The Atmosphere and Climate Modeling - Introduction to Carbon Capture and Sequestration - page 94

permafrost in Arctic regions, yet we have very little understanding of how
much methane will be released from the permafrost if the temperature
increases. However, the potential effect on the climate can be enormous,
so adding such an effect to our models will make the models more real-
istic but also much more uncertain. To paraphrase the former US
Secretary of Defense, Donald Rumsfeld, we could say that one aspect of
climate research consists of turning “unknown unknowns” into “known
unknowns.”
From a public perception point of view this is a diffi cult message; our
models have improved enormously and we have signifi cantly more cer-
tainty that the outcomes of model predictions are accurate. Yet we con-
tinue to have uncertainties about the numerical values associated with
our predictions. Communicating these uncertainties requires special
attention [2.18]. The conclusion that “we are uncertain whether the tem-
perature in 2040 will increase by two degrees” is not helpful in public
discourse. More appropriately, the phrase “we are uncertain whether a
temperature increase of two degrees will be reached by 2030 or by
2050,” expresses confi dence in the outcome, and appropriate uncer-
tainty in the numerical estimates.