For the last three years I’ve been struggling with weather forecasts in the Bristol Channel. The difficulty is finding a forecast representative of conditions out in the channel, when the popular forecast sites, such as XCweather or the Met office, need a land-based location in order to supply a forecast. I’ve tried playing about with different locations- Penarth is sheltered in a westerly wind so, say, I’ll get a prediction for Cold Knap in Barry- but it’s a very imprecise art. You feel from experience that there’s going to be more wind in the channel than the land-based forecasts say, but how much? and will it be too much?
Well, the way to address that is to look at a model rather than a point forecast. There are a number of ways to do this. Of course you can look at isobar maps, published by the Met office amongst others. But these are small scale/large area models covering the UK and its continental shelf, and they don’t show any of the local effects you can expect to see in the Bristol Channel.
The next step is to look at a computer-generated weather model. Here we are hampered in the UK becase the Met office doesn’t make its own models readily available. What is available is the American GLobal Forecast System (GFS) model, best presented by passageweather.com. This model, however, is still very small scale, and although it sometimes shows increased wind in the channel, it’s difficult to really trust it.
So recently I’ve tried a paid-for service supplying a high-resolution computer model forecast in the form of a Gridded Binary (GRIB) file, from a company called PredictWind. The quality of the forecast data is a revelation.
Take a look at the high resolution Grib output from PredictWind below. Stronger winds are shown by the hotter colours and larger arrows; the coloured lines are wind strenth contours. Black lines are pressure isobars.
This show very clearly the way in which the Bristol Channel funnels a westerly wind, and how light winds near land can suddenly be replaced by very strong breezes just a little way out on the water. Every Bristol Channel sailor knows this is what really happens here, but no other weather model I’ve seen is able to capture it so clearly. Looking up forecasts for a land-based location won’t really impart this sort of information.
When you can properly see the funneling effect, it’s going to make you think again about the timing of your passages, and increase your confidence in passages planned when the funneling is absent. PredictWind is expensive (very expensive- its monthly cost is not far off a full subscription to Sky). But most sailing mistakes are weather mistakes, and I think there’s a huge safety improvement to be made in departing port with a weather model of this standard under your belt. The Grib file includes three hourly updates for days ahead, and if you download one immediately before setting off, you should be sailing the Bristol Channel with a higher degree of confidence that the weather you expect for a day’s sail is the weather you are going to get.
That’s the plan anyway. We’ll see how it pans out over the course of the season!