At RealClimate, we have more than once been accused of being imbalanced — criticizing those who would deny the basic science of climate change, while leaving inflammatory statements by what might be called the “environmentalist side” without comment. It’s not an entirely a fair criticism, because there is a world of difference between the willful obfuscation of science and the naive exaggeration of it. There are however plenty of silly, and sometimes outrageous, claims made – see e.g. the Telegraph on Jan. 3rd — and we probably ought to do a better job of calling these out, particularly when they show up in prominent places. So to inaugurate the New Year, I humbly offer a rant about a minor but illustrative example that I happened to notice because there was a link to it on Nature Reports Climate Change.
The subject of the linked article, in the British online newspaper The Independent, is the decline of various bird and butterfly species in England. The article, entitled Changing climate devastates UK species, reports that “insects in particular, and creatures that feed on insects…were sharply reduced in numbers” due to a “cold late spring, a wet summer, with few sunny days, and the long dry autumn….” Now I have no reason to doubt the accuracy of the claim that 2008 was a hard year for UK insects and insectivores. But this is weather we’re talking about, not climate. And while it is true that at least one prominent study shows that there has been an overall increase in rainfall in the latitude band that includes the UK, and that climate models reproduce this trend (see e.g. the Zhang et al. article in Nature, in 2007), one cannot, as we are fond of pointing out, attribute a single, or even several individual extreme weather events to “climate change”.
Indeed, Peter Stott, a co-author on the Zhang et al. study noted, in reference to 2007 (the wettest summer on record in the UK) that “This latest study cannot make the link between climate change and what we have experienced so far this summer.” Moreover, most projections actually suggest drier summers in the UK in the future, though with increased convection (so less total precipitation, but bigger rainstorms).
Another thing that bugs me about the Independent article is the suggestion that climate is becoming “more unpredictable”. I suspect what is meant here is that we used to know what a mean season and normal variations were, and now we don’t. That’s valid, since the baseline climate is changing. But saying it this way — that “climate is becoming more unpredictable” is misleading. In fact, climate may, if anything, become more predictable as anthropogenic forcing becomes even more dominant (as greenhouse gas concentrations increase), relative to natural forcing and variability. And what is definitely not the case — but might be inferred from the article — is that weather is becoming more unpredictable. Weather prediction is based on observations just a few days in advance — climate and climate trends have nothing to do with it.
The point here is not that we shouldn’t be concerned about the fate of insects and birds in the UK (that would be the kind of conclusion that only the most willfully ignorant would draw.) They have been in decline for a long time (mostly due to land use change and pesticides) and there is little doubt that climate change will continue to add insult to injury. But it is simply wrong to confuse a year or even two years of unfavorable weather with a change in climate, and it is irresponsible to headline an article that is really about weather with the provocative juxtaposition of “climate” and “devastates”. Doing so gives the average reader the sense that their personal observations about “weird weather patterns” or fewer sightings of Parus caeruleus represent definitive manifestations of climate change. The fact is, climate changes are — so far — small enough in most places, relative to the natural variability, that one’s personal experience is a very poor guide to what is happening over the long term (observations of sea ice changes by those that live in the high Arctic notwithstanding).
#149 Gavin
My point wasn’t that 30 years was a magic number. I don’t know there is a magic number but, if there is one, is probably at least 30 years and not less.
You need to provide some criteria beyond “noise” for seeing “clear trends” or anyone could see a clear trend that no year has been warmer than 1998 and claim warming has stopped. Not that I am making that claim.
Jim Cross. I think that the problem was the paucity of detail in your original off-the-cuff. I sort of figured you were just denying that my 10 years was useful for climatology. Not that you were making some sort of dig at how long you had to average to get climate rather than weather.
You were rather the opposite side of what I was saying with “take 10 years”. I was picking out that 8 years is a strange number and that, currently, 10 years which would be more sensible would show warming quite strongly and that someone who didn’t use 10 rather than 8 (both stupid numbers, 10 just being a little less stupid than 8) was obviously cherry picking.
You were as far as I could tell saying 10 was stupid. If I was wrong, you would be saying nothing that wasn’t trivially proved wrong.
I wonder if anybody could offer an informed opinion about the content of the University of Birmingham’s press release (http://www.newscentre.bham.ac.uk/press/2009/01/Ice_Age_Press_Release_02_01_09.shtml) that was issued to accompany the publication of the paper by Bao et al (http://www.sciencemag.org/cgi/content/abstract/323/5910/119)? It seems to me that the references to the current warming and potential geoengineering are rather tenuous and certainly not mentioned in the paper – in fact they seem to originate from a throwaway comment on the web pages of one of the authors: http://www.gees.bham.ac.uk/staff/fairchildresearchglacial.shtml
[Response: I read both, and while the link to present day concerns (geoengineering, CO2 levels) are tenous, these researchers are not the only ones to try and make their work more relevant. Most importantly nothing in those described links is wrong, and there is no suggestion that CO2 increases could cause a new snowball Earth – which was the main thrust of the Telegraph piece. In general, adding context to a press release that isn’t in a technical paper in the literature is fine – and in fact very necessary because journalists are always asking what any new result actually means. In my opinion the journalist did a terrible job here – and it was not an inevitable result of the press release. – gavin]
[Further response: Upon reflection, there is some missing context in the press release. There should have been a line or two explaining how the build up of CO2 was expected as the trigger to ending the snowball episode, and perhaps a quantitative estimate of how much CO2 would be needed to bring us out of a snowball state. That would have been helpful, but still, that doesn’t excuse the topsy-turvy telegraph piece. – gavin]
I wrote: “Is the ‘optimum’ situation really to convert all of the Earth’s biomass to human flesh, and live on a diet of Soylent Green?”
RichardC replied: “I’d ask what type of opportunity one would like for ones OWN grandchildren.”
My answer is that I have no children and had a vasectomy 15 years ago to ensure that I will not have any — it’s a simple, safe and painless measure that I heartily recommend to all males of reproductive age. So the “opportunity” I am seeking to leave for the future is more room for other folks’ grandchildren — and not only the human ones.
And to answer my own original question, in my view the “optimum” situation is a rich, diverse, robust, sustainable biosphere within which human beings thrive as an integral part of the web of life, living in balance with, and with respect for, the other life forms with whom we share this planet.
At present, the human species — one single species — consumes an estimated 40 percent of the total biological productivity of the Earth. I would suggest that for a planetary biosphere considered as a living entity, that represents a disease condition. And given the combined impact of our numbers and the particularly destructive and toxic technologies (e.g. fossil fuels, deforestation, overfishing, industrial monocrop agriculture, etc.) that we currently use to sustain ourselves, it is a life-threatening disease condition.
We need to both move to less destructive and toxic technologies, and we need to scale back our numbers — preferably humanely, because if we don’t, the ensuing course of events will certainly reduce our numbers inhumanely.
“At present, the human species — one single species — consumes an estimated 40 percent of the total biological productivity of the Earth.”
Sources please.
Re Michael: (#155): I’m guessing that SecularAnimist was referring to the Ehrlich calculation reproduced at http://www.dieoff.org/page83.htm. Paul Ehrlich repeated this “40%” number at a talk I went to yesterday, though I don’t know if that now represents his middle estimate or if he hasn’t updated his “high” estimate, or what. I make no comments as to the accuracy of the calculation having not thought through it thoroughly myself.
> Sources please
http://www.google.com/search?q=human+species+40+percent+biological+productivity
Seriously, you should at least try Google before asking for help.
Show that you know how to copy and paste into the search box.
Even just copying the exact string out of the post before yours and pasting that in would have found the information you needed help with.
___________
Whigs dep’t
Chuckle — the first hit, if you paste that exact string (without the quotes) into Google:
World Book Encyclopedia | Atlas | Homework Help
… Today human beings consume, waste, or change about 40 percent of the net …
http://www.worldbook.com/wb/Students?content_spotlight/conservation/why
Homework help indeed.
Jim Cross writes:
Try doing a linear regression of the mean global annual temperature anomalies against the year. If the slope has p 0.05, it’s not significant.
A trend has to be statistically significant.
Michael wrote: “Sources please”
Hank Roberts wrote: “Seriously, you should at least try Google before asking for help”
On the contrary, when I state a statistic it’s really my obligation to provide a citation, which I failed to do.
The 40 percent figure I gave does indeed come from the Vitousek, Ehrlich et al study “Human Appropriation Of The Products Of Photosynthesis” published 1986 in Bioscience, which Marcus referenced above.
Other estimates vary, some being as “low” as 24 percent.
A relevant Google search term would be “human appropriation of net primary production HANPP”.
A couple of articles that give a good overview of the subject:
Global human appropriation of net primary production (HANPP)
The Encyclopedia Of Earth, December 2008
Our share of the planetary pie
Proceedings Of The National Academy Of Sciences, July 2007
Another useful search term:
http://scholar.google.com/scholar?q=“ecological+footprint”
Re Jim Cross @151, “My point wasn’t that 30 years was a magic number. I don’t know there is a magic number but, if there is one, is probably at least 30 years and not less.”
There is a way to determine how long you need:
http://moregrumbinescience.blogspot.com/2009/01/results-on-deciding-trends.html
(Thanks to Hank.)
Animist, I take issue with the idea that humans are a blight. I can imagine a far more [edit – ok, enough of this already]
148 Gavin’s response… Yes, and sea level is a good proxy for GLOBAL (as opposed to atmosphere/surface interface) warming. Note that sea level has risen over the last 1,2,5,8, or even 10 years. There is no reasonable way to state that GLOBAL warming isn’t ongoing.
http://en.wikipedia.org/wiki/Sea_level_rise
#162 Jim
Thanks for pointing me to a site that basically agrees with my post that 30 years is a good time frame.
Apparently others think shorter times are fine with sea ice and glaciers. They didn’t really indicate what time frames were good for Arctic sea ice, Antarctic sea ice, summer ice , winter ice, and which time frames for which glaciers but perhaps we can ask them?
I am only marginally comfortable with 30 years when we have decadal oscillations in both the Atlantic and Pacific and solar cycles of hundreds of years or more.
The only good time frame for assessing a trend is a time frame that includes normal variation in the major factors that could affect the trend. The problem is in knowing what those major factors are. Most climate scientists think they know most of the factor(s) (CO2) so 20-30 years is fine for them.
> 149, 165
No, Jim Cross, that site doesn’t agree with your post, because it’s about climate data not sea ice data.
Gavin pointed out to you the difference. You missed the point, and misread the site, in both cases ignoring the specific kind of data discussed and glomming onto “30 years” as a generality. It ain’t.
Read it again. You don’t need to “ask them” about this — although you can ask for help over there, it’s meant to educate youngsters about how to figure this stuff out for themselves.
#166
Hank,
Obviously, you’ve missed my point.
I understood the site was talking about temperature.
I haven’t read a good justification yet to choose a shorter time frame for evaluating trends for sea ice and glaciers (and which ones?) aside from vague comment about “noise”.
Jim, you don’t need to read a good justification.
Take the numbers.
Go to Robert Grumbine’s site.
Follow the steps he outlines.
You’ll figure it out.
And if you’re going to ask “which ones” again — that’s the point.
You decide. Find the data set you want to understand.
Apply the methods to understand it.
You can’t generalize.
Re: #167 (Jim Cross)
I responded succinctly and correctly to your question in #150. Did you not believe me?
Sea ice data has a bigger signal-to-noise ratio, so we can identify a meaningful trend with less data. That’s the way it is. Really.
No, the site is not talking about temperature, it’s talking about how to determine from the data itself how long it takes for the trend to be discernible from the noise. It uses the temperature trend to illustrate this.
Jim (167), and argument why a short period is wrong can be done by taking it to a logical extreme:
Measure the temperatures 12 hours apart. Once at midnight then at midday. You have now “proved” that there is warming and the earth will be as hot as the sun in three weeks time.
North pole ice extent: measure 6 months apart, once at winter once at summer. You have now “proved” that there will be no ice on the north pole by november.
It doesn’t help that you take them at the same time 1 year apart. Try it with the temperature in your house. One date/time measure the temperature in your kitchen. One year later ON THE DOT measure it again.
WHY ARE THEY DIFFERENT??? Maybe you were late putting the dinner in the oven so the room was warmed by the oven. Maybe it was especially mild weather.
The only way to stop that is to assume that, being “noise” it cancels out if you take enough samples at the same time in the same way.
So it must be “years” else you can’t pick the same season.
It can’t be only two or three years because binomial counting statistics tell you that this tells you NOTHING about a trend.
It can’t be a decade because we KNOW of several cyclic elements that would affect the weather on that time (El Nino and sunspot activity, for example).
So it has to be DECADES of measurements.
Now humans live 70 years, so 100 year measurements is longer than anyone alive will see the end of, and 20 years aren’t many decades.
50 years is two generations and the working life of a human today. It is several decades so will sample several El Nino and sunspot activity cycles, evening out the noise of having selected a year that just happened to have a very different activity cycle. We don’t know of any cycle that takes ~50 years either.
Milankovich cycles take millenia and we can measure what stage along it we are, since the orbitals are a VERY much lower chaotic system.
Now, what year would YOU say should be use and give an example of why it is better than 50.
If you can’t think of one, why do you want to know why a shorter period is not useful? You say you don’t know of a reason for *any* period to be selected, so from that POV, any period someone comes up with is OK.
Unless you don’t like 50 years for some reason. If so, what is it?
Guys, I don’t know if this will help. However, I’ve found it helpful to look at things in terms of a likelihood model for probability. Take the example of a coin toss, calculate the likelihood for n tosses over a range of probabilities of the coin to come up heads (you can do this in Excel). So you’ve got a single-parameter model, and the likelihood will tend to be distributed as Chi-square with one degree of freedom. Now watch how the likelihood for a given confidence level (say 99%) shrinks as n increases. That’s about as simple a model as you are going to get–no noise, one parameter, but it gives an idea how confidence increases with data. Any system will have a certain amount of data it will require to reach confident conclusions. For climate, it’s ~30 years; for ice, stronger signal, less noise=shorter timescale.
Oracle of ReCaptcha reminds us to start our seeds now: tending plants
Mark, I like your considerations on the selection of sampling times. Nothing like good old logic.
I’m a musician and academician by trade, so I can relate this to the question of sampling in digital audio: the minimal sampling rate necessary to avoid creating artifacts in the frequency domain is twice the frequency sampled (and this is still rather lo-fi.)
In the case of the sea-ice trends, of course, we have pushing 30 annual cycles, ca. 3 solar or El Nino cycles, sampled at 365 samples/cycles–more than adequate to capture the waveform with reasonable accuracy.
Kevin, 174, though I find it eternally frustrating that people think that nyquist does anything other than give the lowest oversample needed to identify a frequency. Forgetting that this depends on infinite resolution in amplitude and that nyquist says NOTHING about how well it can reproduce the loudness.
So a CD with 44kHz sampling and 16 bit resolution will, at 22kHz be about as accurate as 2-3 bit resolution and the accuracy of selecting that 22kHz is a warble that could put a poor record deck to shame.
But they think that it is 100% accurate right up to the Nyqist limit.
It ant.
Frustrating, really.
Right–nor will the harmonic structure (waveform) be correct, either.