Modelling is a fundamental component of the scientific process. It can refer to mathematical descriptions of natural phenomena, laboratory-scale physical representations of natural systems, computational realisations of interacting processes, or, more commonly, some combination of all of these.

To a structural geologist, modelling offers the ability to take complicated, slowly evolving, unmanagably large (or small !), inaccessibly hot or high-pressure geological systems, simplify them and transform them to a human scale. This quickly develops into an untuitive understanding of the basic processes themselves which can be taken back into the field.

Numerical models themselves are complex computational machines with multiple, interlinked components. To take advantage of highly parallel computing environments, these models take advantage of specialised libraries and computational toolkits. How is it possible to learn to use such complicated tools ? How do we know whether they are working properly ? How can it ever be possible to be innovative ?

I will give a brief overview of some of the basic principles of good practice in numerical modelling and show how to start developing a sound sense of what modelling does and does not offer to structural geology and tectonics. I will do this using examples drawn from users of the GALE/Underworld family of codes, and demonstrating how to progress from simple conceptual models to more elaborate applications.

Along the way I will try to impart an understanding of why different methods work for different problems, how we always try to match the basic building blocks of the methods to the basic building blocks of the physical processes we are studying, and, as a result, how to tell when the code fits the problem.