Surveying and Reporting Bee Losses

By

William Blomstedt

One of the main problems the bee industry confronted when Colony Collapse Disorder struck a decade ago was the lack of good statistics on bee deaths, both present and historical. It is normal a certain number of honey bee colonies die every year; the acceptable loss rate could be between 5% and 20%, depending on who you ask. Certain eras have seen much higher loss rates due to poor weather, the introduction of new pests or diseases, or something unknown. These loss rates have been mostly anecdotal, and there have been very few, if any, extensive surveys performed. Today the beekeeping industry is at the point where it has the technology and money to better understand this problem, so many smart and dedicated people have moved towards it.

While statistics on colonies are gravely needed, the biology and life cycle of honey bees as kept by beekeepers present many difficulties. A beekeeper can count the number of hives on a given day, but when time starts to move the complications multiply. Hives swarm, splits are made, packages are purchased. If a beekeeper comes out of winter with a hundred hives, splits 50 in the spring, but then 40 of his hives die in the fall, has he lost 40 hives? Or gained 10? Skill and medical treatments also have a large effect. A more adept beekeeper might have kept all of the bees alive. A less skilled one might have killed 90% of them.

Another difficulty is the variation of climes in which beekeeping takes place. Honey bees face a very different life if they have a Mediterranean winter vs. a Scandinavian one, and how can a “winter loss” be compared when a Mediterranean bee is in the midst of a spring honey flow while the Scandinavian one is still buried under a foot of snow? Or later, when the Southern bee is baking in a summer dearth and the Swede is living the high life among the acres of clover? In America’s case, add to this the migratory practices of many commercial beekeepers who move their bees around the great continental clock, and the number of variables seem to overwhelm any possibility of level-headed analysis.

There’s more: the rather large question of data collection. These surveys totally rely on beekeepers voluntarily taking the time to reply, as well as their accuracy of record-keeping or memory. They also depend on the skill and knowledge of the beekeeper, if the questioning goes into the realm of why a hive died, which is actually a very difficult thing to answer. Only a fraction of beekeepers fill out the survey, and this is not likely a cross-section of beekeepers as a whole: my reasoning would say they are less likely to be commercial beekeepers. Many critics cite this as the reason why the recent loss numbers seem so high.

It is also somewhat unfortunate that so many groups are trying to gather survey data. Among beekeepers there is a general weariness from the sheaf of surveys that appear on their doorstep each year: from local bee clubs to governmental surveys and every type of organization in between. Beekeepers, even if they have a desire to help, can only answer so many forms without getting much tangible in return. Also, though each survey may ask for the number of hives living and dead of each year, the distributors may also have different aims, methods and wording, so that they produce numbers incomparable to each other. This adds noise to the channel, making it difficult to understand the true picture of bee health.

With all the downsides mentioned above, the task of simply counting the deaths of honey bees seems daunting. But statisticians are well aware of all the weaknesses mentioned above and have their ways of taking their account in analysis. And here is the crux: we need this data. Anecdotes may point us in the direction of the problem, but they essentially mean nothing. To make rational policy and management decisions, we need to know if bees are really dying in great numbers (or at least more than previously), and what practices and locations are associated with a high death rate.

Once the data is collected, what do we do with it? At the very least it needs to be analyzed and acted upon. I argue it also must be visualized and distributed to the public, for it is the public beekeeper who spent his time filling in the form and also (often) the one who’s tax dollars helped pay for the survey. While data cannot be shown at certain levels of granularity due to the possible jeopardizing of anonymity, having open data should also be an important aspect of these surveys. Open data promotes transparency, cooperation, innovation.

Also, as a cartographer, I argue that well-visualized data can also have a greater impact than reading numbers off a sheet. Akin to statistics, which is able to tease out interesting conclusions from a pile of numbers, sometimes showing the data on a map is able to highlight a geographical trend not easily seen in numbers alone.

Two organizations that are working on this problem are Bee Informed, in America, and COLOSS, which is worldwide but more focused on Europe. COLOSS is a non-profit that gathers researchers, students and extension workers with the shared goal of understanding problems honey bees face as well as promoting bee health. COLOSS has more than 800 members from 94 countries. Their Colony Loss Survey, which began in 2008, was one of the seminal parts of the organization. Though any country is welcome to take part in the survey, the frequent contributors are from Europe and the Northern Hemisphere. More recently COLOSS undertook the giant task of putting together the BEEBOOK, a collection of 1700+ standard methodologies that has greatly unified bee research efforts around the world. This includes the methodologies for collecting bee losses, allowing countries to compare data.

I have been participating in COLOSS for the past few years and I received a grant from the Eva Crane Trust to build a web map with their data. The link is below.

http://wblomst.com/COLOSS/

Both USA and Europe are working hard on this problem. It is unfortunate that there are so many surveys, and I implore these organizations to reduce their overlapping efforts and work together. While I have heard grumblings about how these studies have been going on for a long time and nothing “useful” has come from them, it must be remembered that they take time. The average medical study lasts 10 years before publication. Part of the process is getting a better understanding of the problem, and these surveyors constantly look at the data they have and refine the questions in order to find the right angle of attack. Each year they are getting better. So although this paperwork sometimes seems like a hassle they are worth filling out.