Do Human-modified Landscapes Affect Solitary Bee Diversity, Foraging, and Reproduction in Northern Florida?

2011 Annual Report for GS10-092

Project Type: Graduate Student
Funds awarded in 2010: $10,000.00
Projected End Date: 12/31/2013
Grant Recipient: University of Florida
Region: Southern
State: Florida
Graduate Student:
Major Professor:
Dr. Katie Sieving
Wildlife Ecology / UF

Do Human-modified Landscapes Affect Solitary Bee Diversity, Foraging, and Reproduction in Northern Florida?


New Title: Influence of Land Cover and Farmland Habitat on Bee Diversity, Foraging and Pollen Collection in Northern Florida

Eleven farmers on twelve farm properties agreed to participate in the native bee study in 2011. Field work was active in March, April, May and June. Collections ceased in July when blooming crops became rare. A second field season began in March 2012 and will continue through the fall with a break during the late summer months. Several aspects of the study were successfully implemented in 2011 including collection and processing of: 1) bees from bowls, 2) bees carrying pollen, and 3) plants for the pollen reference collection. Also, plant composition and structure data was successfully collected on farms.

Objectives/Performance Targets


The objective of my study is to determine whether on farm management of vegetation is a better predictor of native bee species richness than surrounding landscape factors. The twelve farms in the study are variable in on farm vegetation management from closely mown lawn-like field verges to weedy verges and purposeful plantings of white clover and sunflowers as crops or ground cover. Do such plantings and vegetation affect the native bee diversity on these farms, or is diversity driven by landscape context? To answer this question, my study has four main objectives: 1) analysis of bee foraging habitat, 2) characterization of the landscape context of farms, 3) analysis of native bee species richness, and 4) analysis of pollen collection on farms by native bees.

Figure 1 describes the focus of my study on the influences of the surrounding landscape and on farm habitat on bee diversity on farms. On the farms themselves I’ll be looking at both diversity of native bees and pollen collection by captured individuals. The farms range from certified organic to low pesticide use.

Performance Targets

•Landscape Context Effects: The landscapes around the 12 farms in the study are dominated by forest (pine plantations and mixed oak woodlands) and agricultural land covers (crops and pastures). The landscape mosaic around these farms also includes developed areas (i.e., roads, parking lots, and houses) and wetlands. See Table 1 for dominant and secondary land covers around farms.

Ultimately, I want to assess land cover at different scales to find out what features correlate best with native bee diversity. I’ll address the following issues:

• What are the implications of using land cover vs. land use in assessing landscapes for bees? For example, pasture is an agricultural land use that is typically characterized by mixed-species grassland land cover along with some trees. Although crops and pasture are both agricultural land uses, pasture has higher potential for bee nesting and foraging than many kinds of crop fields (e.g. grain fields). Similarly, oak forest and pine forest may have different values for bee nesting. The relative values of land covers and land uses need to be assessed as I move further into the landscape analysis.

• What techniques should I use to counter spatial autocorrelation between buffers at different distances around farms? Spatial autocorrelation is the tendency of points or areas of a geographic space to be more similar (positive autocorrelation) or less similar (negative autocorrelation) than expected for random pairs of observations (Legendre 1993). During the landscape data analysis I will be looking for correlations between landscape buffer areas and bees species richness on farms, but spatial autocorrelation between buffers may create stronger correlations than would otherwise exist. A few methods are available to avoid or compensate for spatial autocorrelation.

• Also, although the farms generally grow the same crops there is a lot of variation in farm size. To assess the landscape context of a large-sized farm like Island Grove in the same way as a small farm like Sandhill, buffer zones may have to begin at the farm edge. However, this raises the question of what is the usable landscape for a bee that is only a quarter-inch long , and is it likely that a small bee’s landscape extends far off the farm?

Landscape structure is an important part of my study. Although I have only done a preliminary portion of it for this report (see Accomplishments/Milestones), the results look promising for my final report and dissertation.

• Bee Foraging Habitat: The focus of this part of my study is to quantify the foraging habitat available for bees on the twelve farms. Foraging habitat includes crops that are flowering at different sampling times, farm edge vegetation, and habitat contained on some farms like preserved wetlands. I will determine if such habitats correlate with bee diversity on the farms.

•Bee Species Richness: This part of my study focuses on bee diversity and abundance on the farms. I do not sample bees in the broader landscape as I am interested in the native bees that farms support in some way. Two kinds of sampling contribute to this data: hand netting and bee bowls.

Bee Pollen Choices: Here I focus on a possible pollination precursor- what bees are carrying as they forage on farms. Native bees collect and carry pollen to store for later consumption by their offspring, but the facts that they have this pollen on their bodies increases the likelihood that they may be pollinating flowers as they move between flowers. Figuring out if they are actually pollinating flowers or their effectiveness at pollination is beyond the scope of this study.


Landscape Context Effects: A total of twelve farms now participate in this native bee study. Because a couple of farms were added late, field data collection will continue until May 1st. All bees collected in 2011 have been processed, labeled and stored for identification (which is ongoing). Approximately 20 pollen samples from bees collected in 2011 (mostly in May and June) have been processed so far. I’ve maintained good relationships with all the farmers in the study (by adhering to their instructions regarding farm visits and bowl placement to avoid active work sites) and stimulated more farmer interest in native bees. In the Fall of 2011 I passed my qualifying exams and became a PhD candidate, allowing me to continue with this study.

At this stage I’ve just assessed land cover at a1 km buffer size. As data collection continues multiple buffers around farms will be established between 250 m and 2 km. Dominant land covers around farms occupy more than 51% of the landscape (See Table 1). The landscape context portion of the study focuses on agricultural vs. natural land use categories. Preliminary landscape data shows that five of the farms I’ve sampled are surrounded by forest-dominated landscape and seven are in landscapes dominated by agricultural crops and pasture. Figure 2 shows the landscape around one farm and Table 1 shows the results of the assessment.

Bee foraging Habitat: I’m using quadrat sampling along transects to document the height structure, coverage, and composition of plants in crop fields focusing on crops that are in bloom at the time of sampling (preferably blooms that lead to harvestable fruit or vegetables, but also the bolted greens common in fields in early Spring). I’ve also been sampling around farm edges (approximately 50 meters of sampling to every 200 meters of edge) to characterize the nearest vegetation to crop fields that may contribute to bee habitat A few of the farms in the study contain habitats like wetlands that may contribute to bee habitat. These habitats have also been sampled for structure, percent cover, and composition.

Bee Species Richness: Approximately 300 specimens were collected by hand and through the use of bee bowls in 2011. I’m still in the process of identifying specimens, but Table 2 shows how many species were collected by farm, collection technique, and month. Bee bowls are especially good for capturing medium and small-sized bees, and hand capture with nets tends to favor capture of larger, slower moving bees, so the two techniques compliment each other.

Bee bowls ((i.e., soapy water in UV-painted cups that mimic flowers, see below) are placed along transects where crop plants are blooming. Nine cups are laid out along each transect in a flowering crop to capture bees that pass through the area. I have not been sampling non-blooming crops unless they are located adjacent to blooming crops on a transect.

Table 2 may underestimate the number of species collected because the very small bees in my collection are superficially alike and were therefore more likely to be grouped in this preliminary assessment.

Bee Pollen Choices: Ninety per cent of the pollen samples from native bees collected in 2011 have been processed and permanently mounted in glycerin jelly. There are also more native bees in the general collection that were carrying pollen when captured, but I will not process them for pollen until after they have been identified since the process for removing pollen will probably destroy the specimens. Identification of processed pollen will begin in May. Dried flowering plants from each farm visit in 2011 are stored awaiting processing for the pollen reference collection. I have narrowed my collection focus to those plants that have bee activity at the flowers or at least looked like they had potential for pollen or nectar collection.

Impacts and Contributions/Outcomes

In light of the decline in honey bee populations in North America there is speculation about the ability of native bees to pollinate crops, possibly replacing or at least supplementing honey bee pollination services. This project looks at two critical questions: do native bees carry crop pollen, and is it local vegetation structure around farms or broader landscape context that has more effect on native bee populations? The first question relates to native bees’ potential to pollinate crops and the second will provide basic information on whether on-farm vegetation management would be sufficient to support native bee communities. In other words, the results of this study will help farmers to know whether native bees are dependent on where their farms are located, how vegetation is managed on their farm, or some combination of the two.


Dr. Kathryn Sieving
Dept. Wildlife Ecology & Conservation
110 Newins-Ziegler Hall, University of Florida
Gainesville, FL 32611-0430
Office Phone: 3528460569
Rosalyn Johnson
Grad Student
110 NZ Hall
Wildlife Ecology UF
Gainesville, FL 32611-0430
Office Phone: 3528460569