American Bee Journal • October 2019 Vol. 159 No. 10 • Survey Says … Using Citizen Science to Understand New Hampshire Winter Hive Loss

According to the Bee Informed Partnership (BIP) National Honey Bee Colony Loss Survey, as well as local beekeeper reports, New Hampshire has experienced winter hive loss higher than the national average over the last several years. In order to understand the causes behind these high loss rates, the New Hampshire Beekeepers Association (NHBA) has conducted its own hive loss survey for the last three years running. In 2019, over 400 beekeepers (1400+ hives) with apiaries in over 80% of NH’s communities took part (Figure 1). Since initiating the survey, we have not only experienced a 10% reduction in winter hive loss (Figure 2) but have become much better at conducting surveys and engaging beekeepers. This article illustrates the lessons learned with the findings from our survey.

The key to any survey is participation. Based on feedback from our beekeepers that other surveys are too long, NHBA designed its survey to be completed in under ten minutes. The survey asks both demographic (e.g., apiary location, years of beekeeping experience, number of hives and nucleus colonies [nucs] alive on October 1 and April 1) and management questions (e.g., were hives treated for Varroa destructor? Was feeding done in the fall or winter?). The following techniques helped us reduce the time to complete the survey:

• Create an on-line survey that uses pull-down menus and multiplechoice questions where possible but always add an “other” option so beekeepers can include an answer that is not listed. We use surveymonkey. com to collect responses online; however, we happily accept paper submissions as that is easier for some of our beekeepers.

• Only require answers to the questions that are fundamental to your main goal. Beekeepers may not want to share some of their management practices, but any information they are willing to share is valuable. While our survey only requires answers to the questions about the apiary location and number of hives and nucs alive on October 1 and April 1, we have found that most respondents (85-90%) complete the entire survey.

• Allow beekeepers with multiple apiaries to enter them in one survey session. The first year, we required a person to fill out a different survey for each apiary, which was cumbersome since most of the management data was the same. Recent surveys allow participants to enter up to three apiaries in one session, which has made it much quicker for them to complete.

Participation in our survey has increased each year from 247 beekeepers (1000 hives) to over 400 beekeepers (1400+ hives) in 2019.

Advertise and promote all aspects of your survey. While it is important to get the word out via email, social media, newsletters, and local newspapers that the survey is open to collect data each year, publicizing the results of the survey can be even more important in gaining awareness and participation from the beekeeping community.

When survey results are available, we send a copy to each participant and present the data at local bee club meetings. We also take every opportunity to do interviews with local papers and radio stations about beekeeping, and use data from the surveys to help explain a point or gain awareness. The first year of the survey showed a 65% winter hive loss. When a local newspaper wrote a story about our findings, it grabbed the interest of local beekeepers as well as exposed the public to the challenges facing our NH honey bees. The story was picked up by the Associated Press running nationally. As the articles were published, we were contacted by beekeepers throughout the state asking for more information, giving us additional feedback, and asking to be included in the next survey.

For the last two years, approximately 20% of the survey participants were not members of any bee club. We attribute this high participation rate to the buzz that was created with the various discussions, articles, and radio programs. By getting as many beekeepers as possible to take part, we have a more accurate representation of what is happening in the state.

Often running a contest with a small prize can help increase survey participation, and it can be fun to have a friendly competition among beekeepers. NHBA does two giveaways each year. The first is a free membership for a survey participant who is chosen at random, and the second is a small cash prize awarded to the bee club with the highest percentage of membership participation. Using the participation prize as an incentive, one club was able to significantly increase the number of members who completed the survey, ultimately winning the prize which was used to help fund a summer barbeque.

Once a survey has closed, the next stage is to aggregate the results and analyze the data. Posing questions we would like to answer helps focus the data analysis effort. Since the main objective of our survey is to better understand the high winter loss rates, we would like to ask the question: Why are our hives dying? Unfortunately, because the answer to this question is based on several factors, it is too complex to directly answer with the survey data. Instead, we try to gain insights from direct questions such as: How does the number of varroa mite commercial treatments affect survival rates? How does winter feeding affect survival rates? Why do beekeepers think their hives died?

Compiling survey results takes time and effort. By having multiple choice questions, the analysis becomes easier because you do not have to translate non-uniform information such as those provided in comments. Many on-line tools such as surveymonkey.com provide reports that summarize the results from each survey question, which is a good starting point and requires little work. For more in-depth analysis that requires combining answers from multiple questions (e.g., how does the survival rate differ from those who wrapped their hives and used pollen supplements compared to those who took none of these actions?), you need to use more advanced tools like Microsoft Excel or basic data science practices.

Looking at the individual survey results can be overwhelming; it is very helpful to create a one-page, graphical summary that provides a quick overview of the survey results and key take-aways. Figure 3 shows the snapshot we published summarizing the 2018-19 results.

We all know the beekeeping adage, ask five beekeepers a question and you will get seven answers. While it is very important to get other beekeepers’ opinions and input, presenting objective survey results allows a beekeeper to draw their own conclusions about how this information applies to their situation.

For example, each year, we produce a graph showing the survival rate of hives treated with a commercial varroa mite treatment vs. those that are not (Figure 4). Consistently, the data show that hives that have been treated have a higher survival rate than those that were not.

Comparing the number of hives and apiaries that used commercial varroa mite treatments for each year of the survey (Figure 5), we see that the percentage of treated hives has increased by 20% between 2016-2018, indicating that beekeepers are using the survey results to help make management decisions. We see similar upward trends in the number of commercial varroa mite treatments a hive receives each year, the frequency and method of counting varroa mites, as well as winter feeding.

Our interactive discussions with beekeepers about the survey and its results are one of the most critical components of success. First, it helps us understand if our survey results are representative of the winter hive loss in NH. Any time we discuss the survey results with beekeepers, we ask, does this loss match what you saw in your own apiaries? To date, the feedback has shown that the data we collect is representative of what the NH beekeeping community is experiencing.

Second, we use the feedback to help improve the questions to make them less confusing to the participants. For example, the first year of the survey we asked: How many hives were alive on October 1? Then: How many hives had died by April 1? Participants found this inconsistency confusing. We now ask how many hives were alive on both dates. Another example is the question: What did you use for a winter hive wrap? Since we were interested in all winterization methods, not just wrapping, this question has been changed to: How do you winterize your hive?

Third, we try to gain more insight from the data by answering beekeeper questions. Many beekeepers were interested in the question: How do the management practices of apiaries with high survival rates differ from my practices? In order to provide answers, we divide apiaries into three categories: high survival (66-100%), medium survival (33-65%) and low survival (0-32%), and then calculate the percentage of hives in each group that were managed with a particular practice. The graph in Figure 6 illustrates how varroa mite management practices vary across the three groups in 2017. When the blue bar, representing the high survival group, is higher than the orange (medium survival) and gray (low survival), it indicates that a higher percentage of hives were managed with the practice listed at the bottom. For example, 90% of the hives in the high survival group, 70% of the medium survival group and 56% of the low survival group were treated with a commercial varroa mite treatment.

Another common question beekeepers ask is: Which varroa mite commercial treatment showed the best survival rate? This one is very difficult to answer because the survey collects only apiary (not hive) level data so we do not have all the detail required to give an exact answer. However, we can provide at least some information by calculating the survival rates of apiaries that report using only one type of commercial varroa mite product during the year. Of course, because of the limited sample size, we must be careful in drawing conclusions, but it does provide some insight.

Finally, we ask beekeepers to help us interpret the survey results, especially when the data show something unexpected. Together, we brainstorm about theories and methods for determining if the theory is correct. For example, because the data consistently shows that beekeepers with five or more years of experience have better survival rates than beekeepers with one to four years of experience, we pose the question: Why do you think this is? One hypothesis was that beekeepers with more than five years of experience may use commercial varroa mite treatments more frequently than those with less experience. To test this theory, we looked at the data for the last two years. The 2017-18 data supported the theory, but this year’s data does not. Now, we are coming up with new hypotheses to test.

The survey results can be used to identify areas that require further study or education. We have partnered with the University of New Hampshire (UNH) Cooperative Extension field specialists to use the survey results to secure grants and prioritize efforts geared at improving honey bee health in the state. We created a multi-year NH Healthy Hives initiative to educate and bring awareness to a different management practice each year. The 2018 theme, “Nosema Testing,” was chosen based on data from the first winter loss survey indicating beekeepers might have nosema issues and not realize it. Using a portion of a Sustainable Agriculture Research and Education (SARE) grant, microscopes were purchased. A website was created, and volunteers in each county were trained on nosema diagnosis. Many NH beekeepers have taken advantage of this free program to have their bees tested (https://www.nh-honeybee-health. com/diagnosis). We can see from this year’s survey results that beekeepers are more aware of nosema as a possible cause of hive issues. (See “The Honey Bee Diagnostic Network: Targeting Nosema in New Hampshire,” ABJ, July 2019.)

The 2019 NH healthy hives theme, “Mite Testing” (https://www.nhhoneybee- health.com/mite-testing), was chosen based on indications from the survey that while beekeepers are treating for varroa mites, many are not counting and therefore, may not be treating at the proper time. This citizen science project tracks participating hives for one year (April 2019 to March 2020). More than 100 beekeepers have volunteered for the project. Each performs monthly varroa mite counts on their hives from April thru October and records the results along with basic management information (e.g., was the hive alive? Was it treated for varroa mites? Was there a brood break?). This data will be analyzed at the end of the project to understand if these regular varroa mite checks and resulting management actions improved the survival of the hives.

The data collected in winter hive loss survey and associated citizen science projects has given us insights into our high winter losses and has driven ongoing, proactive discussions about honey bee health by both the beekeepers throughout NH as well as the general public. The fact that we have seen a 10% reduction in winter loss in three years, has made these projects well worth the time and effort. If you are not already running a survey to help understand the state of honey bee health in your area, we invite you to draw on our work to get started. If you are running a survey already, we are happy to share any of the work or analysis we have done and would be happy to collaborate in the future to brainstorm and compare results. The questions and results from each of the three years of our survey are online at our website https://www.nh-honeybee-health. com/winter-loss-survey-results. We are happy to answer any questions about the survey, provide guidance based on our experience, and as always would love to get feedback and ideas on how to improve!

Acknowledgements: Thank you to all the NH beekeepers who take the time to participate in the winter hive loss survey and our other citizen science projects. Also, thank you to George Hamilton and Olivia Saunders, our tireless UNH Cooperative Extension field specialists, who support all of our NH beekeepers by helping write grants, giving excellent advice and guidance, and advocating for all our efforts.

Heather Achilles is a Distinguished Engineer working for IBM Research and has been keeping bees as a hobby for over 10 years in NH. She chairs the NHBA Research Committee which includes leading the NHBA Winter Hive Loss Survey and other citizen science projects that require her analytical and problem-solving background in order to understand and improve the state’s honey bee health issues. Email: heather.nhba@gmail.com