Jump to content
Will it snow? 7 day forecast Weather Radar Will it be cold?
Snow?
Local
Radar
Cold?

Theresnoway

Members
 View Profile  See their activity

  • Posts

    180

  • Joined

  • Last visited

 Content Type 

Posts posted by Theresnoway

    Model Output Discussions 18z 31/01/2016

    in Forecast Model Discussion

    Posted

    Should people who are looking for the 6z GFS to verify, be concerned that the operational for London T850's was almost an outlier versus it ensemble suite for the coming weekend? Looking back at the 0z it was not much better, although joined by a couple of the ensembles.

    There is definitely a trend towards colder, but the mean does not drop below -5, and that is not until Monday 15th February, well into la la land I would suggest.

    Caution required, as ever.

    • Like 2

    Model Output Discussions 18z 31/01/2016

    in Forecast Model Discussion

    Posted

    I'd be very wary of the ECM's solution at 9/10 days, it could be the first to pick up on a new trend, but with little support from the other models, I for one will not be dusting off the sledge quite yet.

    It'll be interesting to see where the operational sits compared to the ensembles, I'm expecting it to be an outlier later in the period!

    • Like 1

    Model Output Discussions 18z 31/01/2016

    in Forecast Model Discussion

    Posted

    1 hour ago, johnholmes said:

    Folk can take this or ignore but it might help some of you that are relatively 'new' to model watching.

    Beyond about T+144 hours do NOT compare each new run to the last one, such as 06 to 00. Nine times out of 10 at those ranges you will get variations from run to run. To get some coherent idea using just the GFS or ECMWF or any other model compare like with like. Compare the 06z this morning with that from yesterday. Is there some kind of consistency? If there is then the model may be picking p a consistent signal. Then look at another time output, 00 or 12z and see if either of them are showing a similar consistent pattern. If they are then it would be fair to assume that the model at all 3 times has probably got the overall pattern correct. Detail would become more apparent as T+00 approaches. If this pattern seems to fit the overall upper air pattern that others are showing, in my case, the anomaly charts, then it is a pretty good indicator that the indicated weather pattern will evolve.

    Try it over a few days to see how you feel about it.

    Hello John,

    It would be really helpful to understand the rational to viewing the charts in this way. I can think of 2 possible reasons, but maybe there is a third.

    1. Anything beyond 144 is too far out for the models to grasp properly, therefore you should always expect changes at this range between the same timed runs?

    2. Different data sets are fed into different runs throughout the day?

    However, I do wonder that if it is for the first reason, then as long as the models (regardless of which run it is), are showing a similar outcome, then the likelihood of that outcome is just as likely?

    Thanks in advance

    David

×
×
  • Create New...