Introduction
The first scientist to address Congress about global warming was Dr. James Hansen in 1988, based on a paper he and colleague S. Lebedeff had published in 1987. In his testimony, he stated that there was a 99% probability that global warming was happening even then, when using the 1950-80 global mean temperature as a baseline. He also discussed the probability of more extreme heat and drought events, and ended his testimony requesting help in improving the climate models of the time.
Climate Model Improvements and Results Over Time
The basis for Dr. Hansen’s data in his Congressional testimony was a simple climate model developed by one of the winners of the 2021 Nobel Prize in Physics, Suki Manabe. In 1968, he and R.J. Stouffer tied the atmosphere and earth surface through balancing radiation going out to space with heating of the lower atmosphere through turbulence. In the model, carbon dioxide (CO2) was doubled in the radiation scheme and allowing water vapor (another greenhouse gas) to increase in in response to the warming. It was constrained by keeping relative humidity constant. This simple climate model reached a new equilibrium temperature 2oC higher than the baseline.
What’s missing from such a simple climate model is considerable. For example, missing feedbacks to the climate system include changes to ice and snow cover, changes to the ocean circulation and sea surface temperatures, slow shifts to climate zones and vegetation cover, more immediate soil moisture changes, and others. The challenge is in having sufficient computer capacity to make these additions.
During the 1970s, the atmospheric portion of the climate models were allowed to interact with simple ocean models, but there were problems with the resulting model climate, including the shutdown of the overturning oceanic circulation (that drives the Gulf Stream and the Atlantic current that warms Europe), and runaway warming or cooling. Adjustments were made in how the atmosphere and ocean interacted in such a way to constrain the model climate. By 1993, Manabe and colleagues showed a potential 3.5oC to 7oC global mean temperature increase before the earth system reached equilibrium with a doubling or quadrupling of CO2, respectively, using such models.
The adjustments to constrain the climate in the initial coupled atmosphere-ocean models became unnecessary by 2008, because of greater available computational time and the inclusion of more (and more realistic) physical processes methods. By 2008, none of the atmosphere/ocean climate models used required constraining adjustments on how the atmosphere and oceans interact. Additional refinements included crude representation of sea ice, clouds, and ice sheets, but more work needs to be done in these areas.
How Have Past Climate Models Done to Date?
In 2020, Hausfather and collegues published an article discussing previous climate model projections with greenhouse gas increases, and how they compare to observations to date. The resulting graphic is shown below for 1970 to 2020:
Before about 1995, natural variability leads to some excursions from the climate model projections available at those times, though the long term trend agrees. After 1995, the observed global mean temperature lies within the climate model projections.