Comments
Empirical Analysis of the Solar Contribution to Global Mean Air Surface Temperature Change
Nicola Scafetta has a new paper in the Journal of Atmospheric and Solar-Terrestrial Physics: ‘Empirical analysis of the solar contribution to global mean air surface temperature change’ (doi:10.1016/j.jastp.2009.07.007)
The Abstract states:
The solar contribution to global mean air surface temperature change is analyzed by using an empirical bi-scale climate model characterized by both fast and slow characteristic time responses to solar forcing. Since 1980 the solar contribution to climate change is uncertain because of the severe uncertainty of the total solar irradiance satellite composites. The sun may have caused from a slight cooling, if PMOD TSI composite is used, to a significant warming (up to 65% of the total observed warming) if ACRIM, or other TSI composites are used. The model is calibrated only on the empirical 11-year solar cycle signature on the instrumental global surface temperature since 1980. The model reconstructs the major temperature patterns covering 400 years of solar induced temperature changes, as shown in recent paleoclimate global temperature records.
The Conclusion states:
Herein I have analyzed the solar contribution to global mean air surface temperature change. A comprehensive interpretation of multiple scientific findings indicates that the contribution of solar variability to climate change is significant and that the temperature trend since 1980 can be large and upward. However, to correctly quantify the solar contribution to the recent global warming it is necessary to determine the correct TSI behavior since 1980. Unfortunately, this cannot be done with certainty yet. The PMOD TSI composite, which has been used by the IPCC and most climate modelers, has been found to be based on arbitrary and questionable assumptions (Scafetta and Willson, 2009). Thus, it cannot be excluded that TSI increased from 1980 to 2000 as claimed by the ACRIM scientific team.
The IPCC (2007) claim that the solar contribution to climate change since 1950 is negligible may be based on wrong solar data in addition to the fact that the EBMs and GCMs there used are missing or poorly modeling several climate mechanisms that would significantly amplify the solar effect on climate. When taken into account the entire range of possible TSI satellite composite since 1980, the solar contribution to climate change ranges from a slight cooling to a significant warming, which can be as large as 65% of the total observed global warming.
The above wide range strongly contrasts with some recent estimates such as those found by Lockwood, 2008 M. Lockwood, Recent changes in solar output and the global mean surface temperature. III. Analysis of the contributions to global mean air surface temperature rise, Proc. R. Soc. A 464 (2008), pp. 1–17 10.1098/rspa.2007.0348.Lockwood (2008), who calculated that the solar contribution to global warming is negligible since 1980: the sun could have caused from a -3.6% using PMOD to a +3.1% using ACRIM. In fact, Lockwood’s model is approximately reproduced by the ESS1 curve that refers to the solar signature on climate as produced only by those processes characterized with a short time response to a forcing. Indeed, the characteristic time constants that Lockwood found with his complicated nonlinear multiregression analysis are all smaller than one year (see his table 1) and the climate sensitivity to TSI that he found is essentially equal to my k1S! Likely, Lockwood’s model was unable to detect the climate sensitivity to solar changes induced by those climate mechanisms that have a decadal characteristic time response to solar forcing: mechanisms that must be present in nature for physical reasons. As proven above, these mechanisms are fundamental to properly model the decadal and secular trends of the temperature because they yield high climate sensitivities to solar changes.
Analogously, my findings contrast with Lean and Rind (2008), who estimated that the sun has caused less than 10% of the observed warming since 1900. The model used by Lean and Rind, like Lockwood’s model, is not appropriate to evaluate the multidecadal solar effect on climate. In fact, Lean and Rind do not use any EBM to generate the waveforms they use in their regression analysis. These authors assume that the temperature is just the linear superposition of the forcing functions with some fixed time-lags. They also ignore ACRIM TSI satellite composite. While Lean and Rind’s method may be sufficiently appropriate for determining the 11-year solar cycle signature on the temperature records there used, the same method is not appropriate on multidecadal scales because climate science predicts that time-lag and the climate sensitivity to a forcing is frequency dependent. Consequently, as Lockwood’s model, Lean and Rind’s model too misses the larger sensitivity that the climate system is expected to present to solar changes at the decadal and secular scales.
I have shown that the processes with a long time response to climate forcing are fundamental to correctly understanding the decadal and secular solar effect on climate (see ESS2 curve). With simple calculations it is possible to determine that if the climate parameters (such as the albedo and the emissivity, etc.) change slowly with the temperature, the climate sensitivity to solar changes is largely amplified as shown in Eq. (10).
This finding suggests that the climate system is hypersensitive to the climate function h(T) and even small errors in modeling h(T) (for example, in modeling how the albedo, the cloud cover, water vapor feedback, the emissivity, etc. respond to changes of the temperature on a decadal scale) would yield the climate models to fail, even by a large factor, to appropriately determine the solar effect on climate on decadal and secular scale. For similar reasons, the models also present a very large uncertainty in evaluating the climate sensitivity to changes in CO2 atmospheric concentration (Knutti and Hegerl, 2008). This large sensitivity of the climate equations to physical uncertainty makes the adoption of traditional EBMs and GCMs quite problematic.
About the result depicted in Fig. 6, the ESS curve has been evaluated by calibrating the proposed empirical bi-scale model only by using the information deduced: (1) by the instrumental temperature and the solar records since 1980 about the 11-year solar signature on climate; (2) by the findings by Scafetta (2008) and Schwartz (2008) about the long and short characteristic time responses of the climate as deduced with autoregressive models. The paleoclimate temperature reconstructions were not used to calibrate the model, as done in Scafetta and West (2007). Thus, the finding shown in Fig. 6 referring to the preindustrial era has also a predictive meaning, and implies that climate had a significant preindustrial variability which is incompatible with a hockey stick temperature graph.
August 20th, 2009 at 11:16 pm
Did you mean to say “Nicola Scafetta” in your opening sentence?
August 21st, 2009 at 1:23 am
Yes, thanks. Now fixed.