How reliable are General Circulation Models as a basis for making sound policy decisions with respect to Climate Change issues? That question grounds many political positions
Over the course of the four decades General Circulation Models have been criticized and refined based on specific types of feedback. One of the more prominent scientific needs in any statistically based model is the need for reliable data because in many instances, data gathering techniques affect the reliability of the model. For example, if global temperatures over time around the world were recorded using different measuring techniques, that fact would decrease the reliability of any forecast.
Another variation on the data reliability theme deals with the issue of balance. The data may be gathered correctly, but could still suffer from a balance problem such as complete information for the Atlantic Ocean or atmosphere but incomplete information for Pacific Ocean or atmosphere. Uneven sampling, as it is also know, like most pollsters understand, tends to create biases in model prediction.
Scientists will tell you they consistently look to improve their data gathering techniques relying primarily on the World Meteorological Organization to set the standard for insuring the consistency of data measurement and compilation around the world.
Scientists also seek other ways to improve their research. There are more than a few ways to approach the issue of the reliability of long term climate predictions, here's one example.
A 2000 review entitled, An Overview of the Results of the Atmospheric Model Intercomparison Project, compare the results of thirty one different models from climate science programs around the world. While the bulk of the report provides analysis of the GCMs, they conclude,
"On the whole, the models provide a credible simulation for the large-scale distribution of the primary climate variables characterizing the atmospheric pressure, temperature, wind, hydrologic cycle and radiation balance, although a number of common systematic model errors are apparent".
Back in 2000, the robustness of the GCMs was still being challenged. Improved super-computer technology along with increased funding for research and ocean data collection helped climate scientists refined their Ocean Coupling Models. By 2004, The Program for Climate Model Diagnosis and Intercomparison was able to publish a comparative review of eleven of the models called, An Appraisal of Coupled Climate Model Simulations. It is a lengthy PDF file.
One of the big picture statistical problems the analysts were looking at was whether or not these latest coupled models were able to provide more reliable one hundred year forecasts. They noted that one of the problems with earlier versions was their lack of reliable long term forecasts. They say,
"Coupled ocean-atmosphere GCM simulations drifted relatively quickly and steadily unless constrained by nonphysical flux adjustments (and in some cases did so even with flux adjustments). In recent years the situation has improved dramatically. This improvement was documented in the most recent IPCC assessment report and is confirmed by the results given above. Although most of the CMIP2+ models employ flux adjustments, both the flux adjusted and the non-flux adjusted models exhibit acceptably small "climate drift" for century-scale simulations."
Posts
