Skip to content

New paper live today: Precision wildlife monitoring using unmanned aerial vehicles

A project I’ve been working on has finally come to fruitition: the drone paper is now live at Scientific Reports.

drone through flock

Jarrod Hodgson with drone, East Island, Ashmore Reef. Photo: Shane Baylis


In it, we show that counts of seabird colonies taken from drone imagery are dramatically more precise than counts of colonial seabirds taken by traditional methods (i.e., counting colonies using a spotting scope), across three species (Lesser Frigatebird, Crested Tern, and Royal Penguin), in environments in the tropical Indian Ocean and subantarctic Macquarie Island.

Given that drones are highly portable and their price is dropping rapidly, drones seem likely to dominate colonial-animal population-counting in the future, replacing traditional ground-based counting in many projects. For long-running colony-monitoring projects that choose to change their approach, there will be a transition to manage here: our project found that drone-counts can give a larger average colony-count than ground-based counts, and the amount of difference between drone-counts and ground-counts can be different between species.

So, we suggest a method to determine whether drones under-count or over-count a new species relative to ground-counts, and if so, by how much. Importantly, we also show how many times a research group will have to double-count (i.e., count using both drones and traditional techniques) in order to make the two types of data comparable at a given level of precision.

We’ve had good press coverage, too: write-ups at New Scientist, the ABC, Australian Popular Science, the Australian Science Media Centre,  and new media requests coming in as I type.

H/t to my co-authors, Jarrod Hodgson, Rowan Mott, Ashley Herrod, and Rohan Clarke.


Three-minute Thesis 2014: “What makes a seabird develop a pioneering spirit?”

As it happens, Monash participates in Three Minute Theses, where once a year, excitable young researchers try to explain the geekiest thing that they know about in three minutes or less, to an audience of bemused non-specialists.

Last time this happened, I signed up. Went to the School of Biological Sciences round. Placed second. Went to the Faculty of Science round. Placed second again. Got invited to present in front of a class of undergraduates being taught scientific communication, on the grounds that ‘the best way to teach communication is to show people communication done well’. Was pretty surprised at how little time it takes to transition from ‘no communication experience’ to ‘being held up as an example of how to communicate’, but took the compliment anyway.

And that is the true story of why (some) undergraduates now give me funny looks in the Student Commons. If you’d like to see what they saw, you’re in luck, because someone took a video. Enjoy!

Inferring maximum lifespan from maximum recorded longevity in the wild carries substantial risk of estimation bias

I mentioned a while ago that I had a paper coming out in a Rather Nice Journal. It’s out!

The paper looks at patterns in Maximum Recorded Longevity data – records of the oldest known individual – across multiple species. Maximum Recorded Longevity data have been widely used to look at differences in maximum lifespan in response to things like body size, phylogeny, diet, and flying ability. However, this paper shows that Maximum Recorded Longevity is not a particularly good proxy of maximum lifespan, and is badly biased by things like the number of individuals that there are data for in each species, whether the species tend to be alive or dead when they are recaptured, and the shape of the mortality curve for each species. Many previously-published studies are badly affected by those biases, and some may present completely false results as a consequence.

In any case, if you’d like to read the paper, it’s available here. If you’ve got questions about it, feel free to ask them in the comments!


h/t to my co-authors, Marc de Lisle and Mark Hauber

Why eradicating invasive mice on the Farallon Islands is probably not a problem

Oh, dear.

I’ve just been directed towards a piece on the HuffPo, titled ‘The Farallon Islands, USFWS, and Island Conservation’s Tax-free Government Contracts‘, which gets riled up about New Zealand’s habit of eradicating introduced predators from offshore islands, and especially about the fact that our technology is now spreading to the US of A. It features juicy little tidbits like:

“This $1.3 million dollar (at a minimum) federal contract with Island Conservation targets non-native mice; however, the aerial broadcasting of loose rat poison pellets from helicopters over an island wildlife refuge will result in thousands of protected, wild animals being poisoned as “by-kill” or “collateral damage.” ”

“Lurking in the background is usually New Zealand’s Department of Conservation and their Island Eradication Advisory Service Group selling a variety of consulting services, including the bait and helicopter pilots necessary. Yes, the government of New Zealand is in the business of manufacturing Brodifacoum, 1080, and other horrific poisons via this factory, known as a State Owned Enterprise, which is a private business that the government owns, and profits from.”

“The practice of using helicopters to carry out poison drops (be it brodifacoum or 1080) is only done by the government of New Zealand. No other country in the world engages in this macabre practice that is driven by a business model that demands that an invasive species “crisis” be manufactured whenever money runs low.”

“No independent, third party auditor monitors the full ecosystem impacts of applying unprecedented volumes of one of the most deadly pesticides to wildlife since DDT, or confirms any of the claims of success made in press releases.”


“The risk for contamination of the Farallon food chain and long term epigenetic impacts to the entire ecosystem have been overlooked and dismissed by USFWS, Island Conservation, and Point Blue Conservation Science.”

Spoiler alert: no.

Tiritiri Matangi (foreground) and Little Barrier Island (background): secretly ravaged by human greed?


If someone were to take this piece of journalism seriously, they’d end up thinking that the New Zealand Government had conned its taxpayers into paying to set up a State Owned Enterprise that produces a poison-bait that the New Zealand Government is the sole customer for, that the Government then buys using taxpayer funding, resulting in the sorts of obscene profits to the Government that are liable to distort parliamentary process and cause all manner of corruption. The big loser in this story is the wildlife, who are slaughtered willy-nilly to line these corrupt bastards’ pockets.

On the face of it, this already seems pretty unlikely. After all, New Zealand is pretty routinely ranked amongst the world’s least corrupt democracies (we’ll beat you yet, Denmark!), so a blatant swindle like this would be quite an aberration from the rest of our behaviour. Then there is the matter of how this equation, which is claimed to describe the New Zealand Government’s behaviour, makes any sense:

“Pay to set up State Owned Enterprise”
+     “Pay for the bait the State Owned Enterprise produces”
+    “Pay for the staff to apply the bait to the landscape”
=     “Profit!”

In the journalist’s defence, I can see where they went wrong. New Zealand has, after all, developed techniques to exterminate small (and not-so-small) furry mammals from large tracts of land. We do, indeed, use aerially-boradcast brodifacoum and 1080 poison-baits to do it. Yes, by-kill is a potential issue. But they seem to have missed a few rather crucial points.

First up, the reason we eradicate invasive species isn’t for profit. Profit for the country might come later, because the ecological restoration associated with eradicating introduced species attracts a whole lot of tourists, but the people who sell poison-bait aren’t the same people who sell tourism packages. We eradicate invasive species because the ecological benefits of eradicating invasive species are huge, and incredibly well-documented. The IUCN (famous for producing the Red List of Endangered Species, amongst other things) lists invasive species as the second most significant cause of extinction worldwide, behind only habitat loss. The damage caused by invasive species is ongoing, with more species being driven extinct by them on a yearly basis. If we remove the invasive species, that threat disappears – and the evidence we have shows that the ecological recoveries caused by removing invasive species are dramatic.

Secondly, if this is a con, then it’s the only con in the history of the Universe where the primary method is to tell everyone everything that is going on, at every step of the process. In government-funded projects in a country as small as New Zealand, there are very few areas of spending that escape the notice of the public. The people running the programmes know that public approval of the process is vital, and that if the public couldn’t see the massive benefits, it would be all to easy to believe that aerial deployment of predator baits was unjustified or too costly. So the aggressively-followed regimen has been: Make it accessible. Publish the data. Make damned sure that everyone can see what happens.

So, if you want to visit one of the sites where we have eradicated invasive species, here’s how you do it:

1) Catch a plane into Auckland, New Zealand’s most populous city.
2) Catch the AirBus (NZ$16 one way; $28 return) to the central city. Get off at the Downtown Ferry Terminal.
3) Walk to Pier 4, and book a ticket to Tiritiri Matangi Island with 360 Discovery Cruises (Adults $69 return; family passes available).
4) Get on the boat. The trip will take about an hour. Don’t drink the coffee; the tea is much better.

If you happen to be in the capital, Wellington, you can visit Matiu/Somes Island instead of Tiritiri Matangi. The process is basically the same, but the ferry tickets are cheaper (adults $23 return) because the island is closer to the central ferry building. Or you can drive to Paraparaumu (about an hour out of the capital) and catch a ferry to Kapiti Island (adults $95).

If you’d like to visit a site during a 1080 or brodifacoum deployment, you can generally do that too. Eradication of invasive species isn’t always feasible, so many popular hiking areas receive periodic predator-control using the same methods as we use for eradication. Signs informing people of the baits’ existence are a familiar greeting to many of these areas – especially just before the breeding period, to knock back predators and give our endangered species the best chance of breeding successfully. Generally the helicopter pilots are good enough to prevent any baits from landing on hiking tracks, but the signs give contact numbers in case any are found on-track, tell people how to identify baits, ask people not to touch any baits that they do see or eat any animals from the area, and especially warn people that dogs should not be brought into the area while baits are present.

In addition to making predator-controlled sites as open as possible to the public, the New Zealand authorities have been very careful to make the scientific data available to the public. So, if you want to know what the likelihood of landscape contamination is, that’s fair enough. The research has been done by a variety of research organisations, including our leading universities and crown research organisations. If you’d like to see the breakdown profile in water for 1080 or brodifacoum, that’s all freely available. If you’d like to see data on the effects of brodifacoum application on our native scavenger populations (which you might expect to scavenge poisoned rats, and thereby be poisoned themselves), you can find a paper on that. If you’d like to see a worst-case scenario for 1080 absorption into plants, or popular food species, or its distribution through the food web, that’s all available. If you’d like to see the original research papers, many of those are also available – though some may require a University library membership.

The upshot of this research is that both 1080 and brodifacoum biodegrade on a timescale of days to months, depending on weather conditions, so uneaten baits don’t accumulate to any substantial degree in the environment. Brodifacoum persists for longer in tissues (especially liver tissue) than on the forest floor, so hunters are advised to discard offal from wild meat shot in areas where brodifacoum has been applied, and active research is underway to find alternatives to brodifacoum to avoid any possibility of bio-accumulation.  There are non-target species that are at risk of being poisoned during these operations. We’ve managed to hugely cut down the bykill rate by doing things like dyeing baits so that they don’t look like food to non-target species and researching the effects of alternative baits on target species behaviour. We’re now at the point where the non-target secondary poisoning rate is detectable, but dramatically swamped by the increase in breeding success associated with predator removal.

A final word: the HuffPo piece ends with a video from the ‘censored’ Graf Boys, who have some objections to the use of 1080. They’re not censored, of course – the keen observer will notice that the Graf Boys have a YouTube channel that no-one is asking them to take down. If anyone actually tried to censor them, that would justifiably cause public outcry. If we were the sort of nation that practised censorship on a large scale, we probably wouldn’t rank amongst the world’s least corrupt democracies.

In reality, the anti-1080 lobby is a consistent, but small, voting bloc comprised mainly of hunters. They are represented in Parliament occasionally. For instance, back in 2006 an anti-1080 lobby group made a deal with a small religious conservative party called United Future (0.6% of the popular vote at the last election; greatest electoral return 6.7% of the popular vote in 2002), but the anti-1080 lobby split from United Future soon after on the basis that they supported the separation of church and state, whereas United Future would like to see Christian doctrine officially supported in New Zealand law. At the moment, the anti-1080 lobby’s greatest support in Parliament seems to be New Zealand First (6.59% of the popular vote at the last election), who quietly list opposition to 1080 as one of their official party stances. The incumbent National-led (47.31% of the popular vote at the last election) government, while dramatically decreasing DoC funding generally, have increased funding for 1080 specifically. Labour (the major opposition party, 27.48% of the popular vote at the last election), and the Green party (11.1% at the last election) both endorse 1080 use, although the Greens specifically endorse actively looking for alternatives.

h/t also to Peter Pyle, for his assessments of the specific case of the Farallon Islands.

[edit: the title was “Why eradicating predators on the Farallon Islands is probably not a problem”, but this lacked clarity. Mice are important predators in island ecosystems, especially of invertebrates and herpetiles. I have made the change to avoid any potential confusion with Burrowing Owls]

Ashmore 2: the pseudo-squall

We took the ‘two tides’ approach to sampling Middle Island last night – out on the high tide, back on the rising tide the following morning (around 09:00). The main tasks for the night were retrieving loggers (from Frigatebirds and Red-footed and Masked Boobies), coaxing some remote-downloading loggers to go, and, if enough time, catching more Lesser Noddies.

Sadly, the night was a bit of a bust, excepting the hunt for Masked Booby loggers. The adult Lesser Frigatebirds to which the GPS units are attached only visit the colony fleetingly, for long enough to feed their chicks, and so are very difficult to catch. The remote-downloading loggers still could not be coaxed into downloading.

Around 03:00, a large thunderhead coming from the South-East caused us to cover all equipment in a tarpaulin, deconstruct the sampling station, and generally cower. However, it dissipated just on reaching us, so only a small amount of rain actually fell on the base – although it did become very windy. Following the pseudo-squall, we slept on the beach until sunrise.

Ashmore 2: The day that keeps getting longer

PM 17/11/2013

On East Island this morning doing vegetation mapping and collecting dead seabirds. The map is now complete except for the island’s high-tide line and the stick-stacks that most of the Red-footed Boobies roost and nest on. The dead seabirds will complement my molecular records from Middle Island – we’ll trim a wing off each to use in analysis, except for the specimens that are fresh enough to offer to museums – those will be frozen whole until we get back to land.

In a somewhat-rare observation, I picked up an Australian Pelican skull and long-bones from East Island. This is not the first observation of the species on Ashmore Reef, but they are rare here, so it is possibly the first specimen. I have GPS co-ordinates of the rest of the skeleton.

New island tick: Splitgerber Cay (a small vegetated cay out to the East of East Island, where we collected two of the Shorebird Team en route to the Diversity at the end of the day).

PM +:

Have cut up and vaccuum-packed/frozen one foot and one wing each from one Masked Booby, two Red-footed Booby, four Brown Booby, four Sooty Terns, four Bridled Terns, four Lesser Frigatebirds, and five Lesser Noddies, in varying states of decomposition. AH and RM assisted with vacuum-packing and labelling, although few others chose to be on-deck for very long.

At the same time as I was cutting up my prizes, one of the ship’s tenders made a trip out past the reef-edge to dispose of food-scraps outside of the marine reserve. It returned with an empty chicken-feed sack and a ‘ghost net’, a discarded or lost fishing net that has been drifting the ocean. It was full of crabs, mostly too small to be entangled. Apparently it is standard practice to hand retrieved ghost nets to Customs, who are ever-present.

PM ++:

Took a ship over to West Island – the twitchers have spotted an Asian Brown Flycatcher (or some other small brown bird with a similar name), and I piggybacked on that tender with JL and JC to go and play with Wedge-tailed Shearwaters. The colony here is small – tens of burrows, maybe around thirty? – but it was my first chance to handle Wedge-tails. Regrettably, we didn’t have gloves, so when I was neck-deep in a burrow and had my hand bitten, I flinched and failed to extract it. It was eventually extracted by JL, who has much more shearwater-wrangling experience. All up, we extracted two wedgies tonight – one with a data-logger, which we retrieved, and one new bird, which we banded.

In other news: JL and JC have found Jesus. He is about 7 cm tall, and made of plastic. Similarly, they have found a T-rex skeleton and a whole chicken.

We will head out to Middle Island for night-work tonight. The only thing I have still to do on Middle is look for Lesser Noddies, which is a team activity, so I will switch between helping other groups in exchange for some help Noddy-hunting. I imagine we’ll be back on the falling tide, slightly after midnight.

Ashmore 2: in which maps are generated

AM 17/11/2013

We never quite made it out to bleed Lesser Frigates after drinks with Customs. We headed out for an 09:30 -> 19:30 day the following day instead. I mapped the vegetation on Middle until ~14:00, then sat under shade, panting heavily, until ~15:00. I then recovered myself enough to scribe for AH and RM as they bled Lesser Frigates.

The method for mapping vegetation on these islands is very simple: identify patches of vegetation (Calthrop and Argusia, and a scattered low grass, are the only things here in this season), and walk around the edge of each patch with a GPS unit set to record tracks. Then walk the high-tide line of the island with another track, and mark points of vegetation that are too small to be recorded as patches. Later, we’ll have to fiddle in ArcMap to turn the tracks into polygons and overlay them on other maps, but the data-generation itself is very simple.

Today, the plan is to vegetation-map East Island while the Shorebird Team count birds on the beach, and to pick up dead specimens to bulk up my East Island collection.