cloudy cloudy

Author Topic: Wxsim modelling in the UK  (Read 7444 times)

0 Members and 1 Guest are viewing this topic.

Offline broadstairs

  • Stuart
  • Posts: 7,034
  • Broadstairs, Kent, UK
    • Broadstairs
Wxsim modelling in the UK
« on: March 30, 2005, 03:19:42 PM »
I am wondering how other wxsim users in the UK are getting on. Specifically how are the forecasts turning out? I have been playing for a few days since Tom sent my customisation data and have been getting reasonable results on temps, wind direction and speed but I am finding the amount of cloud predicted to be rather more than expected espcially 3 or 4 days out. How are others finding it. I'm directing my comments at UK users mainly since there is much less freely available data for the UK than in the US in particular. If any UK users have found other data sources for the UK please can they share them with me?

BTW I have found Tom (Wxsim author for those who dont know) to be very helpful in just the same way as Brian is, very refreshing to find people like this supporting software!

Offline cirrus

  • Posts: 1,097
  • Davis VP2 Plus
  • The Dalles, OR USA
    • Mid Columbia Weather
Re: Wxsim modelling in the UK
« Reply #1 on: March 30, 2005, 09:22:44 PM »
Hi !
I know I'm from the U.S., but I was wondering if you have tried using Ready Meteogram data from the NOAA Air Resources Lab @ http://www.arl.noaa.gov/ready/cmet.html ? It's world wide
and the is data  free. It is quite accurate for cloud cover (and everything else)....for me anyway. I have found FOUS can "exaggerate" the results

Offline broadstairs

  • Stuart
  • Posts: 7,034
  • Broadstairs, Kent, UK
    • Broadstairs
Re: Wxsim modelling in the UK
« Reply #2 on: March 30, 2005, 09:25:26 PM »
Thanks Kevin, yes Tom pointed me at that when I started up and it is very good.

Offline cirrus

  • Posts: 1,097
  • Davis VP2 Plus
  • The Dalles, OR USA
    • Mid Columbia Weather
Re: Wxsim modelling in the UK
« Reply #3 on: March 30, 2005, 09:31:07 PM »
My only problem with Ready, is deciding which meteogram is proper for the time of year.

Offline broadstairs

  • Stuart
  • Posts: 7,034
  • Broadstairs, Kent, UK
    • Broadstairs
Re: Wxsim modelling in the UK
« Reply #4 on: March 30, 2005, 10:01:49 PM »
So far I'm just using the one Tom pinted me at. I dont know enough yet to make any other choice :?

Offline G4GJR

  • Posts: 11
  • Newport Pagnell, Bucks, UK
    • Newport Pagnell Weather
Re: Wxsim modelling in the UK
« Reply #5 on: April 14, 2005, 01:33:25 PM »

Hi All,

The READY data is very useful, Tom has suggested in his instructions which Meteograms to use. I quite often use this data to produce a forecast without WXSIM. I can get reasonable results but need to modify the raw data to get it right - and that's the difficult bit!. Better to leave it to WXSIM!!

Terry (G4GJR)
Newport Pagnell (home of the ASTON MARTIN)
North Buckinghamshire
UK

http://www.janter.demon.co.uk/

Offline administrator

  • Administrator
  • Posts: 10,874
  • Get answers to your questions - use Forum SEARCH!
  • Blackpool, UK
    • Weather-Watch.com
Re: Wxsim modelling in the UK
« Reply #6 on: April 14, 2005, 04:42:21 PM »
I also use the READY data. I use the GFS model (not all models cover the UK) and pick the Default+winds version (usually the 84 hour model).

The big problem I see is that when the READY model says that the wind direction will change significantly a couple of days out, WxSim wants to use MOS data to update it's advection info. I don't know what MOS data is, but I do know it isn't available in the UK. At this point I import the (regional) Synop data for the new wind direction, but I'm not sure whether this is sensible or not because the winds over the intervening couple of days would probably have changed the conditions from the regional stations I'm now using. This is where a bigger model works better because I assume it would have worked out the effects over time of changing conditions in numerous cells rather than using advection over longer distances/periods.
Chris
Weather Forum Administrator

Offline Tom Ehrensperger

  • Posts: 2,659
Re: Wxsim modelling in the UK
« Reply #7 on: April 15, 2005, 03:39:51 AM »
Hi All,

In response to your various posts about READY, MOS, advection, etc...

First, I'm very glad READY exists. especially for the GFS model, which is available worldwide!

U.S. users have a few other products at their disposal:

FOUS: a dense block of code, with information a lot like READY's (but not quite as versatile), from the NAM and NGM models.

MAPS: very up-to-date model upper-air soundings for almost all North American airports, based partly on RAOB soundings, but also on aircraft data.  A nice, and often superior substitute for RAOB data.

MOS: 'Model Output Statistics', forecasts specific to each of hundreds of U.S. surface stations.  This is forecast data which comes in very handy as surrogate advection data in WXSIM after winds have shifted.

For those of you in the UK, or elsewhere outside North America, now that you know what you're missing, the question is how to best handle things in WXSIM.

Fortunately, your options are still pretty satisfactory.  The GFS model seems to be quite good overall, and if you do find systematic errors (i.e. too cloudy, too much rain, etc.) you can make modifications right there on the Interrupt Planner, by selecting the item you want to change at the top of the form, and then clicking on the graph with the mouse to change that item's curve.

As for MAPS, you're not missing too much, because the UK and NW Europe have plenty of dependable upper air sites among which WXSIM can interpolate, almost always less that 12 (and sometimes less than 2) hours old.

Regarding MOS - you are missing something useful there.  MOS is great for advection surface data after wind shifts - say, several hours or a day or two into the forecast, when the original METAR or SYNOP data used at the start is rendered useless by the passage of time.  In fact DO NOT use METAR or SYNOP for advection after a wind shift.  It's as likely to hurt as to help!

So, what DO you do for advection data after wind shifts?

To play it safe, you can just use Neutral advection.  Temperature changes may be a bit understated, with cold spells not quite cold enough and warm spells not quite warm enough, but since surface temperature gradients in the rather maritime climate of the British Isles and NW Europe are often small, the errors may not be significant, especially since cloud cover, precipitation, and upper air temperatures (taken care of rather nicely by the READY GFS data) are usually more important.

For a somewhat more 'proactive' approach, I suggest the Default (Frontal Codes) option.  It uses WXSIM's grid of worldwide climatological normals, along with wind direction and the local site's degree of 'continentality' and current conditions, to make an 'educated guess' about the likely upwind temperature and dew point profiles.  If you don't like the first guess it comes up with, you can use the horizontal scroll bar on the right side of the Advection form to change the 'frontal code' (in increments of 0.1) to whatever looks most realistic.

Of course, you might ask how you're supposed to choose what's most realistic!  This is where experience with using the program in your area can help.  Make a habit of inspecting the upwind temp and dew point profile graphs at lower right, whever you import surface data.  This, along with looking at data plots (in WXSIM or elsewhere) will give you a much better idea of 'realistic' in your area.  Hopefully, the first profile WXSIM proposes will be fine, but as you learn more you may well want to alter it.

There's also the older 'Two Upwind Sites' option, which I seldom use, and the Direct Click option, which I use mainly for experimental and developmental purposes.  You might want to play with them, but I think Neutral and/or Default will be your usual choices.

Note that WXSIM, by its very nature, requires a degree of interaction and participation from you.  This is both a weakness and a strength (hopefully more the latter!).  It's a drawback in that it requires care on your part and is not really a completely independent forecast.  The nice thing about this is that it can benefit from your forecast skills, and also help develop your forecasts skills as it forces you to pay particular attention to variables that matter to the forecast.  It's almost a 'symbiotic' relationship between the user and the program, hopefully yielding something better than either could alone.  The scenario testing (multiple runs with slight variations) can be very instructive as well as providing a range of likely outcomes.

I hope that helps!  Please feel free to ask questions and offer comments and feedback!

Tom 

Offline DMrqwq

  • Posts: 77
Re: Wxsim modelling in the UK
« Reply #8 on: June 27, 2005, 10:37:32 PM »
Hi folks

I have had just downloaded WXSIM and havent got a clue how to  work it! Any help?

Cheers

Dale

Offline administrator

  • Administrator
  • Posts: 10,874
  • Get answers to your questions - use Forum SEARCH!
  • Blackpool, UK
    • Weather-Watch.com
Re: Wxsim modelling in the UK
« Reply #9 on: June 28, 2005, 07:23:26 PM »
It not really a plug and play piece of software. My best suggestion is to print out the manual and then sit and methodically work through the examples given in it. That will give you some idea of how to drive it. However, there are loads of options, some of which are highly technical (or at least they appear to be to me) and I suspect you could spend many hours working out what all the options mean and how to make use of them.
Chris
Weather Forum Administrator