Apiaries in the German Bee Monitoring show surprising gains over the winter
Thursday 07 April 2011
The scientific journal Apidologie published a paper in 2010 titled The German bee monitoring project: a long term study to understand periodically high winter losses of honey bee colonies. Pesticide manufacturers, their lobby organizations and government officials are showering this work with praise. Wherever in the world beekeepers raise concern about the impact of certain insecticides on honey bee colonies, they are being told, that a long term study by very capable German scientists have found no evidence of such a connection.
Two independent german scientists, Dr. Hoppe and Dr. Safer, published a critical and thorough review of the data collected and statistical methods used to reach such conclusions and documented a number of very serious flaws.
One of many problems with the paper is the fact, that it seems to suggest, that a number of apiaries managed to significantly gain in strength over the winter.
In the paper, the authors claim There was no significant difference in overwintering quotient between apiaries with no pesticide residues in the bee bread and those with higher amounts of residues. They also state, that the hypothesis could not be verified that intensive contact of honey bee colonies to oilseed rape has a negative influence on overwintering.
In support of this finding, we find a chart labeled Figure 6 showing the Relation between amounts of rape pollen in honey harvested in summer 2006 and the overwintering coefficient of the colonies in the subsequent winter 2006/2007.
What is interesting about this chart is the fact, that there are a number of data points showing an overwintering coefficient greater than one. Each data point represents not a single colony, but presumably a group of ten monitoring hives. Some of these groups must have been able to gain significantly in strength in winter. Upon further review of the circumstances surrounding the German Bee Monitoring there may be some explanations for this surprising result.
1. The data may have been entered incorrectly or misallocated.
The paper was originally submitted on the 18th of November 2009, revised and resubmitted on the 17th January 2010 and accepted for publication on January 30th 2010. Independent scientist, Dr. Hoppe, did visit one of the authors, Prof. Kaatz, on July 19th 2010 at the beekeeping institute in Hohenheim to take a look at the database of the bee monitoring project. It turned out to be difficult to get answers to fairly basic questions. Shortly thereafter, on August 2nd 2010, an email was sent to all participating experts of the project. A deadline was set for October 1st 2010 to reenter a significant part of the data. It turned out that from 2005 until mid 2010 serious errors in the database remained unnoticed. Therefore the condition of the database as of the time of the publication is highly suspect.
2. The method for estimating the strength of the colonies may have been too imprecise. The strength may have been underestimated in the fall or overestimated in the following spring.
The standard method for estimating strength of a colony was developed by the Swiss Agroscope-Liebefeld-Posieux Research Station ALP and published in Apidologie in 1987. It involves counting the number of bees on a dm2 of occupied honeycomb surface. The method used by the authors of the German Bee Monitoring is known to be error prone and has to our knowledge never been validated. The authors describe their method as follows: In general the following procedure was used: all hives were opened and from two-story hives the upper magazines were tilt forwards. By doing so, all spaces between the combs could be inspected and the numbers of combs covered by bees were recorded.
When the Liebefeld method was published in 1987, the Swiss scientists had this to say about the now obsolete method estimating colony strength.
The commonly used method for estimating colony strength by counting frames or spaces between frames occupied by bees is imprecise and results, in our experience, in considerable errors in judgement.
Among the well known problems associated with this outdated method is the fact that the same number of bees will occupy a different number of frames depending on the outside temperature, which unfortunately was not recorded. The poor choice of method may explain at least part of the surprising results.
3. The timing of the estimates, especially coming out of the winter is defined vaguely and contradictory. In the publication about the German Bee Monitoring :
To avoid overestimating the population size of the overwintered colony the population estimation had to be performed prior to the emergence of the first spring brood. Therefore, the last accepted period for measuring the starting population was the 15th week of the year.
On the one hand, the estimate should take place prior to the emergence of the first spring brood, on the other hand the cut off date was the middle of April, when in some regions in Germany the first swarms may hang in the trees. Such ill-defined timing cannot be expected to produce meaningful data.
Even the definition of the overwintering quotient provided in the paper itself would suggest, that gaining of colonies over the winter should not have been plausible to the authors:
The quotient of the population size before and after the wintering of the colonies were calculated as “overwintering quotient” and represented a mea- sure of the weakening of the colonies over winter.
Had something other than weakening be expected by the authors, the term "changing“ or the phrase "weakening or strengthening“ would have been appropriate in the definition of the quotient.
4. The actual number of bees may have been estimated correctly, but bees could have been added from other colonies that were not part of the project.
This would be a serious flaw in the methodology. The publication speaks to training of the participants. We have not seen project wide training materials containing the precise instructions to the beekeepers.
Of course, there are a number of scenarios, that could explain the strengthening in a few individual cases. Collapsing colonies in the vicinity may have been a source of additional bees. Hives at the ends of a row of colonies often profit from disoriented bees. But it would be far fetched to assume this would happen to whole apiaries.
The legend for Figure 6 unfortunately does not tell us precisely whatʻs behind each dot. Given the number of colonies in the project, it is fair to assume, that the groups of 10 hives of the project are shown, making each dot with an overwintering coefficient greater than one quite implausible.
At least part of the German Bee Monitoring seems to have fallen victim to the old rule of data processing: "Garbage in - garbage out".
Applying sophisticated statistical analysis does not make the underlying data any better.
To avoid any misunderstandings - the problems identified here are not the fault of the participating beekeepers or extension experts, who participated in good faith and relied on the scientists to properly structure the project. The overwintering coefficient is a good example to illustrate the shortcomings of the project, which are in sharp contrast to the claims of scientific excellence. The analysis of Dr. Hoppe and Dr. Safer (German Bee Monitoring - Claims and reality) provides further evidence, that the substance of the project may not support the wide ranging conclusions.