Dr Tony Butt world expert in Oceanography, long time contributer to Surfers Path Magazine on scientific issues related to surfing and general big wave hell man, has been writing a series of articles on the History of Surf Forecasting. He interviewed me about what we've developed at Surfcore and our unique approach to this problem. The following is the extract from the article which is in the latest Surfers Path (issue 60 p30-36).

Nearshore Surf-Forecasting and an 'Expert System' - by Dr Tony Butt

Now, while scientists are busy 'working on the above improvements to the large , ' off-the-shelf' models like the WWW3, surfing websites are continuing to look at how to improve predictions specifically for surfers - how to short-circuit that gap between the output of the models and the size and quality- of the waves we surf. While that gap is getting progressively narrower, it still exists and is still being filled by our own - 'mental algorithms' based on intuition, experience and local knowledge. Even if the output of the models gave a perfect prediction of the offshore conditions, we would still not be able to translate that into a perfect prediction of the waves breaking on our beach.

Surfers Path issue 60 - out now

Instead of waiting for the research teams working the WW3 and SWAN models to come up with something more useful to us, how about transferring our our collective intuition, experience and local knowledge into some sort of computer model? A kind of database containing the unique characteristics of a number of surf spots, that can be looked up to find the likely surf at each spot according the predicted combination of offshore conditions. As a simple example, the WW3 might tell us that, 200km off the coast, the swell will be coming from the northwest with a significant height of 6ft and a peak period of 15 seconds. Surfers at spot 'A' know that under those conditions, their beach break will contain 3ft close-outs since the waves won't be of the right direction to refract around an offshore shoal, and produce the A-frame peaks they get on a west swell. Surfers at Spot 'B', on the other hand know that, with a 15-second period, the waves will be sucked into their reef to produce solid 10ft surf. If that kind on information could be compiled into a database for a large number of surf spots, then we'd be well on out way to an automated version of that mental algorithm I keep talking about. It would be a kind of surf-forecasting 'expert system' with its own 'knowledge base and 'inference engine'.

Michael Fish - How we used to do it...(from the Culture gallery)

Well, that's exactly what Sam Blackmore and Jim Higson from surfcore.co.uk have been working on. They have developed a database containing information on 60 beaches around the highly convoluted coastlines of England, Scotland, Wales and Ireland. Factors taken into account include type of break, geographical position, swell-angle tolerance, local and underlining predominant sea states, and more. For each site, these specific local variables are fed into the algorithm, together with the forecasted local wind conditions and the outputs of a deep-water wave model, to arrive at an estimate for the conditions on that beach. Blackmore and Higson have thoroughly tested their algorithms using breaking wave height as a benchmark parameter, and claim usefully accurate results for over 90% of the time. Once the local conditions have been determined, the result is fed back into another database containing the 'ideal' conditions for each specific site. The closeness of the predicted conditions to the ideal conditions is then used to derive a star-rating system. To minimize subjectivity and to avoid generic definitions of 'good' and 'bad' surf the data for this database was gathered from the local surfers themselves.

There are no plans to expand the service beyond the British Isles, which means they can afford the luxury of using a deep water wave model specifically designed for UK waters. This model has much tighter grid spacing than the global WW3, and is driven by wind information derived from a UK specific atmospheric model.

Tony Butt - Taking the drop at his home break(from the Offshore gallery)

Meanwhile, Ben Freeston and Nick Lott from magicseaweed.com have taken a more direct approach. They've been working on narrowing that black box right down by bringing its outer edge closer to shore. They have successfully 'bolted' the SWAN shallow-water model onto the WW3, and are now running an experimental version of this combo model for UK waters. The SWAN model is initialized using the full directional-spectra data from the WW3, available at certain individual points around the coast. SWAN then continues the wave generation and propagation process in shallow water, keeping all the spectral information virtually right up to just beyond where the waves are breaking. The grid resolution of the shallow-water model, since it only covers a relatively small area, can afford to be much tighter than it deep water counterpart. At the moment, Freeston and Lott are running SWAN at a resolution of approximation two miles, and working at tightening that down even further

To run a model of this type, detailed bathymetric data must be input for the entire shallow-water 'domain' between the WW3 output points and the coast itself. Data of this type is not available for every coastline around the world, simply because not all the bathymetery has been measured accurately enough. However a surprisingly large amount is readily accessible, including most surfing areas. Once the UK-waters version has been fully proven, plans are to put the system in place for other surfing ares in Europe and, hopefully, the rest of the world.

To improve their forecasts even further, the magicseaweed team are also designing their own 'expert system'. Users of the website will be encourages to provide on-the-beach observations which will be collected and built up into a database. Once enough observations have been collected to be statistically 'robust', these will be used alongside statistics of the model output itself to feed information back into the model in order to fine-tune it.

Apart from the two teams mentioned above, several other people around the world are simultaneously working on tackling the problem in their own ways...

www.surferspath.com

If you are interested in learning more about Surf Forecasting Tony Butt's book Surf Science available in our shop will help you sort the chop from the waves.