A number of us have been debating whether the agreement between Hansen et al 1988 predictions/projections and later data is good, bad or indifferent. Naturally, the assessment depends a bit on the extent to which the applied forcings used in the projections match the data.
Some confusion has arisen regarding the forcings Hansen actually applied to his computations. On would think this is an easy question, but it is not, for a variety of reasons. In this post, I’m going to simply cut and paste what the 1988 paper actually says. (If you want to read what Hansen thinks of his level of success, he compared his predictions to temperatures in a 2006 paper. I may discuss what he says in that paper later if the conversation veers in that direction.)
Those interested can get the 1988 paper discussed in this post, it’s available from NASA GISS. The abstract is online.
When writing, I’ll assume anyone interested in this discussion is somewhat familiar with the paper in question.
What the Abstract says
The paper presents results of GCM computations with three different sets of forcings.
These experiments begin in 1958 and include measured or estimated changes in atmospheric CO2, CH4, H2O, chlorofluorocarbons (CFCs) and stratospheric aerosols for the period from 1958 to the present. Scenario A assumes continued exponential trace gas growth, scenario B assumes reduced linear growth of trace gases, and scenario C assumes a rapid curtailment of trace gas emissions such that the net climate forcing ceases to increase after the year 2000.
Section 3: Control run is perpetual 1958
Hansen et al 1988 simulations began with a 100-year control run which with forcings equal to those thought to exist in 1958. At the end of this control run, one would expect the “model earth” to fall quasi-equilibrium from a thermal point of view. (In contrast, when forcings vary, the earth may temporarily be too hot or too cold relative to the expected long term average for the instantaneous forcings applied.)
Here is how Hansen et al described the forcings:
Section 4: Forcings during transients.
The body of the text describes forcings in section 4. All forcings are described relative to the perpetual 1958 control runs. (I consider this baseline important because the temperature increase relative to perpetual-1958 should be strongly affected by the forcing history relative to forcing in 1958.)
I’ll paste in images of the actual text below:
Some of the confusion seems to stem from wording which involves discussing the growth rate in emissions growth rates. (So, Hansen is describing the second derivative of the emissions rate. The emissions growth rate is the first derivative of the emissions rate. )
- Scenario A: The GHGs emissions rate is positive in 1958, and this emissions rate grows over time. So, the emissions rate increases over time.
- Scenario B: The GHGs emissions rate is positive in 1958, and this emissions rate grows over for a period of time. Eventually, the growth rate in in the emissions rate slows, and the emissions rate matches the 1988 values. After that, the emissions rate is positive, and constant. This would match reality if every year, humans continually added emissions at the rate we added them in 1988.
- Scenario C: The GHGs emissions rate is positive in 1958, and this emissions rate grows over for a period of time. But then, in 1990, we drastically reduce the rate at which we add GHG’s, and by 2000 entirely stop adding GHGs. (Or, more precisely, the emissions rate is just equal to the rate that the earth removes GHG’s after 1990.)
(This will be presented in graphical form later.)
Hansen et al elaborate:
Note the “Scenario B is perhaps the most plausible of the three cases”, is the sentences used to indicate Hansen thought this was the most plausible case. Steve M at Climate Audit noted that Congressional testimony was worded as follows:
If I were forced to choose one as the most plausible, I would say Scenario B. My guess is that the world is now probably following a course that will take it somewhere between A and B. (p. 51)
Section 4.2: Stratospheric Aerosols
In addition to varying the forcing due to green house gases (GHG’s) Hansen et al included the cooling effect of volcanic activity. This is discussed below:
Pinatubo erupted in June 1991. So, the first volcano eruption came earlier than in the those included in Scenarios B & C.
The Graphs
The discussion of forcings in the text was also presented in graphical form, and translated into an equilibrium temperature without feedback, denoted ΔTo. I have not located the numerical value of scale factor to convert ΔTo to actual forcings in Watt/m2; but found a citation to a previous paper. (Equation 3 and definition 4 in that paper seem to suggest how one might compute this conversion factor; if any reader can identify the precise numerical value Hansen used, I’d thank them! )
Note that, according to theory, if man emits no green house gases and nature neither adds nor subtracts what is currently in the atmosphere, the equilibrium ΔTo will be constant. So, ΔTo sort of means “man stops emitting GHGs”. (More precisely, a constant level of ΔTo means man is emitting GHG’s at exactly the rate “Gaia” removes them.)
Examining the center panel, we see that the forcing for Scenario A grows exponentially, and had diverged visibly from Scenarios B and C by 1980; ΔT0 in the chart is non-linear, and curves up. This curvature is due to the growth in the emissions rate.
Before 1980, the emissions rate for scenarios B and C both grow, resulting in up-ward curvature in ΔTo near 1960. Somewhere near 1984-1988, the emissions rate becomes constant for “B” resulting in a linear trend for ΔTo.
The emissions growth rate becomes negative for “C” around 1990, then around 2000, the actual emissions rate for “C” is zero. That is, after 2000, not only doesn’t the emissions rate grow, the emissions rate for GHG’s is zero. (More precisely, after 2000, the human contribution to emissions is balanced by sinks supplied by “Gaia”.)
Emissions rates have not been curtailed to the nearly zero; Scenario C did not happen by any stretch of the imagination. The open question is whether Scenario B or A are closer to reality.
What does it say in the Appendix B?
Hansen provided further details about the forcings in Appendix B.
They begin by discussing trends from 1958:
As you can read, the forcings for Scenarios A, B, and C are differ even in 1958. Scenarios A experiences greater forcings during the 60s, and so its temperatures would be expected to rise more quickly than those of B or C during that period. Because the earth’s climate has a non-zero time constant, this difference would be expected to influence the absolute magnitude of the temperature anomaly predicted in later years– even up to the present.
(It’s for this reason, I think rebaselineing forcing when comparing to reality makes no sense.)
So, it would appear the growth rates of GHGS in the scenarios mostly matched observed growth rates of CO2 from 1958-1981. However, Scenarios A used some speculatively high estimates for some gases; the other scenarios used lower values.
Scenario C mostly tracks B early on, but then the emissions rates decreases until emissions rates effectively hit zero.
That’s what the 1988 papers says!
I think I’ve regurgitated what the 1988 paper actually says. I haven’t included comparisons to reality. I’ve obtained Hansen’s 2006 paper discussing his interpretation later, after data arrived. I may discuss that soon. But for now, this information should be useful for those trying to understand how Hansen described the forcings in 1988; it may also help clarify what terms like the growth in the emissions growth rate means!
Now, feel free to discuss! Or suggest other bits of information we can find and compare.:)
Lucia, I spent a fair bit of time trying to decode this earlier this year. This post http://www.climateaudit.org/?p=2645 was what I arrived at.
Earlier posts included http://www.climateaudit.org/?p=2630, http://www.climateaudit.org/?p=2638, http://www.climateaudit.org/?p=2602.
These posts provoked vicious attacks from Hansen’s original bulldog service provider, Gavin, and the usuals.
Lucia comments: I thought Tamino was the bulldog. 🙂
Lucia, I discussed this exegesis earlier this year with http://www.climateaudit.org/?p=2645 being what I arrived at.
Thanks Steve– I see you found the conversion factor for dT to forcing. From http://www.climateaudit.org/?p=2645
Yeah, it sure wasn’t easy to find the conversion factor though.
In the CA article, you cite Hansen 1998. That paper indicates Hansen thought we were following trajectory C in 1998. If so, then that could explain why the model predictions seem closer to C than B!
Lucia,
At the end of his RC article on Hansen’s projections, Gavin links to files of the numerical values of the forcings, which presumably includes the actual numbers that H used. Here is the text file scenario file of actual gas concentrations since 1958. Here is the listing of effective forcings since 1988. Here seems to be the current GISS E version.
Nick– Interestingly enough, Gavin put those links there at my request back in January or December. :).
I thought on the previous thread, after I plotted those actual forcings starting in 1958, you suggested the “actual forcing” file Gavin provided is actually the current values used in Model E? So they may or may not be actual forcings?
Also, I plotted the ABC forcing Gavin provides and they don’t match the curves in the Hansen paper. Gavin’s forcing files don’t include the volcano eruptions. Back in December, I discussed some of the values with Gavin. I know he got the temperature data by asking around, and found someone who’d digitized the data off the paper itself. I don’t remember discussing where the ABC forcings are from, but the provenance may be similar.
I’d email Gavin, but I suspect he’d be a bit grumpy about my asking at this point. Or… maybe not. Who knows.
Lucia,
Yes, the “actual forcing” file that Gavin provided is the same as the GISS E doc from here. But I’m not sure why you suggest they may not be ‘actual forcings”. I think they are, as near as the GISS E people can estimate. I’m guessing that what you plotted was just the sum of those individual forcings.
I have to say I didn’t understand what Gavin was doing with volcano-free forcings. It seems to me that we just have to try to match Hansen’s scenarios with what actually happened, whatever caused it.
I thought that’s what you were suggesting when I plotted them on the last thread!
Yes,in the previous post, I plotted both the sum of the forcings in that file and the sum plus the volcanic forcings. (Did I finally show the whole thing? Anyway, I can.)
I assume GISS does use whatever they think are the best real forcings.
The only slight difficulty is the file says “effective” forcings. There is some distinction between “effective” and “actual”, so if I show the whole plot, I’m not sure if we will be comparing “effective” to “effective” or “effective” to “actual”. It matter a little, since the modifier makes a bit of a difference to the plots!( That’s also one of the reasons I wish Hansen included the actual numerical value to convert dT to forcings in his paper!)
Lucia – thank you for posting this. It helps to understand things better even if we haven’t yet approached the question of which Scenario best matches what really has happened relative to GHG emissions. My guess is it is something slightly greater than B, but definitely not as high as A. C is definitely out.
Also, another minor suggestion, maybe the phrase “increasing GHG growth rates” would be easier to undertand than “growth in growth rates.
BobN
People are overcomplicating things massively in this discussion. In simple terms we now have a genuine, gold plated performance record for climate projections premised on the greenhouse forcing theory. Three scenarios were produced to reflect three different possible states of the world. These were categorised by the author as:
Emissions continue to grow (from 1988)
Emissions remain constant (from 1988)
Emissions are “drastically” cut (from 1988)
The best exposition I have seen of the relevant data is here:
http://forums.randi.org/imagehosting/6867486ca56737141.jpg
Leave aside all the complexities of different forcing agents, emissions versus concentrations etc etc. Look at the chart and try this simple test. I will put my answers after each one. You reach your own conclusions.
1. Which of the three scenarios produced in 1988 was most closely tracked by subsequent observations up to present time? [A: emissions “drastically” cut]
2. Which scenario state of the world would best characterise the actual state of the world over the last 20 years? [A: something between “emissions continue to grow” and “emissions remain constant”]
3. Based on these two responses would you say that the model employed to generate the scenrios in 1988 was valid? [A: No]
It’s as simple as that. The model employed to produce that 1988 paper is clearly invalidated by objective observations – a true test of skill.
Paul-
You have concluded that an AGW model is deficient. That is, of course, not possible. Either the data is wrong or we are experiencing a series of “short-term masking effects” that are disguising Nature’s overwhelming desire to conform to the model. Go cut down a bristle cone pine and carve a hockey stick, big guy. It will give you something soothing to do until they come to arrest you.
Been lurking around these websites for a while. Zero met/climate experience and my computer degree is 25 years out of date (went into the military in a non-related field) but, I have read various estimates for the non-AGW rise in temp since the LIA of .5-1.0C per century. Was this included in Hansen’s model? It reads like this is just GHG forcings. If not, then each scenario should be .25-.5 degrees higher over the 50 years since 1958.
Or is the non-AGW rise a fallacy that “deniers” (and as time goes on I find myself leaning that way) grab on to. Even 25 years ago, garbage in-garbage out applied to everything you did on a computer and, if we knew what the professor wanted, we could write mountains of poorly documented but impressive code and arrive at the correct “answer”. After all, who would go through 1500 fortran punch cards to check the result?
R. Peterson:
Hansen’s model isn’t specifically based on temperatures. It is based on conservation of mass, momentum, and energy. However, due computational constraints, and incomplete understanding of some physical processes, the physical processes approximated to some degree. Due to uncertainty associated with these approximations, and uncertainty in the forcings used to drive the model, the predictions must be compared to out of sample data to determine the level of accuracy one might expect from forecasts.
I don’t know if the model has been run and compared to Little Ice Age data. Since there is some debate over the precise global temperature during that period, it would be difficult to get a complete test.