The first Blooms for Bees journal paper ‘Evaluating the ability of citizen scientists to identify bumblebee (Bombus) species’ has been published in PLOS One this week. You can read the full paper on the PLOS One website.
In this paper, we use the 4,200 bumblebee sightings submitted by participants during 2017, to explore the ability of citizen scientists to correctly identify bumblebee species. Looking at the data in this way was important because the quality and accuracy of citizen science project data is poorly understood and is often of concern to scientists and policy makers.
We have been able to demonstrate that citizen science projects can generate high quality data, even for a group of insects that is challenging to identify, and our results highlight the essential role of expert verification.
Although we found that recorder identification ability was generally low, we did find that recorders’ skills improved over time, probably as a result of experience and the verification feedback provided. This suggests that submitting records is a great way to develop your bumblebee identification skills.
Unsurprisingly, recorders’ identification ability varied depending on the bumblebee species in question. Recorders were most able to correctly identify the common species with distinct appearances, namely the Tree Bumblebee, Common Carder Bee, Red-tailed Bumblebee and Early Bumblebee.
Identifying the Garden Bumblebee proved particularly challenging for recorders.
The yellow and black banded species proved more challenging and were often confused with each other. In the case of the Garden Bumblebee, only 5% of the 474 records submitted as such were confirmed as correct by our expert verifier. The majority of the records (70%) were reassigned to the similar-looking Buff-tailed and White-tailed Bumblebees.
We compared the Blooms for Bees data to that generated by the Bumblebee Conservation Trust’s BeeWatch project. We found that BeeWatch participants were slightly more able to identify bumblebees, with 49% of the verifiable records being correctly identified by recorders, compared to 44% in Blooms for Bees. Considering that many of the Blooms for Bees participants had no previous experience of bumblebee identification, we think this is a great result as bumblebees are notoriously challenging to identify.
It was not possible to verify 20% (833 records) of the bumblebee records submitted to Blooms for Bees because supporting photographs were either missing or of poor quality. We would therefore like to encourage recorders to submit photographs wherever possible, even if the images are not perfect, as in some cases we were able to make positive identifications even from blurry photographs.
A blurry Early Bumblebee captured in flight.
The data collected by Blooms for Bees participants has helped us identify where the greatest misidentification issues occurred and has enable us to make recommendations for the development of high-quality training resources. We hope this will help support recorders in future bumblebee monitoring citizen science projects and contribute to improved recorder ability.
We are grateful to everyone who has submitted data to the project over that last two years. A further paper which focusses on gardening recommendations is currently in development and we look forward to sharing our findings.
Full citation: Falk S, Foster G, Comont R, Conroy J, Bostock H, Salisbury A, Kilbey D, Bennett J & Smith, B. (2019) Evaluating the ability of citizen scientists to identify bumblebee (Bombus) species. PLoS ONE 14(6): e0218614. https://doi.org/10.1371/journal.pone.0218614