Exploring relationships between pollinators and canola on the Palouse

Final report for SW18-031

Project Type: Research and Education
Funds awarded in 2018: $207,134.00
Projected End Date: 03/31/2022
Grant Recipient: Washington State University
Region: Western
State: Washington
Principal Investigator:
Dr. David Crowder
Washington State University
Expand All

Project Information


Canola is expanding rapidly within the Palouse region of eastern Washington and northern Idaho, with acreage in Washington State alone growing from 15K to 135K acres from 2012 to 2022. Canola is being adopted by growers because of its potential to diversify dryland cereal-legume rotations and due to its ability to be sold as a food, biodiesel, and feed crop. Canola crops also provide a large and persistent source of pollen and nectar, and canola flowers are visited by honeybees and diverse native species including bumblebees, mason bees, and sweat bees. While many canola fields in the Palouse region are surrounded primarily by dryland agriculture, others are near sizeable tracts of remnant prairie and other insect-friendly habitat. Our preliminary data suggested that canola yields appear to increase at sites close to natural habitat, possibly due to increased bee abundance in these areas. This project was designed to further test the hypothesis that increasing natural habitat surrounding canola fields promotes increased bee abundance and seed yields. Secondly, our project explored how variation in canola plant traits affected attractiveness to bees and plant yields. Third, we assessed how abiotic and biotic stressors affected canola traits important for bee pollinators. Our research objectives were linked with active extension efforts to determine how growers might be able to modify their management practices to promote high pollinator activity on their fields while maintaining adequate insect pest management. Overall, our project promoted a better understanding of the role of pollinators in canola crop production, and baseline data to further improve canola production and bee health. Our project met several goals of Western SARE: (1) promote good stewardship of natural resources (by providing information on how growers can promote pollinator diversity and quality canola); (2) promote crop, livestock, and enterprise diversification (by aiding aided growers increase their enterprise diversity and pollinator diversity through canola production); and (3) examine the environmental implications of adopting sustainable agriculture practices and systems (by exploring how farming practices affected pollinators and canola crops).

Project Objectives:

(1) Explore relationships between landscape structure, bees, and canola yields

(2) Determine how farm management practices affect canola traits and pollinator attraction

(3) Assess how abiotic and biotic stress affect canola traits and pollinator attraction

(4) Educate growers on pollinator management in canola


Click linked name(s) to expand/collapse or show everyone's info



Studies from Canada and Europe suggest that yields and quality of canola increase when considerable natural habitat surrounds fields, potentially due to increased bee abundance in such landscapes. However, similar relationships have not been assessed in the Pacific Northwest USA, where canola acreage is increasing rapidly. Our project tested the hypothesis that diverse crop landscapes of the Palouse region (eastern Washington/northern Idaho) would have more diverse and abundant wild bee populations, and that canola yields would be highest in such regions. We also predicted that canola producers would have lower reliance on honeybees in more diverse landscapes. By sampling canola fields throughout the Palouse, our project gathered important baseline data on the abundance and diversity of wild bee populations at canola fields across landscapes of varying diversity. These data were used to test the hypothesis that bee communities, and resulting canola yields, were impacted by the characteristics of the surrounding landscape.

Our preliminary data also suggested that canola flowers with high sugar content attract bees, which leads to higher yields. Other traits, however, such as nectar volume and and flower size, might also affect bee attraction to canola flowers. Given that the Palouse is a highly fragmented matrix with many crop fields and little natural habitat, growing highly attractive canola is critical to attract bees that might forage over long distances. Our project was designed to test which canola traits affect attraction to pollinators. We predicted that floral/nectar/pollen traits can be influenced by factors including crop variety, crop rotation, soil quality, and the level of abiotic and biotic stress.

Materials and methods:

Objective 1 - Explore relationships between landscape structure, bees, and canola yields

Working with a network of growers, in each year (2018 to 2020) we located 8 winter canola and 8-12 spring canola fields to sample, each separated by at least 2km. Co-PI Sowers, the extension lead on the project, has relationships with most canola producers in Washington and Idaho. She facilitated the selection of farm sites and cooperators throughout the project along with the graduate students and postdoctoral researchers in PI Crowder's laboratory. In each year, each field was sampled twice for bees using bee bowls filled with soapy water and blue vane traps. Traps were deployed in the morning on sunny days along the edge of a canola field and left for 24h. From the samples, we identified all bees collected and determined the abundance of each species and the diversity (# of species) of the overall community. We also collected landscape and weather data from each site from the USDA Cropland Data Layer maps and NOAA. Data on nectar sugar content, nectar volume, and flower petal size were collected from 20 random plants in each field. Canola yield and agronomic practices related to tillage and irrigated were obtained directly from growers at the end of each field season.

We analyzed factors affecting bee abundance and richness using a statistical approach known as structural equation modeling, which allows for the analysis of multiple response variables at the same time (see Figure 1 link below). Structural equation modeling also allows for the assessment of both direct and indirect relationships among many explanatory and response variables (see Figure 1 link below). In our analyses, we explored how farm management factors (tillage, and irrigation), seasonality (early vs late), and landscape characteristics affected the traits of canola crops and resulting bee abundance and diversity. We categorized landscapes based on a simple metric of the percentage of “natural habitat” (which in our region largely consists of remnant prairie or shrub-steppe habitat), a metric that has been widely used in studies of landscape effects on bees. Within our structural equation modelling approach, we also considered how traits of canola flowers themselves (nectar volume, petal size, sugar) interacted with agronomic and landscape factors to affect bee communities. Based on our hypotheses, we expected to see an increase in bee abundance and diversity at sites with more natural habitat. We also expected to see fewer bees in canola fields with tillage, as this may disturb nesting sites for ground-nesting bees.

Figure 1. Predicted relationships among bee pollinators and characteristics of canola crop field sites

To follow up on these analyses, we analyzed canola yields as a function of bee abundance, natural habitat, tillage, irrigation, and the interaction between these variables using linear models . These analyses were conducted to determine how bee communities were affected by canola production practices and landscape characteristics, and whether bees influenced canola yields when considered relative to other factors such as tillage and irrigation.

Objective 2 – Determine how farm management practices affect canola traits and pollinator attraction

On each farm in Objective 1, we sampled floral traits that may affect bee attraction to canola fields. These data were collected once per year during peak bloom from 20 canola plants per field. The variables measured were: (1) nectar sugar content; (2) nectar volume; (3) number of flowers; and (4) flower size. Nectar production was quantified from the inner (lateral) nectaries only. One sample of nectar per plant was used for analysis, comprised of nectar collected into a microcapillary tube from 5 flowers. The length of the column of nectar in the microcapillary tube was divided by the number of flowers sampled to yield mean volume (μL) per flower. We assessed nectar sugar content with an evaporative light scattering detector. Towards the end of the growing season, seed pods were harvested just before ripening. After all pods were harvested, we scored the total number of pods per plant and weight of seeds. Remaining aboveground plant biomass was then harvested, dried at 70°C for 48hr, and weighed. Finally, we analyzed the oil and chlorophyll contents of seeds. Working with growers, we collected data on agronomic practices from each farm, including: (1) canola variety; (2) Round-up Ready/not; (3) neonicotinoid use; (4) diversity of crop rotation; and (5) soil traits (pH, organic matter, nutrient levels, microbial biomass).

We analyzed effects of different explanatory variables (agronomic practices, landscape factors) on canola floral traits using linear models. This allowed us to determine specific management practices, or canola varieties, that promote desirable floral traits. We expected that the the diversity of the crop rotation and soil traits would affect floral and nectar traits of canola and determine the attractiveness of different varieties. We expected that fields with more desirable traits would attract a greater abundance and diversity of bees, which would promote high yields and quality.

We also used ANOVA to determine if floral traits differed between canola varieties. Field was included in all analyses as a random variable (to control for variation across the farms), and separate analyses were be conducted for both spring and winter canola. We also used ANOVA to determine if rates of pollinator visitation (which is a proxy for pollination services) are affected by canola variety. These trials determined the varieties that are naturally most attractive for pollinators, and identified the mechanisms (i.e., increased nectar sugar content) that cause the difference in attractiveness.

Objective 3 - Assess how abiotic and biotic stress affect canola traits and pollinator attraction

Plants have multiple traits that attract pollinators, including nectar as a food reward and flowers as a visual attractant. Floral traits that attract bees and other pollinators can be strongly influenced by abiotic and biotic stress, such as drought or pathogens. In this objective, we explored how an abiotic and a biotic stressor (drought and a fungal pathogen) affected several traits that may signal or attract pollinators in three varieties of canola.

We assessed effects of drought and a fungal pathogen on canola varieties with four greenhouse experiments. In each experiment, we planted two canola seeds 1.9 cm into potting soil in 8.9 cm deep plastic greenhouse pots. After 14 d, plants were culled so each pot contained a single plant. For all drought stress experiments, water treatments were started at planting. Fully watered plants received 100 mL of water, three times per week, and droughted plants received 75mL of water three times per week. For the fungal pathogen experiment, we obtained oats that had been inoculated with Rhizoctonia AG8, prepared by the WSU and USDA plant pathology labs in Pullman, WA.

Our first experiment tested effects of drought on the canola variety HyClass930, the most widely grown variety in the Palouse region. This experiment had two treatments: (i) full water (n = 50) and (ii) drought (n = 50). The second experiment expanded the study of drought by applying both water treatments to three canola varieties (i) HyClass 930, (ii) NCC101S, and (iii) InVigorL233P. Both NCC101S and InVigorL233P have recently become commercialized as high yielding varieties, with InVigorL233P marketed for drought tolerance. There were 144 plants total in this experiment (24 replicates × 3 varieties × 2 water treatments). Our third experiment tested effects of the fungal pathogen, Rhizoctonia solanum AG8-C3 (Rhizoctonia), on the HyClass930 variety. During planting, we added ground oats inoculated with the fungus to the soil (1:10 ratio by mass of oats to soil) following methods described by Babiker et al. (2013). We added clean (uninfected) oat powder (prepared identically to the inoculated powder) to account for soil textural changes in control replicates. There were 144 plants total in this experiment (24 replicates × 3 varieties × 2 Rhizoctonia treatments).

We employed a factorial combination of drought and fungal pathogen treatments across three canola varieties for the fourth experiment. The same three varieties previously described received drought and/or pathogen treatments as outlined in the previous three experiments. There were 288 plants total in this experiment (24 replicates × 3 varieties × 2 water treatments × 2 Rhizoctonia treatments). In this experiment, we culled 12 plants from each treatment group on the 21st day after seeding and preserved the plant tissue for molecular testing on gene transcript responses related to each treatment.

We collected the following data for each experiment: (i) days to flower: number of days from planting to opening of the first flower, (ii) peak flowers: number of flowers open on the fifth day after the first flower was recorded, (iii) average petal area: average area (mm) from three randomly selected petals, (iv) average nectar volume, and (v) nectar sugar concentration: averaged from nectar from three randomly selected flowers on each plant.

We used ANOVA to assess if the various experimental treatments affected floral traits in each experiment.

Objective 4 - Educate growers on pollinator management in canola

(A) Field tours - Each year our team participated in 4-8 “canola field tours”, where we talked to producers about pollinator conservation and the role of pollinators for canola crops. We had previously found that producers are more likely to attend field days if they are given information on many subjects, and we organized each event with other research and extension scientists working in canola at WSU to develop a broad program. While one of our main goals was to spread information on pollinators, by developing a broad agenda for each field day we attracted more producers.

In each year of the project, we also conduct 2-4 field tours on farms with spring canola, and 2-4 field tours on farms with winter canola. This allowed us to capture variation among these crop types while also attracting a broad group of interested stakeholders. Field tours were spread across a broad geographic area, and spaced out in time as much as possible, to ensure that we reached as many producers as we could during the project.

(B) “Stop and Talk” Events – One innovative format that we included in our outreach was “Stop and Talk” events. We encouraged canola producers to directly contact our project team if they had time-sensitive issues related to canola production, particularly issues related to pollinators, integrated pest management, or integrated pollinator management. After speaking to producers on the phone, we arranged a site visit within a week after the call, and we invited neighboring farmers to join us. During these site visits we scouted the canola field and helped assess and mitigate their production challenges. Over the course of the project we conducted approximately 20 "Stop and Talk" events. As canola is a growing crop, building a network of producers that our team regularly interacted with was a major boon for extension as canola acreage continues to expand on the Palouse.

(C) Extension bulletins – We prepared a bulletin on pollinators in canola to serve as an identification guide and a guide for pollinator management. Prior to this project, the US Canola Association’s Grower’s Manual did not have any information on the management of pollinators, but info from our WSU extension bulletin was incorporated into this manual. One of our growers (Hennings) sits on the board of the US Canola Association and helped ensure that our bulletin was widely shared through their website and reached as many producers as possible. Bulletins were also published ill also be published to the WOCS Website (css.wsu.edu/oilseeds/) and the WSU small grains website (smallgrains.wsu.edu). Both Crowder and Sowers are members of the WOCS and “Small Grains Extension Teams” at WSU, and these site serves as a clearinghouse for information for many canola and cereal-legume producers (which is the group we targeted as potential canola producers).

(D) Websites – We published data to the WOCS website (css.wsu.edu/oilseeds/) and the WSU small grains website (smallgrains.wsu.edu).

(E) Canola Production Guide – We incorporated our research into the Pacific Northwest Canola Production Guide by describing pollinator management for canola and the role of integrated farm management for controlling pests without disrupting pollinators. As described above, we also incorporated our results into the US Canola Association’s Grower Manual to increase dissemination of our results.

(F) Outreach presentations and podcasts – Members of our project team shared results at academic conferences including the 2018 meeting of the American Society of Agronomy, which was held in conjunction with the meeting of the US Canola Association. The two lead PIs (Crowder and Sowers) travelled to this meeting along with producers Hennings, Jordan, and Emtman. By attending the meeting as a project team, we interact broadly with other producers and stakeholders to develop our project and ensure it aligned with national stakeholder needs.

Research results and discussion:

Objective 1 - Explore relationships between landscape structure, bees, and canola yields

In 2018 we sampled bees at 8 winter and 8 spring canola fields. This sampling was repeated at 8 winter and 11 spring canola fields in 2019 and at 8 winter and 12 spring canola fields in 2020 (see link to Figure 2 below). Sites were sampled based on the methodologies described above to collect bees and sample plant traits. At each of these sites we examined floral traits of canola fields and gathered data on agronomic practices used by growers, particularly practices related to tillage and irrigation.

Figure 2. Canola sites that were sampled for bee pollinators and floral traits from 2018 to 2020

Bees collected from each year were identified to species and were also classified into broader taxonomic groups (genera and/or families). We collected individuals from groups such as honeybees, bumble bees, and sweat bees, along with many other wild bee species (see link to Figure 3 below). We found 52 unique species in total, which represents nearly 50% of the bee species known to exist on the Palouse.

Figure 3. Photos of a honey bee, bumble bee, and a mining bee visiting canola flowers

The data on wild bee communities showed several key trends. First, wild bees are not highly abundant in most canola fields, with densities at most fields totaling between 10 and 50 bees collected (see link to Figure 4 below). Some fields, however, supported large numbers of bees (over 750 at one field site). Sites that had the most bees were dominated by small halictid bees, while sites with fewer bees had more honey bees and bumblebees (see link to Figure 4 below).

Figure 4. Total abundance of bees at canola sites used in the study

From our structural equation modeling we found that bee abundance was greater in landscapes with more natural habitat and less development (see link to Figure 5 below). Pollinator diversity was also strongly affected by tillage (see link to Figure 6 below). However, canola yield was not strongly affected by bee communities. The only factors we found that strongly influenced canola yield were tillage and herbivore richness. Fields with intermediate tillage had the highest yields; canola yields also increased in fields with lower herbivore diversity (see link to Figure 7 below). This suggests optimal pollinator outcomes and canola yields intersected in fields with intermediate tillage.

Figure 5. Output of structural equation modeling showing relationships among bee pollinators and canola crops

Figure 6. Bee species richness was affected by tillage practices in canola fields

Figure 7. Canola crop yields were affected by tillage practices and herbivore richness


Objective 2 – Determine how farm management practices affect canola traits and pollinator attraction

Nectar sugar concentration, nectar volume, and flower petal size were widely variable among sites, even for sites that grew the same canola variety (see links to Figures 8 to 10 below. While we were not able to collect nectar from some sites, for sites where we did collect these data both the concentrations and volume varied widely. This variability is interesting because it suggests that farm management practices can produce variable plant traits that are more or less appealing for bees.

Figure 8. Flower petal size at sites with different canola varieties

Figure 9. Flower nectar volume at sites with different canola varieties

Figure 10. Nectar sugar concentration at sites with different canola varieties


Objective 3 - Assess how abiotic and biotic stress affected canola traits and pollinator attraction

We added an objective in 2020 to explore how an abiotic and a biotic stressor (drought and a fungal pathogen) affected several traits that may signal or attract pollinators in three varieties of canola. These studies were conducted to gather further information on how farmer's management of the total agricultural ecosystem may affect pollinators. From our work we found strong differences in flower production, flower petal size, and nectar quantity and quality when plants were exposed to drought or pathogens, and we observed genetic expression change in response to these stressors (see links to Figures 11 to 14 below). Interestingly, while water stress reduced flower petal size and both nectar volume and sugar concentration, each of these traits actually increased in response to stress from rhizoctonia. This suggests that canola plants affected by biotic stressors may increase investment in traits that attract bees, but plants that were water stressed had limited attractiveness to bee pollinators. Furthermore, the effect of both the abiotic and biotic stressor on canola flower traits varied across varieties, suggesting strong genotype by environment interactions affected plant phenotype (see links to Figures 11 to 14 below).

Figure 11. Number of open flowers for different canola varieties exposed to water stress or pathogen stress

Figure 12. Average size of canola flowers for different varieties exposed to water stress or pathogen stress

Figure 13. Nectar quantity for canola flowers for different varieties exposed to water stress or pathogen stress

Figure 14. Nectar sugar concentration for canola flowers of different varieties exposed to water stress or pathogen stress 


Objective 4 - Educate growers on pollinator management in canola

Throughout the project we developed a robust outreach component to connect with producers and other agricultural professionals. One of the graduate students on the project, Rae Olsson, participated in six canola field days in both 2018 and 2020 and 4 in 2020. Postdoc Pfeiffer and graduate student Luppino participated in 3 of these field days and an additional 4 field days in 2021. Olsson, Pfeiffer, and Luppino also led our "stop and chat" series, with specific site-visits to 20 canola growers. These students also developed a social media presence related to research on canola.

We also published the first extension bulletin on bees in canola crops and incorporated our information into the US Canola production guide. See the education sections of this report for more details on our outreach component.

Research conclusions:

Our project found a diverse community of wild bees visit canola, with over 50 total species found during our study. This represents > 50% of the total bee species found in the Palouse. As expected, wild bee communities were more abundant and species-rich in landscapes with more natural habitat; bee communities were also impacted by tillage. Canola is grown in areas surrounded by agriculture throughout much of the northern Palouse and in dryland regions, but is also grown in a landscapes containing natural habitat as you move into Idaho and Montana. The surrounding context of landscapes where canola is grown is an important consideration for growers when determining their need for honey bees.

Overall, we did not find a strong relationship between bee communities and canola yield. This may be due to weak effects of wild bees, or it may be due to the the relative commonness with which honey bees are used on canola. Through our discussions with growers we found that they have strong beliefs that honey bees improve canola yields and most growers deploy hives on their fields. However, growers are increasingly cognizant of measuring wild bees on their farms and are aware of the potential impacts of pesticide sprays and tillage on bee communities. While canola yields seem to be driven more by agronomic practices than by variation in bee communities, our results suggest bees represent a key component of canola ecosystems.

We did find many factors that growers may be able to control that affect the attractiveness of canola to bees. First, managing soil and foliar pathogens appears to help promote growth of canola plants that are attractive to bees. Certain varieties of canola are also considerably more attractive to bees, and in landscapes where growers may wish to benefit from wild bees, these varieties should be selected. Overall, while management of pollinators in canola is not straightforward, our results suggest that pollinators can be conserved within these systems without sacrificing yields or crop quality.

Participation Summary
18 Producers participating in research

Research Outcomes

1 Grant received that built upon this project
12 New working collaborations

Education and Outreach

1 Curricula, factsheets or educational tools
3 Journal articles
15 On-farm demonstrations
2 Published press articles, newsletters
8 Tours
8 Webinars / talks / presentations
8 Workshop field days

Participation Summary:

125 Farmers participated
125 Ag professionals participated
Education and outreach methods and analyses:

In both 2018 and 2019 we had 6 on-farm demonstrations, 4 field days, and 2 farm tours where we showcased different canola varieties and discussed the role of pollinators in canola production. Our outreach was more limited in 2020 and 2021 due to the COVID pandemic, but we offered 4 virtual events in each year, 2 field tours in 2020, and 4 field tours in 2021. We estimate that around 250 individuals attended these events in total.

One of the graduate students on the project, Rae Olsson, produced 3 talks at the WOCS-Canola event and entomological conferences to detail the results of the work. A postdoc involved in the project, Vera Pfeiffer, also presented three talks at academic conferences. The individuals, along with graduate student Mario Luppino, also conducted considerable face-to-face outreach to growers over the course of the project. It is estimated that we conducted nearly 100 site visits over the course of the project, and we met with each farmer at least once or twice per season. These face-to-face interactions proved invaluable in forging connections between our project team and growers while aiding us in identifying current and future research priorities.

20 Farmers intend/plan to change their practice(s)
10 Farmers changed or adopted a practice

Education and Outreach Outcomes

Recommendations for education and outreach:

Pollinators are an important part of canola production, but insect pests may play an even more important role and a yield limiting factor. Unfortunately, many of the key insect pests found in canola (lygus bugs, aphids, cabbage seedpod weevil) are present during bloom when pollinators are also attracted to canola flowers. Increasing sustainability of canola as a viable crop in dryland systems will continue to rely on development of flexible integrated pest and pollinator management programs. Research in the interactions between pollinator and pest management can aid in continuing to diversify landscapes of the Palouse region, and provide an example for other dryland crop regions in the USA.

40 Producers reported gaining knowledge, attitude, skills and/or awareness as a result of the project
Key areas taught:
  • Canola floral traits
  • Bee biology
  • Pollination ecology
Key changes:
  • Farming practices that support healthy pollinator species without harming yields

Information Products

    Any opinions, findings, conclusions, or recommendations expressed in this publication are those of the author(s) and do not necessarily reflect the view of the U.S. Department of Agriculture or SARE.