Statistics

Will Climate Change Kill One-Sixth Of All Species?

Fig 2 from the paper, demonstrating the end of the world.

Fig 2 from the paper, demonstrating the end of the world.

Today’s headline was prompted by Mark Urban’s peer-reviewed paper in ScienceAccelerating extinction risk from climate change“, which assures us that if we don’t “do something”, this is the end for a good chunk of life on earth.

Commenting on Urban’s work, the far-left New York Times said “the real toll may turn out to be even worse.

Yes, global warming could kill one-sixth of all species. It could also kill six-sixths of all species. There is nothing global warming cannot do. Why, it might even kill a negative one-sixth of all species—meaning it could cause an increase in speciation.

Think of that! We’d have new kinds of animals we could serve up for dinner, all because of global warming.

That global warming could kill or create species is a logical truth. It is also a logical truth that without human-caused global warming various species might die. What we’re interested in is the likelihood that one-sixth will die given everything we know about the planet and this solar system. Since all probability is conditional, we can make the prediction sound as dire or trivial as we like by being careful how we select our evidence. How careful is Urban?

He opens his paper by worrying about “the biological costs of failing to curb climate change”. A more scientifically ignorant thing to say is difficult to imagine (but give it time: people are always inventing ways of making mistakes). There is no way for human beings to stop climate change. None. Even if all mankind and all his artifacts were tomorrow to be projected into space, the climate on earth would continue to change. Even if mankind never existed, the climate would continue to change. We know these things with as much or more certainty than we know any other scientific fact. Stopping or preventing climate change is never going to happen.

Here’s another indisputable scientific truth: short of a planetary ejection, there is no way to minimize man’s influence on the climate. The minimum would be no influence, and that is impossible so long as man exists. The best we can hope for is to ascertain what might happen given we were to do this or we were to do that.

Now what Urban did was a “meta-analysis”, which I often say is a statistical technique used to “prove” what individual studies couldn’t. The problem with all meta-analyses is that, since all probability is conditional, and all probability models are also, combining studies is often an exercise in slamming disparate evidence together with a 50 lb. mallet. Urban used a pneumatic hammer. Admittedly, he wasn’t working with prime material. The quality of the papers he examined was (statistically) poor.

Urban says the majority of studies he selected for his meta-analysis “estimated correlations between current distributions and climate so as to predict suitable habitat under future climates.” But others relied only on things like “expert opinion (4%).”

A correlation between the distribution of a species and the climate is not a certainty. Meaning that if the climate were to change, the distribution of a species might not decrease and that it might even increase. Which is to say, each study used by Urban had uncertainty attached. Did Urban take this often great uncertainty into account? No. This means his results are too sure. Add to that the ever-present over-certainty in how the climate will change, and we aren’t left with much. Still, let’s continue.

With his hodgepodge of studies, Urban claims “7.9% of species are predicted to become extinct from climate change” (over what period?). He also says that, right now, the extinction rate is 2.8% and that will increase to 5.2% “at the international policy target of a 2oC post-industrial rise, which most experts believe is no longer achievable”. Experts.

Wait. Did he just say that 2.8% of all species are dying each year? We’ve cataloged about 8.7 million species (and of course there are many we don’t know of), which means that next year we’ll lose a quarter million species. That’s about 700 per day. Let’s hope cockroaches are on the list. But it’s almost certainly a wild exaggeration given that the number of actually documented extinctions since anno domini 1600 is only about 1,200. That’s 0.008 per day. Urban is only off by almost six orders of magnitude.

Plus, he and other greenies are very bad about speciation. How many new species are born daily? How many of these are aided in their cladistic journey because of climate change? I have no idea.

Mass extinctions happen. The Permian-Triassic event 250 million years ago wiped out more than half of everything. Nasty business. Probably caused by space rocks or volcanoes and a combination of things going from bad to worse. The dinosaurs also thought space rocks were scary. We should think so, too. Space rocks kill. Watch the skies.

But don’t watch your breath. Slight increases in atmospheric trace gases aren’t likely to be deadly.

———————————————————————————

Note: There is a hunger to discover the “background” extinction rate. There is no such thing. Look at this picture of extinction events. You’ll see there is no constant rate. The search for one is easy to explain: over-certainty. Something causes each species to go extinct, or to come into existence. This cause won’t be some statistically measured rate.

Categories: Statistics

28 replies »

  1. There is virtually no science left in global warming, so pointing out how unscientific the language is is kind of redundant.

    I have come to conclude we are aliens and that’s why we are to blame for everything. We cannot possibly be part of this planet, that’s obvious. Of course, then ejecting us would make sense……..We need to phone home and return to the mother ship.

    Ban meta-analyses?

    Species is a human created idea, just like race. Nature is just a bunch of animals and plants living together and everchanging. How species numbers and extinctions relates to the real world I don’t know. Long extinct animals show up now and again, so extinct is not necessarily accurate or problematic. The EPA loves breaking down animals into more and more species and sub-species to ban activity. In the real world, how important is the concept of “species”?

  2. All you have to do is to look at the graph and see how nonsensical the claim is. I cautioned radiology residents when teaching them statistics (or trying to teach them statistics) to use their eyes and common sense and not rely on calculators. What looks like scattered shots at a barn wall means no necessary correlation. Yet I kept seeing posters with graphs like that above that were supposed to validate a scientific hypothesis…uggghhh.

  3. What definition of species does Urban use?
    1. Agamospecies
    2. Autapomorphic species
    3. Biospecies
    4. Cladospecies
    5. Cohesion species
    6. Compilospecies
    7. Composite Species
    8. Ecospecies
    9. Evolutionary species
    10. Evolutionary significant unit
    11. Genealogical concordance species
    12. Genic species
    13. Genetic species
    14. Genotypic cluster
    15. Hennigian species
    16. Internodal species
    17. Least Inclusive Taxonomic Unit (LITUs)
    18. Morphospecies
    19. Non-dimensional species
    20. Nothospecies
    21. Phylospecies
    22. Phylogenetic Taxon species
    23. Phenospecies
    24. Recognition species
    25. Reproductive competition species
    26. Successional species
    27. Taxonomic species

    And then there is the whole problem of classifying Intergrades and Hybrids.

  4. 90% of all species that ever existed are now extinct. The next large asteroid will be our extinction event. If we saved the earth from this asteroid would we be redeemed in the view of the greens?

  5. The late botanist, Harold C. Bold, like to remind his students that “nature mocks at human categories.” It also mocks at bad research, but most are deaf to the snickering.

  6. “The next large asteroid will be our extinction event. ”
    Yes, but just remember that women and minorities will be more adversely effected.

  7. That is probably one of the worst-designed visualizations that I have seen. It confuses and blurs the numbers, so much that you don’t know what you are looking at. Which, of course, is perfect for a scientific papers showing how complex the science that was done is – after all, clearly (to those in the know) the graph shows that we’re all doomed.

  8. @Yawrate

    Of course not. In addition to being mean to animals (plants don’t count, otherwise nobody would have anything to eat) , humans would be mean to asteroids as well.

  9. This is not letting me post a link.

    No.

    See Betteridge’s law of headlines

  10. We’ve already had about 0.7 degrees C of warming in the last 150 years and that appears to have produced no extinctions at all. Why would the next 0.7 degrees be any different? In any case, species transferred from their normal habitat by accident or design have often thrived in much warmer climates, for example Rabbits transplanted from Europe to Australia where it’s probably 6+ degrees warmer.

  11. I’m actually kind of hoping for a space rock to smash into the earth. We’re going to need another kickstarter before long.

  12. Dr Briggs

    I normally appreciate your droll and incisive wit without comment but I would like to make a small correction to your article. The Permian extinction event wiped out almost 90% of life on earth. I have put my finger on the precise extinction point in a core cut in a petroleum well. The underlying strata are teeming with marine life but only algae lived in the overlying rocks. The cause of the event was most likely not an asteroid but a gigantic volcanic eruption in Siberia that changed the atmosphere. That the animal and plant world ultimately recovered from this calamity shows the resilience and adaptability of life.
    As we say in Australia, “keep the bastards honest!” So keep on keeping on.

  13. Human extinction would minimize man’s influence on the climate to zero. I don’t know if it is possible though. I do think, optimization problems, be it minimization or maximization, without constraints are less fun and realistic.

    It would be more interesting, at least for me, to find out what kind of data are used to construct a model for the projection of distinction risks and how the distinction risk is defined.

  14. Correction- “extinction risk”, not “distinction risk.” (I am watching a movie at the same time.)

  15. As I recall, the study showed animals were moving either up to higher altitudes, where the air is thinner and colder, or north or south, toward the relative poles. This is probably a sign of a warming climate. For many of these species, there is only so far they can go before they simply face extinction.

    JMJ

  16. JMJ, traveling to your relative poles or pole relatives, I hope soon. Really hoping you run out room to move.

  17. “This is probably a sign of a warming climate.”

    Yes, animals are migrating because there has been a 0.3C temperature increase in these regions over the last 50 years. Yet oddly these same animals survive temperature swings of 20C or 30C from summer to winter. The amount of stupid on display by some people never ceases to amaze.

  18. Thanks, Matt, good stuff as usual.

    But wait, it’s worse. As far as I could tell, none of the studies used in the meta-analysis studied actual extinctions. Instead, it’s a meta-analysis of (from memory) something like 138 computer-modeled estimates of future extinction rates.

    Now, in rare circumstances a meta-analysis can be valuable. In general, however, they are meaningless for a host of reasons. For one among the many reasons that meta-analyses are usually meaningless, what is the average of two apples and five oranges?

    But the paper in question takes being meaningless to a new unexplored level. It is a meta-analysis of a big varied stack of computer model outputs … it’s turtles all the way down. I can’t even begin to imagine the theoretical justification of a meta-analysis of computer runs from a host of dissimilar models.

    Keep up the good work,

    w.

  19. I don’t see a problem with a meta analysis if the meta analysis at least uses a set of objective criteria to filter what it will examine. I.e., minimum sample size and trail length. And then it reviews the results from the set of best studies available.

    But if you have a series of studies based on assumptions and guesswork, the absurdity of doing an analysis on that pile of poo, defies logic.

  20. Just realize (again) that giving subtle hints didn’t work. There are differences between rate and risk! For example

    A cancer (incidence) rate is calculated by counting the number of new incidents for a particular cancer, a specific age group, a certain period, a target population (e.g., State or country), etc.

    But the risk is the probability of developing cancer. It is often predicted employing a statistical model with demographic characteristics (DC) of the observational units (OU) as predictors, and is calculated conditionally on a given scenario or a set of values of the DCs. (Hence such predictive results may not be useful when making decision for a group of people, such as FDA drug approvals.) If the OU is a person, the DC can include smoking status, age, and sex.

    We need to know the type of data used to make correct inferences.

    The extinction risks projected in the paper under attack are not extinction rates. It seems that the calculation in this post mistakens the extinction risk for extinction rate. (If not, it demonstrates Briggs is a closet frequentist!)

  21. “Generally though, the climate of the Pliocene is thought to have been much warmer than it is today. The warmest phase was in the middle of the epoch, the interval between three and four million years ago. The climate was especially mild at high latitudes and certain species of both plants and animals existed several hundred kilometers north of where their nearest relatives exist today.
    http://www.ucmp.berkeley.edu/tertiary/pliocene.php

    “NAGOYA, Aichi, Japan, October 26, 2010 (ENS) – At least 1,200 new species have been discovered in the Amazon ecosystem, at an average rate of one every three days during the decade from 1999 through 2009, the global conservation organization WWF revealed today in a new report.
    “Amazon Alive!: A Decade of Discoveries 1999-2009.”
    Presented to delegates from 193 countries at the UN Convention on Biodiversity in Nagoya, the WWF report details the discoveries of 39 mammals, 16 birds, 55 reptiles, 216 amphibians, 257 fish and 637 plants – all new to science.”
    http://www.ens-newswire.com/ens/oct2010/2010-10-26-01.html
     

    What We’ve Lost: Species Extinction Time Line
    http://www.nationalgeographic.com/deextinction/selected-species-extinctions-since-1600/

Leave a Reply

Your email address will not be published. Required fields are marked *