Which model do you regard as the best and why?

hatrickpatrick

Rather than there being a trend, there seems to be an amazingly diverse range of views here on which models should be relied on more and which are useless, etc. If I was to attempt to suggest a trend over the last few years since I started posting here, GFS and ECM seem to be the forerunners, with UKMO lagging slightly behind and others such as NAVGEM hardly getting a mention. Of the top two, people seem to regard the ECM as more reliable even though it gives less time-detail (24h as opposed to 6h increments like GFS, meaning only three or four non FI charts per run) and the GFS as more prone to abrupt reconfiguration at the last minute after several similar runs. Within that, 0 and 12z seem to be favoured over 6 and 18z runs of the GFS, for which I believe there's a legitimate reason which escapes me at the moment. However, having said that, GFS charts seem to be posted more often as embedded images than ECM charts, so my observation is perhaps inaccurate as to which people prefer.

So which do ye like and which do ye not like? Why so?

maquiladora

You've got to look at all of the runs of all the main models ( GFS, ECM, UKMO and GEM) and take all their output on board. They all have good days and bad days.

I wouldn't bother paying any serious attention to lesser models like NAVGEM etc.

piuswal

I'm sure all serious model producers run verification statistics!

What do they make available publicly and do they provide results that can be compared directly for each major model?

Of course what the ordinary user wants is ; how good any model is at forecasting e. g. rain for his/her area, do they produce anything that goes any way towards answering such questions?

Anyone got any infirmation/knowledge in this area?

hatrickpatrick

piuswal wrote: »

I'm sure all serious model producers run verification statistics!

What do they make available publicly and do they provide results that can be compared directly for each major model?

Of course what the ordinary user wants is ; how good any model is at forecasting e. g. rain for his/her area, do they produce anything that goes any way towards answering such questions?

Anyone got any infirmation/knowledge in this area?

When it comes to weather forecasting I'm primarily interested in isobar charts, but perhaps I'm unusual in that regard. I can usually get a pretty good idea of what the weather will be like by looking at the atmospheric pressure, particularly at this time of year - and it's these forecasts of atmospheric pressure which tend to diverge wildly between models the further out one gets from the present time.

piuswal

hatrickpatrick wrote: »

When it comes to weather forecasting I'm primarily interested in isobar charts, but perhaps I'm unusual in that regard. I can usually get a pretty good idea of what the weather will be like by looking at the atmospheric pressure, particularly at this time of year - and it's these forecasts of atmospheric pressure which tend to diverge wildly between models the further out one gets from the present time.

I'm pretty sure that's one of the parameters, pressure, that's readily measured and easily compared between models.

Rikand

Whatever one they use as the main one on meteociel

piuswal

Rikand wrote: »

Whatever one they use as the main one on meteociel

What parameters do you look at?