stodge

Morning all Still huge divergence in the models as to the projected track. Most take it to the south or SE of Florida but GEM and DVD turn it away NE missing Florida completely while BOM keeps Irma hugging the northern coast of Cuba including a strike on Havana before a sharp turn north as a weaker feature and a landfall to the west of Panama City (the one in Florida, not the one in Panama). GFS and ECM differ primarily on MSLP which are sub-900 with GFS and more in the 920s with ECM.

Evening all While NAVGEM follows GFS, GEM continues the westward drift and after striking Cuba, the storm emerges into the Gulf and is re-energised west of Florida before turning north and impacting the Florida-Louisiana coast west of Tallahassee. Looks as though the 12Z ECM may be following GEM so two distinct options now on the table.

Indeed and further westward corrections could mean a glancing blow for Florida and a direct hit on Louisiana with the storm having further time to strengthen.

Afternoon all The problem for western Europe is storms like Irma not doing what we'd like - missing land, re-curving into the Atlantic and pushing the Azores HP up over us and giving us a benign settled spell. Instead, Irma fizzles in North America and slows the progression of LP out of that region encouraging mid-Atlantic ridging and the trough to sink SE over the British Isles. GFS 06Z OP at T+216: Not exactly what we want to see.

Except that it now mirrors the 00Z GEM: The 00Z ECM slightly further east but the track has veered noticeably west in the past 24 hours with the recurving starting to the south of Florida so the storm moves from north to south either along the Gulf coast, the Atlantic coast or straight through the middle of the state. The Carolinas and the mid-Atlantic states seem to have been spared a hit at this time but you have to wonder IF further westerly movement will bring Irma back into the Gulf of Mexico with all that could follow.

For completeness, here's NAVGEM: https://www.tropicaltidbits.com/analysis/models/?model=navgem&region=us&pkg=mslp_pcpn_frzn&runtime=2017090300&fh=210&xpos=0&ypos=0 GEM 00Z has IRMA slamming into Florida and moving north through the middle of the state. Weaker than the GFS or ECM though with central MSLP at 949 MB compared with the sub-900 MB values seen elsewhere.

Evening all Indeed it is Knocker. I wonder if there is a complex inter-relationship between these major storms which isn't well understood at present. As to Irma, the 12Z models are showing some convergence and two options appear to be on the way out - the "fish" storm and a storm ploughing through the Caribbean into the Gulf of Mexico. Concensus seems to be building toward a NW and later more NNW'ly shift taking Irma through the Bahamas but keeping her to the east of Florida but with landfall somewhere along the east coast of the US looking probable but anywhere from Georgia/Carolina up to New Jersey and even New York all at risk. There's less chance of a clean re-curve back into the Atlantic with Irma making landfall and being absorbed into a trough over Canada. She's looking a very strong storm with central pressure below 920 MB.

Morning all For completeness, here's GEM 00Z OP at T+240: The Gulf of Mexico option is back on the table here with landfalls in Bahama and Florida first. The NHC guidance now suggests a turn back to the WNW at the weekend which will spare Martinique and probably Puerto Rico from direct hits but whether the Bahamas and Florida will be so fortunate is impossible to say at this time.

Irma makes landfall along the Carolina coast and moves up the eastern seaboard reaching New York 12 hours later. ECM sends it into the Gulf of Mexico on the 00Z OP as a severe storm, Plenty of uncertainty and one to watch.

Indeed, for Irma to be a Tropical Storm already with the very warm waters of the mid-Atlantic still to cross makes me think she/it will be a very dangerous storm in a week's time. As you say, predictions on track are varied but GFS, ECM and GEM have all maintained their 12Z projections through to this morning for the most part. ECM is far and away the most worrying as that model takes IRMA through the Caribbean and into the Gulf of Mexico and while it's hard to be sure it looked as though New Orleans might be right in her path but that's still 10-12 days away and could very easily be wrong. Neither GFS or GEM offer much of an immediate threat to land though GFS 00Z would mean a landfall in the far NE of the US or the SE of Canada between Newfoundland and Nova Scotia and that wouldn't be pleasant. So much also depends on shear and how the storm evolves in terms of eyewall development which will no doubt be observed by the recon planes which undertake these missions but I would already be worried if I was on Martinique for example and given how close these islands are both St Lucia and Dominica may be under threat with only a very slight shift in track. For us, IF Irma misses land and re-curves into the Atlantic as a vigorous extra-tropical LP the path would likely pull up a lot of warm air over the British Isles and western Europe and shift the jet firmly north giving us a prolonged benign spell but if she remains a western hemisphere feature, the signs for a descent to cooler and more unsettled conditions approaching mid month are there.

Morning all Changes in low-res this morning dependent on how the models treat IRMA in the medium term. GFS 00Z OP at T+252: This is the critical point. The GFS 12Z OP yesterday had Irma re-curving cleanly into the Atlantic as a vigorous extra-tropical feature and that had the effect of pushing HP up from the Azores over the British Isles and western Europe,. This morning, IRMA is held by a northward ridge from the HP and instead shifts north to make landfall around Maine and passes north to the west of Newfoundland before being absorbed into the LP to the SW of Greenland. The Azores HP doesn't ridge and our weather becomes dominated by LP from the north west. ECM 00Z is very different - here's T+240: IRMA is right on the very left heading into the Gulf of Mexico as a potentially dangerous hurricane. It's prevented from shifting north by the strong HP over the eastern seaboard which steers the storm into the Caribbean and the Gulf of Mexico. A long ridge of HP is over the south of the British Isles but the Iceland/Greenland trough is gathering to push towards us. GEM 00Z OP is different again: As we saw with the 12Z yesterday, GEM is the "slowest" of the three models in bringing IRMA along. Indeed, the three models have stuck to their guns from yesterday in terms of track of this storm. IRMA is well to the east of the Caribbean but the HP over the Atlantic with a weak ridge to another HP over the Upper Midwest "could" mean a westerly turn for IRMA and a strike along the Carolina coast or the storm could break through the ridge. The NHC forecast is for IRMA to turn more WSW as the Atlantic ridge strengthens putting Martinique right in the firing line

It's not dead, I'm here The key to tonight's 12Z output is how the models handle TS IRMA which looks set to become a significant hurricane. GFS keeps it away from land and sends it recurving back north east into the Atlantic as a significant LP which sends HP over the British Isles. ECM is very different sending IRMA into the Caribbean and then north to Florida by T+240. That means no push of HP into the British isles and LP around Iceland can be influential. GEM is probably nearer GFS but slows IRMA considerbaly leaving it still just east of the Windward Isles by T+240. The current track for IRMA as shown by the NOAA has it heading toward Martinique on a WSW'ly course which would take her more into the southern Caribbean. So, plenty of confusion and contradiction and guidance in short supply but IRMA is the key.

Afternoon all For the second day, 06Z OP has picked up this idea of a cut-off LP developing in Biscay and sending a warm SE'ly flow across for the end of next week. To be fair, as others have said, it's a difficult time with plenty of tropical activity to confuse/bemuse/befuddle (delete as appropriate).

Afternoon all Well, certainly a significant change from yesterday here in lowland East London. Light rain around dawn but more significant falls during the late morning and continuing now. Certainly not the torrential stuff of four weeks ago but plenty still to come from the south west as the front makes painfully slow progress east. The main rain band is quite narrow at present (the clearance has passed through Reading) but there's heavier rain developing along the front and that's passing through Littlehampton now with another heavier area just to the east of Cherbourg. I do wonder if this front will slip back west a fraction before it finally clears as mini-occlusions develop along its length.

Indeed, it's raining here in lowland East London, a sharp shower from out of nothing. Looking on the radar, a line of heavy showers is forming from Suffolk through Essex into east and south east London. Not widespread as yet.

Morning all No comment from the usual suspects so I thought I'd have a go. It's once again a mixed bag from the morning output. After what may be summer's last real hurrah today with temperatures in the SE likely to peak in the high 20s, a complete reverse for these parts tomorrow with a lot of rain and temperatures may be in the mid teens at best. The weekend looks reasonable as a long ridge of the Azores HP is encouraged NE by LP coming off the eastern seaboard of the US. That should mean dry, pleasant conditions for most with perhaps chilly mornings and fog patches inland but fine days and temperatures close to or slightly above the late Aug - early Sep values. Into next week and things get more "interesting" as more unsettled conditions look to spread in. The GFS 00Z OP at T+240 as follows: With the Azores HP held out west and pushed north by a small LP now over the islands, it's a NW'ly flow for the British Isles which won't be that warm and the likelihood of showers to northern and eastern parts. ECM 00Z plays with an option we often see which drops the trough between a suppressed Azores HP to the SW and heights over NE Europe: Probably fairly unsettled with rain or slow-moving showers. GEM 00Z OP isn't too far removed from this: While the GFS 06Z OP also seems to be exploring the idea of dropping an LP close to the British Isles in the middle of next week though perhaps further to the SW than ECM which might allow warm SE'ly winds to affect the south and east:

Afternoon all GFS going in a very different direction on the 06Z OP: The evolution into low-res screams autumn with deepening and increasingly vigorous LP systems to the NW and the Azores HP completely suppressed leaving us in a strong WSW'ly flow.

Afternoon all More evidence of mid-Atlantic ridging combined with Greenland heights on the evening models. Here's the GEM 00Z OP at the end of the run: GFS OP takes it much further and puts the British Isles in a cool unsettled N'ly or NW'ly for the Bank Holiday.

Morning all Yes, Hurricane Gert doing us a huge favour by lifting the jet north next week and allowing pressure to rise from the south so an increasingly dry and settled theme to the models this morning. Not hot, it's fair to day but pleasantly warm - the problem is the HP meanders around to the west and sometimes drifts far enough away to allow weak frontal systems to move down from the north west bringing in cooler air. On other occasions, the HP drifts NW and with pressure lower to the south with the Iberian LP, there's always the hint of an E'ly flow for the far south and possible attendant thundery showers but little evidence of a plume or any influx of hotter air at this stage through to month end.

The key difference is how the two models deal with Tropical Storm Gert. I would guess UKMO moves it through quicker absorbing it into the LP which then comes over the UK as illustrated. With GFS it's slower and the LP is slowed allowing HP to develop to the east.

Morning all Looking at the models this morning, I'm struck by the influence of events to the south rather than the south west. A lot of people are looking at the Azores HP but perhaps they should be looking at the African/Iberian LP which GFS in particular makes a lot of trying to bring it further north. The movement of this potentially thundery area forces the HP at first north over us and then splits the HP leaving us under the influence of the trough with heights shunted to the north. GEM phases the Iberian and Atlantic LPs into a single feature as a mid-Atlantic ridge builds behind it at T+204: GFS doesn't quite do the same but the trough from the south is hugely influential as it shunts the HP north: ECM does nothing with this at all and is again out of kilter with the American/Canadian models. The role of the trough to the south is significant as it either keeps the HP to the north or prevents it building across from the south west leaving us in the trough trap as I termed it the other day between the trough to the north or north west and the thundery trough to the south.

Morning all I'm not quite sure what to make of the models this morning with some very divergent evolutions. GEM is the most progressive in terms of bringing the LP through as early as Thursday and then (as yesterday) toys with the idea of somebody's ex TS blowing into an intense LP which drags the jet back north so Frosty has a nice chart to show us all come T+240. Trouble is, neither ECM nor GFS want to play that game. ECM is a tad better than yesterday's 00Z which wouldn't be difficult while GFS is perhaps a notch worse after a promising start though we see plenty of northern blocking and end up with a thundery low developing over NE France. I suspect we'll get a reasonable few days in the immediate future before the next spell of prolonged rain towards the end of the coming week after which it's anyone's guess.

Actually, I didn't think they were too bad. GFS 00Z the best of the bunch with the jet retreating north and for the south of Britain at least, a preponderance of settled weather. ECM 00Z is poor and there's no disguising it and it ends with a mid-Atlantic ridge so we can kiss goodbye to the Bank Holiday if that verifies. GEM 00Z takes quite an intense little feature (somebody's ex-Tropical Storm) and runs it across the Atlantic and that pulls the jet back north with it - let's not forget we are in hurricane/TS season - once again leaving the south close to the ridging Azores HP. It's so often the case much of August is poor but as soon as the Bank Holiday passes, the weather settles and warms and we have a fine spell into early September. Still a lot of options on the table by this time next week.

Afternoon all Very little for fans of heat on offer this morning it seems. Transient ridges and half-hearted attempts to build the Azores HP north-east are all the positives on offer. We are caught in one of three "trough traps" at present. Either the trough is aligned NW-SE through the middle of the British Isles between the Azores HP and the Scandinavian HP or, as GEM suggests this morning, a ridge between heights over Scandinavia and Greenland keeps the trough aligned further south with the jet and sends a succession of LP systems across the British isles bringing wind and rain for all. The other option sees LP from Iberia move up between the two HP areas and bring thundery downpours. GEM 00Z at T+240: That's pretty moderate for fans of summer. ECM goes another route but it's not much better with the mid-Atlantic ridge keeping our air sourced from cool places: GFS 06Z OP at T+234: Far and away the best of the bunch with the jet aligned further north and the SE not looking too bad. GFS 00Z OP at T+240: Very different again. The main cause for hope is that there are so many options on the table and though most of them look bad, the fact none of them seems especially favoured means better options can't be discounted (as I sit here clutching an economy pack of straws).

The main rain is still to the south and east of London but a new pulse of moderate to heavy rain has come SW out of Norwich and is heading for the capital. The stream from Norfolk - perhaps we can call it "The Partridge Streamer" - is showing signs of easing possibly by sunset.