The idea of forecasting the climate is amusing to many, because of the perceived difficulty in forecasting weather even beyond a few days. Indeed, there are theoretical reasons why the details of weather events cannot be forecast beyond about 10-14 days. This is because of the non-linear nature of the equations that govern the atmosphere. Un-measurably small differences in the initial state of the atmosphere will grow to differences that are sufficiently large after 10-14 days so as to render a forecast useless, even from a perfect model of the atmosphere. This sensitivity of the atmosphere's evolution to tiny perturbations is often referred to as the butterfly effect, recognizing that even a perturbation from a butterfly flapping its wings can grow over time into a substantial change in the evolution of the atmosphere.

However, the climate is made up of the average weather, that is, for example, the seasonal average temperature and precipitation at a particular location, the average number of storms during a season etc. Figure 1.1 schematically depicts the key processes that give rise to the climate we experience. It shows that there are a number of factors that can be considered as agents that can change climate - if we can anticipate these forcing factors, and know their consequences for climate patterns, then we have a tool to anticipate fluctuations in climate. Some of the forcing factors on the climate include:

• Changes in received solar radiation
• Changes in land-surface vegetation patterns
• Changes in sea-ice distributions
• Changes in sea-surface temperature patterns
• Changes in atmosphere

It is well known that received solar radiation varies with latitude and time of year, and that these variations drive the average climate variations that we know to be the annual march of the seasons - referred to as the annual cycle. Indeed, this is probably the first example of humans making and using a climate forecast - for example, knowing the usual time of year of the rainy season allowed the first patterns of agricultural practice to become established several millennia ago. For our studies, we are interested in going beyond the annual cycle - to the next step of climate prediction - to try and anticipate the departures from the normal annual cycle. For example, will the rainy season be wetter than normal, or will there be a drought? Will the winter be colder than normal or will unusually mild winds prevail? These departures from the annual cycle are referred to as climate anomalies.