Climate Science

Believable Fingerprints

See allHide authors and affiliations

Science  11 Sep 2009:
Vol. 325, Issue 5946, pp. 1321
DOI: 10.1126/science.325_1321b

Detection and attribution studies of climate change compare climate model output to observed variations of some climate parameter, such as temperature or rainfall, to see if the “fingerprint” of the changes represented by the model matches that of the measured parameter, and thus whether the climate changes can be attributed to a specific cause. Agreement across multiple models suggests that the results are more believable and less prone to distortion by the shortcomings of any particular model. Santer et al. used 22 climate models to investigate changes in the water vapor content over the oceans of the world. Water vapor significantly influences climate in many ways, acting as the largest single source of radiative forcing and as a key component in the process of hurricane formation. The authors found that recent changes in water vapor are robustly linked to anthropogenic forcing, and that the success of a model's ability to reproduce water vapor changes is not linked to its success at representing other climate variables. That means the “one model–one vote” method of predicting climate change may not always be appropriate, even though it turned out that way in this study.

Proc. Natl. Acad. Sci. U.S.A. 106, 10.1073/pnas.0901736106 (2009).

Navigate This Article