Ben Santer to Andy Revkin

7:22 p.m. Update

Ben Santer sent two reactions by e-mail, one of which is here and the other — far more technical — is added as a comment:

The bottom line is that the identification of human effects on climate is a signal-to-noise problem. A human-caused warming signal is embedded in the rich, year-to-year and decade-to-decade noise of natural internal climate variability. Scientifically, we never had the expectation that there would be some monotonic warming signal in response to slow, human-caused changes in greenhouse gases, with each year inexorably warmer than the previous year. In detection and attribution studies, we beat down the large noise of year-to-year and decade-to-decade variability by looking at changes over longer sweeps of time. When you consider the entire satellite era (1979 to present), signal-to-noise ratios for global-scale changes in lower tropospheric temperature now exceed 5 – even for UAH lower tropospheric temperature data (see…”fact sheet“). This is what the discussion should focus on – the signal rather than the noise.

and in the comments Revkin pasted the 2nd item:

Ben Santer of Lawrence Livermore sent this response to my e-mail query along with what I’ve added at the bottom of the post:

I’m currently in the process of updating (with CMIP-5 simulation output) the analysis described in our 2011 JGR paper. Recall that the 2011 JGR paper was based on the analysis of older CMIP-3 simulations of forced and unforced climate change.

Our recent PNAS paper [ http://j.mp/pnassanter12 ] indicates that tropospheric temperature variability on 5- to 20-year timescales is, on average, larger in CMIP-5 than in CMIP-3 models. So based on the analysis of CMIP-5 simulations, it is likely that it will take longer than 17 years to discriminate between internal “climate noise” and an externally-forced tropospheric warming signal.

As described in both our 2011 JGR paper (see paragraphs 36 and 38) and our 2012 PNAS paper (see page 3), there are a number of possible explanations for differences between observed temperature trends and model trends in simulations of historical climate change. Dr. John Christy and Dr. Patrick Michaels claim that such differences are entirely due to model response errors. Such claims are scientifically incorrect. Errors in the imposed forcings – particular the anthropogenic aerosol, stratospheric ozone, solar, and volcanic forcings – remain a serious concern. And as the history of the MSU debate has taught us, we certainly cannot rule out residual errors in the observations.

Jan. 9, 2013 at 7:30 p.m

Congratulations – just noticed that you and Ross have finally got your paper accepted at Atmospheric Science Letters. (It will be interesting to see if there is a response from RC).

http://www3.interscience.wiley.com/journal/4735/home/ForAuthors.html

“Submitted manuscripts should not have been previously published and should not be submitted for publication elsewhere while they are under consideration by the Journal.”