Tuesday
Dec022014
by 
Bishop Hill
Bishop Hill Benny at the Senate
Dec 2, 2014 Climate: Parliament Benny Peiser's testimony at the US Senate is now available on YouTube.
Dec 2, 2014 Climate: Parliament 
Buy
Books
Click images for more details
Currently discussing
-
Dec 28 - derreg
on
Curiosidade na Madrugada -
Dec 28 - kekuss
on
Rushed Office Relocation!
Links
A few sites I've stumbled across recently....
-
Science blogs
-
Political blogs
-
Other blogs
Subscribe
Copyright © 2006-2013, Bishop Hill All rights reserved.
Reader Comments (108)
Hamster
Why are so keen for me to join in?
EM What is the obsession with fitting linear regressions to time series data? Anyone can eyeball the data and see recent temperatures are warmer than 20 or 30 years ago. That is all the regression is telling you. It says nothing about a pause, hiatus, plateau or whatever you call it. If it were to continue for another 100 years the regression would still show a positive gradient.
Edit, wrong thread, possibly wrong blog, tis been a long day!
Hay hum...
Paul Dennis
Actually the regression lines are useful. Over a number of doublings the warming effect of increasing [CO2] would be expected to be natural logarithmic. Over the 34 years Mike Jackson has us observing, the trend is close enough to linear to be indistinguishable.
If the trend data is a good fit to linear it supports my case. If not, it helps falsify it.
The post 1980 data can be interpreted as 18 years of rapid warming, after which warming stopped.
Alternatively the same data could be interpreted as 18 years of moderate warming followed by a brief rapid warming event and a subsequent reversion to the trend.
Since the 2014 data is on the trend line the two alternative interpretations can now be tested.
If the first interpretation is correct, the temperatures in years to come would be expected to remain similar. If the second is correct, one would expect the warming trend to resume. I await thenext few years with interest. :-)
EM you need to start thinking critically like a scientist and not as an activist. My comment about linear regressions stands.
You've suggested a model to me that is temperature is proportional to the log of the partial pressure of CO2 in the atmosphere. Then evaluate this and don't plot temperature as a function of time. Do the plot for the satellite record and do it for the instrumental record. There's instrumental data for the last 100 maybe 150 years, direct measurements of the CO2 content of the atmosphere for a significant chunk of this time and good estimates prior to this.
Compare the data and critically evaluate it. Always be objective, ask questions of it, try to find ways to test your hypothesis. So far all we get is post-hoc arm waving and weasel words such as 'is consistent with'. To me the data is also pretty much consistent with long term persistence in the climate record as a result of stochastic processes.
Maybe there is a co-incidental covariation of CO2 and temperature - correlation is not causation. So also test during periods when there is not a monotonic trend in CO2.
Perhaps one should be looking to test the model with the palaeoclimate record. Unfortunately that record is a complete mess and perhaps needs to be thrown out and a new start made to try and determine past climate using new proxies for which we have a good phenomenological description and constitutive models that relate proxy response to temperature changes. Such a model has to be rooted in the thermodynamic response of a system. For example the isotopic composition of water and ice is determined by the phase diagram for water (notably the liquid-vapour and solid-vapour phase boundaries and the bond energies of isotopic molecules), There's also the isotopic composition of carbonate minerals and now also isotopic ordering in the lattice.
In all such systems it is possible to a-priori determine the response to temperature using the constitutive equations. This is not possible for tree rings, ecological studies based on mutual climatic range etc) where more often than not when sampling is done out of the test/calibration periods the response does not agree with the calibration. Witness the recent tree ring discussions.
I've got bored with BH, WUWT, CA, SS and RC because the debate is now so sterile, rather like children squabbling.'...he said this, no she said this.....'. There's a lot of heat being generated but precious little light.
Paul, you say correlation is not causation but Entropic is quite firmly confused on this matter and believes a hypothesis can be 'tested' by thinking about it and drawing a regression line through observational data - if the line fits, the hypothesis has been tested.
'Since the 2014 data is on the trend line the two alternative interpretations can now be tested.'
PD. Fair posts for a BH entry. Know what you mean re these sites....but I guess as an AGWer I would say that.
To be fair I don't think EM is just doing the simple correlation/causation thing here. If he were say correlating something spurious (say number of people reading Harry Potter books with temperature change) then criticisms would be fair. However CO2 has form in this area so the supposition is entirely rational. Of course you then need to look at other factors but if there were no increasing temperature levels (itself a proxy for overall energy/entropy increase within a closed system) with increased CO2 levels then that would be more strange and its mitigant would need to be looked at.
This just appeared at tamino's