This is a belated open thread for this month, for anything non-COP21 and non-AGU related.
There are reports (probably premature) that CO2 emissions this year have stalled. If this were true, and were to continue, how soon should we expect to be able to discern the effects of this in the atmospheric CO2 data?
Do we understand why temperatures in the Holocene were so abnormally stable, relative to the Pleistocene before it?
Kudos to M. Hollande for continuing a great French tradition.
#1 Wili – Emissions stop. Google throws up lots of stuff.
https://www.princeton.edu/main/news/archive/S38/51/51I69/index.xml?section=topstories (this is if we basically ‘crash-stop’ and emissions fall to zero, today)
This one (https://www.newscientist.com/article/dn28469-carbon-emissions-hit-new-high-and-temperature-rise-soars-to-1-c/) suggests that at the current CO2 level (about 400ppm) we are committed to 1.6 degrees warming (so a further 0.6 degrees). To stablise at 400ppm it suggests we’d have to reduce emissions by 60% overnight.
Tokyo Climate Center has posted the November global temperature anomaly:
What a jump! (not that funny, though)
The changes in annual temperature from one year to the next look like random variations with a mean not statistically different from zero. If these temperature changes represent net energy change each year, then the energy into or out of the system would also be random with a mean of zero. A simulated ‘random walk’ with the same variance as actually occurred invariably leads to much larger temperature trends, positive and negative, than has occurred over the last 150 years. Is there a neat explanation why the temperature increases of the last 150 years are no more than the results of a damped random walk, driven by natural variation in the system? The damping would come from a simple negative feedback.
What caused holocene temperatures to be so stable, relative to the pleistocene before it?
Here’s one suggestion:
The agreement reaffirms a commitment to stabilising temperature rises well below 2C, and even retains the option of limiting warming to 1.5C if possible. But it also confirms national targets that do little more than stabilise global emissions between now and 2030.
Given those emissions, sticking to within 2C will require us to take lots of carbon out of the atmosphere and store it in the ground. The parties to the agreement are, in effect, saying “we’re going to sell this stuff, and we’re going to dispose of it later.”
How do I know? Well, peak warming is overwhelmingly determined by cumulative carbon dioxide emissions. To stabilise temperatures at any level, be it 1.5C, 2C or even 3C, net carbon dioxide emissions must be reduced to zero. Most governments, environmental groups and business leaders now understand this.
And it is acknowledged, albeit implicitly, in Article 4 of the Paris agreement, which calls for greenhouse emissions to be ‘balanced’ by carbon sinks some time after mid-century. But we’re unlikely to hit ‘net zero’ emissions before temperatures reach 2C, and even less likely before they reach 1.5C.
Warming is currently at about 1C and rising by 0.1C every five to ten years. We could slow the warming by reducing emissions, of course. But if we fail to reduce at the required rate – and the inadequate emissions targets indicate this is the intention – then we will be left with no option but to scrub the excess CO2 back out of the atmosphere in future.
Who should take back the CO2? The fossil fuel producers
That is why the deal is like a gigantic take-back scheme. The proof lies in what is not said in the Paris agreement. There is no explicit mention of a global carbon budget for instance, which adds up total emissions since the industrial revolution.
That is despite the fact that all governments have acknowledged, through the Intergovernmental Panel on Climate Change, the reality that stabilising temperatures requires a limit on cumulative CO2 emissions. Certain countries simply cannot accept the suggestion that they may be obliged to leave some of their prized fossil carbon reserves underground.
And why should they? We do not need, and nor have we any right, to ban India from using its coal. We simply need to ensure that, by the time we reach 2C, or 1.5C if that is what is eventually deemed safe, any company that sells fossil fuels, or any carbon-intensive product like conventional cement, is obliged to take back an equivalent amount of CO2 and dispose of it safely to ensure it doesn’t end up in the atmosphere.
Right now, that means re-injection underground: forests can’t be relied on over geological timescales (they might burn down, or even die out and re-release their carbon due to climate change itself). But there are plenty of other creative ideas for carbon dioxide disposal: someone just needs the incentive to do it.
And who better than the owners of the fossil fuel assets at the heart of the problem? Logically, the cost of CO2 disposal should be borne by the seller of fossil carbon….
And as though we needed any more proof — people are really good at identifying problems and making choices that push things the wrong way:
A restaurant in eastern China has been caught charging customers an “air cleaning fee” on top of their food bills.
Diners at the eatery in Zhangjiagang city, Jiangsu Province, were charged one yuan ($0.15; 10p) each to cover the cost of purifying the air inside, the official Xinhua news agency reports. Cities in the region have been enveloped in thick smog in recent weeks, with visibility of less than 100m in the worst-hit areas.
The restaurant’s owners recently purchased an air filtration system to improve the dining experience, and covered the cost by passing it on to customers without prior warning, the report says. But after complaints from angry patrons, the local government intervened to stop the practice, telling the owners that it constituted an illegal charge. A city official tells Xinhua that it wasn’t the diners’ choice to breathe filtered air, so it could not be sold as a commodity.
While those on the receiving end of the charge were unhappy, the idea has been welcomed by many Chinese social media users. “I’d agree to the fee!” declares one person on the Sina Weibo microblogging site, while another says they would happily pay more for clean air. One user suggests that the government should follow suit to deal with the country’s severe smog problems…..
I posted a science paper cite a few days ago — not sure if it made it to publication here — showing that a large part of China’s CO2 and air pollution generally arises from a few inland provinces that burn most of the coal, and recommending focusing their efforts on those high emitting sources.
(That would be rather than, say running more air filters powered by dirty electricity, duh).
James Hansen has called the Paris climate accord a “fraud.” But it may be an important step in a process which achieves the carbon tax that Mr. Hansen hopes for. First all the nations of the world needed to agree to work together toward a common goal, while making clear that each nation would have the flexibility to take such steps as it deemed best to achieve the goal. That is what came from Paris. But then, as the next step, in the very near future, Germany, China and the U.S. should agree to impose a carbon tax at the source. Currently there is a growing momentum in favor of the tax (even Exxon wants it!), it would be consistent with the Paris agreement, other countries would have no good grounds for complaint, and it would be precisely what Mr. Hansen urged in his own speech in Paris: that certain of the major economic powers should impose the tax, and this would force all the others to join in or else miss out on the tax receipts.
Germany has a very strong environmental movement. China is in dire environmental straits and faces submersion of major economic centers. The U.S. is an oligarchy but even its oligarchs are facing reality and will soon instruct the lap-dog politicians to get in gear. Agreement among these 3 on a carbon tax is highly do-able.
The perfect is the enemy of the good. Instead of dumping on Paris, it can be used as a launching point.
Dick Newell (#6)
I’ll just give a quote from Andy Lacis, because I like it:
“This is where it is important to understand that the natural variability of the climate system represents temperature fluctuations about a zero reference point. Moreover, these random-looking fluctuations most definitely are not random variations (as in random walk), such that given enough time, they could move the global temperature arbitrarily far from its equilibrium reference point. Climate GCMs (and the real world) must conserve energy, so arbitrarily large departures for the equilibrium reference point simply can not happen in the absence of external forcing being applied. Thus, given enough time, the unforced climate will approach its equilibrium point, and averaging over a time scale that is longer than the time scales of the fluctuations will serve to define the point of equilibrium.”
(apologies for the double question. The moderation lag had led me to believe the question had been swallowed up)
> Dick Newell …
changes in annual temperature from one year to the next look like random variations with a mean not statistically different from zero
Problem there is you’re using your eyeballs to do the arithmetic, and the eyeball is notoriously bad at statistical trend detection.
You’re making the same claim Dan H. has reposted here over the past 11 years, frequently — and it’s been rebutted repeatedly by people who have taken statistics (and some who teach and write textbooks on the subject).
Have you tried the Search function here? You can start with any of the threads you’ll find on the subject. Try searching on trend detection, and refine as needed. Other useful search terms would include the names of some of the people who can help you learn how it’s done: tamino grumbine
It is not at all clear to me that Holocene was any more stable temperature-wise than previous interglacials. What record are you looking at?
Nov. GISTEMP is out: +1.05, demolishing the next-highest November: +0.8 in 2013. http://data.giss.nasa.gov/gistemp/tabledata_v3/GLB.Ts+dSST.txt
Nov. GISTEMP is out: +1.05, demolishing the next-highest November: +0.8 in 2013. http://data.giss.nasa.gov/gistemp/tabledata_v3/GLB.Ts+dSST.txt .
Hiatus…NO, just sloppy temperature records and ignoring the effect of clouds. I moved here (@ -29 152 in AUS)in 2007 triggered by the melting Arctic and back then the sky was usually clear most days causing relatively high 35+ day temperatures and cool below 10 temps at night. But over the last few years there is much more cloud cover keeping the days cooler mid-20tees and nights warmer above 10. When we do have the odd clear day, the day temperature soars as high as 48 and the night might get close to zero.
Our Weather Bureau (BOM.GOV.AU) only keeps MAX, MIN and AVR, which is hardly a good indication of what is happening here. I moved from Melbourne to this location, cause South AUS is predicted to dry out and here there are lots of big rivers. Also a little North there are beautiful rain forests which are now slowly migrating South.
Surely now that the whole world agrees, we should upgrade our temperature records and capture the complete changes. I have my own weather station with very detailed records, but unfortunately only started in 2012. For mitigation my footprint is already negative, but I am assuming the world will not do as well. I have built an underground facility, so that I may stay alive during those really hot days.
(email response from BOM was that it all just averages out)
So my question to you all is: How do these sloppy records create accurate climate models ? Haha, BOM analysis shows an AVERAGE increase of .15 degrees
We have prima facie reason to believe that increasing the greenhouse gas load in a planet’s atmosphere will increase its surface temperature. Since the correlation between CO2 and dT for the last 165 years is 0.91 (82% of variance accounted for, p < 10^-40), that confirms the prediction of the radiation physics. If that is not happening, and the increase in temperature is due to a random process, you have to explain why the greenhouse effect is not working as it does in the laboratory and on other planets.
“Is there a neat explanation why the temperature increases of the last 150 years are no more than the results of a damped random walk, driven by natural variation in the system?”
The short and neat answer is that the Second Law of Thermodynamics prevents random variation from creating a trend.
But to expand a little.
@-“The changes in annual temperature from one year to the next look like random variations with a mean not statistically different from zero.”
No they do not.
The changes in annual temperature from one year to the next show a complex pattern of autocorrelation and a constrained variance following a power law distribution that is shaped by the thermodynamic limits of the system.
@-“If these temperature changes represent net energy change each year, then the energy into or out of the system would also be random with a mean of zero.”
The temperature changes represent net energy changes that either alter the energy balance of the climate to reverse an internal variation, or as a response to a change in energy input or sinks.
@-“A simulated ‘random walk’ with the same variance as actually occurred invariably leads to much larger temperature trends, positive and negative, than has occurred over the last 150 years.”
Because it completely fails to represent the thermodynamic constraints. You cannot gain enough energy consistently over a century to cause the amount of sea level rise observed. By omitting this basic aspect of physics a random model is inadequate.
@-“The damping would come from a simple negative feedback.”
The ‘damping’ comes from well understood physical constraint, not least the 2LoT. you cannot lose more energy than the blackbody emission rate for a given temperature, and you cannot gain energy except by an alteration in the rate of energy input or loss. Those are real physical factors that constrain temperature change and may make it appear to be ‘random’ and revert to a zero mean.
But that is a superficial resemblance, replaced by extensive knowledge of the real physical factors that constrain the historical record of temperature change.
wili @ 1,
There are reports (probably premature) that CO2 emissions this year have stalled.
Doesn’t the current level of CO2 emissions increase the overall atmospheric level by a touch over 2 ppm per year? If so, the ‘stall’ will do nothing more than stabilise this rate of change.
(I am not any sort of scientist, so there may be feedback loops that I am unaware of, but it seems to me that this statement should be more-or-less true for the next few years. Perhaps decades.)
Thank you for the replies to my comment 6. When I looked at this back in 2007, I calculated the variance of the temperature differences each year for 150 years – it was 0.0134 and the mean was a small positive number, but not statistically different from zero. So, in the absence of any negative feedback, the variance after 150 years would be 150 x 0.134 = 2, so a standard deviation of 1.4°C. The actual temperature moved by about half of this, well within what one might expect of a random process. Of course, there would be a negative feedback trying to return the temperature to some very long term mean. Which might explain why the temperature has moved by such a small amount. Given what randomness can produce, I don’t see how one has any chance of detecting an anthropogenic signal.
Those who support the dangerous AGW position will argue that the ‘hiatus’ is caused by random variations in the system. If random variations can explain this, my point is, why cannot the whole thing be explained by randomness?
As Name Omitted says @15&16 above, GISTEMP has posted for November with an anomaly slightly down on October. At +1.05ºC, it was the hottest November on record and the second warmest anomaly for any month on record. Six of 2015’s months remain in the top twenty anomalies list. Three 2015 months are the warmest anomaly for their month, seven the second warmest, with April 2015 a lacklustre fourth warmest. The average anomaly for the year-to-date stands at +0.845ºC which compares with the last 12 months averaging +0.839ºC and last year’s hottest calendar year on record at +0.743ºC. (We could perhaps have a punt at the full 2015 average anomaly being something like +0.86ºC. For a comparison, the average of the warmest Jan on record, warmest Feb, etc, is +0.88ºC although note that does presently include Dec 2014.)
=21st … .2015.12 … +0.78ºC
15th …. .2015.1 … +0.81ºC
=9th …. .2015.2 … +0.87ºC
=5th …. .2015.3 … +0.9ºC
=45th … .2015.4 … +0.73ºC
=21st … .2015.5 … +0.78ºC
=26th … .2015.6 … +0.77ºC
=45th … .2015.7 … +0.73ºC
=26th … .2015.8 … +0.77ºC
=13th … .2015.9 … +0.82ºC
1st … … 2015.10 … +1.06ºC
2nd … … 2015.11 … +1.05ºC
@10 S.B. Ripman: Very well observed, I think. Hansen isn’t going to drop the carbon fee and dividend concept. Nobody else has to drop it, either. The Carbon Pricing Leadership Coalition was formed in Paris to help to advance the advent of carbon pricing.
Angela Merkel was one of the national leaders, in a good bunch of others, who endorsed the concept of carbon pricing at the very outset of COP21.
On this matter–and many others–Dr. Richard Alley’s talk at the AGU Fall Meeting was most helpful. He did not specifically endorse the concept. He placed the first industrial revolution in a deeply realistic frame, as only a fully cognizant paleo-geologist in Pennsylvania can.
on the stability of Holocene/Pleistocene climate…
I’m not sure the Holocene has been that stable (plenty of regional variation and the Younger Dryas etc) – but the question probably needs to be refined – at least to clarify which 10,000 years of the Pleistocene (lasted about 2.5m.y.) to compare with the Holocene (since 11,700 y)…
Gavin has a paper out on getting transient response from the climate record. http://pubs.giss.nasa.gov/abs/ma06110m.html
Dick Newell @6.
While in the thread above you have been given physical reasons for the annual temperature record not being the result of a ‘random walk’, do also consider the mathematical reasons.
“A simulated ‘random walk’ with the same variance as actually occurred invariably leads to much larger temperature trends, positive and negative, than has occurred over the last 150 years.” (My bold)
That is not true. A large set of the random walks you describe may lead “invariably” to an absolute (either positive or negative) ‘spread of trends’ greater than the global temperature trend for the last 150 years but some of those trends, indeed the majority of them will be not “much larger” but actually smaller than the 150 year temperature record trend.
Then, even if you edit away your asserted “invariability,” you are still left with problems. For instance, to achieve a +/- trend equal to or in excess of the 150 years of recorded temperature, it will almost always require short-term inter-decadal trends that are unseen in the temperature record. A ‘random walk’ with a step-size comparable to the 150 year temperature record is ‘very unlikely’ to attain a large trend over 150-steps without short-term trend changes in excess of those found within the temperature record.
Of course, just as there can always be a lottery winner, mathematically any record could result from one unlikely instance of random events. And just like the lottery ticket matching the drawn numbers, it is really a matter of how unlikely it is as a random outcome. To quantify ‘very unlikely’, the two simple tests I describe here already together put the chances of a ‘random walk’ well below one percent.
#6 Dave. The most compelling evidence is not the temperature change in the last 150 years. The most compelling evidence is the data from the glacial-interglacial cycles, in particular the data since the last interglacial. These provide the best estimates of the warming effects of CO2. The recent data you refer to are consistent with these but by no means as compelling. Look at all the evidence for climate change not just a subset that, taken in isolation, might not be all that convincing.
Well, for one thing, a lot more can happen in 1.8 My than in 12 Ky.
This spike at the end of this graph has been persistent throughout 2015. Does anybody know what has caused it? To me, it appears at least as unusual as the 2011 dip, which was explained as being caused by anomalous rainfall on Australia and South America during the 2011 La Nina.
BTW you do realize that the numbers on co2 ‘stalling’ are almost certainly related to the global recession underway. A drop or deep slowdown in economic activity (China, the developing world, the commodity producers for example) across the world has pushed us into economic numbers which fit the definition of a global recession – though the US is not yet among the countries in that situation. There is naturally a fluctuation in numbers when that happens. If our fearless leaders figure out a way to juice global economic growth there is nothing in our current practices which would indicate that the numbers would not return to trend.
Remedial asked “Do we understand why temperatures in the Holocene were so abnormally stable, relative to the Pleistocene before it?”
The Holocene is merely the current interglacial period in the continuing glacial-interglacial cycle of the late Pleistocene, which is driven by how earth’s Milankovich orbital cycles interact to alter distribution of sunlight incident on earth’s surface at the polar regions. Absent human enhancement of earth’s natural greenhouse effect, global mean temperature would be continuing its long slow decline since peak Milankovitch forcing during the Holocene Thermal Optimum of 8000-6000 years ago, and ultimately descend into another glacial maximum.
That said, the Holocene has been longer and more stable than the three preceding interglacials. It appears to be more similar to the fourth most recent interglacial, known as Marine Isotope Stage 11, of 374k-424k years ago, at which time the Milankovitch forcing matrix was similar to present.
Yes, you are right. The 2nd link I posted suggested that to stabilize [CO2] at the current level, you’d need to reduce emissions by 60% overnight.
I’m not all that clear on how long you can continue to emit at the remaining 40% level before you saturate the sinks and [CO2] starts rising again. Or, indeed, whether this is right at all. I was under the impression that you had to get emissions pretty much to net zero to stop [CO2] increasing.
COAL: May be biased being AUS, but I would like to defend the Coal Industry. Too many statements everywhere about DIRTY coal and EVIL coal. Without coal humanity may not have had its industrial revolution, no laptops to blog about evil coal. Over the years they have served humanity well. Imagine having been one of those young children having to go down a mine each day, so that your community could keep themselves warm. Imagine working in one of those power stations, I have been there and it is not pleasant. The industry has only been responding to our demands and now that coal is meant to be phased out, should be given the chance to remake themselves. The evil thing is our excessive consumption, not those who supply us what we need.
For coal exporters like AUS, I suggest that they give a discount to those facilities doing Carbon Capture in return for long supply contracts and if the target facility does not do Carbon Capture, the industry should be forced to invest the equivalent discount in Carbon Capture research.
The “start here” page was put together in 2007 and has many old links, not all of which work. Are their plans to update to more current information?
Dick Newell: Those who support the dangerous AGW position will argue that the ‘hiatus’ is caused by random variations in the system. If random variations can explain this, my point is, why cannot the whole thing be explained by randomness?
RC: You’re confusing what random can do. You’re thinking of a single variable non-physical system. It goes up or down all by itself at random. But on earth, almost all of the energy stays in the surface/atmosphere/ocean system, with inputs and outputs to space very well constrained (as compared to temperature swings on the surface) Our system is more like S + O = G. (Surface energy + Ocean energy = Global energy.) And G is highly constrained by solar input and the Earth’s albedo. It’s rising at a fairly constant rate.
“Random variations in the system” is something like El Nino. It doesn’t create or destroy energy. When surface temperatures “stall”, it does not and cannot (without an obvious forcing such as a huge volcano) mean that global warming has stalled. Instead, it can only mean that the oceans are heating up faster than they were previously.
In fact, if one assumes that the surface record is extremely accurate, then one could make a good approximation of ocean temperatures by inverting the surface record to determine the squiggles in the upward trend of ocean temperatures. When surface temps rise, ocean temps fall and vice-versa. Though of course they’re both rising in tandem as well.
You also speak of the hiatus in the current tense. It died years ago, and even if it had a teensy bit of breath before, October and November not just killed it but vaporized every atom of its fabled existence.
25 Chris Dudley: Your web site http://mdsolar.blogspot.com leads to https://www.blogger.com/profile/14124764472206647347
which says “Very recently, I’ve gotten involved in a startup that plans to rent solar photovoltaic systems in the residential market. My guess is this is going to catch on. My homepage is where you can sign up.” So recuse yourself on the subject of mitigation,
Gavin’s paper is behind a paywall.
Random walks vs climate forcing: A random walk has no force. A random walk is what a drunk does by accident. There is no rhyme or reason or energy behind anything the drunk does. The drunk may go in any direction and may reverse direction at any time.
The climate is not a drunk. There are forces that act on the climate. They are powerful physical forces with a lot of energy. Climate change cannot be compared to a random walk. Climate is driven inexorably in one direction or another by the powerful forces applied to it. You have to have a mechanics to do science, not just a correlation.
Weather takes random walks with its mean tied to the path of the climate.
Re: CO2 stalling. I read last week the same thing. Well, something’s is going on: 1. They aren’t stalling. 2. They are stalling but El Nino is causing all sorts of CO2 to be emitted from natural systems. 3. They’re stalling, but EN is in full swing and Ch4 from clathrates/permafrost are suddenly rapidly increasing.
I say this because the Mauna Loa is showing a 5.6 to 6 ppm increase since the mid-September weekly low and about a (currently) 4 ppm monthly rise in just about 3 months. Last year it took six months for a 6 ppm rise.
Something’s cooking somewhere.
Re CO2 emissions “stalling”. I went to a talk on CO2 levels recently. The stall happens to coincide with an El Nino. El Ninos tend to cause CO2 levels to rise. So unless the stall continues or a La Nina rapidly develops in 2016 I understand there will be a larger than usual CO2 level rise this coming year leading to the Keeling Curve remaining above 400 ppm at least for the next few generations of the Keeling family.
But I was wondering how much of the stall was down to any recession, and how much was down to the warm weather in CO2-emitting countries.
The answer to your question is physics. Energy is conserved. If a body is warming, the energy must come from somewhere. If that body is a planet, that is a whole helluvalot of energy (about 2.3 billion Hiroshima-style nukes worth) that must come from somewhere.
We know with 100% certainty that CO2 is a greenhouse gas. We know it is capable of explaining the warming. We know that a greenhouse gas will also cool the stratosphere (check!). What, pray, is your alternative explanation, given that the Sun’s output has been decreasing on average since the 1950s and that ocean temperatures are increasing as well?
D21: If random variations can explain this, my point is, why cannot the whole thing be explained by randomness?
BPL: And we answered that question, and the answers apparently blew over your head. So stop asking.
You are ignoring information that would readily rule out “temperature is simple random walk”. And in addition, you appear to be interpreting that failure to reject as evidence that temperature in fact is just a random walk. But you haven’t ruled out an obvious alternative explanation: Your test has too little statistical power to detect a true trend. You need to do a power test before claiming that your test shows anything at all.
First, even for the analysis of the temperature trend in isolation, go get a simple graph of C02 and temperature for the last 150 years http://thinkprogress.org/climate/2015/12/15/3732080/nasa-2015-hottest-year/). Note that there really wasn’t much of a trend in C02 prior to 1950. Also, not much trend in temperature. So by taking 150 years, you have 100 years of no trend, commingled with 50 years of trend, in the hypothesized driver of temperature change. The C02 trend is readily attributable to the post WWII economic boom and concomitant increase in fossil fuel use, i.e., that’s not an arbitrary or cherry-picked start point, it’s the known, correct start point for detecting a temperature trend driven by an atmospheric C02 trend.
Let me push this point to absurdity. We have temperature proxies that go back thousands or tens of thousands of years. Why not do the analysis of trend over, say, the past 2000 years? The answer is obvious: Because there wasn’t any C02 trend to speak of for the first part of that. Your approach differs from that only in that your starting point is not as poorly chosen as the year 1.
And I would bet that your test, like most purely statistical tests that purport to show no trend, has too little power (statistical power) to detect a true trend. So this is not very different from the folks who take a 15 year trend and declare that global warming has stopped. Only, here, the lack of power comes from using too long a timeseries instead of too short a timeseries.
So, have you done a proper power test? If not, you haven’t rule out the alternative explanation, which is that you are looking at the results of a poor test. I.e., your test is a stopped clock, and will always fail to detect trend.
If you want to have some faith that your test is providing you with meaningful information, then, you should do some type of power test. Monte Carlo is easiest. Construct a series with no trend for the first 100 years, then some meaningful linear trend for the last 65 years, tack on a simple error term whose variance is plausible, and apply your test for a random walk. Reshuffle the errors and iterate. If you frequently fail to reject the simple random walk, for a series that you know is not a simple random walk, then your test lacks adequate power.
So that’s straightforward.
But there’s a vastly easier way to reject the hypothesis that temperature truly is a random walk. The temperature trend does not exist in isolation. Analyzing it in isolation ignores a lot of information. In particular, we know the C02 series isn’t a random walk. We know what drives it. In the main, it’s a simple consequence of mass balance — we burn fuel and the carbon has to end up somewhere.
So, ask a second question, again in purely statistical terms. If temperature is merely a random walk, how likely is it that, of all such random walks, it would neatly parallel the C02 series, over the past 150 years. Purely by chance, out of all such possible random walks?
So just adding that one timeseries to the analysis already highlights how unlikely it is that temperature really is merely a random walk.
Then realize that not only has this tight correlation occurred in the present, but in numerous historical episodes (e.g., for at least the last three ice age cycles). In fact, I recall watching a talk by Richard Alley where he made it clear that, as far as can be told from the geological record, whenever the earth’s average temperature was high, there was a lot of C02 in the air, and whenever not, not.
At this point, how likely is it that Earth’s surface temperature is purely a random walk, unrelated to atmospheric C02 levels? Objectively, you’d have to say, pretty unlikely.
Finally, all of this purely statistical analysis really must defer to the folks who keep bringing up the physics. We know the mechanism. Purely statistical tests of aggregate timeseries, as a way to determine whether or not global warming is real, is probably the least strong approach one could possibly take. You’ve thrown out almost all the available information. Just look at the match between the spatial patterns of warming (poles versus equator) and the predicted patterns from physics-based models. Is there any way to reconcile the close spatial correlation with the notion that the global average is a simple random walk? I certainly can’t come up with one.
In short: What you have is probably just a poor test of whether or not something is causing the current rise in temperatures. Or, if you prefer, whether or not there is a temperature trend, as opposed to random walk. There are probably a very large number of such poor tests. Mostly, what they share is a lack of power. No one will take your test seriously because the underlying hypothesis is inconsistent with even the tiniest bit of additional information beyond the aggregate temperature timeseries. And even as a purely statistical exercise of that timeseries in isolation, you would need to do a proper power test before you can claim anything.
In answer to MA Rodgers (26), and my use of the word ‘invariably’, I suggest you try it. Get a spreadsheet, and generate 151 random numbers, normally distributed, with mean zero and variance equal to the variance of the actual temperature changes. Each number represents the net contribution (+ve or -ve) of energy to the part of the atmosphere measured. Then look at the cumulative sums (integration), and you will find that most times it departs from the start point by a lot more than the actual departure in the real data. That is why I used the word “invariably”. I am no climate scientist, but I do understand that what drives the climate is complex, not least small changes in the sun, perhaps larger indirect effects of the sun on cosmic rays and thence clouds, small variation in cloud cover has large effects on energy transfer; who knows how energy moving between the atmosphere and the oceans works; who knows how energy from volcanic action works. The net effect of all these processes creates a trend line with noise. The integration of the noise is the trend. When you generate artificial noise with mean zero, you get some rather dramatic trends. (so in reply to Richard Caldwell 25, I am not assuming a single variable system. This comes a bit rich from those who assume that there is only one control knob on the atmosphere, CO2)
Doubtless more CO2 in the atmosphere in the atmosphere has some, unmeasurable effect, but it beats me how you can distinguish it from the noise. I suppose the answer is that the unverified models tell us.
PS I have included my website this time, just to show that I really am on the side of the environment
Dick Newell, what you are missing is the physics. Any statistical analysis without an understanding of the physical processes of energy transfer that drive climate are guaranteed to lead to misunderstanding how climate works.
A new form of climate denial is being identified: http://www.theguardian.com/commentisfree/2015/dec/16/new-form-climate-denialism-dont-celebrate-yet-cop-21
Re me @ 17. If surface temperatures are the base indicator to react to, like 2 or 1.5 in COP21, how com nobody is concerned about their accuracy ? Am I pissing in the wind ?
#42, Dick Newell–
…look at the cumulative sums (integration), and you will find that most times it departs from the start point by a lot more than the actual departure in the real data.
I’ll take your word for it, but honestly if that’s true I find it not at all surprising. In the real world, there is conservation of energy, and there is the immediate Planck feedback, both of which should tend to stabilize temperature and damp down excursions to some degree. And in the statistical representation of the real world, there is something called ‘autocorrelation,’ which I (admittedly a Bear of Very Little Math) take to reflect those realities.
This comes a bit rich from those who assume that there is only one control knob on the atmosphere, CO2
Er, no, nobody thinks that. The dominant one, yes, but not the only one. As proof that that is the case, please consider Figure 5 in the following link; it’s from the Summary for Policy Makers of the last Assessment Report:
And no, it doesn’t include cosmic rays, for the simple reason that there is no strong evidence that the proposed effect is actually significant in any way. It’s been discussed here. There is still some lab work ongoing, AFAIK, but it’s not looking very promising for the cosmic ray brigade. Or so I understand the state of the art to be, at least.
“However, albeit a zero-effect cannot be excluded and the range of the estimated values is quite broad, the best estimate of the CI-CNM analyses indicates that an increase in GCRs leads to a small increase in mid and high cloud cover.”
That’s from this paper:
2010, and only two citations. And besides that, it’s already departed from the premise of the original work, which involved a longer-term trend due to the rotation of the galaxy; this is looking at Forbush Decrease events, which occur on timescales too short to account for observed temperature trends–absent, at least, evidence of a trend in FD occurrence.
the unverified models tell us
Er, no, the models have been evaluated to a faretheewell. Since there has never been any claim that they are perfect, I’m not sure that ‘verified’ is the appropriate word, but they have been shown to have retrodictive and representative skill. See:
That’s a power point discussion draft of the corresponding chapter in AR 5, and it summarizes what models can and can’t presently do–or perhaps I should say, what models presently do pretty well, and what they don’t do so well. The final chapter is brutally honest about the latter, and the depth with which it discusses the pros and cons is close to brutal for the reader, so be glad I linked the power point version instead.
Dick Newell writes: “This comes a bit rich from those who assume that there is only one control knob on the atmosphere, CO2”
A straw man, on top of “we don’t know everything therefore we don’t know anything” on top of mathturbaton without physics.
Tamino, who is an actual time series statistician, dismantles the climate as random walk nonsense here:
> 17, 45, Theo vdB
You’re misunderstanding one of the old simple basics — how data is collected and how the accuracy increases because we’re using multiple thermometers. Yes, the stations are collecting only simple information. No, that doesn’t mean the models can’t be right.
I’m not sure where to suggest you begin. Possibly the “Start Here” link at the top of the page would be helpful. You need to spend some time reading the basics.
This doesn’t mean there’s no problem.
The point is — understand the tools, don’t blame them.
Theo van den Berg says:
15 Dec 2015 at 10:49 PM
COAL: May be biased being AUS, but I would like to defend the Coal Industry. Too many statements everywhere about DIRTY coal and EVIL coal. Without coal humanity may not have had its industrial revolution, no laptops to blog about evil coal.
I may be a little biased but I would like to defend the following practices that were “beneficial” to humans before we found out they were not:
Tobacco and cigarettes.
Shark cartilage for curing cancer
Heron cough suppressant (my personal favorite)
Mercury for treatment of Syphilis
The Lobotomy as a cure for mental illness
LSD THERAPY for Schizophrenia
Losing weight by ingesting Tapeworms
Cocaine for pain relief and that little extra Boost.
should I include Vitamin Water?
Yes! Those were the good old days.
The term for what you are doing is mathturbation–it may feel good, but it accomplishes no useful purpose.
Let’s take your speculations one by one:
1) “…not least small changes in the sun,…”–Nope! The sun’s output averaged over a solar cycle has been declining now for 6 decades. Next!
2) perhaps larger indirect effects of the sun on cosmic rays and thence clouds,–Nope! Thoroughly discredited by both the CLOUD experiment at CERN and by the fact that temperatures continue to climb even as solar output and the heliomagnetic field weaken. Next!
3) “…who knows how energy moving between the atmosphere and the oceans works;…”–Climate scientists do. Next!
4) “…who knows how energy from volcanic action works…”–Again, Climate scientists do.
Dick, it’s not like this is a new field of study. We’ve known about the greenhouse effect since the 1820s. We’ve known CO2 was a greenhouse gas since the 1850s. The first prediction of anthropogenic warming due to CO2 from fossil fuels dates from 1896.
Now hurry, go read this
before you embarrass yourself or waste our time further.
Mathematical notation provided by QuickLatex
Powered by WordPress
Switch to our mobile site