-
Posts
226 -
Joined
-
Last visited
Content Type
Forums
Blogs
Gallery
Events
Learn About Weather and Meteorology
Community guides
Everything posted by Trom
-
Out of interest, does anyone else seem to find that the Met constantly underpredicts the min low temp overnight? Last night -3 was predicted for my area (South Croydon) whilst the weather station in the garden said -5.4. Tonight it has +1 all evening but the weather station is reading -2.7. I don't notice this with daytime temps. Noticed it over the last couple of years too.
-
Probably quite difficult to do the computations for the NWP models on a slide rule back in the 80s Matt. My ZX81 with its 1K of memory would have struggled. Joking aside we've come a huge way in computing power, and the understanding of the teleconnections and their interrelationships is something that has really come on in my time on here. Sorry mods - meant to post this on the moans and ramps thread please move or delete.
-
Model Output Discussion - Mid Summer Onwards
Trom replied to Paul's topic in Forecast Model Discussion
I would like the answer to that question. I thought previously that the control had the same starting data, and the only difference with the operational run was the model resolution. Given they now have the same resolution but different results there has to be another variable at play. -
Model Output Discussion - Mid Summer Onwards
Trom replied to Paul's topic in Forecast Model Discussion
Mushy am I right in thinking that due to the averaging process in the anomaly charts (across models and time) it would take something to show up on one of the models for a more consistent time (i.e number of runs) before the anomalies would show it? Just checking my understanding. -
Model Output Discussion - A cold start to Spring?
Trom replied to Cheshire Freeze's topic in Forecast Model Discussion
Yes and this second stronger reversal is only just happening now. We know the models struggle with these events from past experiences and the extra levels of vertical resolution do not seem to have noticeably dampened the volatility between runs and models. The hope for me is the first warming puts us in a pattern like you posted and the recent downwelling compounds the easterly potential in the very long term. Could be a late start to Spring. -
Model Output Discussion - Into February
Trom replied to damianslaw's topic in Forecast Model Discussion
With the multiple reversals occurring I wondered if the NWP models would be able to quickly assimilate the impacts of a SWW. They have overall been hugely volatile in the past but I thought the extra verticle levels might reduce that confusion. 2018 remains a bit of an outlier in terms of how quickly the models built the likely impacts into their outcomes and that seems to have been largely due to the rapid speed of the reversal downwards. These events seem to be showing that the models still have a lot to learn. Not surprising I guess due to the lack of historical data available to backtest the algorithms and the fact that each event is rather unique. I'm assuming that the spread of outcomes between runs, in the ensembles, and between models makes the averaging process of the anomaly charts rather meaningless right now. I guess the best way is either to hunker down until a more consistent outcome is determined or focus on the teleconnections. Allround the final location of high pressure looks very undetermined. -
Model Output Discussion - Into February
Trom replied to damianslaw's topic in Forecast Model Discussion
7 days to the reversal and then 2 weeks to 2 months for the impact on the trop. So even with a quick response, we're looking at 21 days. I also remember the models having a lot of fluctuation/volatility when previous SSW events occurred. So I'm thinking at some point in March we will see the impacts. -
Model Output Discussion - New Year and Beyond
Trom replied to BlueHedgehog074's topic in Forecast Model Discussion
I seem to remember from previous events that the impact on the trop was not initially well modelled by any of the models and led to periods of volatility in output before things settled down. -
Model Output Discussion - Christmas week and beyond
Trom replied to nick sussex's topic in Forecast Model Discussion
Seattle's weather is very similar to ours largely for the same reasons - being east of a large pool of water. -
Model Output Discussion - Christmas week and beyond
Trom replied to nick sussex's topic in Forecast Model Discussion
It's roughly a correlation of 0.9 on that graph which can hardly be called terrible in the world of correlations. Again I would point out it is for the entire Northern Hemisphere and as we know the equator and polar areas typically have a much higher correlation than the mid latitudes. I tend to feel that these verification charts lead people to believe that the models are far better at forecasting for the UK than they really are. Remember we are mid lat and to the east of a big body of water. I don't think it's realistic to believe that the 0.9 correlation applies to the UK. I'm not trying to flatter the GFS but really pointing out that all models struggle with our location. -
Model Output Discussion - the cold has arrived
Trom replied to Community Team's topic in Forecast Model Discussion
Lovely analysis. Just using your arrows you can say you want 3 and 4 to prevent 2 to allow 1 -
Model Output Discussion - the cold has arrived
Trom replied to Community Team's topic in Forecast Model Discussion
-
Model Output Discussion - the cold has arrived
Trom replied to Community Team's topic in Forecast Model Discussion
I'd imagine this scenario being so rare means the models have not been back tested with similar data. -
Model Output Discussion - the cold has arrived
Trom replied to Community Team's topic in Forecast Model Discussion
A little worrying that Op and control follow each other so closely -
Model Output Discussion - the cold has arrived
Trom replied to Community Team's topic in Forecast Model Discussion
I posted the images for both 12Z models on the SE pages and it fades South into France. -
Well from a South London point of view Arpege (rolling out now) gives more of a chance than the Arome model due to the pivot location. Also backed by the Met forecast that's gone for more substantial snow showers with a higher probability. Definitely one to be glued to the radar for in my locale. Those to the East and Kent especially look good regardless
-
Model Output Discussion. 3/12/2022. How cold will it get?
Trom replied to chionomaniac's topic in Forecast Model Discussion
You get 78 if you forget the division sign!!