For the second consecutive year, we joined forces with a class at Indiana University’s O’Neill School of Public and Environmental Affairs to host a “Bee Data Hack-a-thon”. We tasked Professor Roger Morris’s class of graduate and undergraduate students with determining whether a relationship exists between yield and bee flight time in a given season. Armed with two sets of data—USDA yield data and NOAA weather data—the students were given 72 hours to reach a conclusion.
The Challenge: Determine if relationship exists between bee flight time and almond yield.
The Resources: USDA yield data, NOAA weather data.
1. No teams were able to find a relationship between bee flight time and yield
Although the datasets provided to students were admittedly limited—yield and weather numbers were aggregated by county—we figured there would be a rough trend between bee flight time and yield. As it turns out, more “bee flight” days during pollination doesn’t lead to increased yields at the end of the season.
2. Further analysis (with more granular data) is necessary
In the interest of keeping things simple, we gave students a very basic equation to calculate “bee flight days”:
Bee flight days = # days [temp. > 55℉] AND [wind < 10 MPH]
In other words, if temperature reached over 55 degrees and the wind was below 10 MPH, at any point during a day, that would count as one bee flight day. This meant that days that were sunny and 70 degrees were counted the same as days that were chilly and rainy.
Additionally, no consideration was given to hive strength or hive stocking rate per acre. The importance of strong hives has been well-documented by researchers, as colonies with higher populations will send more bees to forage.
Putting aside our concerns with the accuracy of the data, it was startling to see that no team found that yields are higher when there’s better weather during pollination season. Lots of growers will tell you they stock 2+ hives per acre as insurance against poor weather during pollination. But these findings suggest that bees get the job done no matter the weather.
Though we're not the first to suggest that growers are stocking too many hives per acre, we also understand that these results don't account for many factors that impact yield. We’d like to repeat this study on a smaller level, where more precise data can be captured on the number and distribution of pollinating bees, as well as weather and yield at the orchard level.
Thanks to Professor Morris and his students for taking part in another successful Bee Data Hackathon!