November 2, 2006

Messy Models

The global warming scare comes largely, if not exclusively, from the outputs of numerical climate models. Some of the models are relatively simple in their design while other climate models are among the most sophisticated computer programs ever built. When the concentration of greenhouse gases is increased numerically, almost all models of climate show an increase in global temperature with the most warming occurring in the Northern Hemisphere’s highest latitudes. Predictions involving precipitation, drought, hurricanes, floods, changes in climate variability, and all the rest vary considerably from model to model. Many greenhouse advocates treat the 2 ×CO2 model simulations as predictions for the future with little regard for shortcomings in the way the models numerically represent the 1,000s of complex processes at work in the climate system.

The Intergovernmental Panel on Climate (IPCC) change warns in their Summary for Policymakers “models cannot yet simulate all aspects of climate (e.g., they still cannot account fully for the observed trend in the surface-troposphere temperature difference since 1979) and there a particular uncertainties associated with clouds and their interaction with radiation and aerosols.” Two articles have appeared in the scientific literature recently that further expose the weaknesses in the model simulations.

As we mentioned previously, two scientists from the University of North Carolina assembled solar output and global temperature data for the period 1900-2000 and the results are published in a recent issue of Geophysical Research Letters. Scafetta and West then used a sophisticated statistical procedure to establish “four scale-dependent empirical climate sensitive parameters to solar variations.” They developed these parameters separately for four periods including 1900–2000, 1900–1950, 1950–2000 and 1980–2000. They found that the control of global temperature by variations in solar output “could have, on average, contributed approximately 75% of the 1900–1950 global warming but only 25–35% of the 1980-2000 global warming. By considering a 20 –30% uncertainty of the sensitivity parameters, the sun could have roughly contributed 35–60% and 20–40% of the 1900–2000 and 1980–2000 global warming, respectively.”

However, they note that the observed solar-global temperature connection is “significantly stronger” than what is predicted by numerical models. They conclude that the models might be inadequate “(a) in their parameterizations of climate feedbacks and atmosphere-ocean coupling; (b) in their neglect of indirect response by the stratosphere and of possible additional climate effects linked to solar magnetic field, UV radiation, solar flares and cosmic ray intensity modulations; (c) there might be other possible natural amplification mechanisms deriving from internal modes of climate variability which are not included in the models.” Otherwise, the models are just fine?

Well, a recent issue of Geophysical Research Letters contains more bad news, or at least more challenges, to the modeling community. A team of British scientists noted that numerical models using historical sea surface temperatures (SSTs) consistently simulate a decrease in sea level pressure in the Indian Ocean region. This drop in pressure was associated with a substantial increase in precipitation in the area. Furthermore, because this area is known to control climate conditions in other parts of the world via teleconnections, the drop in pressure in the models could contribute to results in many other parts of the world.

Copsey et al. noticed a little problem – the actual observed sea level pressure data have been going steadily upward! As seen in the Figure 1 below, the empirical data show a clear upward trend from 1950 to 1996 while the model runs, whether forced by historical SSTs alone or SSTs plus greenhouse gases and aerosols, show a downward trend in this important climatological region over the same time period. The authors conclude “The contrast between the model results and observations suggests a model error.” It is somewhat refreshing to us at World Climate Report that they didn’t conclude “The contrast between the model results and observations suggests an error in the observations”!

Nonetheless, we see articles in the literature every week showing weaknesses and shortcomings in the numerical climate models. Many of these problems are fully discussed in the IPCC reports and elsewhere, but the press coverage given to the two studies reviewed here will certainly remain at 0.0!

Figure 1. Average trends in sea level pressure in the Indian Ocean for the period 1950 to 1996 using numerical model output forced by changes in historical SSTs only (top grouping) and SSTs plus greenhouse gas and aerosol buildup (middle grouping). The observed sea level pressure trends (lower pair) from HadSLP1 and Kaplan datasets used in the Copsey et al. (2006) study.


Copsey, D., R. Sutton, and J. R. Knight, 2006. Recent trends in sea level pressure in the Indian Ocean region, Geophysical Research Letters, 33, L19712, doi:10.1029/2006GL027175.

Scafetta, N., and B. J. West, 2006. Phenomenological solar contribution to the 1900–2000 global surface warming, Geophysical Research Letters, 33, L05708, doi:10.1029/2005GL025539.

No Comments

No comments yet.

RSS feed for comments on this post.

Sorry, the comment form is closed at this time.

Powered by WordPress