What is small, six-legged, feeds on canola, and moves in unpredictable patterns?
Whilst this may sound like the setup to a punchline, the answer is no joke. Aphids are major pests of canola in Australia, costing millions in crop damage annually.
The green peach aphid (GPA) in particular causes serious economic losses as a major vector of plant viruses such as the turnip yellows virus (TuYV). Compounding the issue, GPA has evolved resistances to many commercially used insecticides.
Using insecticide applications strategically, targeting specific areas and times, is thus advisable over blanket sprays to reduce resistance, pest and virus risk.
But if we want to target our insecticide use, we need reliable information on when and where to spray, to make sure we are targeting high risk regions and periods.
Training, learning and testing
In a study led by Cesar Australia researcher Dr. Alex Slavenko, recently published in the Annals of Applied Biology, we used state-of-the-art machine learning models to predict which environmental conditions were favourable for aphid flights during autumn and early winter – the key period when aphids, particularly GPA, fly into newly sown canola crops.
Machine learning is a powerful tool to create predictive models, but requires a large amount of data to “train” the model, allowing the computer to “learn” the relationships between variables, and to “test” the model by assessing how well it can make predictions on new, untrained data.
We used an extensive dataset drawn from 6 years of surveys across the grain growing regions of Western Australia, Victoria and New South Wales (as mapped below). We collected from more than 200 canola fields and deployed over 800 yellow sticky traps to capture flying aphids.
We determined that aphid flights are associated with patterns of rainfall over the preceding summer, which affect the available “green bridge” where aphid populations build up before crops are sown. Additionally, daily weather conditions also affect the likelihood of aphid flight.
These results enabled us to create general rules-of-thumb about when to expect greater aphid numbers. Interestingly, we saw slightly different patterns in Western Australia compared with Victoria and New South Wales.
In both regions, conditions that were warm, humid, with low winds following rainy summers were found to be conducive to aphid flights in autumn and early winter.
In Western Australia, aphid numbers continue to increase and peak later in the season compared to Victoria and New South Wales, where aphid numbers are predicted to drop as winter conditions set in.
While these findings should not and cannot replace direct field monitoring, they offer valuable insight that may prove useful in planning effective monitoring programs in canola fields. This is important, as we also found that if a single aphid is captured in a sticky trap, there is a 1 in 4 chance it carries TuYV- a probability that rises with increasing aphid counts.
So, I’ll ask again, what is small, six-legged, feeds on canola, and moves in unpredictable patterns? Thanks to this research, not an aphid anymore…!
Acknowledgements
Thanks to Samantha Ward and Paul Umina (Cesar) for reviewing the article.
This research was conducted by Alex Slavenko, Marielle Babineau, Anthony R. van Rooyen, Benjamin Congdon, Paul A. Umina and Samantha Ward. This research was undertaken as part of a Grains Research and Development Corporation investment, ‘Insecticide resistance in the green peach aphid: national surveillance, preparedness and implications for virus management (CES2001-001RTX).
The full study can be accessed at https://doi.org/10.1111/aab.12911.
For more information, contact Dr. Alex Slavenko at aslavenko@cesaraustralia.com
Cover image: Photo by Marielle Babineau, Cesar Australia
