Producing computer models that accurately reflect the complexity of the Earth's climate is an immense challenge. Even when focusing on the priority areas of natural variability and the impact of human activities, the task is daunting. Extending it to cover all the potential causes of climate change may prove overwhelming. So the approach must be to develop systems that inspire confidence in our ability to address the priority areas, and then go on to explore the sensitivity of the climate to other factors. The only way to do this is to create detailed computer models of the global climate, to evaluate the various hypotheses. The first stage is to establish whether these models are capable of producing a realistic representation of the climate and can respond to the most obvious quantifiable perturbations. The next stage is to check the relative importance of other possible causes of climate change, both to ensure we have correctly identified the priority areas and also to gain a better understanding of past events and decide how much confidence to attach to predictions of future changes.

Given these broad aims, the objective of this chapter is to describe the essential features of computer models of the climate and then to assess their performance in terms of reproducing known features of the climate. After this, the ability of the models to handle the natural variability of the climate and the possible impact of human activities will be considered.