify and improve existing anthroscapes to maximize pollination ecosystem services for the future. In this way, novel ecosystems can be designed to act as bee source habitats, which will contribute to ecological resilience. These findings suggest that habitat continuity should be studied at this scale or coarser. Through mapping, observation, and analysis this type of research helps us to spatially understand how well various bee genera could persist in urban habitats. The term ‘bees’ encompasses a diverse suite of bee types, each with their own foraging preferences, dispersal abilities and other habitat requirements. Ultimately, this study provides insight into which urban habitats provide the best floral resources to native California bee genera to promote habitat continuity and combat habitat fragmentation.Some gardens are providing greater beneficial floral resources to foraging bees, such as the MWB and STOR gardens in the Arboretum than others. As seen in Chapter 1, both the MWB and STOR gardens far outperformed all other Arboretum gardens and are in effect acting as beneficial novel ecosystems, created and managed by people. The most ecologically valuable Arboretum gardens for bees showed high floral diversity in this study. This trend was also found by Parreno et al. where diversity, abundance, and health of wild bees was found to be the primarily driven by land-use modifications and that pollinator conservation depends, in part, on remedying negative land-use trends. It is important to preserve natural and human made habitats to help harbor bees, other beneficial insects and benefits to other ecosystem services . On the other hand, some gardens in the Arboretum represent poor habitat and did not support bee pollinators in terms of forage activity. The poorest gardens for bee foraging, with less than bee foraging visits for the year included: ACAC , COAS , ERIC-N , MOUN , and WALN . In comparison with other garden rankings, the gardens listed above were far below the average bee to garden annual visitation rate average of 132. Extrapolating garden variation findings to the greater context scale shows how important the success of each garden is to habitat network connectivity. By examining bee foraging associations prior to designing gardens , we can ensure that, for example, a designer would not opt for Acacia trees since they do not seem to hold much value to bee foraging.