How are Hansen's Global Warming predictions panning out?

[FL_Prius_Driver]

Originally Posted by chogan2

His mid-level scenario predicted warming at an average of 0.24 degrees centigrade per decade.

So, take a decade on the left half of the chart and average the heights of the data point, and plot that at the mid-point of the decade. (By eye, I mean.) Then do the same for a decade on the right. Those two mid-points are 10 years apart. Based on Hansen's projections, the more recent decade average should be about 0.24 degrees centigrade higher than the previous decade average.

Or, as I said above, the points on the left should average to just below the line, the points on the right should average to just above the line. Which is how it looks to me.

That's the proper way to evaluate this by eye, not the incorrect 0.5 degrees as posted above, and certainly not by looking at the start year and end year in isolation.

Thanks for the response. Unfortuately while there is lot of info on the Hansen predictions, I have little info on the OP's chart methodology of coming up with tempertures. Regardless, I do not want to turn this into a chase for details since the prime purpose was to measure the temperature of the PC enviornmental forum. In that case, it's lower than years past.

[amped]

Originally Posted by Godiva

Substitute people still smoking with the damage the increase in the price of gas is causing. We could have been conserving and developing alternative technologies years ago.

Oil and the Big 3 conspired to limit CAFE, dismantle the CA electric car legislation, etc.

And considering the harm to the economy and the air everyone breaths....it goes way beyond "second hand smoke". "Non-smokers" are being harmed by their decades of obstruction and deceit.

Same tune different lyrics.

Toyota alone among the Big 3 embraced and supported the newly mandated CAFE standards. Care to change your tune?

[tripp]

I think she's referring to the American "Big 3", not the actual Big 3.

[TimBikes]

Originally Posted by Godiva

Hansen thinks oil companies should be sued just like the Tobacco Companies for not only failing to do anything to mitigate climate change they were contributing to but going beyond failing to inform to actively covering up the damage they were doing.

Does he also want to sue anybody who has ever driven a car? That seems only fair.

[TimBikes]

Originally Posted by chogan2

Despite your coy posting, you want us to look at a year and compare it to another year. Ie, compare the red dots.

Sorry, it just doesn't work that way. As I've posted here in the past, you need to average a lot of years to be able to say whether a trend in global temperature is or is not statistically significantly different from zero, because there are large annual temperature changes due to "weather" (meaning, all kinds of short-term phenomena).

Realclimate makes this point pretty clearly, and those guys actually run models for a living, so they ought to know the underlying statistics. Here's their discussion of "short term trends".

RealClimate

Here's their analysis of Hansen's two-decade-old projections, with a lot of technical detail.

RealClimate

In both pieces, they make it clear that it's nonsensical to compare individual years. Why? Because the random variation due to weather is very large compared to any systematic trend due to global warming. Because you are looking at noise if you look at individual years.

Exactly how nonsensical is it, to look at any one year? Well, how large is the annual variation in weather compared to the annual projected warming trend? That is, how much are you confusing weather with climate, if you look at individual years?

The answer is, almost all of of the year-to-year fluctuation is weather. So that if you focus on individual years, you are (almost entirely) talking about weather, not climate.

Let's put some numbers on that. The middle-range scenario in Hansen's projections (the one that most closely matches the C02 forcings that actually occurred) projected warming of something under 0.25 degrees centigrade per decade. That's an average annual impact of 0.025 degrees per year. Now, look at the graph, and see that the typical annual variation in global temperature is easily 10x larger than that. That's why you can't look at individual years and say much of anything. If you look at any one year, you are looking at weather, not climate.

If you want to start looking at climate, average the years to get rid of the transient weather component. If Hansen's prediction were in the ballpark, you'd see the average datapoint on the left side of your graphs falling just below the zero line, and the average datapoint on the right side of your graphs falling just above the line. I'm not going to take the time to plot them, but that sure looks about right to me. The average of the temperature datapoints, as shown in the graphs, appear reasonably close to what Hansen projected.

If it stays cool, sure, I'll change my mind. Given that the current La Nina appears to be breaking up now, and that we're starting a new solar cycle as of this spring, it seems unlikely that global temperatures will stay low. But facts are facts. If the facts change I will change my mind. But it's just nonsense to look at individual years.

So, my conclusion, based on averaging the datapoints by eye, is that Hansen was in the ballpark. But you don't have to take my work for it, you can read what the experts in climate have said, as in the realclimate analysis above.

I agree it is non-sense to look at individual years, which is why I presented the entire graph, not the beginning and endpoint values. But as anyone can eyeball it, there is trouble in global warming paradise. And if you don't like the graph, you can go to the dataset and run the regressions yourself. If you do, you will see some interesting things.

First, I ran the global average value regression and it shows a slope of 0.14 C / decade from 10/79 - 6/08. This is substantially below Hansen's projections and since it covers the entire 30 year satellite record, is hardly "weather" and is not cherry picking the data since it is the whole record.

Second, when you plot the trendline you notice a very interesting step function in 1998, which just so happens to be a strong El Nino year. If you break the data set to pre/post 1998, you find the trend for pre-1998 to be 0.024 C per decade, 1/10th what you suggest is Hansen's value. For 1998 onward the trend is 0.06 C per decade, about 1/4 of the Hansen value.

So either way, he is far from the mark. And unless one can explain how CO2 caused a step function in a single year that drove a sudden rise in global temperatures, I'm betting on mother nature (and El Nino) and against Hansen. But even if you ignore the step function, Hansen comes up something like 80% above reality.

[chogan2]

Timbikes, if you think you can test this better than the pros, you need to start by looking at the right data. As I understand it, Hansen's prediction was 0.24 degrees centigrade per decade for the earth's average surface temperature. So you need to look at average surface temperature data. We all know that not all parts of the atmosphere and not all parts of the surface are expected to warm at the same rate. So, you need global surface temperature data, not some other data, such as the data your link pointed to, which was somebody's posting of data labeled as lower troposphere temperatures. And you need to look at the period Hansen was discussing, which was 1984 forward, not whatever number of decades happen to show up in the dataset, as you have done.

So, whatever that data set is, it's not labeled as surface temperatures, and it doesn't match the NASA data on surface temperatures, which can be downloaded here.

Data @ NASA GISS: GISS Surface Temperature Analysis (GISTEMP)

I think that people who model climate for a living are smart enough to understand and account for things like El Nino events, or the current La Nina event. I don't have enough ego to think that a few minutes of my time doing some half-baked (ie, univariate) analysis will somehow provide a more definitive answer than theirs.

Having said that, I downloaded the GISS mean global (land+water) surface temperature data and ran my own regression. Annual observations, linear OLS regression predicting surface temperature anomaly as a function of year. From 1984 (the start year for Hansen's original model (per the realclimate article I posted earlier) to the present (2007, because there is no annual 2008 number yet), I get a trend of 0.21 degrees centigrade per decade, with a standard deviation of 0.027 degrees centigrade. Hansen's 20 year old prediction is well within two standard deviations of the mean.

(In case you wonder why I used annual data, I have to do that to get any reasonable estimate of the standard deviation in this simple analysis, as the monthly data are strongly serially correlated. The monthly data should give me roughly the same mean but (with OLS) would substantially overstate the true precision.)

(Further, the exact start point does not much matter, plus or minus a few years. If I were to start in 1988 instead of 1984 (as Hansen's model did, per the realclimate discussion noted above), the estimated slope would drop from 0.21 to 0.20 degrees centigrade per decade, which I would characterize as a negligible difference.)

For what it's worth -- and it isn't much -- I have attached the plot. Black line is the actual annual global (land and water) surface temperature anomaly (ie difference from baseline period), pink line is the linear regression fit, blue line is Hansen's predicted rate of change. (I've centered both of those straight lines on the midpoint of the data -- so I'm only testing the slopes of the straight lines, not their intercepts as well. I have to do that because the regression line will automatically be centered on the midpoint of the data unless I want to take the time to monkey around with it, which I do not.)

So, a) you can't reject Hansen's 20-year-old prediction for average global surface temperature increase based on this half-baked (ie, univariate regression) analysis, at least not if you actually look at surface temperature data and look at the period Hansen was discussing, b) by eye, the prediction looks pretty good to me, but I'm an economist so my standards are pretty low in terms of predictive accuracy, and c) the world has moved on, and there is little reason to care whether a 20-year-old prediction using methods that are crude by modern standards has or has not stood the test of time. That said, I think Hansen's work has stood the test of time so far. Again, for what that's worth. You'd be far better off reading what the actual climatologists say about it, for example, on realclimate.org.

[Alric]

Originally Posted by TimBikes

I agree it is non-sense to look at individual years.

Even looking at data since only 1979 is cherry picking. But why break it into segments like you did? Your first graph with the most data points is the most valuable.



Even if he is off like you say temperature is still going up. Good enough for me.

[PriusSport]

The country has been governed the past 8 years by a reactionary. Not a conservative. And before that, Clinton was hamstringed by a reactionary congress. The Republicans have gone beyond conservative in their move to the right. They have eroded the tax base, and continued uncontrolled spending with borrowed dollars. And ignored the scientific warnings about environmental abuse.

If the Democrats take over in the fall--and they must--lots will change, because things have to change if we are to survive. It has really become that apocalyptic. It is a question of the rational Center taking over American politics again.

[TimBikes]

Originally Posted by PriusSport

The country has been governed the past 8 years by a reactionary. Not a conservative. And before that, Clinton was hamstringed by a reactionary congress. The Republicans have gone beyond conservative in their move to the right. They have eroded the tax base, and continued uncontrolled spending with borrowed dollars. And ignored the scientific warnings about environmental abuse.

If the Democrats take over in the fall--and they must--lots will change, because things have to change if we are to survive. It has really become that apocalyptic. It is a question of the rational Center taking over American politics again.

Huh? Are you sure you posted in the right thread?

[chogan2]

Just this week, realclimate.org posted a nice analysis of the recent surface temperature trend after filtering out the ENSO (El Nino) variation.

RealClimate

So you don't have to eyeball the data and toss out selected years, you can see what climate scientists actually make of it in detailed modeling.

Their conclusion:

"The basic picture over the long term doesn't change. The trends over the last 30 years remain though the interannual variability is slightly reduced (as you'd expect)."