The role play on modeling was great fun to follow. It appears that by now, participants enjoy themselves in immersing themselves into roles they personally do not support. Yes, this was exactly intended – to explore unusual perspectives and viewpoints and by looking at the topic from a different perspective carve out new aspects.
Being a model developer myself, I know and appreciate the power of modeling as an important tool for knowledge gain, forecasting, (virtual) experimentation and even communication. But, sure enough, all modeling results need to be taken with care and need to be interpreted in the right context. A model is never accurate! That is why it is called a model. Reality is always more complicated.
However, with a reasonably well designed and validated (i.e. making sure a model represents reality and is fit for its purpose) model uncertainty can be quantified! Running an ensemble of simulations allows to study the sensitivity of a (modeled) system on initial conditions or assumptions. With this, one can assess the effect of certain measures (e.g. carbon emission cuts, geo-engineering measures, etc.) and their interaction with other processes that may render them useless. Since we have no climate laboratory (like a wind channel or a chemistry lab) we need to rely on modeling in this case.
Finally, we need to make sure that we do not fall into the trap of confusing projections with predictions. In climate modeling we are not able to predict the climate of the year of 2100. But we are able to run projections that show (aspects of) the effects of certain decisions in future situations. This is a useful and powerful application of modeling even though uncertainty is involved and cannot be diminished.
Hello there!
While I do agree that modelling can provide us with many insights that go beyond merely going with our gut feeling, especially in policymaking I belive that there is a strong risk to fall for the illusory perception of knowledge, which is what models cannot provide. This applies less in the natural science sphere, but even more so when human-made socio-economic parameters enter the model. For a brief critique on IAMs, for instance see: http://web.mit.edu/rpindyck/www/Papers/MisuseClimateModelsREEP2017.pdf
Best regards,
Adrode
No model is safe from misuse of course. However, when applied thoughtfully models provide great value in cost-efficient testing and verification. And “certainly” this applies to physical models in the same way as to other models.