CHRIS TIBBS ON LOCAL WIND EFFECTS
All the time we are getting access to higher and higher resolution GRIB files. So GFS now offers ¼ degree resolution [one weather point every 15 miles] for the global model. A decade ago that sort of accuracy would have cost a lot of money and was not really attainable for the ordinary sailor, but now it is free, or often on a very affordable subscription for the highest resolution.
That said, we are always going to have local differences that are simply not picked up by GRIB files at the moment. That often depends on what the local effects are, some are dealt with quite well, others less so.
FRONTS
If you have a fairly active front, it is pretty safe to assume the front will give you a more abrupt change in the wind direction than that being shown on a GRIB file. Most sailors will have experienced the very sudden change in windspeed and direction you get as that front arrives. The GRIB files, however, will usually smooth that out over a number of hours. The prediction it gives for the front coming over and the subsequent change will be accurate, but it will not show how sudden that change will be.
SEA BREEZE
Sea breeze is an area where you have to make a judgment about the complexity of the situation. The sea breeze in the Solent is different to a few miles down the coast in Brighton, for example. The Solent sea breeze is famously complicated, coming separately from either side of the Isle of Wight, and that complexity is unlikely to be picked up, so you do need to use your own skill and judgement to determine what you think is going to happen.
Much of the problem comes down to what the GRIB models see as the land or sea mass. The Solent models struggle with the relatively small piece of water between two pieces of land. The models tend to smooth out the differences across that area, so you can see the sort of scale where land mass can have an effect without the models picking it up.
STORM CELLS
Something else that GRIBS really do not pick up are things like thunderstorms, or smaller clouds, as they are too small for the scale within the model. It will pickup a cold front, which you might assume would bring the possibility of showers on that front, but it won’t pick up wind effect of things to do with clouds and showers.
RAINFALL
If you’re using GRIBS to get your forecast, and you see a very active front due to arrive, it is a good idea to then turn to an inshore waters forecast or the Met Office (or similar) to see what the rainfall prediction is. That rainfall prediction can be quite accurate in short-term forecasting and will provide a better idea of the time that front is likely to hit and so the time you are looking for the wind to change abruptly, as it will arrive with the rain cloud.
HEADLANDS
Depending on how big a headland is, GRIBS will struggle here too. It is easy to check this ahead of time. If you look at the GRIBS files historically, you will be able to see whether there is any bend being predicted round a headland. All headlands will have some acceleration effect and some wind bend effect, so if the GRIB file never picks this up, irrespective of true wind direction then you know it is something to look out for.
TIDAL ACCELERATION
Finally, clearly tide will not be factored into any of this and can cause a significant increase in apparent wind. Again, round a headland where the tide is accelerating, the combined effect can be up to 15 knots wind speed increase. If the headland is too small to be picked up by the GRIB files and you are going to have the tide underneath you, that is going to account for a significant discrepancy.