Predicting the track of open ocean swells in deep water is almost easy. Waves travel in very predictable patterns, height slowly decreases as the swell propagates. Waves of different periods separate and arrive at different times. Given an understanding of the strength of the original storm you could calculate roughly the size and arrival time of the swell with some fairly simple maths and without needing a computer. What happens as these waves enter shallow water is where a large part of the challenge to us as surf forecasters lies.
Shallow water is defined in relation to the period of the wave. Longer period swells have longer wavelengths and a wave will start to 'feel' the seabed at depths of around half it's wavelength. So, for example:
10 Seconds 78m
17 seconds 225m
Now these values are considerably deeper than the seabed surrounding most surf breaks, so even that low period swell will start to feel friction with the seabed quite some distance offshore. The simplest effect of this friction is to slow the wave and for it to lose power, more complex effects arise because, of course, the sea bed is far from level and even.
Bathymetry is simply the shape of the sea floor. Just as we have valleys, mountains and flat plains on land the sea bed is similarly defined by troughs, trenchs and undersea mountain ranges. Just like on land some of these features are small, local and potentially changeable: a patch of reef, a changing channel in the sand or shallow water around a nearby pier or headland. Some exist on a massive and (for our timescales) permanent basis: Deep mid Atlantic trenches, a deep water trench funnelling surf to the SW French coast, the deep ocean surrounding the Hawaiian islands, an undersea bank occasionally creating huge waves off the coast of California.
This is a key concept for us as surf forecasters. We've mentioned that as waves enter shallower water they will start to be affected by their interaction with the sea bed. Now imagine a wave heading towards a headland. It's likely that the water around this headland is shallow and that the part of the wave entering it will slow down, but the same wave, further away from the coast is in deeper water, it hasn't slowed. This causes the wave to change it's direction and start to bend towards the headland, in the same way that applying the brakes on only one wheel of a car will slow that side of the car and cause it to turn towards that wheel. It's a simple and elegant concept that modifies the way waves behave on both a large scale in very deep water (which our swell model will attempt to calculate) and on a very very local level as a wave hits a shallow sand bank or reef.
Now imagine a big sand bank of slab of reef sitting in shallow water. As the unbroken wave approaches this it slows quickly, but either side in deeper water it continues faster. This creates the same bend in the wave we described above, on both sides of the bank. Most surfers will have experienced this in the water, you see the line of a new set wave approach the beach and as it starts to reach the bank it'll start to 'bowl', you'll also have noticed as it does this the wave will start to increase in size over the bank, all that energy bending in towards the shallow water is 'focused' and the wave will be bigger than other waves on the beach.
Likewise as a swell bends round a headland or into a bay with deep water in the middle the edges of the swell slow and the wave spreads out, in the same way that propagating waves lose size because the same amount of energy is covering a wider area these waves will be smaller than they would have otherwise been. This is often the case at point breaks, they may offer a long ride and perfect wall but possibly break smaller than other beaches in the area.
Now bear in mind that longer period waves are travelling faster to start with and react with the sea floor at greater depths than shorter period waves and you'll appreciate that longer period waves are affected more by refraction than shorter period waves. This can mean that they focus more and make bigger breaking waves. It can also mean they bend into areas that might otherwise miss a swell. It's for this reason that a decent long range groundswell, with it's longer period waves arriving first might initially show best on breaks that really rely on this refraction, that tucked away reef for example – where the latter part of the swell might peak on more exposed beaches with spots that worked earlier on showing less size as the swell progresses.
Of course refraction will only allow waves to bend so far, put a harbour wall jutting out in front of incoming swell and you'll most likely create flat water inside. In exactly the same way incoming swell can be blocked by small local features like piers and headlands and by larger scale obstructions like offshore islands. Often the direction of the swell can be critical in calculating how this will happen. Again the swell model deals with much of the larger scale interpretation, but checking out Google Earth and starting to look at how the path of an incoming swell might be obstructed is an important part of working out how to forecast for your local spot.