An Interesting and Important Discussion
Steve Fitzpatick is summarizing important aspects of radiative energy transport and its interaction with material in the Earth’s atmosphere, at Jeff Id’s tAV. I think many of these phenomena and processes are represented by parameterizations.
As I have noted from time to time, these interactions are frequently used to tune hindcasts so that a better representation of the past response of the Earth’s climate systems is obtained. And that I find this situation to be somewhat disconcerting due to the fuzzy nature of the parameterizations that are used to characterize these inherently complex interactions.
Let’s consider an example as follows. Take the steady-state isothermal flow of an incompressible fluid through a straight round pipe and for which a known flow rate is imposed. We want to check the predictions of a model of the pressure gradient along the pipe with measured data. For this flow as described, we know that the friction factor between the pipe wall and the fluid is the most critically important aspect of the response of the pressure gradient. The physical phenomena and processes of importance relative to the response of interest are known without a doubt.
If the predictions by the model do not agree with the measured data we know exactly which aspect of the modeling is in error and we can adjust that single parameter with assurance that we can obtain the correct results solely by the actions of this single parameter. We get the correct response for the correct reason. We know we have gotten the correct response for the correct reasons based on both theoretical and empirical information. Especially, empirical data have shown the direct connection between the pressure gradient and the wall friction factor for several hundred years.
In my opinion aerosols, and their critically important interactions with radiative energy transport, are among the more inherently complex aspects of real-world, radiative-energy transport in the Earth’s atmosphere. While theoretical studies have provided insights into these phenomena and processes, these are basically for highly idealized situations. The parameterization of the interactions are at all levels; from estimation of the geometric characterization of the aerosols, to the numbers of particles, to connections with several important aspects of clouds, and finally to the interactions with radiative energy transport. See Chapters 7 and 8 of Parameterization Schemes: Keys to Understanding Numerical Weather Prediction Models [Paperback] and Chapters 5 and 8 of Thermodynamics of Atmospheres and Oceans, Volume 65 (International Geophysics) [Hardcover].
When so many different aspects of complex processes are described by parameterizations, it can be difficult to ensure that tuning of parameterizations can be directly linked to the response of interest. This is especially true, in my opinion, when the response of interest is a very high level solution functional in which all aspects of the total problem are mapped to a single number. It is very disconcerting that parameterization of the geometric details of aerosols, for example, have been linked to representation of the response of the global-average surface temperature.
No comments yet.
Models Methods Software 
Leave a Reply