Looking at the temperature trends from 1900-2008, it is not clear that there is a carbon dioxide signal.
In a recent post I looked at how base periods can create an artificial upwards temperature trend;
The 20thC featured 2 El Nino dominated climate patterns (+ve PDO), and one La Nina phase (-ve PDO) from 1940-1976. The temperature trend in the first +ve PDO is almost identical to the temperature trend in the second +ve PDO;
The similar slopes at the beginning and end of the 20thC represent warming WITHIN the +ve PDO’s, while the lower starting point for the first +ve PDO is an artifact of the 1951-1980 GISS base period. The GISS graph also shows post 1998 temperatures as increasing. This is contradicted by the other temperature data collectors, which show a decreasing trend consistent with the emergence of another –ve PDO post 2001 (discussed below). The issue is, what would be the temperature trend be with ENSO removed and what part would CO2 play in causing that residual trend?
In a recent paper, David Douglass and John Christy isolate a temperature trend due to CO2 forcing, independent of feedback (ie: the enhanced greenhouse) and natural factors such as ENSO and volcanic effect;
Douglass and Christy’s (DC) study is based on 1979-2008 UAH non-surface data. After extracting ENSO, volcanoes and allowing for latitude band effects, they isolate a CO2 signal of+0.070g/decade; where g is the gain due to any feedback. In respect of ‘g’ DC note “there is general agreement among climate scientists for the case of no feedback”. (p3).
DC estimate there is an undeducted solar irradiance forcing (SF) of 20% (p10), or +0.014C per decade. This generally agrees with AR4’s figure for SF of +0.12Wm-2, which translates to a temperature of +0.16C per century (see Chp 2 pp 187-193). AR4 has reduced this SF figure from TAR’s estimate of +0.3Wm-2, or a temperature increase of approximately 0.4C PC (see 188.8.131.52; FIG 6). The AR4 amount for SF is based on the period from 1750-present, but, according to FIG 2.17, the bulk of the SF has occurred in the 20thC. DC’s SF estimate seems about right then.
So, deducting DC’s SF from +0.07 – +0.014 = +0.056C PD for a CO2 signal in the period 1979-2008.
However, DC note that “the global atmospheric temperature anomalies of Earth reached a maximum in 1998 which has not been exceeded during the subsequent 10 years”. (Abstract). As noted above, GISS is showing increasing post 1998 temperature, so what is happening in the 21stC?
In an analysis based on the period 2001-2008 Lucia also removed ENSO from 5 of the temperature indices;
For a full discussion of Lucia’s analysis see;
Lucia has applied 2 statistical approaches to all post 2000 data, GISS, HadCrut, NOAA, UAH and RSS, and obtained a combined result for OLS of -0.3C(+-1.6) PC, and for Cochrane-Orcutt, -0.6C(+-1.5) PC.
Averaging the 2 methodologies gives an ENSO free temperature trend for 2001-2008 of-0.45C or a decadal trend of -0.045C. Lucia has not adjusted for volcanoes as there were no proximate eruptions, or for SF. If an offset for SF of +0.014C is made, this would produce an underlying cooling trend of -0.059C PD, presumably due to CO2.
So, in summary:
1. AR4 notes that “Most of the observed increase in global average temperatures since the mid-20thC is very likely due to the observed increase in anthropogenic gas concentrations” (Executive Summary, CHP 2)
2. AR4 allocates a Radiative Forcing to the combined GHG’s of 2.63Wm-2; CO2 is allocated a RF of 1.66Wm-2, or 2/3’s of the total RF.
3. The RF for CO2 is estimated by AR4 to lead to an increase in temperature from a doubling of CO2 of ~ 3C. CO2 has increased ~ 40% since 1900. This should have produced a temperature increase of 1.2C or 0.12C PD.
4. Applying AR4’s quotient for CO2 RF of 2/3 to the findings of DC and Lucia we obtain the following CO2 signals; DC = +0.056 3 x 2 = + 0.037C PD for the period 1979-2000; for Lucia = -0.059 3 x 2 = -0.039C PD for the period 2001-2008.
5. A further complication applies to the first ½ of the 20thC temperature trends. There was less CO2 and GHG’s prior to 1976, yet the temperature trends at the beginning of the 20thC, as shown by GISS above and HadCrut are very similar; http://i32.tinypic.com/2s01m5y.jpg
6. Then, of course, there is the 30 year decline in temperatures from 1940-1976 when CO2 was increasing.
7. DC and Lucia have found a CO2 signal. It is inconsistent, I draw 3 conclusions;
a) The inconsistency found by DC and Lucia reflects the contrary movements of CO2 and temperature apparent during the rest of the 20thC and history generally.
b) IPCC forcing estimates for CO2 are grossly over-inflated. Even more so when enhanced greenhouse, “g”, is quantified with +ve feedback.
c) In respect of “g”; if the CO2 signal is larger than that found by DC and Lucia, then –ve feedbacks would have to be much greater. These –ve feedbacks cannot be aerosols (see DC p 12), or ENSO as suggested by Keenlyside et al. Perhaps climate sensitivity to SF is greater than AR4 assumes.