I’ve moved to WordPress: http://bobtisdale.wordpress.com/

Tuesday, November 24, 2009

What’s Wrong With This Graph?

I’ve moved to WordPress.  This post can now be found at What’s Wrong With This Graph?
#################
The following graph of global temperature anomalies is from the “The Copenhagen Diagnosis, 2009: Updating the world on the Latest Climate Science.” The report was released today with obvious intent. I scrolled as far as Figure 3, then copied it, and closed “The Copenhagen Diagnosis”. I’d had enough. Please feel free to answer the question asked in the title of this post.


http://i45.tinypic.com/2w2ih5y.png
Figure 3 of the “The Copenhagen Diagnosis”

Links to “The Copenhagen Diagnosis, 2009: Updating the world on the Latest Climate Science.”
High resolution PDF (23.3 MB)
Low resolution PDF (3.3 MB)

UPDATE (November 25, 2009): In the comments, Alessandro noticed what I did. He wrote, “The smoothing is not declared and probably the final part ('2000) is corrected.” But I will make my case that those smoothed curves are trends of some sort.

As Alessandro mentioned, they don’t identify the smoothed data. Do the smoothed curves represent filtered data or polynomial trends? If you were to look at the long-term surface temperature anomaly graphs (land, sea, and combined) in Chapter 3 of the IPCC AR4…
http://ipcc-wg1.ucar.edu/wg1/Report/AR4WG1_Print_Ch03.pdf
…(The authors of the "Copenhagen Diagnosis use the IPCC AR4 as a reference) you’d note that the IPCC’s smoothing methods do show plateauing of temperatures toward the end of the data, in 2005. Refer to Figures 3.1, Figure 3.4 Cell A, Figure 3.6, and FAQ3.1 Figure 1.

Figure 3 from the Copenhagen Diagnosis above, however, does not show the decreasing rate of rise. The above graph shows relatively straight lines over the past decade+. This leads me to believe that the smoothed curves are some type of polynomial trend lines. This is confirmed by the discussion of Figure 3 on page 13 of the “Copenhagen Diagnosis”:

“IPCC AR4 presented ‘an unambiguous picture of the ongoing warming of the climate system.’ The atmospheric warming trend continues to climb despite 2008 being cooler than 2007 (Figure 3). For example, the IPCC gave the 25-year trend as 0.177 +/- 0.052 deg C per decade for the period ending 2006 (based on the HadCRUT data). Updating this by including the last two years (2007 and 2008), the trend becomes 0.187 +/- 0.052 deg C per decade for the period ending 2008. The recent observed climate trend is thus one of ongoing warming, in line with IPCC predictions.

“Year-to-year differences in global average temperatures are unimportant in evaluating long-term climate TRENDS. During the warming observed over the 20th century, individual years lie above or below the long-term TREND line due to internal climate variability (like 1998); this is a normal and natural phenomenon. For example, in 2008 a La Niña occurred, a climate pattern which naturally causes a temporary dip in the average global temperature. At the same time, solar output was also at its lowest level of the satellite era, another temporary cooling influence. Without anthropogenic warming these two factors should have resulted in the 2008 temperature being among the coolest in the instrumental era, while in fact 2008 was the 9th warmest on record. This underpins the strong greenhouse warming that has occurred in the atmosphere over the past century. The most recent ten-year period is warmer than the previous ten-year period, and the longer-term warming TREND is clear and unambiguous (Figure 3).” [Caps are mine for emphasis.]

So if these are, in fact, trend lines, the authors of the “Copenhagen Diagnosis” have created a double-edged sword for themselves. Trend lines can be projected forward in time for use in forecasts. That is one of the intents of trend lines. (The trend line options for EXCEL provide the option for extending the trends.) So eyeballing the trend forecasts onto Figure 3, we see a continuous warming over the next decade or two, as the authors intended.
http://i45.tinypic.com/2qiojgn.png
Figure 3 of the “The Copenhagen Diagnosis” With Forward Trend Projections

BUT (BIG BUT) those same trend lines can also be used for projections into the past, for hindcasts. Are they now reinstating the Medieval Warm Period?
http://i48.tinypic.com/vpgbj4.png
Figure 3 of the “The Copenhagen Diagnosis” With Backward Trend Projections

And one last note: As quoted above, the authors of the “Copenhagen Diagnosis” stated, “the IPCC gave the 25-year trend as 0.177 +/- 0.052 deg C per decade for the period ending 2006 (based on the HadCRUT data).”

Actually, the IPCC lists that trend for the period ending in 2005, not 2006. Refer again to FAQ3.1 Figure 1 in AR4 Chapter 3. In the description, they write, “Linear trend fits to the last 25 (yellow), 50 (orange), 100 (purple) and 150 years (red) are shown, and correspond to 1981 to 2005, 1956 to 2005, 1906 to 2005, and 1856 to 2005, respectively.” The specific trend values are listed in the artwork.

26 comments:

HR said...

Darn I wanted you to tell me!!
I got to your website because I googled "NOAA reconstructed sea surface temperature". I thought it strange to mention SST and not actually plot it.

I'm going to guess that's the issue!!

Charlie said...

What's wrong is the innovative smoothing. More specifically how they handled the smooth curve for the last few years.

It's even more disingenuous than the Rahmstorf smoothing hack used in the Copenhagen Synthesis Report a few months ago.

Charlie said...

Hmmm. It is also weird that most (all?) datapoint in the baseline period are well above 0 degrees anomaloy.

But this is just a trivial mislabeling of the Y axis. Nothing to worry about. Right?

HR said...

I just wanted to ask a further question about the CO2 section of the Copenhagen Diagnosis. If this is not relevant to you can you point me in a direction were I might get it answered?

It seems like simple mathematics to me but there's a good chance I'm wrong.
They say that the global rate of CO2 emmission has increased sinced 1990, they also say that the rate of sequestration of CO2 by natural sinks has reduced by 5% over the past 50 years. But when you look at the graph showing CO2 in ppm then it appears that the rate of increase in CO2 in the atmoshpere is steady (i.e. the trend is linear).
Doesn't the first two suggest that we should be seeing an increase in the rate of accumulation of CO2 in the atmosphere?

Anonymous said...

Hi Bob,

I put together a plot using what I feel are equivalent parameters: 1) offsetting the temperatures by 0.3C to have the period 1880-1920 be around 0C, and 2) plotting of the 12-month means:
http://www.woodfortrees.org/plot/hadcrut3gl/from:1850/offset:0.3/mean:12/plot/gistemp/from:1850/offset:0.3/mean:12

Are you referring to the Hadley temperatures around 1860 dipping below those in the 1900-1920 period in the Copenhagen graph? Yeah I can see that, but the overall longer term means for the Hadley data during the period 1860-1900 are pretty much in line around the 0C area.

Even with that slight difference, I still see the general 0.8C temperature change for the last 100 years being pretty much in agreement on both graphs (both the one you copied and the one I plotted).

Dennis H.

Anonymous said...

Bob

Rather ironic given Jones' comment about not changing the anomaly comparison period until after he'd retired.

Anonymous said...

hide the decline?

Stewart said...

It is just plain dodgy! Apart from the smoothing, the obvious problem is the use of the 1880-1920 mean.
Look at the graph - 1880 starts just after a spike in temperatures, and follows a nice downward slope until 1920.
The choice of this period seems, scientifically speaking, arbitrary and unjustifiable. Speaking in marketing terms though, it's great - since it makes the anomoly look bigger!

As a point of reference, AR4 used 1961-1990 on page 31, and 1901-1950on page 40 - both somewhat more reasonable baselines. I think 1961-1990 is generally considered standard for talking about anomolies in today's terms.

Stewart said...

Using the mean period of 1880-1920 is dodgy. IPCC AR4 used 1961-1990.

1880-1920 seems entirely arbitrary and simply calculated to exaggerate the warming trend.

Espen said...

Charlie is right, what's wrong is the thick smoothed lines, which look like centered running averages, but which obviously is something different since the period must be in the ~30 year range and they go right to the end of the graph. Some variant of Rahmstorf's filters ("non linear trends") again?

Unknown said...

I would suspect the question to ask, (given that many high temperature records around the World are in the late '30s), what happened to the '30s & '40s?

Dave.

Alessandro said...

The smoothing is not declared and probably the final part ('2000) is corrected.

Demesure said...

Like Charlie said, the smoothed curves span is same as raw points', which is only possible with "innovative" smoothing (padding with hide-the-decline synthetized points).
And the post 70s warming is about nearly 50% more than what is found in archives (see same graphs in wayback machine)

Nothing new in the graph, except more flagrant cheating.

Anonymous said...

How could the two independently prepared temp series effectively merge for the last 30 years without collusion?

BarryW said...

Oh jeez, if the graph is relative to the 1951-1980 time frame then it's zero point should be there not at the beginning of the graph. Crappy propaganda.

Basil said...

Well, as others have observed, the smoothing is suspicious. But contra some comments, no padding is necessary to smooth through the end of the series. Here is straight forward Hodrick-Prescott smoothing applied to HadCRUT3, through October 2009:

http://i45.tinypic.com/243s3uv.jpg

So what is missing, of course, is the downturn since 2004. I.e., it is another clever attempt to "hide the decline." Even if one questions "my" smoothing, we have Trenberth acknowledging that global warming has at least stalled, which is obscured by this smoothing.

Finally, the base period is silly.

Basil said...

Something else to point out about the smoothing: it pretty much eradicates the 1940's "blip" that these blokes are so worried about. It is a pretty odd smoothing algorithm that has ALL the data points above the smooth during this period. I think we can pretty much rule out any kind of "least squares" smoothing technique here.

Unknown said...

this thing is a piece of work!

in addition to the undeclared smoothing and end point correction Charlie and Stewart point out a couple of elementary issues:

1. 1880-1920 and what is the justification? "Yields the most impactful result" does not count.

2. My eyeball says there ain't no way the anomaly for the declared calibration period is zero.

Bob Tisdale said...

HR: You asked, "I just wanted to ask a further question about the CO2 section of the Copenhagen Diagnosis. If this is not relevant to you can you point me in a direction were I might get it answered?"

Sorry. I don't pay attention to CO2. I don't find any relationship between the increase in it and the variations in Ocean Heat Content.

Anonymous said...

Bob: "BUT (BIG BUT) those same trend lines can also be used for projections into the past, for hindcasts."

I think you are becoming a little too paranoid and cynical, Bob.


Are you saying that if you or any data analyst ever uses trend lines on cyclical or linear data bases to better understand how the data has varied over time, should always be interpreted as the author purposely intending the reader to always consider it a forecast tool far into the infinite future?

Does that mean that any of your SST charts that have linear trend lines and decadal changes plotted on them should also be interpreted as your intention to use that linear trend to predict far into the future too?
I will remind you of this the next time you use linear trend lines on your charts.

My own scientific interpretation of the plot would be that since cyclical patterns obviously occurred in the observed data, I would then safely assume that some sort of cyclical pattern would also occur well into the future and the latest linear trend would not. It is very likely that the polynomial equation and trend lines will be modified in the future as new data is included in the chart.

Dennis H.

Bob Tisdale said...

Dennis H: You wrote, "I think you are becoming a little too paranoid and cynical, Bob."

Paranoid, no--but I'm a skeptic, Dennis, so I always look for underlying motives.

You wrote, "My own scientific interpretation of the plot would be that since cyclical patterns obviously occurred in the observed data, I would then safely assume that some sort of cyclical pattern would also occur well into the future and the latest linear trend would not."

But based on your opening clause, you've been trained as a scientist, Dennis. The intended audience for the "Copenhagen Diagnosis" (clever title, I must admit) are not scientists.

Anonymous said...

Bob: "But based on your opening clause, you've been trained as a scientist, Dennis. The intended audience for the "Copenhagen Diagnosis" (clever title, I must admit) are not scientists."

Yes, but neither are most of the commentators on Accu-Weather's Global Warming blog site that religiously follow (and believe with all their hearts) that the data and one-sided political rants presented on ICECAP's propaganda site as being the truth. But I don't see you providing the same sort of "public service" and criticizing any of their obviously underlying, biased motives with their graphs and charts (mostly with cherry picked short-cycle trend lines) with the same sort of data implications on the opposite side.

If you want to come across as being fair and show some "neutrality" on your web site, and in order to educate the non-scientists out there, then it would be nice if you apply the same level of scrutiny and criticism with those biased data charts and stories they produce and call them out just as well.


Case in point, all the trend lines produced since the 1998 El Nino peak by groups like ICECAP that were trying to convince their non-scientist readers and believers that the short-term trends that they wanted their believers to all concentrate on (and to basically ignore all the long-term trends) and that these short-term trends would continue for some period into the future (using "hand-waving and smoke-and-mirrors" data correlations with things like solar minimums and sunspot activity, etc.) to send us into the next "Maunder Minimum". The problem is that their biased prediction trend lines were pretty much wiped out very quickly by the most recent rise in global temperatures during this moderate El Nino pattern (even though solar activity has still yet to come off the bottom: http://www.swpc.noaa.gov/SolarCycle/). That is why the current CRU email scandal appears to have come during a conveniently well-timed period so that groups like ICECAP will not have to own up to their recently poor predictions based on imagined correlations and short-term trends.

The long-term global temperature data trends will not be changing as the result of anything coming forth from all this public discussion over those cherry picked emails. The long-term trends are still intact, even with your little criticism with the Copenhagen charts. The magnitude of these trend lines may change slightly over time through newly identified cyclical pattern changes using further polynomial analysis, but the general (big picture) positively (upward) global temperature trends are and will continue to be in effect for some time to come: http://www.woodfortrees.org/plot/
hadcrut3gl/from:1850/offset:0.39/
mean:12/plot/gistemp/from:1850/
offset:0.3/mean:12/plot/uah/
from:1970/offset:0.54/mean:12/
plot/rss/from:1970/offset:0.54/
mean:12.

Those long-term temperature trends will continue with or without the created distraction of a public scandal or your one-sided scrutiny of the underlying motives of data analysis, besides your own.

Dennis H.

Bob Tisdale said...

Dennis H: You wrote, "Yes, but neither are most of the commentators on Accu-Weather's Global Warming blog site that religiously follow (and believe with all their hearts) that the data and one-sided political rants presented on ICECAP's propaganda site as being the truth. But I don't see you providing the same sort of 'public service' and criticizing any of their obviously underlying, biased motives with their graphs and charts (mostly with cherry picked short-cycle trend lines) with the same sort of data implications on the opposite side."

This post is solely about the graph in the "Copenhagen Diagnosis" and the verbage attached to it. Nothing more, nothing less.

And why would you expect me to be neutral? You know my position very well. We've been discussing these same topics for some time now.

Anonymous said...

Hi Bob,

Oh yeah that's right, the ICECAP site (specifically Joe D'Aleo) has used some of your postings on their propaganda site, so you now feel obliged not to criticize any of the underlying motives of the trend lines they use on their graphs. Are they scratching your back too with any financial kickback when they copy and use your posts? See, I can be a skeptic when it comes to those types of underlying motives too.

I can't remember your exact words, but somewhere in the back of my mind when I first starting bantering with you on the ACCU-Weather site about your own web site, you told me that all you were trying to do was provide data analyses without any expressed personal opinion or commentary. By simply ignoring and not criticizing the same data analysis techniques being used by your "friends" that you are purposely emphasizing in this posting about the same techniques being used by your "enemies" makes you look less credible as simply a data analyst you portrayed yourself as being.

Dennis H.

Bob Tisdale said...

Dennis H: Regarding your November 27, 2009 1:56PM comment, I have had my motivations questioned by AGW proponents, who assumed they were based on politics. They are not. Now you’re accusing me having financial motivations, of receiving kickbacks from IceCap? That’s a ridiculous debate tactic, Dennis. If you knew my financial situation, you’d think it was as funny as I do.

Now, if you had read my comments on this thread at WUWT...
http://wattsupwiththat.com/2009/09/03/california-wildfires-caused-by-cooler-pacific-la-nina/
...and read the exchange between Joe D’Aleo and me on the following thread at WUWT...
http://wattsupwiththat.com/2009/09/26/correcting-the-surface-temperature-record-for-uhi/
...you’d have discovered that your assumption that Mr. D’Aleo and I share the same viewpoints is in error.

You wrote, "...you told me that all you were trying to do was provide data analyses without any expressed personal opinion or commentary."

You'd have to find that exchange for me at Accuweather, Dennis. I don't recall exactly what I wrote or in what context it was written. Most of my early SST posts were simply descriptions of the graphs, the of data. That has obviously changed. I can even tell you when it changed. That changed when I discovered the step changes in East Indian and West Pacific SST anomalies and in the TLT anomalies of the Mid-To-High Latitudes of the Northern Hemisphere. That changed when I realized the significance of those step changes.

Regards

Anonymous said...

"They say that the global rate of CO2 emmission has increased sinced 1990, they also say that the rate of sequestration of CO2 by natural sinks has reduced by 5% over the past 50 years."

This is one of those misleading IPCC statements. Uptake of CO2 by both land biosphere and oceans is higher than ever in recent years (about 4 GtC/yr), however since the amount of man-made CO2 put out has gone up, the proportion that has been sequestered has gone down from about 50% several decades ago to about 45% now.

If you think about it, though, the efficiency of carbon sinks depends on how much CO2 is in the atmosphere, not how much we are putting out in a particular year. If you plot THAT trend, it suggests ever greater ability to sequester CO2 as the CO2 ppm ratio increases.

Donations

Tips are now being accepted.

Comment Policy, SST Posts, and Notes

Comments that are political in nature or that have nothing to do with the post will be deleted.
####
The Smith and Reynolds SST Posts DOES NOT LIST ALL SST POSTS. I stopped using ERSST.v2 data for SST when NOAA deleted it from NOMADS early in 2009.

Please use the search feature in the upper left-hand corner of the page for posts on specific subjects.
####
NOTE: I’ve discovered that some of the links to older posts provide blank pages. While it’s possible to access that post by scrolling through the history, that’s time consuming. There’s a quick fix for the problem, so if you run into an absent post, please advise me. Thanks.
####
If you use the graphs, please cite or link to the address of the blog post or this website.