Raven drew my attention to something posted at NASA GISS, which implies that, for all practical purposes, normal variations in solar irradiance don’t affect their predictions of future warming trends.
On their web site discussing 2007 temperatures, NASA GISS displays a figure showing the variation in solar irradiance measured over time. I’ve reproduced that figure to the right. In that figure we see the irradiance has an 11 year cycle and an amplitude of 0.15 W/m2.
NASA GISS then places the magnitude of this oscillation in context predicting future warming, writing:
The solar minimum forcing is thus about 0.15 W/m2 relative to the mean solar forcing. For comparison, the human-made GHG climate forcing is now increasing at a rate of about 0.3 W/m2 per decade (Hansen & Sato 2004). If the sun were to remain “stuck” in its present minimum for several decades, as has been suggested (e.g., Independent story) in analogy to the solar Maunder Minimum of the seventeenth century, that negative forcing would be balanced by a 5-year increase of GHGs. Thus, in the current era of rapidly increasing GHGs, such solar variations cannot have a substantial impact on long-term global warming trends. Furthermore, recent sighting of the first sunspot of reversed polarity (reported Jan. 4 by, e.g., SpaceWeather.com and NOAA) signifies that the ~ 4-year period of increasing solar irradiance is about to get underway. Emphasis mine
I may be mistaken, but I think NASA GISS is saying that the effect of solar variations is currently so small compared to the effect of GHG’s that solar variations can’t throw their predictions off ‘substantially’, (whatever ‘substantially’ might mean in percent. 🙂 )
The definition of ‘substantial’ is not quantified. So, I thought I’d look at a graph of forcings, I plotted from a file made available at Real Climate by gavin.
Recall that ‘forcing’ is defined as a level of irradiance above a baseline set at some point in time. In the plot to the left, the total estimated forcing experienced by the earth’s climate relative to a baseline in 1880. If we examine this graph, the total forcing due to all factors– including solar, ghg’s, stratospheric aerosols (due to volcanic activity) and what not is illustrated with a thick pastel green line.
The variability due to solar is illustrated in a kelly green line which oscillates up an down every 11 years.
If you examine the line for the total forcing, you can see the small magnitude of forcing due to solar oscillations riding on the total forcing curve; they barely make any difference. (In contrast, if volcanos start exploding much more frequently than in the past, this can have a noticeable impact — though, my impression is the predictions for the mean include volcanic activity at some past historic rate.)
So, basically, if I understand NASA GISS correctly, they are saying that, given how very much larger in magntide the other total forcing are, the oscillations due to the solar cycle no longer impact future temperature trends in any ‘substantial’ way. And, of course, NASA GISS is also saying that these also don’t impact their projections or predictions for future temperature trends.
So what does this mean with respect to validation or falsification or predictions?
Kim asked what this implies with regard to my discussion of hypothetical future weather that could falsify projections.
Recall, I said that a 10 year run of weather with a 0.0 C/century slope in GMST, obtained by OLS would statistically falsify the IPCC projected central tendency of 2.0C/century, and any value on the high side of that value. (It would not statistically falsify anything lower.)
With regard to a hypothesis test, and the cut-off for a falsification means is,
the NASA GISS implies that, even if the sun gets stuck in a minimal solar cycle, NASA GISS currently says that will have no substantial effect on the future trends. That presumably means that my hypothesis testsneed not be modified, even if the sun were to get “stuck” in a Maunder Minimum. The effect of the sun getting “stuck” is expected to have ‘no substantial’ effect.
So, in reply to Kim’s specific question: The cut-off for falsification of a 2.0 c/century or higher trend remains 0.0 C/century.
Will the observed trend be 0.0C/century or lower measured between 2001-2010? Who knows? . We must wait 3 years more years accumulate 10 years of data.
If the IPCC projections for the central tendency are correct, and trend really is 2.0 C/century, presumably it will warm up later this summer, fall, winter… or sometime next year. The IPCC projections will then, not be falsified.
Oddly enough, the way statistics work even if the absolute total complete denialists are correct, there is roughly a 50% change the 2.0 C/century trend won’t be falsified to the 5% confidence level in any given 10 year run!
Falsifying incorrect null hypotheses ain’t easy! (Which is why it took a long time to statistically falsify the 0.0C/century null hypothesis in the first place.)
But to explain why falsifying ain’t easy, I’ll have to discuss β error. Which I will do later. 🙂
End notes:
This is hopefully a link to a page with the excel file I used to create the graph: Forcings In Figure.xls If I didn’t screw up, this data is identical to the data at Real Climate.
“The modelers say that higher greenhouse warming produces recognizable changes in the Walker circulation. What better data is there to test such model results than the tropical rain products from TRMM?”
From Pielke
Thanks fred! I’ll look into that data product.
So you have solar forcing numbers, right? Hmm…
Shaviv 2005, On climate response to changes in the cosmic ray
flux and radiative budget:
i.e. ten times the NASA numbers! Interesting, huh? I’d be curious to see the effect on the model of increasing the solar forcing by an order of magnitude…
http://www.phys.huji.ac.il/%7Eshaviv/articles/2004JA010866.pdf
Andrew– If you are correct, then that magnitude would not be neglibile. However, to do any sort of validation/ falsification, you need to be far enough along to predict forcings from year to year.
Is anyone anywhere near able to provide a prediction for say, 2008? 2009? That would be a hurdle you need to get over to give your theory predictive power.
That is the crux of it! You’d have to ask a solar scientist, or perhaps and expert in cosmic rays. I really couldn’t tell you. Indeed, there are multiple predictions of the magnitude of the next cycle. Until we get better at that, it will probably have to all be postdictions (which are pretty meaningless).
I trying to wrap my head around the values of various forcings. Can you put the values in a spreadsheet/text file for me? I want to test the potential effect.
Lucia,
Joe d’Aleo at IceCap recently published a graph that plots CO2 concentrations against Hadley and MRU temperature records for the period 01/1998 -01/2008. The R2 values were 0.001 and 0.005 respectively, hardly something to hang your hat on. You would expect the fellows at GISS and RC to have seen that. Methane ppmv values are declining, cloud cover appears to be increasing, so on what basis do they continue to show the levels of purported GHG forcings?
Andrew– Postdiction isn’t meaningless. It’s necessary to postdict first. But, post-diction isn’t the same as pre-diction. So, to be really confident you’ve shown something, you need to prove you can predict.
With climate stuff, that means you often need to wait for the world to go around the sun a few cycles.
I suppose your right. I’ll see if I can figure out what we should expect sun-wise over the next cycle, and perhaps get back to you. I not usre if you got my request for the radiative forcing numbers above. I’d love to have them, if that’s okay. 🙂
Andrew– I added a link to a page with the excel file above.
I hadn’t seen those RealClimate forcing files before, so I thought I’d play with them a little bit. This is all just in fun — I’m not going to make any grand conclusions so please don’t misinterpret:
First I noticed that most of the forcings (all but solar and GHG) were nearly constant from 2000-2003 (at least). Second, the GHG forcing is increasing almost linearly. The solar forcing was of course following the solar cycle.
I decided to extend the forcing out from 2003 (the date the file ends) to 2010. For the GHG forcing I simply extended the linear trend from the 10-years prior 2003. For the solar forcing, I copied the data from the end of solar cycle 22 and start of solar cycle 23. That is, I copied the data from 1993 to 1999 into the rows for 2004 to 2010. All other forcings were held constant at their 2003 values.
So, after doing all this, here’s what I found:
– forcing slope from 1970 to 2003: 0.42W/m2/decade (call this the nominal slope)
– forcing slope from 2000 to 2007: 0.14W/m2/decade (~1/3 the nominal slope)
– forcing slope from 2008 to 2010: 0.76W/m2/decade (~2x the nominal slope)
– slopes on other short periods are hard to define because of volcanoes
If the real forcing is anything like the values above, the next few years could interesting.
Hopefully we don’t have any major volcanic eruptions…
I’m curious, what does Lumpy predict from these hypothetical forcings?
Lumpy Mach 1:
From My prediction.
As you know, moscher has totally mocked my prediction for 2008. 🙂
I didn’t extrapolate using the model with the more precise data for volcanos yet. I got distracted when Mr. Rabett posted his attempt at Taylor series, and then later when I thought about Roger Jr’s question and realized it was possible to answer it.
So.. eventually, back to Lumpy.
It’s too bad I can’t include images here.
On second thought, it’s probably a good thing. 🙂
Our post-2003 hypothetical forcings do not agree. (I’m using your “Equilibrium Temperature” to be proportional to the net forcing). Is it correct to say that the solar cycle is not included in your forcings? If so — and if you’re inclined to satisfy my curiousity — does overlaying a solar cycle estimate have much of an impact?
Actually John V– at this point, I’d have to look in the file. I created that figure when I made my first version of lumpy (which had a pretty darn good correlation coefficient– which astonished me.)
Then, I decided I needed to make a prediction because no matter how bad or new a model is, or how unread a blog, I figure one should predict 2008 temperatures in January!
I’m learning to upload the specific excel spread sheets to match files as I blog more and more. Otherwise, I can’t keep track. (But, I’ll try to hunt that down for you.) Since this is a hobby blog, and I’m fiddlign with different brain-farts, if I don’t upload the files to my blog archives, I’ll end up worse that… worse than… He who shall remain unnamed….
Equlibrium tempearture is supposed to be forcing scaled by sensitivity to give a temperature.
I find it somewhat disturbing that NASA neglects the impact from cosmic rays on clouds. It is close to two years since the link was proven beyond any reasonable doubt in the Harrison et al paper.
(See http://www.cosis.net/abstracts/EGU06/07661/EGU06-J-07661.pdf)
Svensmark and others might be wrong about the magnitude of the effect (and even the mechanism) but to claim in 2008 that solar activity does not affect clouds is a disingenuous. Hence any claim that the solar forcing is only +/-0,15 W/m^2 is downright wrong.
Avfuktare Vind —
Thanks for the reference. I’ll give it a read.
I think, there NASA GISS are very careful in picking their words. The variations in the irradiance from the sun itself only vary that much. They discount cosmic rays by simply not mentioning it.
I know that the bloggers at Real Climate just don’t believe it. As for me: I know absolutely nothing about it! (No one can know everything. I often pretend I do, but then someone says “cosmic rays” and I have to admit I know nothing about it!)
Hi lucia, thanks for the data. I decided to test the effect for myself (the hypothetical effect) and the result is quite interesting! If and I stress if the CRF-LACC link is as strong as Nir says, the effect on the overall forcings would be quite large indeed! since 1/.15 is equal to 6.66… the indirect forcing would, theoretically be almost seven times the irradiance forcing. Adding the effect in produces this:
(The y axis is W/m2, I think, and green is with, blue without)
Now I just have to figure out how to project this forcing into the future (along with other forcings) and maybe we could use Lumpy to make competing predictions! I’ll get back to you on this idea. Lumpy is your baby, so I’ll understand if you don’t want me toying around with a “solar” version.
Agh! the image didn’t show up!
Here’s a link, maybe it will work:
Andrew– send the link to the image and I’ll insert for you.
Oh, anyone can make a Lumpy! I was planning to get various Excel macros sorted out and post it with instructions on how to make your own Lumpy. But, when a person does that, they need to have it truly ready because I know from experience, answering questions about how to modify Excel spreadsheets on a blog can end up sucking up a lot of time– particularly if the spread sheet is unclear or buggy.
When you read about cosmic rays and clouds it is worthwhile to have in mind Svensmarks hypothesis. Several sources – and you will find most of them at realclimate – have tried to debunk the hypothesis by misstating Svensmarks claim. So just for clarity:
Svensmark claims that cosmic rays of a certain energy (roughly 10 GeV) are affecting cloudiness in the lower atmospere and mainly only over the large oceans far from land. The particular energy range is needed to create the secondary particles that ionise cloud seeds in the lower atmospere. The effect if little at high heights and close to land as cloudiness is not mainly controlled by ionisation there.
Harrisons et als paper unfortunately concerns the british islands and hence cannot be used to establish the magnitude of the effect (as the magnitude can only be assessed far from land). Svensmark himself claim several watts per sq m over the 20th century but I have seen no independent assessments.
A second issue: the speed of earths rotation has increased quite significantly over the last decade. From my perspective that should tell us something about the heat of the oceans, and presumably faster rotation means colder oceans, though it could also mean changes in floating ice or in glaciers. I know I wont find time to do a proper calculus anytime soon (which I think is needed: the density of ice vs water is quite significant, so rather small changes of ice might skew the over all picture; on the other hand most of the ice is quite close to the rotational axis). Are you or somebody else aware of someone that tried to extract something out of the speed of rotation, or have the time to do a proper integration?
I’m not aware of anyone who has tried to do anything to consider changes in the speed of rotation of the earth on climate or weather. After you posted, I googled, and found this NASA article. It just discusses detecting the linkage between changes in the atmospheric motions and the earth’s rotation.
Avfuktare Vind says:
Where is the data that supports this statement? I tried google and could not find anything.
Raven, I don’t have the exact data but you can see for yourself that the number of leap-seconds are down from one every 1,5 years to one very 5 years or so.
Wikipedia shows the difference between leap-second adjusted time and non-adjusted time here: http://en.wikipedia.org/wiki/Image:Leapsecond.ut1-utc.svg
It is clear that earth slowed down significantly from 1999 and has kept the low speed since.
I think it would be interesting to investigate what this tells us about ocean heat content.
The earth orientation center has some data, but unfurtunately I suspect that there aren’t data there prior to 1979. (http://hpiers.obspm.fr/eop-pc/) (and heres a nice graph: http://hpiers.obspm.fr/eop-pc/earthor/utc/leapsecond.html ) Atomic time goes back to 1955 if my memory doesn’t fail me, so there should be better sorces somewhere.
The eart speeded up in the seventies, slowed a little in the eighties, speeded up again in the nineties and slowed significantly after 1999. I.e. it follows pretty well what the IPCC tells us the temperature anomaly is.
Interesting.
Here is a piece of data that supports the idea that the earth has been spinning faster since 1998:
http://www.gsfc.nasa.gov/topstory/20020801gravityfield.html
The story I linked does not offer an explanation for the increase in the bulge. However, basic mechanics suggests that earth spinning faster == bigger bulge at the equator.
I certainly think that the IPCC has some explaining to do. The “missing” heat in the atmosphere must shurely have gone into the ocean if the radiative imbalance is really some 0,85 W/sqm. But with no change in sea level rise, no positive temperature trend in measured SSTs or ARGO and an acceleration of earths spin it is downright unbelievable that the ocean has taken up the heat. And with neither oceans nor atmosphere getting any warmer the concept of “commited warming” appears to be mere hand waving.
I looked at http://sealevel.jpl.nasa.gov/science/timeseries/199611_G.html where you can study SST and SSH anomalies. There seem to be significant correlation between temperature and height, even though underlaying sea water must be as important as the SST (though much slower in response of course).
Begs the question even more: where is all the heat that is supposibly trapped beneath green house gases?
Avfuktare Vind —
Within the framework of conservation of energy, and AGW, committed warming is a concept that makes sense. Of course, we need to keep up with validation and compare predictions to data as it comes in, but the general idea of committed warming has a firm theoretical basis. (The question is: how large might this be?)
I’m not sure I follow a lot of the earth/ spinning energy arguments. Someone needs to sit down and run some numbers to estimate how much a change in the earth’s rotation rate would increase or reduce thermal energy. I haven’t done it.
Lucia, thanks for the reply.
If the radiative imbalance really was +0,85 W/sqm we should see an accumulation of joules in the climate system but I find no evidence of this. This is what I meant.
I do agree that a relatively long lasting forcing such as CO2 means that emissions have a “relative commited warming”, i.e. all other equal we should experience warming. But alas, in nature all other aren’t equal, other forcings vary as well, some of which we know very little. Hence the idea as expressed by the IPCC – that temperature will continue to rise even if we stop emitting carbon dioxide today – has little bearing on reality. In real life, at least from what we can measure, we have a slight to moderate negative radiative imbalance today – not a positive dito. And so temperatures will remain or decrease if we stop emitting CO2 today – if all things remain equal (which they don’t).
So in short, I do agree that within the framework of conservation of energy “commited warming” is a reasonable concept. As promoted by the IPCC, I don’t think it is reasonable.
I have a sign convention question: by “negative” do you mean we are gaining more than losing? Or loosing more than gaining?
My understanding from the few direct measurements of radiative imbalance, we are gaining slightly more energy than we are losing. I have no citations for this.
Do you have any– I’d be interested in reading these.
Lucia, I meant the same convention as the IPCC, positive=warming in pipeline.
Raven/others
I looked at this graph: http://www.fao.org/DOCREP/005/Y2787E/y2787e03b.htm#FiguraB showing detrended delta Ts vs the Length of Day (LOD). Indeed the correlation is striking, but the timing seem wrong: the LOD changes BEFORE temperature which is not what I expected (unless it takes several years for warming to propagate from oceans to land, but then we would need to assume that essentially all significant forcings are active over the big oceans, and not even Svensmark claims that).
I also want to caution that ocean currents may redistribute the mass of the ocean and thus affect LOD without any warming and cooling occuring. (I don’t know if such a redistribution is significant, just want to be clear that theres an additional source of error, above the behaviour of ice which I mentioned before).