I read with interest David Kramer’s piece on geoengineering (Physics Today, February 2013, page 17). I must say, I am more alarmed by what the geoengineers in his report are proposing than by the climate changes that are taking place. I believe geoengineers are removed from scientific reality. They ignore the fact that the climate system and its components—clouds, hurricanes, and so forth—are highly nonlinear and thus very sensitive to the initial conditions and to changes in the parameters. Nevertheless, one could study the system’s response in a probabilistic way when certain parameters are changed or when we introduce fluctuations, if the relationships among all the components are known exactly.
And here lies the whole problem with geoengineering. The formulation of the climate system and its components is only approximately known. More than 30 climate models are floating around in the climate community, and their predictions about general dynamics simply don’t agree with each other. In a recent publication,1 we considered 98 control and forced climate simulations from 23 climate models and examined their similarity in four different fields (upper-level flow, sea-level pressure, surface air temperature, and precipitation). We found that except for the upper-level flow, the agreement between the models is not good. Moreover, none of the models compares well with actual observations.
One person in the Physics Today story said that geoengineering may result in changes in various weather patterns, but nobody knows what the changes are going to be and how they will affect the climate system. If the warming in the Arctic is a big event to mitigate, then it will require a significant “geoengineering” effort. To me, that means significant changes will occur elsewhere. Who can say whether those changes will be less serious than those taking place now? How can geoengineers talk about modifying clouds and albedo when clouds are represented in the climate models as mostly linear parameterizations?
Kramer’s report did not mention hurricanes, but geoengineers also propose to dissipate them. Hurricanes are unique in the climate system because they represent major self-organization. As physicists well know, self-organization occurs in dissipative systems in which energy is not conserved but instead is exchanged with the environment. Hurricanes involve huge amounts of energy. Scientists have little idea how the atmosphere and the ocean will be affected if that energy is not allowed to be exchanged.
I would not have a problem with geoengineering if the physics and dynamics of the climate system were well known. Climate scientists have a good idea of the large-scale flow of ocean currents, but detailed measurements are not available. They know the basic physics of cloud formation and its thermodynamics but do not fully understand detailed cloud microphysics or the complex connections between climate and ecosystems. And with complex nonlinear systems, details are important. So we need to make an effort to improve our understanding of our climate system and its components before we try to operate on it. We can engineer a car or a plane because we know the underlying physics of motion, combustion, and flight, and we understand the role of every component. Can geoengineers say the same about climate?
