Nautical Natters - 14th Sept.

I'd just like to thank Dave Lee for last night's 'Nautical Natters' on The Tides and Winds of the River Exe. Good to see so many new faces there as well as a few 'lee-bow deniers'! Even if you've been sailing since Noah bought the first RS Ark (whatever did happen to that early SMOD concept?) it's good to attend events like this to remind you of what you've forgotten or neglected over the years.

I'm sure that everybody came away from the evening with something new or refreshed to take on the water next time.

Thanks, Dave.

Hi Alistair, I'm glad you found it useful and thanks for your support and feedback. Also, thanks to everyone who contributed when I put them on the spot during the talk, particularly Mark Elkington for fielding the Lee bow question

Yes, thanks Dave for an engaging talk.
Talking about forecasts, I personally prefer the UKMet for our local forecasts over the NWS GFS derived forecasts (Wind Guru etc.). However, on the day, as Dave said, one has to update any of these by applying local knowledge and "Nowcasting". I find it particularly useful to look at the latest synoptic charts (I use meteocentre.com/analysis/map-surface.php?lang=fr&map=UK&area=uk ) This site also has useful links for radar and satellite and extended various model data inc. ensembles.
There is another good sailing weather site at Predictwind.com which has the GFS and ECMF at 8km res. (!km if you pay!) - for up to a week ahead.
Referring to forecast models generally, I saw an interesting comparison study of the US model versus others for Hurricane Irma. The ECMF seems to do consistently well for Atlantic TC prediction - much better than the NWS model - also better than the UK model although I believe this is usually quite good on TCs (Mark?)
If anyone wants to view the article, the link is at: mashable.com/2017/09/14/hurricane-irma-w...els-gfs-vs-european/

Hi Roger
You are right that the GFS model which is the basis of many web weather sites is probably one of the least accurate of the main weather models. I'm clearly a little biased here but most independent assessments of model performance puts ECMWF, Met Office and the Japanese Met Agency as the top performers. It varies a bit over time and what specific performance measures you are looking at. Most of the comparison studies I have seen focus on temperature, precipitation and extreme events - so its hard to say which is best at accurately predicting a 10 knot south westerly rather than a 12 knot west south westerly.

A few years ago I kept notes of the Met Office friday morning predicted wind for the following Sunday and the actual wind during the race. In general it was pretty good indicator, but more than three times more likely to be 5 knots less than forecast than 5 knots more than forecast.

Mark