Temperature Rising
On a recent ABC interview, a well known scientist made the following statement about global temperatures.
Actually, there has been cooling, if you take 1998 as your point of reference. If you take 2002 as your point of reference, then temperatures have plateaued. This is certainly not what you’d expect if carbon dioxide is driving temperature because carbon dioxide levels have been increasing but temperatures have actually been coming down over the last 10 years.
Is she right?
The assertion led denialist commentators such as Andrew Bolt to claim temperatures are reducing and that there was a conspiracy of silence about the dropping temperatures. A graph of global temperatures (from two sources) is displayed below. She is correct on both counts! Temperatures are indeed lower than in 1998. She is also ostensibly correct that temperatures have plateaued since 2002.
It’s all a matter of your point of comparison. The baseline of 1998 she chose was a big peak, which was apparently associated with a large El Nino event. There are lots of these events that need to be taken out before you can see “the trend”. How about the plateau idea? This assumes that 2002 and 2007 are somehow similar in all respects. But if you take into account the 11 year sun cycle (which you can just about see) there should have been lower temperatures in 2007. So a plateau is actually a rise compared to the expected fall. See here for a robust discussion of whether the public comments were consistent with the data.
Even the graph above, which begins in 1978, gives a very different picture to the one below, which starts in 1850. (The trend lines were not added by me, but they don’t look unreasonable).
Or how about going back to the time of Christ? This is mainly based on tree ring data. It gives quite a different perspective.
This all makes me start to question the wisdom of the old adage a picture is worth a thousand words. After removing estimated effects of El Nino and El Nina events and the solar cycle, we might have a picture that would tell us something. (Can anyone point me to such a graph?) But the issues of time period and point of comparison still remain. The problem with graphs is that they can be easily selected to tell just about any story.
OK. I am on a Google inspired roll. Let’s go back 400,000 years! When I see data like this I am filled with a combination of awe and skepticism. It is based on analysing the gas bubbles in ice cores. Let’s accept the science behind this and suppose that it is a reasonable measurement of historical temperatures.
Have a look at the right hand axis. We are talking variations of the order of 10 degrees. The data from the time of Christ that we saw in the third graph above is the uninteresting little stable patch at the right hand extreme.
The graph also shows a time series of CO2 concentration on the same graph. There is a pretty clear association between temperature and CO2. Association does not mean causation but climatologists have theoretical reasons to thing that CO2 drives temperature. And it there is any reverse causality it means a feedback which makes the causality stronger. There is a very obvious cycle in CO2 which roughly predicts high levels of CO2 are to be expected during this millennium, even if human kind were still living in trees. But look at the very end of the red curve. The vertically increasing bit at the end is not a plotting error! If you accept that the association is causal, and the accuracy of the above graphh, you would have to say we are in serious trouble if CO2 levels out at around 370ppmv.
The only argument against is that either the association is not causal (which not many people accept) or that the increasing in CO2 is not as large as it appears in the above plot. This is certianly possible because the time series of CO2 combined ancient ice-core data with recent direct measurements. So a systematic bias in the ice-core CO2 estimates could mean that recent CO2 levels are not historically high.
You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.
April 17th, 2008 at 12:59 pm
The last graph - 400,000 years of ice core data - also displays another statistical feature. The earlier data, more than 200,000 years ago, seems to show less temperature variation as is evidenced by the blue curve appearing to be thinner. Almost certainly this is an artefact due to there being less data available. Given that any discussion of current trends must be relative to typical variability, this is a real problem.
For example, it is often reported that the current rate of change is greater than ever before, but is this verifiable if earlier data records were not frequent enough to demonstrate that a change was quick? I suspect the current rates of change are very high spikes like the current CO2 level are highly unusual, but I have to retain a statistician’s scepticism and say it is not proved as well as I would like.
May 1st, 2008 at 11:12 pm
Would be good to mention that in the ice core data, the temperature rises *precede* the CO2 rises (by about 800 years). The IPCC report has a dense page explaining it.”
May 5th, 2008 at 11:46 am
One wishes that discussion in the media was as well informed as the contributions on this blog to date!
There’s been two media contributions in the past two weeks that comment misleadingly on current temperature trends. Michael Duffy has an article in last Saturday’s Sydney Morning Herald (p.37) that is
headed: “New climate figures would make a great debate - if anyone reported them!”. His point is that there is no trend. He does not point out that solar radiance was at its peak in late 2001 or 2002, then dropping to a low in 2005, with an apparent further slight drop at the beginning of 2007. Thus, in the absence of changes arising from other sources, temperatures ought to have been dropping. Michael Duffy is clearly not interested in an informed debate!
There are a couple of very nice graphs (Figures 4 & 5) on the web page:
http://www.pmodwrc.ch/pmod.php?topic=tsi/composite/SolarConstant
Previous minima (since 1977) were in 197?-1977, 1984-1987 and 1994-1996, roughly.
There’ve been recent signs of new sunspot activity. Solar cycle 24 has officially started. Sometime in the next couple of years, we can expect a double whammy to kick in - rising solar radiance now adding to greenhouse gas effects. The next solar radiance maximum is expected in 2011 or 2012. For high and low predictions through to 2020, see the web page:
http://www.swpc.noaa.gov/SolarCycle/SC24/index.html
Then there were comments from Don Aitkin, former VC of U Canberra, on yesterday’s Ockham’s Razor, following up from his talk a week earlier:
http://www.abc.net.au/rn/ockhamsrazor/default.htm
He makes the truly remarkable statement that:
“A new sunspot cycle is predicted to keep temperatures down over the next decade or so.”
Expected by who? Not till after 2012 or 2013, for sure!
On the ice core data, the explanation for the temperature cycles of around 90,000 years is slow changes in the pattern of the earth’s movement around the sun, and in a top-like movement of its axis. Once temperature changes start, CO2 seems to get released into the atmosphere, from sea and/or grounds sources, and there is a feedback effect that reinforces the temperature changes. The IPPC’s dense page (p.446; orbital forcing is discussed on p.445) is saying, in a convoluted manner, that the details are not well understood. Then one has to explain how the CO2 gets absorbed as the temperature drops.
This all makes sense, even if the details still have to be filled in.
After all, it seems unlikely that regular ~90,000 year visits from space travelers would have somehow generated the CO2, or that volcanos timed major eruptions with the right kind of regularity!
The current experiment with the earth’s climate (CO2 levels now almost 100ppm above the peaks over this earlier time) has no precedent in the past 600,000 years of earth history.
May 14th, 2008 at 3:53 pm
John Henstridge,
I have been reading “With Speed and Violence” by Fred Pearce, a New Scientist author and former editor.
http://www.amazon.com/Speed-Violence-Scientists-Tipping-Climate/dp/0807085766
His thesis is evidence now shows that past natural climate change has been frighteningly rapid and destructive. Once nudged out of a stable state - often by small “forcings” - local or global climate often passes a “tipping point”, with changes happening “virtually overnight”. This realisation of the suddeness of change has only dawned recently, partly as a result of increased resolution of proxy measurements.
Time resolution is also the main issue with variability of the last graph. The highest resolutions are found in ice from regions of high snowfall, but these rarely contain the the oldest ice. Additionally, gas diffusion limits the time resolution of ice cores.
Chris, the Medieval Warm Period (MWP) and the Little Ice Age, were largely local events, concentrated in the Northern hemisphere particularkly around the north Atlantic, which seems to be the measurement site of the plot you have chosen. They are often used by denialists to claim the earth was as warm or warmer in the Middle Ages as it is now.
http://en.wikipedia.org/wiki/Image:2000_Year_Temperature_Comparison.png
And yes, James Franklin temperature rises precede CO2 rises by several hundred years. The triggers for many of the changes during the ice ages are thought to be Milankovitch cycles:
http://en.wikipedia.org/wiki/Milankovitch_cycles
However the small shifts in regional and global heating can only account for a small fraction of the climate change during the ice ages. The rest is thought to be due to positive feedbacks, of which two dominate: CO2 and albedo changes - with methane a distant runner up. As John explains it is not clear what drives CO2 into the atmosphere, but the albedo changes are more prosaic. We are seeing them at the moment - meltwater ponds on ice and vegetation absorb much more radiation than ice and snow. That small initial “forcings” can trigger positive feedbacks that drive climate into another state emphasises the unstable and chaotic nature of our climate system. It’s also scary.
June 2nd, 2008 at 8:58 pm
As I understand it the CO2 changes are due to ocean temperature changes. As the temperature drops CO2 becomes more soluble in water and much of it is absorbed by the oceans. This decrease in atmospheric CO2 leads to further reduction in atmospheric temperature in a positive feedback loop. The delay of atmospheric CO2 concentrations behind temperature changes is due to the long time it takes for CO2 to get into the deep ocean rather than the surface layers.
I understood that they have also analyzed recent air bubbles in the ice. The spike at the end is certainly real. It also has a different carbon isotope composition, fingering a human origin for the increase. Any previous similar spikes would have shown up in the ice cores because it would take a long time for the CO2 to completely return to pre-spike levels.