This month’s open thread. Current topics are focused on the laughingly bad Daily Mail article by David Rose, the fallout from the Wall Street Journal’s latest regurgitation of why no-one should ever do anything ever. And perhaps someone might want to audit some of David Whitehouse’s arithmetic and reading comprehension…
Or anything else. Within reason.
I am hoping someone here can help me.
Recently, I’ve seen quite a number of “skeptic” bloggers using an argument based on the null hypothesis against climate change – “the null hypothesis is that global warming is natural”. The implication of this is clear – they’re arguing that scientists have to PROVE the anthropogenic cause before they’ll accept it. This is surely pretty well impossible.
Based on my recollection of my own research from well over 20 years ago, the null hypothesis is specifically related to a statistical test in a scientific investigation and cannot be applied to either prove or falsify a theory which is based on many lines of evidence. Clearly, the null hypothesis can be usefully employed to test some predictions based on a theory and this may be where the misunderstanding has arisen.
The only obvious reference to this that I’ve found is the following from Coby Beck:
http://scienceblogs.com/illconsidered/2006/03/null-hypothesis-says-it-natural.php
Can anyone point me to anything which expands on Coby’s argument?
Thanks
Re sea level rise:
If anyone knows of any major new studies now being undertaken on sea level rise, please give me the details.
When deniers claim an ice age was being predicted in the 70’s, I remember reading here at RC about a particular researcher who was somewhat perplexed over the fact that his research indicated it should be warming. As I remember, he theorized that it was aerosols that were buffering the warming but he believed that in the next several decades the aerosol cooling effect would finally be overwhelmed by CO2. Of course this researcher got it right. I’ve been trying to find out who he was. Can anyone help?
> lame
arithmetic
Does anyone know of any update of this graph?:
http://maps.grida.no/go/graphic/temperature-trends-and-projections
It’s the reconstructed temperature over the last millenia plus the observed temperature plus the IPCC projections, so that you get a long-term perspective of future projections. But this one is based on IPCC 2001. ¿Is there a similar graph with the IPCC 2007 data? Thanks!
The null hypothesis is NOT that “natural” variation is true. The null hypothesis is that there is no significant trend. For example, let’s say human-emitted aerosols and carbon precisely balanced each other in a particular set of measurements. There would be no trend and the null hypothesis of “no warming (or cooling) would be accepted for that particular test. That would not at all mean “natural variation” is true.
The deniers here are trying to make rather more out of the term “null| than is contained within it.
jgnfld @ #106
Yes, that’s my understanding too. The problem is that it’s quite difficult to find any authoratative sources which tackle this issue as it relates to attribution in climate change. Most sources simply describe the null hypothesis as it would be applied in laboratory tests, where most variables can be controlled.
If the factor of 100 is the problem, then it’s really lame. David Whitehouse doesn’t specify any unit, and GISTEMP isn’t a measure of temperature anomalies, but an index. An index can be scaled by any factor without loss of information, if used consistently. You just need the correct unit if you want to use GISTEMP as an estimate of quantified near-surface temperature anomalies.
Paul Briscoe,
One way of interpreting the denialist’s straw man is to ask how likely it is to see a 35 year rising trend in temperatures with the slope, etc. that we currently have, along with there being no increase in total solar irradiance, etc. If you look at the reconstructed temperature series, the current epoch stands out like a…well, hockeystick.
I suppose even getting these idiots to admit it’s warming is progress. Now we have to get them to accept conservation of energy.
Paul Briscoe,
In attribution studies, the null hypothesis is traditionally one of no human influence on a particular climate variable, and thus when something is reported along the lines of “the greenhouse gas increase response very likely contributes to greater than half the total observed warming,” it means that the null hypothesis is rejected at the 10% significance (or sometimes even more confident) level. Note that this practice can bias successful attributions toward well-modeled or well monitored regions of the globe and also variables of interest (e.g., temperature vs. precipitation). Some like Kevin Trenberth have recently suggested to have a null hypothesis of human influence given the massive accumulation of evidence, thus putting the burden on those who say there isn’t a human influence, but this is a large departure from normal practice.
A lot of the attribution studies out there are actually quite conservative in the attribution (e.g., by using no prior assumptions about the response to a particular forcing) and a wide range of methodologies have been employed to declare a substantial influence from humans with high confidence. Chances are that your skeptical correspondents don’t really have an idea themselves of what would constitute a “proof” that the null hypothesis is rejected, or just exactly what would convince them.
We don’t have an alternative Earth to sample and experiment on, so the next logical step is to use physics and look for patterns in space and time that can help guide statistical tests of attribution. When you do this, there is really no study that has withstood scrutiny that can explain the bulk of modern warming without the human influence.
By the way, the AR4 has a pretty good discussion of attribution, and I am under agreement to not talk about it, but AR5 looks to be even better. But a number of recent papers do a good job too (some by Hegerl are good, eg., Hegerl and Zwiers, 2011, Hegerl et al 2011, ERL, or the 2010 “Good Practice” paper on attribution studies; Reto Knutti typically does a good job on some of this stuff as well).
79, Barton Paul Levenson, you haven’t been here much lately, so welcome back.
Hint: You need 30 years or more to establish a climate trend.
James Hansen gave his first AGW warning in 1988 (or thereabouts, perhaps his Congressional testimony was not his first.) At that time, what trend based on 30 years had been “established”? Has it been confirmed by subsequent events?
One trend that had been “established” (i.e. presented in public) by 1988 is the empirical fit of the linear trend plus cosine curve, and that has been “confirmed” within rough limits by the subsequent rise and approximate plateau. The limits were “rough”, as I termed them, due to the combination of random variation in the observations and the fact that only 2 cycles of rise and plateau had been observed.
46, Grant. Thank you. I didn’t notice it earlier.
There’s a whole lot of garbage being put out on Dr. Mann on DotEarth. I find it wholly despicable that people have been made to believe that he has “persecuted” phony skeptics and that all the personal threats he and others have received, and the unrelenting quasi-legal attacks, are somehow equivalent.
http://dotearth.blogs.nytimes.com/2012/02/02/a-shameful-attack-on-free-speech-by-a-group-claiming-to-speak-for-coal-dependent-workers/
I know this is old news, but the posse is in full cry there.
Andreas @108
GISS may call itself an “index” and you may infer whatever you will from that ‘calling’. But to state that GISS “isn’t a measure of temperature anomalies” begs the question “So what does it measuring then?”
And do be advised that a non-answer on this one would rate rather high on the Whitehouse Scale of stipidity.
On the “null hypothsis”, phony skeptics are fond of presenting phony science. Toss in a few sciencey looking terms, find a little language that makes the moon seem like it might be made of green cheese, and voila.
If anyone manages to prick their balloon, they immediately ask for data and evidence, but *their* data is “real”, as real as PR and money can make it appear, including a whole lot of insider links that add up to perverted hot air, the opposite of skepticism.
While I’m sure it is useful to real scientists to get a real response to this gobbledygook, it has limited utility in moving the conversation forward. It’s effect, even if demonstrably wrong, is to use up the time and energy of innocent bystanders who would like to help and skilled scientists who have better things to do with their time, such as finding solutions and advancing the state of knowledge.
Tiresome.
Judith Curry is at it again, arguing that we can’t trust the global record at all:
http://www.skepticalscience.com/news.php?p=3&t=104&&n=1261
Thanks, everyone, for your helpful comments. Thanks especially to Chris Colose for pointing me to AR4. The problem I can foresee with this is that deniers will dismiss it because it relies heavily on models, which they don’t accept!!!
Sadly, Susan Anderson is probably correct. I think an effective response would be useful to have for those of us who attempt to support the science, but it is ultimately futile where “phony skeptics” are concerned.
SM 111: James Hansen gave his first AGW warning in 1988 (or thereabouts, perhaps his Congressional testimony was not his first.) At that time, what trend based on 30 years had been “established”?
BPL: Rising temperatures. Temp. data was available back to 1850, and 1880-1987 is N = 138 years. Of course, Hansen probably relied on the NASA GISTEMP series, which only goes back to 1880 — N = 108 years.
Paul,
And she is really, really embarrassing herself. Almost every point she brings up has been demolished either in formal or informal studies. As a professional she really should be aware of this.
Re #114, MARodger:
“But to state that GISS “isn’t a measure of temperature anomalies” begs the question “So what does it measuring then?””
It measures different kinds of temperatures (2m air temperatures, SST, at some points nothing) and blends them together. The result is an index, not a property of a real object. It is useful for estimating near-surface temperature anomalies, but it doesn’t measure them.
From skeptical science:
“As a general announcement, a spammer (jdey123/mace) has been masquerading as Judith Curry on this thread. ”
That makes more sense. The comments struck me as too ridiculous, even given Curry’s penchant for saying some absurd things.
Paul @117, Susan @115, I’m in general agreement. Sometimes, I try to use humor to get the point across.
You go to buy a cut of meat from a butcher, but you notice that his thumb is on the scale when he weighs it. You object. He counters “You’re accusing me of cheating! The null hypothesis is that my thumb is carefully placed on the scale so as not to change the result. Besides, you have no evidence of how much a thumb weighs, when still attached to a live human.”
“No!” you reply, “The null hypothesis isn’t that the laws of physics have been abolished in your case. I don’t know how much your thumb weighs, but it weighs something more than zero, and I don’t want to pay for it.”
Says the butcher, in a huff, “You need to keep in mind that I’m a skilled professional. In fact, I’ve got a study published in the Journal of Practical Adhesives showing that a butcher’s thumb is sticky, on the average. It’s capable of pulling the scale upward. Here, I’ll give you a free reprint, courtesy of the Heartmeat Institute.”
I could go on, but Monty Python probably does the routine better – with a dead parrot.
#119
That ‘Curry’ was fake.
http://www.skepticalscience.com/news.php?p=3&t=104&&n=1261#73531
Andreas @120
Well it’s back to school for me. I never knew there were “…different kinds of temperatures.” All these years I have obviously misunderstood the second law of thermodynamics!
What you seem to be trying to say is that because GISSTEMP is a combined measure (and I will insist that it is the product of measurement and thus a “measure”) of Land-Surface Air and Sea-Surface Water Temperature Anomalies, this invalidates it and makes it but a useful ‘estimate’, “not a property of a real object.”
You add to this argument the lack of ubiquity of such measures (likely in a geographical sense) as though this add further weight to your contention.
However I was taught that all measurements are ‘estimates’ and it remains so regardless of how up-front and personal you get with the thing you are measuring. And even a simple headcount can be subject to error.
I thus dismiss your argument of GISSTEMP not being a measurement of the properties of a real phenomenon.
Further, making recourse to ‘Ah but this is an index’ is a simple & worthless digression (even if GISSTEMP uses the ‘index’ word itself).
Given then that GISSTEMP is a measure ‘useful for estimating’ global temperature (as is HadCRUT3), why do you legitimise the erroneous scribbling of Whitehouse, which was you message @108?
@ John P Reisman #70 and 84# (Sorry for doubling up. Didn’t think my first post went out)
Thanks for the link. Just what I needed.
I am replying to a sceptical letter (Global Warming cancelled!!) in my local newspaper.*
Bloody freezing in the Ukraine (200 deaths reported from cold – mostly “street people” I gather) while it looks like 15C above normal around the American Great Lakes and Hudson Bay. Skating on Lake Eire cancelled?
Theo H
* My local/regional newspaper covers Exeter in Devon, England, and the UK Met Office is based in Exeter. So I hope I get my rebuttals right. :)
@120…and I suppose you think a ruler measures something “real” as opposed to being an index.
ALL measurements are indices of one sort or another and ALL measurements are measured with error. Counting tulip bulbs is measured with error as if you get enough people to count a bag of them, you will find that not everyone comes up with the same number, or records it wrongly, or something such that errors occur in the measurement process.
This is such an elementary fact that you simply cannot go further in science until you internalize it.
Bird numbers plummet around stricken Fukushima plant
118, Barton Paul Levenson: BPL: Rising temperatures. Temp. data was available back to 1850, and 1880-1987 is N = 138 years. Of course, Hansen probably relied on the NASA GISTEMP series, which only goes back to 1880 — N = 108 years.
And based on the full data set, what “trend” was Dr. Hansen warning of? The linear trend or the linear trend + cosine? His warnings seem to have departed from both of them.
Today (3/6/2012) the American comic strip called Doonesbury is starting up a story designed to ridicule climate deniers that make up facts.
http://www.doonesbury.com/
It will be interesting to see how far they go with it.
Does anyone how to contact whoever is moderating over at Joe Romm’s? I was a semi-regular commenter there for years, but for the last several months my comments simply never appear. I certainly didn’t post any denialist crap to earn a ban – I’m a climate realist (which, as Neven would say, means “alarmed”). But I’ve somehow been effectively blacklisted.
Thanks!
Susan #113, there seems to be an intensification of a crafted tu quoque strategy, to effectively distract attention from the science & the awareness of the climate threat, thus increasing the threat to humans of this & future generations.
One piece: WSJ signatory Bert Rutan’s views about aliens (see Russell #71) as told to Wired, presumably on the record, and the (seemingly popular in some circles) new-age-leftie how-to-fix-our-problems-&-be-sustainable-in-future film Thrive about aliens.
“We will be judged by those who come after us, both by what we did do and what we didn’t do, in the time given to us.”
sidd at 48:
That transformation just converts the data to mean anomalies and expresses them in units of local standard deviations. That only shifts and squeezes the distribution; it stays positively skewed, principally because rainfall can’t be negative. The longer the rain duration analysed the nearer the distribution will be to gaussian (that central limit theorem again), e.g. most annual rainfalls are nearly normally distributed (not in arid areas).
144 x 72 x 2.5 deg, so that is global, land only? Very interesting.
GlenFergus & sidd — Good generalizations of normal are the stable distributions; indeed the normal distribution is a (rather special) instance of a stable distribution. There is at least one study which fits certain rainfall statistics to the general one-sided stable distribution; the important coefficient in that study indicates a nearly normal, but one-sided of course, stable distribution.
The stable distributions generalize the c3entral value thereom in such a way that any given stable disribution can be thought of as the sum of stable distributions with the same major parameter. These distribut6ions, for me, seem far better motivated than othr, eaarlier attempts to fit well-known distributions, gamma and so on.
Mr. Fergus:
these include both land and ocean
sidd
http://www.nasa.gov/mission_pages/NPP/news/missing-energy.html
“10 years of data — 2001 to 2010 — from NASA Langley’s orbiting Clouds and the Earth’s Radiant Energy System Experiment (CERES) instruments to measure changes in the net radiation balance at the top of Earth’s atmosphere. The CERES data were then combined with estimates of the heat content of Earth’s ocean from three independent ocean-sensor sources.”
Thanks David, interesting. I have cause to want to do some more stuff with annual rainfall distributions shortly, so shall have a close look. Stable distibutions are generally fat-tailed, no? But annual rainfall distributions aren’t necessarily. I’ve been thinking just Pearson IV.
BTW, for shorter durations, the Australian BOM is currently revising all our design intensity data; no small task over a whole continent. They’re just fitting LP3s to partial series, same as was done 25 years ago: http://www.ncwe.org.au/arr/p1.html.
G.
Am I missing something??
There is a lot of debate above as to whether the last 15 years temperatures show a flat trend, or a slight increase – and whether the trend should be taken over 30 years rather than 15.
Isn’t the point that in a period when humans have generated more CO2 more quickly than ever before (india, China, brazil, etc, coming on stream big time) we have seen little increase in temps that reflects that additional output of CO2.
As I understand it, there is a general acceptance that the planet continues to warm “naturally” – so, if man made CO2 isn’t resulting in big increase in temps – then the “Natural” warming can’t be there at the moment?
Do we understand what is happening at all?
James.
Look at the trend from 1977-87. Look at the trend from 1987-1997. Now look at the trend from 1977-1997. Temperature data are noisy. If you want to spot trends you need to take periods longer than 30 years.
Again, where the hell do you get that the planet is “warming naturally”? According to the phase of Milankovich cycles, we ought to be cooling. According to the fact that we just came out of the deepest solar minimum in >100 years, we ought to be cooling now. And given the intensity of the brown cloud coming off of Asia, we ought to be cooling now. We are not. We are still warming. Get the science and forget about talking points.
James wrote: “Do we understand what is happening at all?”
Who are you calling “we”?
The multiple false assertions in your comment suggest that YOU do not, in fact, understand much of what the world’s climate scientists DO understand.
I’m constantly amazed by the obscure and unlikely outlets that ‘sceptics’ can find to promulgate their views. The latest example being the Catholic Herald:
“Is the ‘anthropogenic global warming’ consensus on the point of collapse?”
>Ray Ladbury @ 109: “Now we have to get them to accept conservation of energy.”
I agree, and not just as sarcasm. I might add that basic physics such as conservation of energy sometimes seems rather
attenuated in statistical climatology.
>Chris Colose @ 110: “Some like Kevin Trenberth have recently suggested to have a null hypothesis of human influence
given the massive accumulation of evidence, thus putting the burden on those who say there isn’t a human influence, but this is a large departure from normal practice.”
Might that nudge you in a Bayesian direction with physics assigned a prior of 1.0?
>Paul Briscoe @ 117: “The problem I can foresee with (attribution) is that deniers will dismiss it because it relies heavily
on (current data, paleo data, physics and “models” ie doing the calculation) which they don’t accept!!!”
Deniers aren’t news. Try to increase the number of people who get it as opposed to being lost in the fog.
> James @ 137: “Am I missing something??”
As best I can decipher from your comment, yes.
[Aside: You say “As I understand it, there is a general acceptance that the planet continues to warm “naturally” ” I don’t know where you get this.]
What you may be missing is planet earth and physics, notably the aforementioned conservation of energy.
First, as a land dwelling creature not that our land surface temperature is going up. See the Muller “BEST” result, and how some interpret the relentless rise as always falling. Next, recall that (1) land is only about 30% of our planet’s surface, (2) the oceans are more dynamic than solid ground, and (3) the earth as a whole is a dynamic physical system. The overall surface temperature pattern is increasing with slightly different annual variation than land alone and requires some attention to ocean dynamics. The largest ocean is the Pacific, and its largest area is the south Pacific and this is the home of the largest oceanic contribution to short term variation to overall surface temperatures, the El Niño Southern Oscillation (ENSO).
We’ll take a look at how ENSO works, but first here is some information to help you anticipate the importance of it. The term “global warming” is used in scientific literature to mean the increase in the global average surface temperature, but this is only a small part of the warming of the planet. About 90% of the warming goes into ocean heat content. Another chunk goes into melting ice – many cubic kilometers of it. Some seeps into the solid earth.
How does ENSO work? Atop the south Pacific there is a large pool of warm water know as the Pacific warm pool. Circling the globe on either side of the equator there are very important air currents known as the trade winds. The strength of the trade winds varies a lot from year to year. Why this happens is beyond the scope of this comment. When the trade winds are average, the Pacific warm pool is of average (and considerable) size. When the trade winds are weak the warm pool spreads over more of the ocean surface, leading to a warmer average surface temperature. When the trade winds are strong there is much upwelling of cold water along the Pacific coast of South America. Wind driven surface currents push the warm pool towards barriers to the west while spreading cooler water across much of the Pacific, leading to a lower average earth surface temperature. The warm pool is meanwhile compressed laterally and extended downward. During this time, known as La Niña, more than the usual amount of sea surface water goes down to mid depths and adds to ocean heat content. Thus there is more planetary warming just when the “surface temperature only” view indicates less. None the less the most recent La Niña year (2011) was much warmer than any La Niña of the whole previous century. That tells you that despite annual variation, global warming continues unabated.
> If you want to spot trends you need to take periods longer than 30 years.
The amount of variation in any particular data set determines how much of that particular data you need to look at, to say there’s probably a trend in that particular data set.
It’s a specific calculation, something taught in Statistics 101.
For annual (once a year) data, that means how many years you need.
If you were flipping coins, it would mean how many coin-flip events.
You can do this yourself. There’s a good high-school-level discussion here with examples:
http://moregrumbinescience.blogspot.com/2009/01/results-on-deciding-trends.html
I would like to see some discussion on Hansen and Sato’s latest papers, since together they are very scary. Almost “game over” scary.
1. “Paleoclimate implications for human-made climate change” at http://pubs.giss.nasa.gov/abs/ha05510d.html concludes that we cannot push temperatures more than one degree C above the Holocene average (pre-industrial) or we will set off rapid warming and sea-level rise from the Greenland and West Antarctic ice sheet.
This is based on an analysis of previous warm interglacials, arguing that globally they were only fractionally warmer than the Holocene peak, but with major polar amplification. Also, that sea level rise was very fast during these interglacials (several meters per century).
2. “Earth’s energy imbalance and implications” at http://pubs.giss.nasa.gov/abs/ha06510a.html implies that the negative aerosol forcing has been underestimated by most climate models, since their response function to warming is too slow. The consequence is that a large part of current CO2 forcing is being masked by aerosols.
Put these together and we have a huge, and apparently insoluble, problem. Point 1 requires an extremely rapid decarbonisation and carbon sequestration program, since we are almost at 1 degree of warming already. Point 2 then states that if we stop releasing CO2, then we will remove the aerosol masking (most of which comes from fossil fuel burning) and expose ourselves to the full forcing from CO2 already released. This would cause even faster warming.
So it’s “damned if we do” and “damned if we don’t”. Anyone see a way out of this? Does it mean that we are now forced to try geo-engineering to give ourselves some hope? Could we even afford to do that if we are rapidly decarbonizing at the same time?
Paul A 140’s link
Downtown flake city.
Dr Nick Bone @143
The Hansen & Sato paper (presently off line so this is from memory & the abstract) seemed to me a bit too pesimistic with the acceleration of sea level rise. They discuss doubling of the rise (every 10 years in the abstract, the paper considers other doubling periods) based on a short record of GRACE data.
I would argue that the 10 yr doubling rate of rise would soon become impossible unless it is the result of unmelted ice. From the present 1.5mm rise from melting, the 10 yr doubling would yield 24mm pa by mid century & the energy to melt that amount of ice starts to become massive.
That doesn’t mean sea rise will not hurt, just that without a massive source of icebergs waiting to topple into the ocean, sea rise is a more multi-century problem than a multi-decade one.
> “…sea rise is a more multi-century problem than a multi-decade one.”
If we continue with accelerated emissions a big problem before 2100 is likely. But Roger, always keep in mind that the attempt to burn all the carbon will be stopped well before 2100 by some combination of people power and climate shock. The faster people power grows, the less climate shock is needed. YOU must be part of the reason sanity prevails sooner rather than later. Yes, YOU.
Meteorologist Jeff Masters, interviewed by Christine Shearer at Conducive Chronicle (emphasis added):
This morning, NPR’s Morning Edition ran a report on the multi-year drought that is “ravaging” much of Mexico. Not a single mention that the Mexican drought is part of a world-wide pattern of such droughts. Not a single mention that increasing frequency and intensity of such droughts is a long-predicted result of anthropogenic global warming. Not a single mention of climate change.
‘sceptics’ … the Catholic Herald
It’s a conservative newspaper — not a publication of the church.
—
A bit on clouds:
“One potential geophysical mechanism for changes in atmospheric transmission from one solstice season to the next involves cloudiness. The possibility that clouds are linked to solar activity has appeared in the literature, although the concept is controversial and conflicting results exist (Svensmark, 1998, 2007; Sloan and Wolfendale, 2008). As noted previously, clouds over the South Pole vary erratically….
“When all 17 years enter the regressions, there is clearly no link between the standard deviations and the solar cycle….
“… mean irradiance ratios and their standard deviations show no linear dependence on year over solstice periods from 1992 to 2008. This indicates the absence of a trend in cloudiness during the duration of the measurements….”
There are variations and more to look into, it’s an interesting location for observations.
http://www.atmos-chem-phys.org/11/1177/2011/acp-11-1177-2011.pdf
Atmos. Chem. Phys., 11, 1177–1189, 2011
http://www.atmos-chem-phys.net/11/1177/2011/
doi:10.5194/acp-11-1177-2011
Solar irradiance at the earth’s surface: long-term behavior observed
at the South Pole
J. E. Frederick and A. L. Hodge
Department of the Geophysical Sciences, University of Chicago
11 February 2011
GlenFergus @136 — Well, the normal distribution is a stable distribution. :-) The one-sided stable distributions indeed rolloff in a heavy tailed manner. I opine that the stable distribution properties are suggesting something important about extreme rainfall events so perhaps your data series just are not long enough yet.
The study I found was from South Australia metorologists for a time series of Melbourne rainfall data.
re dale 103
James Hansen in 1976 published : Greenhouse effects due to man-made perturbation of trace gases.
http://pubs.giss.nasa.gov/abs/wa07100z.html