Locusts and robots
by Tim Groenendyk
In an effort to develop a system that anticipates the movement of destructive locust swarms, Stephen Simpson has armed himself with robotic planes, super computers and his second ARC Linkage grant.
A solitary locust, understandably, does not pose much of a threat to mankind. However, when locusts cluster together they change colour, and demeanour. They form destructive ‘marching bands’ in juveniles, or ‘flying swarms’ in adults, devastating crops.
“If you increase the population density just a little bit in a given area you’ll reach a critical threshold, and almost like a disease running through a population, the gregariousness, which is their swarming form, will spread very rapidly,” said Simpson.
“The cohesive mass movement of a group of locusts and the speed and direction in which they go emerge as a result of individual locusts responding to their nearest neighbours and aligning with their neighbours, even though each individual has no idea how big the swarm is or where it’s going."
Using mathematical models, Jerome Buhl, a member of Simpson’s team, can simulate a swarm of a million or more individual locusts and explore what the interactions might look like.
“We’ve done that by making our own CUDA supercomputer for $3000 by wiring up in parallel a series of graphics cards from gaming machines," said Simpson.
To prove that their computer models demonstrated what occurs in the field, Simpson and biological sciences research colleague Greg Sword collaborated with field robotics expert Salah Sukkarieh and his group.
Because Australian plague locusts are too small to radio collar, as Simpson and Sword had previously done on Mormon crickets, the interdisciplinary team had to innovate an individual locust tracking device.
“Salah’s team’s solution was to mount a small crystal on the back of individual locusts and release them into the band,” explained Simpson.
“Then a fixed-wing unmanned aerial vehicle, a robotic plane, mounted with a powerful stroboscope and a detector, catches the reflection coming back from the crystal on the back of the target individual."
The robot can then follow an individual, and produce data showing where the locust has moved, how it has moved in relation to other tagged locusts, and how quickly.
Simpson recognises that the Australian Plague Locust Commission have been invaluable to this research effort.
“They have an unprecedented knowledge of where locusts are,” he said.
“But they also know the farmers, and can organise access for us to particular properties. And they’ve provided us with logistical support by way of aerial surveys and ground support."
These routine surveys can locate bands of locusts across vast regions.
“They also have a fantastic database from their previous surveying efforts and we’ve used that to add value to our projects on numerous occasions.”
Talk of another partnership arose during the course of this research project. Last November, during field research, Simpson and team observed a mass locust die off caused by an unknown disease.
“We’re able now to start to look at alternative bio-control agents, naturally occurring pathogens, which under certain circumstances maybe very helpful as controls. The Commission are committed to such an application."
The aim of all this work is not to cull the entire population of locusts but to strike a balance between protecting economical interests – farms, crops and pastures – and acknowledging locusts integral role in the Australian food chain.
“Locusts are relied upon on as food pulses by many of the native insectivorous animals,” explained Simpson.
“But if you can just knock the population size down by a relatively small percent you may be able to push it below that threshold of swarm forming gregariousness. It’s just about being more environmentally clever in the way you deal with them.”