Sandy
Well-known member
They still play Sailing By.Haven't listened to the UK Shipping Forecast in years ... I'll bet they don't play Sailing By any more ......
They still play Sailing By.Haven't listened to the UK Shipping Forecast in years ... I'll bet they don't play Sailing By any more ......
Because it's not all free, and because GRIBs aren't a good way to show weather. The free ones are not great resolution so you'd end up paying anyway if you wanted to get good information. A grid is a terrible way to show a weather system and how it's going to move, the forecast tells you where the system is, what direction it's going and how fast. Using a GRIB you'd need to reverse engineer that information from your gridded info otherwise you're just looking at grid points and assuming what it says will happen is going to happen, and that's not usually the case.Why pay when it is all free, paid for by taxpayers?
Mostly the commercial GRIB products are derived from the GFS. They are just re-hashing the data. You get no more information. There is little difference between models for the first 4 or 5 days. Watch Windy.com to see a comparison between GFS, ECMWF and the German ICON-EU, their 6 km Europe model. Differences are usually within model uncertainty. If one is significantly different from the others then the situation is more unpredictable than usual. Claims of high precision and high accuracy are nonsense. Weather is never precise. Remember that the effective resolution of ANY NWP model is about 5 grid lengths.Because it's not all free, and because GRIBs aren't a good way to show weather. The free ones are not great resolution so you'd end up paying anyway if you wanted to get good information. A grid is a terrible way to show a weather system and how it's going to move, the forecast tells you where the system is, what direction it's going and how fast. Using a GRIB you'd need to reverse engineer that information from your gridded info otherwise you're just looking at grid points and assuming what it says will happen is going to happen, and that's not usually the case.
Pads are still available I understand. Otherwise download them off the internet. ?Shipping forecast nostalgia.
A million years ago I was taught how to write the forecast in shorthand and then draw a synoptic chart based on that with isobars, fronts, etc. Used to get pads of a pre printed format to draw on.
Was never a great success!
I wrote the RYA training manual on this some 50 years ago. It was a hairy exercise but life has moved on. Short term, small detail is still a problem. Outlooks for several days ahead we’re a pipe dream in those days. When cruising nowadays we can often be fairly sure about plans 5 or y days ahead.Shipping forecast nostalgia.
A million years ago I was taught how to write the forecast in shorthand and then draw a synoptic chart based on that with isobars, fronts, etc. Used to get pads of a pre printed format to draw on.
Was never a great success!
I think you’re confusing some things there. The higher resolution reflects the additional processing it doesn’t offer higher precision. The output is better because more data is used and processed. The model is the same but the output can and often is different. Lower resolution is cheaper to process and often good enough. In both cases gribs need to be interpreted back to weather though.Claims of high precision and high accuracy are nonsense. Weather is never precise
It all depends on what you mean by accuracy and precision. In the 1970s the Met Office large scale model used a 300 km grid. Its fine scale model used 100km. Their global model is now 10 km and the detailed model about 1.5 km. in the late 70s, we were just starting to see skill at 5 days ahead in model output. Now TV forecasters, sometimes, have confidence up to 10 days. Obviously, more data at higher resolution and models able to calculate the physical processes meaningfully at high resolution produce better, more accurate forecasts.I think you’re confusing some things there. The higher resolution reflects the additional processing it doesn’t offer higher precision. The output is better because more data is used and processed. The model is the same but the output can and often is different. Lower resolution is cheaper to process and often good enough. In both cases gribs need to be interpreted back to weather though.
That was sort of my point in the first place. Knowing where a weather system is and its direction is more useful than a bunch of points which you have to manually build back into knowing where the "weather" is likely to be and where it's going. It's of no use to me knowing that at a given grid position the model thinks the wind will be 13kt westerly at 4pm unless I place complete trust in the accuracy of the model, which I don't. If I know that something is blowing through just north of me then I can at least make an informed choice about risks over time.The point that I am making is that Weather is not precise
Surely, that is what GRIB viewers do for you. A good viewer, XyGrib is one, will give you isobars, rain areas, wind speed and direction etcThat was sort of my point in the first place. Knowing where a weather system is and its direction is more useful than a bunch of points which you have to manually build back into knowing where the "weather" is likely to be and where it's going. It's of no use to me knowing that at a given grid position the model thinks the wind will be 13kt westerly at 4pm unless I place complete trust in the accuracy of the model, which I don't. If I know that something is blowing through just north of me then I can at least make an informed choice about risks over time.
Thanks, Frank, for your clear explanations on these weather threads. It's great having someone on the forum who has not only been a career weather forecaster, but is a sailor too.It all depends on what you mean by accuracy and precision. In the 1970s the Met Office large scale model used a 300 km grid. Its fine scale model used 100km. Their global model is now 10 km and the detailed model about 1.5 km. in the late 70s, we were just starting to see skill at 5 days ahead in model output. Now TV forecasters, sometimes, have confidence up to 10 days. Obviously, more data at higher resolution and models able to calculate the physical processes meaningfully at high resolution produce better, more accurate forecasts.
However, when Predivtwind, as the best known example, quote accuracy and precision, they are referring to verifications at a point. That is what you and I will experience on our boats. The point that I am making is that Weather is not precise. Nor are models. The highest resolution global model grid is around 10 km. effective resolution is about 50 km. Models have to have a built in smoothing. Grid lengths of high resolution models such as the Spanish use of HARMONIE use a 2.5 km grid. Effective resolution 12 or 13 km. the French AROME grid is similar to the U.K. Try using the Meteociel site Ensemble link to see the spread of values of wind speed.
PS. I do not know if this will work. Look at Meteociel.fr - Tableaux PE-AROME. That is the AROME ensemble, winds are in km/hr. Look at the spread of wind speed. What do you mean by accuracy?
I depend on a rock on a string. See Weather rock - WikipediaThanks, Frank, for your clear explanations on these weather threads. It's great having someone on the forum who has not only been a career weather forecaster, but is a sailor too.
My bit of seaweed has become redundant now I live in a place where it hardly ever rains!
GROAN!I think you'll find that it's the rock that depends on the string...