from: "Lockwood, M (Mike)" <mike.lockwoodatXYZxyzc.ac.uk>
subject: RE: Quick reply and another quick question ....
to: "Phil Jones" <p.jonesatXYZxyz.ac.uk>
Yes mostly the difference between TIM/SOURCE is the zero-level offset - although TIM does show very slightly less relative decline into the low values of the current solar minimum than does SoHO/VIRGO etc. (but this is a small relative difference)
As to the merits if the absolute calibration of TIM - it should have been the best yet and so it was a bit of a shock when it gave so much lower values than the previous instruments. In truth I havnt read much which would confirm it really is correct. I remember reading a paper by Kopp (Solar Physics, 2005 I think) who revised the expected error budget based on in-flight calibrations and they quote a 1-sigma error which was about a tenth of the difference to VIRGO and ACRIM, respectively. They were still very cagey about stating outright that their instrument was right and the others were wrong at that stage. I'd put that more down to worthy scientific caution than any insecurity on their part: but I have no more than that to go on and would be interested if there is a more concrete statement to use now.
I have a question for you in this context. I gave a paper at NAM 2 weeks ago where I talked about the low solar minimum. Some folk are already latching on to the "stalled rise" in your global temp anomaly and saying it is a response to the TSI decline to the current minimum. So I used your HadCRUT3 to take a look at this. What I did is so far from rocket science, I can explain in a few sentances......
I took a range of timescales T on which one tries to define a trend. I then took running means over each T and looked at the pdfs of past deviations from those running means. I the assumed a range of trends applied over the last T/2 years and worked out the probability of the observed integrated deviations from each assumed trend from the past deviation pdfs. I found that if you take T = 6 years then yes the most likely recent trend is down (-2 deg/century over the past 3 years) but the distribution is very broad indeed and so probability of a rise of say +2degrees/century is still considerable If you take T = 9 years the most likely value is zero and if you take T of 15 years or above the most likely current trend is +2 deg/c and the distribution is very narrow, so there is only an infinitessimal probability that the trend is really down. I know this is intuitively very obvious (to all but the most obtuse) but is it worth writing up to make it clear and quantified? If so, do you have or know of similar papers?
From: Phil Jones [mailto:p.jonesatXYZxyz.ac.uk]
Sent: Thu 07/05/2009 10:36
To: Lockwood, M (Mike)
Subject: Quick question
Heard a talk last week at a GCOS meeting in Geneva by Bob Cahalan.
I presume you accept that the Tim instrument on SORCE is giving the
of TSI? Presumably also everybody would agree that all this does is lower
TSI by about 5W/m**2 to ~1361.
What this means for the planet and for climate models is a quarter
of this (so about 340 incoming).
The next IPCC Report runs (for AR5 in 2013) will still be using
what they've recently
used - to recalibrate a GCM takes a year or two.
I'm sure there will be some who'll say that GCMs shouldn't be
believed as they are not
using the wrong value for TSI. There will also be confusion about
drop of 5 and forgetting
to divide this by 4!
Prof. Phil Jones
Climatic Research Unit Telephone +44 (0) 1603 592090
School of Environmental Sciences Fax +44 (0) 1603 507784
University of East Anglia
Norwich Email p.jonesatXYZxyz.ac.uk
Scanned by iCritical.