News for hackers. Stuff that matters.
Re:Australia fires: Misleading maps and pictures go viral by jcranmer on Tuesday January 07, 2020 @09:12PM (#21987589)
The intellectually honest thing to do here would be to either try to find some precedent study of wildfire mortality for wildlife to get a coarse estimate on what happens, or to admit that you don't know. Instead, we have an "affected" number that is technically not presented as "killed", but we-kinda-sorta-hope-you-interpret-it-as-such.

I'm personally frustrated that no one has been able to dredge up an actual study on wildfire mortality patterns.

Re:Australia fires: Misleading maps and pictures go viral by nl on Tuesday January 07, 2020 @10:14PM (#21987931)
Small mammal populations in a eucalypt forest affected by fire and drought [2009][1] tracks 5 species in a particular area from 1970 until 2005.

Only the abstract is available (unfortunately), but by my reading of that it shows that a 1973 bushfire wiped out all 5 species in the area ("Following the 1972 fire, numbers fell to the lowest level recorded during the study and each population subsequently disappeared from the plot between the 1973 and 1974 winter censuses").

In further reading (I read a lot of abstracts trying to find one that had numbers!) it appears this pattern of some surviving the immediate fire but 18 months later having disappeared dur to predators and lack of habitat is fairly common.

[1] https://www.publish.csiro.au/WR/WR08086

Re:Australia fires: Misleading maps and pictures go viral by wutbrodo on Wednesday January 08, 2020 @03:40AM (#21989582)
Thank you, this is _exactly_ the kind of analysis I'm interested in! (but unfortunately, far too nuanced to be comprehended or covered by media with any sort of reach).

> it appears this pattern of some surviving the immediate fire but 18 months later having disappeared dur to predators and lack of habitat is fairly common.

I don't think your reading of this is quite correct (or perhaps the implication that I'm incorrectly reading into it). The species didn't quite "disappear", as all of the ones claimed to have done so after the 1972 fire were able to be tracked during the 1980 fire:

> Following the 1972 fire, numbers fell to the lowest level recorded during the study and each population subsequently disappeared from the plot between the 1973 and 1974 winter censuses. The less intense 1980 fire did not lead to extirpation, but numbers of A. agilis, A. swainsonii and R. fuscipes declined as drought conditions persisted through 1983.

This suggests to me that they were repopulated according to the carrying capacity of the habitat (which itself recovered), either from areas outside of the study area or from the small populations that remained (the abstract is careful to repeatedly say "disappeared from the plot", a qualifier that you dropped despite it being significant, as sampling has error).

These dynamics are what need to be studied, or at least extrapolated from. The study you shared is based on an 0.02 square mile study area (ಠ_ಠ), of a single ecosystem type; It's possible that the vast scale of the current wildfires mean that recovery from external populations is much less likely, but the claim under discussion here is very far off either way. (For more information about these dynamics, the "Population refuges and immigration" section of this paper is interesting).

At any rate, the centrally-trumpeted claims that these animals have already died based on habitat destruction is plainly false.

I guess I should note that I don't think pushing back against these estimates suggests that this isn't a tragedy, or that it rebuts calls to action against climate change or mitigation or whatever. I just don't think it serves anyone to misrepresent the facts, or in this case, for such a sloppily-calculated, off-the-cuff figure to be spread so widely as gospel. This is exactly the kind of thing that gives anti-science skeptics traction. OTOH, maybe I'm typical-minding too hard here: most people aren't interested in an accurate view of reality in any case, and maybe being dishonest and plying the public with bullshit figures is how things need to be done. Maybe the harm from giving stupid denialists fodder is outweighed by the benefit of inciting stupid non-denialists to care. ¯\_(ツ)_/¯

Re:Australia fires: Misleading maps and pictures go viral by Cogito on Wednesday January 08, 2020 @07:59AM (#21990975)
The author of the 480 million affected animals estimate has co-authored an article addressing this in more detail and with lots of references.

I've posted it at https://news.ycombinator.com/item?id=21990825

Re:Australia fires: Misleading maps and pictures go viral by wutbrodo on Wednesday January 08, 2020 @05:54PM (#21996649)
Thanks for doing the legwork here!
Re:Australia fires: Misleading maps and pictures go viral by Cogito on Wednesday January 08, 2020 @06:23PM (#21996874)
You're more than welcome!

I aim to be sceptical when taking on new ideas or information, and all the questions people have are very reasonable.

It's hard to even imagine the size and ferocity of these fires.

The problem is when being sceptical twists into being dismissive. It's actually really easy to do a bit of research and contact professors - but it takes time.

When I first clicked through to the University of Sydney's statement, and clicked on Professor Dickman's profile, I was left reeling by the sheer volume of research the man has done on exactly these problems. In our emails he had an anecdote about jumping over fire fronts while researching them when he was younger.

The scariest part is that the estimates given were for less catastrophic events. They are a low bound based on previous understanding, and this is a far worse event than ever seen before. I highly recommend people reading the linked article, and follow the references therein.

Re:Australia fires: Misleading maps and pictures go viral by nl on Wednesday January 08, 2020 @10:24PM (#21998283)
Do you know how he reconciles the vastly different death rate estimates from the 2009 Black Saturday bushfires with these ones?

Black Saturday: 1M animals killed with 400K hectares burnt = 2.5 animals/hectare

These fires: 1B dead with 8.5M hectares burnt = 117 animals/hectare

I don't know how the Black Saturday estimates were done, just that the source was the RSPCA.

Re:Australia fires: Misleading maps and pictures go viral by nl on Wednesday January 08, 2020 @04:19AM (#21989782)
I don't think your reading of this is quite correct (or perhaps the implication that I'm incorrectly reading into it). The species didn't quite "disappear", as all of the ones claimed to have done so after the 1972 fire were able to be tracked during the 1980 fire... This suggests to me that they were repopulated according to the carrying capacity of the habitat (which itself recovered), either from areas outside of the study area or from the small populations that remained

Yes, this is my reading too. I didn't mean to imply that they became extinct, just that they were wiped out from the area by the fire.

At any rate, the centrally-trumpeted claims that these animals have already died based on habitat destruction is plainly false.

I don't understand what this means ("these animals have already died based on habitat destruction is plainly false")?

This study showed that the fires wiped out the animals under survey. If the same pattern is shown here we'd expect the majority of small mammals to be dead in the fire areas.

Re:Australia fires: Misleading maps and pictures go viral by wutbrodo on Wednesday January 08, 2020 @05:29AM (#21990140)
> I didn't mean to imply that they became extinct, just that they were wiped out from the area by the fire

Right, what I meant was that it's easy to infer from your phrasing (as I did) that the animals were wiped out (even if just locally) in a way whose implications wouldn't include the population being restored (or mostly restored) within a couple years, as was the case for most of the populations studied (a couple of them did permanently dwindle to zero). I don't think this was necessarily implied by your comment, but I thought it'd be better to clarify for our and others readers' benefit.

> I don't understand what this means

Sorry for the poor phrasing. I'm referring to the central claim parrotted by news articles, that half a billion animals _have_ died during the fires, as opposed to the forecast that half a billion will die due to habitat destruction (which I claim in the rest of the comment is itself a suspect estimate).

> This study showed that the fires wiped out the animals under survey.

I don't believe this is accurate. They sampled ~zero animals in the study area, which is easily (and much more likely) explained by the animals spending less time in the burned area until the habitat was restored. The surviviros could easily have been leading fulfilling lives 300 yards away (bearing in mind that the study area was a 700x700 ft square). "no animals don't spend time in recently-burned areas" is a substantially different finding from "burned areas cause total animal death".

This is why the comparison to the wildfires is hairy. The effect depends a lot more on migration dynamics than simple assumptions of total death based on carrying capacity per species. An additional wrinkle for expanding the dynamic in this study to the scope of all animals affected by the wildfire is the fact that it's limited to small mammals: if I'm not mistaken, it's a fairly basic tenet of ecology that a collapse in prey population due in part to predation is accompanied by a boom in predator population, for obvious reasons (followed by oscillations in both populations until equilibrium is reached).

Re:Australia fires: Misleading maps and pictures go viral by nl on Wednesday January 08, 2020 @07:05AM (#21990623)
I'm referring to the central claim parrotted by news articles, that half a billion animals _have_ died during the fires

Well that seems to be what the claim actually is[1].

I don't believe this is accurate. They sampled ~zero animals in the study area, which is easily (and much more likely) explained by the animals spending less time in the burned area until the habitat was restored. The survivors could easily have been leading fulfilling lives 300 yards away (bearing in mind that the study area was a 700x700 ft square). "no animals don't spend time in recently-burned areas" is a substantially different finding from "burned areas cause total animal death"

I don't have access to the full report, but this would surprise me. Have you ever seen what a bushfire does?

Large ones have fire fronts tens of kilometers wide and anything in bushland dies. Here[2] is a before and after pic on Kangaroo Island from the fires there that are happening at the moment. Here's a picture of a car after a bushfire went over it[3]. That silver on the ground is the wheels and engine block melted. Bushfires in Australia are nasty because the Eucalyptus oil makes them very hot - the radiant heat can kill 400m away[4].

This fire is worse than normal too, because the fire fronts are so wide. There isn't some place 300 yards away for animals to escape to.

[1] https://sydney.edu.au/news-opinion/news/2020/01/08/australia...

[2] https://www.reddit.com/r/australia/comments/ekt0ra/kangaroo_...

[3] https://www.msn.com/en-au/news/australia/blaze-hot-enough-to...

[4] https://www.science.org.au/curious/bushfires

Re:Australia fires: Misleading maps and pictures go viral by Cogito on Wednesday January 08, 2020 @08:00AM (#21990983)
Replying to you directly here because it seems like something you would be interested in.

The author of the 480 million affected animals estimate has co-authored an article addressing this in more detail and with lots of references.

I've posted it at https://news.ycombinator.com/item?id=21990825

Re:Australia fires: Misleading maps and pictures go viral by nl on Tuesday January 07, 2020 @10:04PM (#21987873)
A fairly old book[1] has bird mortality rates between 40% and 54% for similar fires.

[1] https://books.google.com.au/books?hl=en&lr=&id=4f8anIoS33MC&... (page 358)

Re:Australia fires: Misleading maps and pictures go viral by Cogito on Tuesday January 07, 2020 @10:02PM (#21987867)
I couldn't find the publication myself (though I now see others have posted it in other threads [0]) so I emailed Professor Dickson to get his input.

In the meantime I looked for some studies, and found two that are interesting and include some background on the subject.

The Responses of Mammals to La Niña (El Niño Southern Oscillation)–Associated Rainfall, Predation, and Wildfire in Central Australia [1]

> The responses of vertebrates to wildfires are complex, reflecting the varying food and shelter requirements of different species, their exposure to predation, and other factors. Vertebrates that shelter above the ground and are not particularly mobile are often killed directly by fire (Chew et al. 1959; Newsome et al. 1975; Silveira et al. 1999; Simons 1991). Conversely, species capable of extensive movements such as large predators and birds frequently escape the flames and may increase in abundance after wildfire when food resources become available (Loyn 1997; Newsome et al. 1975).,Burrowing herbivores and omnivores are likely to survive fires and, depending on predation, may increase in abundance after postfire flushes of vegetation (Newsome et al. 1975). Postwild-fire decreases in the abundance of some species have been attributed to a lack of food resources and increased exposure to predation (Groves and Steenhof 1988; Lawrence 1966; Newsome et al. 1975, 1983).

Immediate and short‐term responses of bird and mammal assemblages to a subalpine wildfire in the Snowy Mountains, Australia [2]

> Over 35 days in January–February 2003, wildfires burnt across much of the subalpine/alpine landscape of south‐eastern Australia, including about 70% of the land above 1500 m in the Snowy Mountains. At the time of the fire, studies of two subalpine faunal assemblages were being undertaken. The opportunity to resurvey the assemblages was taken in order to examine the immediate impact of fire in an environment where it is uncommon but predicted to occur increasingly with global warming. A study area in the Whites River Valley, where the number of bird species was counted monthly from 1996 to 2003, and weekly in late winter–spring from 2000 to 2003, was burnt in one fire. As well as the expected decrease in the number of individual birds, the fire resulted in an immediate decrease in the number of bird species, unlike in previously studied montane forest, with only the regularly wintering species plus the olive whistler and the ground‐feeding flame robin remaining. During the post‐winter avian immigration, few regular spring migrants appeared on burnt sites despite their nearby presence on the unburnt side of the valley. Five of six small mammal trapping grids were burnt. As with fires at lower altitudes, there was an immediate reduction in mammal numbers on burnt grids following the fire, but in addition, one species, Mastacomys fuscus, declined further in the ensuing 2 months both on burnt and unburnt sites. Numbers of Antechinus swainsonii and Rattus fuscipes stabilized until autumn/winter when there was a further decline due to the unavailability of subnivean space to allow winter foraging, allied with a concentration of fox predation on areas still carrying populations of small mammals.

[0] https://www.researchgate.net/publication/318029981_Impacts_o...

[1] https://academic.oup.com/jmammal/article/86/4/689/876336

[2] https://onlinelibrary.wiley.com/doi/abs/10.1111/j.1442-9993....