Determining The Relative Value of Hives for Almond Pollination
Contents
A Revisit of Sheesley and Poduska. 1
Sheesley and Poduska’s Data. 2
Determining the Relative Value of Hives for Almond Pollination
First Published in ABJ October 2018
Randy Oliver
ScientificBeekeeping.com
As we approach time to line up almond contracts, we should be setting prices based upon hard data. Growers pay for fertilizer by the pound, water by the acre-foot, and for labor by the piece rate. So why the heck don’t they pay for pollination services in the same manner? In today’s almond pollination market, the best beekeepers are not being paid for what their bees are truly worth, and bargain-hunting growers are getting screwed.
Half my annual income comes from renting bees for almond pollination—I’ve been doing it for nearly 35 years. These past couple of years, we’ve finally broken through the “glass ceiling” of $200 for the rental of strong hives. This is because the supply of bees has barely kept up with the continually-growing demand for hives to pollinate California’s now more than 1 million acres of almond orchards. It’s not that beekeepers are gouging the growers—it’s just become more expensive to consistently produce strong hives by February 15th.
But not all growers are willing to pay $200, and some beekeepers (including myself) will take less for weaker colonies. The question then is, what are the relative productivities of colonies of various strengths as far as the pollination service that they provide?
The best agricultural management decisions are based upon sound interpretation of scientific data. Yet I again and again find almond growers basing their decisions upon advice derived from less-than-rigorous interpretation of a single landmark study done in 1970—that of Sheesley and Poduska [[1]]. For example, a 2017 publication from the University of California [[2]] stated:
In a 1970 study on colony size and pollen collection, the 8-frame colony did three times as much pollination work as the 4-frame colony. Beyond about 12 frames of bees, increased colony size does not seem to increase foraging.
Is it really true that an 8-framer is three times better than a 4-framer? Or that growers are wasting their money on colonies of greater than 12-frame strength? To see, we’ll need to perform…
A Revisit of Sheesley and Poduska
Sheesley performed a meticulous and labor-intensive study. The paper is open access, and is a worthwhile and easy read that gives you an idea of the state of almond pollination 48 years ago.
Practical application: it’s disappointing to see how little progress we’ve made in applying the authors’ excellent recommendations to almond pollination contracts.
Briefly, they performed the study shortly after the California Beekeepers Association changed their minimum colony strength standard for almond pollination to 4 frames covered with bees. Their study also spanned one year in which colonies were relatively weak overall, followed by a season in which the hives were much stronger. I feel that it’s worthwhile to show just how much colony strength could vary from year to year then, as it does now, so I’ve copied their results below:
Colonies were much stronger for almond pollination during 1970 than in 1969. A combination of three factors can be credited for this difference. Many beekeepers fed their colonies during the fall and late winter of 1969 to build colony populations for almond pollination. The late spring rains of 1969 in Fresno County provided more than the normal weed and range plant pollen and nectar sources. This allowed beekeepers to enter the 1969 winter with much stronger colonies than in the previous year. An awareness of the minimum colony strength standards adopted in 1968 in California may have encouraged the combining or culling of some colonies prior to the 1970 almond bloom period.
Practical application: beekeepers will provide what the growers demand and are willing to pay for.
The table above indicated that beekeepers could indeed supply strong hives in mid winter nearly half a century ago—but that was before varroa, and when there were only 1/7th as many bearing acres of almonds as there are today [[3]]. Strong colonies today don’t come cheap, and almond pollinators are forced to put more and more money into mite control, transportation, and feeding, as well as to continue to combine or cull weak colonies in order to meet the demand for 8-frame or better hives.
Sheesley and Poduska’s Data
The researchers used the amount of pollen collected by a hive as a proxy for pollination services rendered, and graded each hive at least four times in order to assign it an accurate strength, as determined by the number of frames covered with bees (standard California almond grading). I’ve retyped their data below:
Note the two highlighted outlier values, based upon data from only 3 and 5 hives. Also note the far greater amount of data collected from hives of 8-frame strength or less, indicating that interpretation of data for that strength range will be more robust than for stronger hives. The human eye is not very good at interpreting a bunch of numbers in a table—we do a much better job with recognizing patterns when numbers are presented in graphical form, as in the chart below:
Figure 1. Although a column graph might actually be more appropriate for presenting this data, a line graph helps the eye to pick out the relationship between colony strength and amount of pollen gathered. Sheesley’s data was collected during three different time periods and conditions, and each data point is the mean of different numbers of hives. In 1969, overall colony strength was weaker—perhaps contributing to greater pollen collection due to less competition. The circled outlier values really show up, but they were taken from only a few hives. It’s outliers such as these that add to the difficulty of interpreting a data set.
In order to account for the seasonal differences, I adjusted the three data sets by dividing each set by the total amount of pollen collected the 4 – 11-frame colonies in that group. This allowed me to group all the data into one set for better analysis:
Figure 2. Seasonal adjustment makes the data less messy, and allows for generating trendlines. And that’s what we’re interested in—whether there is a predictable relationship between colony strength and the amount of pollination service performed.
Unfortunately, the adjusted data shown above still over-represents those outlier values from a relatively few 12- and 13-frame colonies. The data in Table 1 are the means of measurements from the indicated number of hives–thus we don’t know the degree of variation in the original data. As colony strengths increase, those means are calculated from fewer measurements, so there is more variability in them. Any beekeeper knows that there is tremendous hive-to-hive variation in performance; what I’m trying to do is to tease meaningful information from this data set.
Practical note: the more that an author “works” the data, and the more convoluted the statistics, the more leery I generally am of the interpretation. That’s why I’m walking you step by step through my analysis of this important data set.
So I reworked the data by weighting it—by plotting all 355 datums individually in the scattergram below. Although you can’t tell by looking at it, each point may represent up to 28 values [[4]]. By doing this, I could then have Excel calculate weighted regressions, which gave less weight to the few outlier values.
Figure 3. Now we’re getting somewhere! Note in the above scattergram that the relationship between colony strength and performance is essentially linear up through 16-frame strength. That said, I used the even better-fitting polynomial curve for further analysis, since it accounts for the changing brood-to-adult bee ratio over the range of colony strengths. Keep in mind that the data for colonies up to 8-frame strength is by far more robust than that for colonies of greater strength.
Note in the above chart the very high R2 values for both the linear and polynomial regression curves—indicating that the direct, and nearly linear, correlation between colony strength and pollen collection is quite strong (the higher the R2 value, the stronger the correlation, with a value of 1 indicating a perfect match). For general purposes, let’s look at the linear regression, which fits the data closely overall, as well as reflecting other data that I’ve seen [[5]]. In the table below, I’ve used its formula to show how easy it is, for practical purposes, to compare a hive’s pollination value relative to its frame strength:
In the table above, compare the amount of pollen collected by a 4-frame colony as compared to that by 8-, 12-, and 16-framers. I calculated the equivalent stocking rates for use in drawing up pollination contracts.
Practical application: my analysis of Sheesley and Poduska’s data suggests that it has been widely misinterpreted. It certainly appears that there is a direct and nearly linear relationship between the number of bees in the hive and the amount of pollen collected, regardless of colony strength. Thus, an 8-framer provides roughly twice the pollination service of a 4-framer, and a 16-frame colony does as much work as four 4-framers. California’s favorite pollination broker, Joe Traynor, has long argued that one strong colony per acre may well provide adequate pollination services under most conditions.
OK, I finally got that off my chest! However, since the highest brood-to-adult bee ratio occurs in the 6 – 10-frame strength range [[6]], it would be biologically expected that colonies of that size would have the highest demand for pollen by both the nurse bees and the newly-emerging workers and drones, and thus be more efficient pollinators on a per-bee basis. This is reflected by the polynomial regression equation, which I then used for all following calculations.
Practical application: what both beekeepers and growers are most interested in are the relative dollar values for colonies of various strengths at providing pollination services. So let’s do the math.
As explained by Sheesley, almond growers may overpay for services when they rent weak colonies, as well as underpay when they rent very strong hives. So I plotted the dollar value of expected pollination service performed (as measured by amount of pollen collected) relative to colony strength, using the benchmark figure of $180 for 8-frame hives [[7]]:
Figure 4. Assuming a benchmark rental rate of $180 for a colony of 8-frame strength, we can calculate the expected performance valuation for colonies of other strengths (shown as the dollar figures above each colony strength). It’s easy to see that growers these days are typically overpaying for 4 – 6-frame hives, and grossly underpaying those of us who provide colonies in the 12 – 16-frame range.
Practical application for almond growers: yes, it would be equally cost-effective for you to pay a beekeeper $240 per hive for 11-frame colonies stocked at 1.5 hives per acre compared to paying $180 for 8-framers stocked at 2 hives per acre. And the $240 for those 11-framers would be a much better deal than that “bargain” you got by paying $140 for 6-framers.
Sheesley’s Conclusion
Sheesley and Poduska didn’t elaborate on the interpretation of their findings to any great degree—the oft-repeated extrapolations appear to have grown with repetition, rather than being based upon detailed analysis. But way back in 1970, the authors were quite clear on one thing—that pollination fees should be based upon colony strength, rather than amount of woodenware delivered to the orchard:
Beekeepers who provide strong pollinating colonies need to be paid for the additional management expenses involved. Almond growers and beekeepers both can profit by adopting a multiple rental price structure for almond pollination based upon colony strength. It seems logical that written almond pollination agreements in the future should financially encourage the use of strong honeybee colonies.
Those words were written nearly 50 years ago! Why the heck doesn’t every grower pay on a minimum agreed-upon frame strength—I’ve filled such contracts for 35 years, but for some reason they haven’t caught hold in general. Those contracts call for a target stocking rate of frames per acre, and pay the beekeeper by the graded average frame strength actually delivered, with a maximum strength top out so that the grower doesn’t break his budget. Such a contract ensures that the grower gets exactly what he pays for, and that the beekeeper gets rewarded for the cost of preparing his hives prior to delivery. It’s a win-win for both players.
Take-home message to growers: for pollination services, pay for, and stock the orchard, by number of frames of bees per acre, rather than number of hives per acre. You and the beekeeper can both benefit from pollination contracts that make it worthwhile for the beekeeper to provide strong, healthy hives that will provide the best pollination services. And you can actually SAVE MONEY by paying for very strong hives, and stocking them at a lower number per acre.
Citations and Notes
My apologies for misspelling Poduska in the ABJ version of this article.
[1] Sheesley, B & B Poduska (1970) Strong honeybee colonies prove value in almond pollination. California Agriculture 24(8): 5-6. http://calag.ucanr.edu/archive/?type=pdf&article=ca.v024n08p5
[2] Connell, J (2017) Honeybees, Colony Strength, and Beekeeper Challenges. University of California Agriculture and Natural Resources http://www.sacvalleyorchards.com/almonds/pollination/honeybees-colony-strength-and-beekeeper-challenges/
[3] Johnson, DC (1987) Fruits and Nuts, Bearing Acreage, 1947-83. National Agricultural Statistics Service, U.S. Department of Agriculture. Statistical Bulletin No. 761.
CFDA (2018) 2017 California Almond Acreage Report https://www.nass.usda.gov/Statistics_by_State/California/Publications/Specialty_and_Other_Releases/Almond/Acreage/201804almac.pdf
[4] Each value represents the mean pollen collection for all colonies in that test group that were of the same frame strength.
[5] Oliver, R (2012) 2012 Almond Pollination Update. ABJ April 2012 https://scientificbeekeeping.com/2012-almond-pollination-update/
[6] See Fig. 3 in https://scientificbeekeeping.com/understanding-colony-buildup-and-decline-part-4/
[7] Goodrich, B & RE Goodhue (2016) Honey bee colony strength in the California almond pollination market. ARE Update 19(4): 5-8. https://s.giannini.ucop.edu/uploads/giannini_public/4c/be/4cbee741-f4aa-4f11-96bf-d8a84681051b/v19n4_2.pdf