What was done
The author reviews the current status of the climate modeling enterprise, especially as related to coupled general circulation models (CGCMs), and gives his best estimate of where the field is headed in the future.

What was learned
The author reports that even though the climate models of today are wonderful scientific achievements, there is still a lot that needs to be accomplished before they can do all we would like them to do, like tell us the consequences of continued increases in the air's CO2 concentration.  Some of the areas the author lists as being far from adequately treated in the models are ocean-atmosphere and soil-vegetation-atmosphere interactions.  Within these categories are questions about water vapor feedback, changes in cloud optical and precipitation properties caused by changes in the spectrum of cloud condensation nuclei, and potential changes in North Atlantic Deep Water formation.  He also lists a number of vexing questions related to weather extremes.  "Will northern mid-latitude storms intensify, and how will their main tracks shift?"  "Will tornadoes and thunderstorms become less frequent but more violent or vice versa?"  "Will tropical cyclones be less frequent but more intense if the ocean surface warms in the tropics?"  And he notes that we will "have to wait for answers," as today's best CGCMs are still not sufficiently realistic to enlighten us on these important subjects.

What it means
Although the author notes - indeed advises - that "we must continuously evaluate and improve the CGCMs we use," he acknowledges that today's model results "are now used by many decision-makers, including governments."  Clearly, this state of (world) affairs is disturbing, as national and international policy is being made on the basis of vastly imperfect mathematical representations of a whole host of real-world physical, chemical and biological phenomena.  Although many may think that what we know today is sufficient for this purpose, the questions posed by the author, for which we still lack definitive answers, should demonstrate that this assumption is not true.