We’re over halfway through this process, so it might just be time to try and explain what on Earth we mean when we use the word ‘modelling’. I don’t speak for everyone else in the group, but I can at least explain what I think it means.
A model car is made to look like a real car. Since it’s smaller, there will have to be compromises – maybe the doors don’t open, the wheels don’t spin. However, if you wanted to know what a car was, looking at a model car would be a good place to start.
A scientific model doesn’t necessarily look like anything – it might just be a bunch of numbers on a screen. But it aims to be a good reflection of how something behaves. A weather model, like those run by the Met office or the BOM, has rain, sun, clouds and wind, and these things all behave pretty much the same as they do in real life. Wind can blow clouds around, some clouds make rain. Weather models are so good at mimicking real weather that they can make predictions. And these predictions are good enough that people will read them, and (maybe) trust them.
Weather models are incredibly complex, and are run on supercomputers. I don’t see much of a story in such big, predictive models. But modelling isn’t just about number crunching and whether you should reschedule your barbeque. Modelling is a mindset. A good model uses the available information to give appropriate advice.
Currently we’re mostly looking at threshold modelling. Threshold modelling looks at big changes in systems, and the tipping points that cause them. Imagine a forest – if you cut down one tree in ten, the forest will stay as a forest. After a decade or so, new trees will fill the clearings, and the system will be pretty much the same as you found it. If you chop down nine trees in ten, the ecosystem will change dramatically. Grasses will shoot up; grazing animals will move in and keep young trees from growing larger. The forest becomes shady grassland, and it will take more than just time to change it back to a forest.
Although these thresholds are not precise, they are definitely real, and observable. More importantly, at least from my point of view, they are much more understandable than the equations that predict our weather.
I guess my question for today is: How accurate is this description of modelling? What does it get wrong, and what does it leave out? And how similar is it to your understanding of modelling?
– David Shaw