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

Paul

Site development
 View Profile  See their activity

  • Posts

    12,348

  • Joined

  • Days Won

    53

 Content Type 

Posts posted by Paul

    Model discussion highlights

    in Forecast Model Discussion

    Posted

    31 minutes ago, Scott Ingham said:

    100% yeah this is all emotions and past disappointments scewing proper analysis of what’s going on. All the available data points in the right direction bar like you say the odd run.

    Not wanting to pee on anyone's parade here, but proper analysis would surely be far more nuanced than that, especially going beyond the weekend in terms of specifics of the cold's longevity, how far south it reaches/stays & any snow risk around that. All the data points don't point in one direction. Many point in a similar direction, but there are operational runs (namely the GFS today) and plenty of ensemble members from various models which show differing solutions.

    Yes, the colder outlooks in the shorter to medium term do look more likely right now and I'd certainly take the UKMO and ECM above the GFS 9 times out of 10, but that's not the same as the colder/coldest options being the only possibility on the table. 


    Original post: https://community.netweather.tv/topic/99706-model-output-discussion-into-2024/?do=findComment&comment=5004114
    • Like 1
    • Insightful 1
    • Popular Post
    • Popular Post

    Model Output Discussion - Into 2024

    in Forecast Model Discussion

    Posted · Edited by Paul

    31 minutes ago, Scott Ingham said:

    100% yeah this is all emotions and past disappointments scewing proper analysis of what’s going on. All the available data points in the right direction bar like you say the odd run.

    Not wanting to pee on anyone's parade here, but proper analysis would surely be far more nuanced than that, especially going beyond the weekend in terms of specifics of the cold's longevity, how far south it reaches/stays & any snow risk around that. All the data points don't point in one direction. Many point in a similar direction, but there are operational runs (namely the GFS today) and plenty of ensemble members from various models which show differing solutions.

    Yes, the colder outlooks in the shorter to medium term do look more likely right now and I'd certainly take the UKMO and ECM above the GFS 9 times out of 10, but that's not the same as the colder/coldest options being the only possibility on the table. 

    • Like 26
    • Insightful 5

    Moans, ramps and banter

    in Spring Weather Discussion

    Posted

    I do think sometimes, some people miss one key point: It's a discussion forum.

    If you're looking for a concise local forecast, a forecast with no bias towards where people live or what weather type they're interested in, you aren't going to get that in a discussion thread. 

    There are lots of great places to get those things, inc on this website. But to expect the model thread to be an open, active, fun and busy thread, along with being perfectly concise  and with no bias ( either based on poster location or weather type preference) is just an impossible ask, and always has been.

    • Like 8

    SE, London and East Anglia Weather Discussion - Dec 2023 on

    in SE, London and East Anglia Weather Discussion

    Posted

    2 minutes ago, dubmuffin said:

    Surely these showers are not coming from out in the North sea and Europe, but they being created around the estuary, so as long as there is a big temp difference between the upper air and the sea, and the flow remains favourable, they just keep coming?

    Yep, to an extent - more info here:

    AdobeStockSNOWcoastLomond2.jpg

    favicon.ico Sea Effect Snow - the UK version of Lake Effect snow. Very cold air across the North Sea

    WWW.NETWEATHER.TV
    Very cold easterly flow, found when a high pressure sits over Scandinavia and brings air to the UK from Siberia. If conditions are right, significant snow can result.

     

    • Like 5
    • Thanks 1

    Model Output Discussion - Into 2024

    in Forecast Model Discussion

    Posted

    8 hours ago, Battleground Snow said:

    Do you think that because the EPS are run at the same high resolution as the OP, that it sometimes times causes them all to see the wrong solution early on in the run?

    It always seems like the mean and the op are very close up to around 144, where as the gfs can be quite different to it's mean at that point.

    It's quite amazing the difference in the cold pool distribution at 72 hours compared to the UKMO, but then again I guess we only notice it because it's over the UK, when it's a micro detail globally 

    8 hours ago, Met4Cast said:

    Good question.. honestly that's not something I know the answer to. I have very limited knowledge on how modelling actually works beyond the basic "observational data in >> forecast out". It's an interesting idea though and I'm sure @Paulmay be able to answer the question as he knows a fair amount about how modelling actually works. 

    In theory you'd think higher resolution would be better at picking up small lows etc vs lower resolution, but I guess if it goes wrong then it could go very wrong? Not sure!

    I don't think running at a higher resolution would be likely to make the EPS members more likely to go wrong earlier in the run. I suppose it is possible that in having the resolution to forecast smaller features, some perturbations could be more liable to go off on tangents based on them, but having 50 of them should be a decent barrier to stopping any of that skewing the mean etc.  

    I think in terms of the op being closer to the mean later in the run than the GFS. The fact the EPS has double the number of perturbations probably helps with that, in smoothing out the mean from outliers that may otherwise skew it somewhat if less perturbations were being run. 

    • Like 2
    • Thanks 2
    • Popular Post
    • Popular Post

    Model Output Discussion - Into 2024

    in Forecast Model Discussion

    Posted · Edited by Paul

    Using the insightful reaction - please read.

    The insightful reaction is for insightful posts, and those with enough of those reactions are copied to the model highlights thread.

    It's important that the insightful reaction is used only on that basis in this thread. Some people are seemingly using it for other reasons (for instance, sarcastically at times), which needs to stop. For those who repeatedly do this, we'll have no option but to stop them using any reactions, so as to protect the integrity and usefulness of the highlights thread.

    • Like 22
    • Thanks 2

    Model Output Discussion - Into 2024

    in Forecast Model Discussion

    Posted

    13 minutes ago, WYorksWeather said:

    To be fair those are 2m temperatures, which probably don't take sufficient account of geography. Stations in frost hollows or in the presence of snow cover will probably be quite a lot colder.

    They absolutely do take into account the geography - the models have heights, surface type, area type (urban vs countryside etc) built into them. But, they are bound by their resolution when it comes to that - so the GFS at 0.25 degrees will be averaging the geographical nature of the grid square over that, which is about 17km or so in our neck of the woods. So, it's not going to pick up small features, frost hollows and the like. 

    • Like 1
    • Thanks 1

    Model Output Discussion - Into 2024

    in Forecast Model Discussion

    Posted · Edited by Paul

    6 minutes ago, johncam said:

    Surely they get to see the 12's long before us??

     

    6 minutes ago, Mcconnor8 said:

    No, why would they?

    They may have earlier access to their own models, and probably to some of the others, but we're talking minutes rather than hours here. The models take the time they take to run,  are initialised with data at their run time, then run sometime after that. So there isn't a great deal of flex time wise. 

    • Like 6
    • Thanks 2

    Model discussion highlights

    in Forecast Model Discussion

    Posted

    21 minutes ago, claret047 said:

    Do all models use the same resources to collect their data in order to do their modelling? The reason I ask is that I believe the GFS is an American model (please correct me if I am wrong) and wonder if they have more information as far as modelling the area around Greenland. If so perhaps they can predict that area better than ECM or UKMO. Hopefully that is not the case as we all want the evolutions of the other two models to be correct. 

    They will also use very similar sources of global weather data for their starting conditions, but there will be some differences in terms of the sources and the way they use that data + of course the physics in the models themselves will be different. The location of where the model is run from isn't really a guide to where the models are strongest, though - for instance the ECM and Meto models are generally very good performers for US hurricanes despite them being made by European-based organisations. So, I'd not set any stall on that side of things. 


    Original post: https://community.netweather.tv/topic/99706-model-output-discussion-into-2024/?do=findComment&comment=4996234
    • Like 1
    • Popular Post
    • Popular Post

    Model Output Discussion - Into 2024

    in Forecast Model Discussion

    Posted

    21 minutes ago, claret047 said:

    Do all models use the same resources to collect their data in order to do their modelling? The reason I ask is that I believe the GFS is an American model (please correct me if I am wrong) and wonder if they have more information as far as modelling the area around Greenland. If so perhaps they can predict that area better than ECM or UKMO. Hopefully that is not the case as we all want the evolutions of the other two models to be correct. 

    They will also use very similar sources of global weather data for their starting conditions, but there will be some differences in terms of the sources and the way they use that data + of course the physics in the models themselves will be different. The location of where the model is run from isn't really a guide to where the models are strongest, though - for instance the ECM and Meto models are generally very good performers for US hurricanes despite them being made by European-based organisations. So, I'd not set any stall on that side of things. 

    • Like 14
    • Thanks 1
    • Insightful 4

    Moans, ramps and banter

    in Spring Weather Discussion

    Posted

    Just wanting to follow up on the note at the start of this thread and now at the top of each page in here. I understand when a group of people come together, online, offline or whatever, we're not all going to get on, enjoy each other's company, will have clashes of personality, opinion and so on. But, this and other threads on the community cannot be used to moan about or criticise other members. 

    There is the facility to block other members posts (I know it's not perfect with quotes etc, but it's still helpful), there's more info on how to do that here:

    Beyond that, if you're thinking someone is trolling or disrupting things, please do report their posts, it's the quickest and best way to draw the teams attention to potential issues.

    If you do have any additional concerns or ideas you'd like to discuss, please head over to the private team contact forum and message us - all threads in there can only be seen by the person starting them along with the team:
    https://community.netweather.tv/forum/171-contact-the-forum-team-private-forum/

    • Like 5
×
×
  • Create New...