Weather. Clearly. Explained.

B

BelleSerene

Well-Known Member
Joined
19 Sep 2005
Messages
3,423
Visit site
I'm trying to get a more formal handle on how weather works, and weather forecasting from the cockpit.

Much on the subject is written by people who don't really understand it. So it's hardly enlightening.

Some is written by people who understand it but aren't gifted in communication. Unravelling multiple forward references and sifting through the guff for the reasoning can be hard work.

Can the enlightened on the forum recommend something concise and meaningful? In written or DVD form for dummies perhaps?
 
Kemp and Young Notes on Meteorology.

Was written many decades ago, based on the syllabus for the old 2nd mate FGN.
There are others but I have seen none which explain the bare bones and principles in a more simple way.
I also found my PPL meteorology book very good. similar stuff different focus.
 
Uricanejack said:
Kemp and Young Notes on Meteorology.

Was written many decades ago, based on the syllabus for the old 2nd mate FGN.
There are others but I have seen none which explain the bare bones and principles in a more simple way.
I also found my PPL meteorology book very good. similar stuff different focus.
Click to expand...

Worth PM'ing Simon from the weather service who posts on these forums for his recommendations - believe he's also written a book...

http://www.ybw.com/forums/member.php?3286-simonjk

Dr. Simon Keeling
www.WeatherSchool.co.uk & SailingWeather.co.uk
 
I learnt as part of Geography at school. Why not take a look at the current school books, they are written pretty well these days.
 
There is a problem in that "explaining" the weather involves trying to reduce a massively complex system to something simple enough to express in a reasonably concise form. And the system is not just complex in terms of the number of variables, it is also mathematically complex in that a lot of the relationships between variables are non-linear. It is almost certainly impossible to create reliable forecasts more than a week ahead, except in VERY stable circumstances that are unlikely around the British Isles! Tiny changes in the input variables (i.e.the measurements of meteorological variables now) can create massive changes in the result a few days down the line. The butterfly flapping its wings in the Amazon that gives rise to a storm in the English channel is a BIT over the top as an illustration - but not an awful lot!
 
AntarcticPilot said:
There is a problem in that "explaining" the weather involves trying to reduce a massively complex system to something simple enough to express in a reasonably concise form. And the system is not just complex in terms of the number of variables, it is also mathematically complex in that a lot of the relationships between variables are non-linear. It is almost certainly impossible to create reliable forecasts more than a week ahead, except in VERY stable circumstances that are unlikely around the British Isles! Tiny changes in the input variables (i.e.the measurements of meteorological variables now) can create massive changes in the result a few days down the line. The butterfly flapping its wings in the Amazon that gives rise to a storm in the English channel is a BIT over the top as an illustration - but not an awful lot!
Click to expand...

Yes, same as someone who asked me to explain integral calculus "just the basics".

Instead of weather worrying, it's much better to get used to all sorts of hellish conditions. I once spent a day convincing the entire crew that the wind indicator was complete rubbish, and it couldn't be blowing 35 knots, no way. Another time it was due to be 40knots out of Gibraltar, everyone says hum better wait till tomorrow, but i went anyway, and it was fine, downwind, what the heck.

For the practical types, improving gelcoat repair techniques is a good option, and saves getting out the fenders even if it's a bit windy in the marina. Actually for the price of fenders, you can do a huge number of gelcoat repairs.

Easily the best way to find the weather "from the cockpit" at least in a marina is to strike up a conversation with a neighbour, the older the better. Tell them your plans in brief outline "we're planning to leave at 10am and head for Cherbourg" or whatever, and you have to get on with stowing things, nice talking. Within 20minutes they'll be back with full passage plan, tidal considerations, everything including the valuable info that it's likely better to go a bit later like perhaps 1pm for the tides, or the following day to doge the bumpy stuff.

Finally, get a Grib file to check the weather and plot a course going IN BETWEEN all the nasty wind arrows! Sorted.
 
Basic weather patterns are easy to understand from the idealised versions given in books and there are lots of good books on it; any of the RYA books will do. The trick is learn the idealised weather patterns and sequence of weather for depressions and anti cyclones; this is very defined. After that simply start to watch forecasts and compare to your actual conditions (wherever you are, sailing, driving to/from th office, cutting the grass). You will soon pick it up and develop a 'weather eye'. Once the basics are understood you can start to look into the physics behind it. The real skill is interpreting what the trends around you are predicting: change in wind speed / direction, change in pressure, change in clouds, temperature, rain etc. You will soon get the hang of it just by understanding the sequence of weather associated with the basic idealised models. This is the method used for centuries. Look at a synoptic chart every day as near to the same time as possible, I found this greatly helped when I was learning.

Here is the link to the RYA eBook http://www.rya.org.uk/shop/pages/product.aspx?pid=E-G1(RYADefaultCatalog) The paper version is also available.
 
Last edited by a moderator:
Sandy said:
I learnt as part of Geography at school. Why not take a look at the current school books, they are written pretty well these days.
Click to expand...

Meteorology is a branch of physics, not geography.

I can recommend Reeds Weather Handbook. Written by an active sailor who had a career in meteorology with ten years as a Met Office senior forecaster.
 
When we were in the Clyde in late June and most of July last year it was easy.

When you could see about half a mile-50% of the time, it was going to rain.

When you could not see half a mile-the other 50% of the time, it was already raining.

Simples.................................
 
BelleSerene said:
I'm trying to get a more formal handle on how weather works
Click to expand...
The weather does not work, weather just happens and the mechanics defy human comprehension.

Weather forecasting is a pseudo science but hope is on the horizon in the form of computer programmers. IBM paid $2 billion for a large weather forecasting company, experts reckon they were not buying weathermen and were actually after the large data sensing network and the historical weather data archived to train their IBM neural networks. Most weathermen will be out of a job within 10 years as AI neural networks finally make sense of the weather.

http://www.informationweek.com/stra...-together-forecasting-and-iot-/d/d-id/1325362

Hyperlocal weather forecasts coming to a mobile phone near you...
http://www.marketwatch.com/story/ibm-finally-reveals-why-it-bought-the-weather-company-2016-06-15
 
BlowingOldBoots said:
Basic weather patterns are easy to understand from the idealised versions given in books and there are lots of good books on it; any of the RYA books will do. The trick is learn the idealised weather patterns and sequence of weather for depressions and anti cyclones; this is very defined. After that simply start to watch forecasts and compare to your actual conditions (wherever you are, sailing, driving to/from th office, cutting the grass). You will soon pick it up and develop a 'weather eye'. Once the basics are understood you can start to look into the physics behind it. The real skill is interpreting what the trends around you are predicting: change in wind speed / direction, change in pressure, change in clouds, temperature, rain etc. You will soon get the hang of it just by understanding the sequence of weather associated with the basic idealised models. This is the method used for centuries. Look at a synoptic chart every day as near to the same time as possible, I found this greatly helped when I was learning.

Here is the link to the RYA eBook http://www.rya.org.uk/shop/pages/product.aspx?pid=E-G1(RYADefaultCatalog) The paper version is also available.
Click to expand...

Exactly what I did. The RYA book (unless they've ditched the wonderfully communicative watercolour Illustrations and replaced them with hard mechanical computer graphics) sets out the essential theories of main weather systems and local effects very clearly.

Then do as BOB says and follow the forecasts daily.

Don't forget to look at the sky and guess the humidity!
 
jonjo5 said:
The weather does not work, weather just happens and the mechanics defy human comprehension.

Weather forecasting is a pseudo science but hope is on the horizon in the form of computer programmers. IBM paid $2 billion for a large weather forecasting company, experts reckon they were not buying weathermen and were actually after the large data sensing network and the historical weather data archived to train their IBM neural networks. Most weathermen will be out of a job within 10 years as AI neural networks finally make sense of the weather.

http://www.informationweek.com/stra...-together-forecasting-and-iot-/d/d-id/1325362

Hyperlocal weather forecasts coming to a mobile phone near you...
http://www.marketwatch.com/story/ibm-finally-reveals-why-it-bought-the-weather-company-2016-06-15
Click to expand...

Now that will be interesting.. I understand the mechanics, but I wonder how Watson will handle chaos theory/random element... I'm assuming they are hoping to throw so much data at it that Watson will just get 99.9% (name your high percentile) right???
 
Fantasie 19 said:
Now that will be interesting.. I understand the mechanics, but I wonder how Watson will handle chaos theory/random element... I'm assuming they are hoping to throw so much data at it that Watson will just get 99.9% (name your high percentile) right???
Click to expand...
The technology that IBM used to create Deep Blue and defeat Gary Kasparov 20 years ago is considered ancient and Micky Mouse these days in Artificial Intelligence circles.

The leading edge of AI no longer uses hand crafted logic, instead convolutional neural networks learn to recognize patterns and features without human guidance from millions of examples in a data archive. Another branch of neural nets known as recurrent nets support the notion of the passage of time and help the net detect movement and infer velocity. A more recent advance is called a regenerative net, here the network is driven in reverse and challenged to imagine what happens next based on the current state of trained comprehension. The final step of AI weather forecasting might involve reinforcement learning, here weather prediction would be presented to an AI learning agent in the form of a game with rewards for accuracy, a weather prediction agent encoded as a game playing policy in a neural net could move weather systems ahead in time and get positive or negative reward depending on divergence from archived data. Over millions of training episodes the neural net policy will extinguish bad thoughts on how the weather moves and refine good thoughts.

I have not read any description of IBM's plans or how AI weather prediction works in general but imagine they will be employing some combination of the above.
 
jonjo5 said:
The weather does not work, weather just happens and the mechanics defy human comprehension.

Weather forecasting is a pseudo science but hope is on the horizon in the form of computer programmers. IBM paid $2 billion for a large weather forecasting company, experts reckon they were not buying weathermen and were actually after the large data sensing network and the historical weather data archived to train their IBM neural networks. Most weathermen will be out of a job within 10 years as AI neural networks finally make sense of the weather.

http://www.informationweek.com/stra...-together-forecasting-and-iot-/d/d-id/1325362

Hyperlocal weather forecasts coming to a mobile phone near you...
http://www.marketwatch.com/story/ibm-finally-reveals-why-it-bought-the-weather-company-2016-06-15
Click to expand...



There is nothing pseudo about meteorology. Atmospheric physics are well understood. However there is no prospect of totally accurate weather forecasting. Application of neural networks is nonsense. If you really understand the physics you will know that large scale patterns may be predicted well ip to 6, 7 perhaps 8 days ahead. I use the word "may" advisedly. Detail within the large scale is predictable in principle for shorter period, maybe a few hours for something the size of a large thunderstorm. Detail much smaller than that - say a few km in size is unlikely ever to be predictable. For example, taking the argument to the extreme, gusts and squalls will only ever be predicted implicitly.

Many sailors and some providers of forecasts do not understand the limitations inherent in weather forecasting.
 
franksingleton said:
There is nothing pseudo about meteorology.
Click to expand...
It is a notoriously sub optimal science, perhaps the underlying science is junk and completely inadequate for predicting how trillions of energetic molecules will interact in the atmosphere.

franksingleton said:
Atmospheric physics are well understood. However there is no prospect of totally accurate weather forecasting.
Click to expand...
People will settle for noticeably more accurate. It is a competitive world out there, if a rival science delivers better weather forecasts the old guard will be out of a job.

franksingleton said:
Application of neural networks is nonsense.
Click to expand...
Until 5 years ago the 30 year old science of speech recognition claimed the same. Today neural network based speech recognition has replaced the previous technology.

Plenty of experts are excited about neural network weather forecasting, it seems you are not the expert you think you are.

http://www.sciencedirect.com/science/article/pii/S221201731200326X

http://dl.acm.org/citation.cfm?id=2744521

http://www.nssl.noaa.gov/users/brooks/public_html/papers/hallbrooksdoswell.pdf
 
You must log in or register to reply here.

Share this page

Top