We had a guest post by Bob Vislocky that covered this topic well a month ago. but here is a good write up reposted from Paul Homewood’s blog.
Study Claiming Insect Decline Due To Global Warming Is Based On Faulty Temperature Data.
January 20, 2019
By Paul Homewood
h/t Joe Public/Dave Ward
Many thanks to some real Sherlock Holmes work by Joe and Dave.
The article refers to this study by Brad Lister last year:
Significance
Arthropods, invertebrates including insects that have external skeletons, are declining at an alarming rate. While the tropics harbor the majority of arthropod species, little is known about trends in their abundance. We compared arthropod biomass in Puerto Rico’s Luquillo rainforest with data taken during the 1970s and found that biomass had fallen 10 to 60 times. Our analyses revealed synchronous declines in the lizards, frogs, and birds that eat arthropods. Over the past 30 years, forest temperatures have risen 2.0 °C, and our study indicates that climate warming is the driving force behind the collapse of the forest’s food web. If supported by further research, the impact of climate change on tropical ecosystems may be much greater than currently anticipated.
Abstract
A number of studies indicate that tropical arthropods should be particularly vulnerable to climate warming. If these predictions are realized, climate warming may have a more profound impact on the functioning and diversity of tropical forests than currently anticipated. Although arthropods comprise over two-thirds of terrestrial species, information on their abundance and extinction rates in tropical habitats is severely limited. Here we analyze data on arthropod and insectivore abundances taken between 1976 and 2012 at two midelevation habitats in Puerto Rico’s Luquillo rainforest. During this time, mean maximum temperatures have risen by 2.0 °C. Using the same study area and methods employed by Lister in the 1970s, we discovered that the dry weight biomass of arthropods captured in sweep samples had declined 4 to 8 times, and 30 to 60 times in sticky traps. Analysis of long-term data on canopy arthropods and walking sticks taken as part of the Luquillo Long-Term Ecological Research program revealed sustained declines in abundance over two decades, as well as negative regressions of abundance on mean maximum temperatures. We also document parallel decreases in Luquillo’s insectivorous lizards, frogs, and birds. While El Niño/Southern Oscillation influences the abundance of forest arthropods, climate warming is the major driver of reductions in arthropod abundance, indirectly precipitating a bottom-up trophic cascade and consequent collapse of the forest food web.
https://www.pnas.org/content/115/44/E10397#F1
Sounds like an open and shut case eh?
Lister even adds these temperature graphs to his paper:
Given that, as the paper itself admits, warming in the tropics in theory should be much less than elsewhere, claims of a 2C increase since the 1970s did not pass the sniff test. Fortunately Joe Public decided to go away and check the actual data used by Lister.
His findings should alarm anybody who believes in the integrity of science and peer review.
Let’s take a closer look at those graphs.
All the temperature data used comes from just two sites, El Verde and Bisley (not Bisely). But Bisley data only begins in 1993, so the comparisons with the 1970s rely solely on El Verde, the data for which is sourced from here.
http://evfs.ites.upr.edu/research/meteorological-data
And this is what the Station Metadata has to say:
http://luq.lternet.edu/data/luqmetadata16
In short, the data Lister uses prior to 1992 is worthless, and were substantially understated in comparison with those that followed.
If we begin the chart in 1992, we find that temperatures have actually been dropping, and not increasing.
The metadata states that temperatures showed an abrupt increase in 1997, as the adjustments were ended. However, this step up is not apparent on Lister’s graph. Rather, the step up is in 1992. This seems to indicate that Lister uses the unadjusted data.
The metadata also links to the dataset:
http://luq.lternet.edu/data/luqmetadata16
And this provides some handy graphs.
Below is the chart of annual temperatures from 1993 to 2013. [The data only runs to Feb 2014]
This shows the same pattern of declining temperatures since 1992. We know that the data since 1997 is fully reliable, and this too shows a declining trend. Note too that Lister’s graph indicates temperatures since 2013 have been lower still for three of the last four years.
https://climhy.lternet.edu/plot.pl
This still leaves us the problem of Bisley, where data starts in Feb 1993. CLIMDB only have data till Nov 2010, so full annuals are only available from 1994 to 2009:
https://climhy.lternet.edu/plot.pl
For some reason, Bisley shows a rather different picture to El Verde, although there is no obvious warming trend. Given that El Verde and Bisley are only a few miles apart, it is obvious that the divergence between the two is due to dodgy data and not real.
To sum up, we have a paper which makes bold claims that arthropods have been declining at an alarming rate since the 1970s, and that the cause is climate warming.
Yet these claims are based on long term temperature data, which, according to the organisation that actually maintains the data, is not reliable and should not be used for long term trends.
The only reliable data covers the period since 1992, and this shows declining temperatures. Even this dataset is not consistent with the Bisley one.
Clearly the whole study is worthless, and the paper should be withdrawn.
There are some alarming facts about all of this:
1) Why did the researchers not suspect that the temperature data looked hopelessly wrong at the outset?
2) Why did peer review not do the basic checks that I did?
3) The study carries out some mindbendingly complex statistical analysis, linking arthropod decline to rising temperatures. But how can this analysis have been robust, when the temperature data was hopelessly wrong?
The conclusion is that the faulty temperature data matched the researchers’ expectations of climate warming, and consequently they never bothered to crosscheck. It would after all have been extremely simple to have asked the people who maintain the data.
Whether or not arthropods are in decline I have no idea. But by blaming non existent climate warming, there is a very real danger that the true cause is being missed. Indeed, looking at those graphs, it may well be climate cooling that is responsible.
I plan to contact PNAS, who published the paper, to request that it be withdrawn.
FOOTNOTE
I have also crosschecked the temperature data from 1975 to 1991, available here. This correlates with Lister’s graph.
So there are no actual temperature records covering the period in question? A minor little obstacle.
What the abstract did not go into was the collection procedure used in the 1970’s study they were comparing to. Was it the same, or different, from the procedure used in this study?
Their methodology appears to have a few bugs.
When we’re they there doing their data gathering? Their paper was submitted for publication in January 2018. In September 2017 Hurricane Maria ran through that very area at Cat3 strength. A few weeks prior to Maria, Irma also passed through the area. If they arrived after Maria had passed, it would leave little to wonder exactly why their was no birds or bugs.
No bugs missing in my area. They are, in fact, nastier than ever at times. The Little Black Ants decided to take up residence somewhere warm and spend a good deal of time harassing my kitchen sink, in search of water.
Obvious question here: If PR bugs are seemingly declining, did it occur to Mister Lister that maybe they went elsewhere, and the birds, since they eat bugs, followed them? That would certainly be the first thing I’d check, because the PR is a rather large chunk of landscape. Maybe the locals sprayed so much bug spray that those people killed them off.
If this Lister feller got a grant for this study, I’d want my money back.
I suspect insects species go thru the same up & down cycles like higher animals do — like the old rabbits/coyote or elk/wolf population vs time graphs.
“driving into the forest” sounds a bit suspect. Maybe the local fauna have learned to stay away from vehicles.
Any chance the area was affected by one the frequent hurricanes passing over?
After Cyclone Yasi there was a significant change in the bird life of my bit of coastal forest. A number of regulars stayed away for months, but the kookaburras immediately capitalised on the incident, cleaning up everything that had been blown out of the tree-tops. Including a few frogs of a species that was supposed to be extinct.
(Extinct = living above the level that researchers can reach.)
Two data points, almost 50 years apart, with no investigation into the events prior to and coincident to the two readings.
No true scientist would attempt to generate a trend from such scanty data.
However such misconduct is par for the course in climate science.
Then to cast the Blanket Blame on GW without eliminating other possible causes is sloppy at best
The integrity of science and peer review cannot stand in any way between a scientist and his funding and grant money. Computers, graphs and statistics stand near at hand to assist the favourable outcome for the grantee.
Was any spraying done to reduce insect activity ?
Mosquitos aren’t the only insects killed when spraying is done …
It’s funny how we are told good harmless insects are being destroyed by global warming.
but the bad pest insects will increase in numbers and decimate our crops.
Any insecticide use by the government or ? ?
Next, we need a replication of the measurements of the populations of insects, birds, lizards and frogs. The entire climate file is an apriori collection of nonsense. The publish or perish drive plus the unbelievable 10s of thousands of practitioners, results in “reasoning” what sorts of things could be effected by a rise in temperature thereby creating an exhaustive list of topics. Then studies are concocted to show that something terrible is happening. The proof of this approach being used is to be found in the eclectic hodge podge of topics. Remember when the Pause came to be accepted that there were 55 explanatory causes for it, all of which supported CAGW! Natural variation was not among them in any real way, because long ago the synod of climatescientist whips had ruled climate variation to be tiny and anyway cancelled out. This why a Canadian mining engineer has been able to have premier climate papers retracted or withdrawn.
A 98% decrease in insect numbers from a 2C increase in high temperatures doesn’t pass the laugh test, much less the smell test.
And that’s assuming they could actually get good temperature data, which they didn’t bother to.
well….the paper was very effective at what it set out to do
google…….Puerto Rico insects
…it’s now official….all over the internet….hyperalarming, massive insect loss, devastating, decimated, mass extinction…..all the usual flash card words
which was its only purpose. Alarmism.
Well done — the devil is always in the details — and activist scientists and those with preconceived “causes” never carefully check the details if the data “confirms” their bias.
Nothing will happen unless a full-tenure NAS member submits a formal rebuttal to the Editor. The NAS is a good-ole boys club. It only listens to its own members while pontificating and using fast review methods to get papers out quickly. A NAS member uses PNAS as a review-light source to get papers out quickly, even if junk.
The PNAS Editor in Chief or Sr Editor would then be obligated to submit the rebuttal to the original authors for a counter-point, and then publish both. If the original authors could not adequately defend whatever the rebuttal shows, then the Editor would be strongly pressed by ethics to publish a “Letter of Concern” on the original work or even outright retract the paper if the original authors remained silent.
All you found is that a 0.4 correlation coefficient is called close to random noise, and shows no publishable correlation. The P-value has not meaning in the face of that kind of correlation and confuses a population study around a mean with a linear least squares line, which is a very different thing.
Pet peeve warning!
Many ‘scientific’ papers combine the following 3 contradictory statements:
1) the science is settled (and has been for some time)
2) new research shows it’s worse then we thought previously (when the science was already settled)
3) more study is needed (to finally understand what tf is going on)
These 3 statements are completely contradictory to each other but are in almost all climate ‘science’ papers.
When reviewing a paper the first check should be that it is internally consistent or as a minimum one part should not contradict another part.
How these statements get through peer review is completely beyond me.
Any self respecting reviewer should send the draft straight back as these internal contradictions make the paper non scientific almost by default . Furthermore, any journal that publishes these papers is also no longer scientific.
Might be an idea to ask for retraction of all paper which contain these three points. That would clean up the scientific literature quite nicely 🙂
All the best, have a great day!
WWW
**How these statements get through peer review is completely beyond me.**
Read the Wegman Report on Climateaudit.org
Excellent plan and a very valuable disputation of these claims.
Insect collections and worse, alleged insect collectors…
My understanding is that they cordon off a small workable square; e.g. 1 meter by 1 meter and count all of the insects they find.
It is also my understanding that they put a net under a tree, then spray the tree with insecticide to kill all of the insects, allowing them to drop to the net.
Both methods are dependent upon honest site surveys and serious attempts; both methods are also susceptible to wild carousers celebrating their I&M vacations in Puerto Rico. (I&M is an abbreviation for ‘intoxication and malingering’.
It is doubtful that Puerto Rico’s insects are in decline while Virginia forests are not suffering insect declines.
“For some reason, Bisley shows a rather different picture to El Verde, although there is no obvious warming trend. Given that El Verde and Bisley are only a few miles apart, it is obvious that the divergence between the two is due to dodgy data and not real.”
Think about that. then think about it again.
I dont know why these guys used data from sensors located on rooftops / 20m towers
Good catch.
In the end I suspect they will look at the other temperature data
Tmax is only up about .8C since 70s
“I dont know why these guys used data from sensors located on rooftops / 20m towers”
I think because the site is in the actual forest they are studying.
San Juan has a record going back pre-1900. It shows quite a lot of warming.
San Juan actually shows a slight cooling trend in Max Temp for the period of study (1976-2013), as does Pico del Este which is a coop site located within the forest. See my write-up on this same topic from last month for the graphs:
https://wattsupwiththat.com/2018/12/21/debunked-insect-decline-in-rainforest-caused-by-climate-change/
Thanks, Bob, for showing Nick has a tendency to omit contrary evidence; here he ignored data concerning the topic of note and confused people with data (?) covering a different period. Climate science at its best.
The GHCN station worksheet for San Juan is here. It shows (unadjusted) very strong warming for TAVG – over 2° since 1890. There was a burst of warmth in the 1970’s which would bring down the trend for the period since. Homogenisation also reduces the trend.
May I also note that, for Bisley, the data begins in the particularly cold post-Pinatubo years? Starting the data just 4 years later makes the trend much less alarming.
The metadata states that temperatures showed an abrupt increase in 1997, as the adjustments were ended. However, this step up is not apparent on Lister’s graph. Rather, the step up is in 1992. This seems to indicate that Lister uses the unadjusted data.
Nope. Are we looking at different graphs? I see no step up in 1992, the step up is more in 1995-1997. He is using the adjusted data. Which is still wrong, of course, and more so since the metadata specifically alerts against it.
See image:
http://www.elsideron.com/ElVerde.png
Once more, and again and again, we see little to no proper use of error bounds or confidence limits, conducted with proper methodology and rigor.
Proper, classic error analysis is the main way that suspect data can be recognised early and easily. It is one of the main reasons for its existence. Lamentably, we see too much of the lip service style of error analysis, often merely the stats read off an Excel graph, that amount to no more than virtue signalling by error numbers. Look, my error numbers are low, so this has to be a good paper.
For the present paper, a proper, classic error analysis would have compared the past temperatures at these 2 sites with regional temperatures. A difference would have been found. If the reason for the difference was not explained, the error bound should have been wide enough to incorporate most of the site temperatures and the regional temperatures. A reader of the paper might then have remarked. “Look, the temperatures are quoted at 2 s.d. being +/- 2 deg C” as the case may be, an immediate alert that not all was well.
At dome future time when the penny drops, there will be major benefits to the credibility of climate science once proper, classic error analysis becomes customary. Geoff.
2011 and 2013 when the low arthropod numbers was recorded were apparently exceptionally cool years. Arthropod activity is very temperature-dependent, so it seems to me that the cold is a quite plausible explanation.
Apart from the other criticisms above,
If they can’t write good English I doubt they can do good science.
And, appearing cynical, but knowing how most humans behave in the real world, I’ll bet dimes for dollars they don’t even still have the data from when they visited in the 1970’s.
It doesn’t pass the smell test on several counts.
Plant insects have temperature ranges they do well in. For purposes of the following I’ll refer to their current niche ideal as “ambient” temperature & experimentally increased comparison as “elevated” temperature. My thinking is the following addresses some of the O.P. theme of insect population potentially being under change.
Because elevated CO2 (eCO2) is a topic at WUWT I am going to give one example of a plant feeding (sucks from plant vasculature) insect (specific rice leaf hopper) when current level of CO2 is factored at ambient or elevated temperature. This will be cross compared when eCO2 is factored at ambient or elevated temperatures.
Adult weight of this rice leaf hopper was greatest under eCO2 with ambient temperature (temp.) ; but adult weight was almost the exact same as under CO2 with ambient temp. The lowest adult weight (for this specific insect, I am not asserting for all kinds of bugs) was under eCO2 with elevated temp.; & the decrease from maximum weight was a lot.
Female longevity of this rice leaf hopper was greatest under CO2 with ambient temp. (she likes it like that). The 2nd longest female longevity was under eCO2 with elevated temp.; while longevity was similar under both CO2 with elevated tempe, as well as under eCO2 with ambient temp.
Number of eggs laid by a this rice leaf hopper was greatest under eCO2 with elevated temp. ; in fact almost double the number of eggs laid under the least case. The least number of eggs laid/female was under ambient CO2 with ambient temp. Both under CO2 with elevated temp., as well as under eCO2 with ambient temp. females laid the same number of eggs.
Nymph (growing) rice leaf hopper stage duration was completed the fastest under eCO2 with elevated temp. All other CO2/eCO2 + temp. resulted in the same number of days growing as a nymp.
In other words, lineal extrapolations are not always correct when applies to living organisms.
I’ll try to elaborate.
For example: eCO2 alters the leaf carbon ratio higher relative to leaf nitrogen level & we see evidence that chewing insects (“herbivore” insects) tend to eat more in order to meet their metabolic need for protein (nitrogen based). This is not always (not lineal) happening at every feeding episode for every insect, nor at every stage of life for the insect.
Some insects will self-select their meal (sometimes) to get a ratio of carbon to nitrogen that is different than what they “needed” before (ex: egg development). They will also balance their intake of carbohydrate at certain times based on the water content of the carbohydrate food source to balance the amount of water intake involved; they can only “load”their system with so much water at a time.
Aphids are sucking insects that bedevil plants. Yet not all kinds of aphids feed heavier ( a lot of kinds do of course) on eCO2 grown relatively higher carbon to nitrogen ratio leaves.
Aphid insects are not chewing into the leaf cells to ingest nutrients; they are sucking in dissolved “sugars” (& amino acids). In other words aphids are already taking a leaf compound awaiting transport out. The leaf has this (aphid’s nutrient fluid) whether CO2 elevated or not, so the aphid (& other sucking insects, like the cited rice leaf hopper) do not “eat” more (unlike “herbivore” insect munching) just because out in the leaf tissue eCO2 boosted the % of carbon.
How many insects do you see in winter? How many do you see on a hot summer day?
Guess what… the very same study counts more insects in the hot July than in the merely warm January. But somehow the author blames the multi-year reduction on a hotter climate. Go figure.
Didn’t Puerto Rico just experience a devasting hurricane last year, or did I imagine it? Seems to me it would matter immensely if his visit was before or after. I’m sure the hurricane must have decimated the wildlife.