Between 2001 and 2015, the National Red Imported Fire Ant Eradication Program, managed by Biosecurity Queensland, wasted $400m of public money chasing the last fire ant to eradicate it while the infestation blew out from 40,000ha to 400,000ha. In 2015, the Australian Agriculture Ministers’ Forum commissioned a review to determine the course for the future - continue with an eradication effort or revert to a program of containment and control based on a comparison cost-benefit analysis of the two options. The review panel, chaired by Bill Magee and advised by modellers from Monash University, could not conduct a cost-benefit analysis of the options because Biosecurity Queensland does not collect reliable and consistent data on the program. Nevertheless, with no data, the review panel said eradicating fire ants from south-east Queensland was still feasible with up-dated remote-sensing surveillance technology, ie helicopter surveillance for fire ant nests. Technology to detect fire ant nests from the air has never existed, does not exist now and is not likely to exist in the foreseeable future. In 2008, modellers from Monash University said fire ants were spreading faster than the program was finding them and advised using remote-sensing surveillance technology to find them by helicopter. They said the technology was already being used in the USA. Scientists in the USA use the technology to study known nests for research purposes. They do not use the technology to find the nests in the first place. In 2009, the Roush scientific review panel, cautioned against using technology, saying if it could not distinguish between other warm elliptical objects like cow pats and rocks, it will virtually identify everything as a fire ant nest and if it misses small nests, those nests will spread. They also said there are huge technical difficulties in integrating the electronics of the three complex cameras to produce a single useful image and the system would need huge capacity to store and analyse up to a terabyte of data from each of three cameras each day. Nevertheless, between 2009 and 2015, Biosecurity Queensland trialled the remote sensing technology. It falsely identified thousands of cow pats and rocks as fire ant nests and by the end of 2015, it had identified only 38 colonies. In the meantime, it missed thousands of fire ant nests as the infestation blew out of 93,000ha to 300,000ha. In 2017, on the recommendation of the Magee review panel, Biosecurity Queensland commissioned a company to advise on the development of an up-dated remote-sensing surveillance system. The company said almost exactly what the Roush review panel had said in 2001. They said the basic premise of the technology, that fire ant nests could be detected with thermal imaging because they are warmer than their surrounds during cooler months, was wrong. Soil clumps, cow pats, rocks and wood were also warmer than their surrounds and were identified as fire ant nests. The technology was effectively identifying everything as a fire ant nest, but was missing actual fire ant nests because the technology identified them as clumps of soil. The company advised that any future remote-sensing surveillance technology needed to collect five sources of data, not just the three as previously, and they recommended cameras that could collect that data. They also advised that any new remote-sensing surveillance system would need: • additional hardware to synchronise the different cameras from different manufactures. • CASA approval to mount the system on a helicopter. • A GPS system that could access both US and Russian satellite constellations, • Software to facilitate inflight camera configurations. • A high speed, high capacity storage system to store two images per second for each of the four cameras for up to 6 hours a day. • Approval from the US State Department to ensure any cameras or components imported from the USA met International Traffic Arms Regulations to control the use of defence related technology. • And the helicopter would need to fly higher and faster than previously. Little wonder Biosecurity Queensland has not conducted remote-sensing surveillance for fire ant nests since 2015. There has has never been an effective remote-sensing surveillance system for detecting fire ants by air, there is not one now and there is not likely to be one in the foreseeable future. But Bill Magee got the plumb job as Chair of the National Red Imported Fire Ant Eradication Program’ Scientific Advisory Group, and Dr Wendy Craik, who presented the Magee review to the Agriculture Ministers’ Forum in July 2017, got the plumb job of Chair of the program’s Steering Committee. The Queensland public just got more fire ants Time for a Royal Commission. 21st April 2020.
By 2015 Biosecurity Queensland had wasted $400m of public money on the National Red Imported Fire Ant Eradication Program, in a futile attempt to find and kill the last fire ant instead of tightly containing and managing the 40,000ha infestation that existed in 2002.
By 2015, the fire ant infestation in south-east Queensland covered over 400,000h, ten times worse than in 2002. The Australian Agriculture Ministers’ Forum (AGMIN), who have the final say on the program, were asking serious questions about its future – should they continue to fund an eradication effort or should the program revert to containing and managing the infestation – as fire ant experts from the USA had advised in 2001?
AGMIN commissioned an independent review of the program to answer that question. They asked the review panel to advise the best option and the appropriate level of funding for that option, based on an assessment of the program’s progress from 2001 to 2015 and on a comparison cost-benefit analysis of the two options. Mr Bill Magee, Director of Magee Consultancy Services Pty Ltd, specialising in plant biosecurity, was selected to chair the review panel and Magee appointed Dr Daniel Spring, a computer modeller from Monash University, to conduct the analyses. Spring was familiar with the program, having conducted modelling in 2008 that showed that fire ants could not be eradicated while they were spreading faster than they were being found. He recommended the program commence trials with remote-sensing surveillance technology, ie helicopter based surveillance, to detect fire ant nests more quickly from the air. He said the technology was already being used in the USA.
In 2016, Magee reported that the review team could not analyse the program’s progress because it does not collect reliable and consistent data on the efficacy of its treatment and surveillance regimes. For the same reason, the review team could not conduct a comparison cost-benefit analysis of continuing with an eradication effort or reverting to a program of contain and control. The reviewers acknowledge they could not recommend an optimal approach, only give guidance.
Nevertheless, with no data to support their recommendation, but with a history of supporting trials of a remote-sensing surveillance technology from as early as 2008, the Magee review team said it was still feasible to eradicate fire ants from south-east Queensland with an up-dated version of the failed remote-sensing surveillance technology that had been decommissioned in June 2015. And they recommended continuing the eradication effort for another ten years with another $400m, simply because that was how much it had cost so far.
On 26th July 2017, AGMIN ratified a new Ten Year National Red Imported Fire Ant Eradication Program with a budget of $411.4m on the recommendation of an intergovernmental panel, chaired by Dr Wendy Craik, who had been invited to attend that meeting of AGMIN. AGMIN also approved the membership of the Steering Committee that would provide strategic oversight of the new program. Dr Craik did not appear on the National Biosecurity Committee’s list of potential candidates for the position of Chair of the Steering Committee in 7th June 2017. But she got the job anyway. And Bill Magee got a nice job as Chair of the Scientific Advisory Group of the new Ten Year Plan National Red Imported Fire Ant Eradication Program.
Remote-sensing surveillance technology to detect fire ant nests from the air by helicopter has never existed, does not exist now and is not likely to exist in the foreseeable future.
The technology has never been used for fire ant surveillance in the USA. In 2008, US scientists were trialling the technology to monitor the behaviour of known fire ant nests for research and development purposes. Their trail showed the technology could detect around 80% of known nests, but still 17.5% of them were false alarms – a significantly high false positive rating. US scientists do not use remote sensing surveillance technology to detect those fire ant nests in the first place; ie for surveillance purposes. The critical rating for a surveillance technology is not its false positive rating, ie the number of times something other than a fire ant nests is identified as one, but its false negative rating, ie the number of times real fire ant nests are missed, because they are identified as something else or because they were too small to detect.
Biosecurity Queensland’s trials between 2009 to 2015, to create ‘world first’ remote sensing surveillance technology for fire ants, something not used anywhere else in the world, were a disaster.
In 2009, then ALP Minister for Agriculture, Tim Mulherin announced that Queensland was ramping up its fight against fire ants by developing world-first technology to detect fire ants from the air. The technology was supposed to be simple. On the premise that, during the cooler months, fire ant nests are warmer than the surrounding land, they could be detected with thermal imaging cameras. Biosecurity Queensland created its own remote sensing surveillance technology with three cameras to collect visual images of fire ant nest shaped objects, infra-red images of warm fire ant nests, and near infrared images of the amount of chlorophyll in the area because fire ants remove vegetation from their mounds.
The first trial in 2009 was a disaster. The three cameras could not focus on the same spot at the same time so they produced blurry images. At best, they could detect 30% of nests that were bigger than 30cm. The second trial did not start until November 2011 and did no better. The system could not integrate the electronics of the three cameras to produce a single image and the computer program that was supposed to analyse the images to identify fire nests identified 5,200,000 possible nests. In other words, it identified rocks, cow pats and other rubbish as fire ant nests. The third trial in 2013 identified 350,000 possible fire ant nests. Field staff inspected 67,000 of them and found only three infested properties and a total of six nests. The Courier Mail had a field day saying ‘$7.2m, 2 years, 85,000ha and this is all they found!’
In August 2014, Biosecurity Queensland was still touting the wonders of the remote-sensing surveillance technology on its Facebook page. ‘Can you believe that we can pinpoint fire ant mounds from our chopper?’ they said. Well, they couldn’t. By 2015 the cameras had become old and even more unreliable than they had ever been. A review that year found that, from 2009 to 2015, Biosecurity Queensland’s remote-sensing surveillance technology had detected a total of 38 colonies on 23 sites. Meanwhile, the infestation had blown out from near 100,000ha to over 300,000ha.
This total waste of time and public money should not have happened. Biosecurity Queensland had been advised against using remote-sensing surveillance technology from as early as 2009, for good reasons. But they ignored the advice.
While the Monash team was advocating for remote-sensing surveillance technology in 2008 and Minister Mulherin was touting it as the silver bullet to save the program in 2009, Professor Rick Roush, the chair of the fourth independent scientific review of the National Red Imported Fire Ant Eradication Program, was cautioning against it.
Professor Roush, a specialist entomologist and geneticist and Dean of the School of Land and Environment at the University of Melbourne chaired the fourth independent scientific review of the program. The findings of the Roush review panel were clear, damming but not surprising. They were alarmed that the fire ant infestation had blown out from 28,000ha in 2001 to what they called ‘an all-time high’ of 93,000ha in 2009. This, they said, was evidence that the program’s surveillance methods could not detect new infestations and the program’s treatment methods were questionable because the infestations are recurring in treated areas. In conclusion, they said fire ants could not be eradicated from Brisbane using current methods.
They said the primary reason why fire ants were then infesting 93,000ha was because the program could not find fire ant nests. And they were not impressed with Biosecurity Queensland’s plan to use remote-sensing surveillance technology to find fire ant nests by helicopter.
The review team remarked that this was the third time the program had considered using the technology. They said the biggest problem with the technology was the likelihood it would not be able to distinguish between fire ant nests and other warm elliptical objects like cow pats or rocks. If the technology could not control its false positive rating and it effectively detected fire ant nests everywhere, there was no point to it. And if it declared areas to be fire ant free when it has not detected small or incipient nests, those nests would eventually spread.
They were also concerned that because the technology could only be used between May and September when the ground was cool enough to detect warm fire ant nests and the helicopter could not fly during wet or windy weather, it was still likely fire ants would continue to spread faster than the program was finding them.
On top of that there were huge technical difficulties in integrating the electronics of the three complex cameras (to collect visual, infrared and near infrared images) into one useful image, and the system would need huge capacity to store and analyse the one terabyte of data each camera captured each day.
Finally, they said there were significant challenges with this technology that needed to be rigorously assessed as soon as possible. If the technology could not identify fire ant nests, there was no chance for a future eradication program at reasonable cost.
But the program continued with its failing eradication effort and by 2015, the infestation had blown out from 93,00ha in 2009 to 400,000ha in 2015.
In 2016, AGMIN accepted the recommendation of the Magee review, presented to AGMIN by Dr Craik, to continue to fund an eradication program with up-dated remote-sensing surveillance technology.
In September 2017, Biosecurity Queensland claimed that remote sensing technologies were an essential part of a suite of surveillance tools used by the National Red Imported Fire Ant Eradication Program in south-east Queensland to find and eradicate fire ants, and claimed their trials of remote sensing surveillance technologies had shown it was feasible to detect fire ant nests from the surrounding terrain using multi-spectral airborne sensors. In September 2017, Biosecurity Queensland sent out tenders for organisations to undertake research to identify the spectral bands best suited to maximising the probability of detecting fire ant nests and minimising the problem that had plagued earlier trials – the huge number of false positive findings (many thousand warm cow pats and warm rocks), and to give advice on developing a camera system for capturing the data.
In June 2018, Outline Global Pty Ltd won the tender, completed their field trials by August 2018 and reported in December 2018. They said much the same as what the Roush review had said in 2009.
They said the basic premise of the technology was wrong. The premise that fire ant nests are warmer than their surround areas and their ‘heat signature’ was the most important data for detecting them, was completely wrong. They found a very high number of other features in the landscape also exhibited positive temperature differences relative to their surrounds: eg bare dirt or soil clumps, cow pats, dry vegetation, green vegetation, rocks and wood. These gave the technology a high false positive ratings. They also found the technology identified actual fire ant nests as bare dirt of soil clumps, giving the technology a high false negative rating as well.
They said, to reduce the number of false positive findings, a system needed to collect data on a minimum of four spectral bands, not just the three used in earlier trials. They recommended a system that would collect five sources of data – colour (red, green, blue), near infrared and ultraviolet data and high resolution short-wave infrared and long range infrared data and recommended of four cameras that, together, could collect that data.
However, they noted many factors that would impact the performance of any future remote-sensing surveillance system that were outside the range of their research. They were:
Remote-sensing surveillance technology to detect fire ant nests from the air by helicopter has never existed, does not exist now and is not likely to exist in the foreseeable future. But Biosecurity Queensland has wasted millions of dollars of public money in its attempt.
Mr Magee and Dr Craik got plumb jobs advocating it, but all the Queensland public got was more fire ants. The fire ant infestation is now out of control. It covers over 500,000ha as fire ants are continually found beyond the operational bounds of the fire ant program and re-infesting well treated areas.
Time for a Royal Commission