By the way, I've also noticed a tragic inability of mathematical models to predict non-linear or exponencial change.  For instance, sea level rise exponencially increases the odds of inland flooding due to storm surge.  Another example is that because future natural methane emissions (from melting methane hydrate) can't currently be accurately estimated, it is simply not included in climate modelling.

Frankly, I feel sorry for any poor fool who depends upon current climate models to plan for the future.  Instead, an insurance model of risk should be adopted, where money is spent to plan for and adapt to the worst case senario.  On thing ought to be perfectly clear: non-linear events (black swan events) are low probability/high consequence, and will inevitably eventually occur.  Tomorrow is almost always like today, until it isn't.  Do you really need a model to tell you that?