Jump to content
  • Member Statistics

    17,608
    Total Members
    7,904
    Most Online
    NH8550
    Newest Member
    NH8550
    Joined

December 2024


TriPol
 Share

Recommended Posts

44 minutes ago, Dark Star said:

Awful.  Not that the models are meant to be accurate at that distance in time, but to be completely opposite would say not to even run them for that period.  I guess some day when "AI" is built into the models which would allow them to correct themselves when it sees its own stupidity?

I’m using the AI models right now and outside of clear situational bias situations, the EC Ens is beating them fairly often and by a pretty wide margin  
 

I met with an AI specialist at an industry event in April and the takeaway was that the AI models right now aren’t operationally sound and should be used as a situational risk assessment datapoint rather than relied on for operational use.

 

You can hindcast and use the reforecast if you want I guess, but I think the AI will have much more use in initial conditions and perturbations before it ever becomes operationally superior in intra run correction 

  • Like 3
  • Thanks 1
Link to comment
Share on other sites

Given how conventional weather models are designed and operate, how is "AI" any different? A conventional weather model is an objective prediction not based on any human input (after humans give model latest data), so I can't really see how "AI" is anything different? 

What does an AI weather model do that a conventional weather model (in use since 1970s) does not do? (and if they aren't as good, what's the point of them, what can in future allow them to outdo the ones we use now?) And what prevents operators of those models from incorporating same advances? Why would one not do so? 

I guess I am a bit skeptical about the real nature of "AI" and I suspect it is just jargon for "different technique" since we already have weather models that are objectively derived simulations of a future state of the atmosphere, what can AI possibly be other than that same thing? 

Link to comment
Share on other sites

2 hours ago, NittanyWx said:

I’m using the AI models right now and outside of clear situational bias situations, the EC Ens is beating them fairly often and by a pretty wide margin  
 

I met with an AI specialist at an industry event in April and the takeaway was that the AI models right now aren’t operationally sound and should be used as a situational risk assessment datapoint rather than relied on for operational use.

 

You can hindcast and use the reforecast if you want I guess, but I think the AI will have much more use in initial conditions and perturbations before it ever becomes operationally superior in intra run correction 

Posts like this are why I still skim through these threads - great insights, thanks.  

  • Thanks 1
Link to comment
Share on other sites

12 minutes ago, Snowlover11 said:

Just got to work you, would think a blizzard is inbound, they literally salted the entire parking lot with massive amount of salt.:lol:

Same thing around here.  It is nuts.  Never used to be this way.

I get it if you want to put some salt down but several places as of late this afternoon were literally dumping it.  I came out of an office building and had to do a double take to realize what was on the ground.

Link to comment
Share on other sites

2 hours ago, NittanyWx said:

I’m using the AI models right now and outside of clear situational bias situations, the EC Ens is beating them fairly often and by a pretty wide margin  
 

I met with an AI specialist at an industry event in April and the takeaway was that the AI models right now aren’t operationally sound and should be used as a situational risk assessment datapoint rather than relied on for operational use.

 

You can hindcast and use the reforecast if you want I guess, but I think the AI will have much more use in initial conditions and perturbations before it ever becomes operationally superior in intra run correction 

Can you comment on whether the ECMWF AI model one sees in the model sites, like Pivotal (ECMWF-AIFS, for AI-enhanced deterministic global forecast) is the Google DeepMind GraphCast model referred to in the article below about AI and weather forecasting?  The quote in one of the embedded tweets in the link below says, "The Google DeepMind GraphCast model is available on the ECMWF platform. It's open-source and can predict weather conditions 10 days in advance with better accuracy than current methods."  If so, certainly the article below is at odds with your commentary on their accuracy.  

https://www.weforum.org/stories/2023/12/ai-weather-forecasting-climate-crisis/

Link to comment
Share on other sites

27 minutes ago, Bxstormwatcher360 said:

Some good echos are starting to build to our west and south. Pretty robust for a clipper i might add. I think it switches to all snow when the heavier stuff gets in.

Those heavier echoes are BS, it's barely a flurry under them.

Link to comment
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
 Share

×
×
  • Create New...