This is interesting

Bad news is always more newsworthy than good. The widely reported finding that insect abundance is down by 75 per cent in Germany over 27 years was big news, while, for example, the finding in May that ocean acidification is a lesser threat to corals than had been thought caused barely a ripple. The study, published in the leading journal Nature, found that corals’ ability to make skeletons is “largely independent of changes in seawater carbonate chemistry, and hence ocean acidification”. But good news is no news.

And bad news is big news. The German insect study, in a pay-to-publish journal, may indeed be a cause for concern, but its findings should be treated with caution, my professional biologist friends tell me. It did not actually compare the same sites over time. Indeed most locations were only sampled once, and the scientists used mathematical models to extract a tentative trend from the inconsistent sampling.

First time I’ve heard that that paper is anything less than pure and perfect science.

Anyone got more?

58 comments on “This is interesting

  1. “the scientists used mathematical models to extract a tentative trend from the inconsistent sampling”

    How the do people get away with that? You can’t “model” results. The whole fucking point is that you measure them.

  2. “the scientists used mathematical models to extract a tentative trend from the inconsistent sampling”

    Hey I can make u stuff to.

  3. Nobody with an interest in the global warming scam will be at all surprised at such worthless, dishonest fake science.

  4. How the do people get away with that? You can’t “model” results. The whole fucking point is that you measure them.

    Standard practice with Greens.

  5. The headline and green twittersphere is replete with the usual cataclysmic hype viz: a carefully designed peer reviewed study has definitively shown a 75% decline in insects. Insectogeddon as reported in one newspaper.

    SJW says read the paper and helpfully provides a link. This is most enlightening. Here are three direct quotes from the methods section of the paper.

    “Most locations (59%, n = 37) were sampled in only one year,”
    “Our data do not represent longitudinal records at single sites, suitable to derive location specific trends”
    “However, the data do permit an analysis at a higher spatial level, i.e. by treating seasonal insect biomass profiles as random samples of the state of entomofauna in protected areas in western Germany”.

    To paraphrase: we have done some random sampling of insects at a variety of sites at different times, direct comparison over time is not possible so no trends can be measured. However if we construct an unverified mathematical model we ca convincingly show the end of the world is nigh

  6. Well, 75% seems a lot to attribute to the imperfect methodology. I know a lot of you are engineers and such, but experiments on living stuff generally are less accurate than in those fields because we have a lot of uncontrollable variables.

    Looking out my window I can tell the population of various birds, squirrels (ahem), rabbits and so easily varies by that much over the course of a few years. So perhaps for insects too.

    Criticising plos one as a “pay to publish” journal is for the ignorant. Many good journals raise page charges.

  7. If 75% of flying insects had gone to meet their Maker it wouldn’t need a fucking load of mathematical bollocks to make the knock-on effects on nature of that circumstance all too obvious.

    It is another pathetic attempt to render the credulous even more frit and malleable for Greenfreak marxian manipulation.

  8. I didn’t notice this year those two or three days in Summer when the entire world seems to consist of flying ants, so perhaps we really are doomed after all.

  9. @Ecksecutioner.

    No, it wouldn’t. Which is why figure 4 shows real unmanipulated data from the locations they had multiple sampling years. One shows steady state, 25 show reduction over time, none show an increase over time.

    The methodological limitations aren’t likely to turn up that kind of consistency, so it’s pretty damn convincing to me at least that something is happening.

  10. I read about what the German insect fanciers were doing, and they seem legit. They are quite open about the limitations of their data, and AFAIK there’s no money in advocating for hoverflies.

    I, too, have noticed that barely any insects splat my car window in summer these days. Just a decade ago, I had to keep a squeegee ready for long drives. It hardly gets used now.

    Dunno what, if anything, it all means, but this is the sort of stuff I think we’d want our environmental scientists investigating, rather than more global warming bullshit.

  11. Biggie–there were fewer flies and bluebottles (the real kind: far too many of Plod unfortunately) around this year. Likely the ;product of three largely cool, wet, crappy summers. Is that the “Death of Grass”?

    I doubt it, all teutonic bullshit to the side. Where are all the dead birds with 75 % fewer flying insects to eat? Nor has there been any shortage of snails/slugs etc this summer. As wet weather suits them.

    You freak out if you like. I’ll keep things Tea –to use a stupid advertising phrase.

  12. Bad news is always more newsworthy than good, because we’re descended from violent troupes of furiously masturbating monkeys who touched a space monolith.

    Or so Richard Dawkins says.

    Think about it: bad news = possible threats to the pack. No wonder it gets more attention. Good news is a nice-to-have, but what humans really want is no bad news.

  13. Ecks. There is some data. It is imperfect. We don’t know what the cause of the change is. We don’t know what the consequences of the change are. Sensationalist media, to the shock and awe of absolutely no one other than Ecks, says it means the end of the world.

    You are the one freaking out.

  14. I, too, have noticed that barely any insects splat my car window in summer these days. Just a decade ago, I had to keep a squeegee ready for long drives. It hardly gets used now.

    Come to France and go for a drive in August. Bring your squeegie, you’ll need it!

  15. Tim – thanks, but no thanks.

    It’s hard to scrape a stripey jumpered cyclist off the grille of a Jeep, and the onion smell never goes away.

  16. TN

    And if you don’t wash them off the new water-based paintwork of your vehicle they eventually ended up printed there forever…

  17. @Rob – We had that in Cornwall this year – even the beaches were crawling with them below the tide line… how they didn’t all drown is beyond me!

  18. As others have said and it *is* worth repeating there was apparently no time series data… and then they leap to extrapolating from random garbage to a crisis….

    If there’s one thing that biologists really are not as a rule skilled at – it’s mathematical modelling and the associated requirement for validation. There is some good work done but from what I see there is much naive use of statistical tools and journalist (NGO even) friendly conclusions are popular.

    The complete + utter imbecility of devising a model without elaborating how it will be validated and then claiming it’s a predictive tool simply leaves this sometime biological scientist just fuming…..

    In the unlikely event that this work is challenged – I presume that the only laptop with the data on its hard drive will be stolen like the “baby fishes eat MacDonalds plastic garbage” from Sweden that is still doing the rounds – even though the actual paper has been retracted.

  19. tomo,

    Agree wholeheartedly but in this case the blinding equations are totally unnecessary. The limited time series data is entirely convincing in its own right.

    While predictive modelling is pretty well established (used in evaluating prognostic factors in clinical trials, where data is usually even more limited), it looks like the whizz bang maths is an attempt to tease far too much information out of limited data. And as you say without some validation, it’s impossible to determine which of those predictions might (emphasis on “might”) be useful, and which can be written off as sheer randomness.

  20. And as the bluffer’s guide to science said, to paraphrase, “We used the Neveredgof-Smirnof exact test for paired categorical ranked goodness-of-fit because it gave us P<0.05".

  21. The problem with all these studies is that the politicians are as clueless as the journalists, and real bans are applied, whose effects are far more serious than any of the non-existent crises they were enacted to solve. Just look at the ban of neonicitinoids. One day a failed crop harvest will bite the EU in the arse. Yet another reason to leave that doomed demographic and economic death cult.

  22. BiG

    Yeps… the maths is generally wheeled when the outcome is either unclear or undesirable. Lost count of the utter bollocks reports I’ve plowed through over the years.

    A 75% slump in flying insects would very obviously have a dramatic impact on several bird species (and to be fair the authors hint at that) – I’d be more sympathetic if the crew had attempted to corroborate their findings with some bird crews… – leading to at least some partial validation of their estimates.

  23. @ SJW
    Can you lend me a magnifying glass? If so, I’ll try!

    I just cannot read some of that, especially the tables – and how can I check anything if the numbers are illegible? There is something seriously wrong with anyone who publishes illegible stuff whether or not it agrees with my prejudices.

  24. Biggie–You have an unnaturally low freak-out threshold. Probably then result of being near the fall-out from Germagreen scum. Amongst the most virulent poison(ers) known.

  25. Perhaps it’s the same thing that caused the bees to die out and no crops to be pollinated, leading to the worldwide famine we just had.

  26. Don’t know about Germany, but here in the UK we’ve had a bumper crop of just about every type of fruit. Something’s been pollinating them!

  27. It did not actually compare the same sites over time. Indeed most locations were only sampled once, and the scientists used mathematical models to extract a tentative trend from the inconsistent sampling.

    Seems they are qualified to be doing climate science.

  28. I also some time ago noticed the lack of squashed bugs on my windscreen.

    In addition, I know that there used in my youth to be far more flies around because getting them buzzing around the house was practically a daily occurrence and now it does not happen at all. Perhaps it was because I used to smell of shit as a small boy, which is entirely possible

  29. To expand on biCR’s comment, the question is; not how many insects you’ve got but how many insects you need. Bees didn’t evolve to pollinate flowers, they evolved to create more bees. Once a bee’s pollinated a flower, as far as the flower’s concerned it’s job done. It doesn’t need more bees, rummaging for bee scoff sticking their heads up its stamens. One blowfly will lay enough eggs to maggot infest a dead…retired accountant from Ely. It doesn’t require a swarm of them.
    Truth is, insects breed like,,,insects. or even more than rabbits. As was mentioned above, ants overdo even this. One suspects the minimum insects required for Gaia to carry on doing her stuff is a lot less then the insects we get.

  30. I blame all the wind farms. Stands to reason, dunnit. Anything flying around gets mashed up by the blades.

  31. Can you lend me a magnifying glass? If so, I’ll try!

    Drop by and I’ll be happy to do that. However, there’s a “larger image” option on all the tables.

  32. @john77, October 23, 2017 at 3:48 pm

    Can you lend me a magnifying glass? If so, I’ll try!

    I just cannot read some of that, especially the tables – and how can I check anything if the numbers are illegible?

    In browser toolbar View, Zoom

    Shortcut Ctrl+ and Ctrl- or Ctrl & use mouse wheel

  33. @BraveFart, October 23, 2017 at 5:32 pm

    I also some time ago noticed the lack of squashed bugs on my windscreen.

    In addition, I know that there used in my youth to be far more flies around because getting them buzzing around the house was practically a daily occurrence and now it does not happen at all.

    Flies – yes massive decline in last 30 years. Those blue UV fly shockers rarely seen now. Spiders which feed on flies also down.

    My theory: lack of fly food – dog-poo, food in open bins.

    The H&S & Council-fines nazis are responsible

  34. The NHS data site has this on hospital admissions related to insect bites for the period 2009-2014.
    http://content.digital.nhs.uk/media/14856/Insect-Bites2009-2014/xls/Insect_Bites_2009-2014.xlsx
    If anyone can find the data for pre-2009 I’d be grateful.

    Although it’s a proxy for insect numbers, and needs controlling for population growth and people being fannies ( can I have a grant please? ), what’s interesting is the variation in the ‘wasps, bees and hornets’ category from 500 one year to 1100+ in the next.

  35. “I didn’t notice this year those two or three days in Summer when the entire world seems to consist of flying ants, so perhaps we really are doomed after all.”

    They were all here, we had the worst infestation I can remember.

    We also have an Insect-a-clear in the kitchen it was was just as busy as I can remember.

  36. “How the do people get away with that? You can’t “model” results. The whole fucking point is that you measure them.”

    Everyone models stuff. A model is how you relate a measurement instrument to what it is measuring. (For example, an eye only measures photons hitting the retina. To turn that into an understanding of the world, you need all sorts of assumptions about light traveling in straight lines, bouncing and refracting and so on.) The question is, is the model accurate enough to be useful?

    Sampling sites once isn’t necessarily a problem, either. This is like doing opinion surveys on the street a year apart, and someone objecting that most of the people in your surveys were only sampled once! When you went back a year later, it was a completely different set of people walking down the street!

    The real question is whether the samples were selected independently of the variables being studied. For example, a potential problem here is that they picked conservation areas as study sites – so the question here is about whether sites are more likely to become conservation sites (and hence be sampled) if the wildlife there is rising/declining? Also, were samples taken at the same time of the year towards the start and end of the trial period – given that most insects die off in the winter? They did try controlling for local climate, but it’s not at all obvious that the dependence will be linear on the averages they pick to represent it. I could probably go on, but I’m not interested enough to bother. Treat with caution.

  37. In addition, I know that there used in my youth to be far more flies around because getting them buzzing around the house was practically a daily occurrence

    Remember butchers’ used to use flypaper?

    Put me right off my sausages.

  38. @Chris Miller “Perhaps it’s the same thing that caused the bees to die out and no crops to be pollinated, leading to the worldwide famine we just had.” – well yes, and of course honey became unavailable in the shops, and twenty quid a jar if you could find any….

  39. Plenty of flies in France, just driven Calais to Metz, car is plastered with them. Perhaps this tells us something about the French…

  40. Arthur Dent: “To paraphrase: we have done some random sampling of insects at a variety of sites at different times, direct comparison over time is not possible so no trends can be measured. However if we construct an unverified mathematical model we ca convincingly show the end of the world is nigh”

    Not at all, read what they say: “the data do permit an analysis at a higher spatial level, i.e. by treating seasonal insect biomass profiles as random samples of the state of entomofauna in protected areas in western Germany”

    They are treating as random samples from the same kind of area across time and seeing a decline. This is a valid approach to the data they have, not an “unverified mathematical model”.

    I think the conclusion that flying insect numbers have decline will be unsurprising to anyone that works in the countryside on a daily basis. Insect traps that I personally monitor do not catch the same numbers today that they did even 10 years ago. On a summer day my windscreen is not as totally covered in squashed insects and bugs as it was driving around in the 1990s. These effects are not hard to see.

    The magnitude of change and the reason are unknown, but there is certainly a decline. And it is something we should not pretend is not happening. Working out why it is happening is worth doing.

  41. Indeed I quoted that specific paragraph. However you cannot take samples from a small but structured sampling programme and then interpret the data,after the fact, as if it was coming from a set of random samples. Statistics can’t be mangled in that way. Their samples were patently not random ones

  42. Don’t know about Germany, but here in the UK we’ve had a bumper crop of just about every type of fruit. Something’s been pollinating them!

    Germany had a record fruit harvest in 2014, then a a significant drop to 2015 and a similar harvest to ’15 in ’16.
    http://www.fruitnet.com/eurofruit/article/171313/german-produce-stats-highlighted

    There seems to be a big drop from ’16-’17 but these are only prelim figures and I don’t know whether one can expect them to go up.

    https://www.destatis.de/EN/FactsFigures/EconomicSectors/AgricultureForestryFisheries/FruitVegetablesHorticulture/Tables/2_4AreasQuantitiesHarvested.html;jsessionid=4EC5E2DC64E45C5E3A1E3E328C9ADCD8.InternetLive2

  43. I think it may have been Rutherford who said that “if you have to use statistics to interpret the results of your experiment you need to redesign the experiment!”.

  44. @ Pcar
    Thanks
    The zoom fundtion in my browser is hidden so I didn’t know it was there until I searched.
    I think the arguments for a major decline look pretty convincing – the like-for-like comparisons on 26 sites support the numbers produced by their complex model.

  45. Arthur Dent “Indeed I quoted that specific paragraph. However you cannot take samples from a small but structured sampling programme and then interpret the data,after the fact, as if it was coming from a set of random samples. Statistics can’t be mangled in that way. Their samples were patently not random ones”

    They don’t treat the data as if the samples are simple random samples and they do take account of how the data is structured. From the point of view of looking at total biomass in various kinds of protected areas the way they treated the data is absolutely fine.

  46. Anybody know of any population surveys of swifts, sand and house martins, swallows or bats that cover the same time frame and geographic locations ?

    There would be the validation of the 80% reduction in flying insects as a slam dunk.

  47. Living within a few hundred metres of a river, and with plenty of greenery around (and a restaurant downstairs) I was staggered by the near-absence of just about any type of insect this summer; until my neighbour pointed out the family of bats that now lives in our roof space.

  48. My unscientific survey shows that there are a lot more flying insects in the city of Pittsburgh than there were 15 years ago. Back then many local governments(We have 130 individual municipalities plus other level governments in Allegheny county for total of 197) still sprayed poor neighborhoods with pesticides. Outside of the city there appear to be far fewer, judging from the streaks on my windshield.

Leave a Reply

Name and email are required. Your email address will not be published.