I'm a professor at U Michigan and lead a course on climate change problem solving. These articles often come from and contribute to the course.
By: Dr. Ricky Rood , 4:42 AM GMT on February 13, 2008
Models(4) Iconic Figure:
Of the figures that I consider the Iconic Figures of climate, there is one based totally on models. A recent version of this figure from the IPCC 2007 is given here.
Figure 1: Observations and simulation of the past century from the IPCC 2007 Technical Summary (Working Group 1) (largish PDF).
This is a figure of, approximately, the last century. In this figure there are three traces. One of traces, the black one, is of the observed, globally averaged surface temperature record. In the bottom figure is a blue curve, which is a model simulation that does not include anthropogenic (human-related) forcing. That is, it is “natural” forcing. In the top curve there is a red curve that is a model simulation that includes both natural and anthropogenic forcing. The point of this figure is that both natural and anthropogenic forcing is important, and that the recent warming requires the inclusion of anthropogenic forcing to simulate the recent observed temperature increase.
Forcing: For the purpose of this figure, “forcing” are those things that change the ability of the Earth to absorb or reflect radiative energy. Another “forcing” is the radiative energy that comes from the Sun. “Natural” forcing starts with the variability of the Sun. Of special importance in the realm of natural forcing is the impact of volcanic eruptions. Large volcanic eruptions put aerosols into the atmosphere. Aerosols above the Earth’s surface can reflect more solar radiation or they can absorb radiation in the atmosphere. These help cool the surface of the Earth. Aerosols also impact the infrared radiation; that is, the radiation emitted by the Earth back to space. Other natural forcings include water in the atmosphere, in all phases, and carbon dioxide. In general, these model experiments assume that the amount of carbon dioxide in the atmosphere prior to, about, 1850 is “natural.” Of course, the amount of solar radiation that is reflected by the surface is also included – ice and land.
In contrast to “natural” forcing is anthropogenic or human-related forcing. This is change in the forcing relative to the natural forcing. The most important of the anthropogenic forcings is due to carbon dioxide, which is calculated as the additional forcing due to the increased amount of carbon dioxide relative to the “pre-industrial” amount of carbon dioxide. Pre-industrial forcing is linked to about the year 1850. There are other greenhouse gases like methane, nitrous oxide, and the chlorofluorocarbons. Nitrous oxide increases are largely related to use of synthetic fertilizers. Other anthropogenic changes in the radiative balance of the Earth are related to changes in reflection at the surface due to how we use land.
The Plot: Here is my description of this plot. The dark red and the dark blue lines are averages from many model simulations. The light lines that surround the dark lines are all of the individual simulations. Prior to 1950 the natural and anthropogenic simulations are not much different from each other. After 1960 only the plot with anthropogenic forcing follows the temperature observations. Perhaps more importantly, the natural and anthropogenic curves diverge from each other as time goes along.
The light lines surrounding the dark lines give some idea of model variability. It is notable that, for the most part, this variability covers the range of variability in the observations. The models do not follow, point by point, the shorter scale variability in the observations, for example between 1920 and 1930. The models have variability, such as the El Nino – La Nina and North Atlantic Oscillation. The spread of the models suggests that the model variability covers this range of variability, but the models are not tracing this variability on an event-by-event basis. The comparable spread in the models and the observations also serve as a sanity check that the models represent variability in the same range as the Earth’s climate.
The simulations do show the impact of several large volcanic eruptions. The volcanoes do cause cooling of the globe. Volcanic eruptions, and especially the well observed Mount Pinatubo eruption in 1991, provide opportunities to evaluate processes in models.
It is also of interest to examine where the models and the observations do not agree. A most interesting period is from 1935-1940, a period when the planet was warm. (Thanks to crucilandia for pointing a reference to get me started.) A substantial literature is developing that examines this period. It seems to be associated with substantial Arctic warming. It is a period that demands more study. The cooling that all of the models calculate about 1915 is also interesting.
An important take away message from these simulations is that there are factors other than carbon dioxide that cause temperature variability. Hence, carbon dioxide and temperature are not necessarily correlated on shorter scales of variability. (This is a like my wave metaphor on this blog. )
Conclusions: This is a figure open to interpretation. Personally, I find this figure compelling. I know how difficult it has been to develop the models and to specify the forcing. There is also a huge depth of analysis at different levels of detail and averaging that support the conclusion that it is only with increasing carbon dioxide forcing that the recent temperature increase can be explained.
Others can look at this plot, and come to a different conclusion. One issue that many raise is what about the treatment of aerosols? This is a process in models which has substantial uncertainty in its quantification.
Looking forward to the comments.
Here are the previous blogs on models.
Uncertainty and Types of Models
Models (1) Assumptions
Models (2) Forgotten Layers
Models (3) Predictable Arguments
The views of the author are his/her own and do not necessarily represent the position of The Weather Company or its parent, IBM.