A typical morning for a wildlife rescuer begins with a patrol of the roads and the removal of dead animals that were hit overnight.
For Surf Coast Wildlife Rescue founding member Jason Cichocki, the nightly victims usually involve a kangaroo or wallaby, multiple birds and the occasional echidna.
He estimates about 250 kangaroos and wallabies were being killed each year on a key thoroughfare at Anglesea connecting Victoria's Great Ocean Road to inland towns.
But he said that number had more than halved after "virtual fences" were installed along 12.4 kilometres of Forest Road.
"What I noticed in the second year was the animals were what I would call habituating to the fact that the noise meant that a vehicle was coming," he said.
The so-called virtual fence looks a lot like the reflective posts you might often see on the side of the road, but these posts are green in colour.
Placed at 25-metre intervals, the posts have devices attached to them that are activated by approaching headlights from dusk until dawn.
The devices then emit a sound, paired with flashing blue and yellow LED lights, which alert animals and form a virtual barrier.
The Anglesea trial is in its third and final year and Mr Cichocki said he had also observed a reduction in the number of kangaroos feeding on the side of the road.
"At night time we would quite regularly have mobs of kangaroos on both sides of the road … in the second year of the trial we don't see that anymore," he said.
Wildlife road toll rising
The number of wildlife hit by a vehicle on Victorian roads has increased by about 25 per cent in the year to June 30 this year, compared to the previous year.
Wildlife Victoria recorded 12,129 accidents in 2022-23 and 15,206 in 2023-24.
The highest number of reports are from the Macedon Ranges Shire Council, Hume City Council, Yarra Ranges Shire Council, Whittlesea City Council and Nillumbik Shire Council.
"Some of the common things we see impacting native animals are climate change, habitat destruction and urbanisation," Wildlife Victoria public relations manager Aleisha Hall said.
"Our population is booming and with that comes the need for development for residential and commercial,
"We're really seeing an increase in those peri-urban areas where a lot of our native wildlife live so they're being pushed closer to the roads."
The Anglesea trial was aimed at reducing the number of kangaroos being killed, but Mr Cichocki said it had also helped several other species.
"At least five birds a day were getting squashed on this road … after fitting virtual fencing, in the very first year I went from 1,800 plus birds that I would have to remove off the road to only five," he said.
"We would also regularly see at least five to 10 echidnas hit and killed on Forest Road every year but in the last two years since virtual fencing has been fitted, I've not had a single echidna hit and killed."
The technology has been used across Europe to reduce the number of deer killed on roads.
Wildlife Safety Solutions introduced the technology to Australia, installing the fencing in 41 locations.
The Anglesea trial covers the greatest area, with smaller installations tested at 13 sites across Victoria, including Phillip Island, the City of Casey, the Mornington Peninsula and Mount Buller.
But not all sites reported as great of a reduction in roadkill numbers as Anglesea, with some even doubting the effectiveness of the technology.
Mixed results on effectiveness
A three-year trial on Phillip Island between 2019 and 2022 compared roadkill numbers in fenced sections with unfenced sections and found there was no significant effect of the virtual fence on roadkill rates.
"As a result of this, alternative solutions are being investigated," a Bass Coast Shire Council spokesperson said.
"A research project to alter speed limits is currently being initiated."
A study will be published on the trial, which was a collaboration between the local council, Victoria University, Phillip Island Nature Parks, Regional Roads Victoria and citizen scientists.
Victoria University environmental science lecturer Christine Connelly said there was no robust-enough study that proved the effectiveness of virtual fences.
"We need both more data on these virtual fences to determine whether or not the claims that are being made about their effectiveness are correct and we need to look at a whole range of different species," she said.
"[A study] might be measuring before and after a fence goes in but if they're not actually looking at comparing an unfenced area [with a fenced area] through the same period of time then other confounding effects could be impacting their results."
In contrast to Mr Cichocki, Dr Connelly observed a different kind of habituation where wallabies were no longer alarmed by the virtual fence.
"So maybe looking at a technology that has more species-appropriate sounds or perhaps even different sounds, might be a better deterrent," she said.
Alternative solutions to reducing roadkill
Across Australia, other technologies to drive down roadkill are being investigated, including the RooBadge – a collaboration between car manufacturer Volkswagen and the University of Melbourne.
The RooBadge is an accessory placed on the front of a car that uses AI to compare its GPS coordinates to the distribution of kangaroos in an area, emitting a sound to deter the kangaroo species.
University of Melbourne researcher Helena Bender said the RooBadge used natural sounds, unlike the virtual fence, which could reduce the risk of habituation.
"We're currently doing trials out in Plenty Gorge and we will do some at a couple of the other state parks, we're doing very much controlled trials at the moment," Dr Bender said.
Wildlife Victoria is also working with councils on its Wildlife Road Toll Reduction Project to determine road safety strategies that could be implemented to reduce roadkill including exclusion fencing, LED road signs and underpasses and overpasses.
But Victoria University's Dr Connelly said the most effective solution to reducing roadkill was by slowing down on the road.
She said she would like to see a trial across different sites looking at dynamic speed reductions.
"So that might be having different speeds at day and night or at dusk and dawn, than at other times when maybe the likelihood of roadkill and collisions with wildlife are lower," Ms Connelly said.