5.2 Why do we use models and scenarios?

Models and scenarios help us to make decisions. They are a way to simplify complex systems to help make decisions. Models are only as good as the

• quality of the input assumptions

• validity of the mathematical assumptions used to describe the system

• way that the model treats uncertainty

• way that users interpret the findings or results.

Answering the question about what more global warming will mean later this century requires some insight to future emissions. There are huge uncertainties here; for example, to do with future rates of economic growth and population in different parts of the world, the development of new technology and so on. One thing we know for sure (unless we are very lucky!) is that when we try to say things about what the world will look like in the future, we'll get it wrong.

Scenarios help us to explore a range of possible futures, which allows us to explore uncertainties more fully, based on the way that different and uncertain variables might change and interact. But we should always remember that scenario builders and modellers are all human. So it's inevitable that subjectivity, values, beliefs and feelings are embedded in all models and scenarios.

To an extent these tendencies can be evaluated through rigorous testing of models, for example by running hundreds of different models for the same set of input conditions and then by comparing the results – and to do this several times for several sets of different input conditions.

Nevertheless, this doesn't mean that that for example the central results are any more or less reliable than the extremes – they might for example reflect widely held prevailing social norms for the type of people who are inclined to develop and test models!

Models are the best tool we have for understanding the future direction of climate change. If anything they are more conservative than alarmist. Their veracity is tested by ‘predicting’ known climate or the impact of particular events (such as the eruption of Mt Pinatubo), which the best models do very well.

Or, as the IPCC put it in their Third Assessment Report (2001):

Additional information can be found in these web-based resources

• ‘How reliable are the models used to make projections of future climate change?’ from the IPCC.

• Briefing from NASA on the role of models in predicting future climate.

• RealClimate on climate models.

• DIY climate change modelling tools and resources for education and research at Educational Global Climate Modeling from Columbia University.

• Your chance to help the climate modellers! – by allowing them to exploit the computing power of millions of PCs by remotely running a program on your computer (no catch, no hassle for you). See Climate Prediction for details.

• Article from New Scientist ‘Why economic theory is out of whack’ discusses the role of assumptions made in models and the recent financial crisis, and a tendency to conservatism.