Help with Forcasting

[nking]

Hi,

I’m a new member of the forum having got a PWS for Xmas. Now that I have the equipment setup and collecting data I now want to set about understanding cause and effect on my local weather. I’m not looking to become an expert but I would like to be able to eventually predict weather for the coming 24 hrs using the PWS data only. Yes, I could just use the met office or similar but that takes away the fun of pitting your wits against nature. I have found a few useful sites to help me http://www.xcweather.co.uk/GB/observations http://www.weatherfaqs.org.uk/ ,but I’m sure that there are others and I hope that some members might point me in the right direction. Are there any good books that would help?

I know that Cumulus uses the Zambretti algorithm (http://www.meteormetrics.com/zambretti.htm) and whilst impressive I do find the descriptions extremely subtle and feel more detail is possible (this comes from a complete novice!! ). I hope that one day I’ll be able to contribute an improvement to this area of Cumulus.

Cumulus appears to provide real time predictions and does change frequently (I did find reference in the forum to making a change to the ini FCPressureThreshold, is this still valid?) and wonder if there might be an easy way of capturing this data so that one could look back and see how accurate the forecast was and how often it changes etc.

I’m sure there are some experienced forecasters in the forum who will give me guidance.

[steve]

I'm inclined to think that you will never be able to do accurate forecasting using just the local data you get from a PWS. But then again, my VP2 doesn't do too bad a job sometimes (better than the Cumulus forecasts) so maybe with some tweaking anything's possible.

The book "Teach Yourself Weather" by Peter Inness claims to be able to get you started with doing your own forecasts, but I don't have a copy - it's only £6.99 delivered from Amazon, so maybe worth a punt; I might even buy a copy myself.

The FCPressureThreshold setting should work.

[beteljuice]

... Cumulus appears to provide real time predictions and does change frequently ...

The 'problem' that the Zambretti algorithm has is not a Cumulus thing, rather it is a 'Fine Offset' thing in that the wind vane is very twitchy and at certain thresholds will change the forecast to and fro.

Also not mentioned in the article on that link (very good BTW), is that the forecast is most reliable at the start of a weather system - typically 09:00 solar time.

[nking]

Steve – Thanks for the book suggestion, I will check it out. I know you’re a very busy man but if you have time or someone else knows the answer to this; Cumulus has a number of calculated fields e.g. dew point, cloud base etc. Is there anywhere I can look at the logic for all calculated fields? Beteljuice mentions the twitchy wind vane but Cumulus seems to do a good job smoothing this out so does this mean that the averaged value isn’t passed onto the Zambretti module?

Beteljuice – (couldn’t find a first name…. Sorry for my rudeness if this is your first name) I thought you might find the Zambretti site of interest, I also came across another algorithm http://www.weather-above.com/Sager%20Algorithm.html alas I haven’t found the logic it uses. Do let me know if you have anything on this Sager logic.

[gemini06720]

I'm inclined to think that you will never be able to do accurate forecasting using just the local data you get from a PWS. But then again, my VP2 doesn't do too bad a job sometimes (better than the Cumulus forecasts) so maybe with some tweaking anything's possible.

Steve, about 6 months ago I bought a package called WXSIM, an Interactive Local Atmospheric Modeling Software.

The package is composed of 2 main programs,
- WXSIMATE "a very handy and time saving tool for collecting Internet and personal weather station data"
and
- WXSIM "a very unique software package for modeling, forecasting, or simply studying weather; WXSIM is perhaps best described as an interactive local atmospheric model - you only needs initialize WXSIM with current data, some optionally downloaded from the Internet, and then "turn it loose" to model the weather for the next few days, but with the option to interact with the program and mix in some of your own forecasting skill and knowledge".

My customized version of WXSIM download Internet data from about 200 different weather data providers within a radius of about 300 Km (such as from automated weather stations, from water buoys, from METARs, etc.). The programs (WXSIMATE/WXSIM) are run 6 times a day and thus create an ever changing updated forecast.

Although some adjustments/customizing were needed at the beginning (such as adjustments for my immediate area and the surrounding areas, including my station's elevation on the north side of a mountain), the 2 programs have been running, as stated above, for about 6 months ... and the forecast produced by WXSIM have been fairly accurate most of the time, but way more accurate than the forecast produced by Environment Canada (the canadian government office responsible for everything weather).

The WXSIM package is not made for everyone (about $200 US) but once setup it needs very little maintenance...

[steve]

Cumulus has a number of calculated fields e.g. dew point, cloud base etc. Is there anywhere I can look at the logic for all calculated fields?

Most of the calculated values in Cumulus use Steve Hatchett's routines: http://www.softwx.com/weather/uwxutils.html

Beteljuice mentions the twitchy wind vane but Cumulus seems to do a good job smoothing this out so does this mean that the averaged value isn’t passed onto the Zambretti module?

Yes, the 'average' direction is used in the forecast calculation.

[beteljuice]

Beteljuice – (couldn’t find a first name…. Sorry for my rudeness if this is your first name) I thought you might find the Zambretti site of interest,

(Zambretti) - You'll find out !
To me - beteljuice
About me - the beteljuice
If you insist on a name string - Bet El Juice (It's a silly story)

I also came across another algorithm http://www.weather-above.com/Sager%20Algorithm.html alas I haven’t found the logic it uses. Do let me know if you have anything on this Sager logic.

I suspect that (or a derivative) is what Davis Instruments are using. The original (The man is still around) involved a big series of look-up tables, and is in most part the Zambretti plus rate of baro change decisions and cloud / sky observations (Which your kit can't do)

The original 'slide rule' and book are no longer available, and in any case were Northern temperate zone specific. (As is the Zambretti 'slide rule')

However the Zambretti algorithm that Cumulus utilises has been given a small amount of flexibility, and should also work for the Southern hemisphere.

I'm begining to suspect that the Davis forecast 'improves' if you ask it to import local METAR (Can anyone say if it does that or not ?)

[daj]

If you insist on a name string - Bet El Juice (It's a silly story)

I always call him Bet. The man plays his cards close to his chest (his choice of course). I did establish he was a he and not a she though, as he slipped up on a post once and gave it away. An international man of mystery

[steve]

I'm begining to suspect that the Davis forecast 'improves' if you ask it to import local METAR (Can anyone say if it does that or not ?)

I'm pretty sure it doesn't have that facility.

[beteljuice]

The man plays his cards close to his chest (his choice of course).

It is of course easy to do some rudimentary checks. However; How do you know that the pointers I give eg. Lady b are not in themselves a misdirection ?

The beteljuice is a neutral, sexless, raceless 'web' nom de plume. Of course I have my own bias / experiences / shortcomings, but I when I "pull someones chain" it is hopefullly recognized as an incentive to learn, or revise ones attitude. (Did I mention a certain degree of arrogance ?)

Also I must admit there is a slight 'glow' when forums, book publishers, coding sites, technical sites acknowledge 'beteljuice'. I think I prefer to be Infamous and anonymous rather than a known quantity.

Those who have real experience of the beteljuice know that the sideways (but true and interesting) option will be the one taken.

Enough information ... back to topic

Thanks for the feed-back Steve.

I forgot to mention that I have a software copy of someone elses work on The Sager Algorithim, (actually worked around some stuff I did), but it is a bit disjointed, and to give it the North / South divide and the 'range' and 'what rate of change' flexibilty requires some serious brain ache - might get round to it one day if nothing else surfaces (why re-invent the wheel ?)

[nking]

I forgot to mention that I have a software copy of someone elses work on The Sager Algorithim, (actually worked around some stuff I did), but it is a bit disjointed, and to give it the North / South divide and the 'range' and 'what rate of change' flexibilty requires some serious brain ache - might get round to it one day if nothing else surfaces (why re-invent the wheel ?)[/quote]

Hi Betejuice,

Could you let me have a copy of the code or point me in the right direction. I would like to take a look (time permitting) and as I'm in the North then this will reduce the brain ache for me.

Not too many replies to my request thus far, it maybe that people like to watch rather than understand?

[beteljuice]

Not too many replies to my request thus far, it maybe that people like to watch rather than understand?

Bless - You must appreciate that if government bodies / shipping / airline magnates with all their global high tech resources are still basically waving a wet finger in the air - then we are still working with sea weed

Here's the link - http://eltiempo.selfip.com/malaga.htm

Enjoy