If a company developed a drug for the treatment of a disease but did not do clinical tests, it would not be prescribed by reputable physicians. Indeed, there are claims by pill companies that promote health benefits, yet the Federal Drug Administration requires adding
“This statement has not been evaluated by the FDA. This product is not intended to diagnose, treat, cure, or prevent any disease”?
There is a clear analog with multi-decadal climate model predictions where no skill has been shown in hindcast predictions of changes in multi-decadal regional climate statistics. As we have reported in our paper
Pielke Sr., R.A., and R.L. Wilby, 2012: Regional climate downscaling – what’s the point? Eos Forum, 93, No. 5, 52-53, doi:10.1029/2012EO050008
“It is ….. inappropriate to present [multi-decadal regional climate forecasts]…… to the impacts community as
reflecting more than a subset of possible future climate risks.”
Skill in multi-decadal regional climate model predictions of changes in climate statistics has not been shown (i.e. there are no “clinical trials” to show that the approach is robust) .
For future studies in the literature and media releases to present their results as anything more than a model sensitivity experiment (and that they should only be interpreted as, at best, a subset of what is plausible for the future climate), they would be guilty of climate science malpractice.
As just one text example of what this means, statements such as “temperatures will increase by 1C”, for example, should be written as “temperatures could increase by 1C”. The use of the term “will” indicates a certainty in the climate prediction which is not correct. The term “could” means the prediction is plausible.
Also, if they still insist on presenting their model results in figures with decadal time periods on them (e.g. 2040-2049, etc), they must make it clear that the results are intended to improve our understanding of climate processes and not an actual forecast for those decades that should be used by the impacts community to represent the envelope of what the regional climate will be decades from now.
Even for those studies that present their results as sensitivity studies, their paper should have an FDA-like disclaimer;
“The multi-decadal regional climate model results presented in this paper have not shown skill at predicting changes in multi-decadal climate statistics. The model results in our study should not be used to quantify the envelope of the risks from climate to societal and environmental resources in the coming decades. Our model sensitivity results are provided only to assist in improving our understanding of climate processes. “
Without this disclaimer in papers, assessments and other communications which report on multi-decadal regional and local simulations of changes in climate statistics, they are committing climate science malpractice.
This label, of course, can be avoided if the researchers provide quantitiative model and observational comparisons of multi-decadal regional and local predictions of changes in climate statistics, and show them to be skillful in terms of what metrics are needed by the impacts community. I invite anyone who has published such a study to present a guest post on this weblog alerting us to such a robust scientific study.