Fantasy Island (fi) Threshold Test

[nodrog]

(Mods move topic if required).

Fantasy Island (FI)

As a recent discussion about Fantasy Island (FI) has occured I think from my experience it is of some value for members to agree with a common understanding to this title.

Fantasy Island could be commited to several time frames and re titled as agreed or just bolted to one single time frame.

All members should use this topic to debate where the origin of FI belongs and this to be the agreed threshold test.

Fantasy Island is a period of time after the 'Model Runs' when the confidence of any forecast is purley speculative. Several Model Runs would have to be consistant to produce a better chance of any weather event actually happening closer to real time. This is where the boundary of real time and fantasy island needs to be established and followed as a threshold test.

Discussion and comments below.

Edited October 9, 2011 by Tangerine Lion

[Jayces]

I believe (though I may be wrong) that the problem with FI is that it varies where it is depending on the what the models are showing. I've seen it range from 4-5 days out to a week to 10 days out.

[johnholmes]

(Mods move topic if required).

.

see my recent post in the model thread?

[Gibby]

For the record I always relate to FI as +192hrs on my daily reports.

[stewfox]

(Mods move topic if required).

Fantasy Island (FI)

As a recent discussion about Fantasy Island (FI) has occured I think from my experience it is of some value for members to agree with a common understanding to this title.

Fantasy Island could be commited to several time frames and re titled as agreed or just bolted to one single time frame.

All members should use this topic to debate where the origin of FI belongs and this to be the agreed threshold test.

Fantasy Island is a period of time after the 'Model Runs' when the confidence of any forecast is purley speculative. Several Model Runs would have to be consistant to produce a better chance of any weather event actually happening closer to real time. This is where the boundary of real time and fantasy island needs to be established and followed as a threshold test.

Discussion and comments below.

Most of the FI will be based on model runs e.g T300, so not sure what you mean FI based on things after the model runs ?

JH explain it far better then I can in the Model thread.

[nodrog]

Stewfox : "Most of the FI will be based on model runs e.g T300, so not sure what you mean FI based on things after the model runs ?"

Thank you Stewfox, many members (new or old) are not sure when people state "that is FI". The FI bench mark seems to change like the seasons which I find difficult to understand. A better explanation to this title and a fixed time frame to such model runs perhaps as you explain.

Some members have stated FI within a week before today and some as you show are several days later.

Thank you for the topic replies, more welcomed as Im sure this title will crop-up in other areas.

Without Prejustice : I dont want the more advanced to take a 'pitty' look on new members to prove they are better thus confusing the matter, this topic is only suggestion, not to be taken into serious debate.

From JOHN HOLMES off the Forecast Model topic :

"how about instead of FI we call it CCL

Cloud Cuckoo Land

or

WFN

Weather Fans Nightmare

??

For those new to the idea of model watching there is nothing wrong in looking and hoping at charts that show the weather type you hope for, be it storms, heat or snow and cold. Just as long as you realise the chances of it occurring increase as the time scale increases.

Forecast accuracy at T+12 is probably over 98%

At T+72 about 75-80%

ByT+144 its around 60-65%

By T+240 50-60% as a broad figure

and by T+384 less than 10%.

If you go on those average figures it will give you an idea of just how likely your hopes are to be fulfilled.

At ranges in excess of T+96 its best to compare all the models, especially GFS, UK Met(out to T+144), ECM out to T+240, and also GEM seems to do fairly well at these later outputs but quite so good closer too.

Beyond T+144 and then other things like teleconnections you here folk talk about need to be given as much weighting as the basic synoptic outputs, things like the 500mb anomaly charts from GFS, ECMWF and NOAA, the first two valid for T+240 hours with NOAA issuing 2 charts one for 6-10 days ahead and another 8-14 days with each evening issue.

Further out then AO, NAO, MJO and ENSO all need watching carefully to see if there is support for the basic synoptic model outputs.

Along with these basic ouputs look at the Ensemble outputs for the different models as well.

For what some of these terms that some of you may not be familar with take a look in the Net Wx Guides section, accronyms etc will give you some basic ideas.

Hope this post is ok admin/mods as it may help some folk get their heads around what is likely to become ever more frantic FI searches as we head into winter? "

Hope you did not mind me moving it to here John.

Edited October 10, 2011 by Tangerine Lion

[Methuselah]

I think, I may well be wrong, that in winter, when super-cold is expected, FI tends to occur as soon as 4 days; the same applies in summer when super-hot is desired? Most other times, I don't pay much attention to what's forecast after about 7 days...It varies.

[johnholmes]

Hope you did not mind me moving it to here John.

not at all

[The PIT]

It varies for me generally t96 onwards. However the best way is too look at ensembles and see when the large variation starts. So looking at today's run this suggests FI starts around October the 16th.

[VillagePlank]

Snip - slightly wrong, will repost later.

Edited October 21, 2011 by Sparticle