A model is just a way to define how something works and test out hypothetical situations. Something as simple as saying speed x time = distance gives us a very very simple model for describing how long it takes you to drive from home to the beach. Straight away we can use this model for a bit of forward planning. We know you live 60 miles from the beach, we know your van can only manage 30 miles an hour, so we know that it’ll take you 2 hours to get there. We’ve created a model and we’ve used it to predict what will happen. You might not have ever driven this route before, but we can have a good idea of how long it’ll take.

A ‘swell model’ is exactly the same in principle, if we know there's a 10ft swell 500 miles away heading in this direction we can work out when it'll arrive (based on how fast it's travelling) and how big it'll be when it gets here.

Of course doing this for the whole globe, for multiple storms and swells, gets a lot more complicated in action. In fact before we even look at the swell model we need to start with the atmospheric model that powers it:

Every day, several times a day, Meteorological agencies around the globe are sharing weather data. Trained observers, automated weather stations and reports from air and sea traffic and data from satellites all combine together to allow us to build up a picture of what the weather is doing right now. This set of conditions is the starting point for a model that attempts to predict what the atmosphere will do over a period of about two weeks into the future. It’s built on lots of tiny bits of maths and physics just like our example above. If we know how hot the ground is, and we know what it’s made of we can probably predict how much water will evaporate from it, if we know what the air pressure is we can work out what the wind will do, so we can work out where this condensed water (cloud) might travel to. If when it gets there the air temperature is colder we can work out how much it will condense and if it hits a certain limit we know it might rain. Now take that picture we’ve built up for the weather from an hour from now back into our model and we’ll have an idea of two hours from now and so on…

Now this has already moved way beyond the scope of what we can calculate ourselves. In fact our understanding of the physical processes evolved before we were able to use it to make such accurate predictions, early attempts by teams of meteorologists to manually perform the calculations necessary failed and it was only with the advent of powerful computers that these models started to become useful. It’s not much practical use creating a forecast for tomorrow if it’s going to take a week to calculate!

So we have super computers powering a complex mathematical model that tells us, precisely, how it expects the atmosphere to change over the next 10-15 days. Crucially for us as surfers it tells us how the winds will blow over the surface of the sea. The next stage is, mathematically at least, somewhat simpler. The ‘swell model’ takes these winds, maps of the shape of the seabed and coastline and runs some more maths to work out what will happen if the winds blow as they’re forecast to.

This swell model sits at the beating heart of almost every surf forecasting site or service. With it we can make specific predictions about the state of the ocean more than a week in the future, but by understanding how it works we can also gain a better understanding of what it is, and isn't, telling us.