Jump to content


  • Content Count

  • Joined

  • Last visited

Community Reputation


Profile Information

  • Gender
  • Location
    South Croydon

Recent Profile Visitors

2,859 profile views
  1. The operational is run in high resolution out to 10 days. The ensemble members are not. If you want to compare the model in the early (<10 day) to the ensembles you'd be better of looking at the control which uses the same initialisation data as the op but runs in low resolution. The impact of high and low resolution on the model can be gauged by comparing operating to control.
  2. I notice you've been mentioning the lack of aircraft data in the data assimilation stage is affecting the accuracy of the NWP. This looks fairly logical, but I didn't know how much of an impact it would have, or the amount of input data that comes from this source. After a bit of hunting on the internet I found a study on the accuracy of the modelling (RMSE) when removing each source of input data (admittedly the focus is short run predictions). Aircraft data is most significant for wind and temperature and the study finds it to be significant. So my conclusion is that it must be a significant source of error in the modelling. Sorry couldn't get the link to the study to work. The title if you are interested is: Observation System Experiments with the Hourly Updating Rapid Refresh Model Using GSI Hybrid Ensemble–Variational Data Assimilation
  3. Given the ensembles have their initialisation data tweaked (control excluded) you could argue that they are wrong from T+0. To my mind the ensembles should be used to give more substance to the operating run. If the ensembles and op follow a similar path then you can take the op as being more reliable. In other words tweaking the initial data had little impact on the output. Also often you can pin-point the time where the ensembles and op diverge giving you an idea of the timescale that uncertainty is creeping in. The greater the spread becomes the lower the confidence in the op becomes. The other big use of the ensembles is to look for clustering to spot potential trends. The other point is that the ensembles are run in lower resolution than the operating run. This difference explains divergence between the control and op which both are run using the same initialisation data. Other than for fun picking an isolated ensemble run adds little.
  4. My god this is getting frustrating - I'm not even concerned about cold right now, I'd just like a little dry. Every time the models seem to show high pressure for the UK it's out in FI and disappears in the next run. Even more frustrating given a pretty unorganized trop polar vortex. My god if we could get some amplification of the pattern rather than this on onslaught of low pressure. This must be 3 months of a relatively similar pattern. Please god just a little high pressure. Walking the dog is becoming a miserable experience. Any kind of block - I'm way beyond caring about cold right now. I can't see anything convincing in the models to suggest it's going to stop. Even the strat is not giving much hope - possible late Jan warmings but that's it.
  5. I use that site too. The radar is good. They also use a combo of GFS and other models (albeit free ones) for their forecasts.
  6. My head office where I have to visit every few months is in LaCrosse Wisconsin currently -24C just before mid-day. Remarkably that's a warm-up for them. A lot of the time you just can't be outside as it's too cold. Issues with frostbite on exposed skin and cold air affecting the lungs. Mississippi freezes and people drive on it and go ice fishing.
  7. Wow big upgrade overnight for snow in the South London/North Surrey area from the Met. Double flake symbols instead of single and extension of the period at both ends. Was 6pm to 4am now showing as 5pm to 8am. I have my finger crossed that they're right.
  8. Interesting height rises to the NE late in the GFS run (day 15 posted) Could this finally be the impact of the SSW hitting the trop? Will need to see if this sustains in future runs and whether it gains any cross model support once it comes into the ECM 10 day timeframe. Thought it was worth posting given the reasonably bullish Met 6+ day forecast
  9. My firm has an office in LaCrosse Wisconsin. There's not a lot going on in most of Wisconsin other than corn and dairy. The temps get so low that it rules out being outside for any period of time as any exposed skin starts to freeze. I have been ice fishing on a frozen Mississippi. Massive cold period about to hit the mid-West in the next few days. LaCrosse will be -20C in the day and closer to -30 at night by the weekend.
  10. Yes, the operational is very close to the mean for 7 days out. Recent runs the op and ensemble means haven't been. Does this herald some consistency and perhaps less volatility on the ECM in future runs? Time will tell.
  11. I think it must be ensemble members from the GEFS
  12. I agree about the no panic it's just that I'd rather see a cold op supported by the ensembles rather than an operational that's at variance with it's ensembles. Hopefully, you are correct and the ECM op moves into alignment with the ensembles. My unease is also compounded by the fact that it's the ECM op that is so out of kilter with its ensembles and GFS output. Another way of viewing it that the volatility caused by Trop/Strat interplay is having a major effect on the reliability of the model output and that higher definition compounds this uncertainty. We did see all models struggle significantly with the quick trop impact last year. This year's been a more normal SSW impact (i.e. slower) but we know it's impacting shortly. Fingers crossed.
  13. Well partly and the fact that the initial input data is tweaked for the ensemble members (control excluded).
  14. I don't think you can ignore operating runs given they are high definition compared to the ensembles and controls low definition. The question is, especially on the ECM, what is the operational picking up on that the lower definition control isn't that causes the direction to be so different.
  15. I always think it's like a break in snooker. You know where the balls should roll but in essence, every break is different. The physics look simple but the outcome is surprisingly variable in snooker. The split SSW is like the white ball and the voretex the reds. You know the vortex will dissipate but exactly where and how remains an enigma.
  • Create New...