Wednesday, September 21, 2016

Common Era Temperature Reconstructions vs. Solar and Volcanic Forcing

... because wiggle matching is so much fun.


The original inspiration for doing these plots goes back a bit, but have recently become topical at Judith Curry's by way of a guest post, Impact of the ~ 2400 yr solar cycle on climate and human societies, written by a fellow called Javier.  It's a lengthy post, and I'm not going to attempt to dig into it here.

Instead I'll simply throw out ...

Some Plots

The method to this madness:
  1. Regress CO2 and Berger 65N summertime insolation over 500 CE-present, where present is the latest year available in the given temperature reconstruction.
  2. Compute a residual.
  3. Regress the residual over 1600 CE-present.
  4. Combine both regressions into a single "prediction" with simple addition.
As should be obvious from the legends in each plot, I did some fiddling with moving averages and lead/lags to obtain better fits.  I've previously done more aggressive manipulations, particularly with 65N summertime insolation to make things fit better.  This time I didn't touch insolation or CO2 at all, just left them as is and tweaked TSI and volcanic aerosols.

The attentive reader may be asking why I didn't regress TSI and volcanic aerosols back to 500 CE as I did for CO2 and 65N summertime insolation.  The answer is because it doesn't work at all for Oppo 2009, and I wanted to be more consistent about my regression intervals than less.

What can we glean from this exercise?  My main takeaway is that paleo reconstructions such as these are difficult and fraught with uncertainty.  That said, the plots above are ranked in order of how much sense they make to me, with Mann 2008 making the most sense from an orbital forcing perspective, although Oppo 2009 doesn't do so badly in that respect either.

Mann and Moberg have the better fidelity to TSI and volcanoes.

That Moberg and Oppo broadly agree on trends during the Medieval Warming Period (for purposes of this post, let's call it between 800-1300 CE) is certainly an argument for something funkadelic going on which cannot be explained by orbital forcing, TSI, CO2 or volcanoes.

What none of these plots diminish is the importance of CO2 from 1800 CE-present, though somewhat ironically, Mann's Hockey Stick returns the lowest CO2 sensitivity of the bunch.  I wouldn't read too much into that particular result, however, because ... well because calling this kind of amateur twiddling an "analysis" is probably being overly-generous.

Update 9/25/2015

There's been some discussion about Liu et al. (2014), The Holocene temperature conundrum.  So I ginned up two more plots as a response:

I shouldn't have to write much else, these plots more or less speak for themselves.

One note is appropriate:  CO2EQ is a calculated composite of CO2, CH4 and N2O data taken from Antarctica.  Gaps between the ends of those records and the present were filled in using the mixing ratios used in AR5 for the CMIP5 historical runs.

Calculation of the composite is as follows:
  1. Take the natural log of CO2 (ppmv) and the square root of CH4 and N2O (ppbv).
  2. Multiply by 5.3525, 0.0315 and 0.1131 respectively, which gives the radiative forcing for each species in W/m^2.
  3. Sum the forcings and divide by 5.3525, which gives the equivalent of ln(CO2).  This is the value I used in the regression.
  4. If desired, raising e to the power of the result of (3) gives CO2EQ in ppmv.