parametrisations

Climate models are constrained with parametrisations: eg rising temperature leads to drought in the tropical rain forests

Unthreaded #8 « Climate Audit 2007

So the problem with climate models is digging out the underlying assumptions in the non-physical parametrisations. the most important one being clouds and precipitation as function of increasing CO2.

Unthreaded #8 « Climate Audit 2007

However if what you saw was blatantly wrong , you could always play with the parametrisations to get a better fit .

Exponential Growth in Physical Systems « Climate Audit 2007

Climate models are constrained with parametrisations: eg rising temperature leads to drought in the tropical rain forests, because it is observed that rising CO2 correlates with drier rainforests but unfortunately that is caused by deforestation .

Unthreaded #8 « Climate Audit 2007

There are no “feedback loops” programmed in, only basic physical parametrisations, often the same as those used for short term weather forecasts.

Unthreaded #14 « Climate Audit 2007

This will also partially answer your question on parametrisations.

Earth's climate crashes in 2013 « Climate Audit 2006

Combined all the likelihoods of the input parameters and the uncertainties of the parametrisations for climate makes it impossible to make a forecast with a 70% likelihood for the next 50 tyears:

The Coming Katrina Anniversary « Climate Audit 2006

He would have seen the distinction between his invertible and non-invertible rational parametrisations only in so far as the procedures that give the latter make special assumptions, and (arguably) in so far as non-invertible rational parametrisations miss some rational points (very easily found ones in some of Diophantus's examples).

Citizendium, the Citizens' Compendium - Recent changes [en] 2009

Diophantus generally seems to prefer procedures that lead to invertible rational parametrisations; when his procedure leads to a non-invertible rational parametrisation (a so-called unirrational parametrisation) it is generally the case that one can show nowadays that no invertible rational parametrisation is possible.

Citizendium, the Citizens' Compendium - Recent changes [en] 2009

However, Diophantus never inverts his parametrisations (or procedures, as he would more likely have seen them).

Citizendium, the Citizens' Compendium - Recent changes [en] 2009