Since first being detected in Brisbane, Queensland, in 2001, red imported fire ants (Solenopsis invicta) have shown themselves to be an extremely damaging invasive pest, affecting agricultural crops, native species and human health (in the form of painful stings and potentially fatal allergic reactions).
Attempts to eradicate this biological invader have to date been unsuccessful, despite concerted efforts. In this week’s online version of Proceedings of the National Academy of Sciences (PNAS), we report the results of a new analysis in which we reconstructed the historical trajectory of the Brisbane fire ant invasion to shed light on the effectiveness of the past eradication strategy – with a couple of surprising results.
Meet the enemy
Red imported fire ants are native to South America, and may have been introduced to Australia as early as 1990. They live in colonies consisting of workers, alates and queens:
Workers are sterile females that carry out a range of tasks, including foraging. They are small (2-6 mm), reddish-brown in colour, and generally live for only a few months.
Alates are winged males and females that leave the colony to mate in flight. Males die after mating, but females shed their wings and become new queens, founding colonies underground.
Queens are capable of laying up to 800 eggs a day, and can live for up to seven years.
In Australia, the ants were found at two well-separated locations on consecutive days, with samples confirmed to be fire ants by the Queensland Museum and the Queensland Department of Primary Industries.
The two incursions – one centred on Brisbane’s main cargo port at Fisherman Islands, and a larger one centred on the southwestern suburbs of Wacol and Richlands – were later shown to be genetically distinct, indicating two separate introductions.
The Queensland Department of Primary Industries immediately mounted an emergency response to delineate the invasion and if possible eradicate the fire ant, despite scepticism that eradication could be achieved, given no other country has been able to eradicate them.
The National Red Imported Fire Ant Eradication Program, launched in September 2001 by Biosecurity Queensland, initially appeared to be highly successful.
But, although the smaller incursion at Port of Brisbane was declared successfully eradicated, the larger incursion in Brisbane’s southwest continues more than 12 years after the program began.
Indeed, although more than $A275 million has been spent on the eradication program thus far, the known infested region has doubled in area since 2004.
What we did
For our PNAS analysis, we wanted to evaluate whether the eradication program is on track to achieve eradication, and to learn lessons that will inform future management decisions. The results were surprising and frustrating in that an opportunity for eradication seems to have been narrowly missed. They also highlight the crucial importance of mathematical modelling of biological invasions.
The plots below show our estimates of the month-by-month trajectory of the fire ant invasion during 1996-2011, in terms of the number of reproductively mature nests and their geographic range. They illustrate just how close the program came to successful eradication, with the total number of reproductively mature nests dipping below 150 around the end of 2003.
The sequence of maps shows the changing density of ant nests in infested areas, which declined and recovered in a similar manner to the total number of nests.
A comparison of these density maps with the areas searched and treated (using baits toxic to the ant) in each year shows that the eradication program never fully delimited the invasion.
There was always an infested area outside the searched and treated areas. This is not to be critical of the program: delimiting the invasion was recognised as crucial from its inception and extensive research was conducted to develop surveillance tools for that purpose. But inferring the boundary is a difficult problem.
If accurate estimates of the invasion boundary had been available, resources could have been reallocated to fully cover the infested region. This could have been done at essentially no extra cost by reducing the number of follow-up treatments in formerly infested areas, as we discuss below.
What we found
Perhaps the most surprising finding was that the southern and western boundaries of the invasion advanced at a steady rate, seemingly unaffected by the sharp drop in the number of reproductively mature nests, as mentioned above.
So the eradication program appeared highly successful in that fewer nests were being reported, but this masked an expansion of geographic range. Such discrepancies may be a general feature of eradication programs, and policy makers need to expect an infestation to continue expanding its range despite declining numbers.
The rate of movement of invasion boundaries should be estimated, and the eradication strategy should be designed to keep pace with the expansion in a preemptive, rather than responsive manner. This demands sophisticated mathematical modelling.
Each month, models should infer areas likely to be infested, even if no nests have yet been found there. These areas can then be treated with baits. We believe our new model provides unprecedented spatial and temporal resolution for this purpose.
Another unexpected finding was that immature nests (those not yet capable of founding new nests) outnumbered mature nests at almost every stage of the invasion. Destroying immature nests is thus at least as important as destroying mature nests. But many of these immature nests are too small to detect and their presence can only be inferred.
One consequence for policy makers is that monitoring and treatment of recently infested areas should continue for at least eight months (the approximate time between founding of a nest by a new queen and the first mating flight by alates) after nests have apparently been eradicated from those areas.
(Until 2006 the practice was to perform three or four repeat bait treatments per year for three years, which may be overkill, given that each treatment is estimated to kill at least 80% of nests. The program now considers five treatments administered over 18 months to be sufficient.)
Newly infested areas contain only immature nests, so the total area infested is larger than the area containing mature nests. So, again, sophisticated models are needed to infer how far immature nests have spread, and the eradication strategy should include a plan for dealing with those invisible advance invaders.
What comes next?
Is the eradication program on track to success? We found that numbers of red imported fire ants were again in decline throughout 2011. But by December of that year – the end of the period we examined – the infested region was larger than ever, with the invasion boundaries still advancing to the south towards the Gold Coast and Scenic Rim and west into Lockyer Valley.
To put it bluntly, at the current rate of expansion only a scant few years remain before some of Queensland’s most popular tourist destinations and most fertile farmlands become infested.
Unless, of course, we can avert this.