Dear Global Forecasting System, why are you the way that you are? Serious hat back on, it has a habit of overdoing windstorms & when an ex-tropical cyclone is added to the mix, that bias increases further. I'm not saying such a strong windstorm is impossible, just not likely.

Hurricane Gabrielle is forecast to pass very near or over the Azores as a hurricane later tomorrow through Friday. The IBTrACS database contains four comparable events in September, the most recent hurricane Gordon in 2006.

A busier interlude has been long overdue this Atlantic hurricane season. Most forecast modelling develops the eastern disturbance into a hurricane, & quite a few take the western one that far too. There's some interesting potential for the two systems to interact next week.

A textbook example of concentric eyewalls in Super Typhoon #Ragasa as seen from Chinese radars. This and other radar loops are available at bmcnoldy.earth.miami.edu/tropics/radar/. #HuajiashaCN

While ex-Gabrielle is forecast to bring some very disturbed weather to mainland western Europe this weekend into next week, it now looks likely that the UK & Ireland will be spared, instead seeing dry & perhaps increasingly warm. GFS picked up on that a bit sooner than ECMWF.

A hurricane heading east-northeast away from Bermuda rarely looks this impressive on satellite imagery. Gabrielle currently has peaked sustained winds of 140 mph. Forecast to start weakening this afternoon, & pass near or over the northern Azores as a cat 1-2 hurricane Thu-Fri.

Not very many major hurricanes have been reliably recorded within 100 km of Gabrielle's current position. In the IBTrACS database, there are 7 instances, most recently major hurricane Sam in 2021.

Gabrielle became a major hurricane at 12:00 UTC today, which means the 2025 Atlantic hurricane season has, to 22nd Sep, seen 2 hurricanes with both becoming major. A rare circumstance, with only four prior occurrences since 1900 (IBTrACS database): 1930, 1935, 1991, & 2009.

At 21:00 UTC yesterday, the 2025 Atlantic hurricane season finally produced its second hurricane according to the NHC. For hurricane count to-date, this season is well short of 'active season' pace. Of the top 12 seasons by this metric, all were on at least 4, most at least 6.

Provisionally a UK maximum of 27.6°C today. For recent whole-Septembers that's right in the middle of the distribution for monthly peak temperature. For 2nd halves of September, it's toward the higher end but unremarkable really. Yet here I am remarking! 😄

This is now the hottest day since 7th September in parts of the southeast, & since late or even mid-August in parts of central-southern England. A brief but large change from of late. Here in NE Dorset, 10th-17th September had a mean daily max 0.9°C below the 1991-2020 average.

Current sea surface temperature anomalies around Europe make for dramatic viewing, with strong to extreme anomalous warmth to the north & south standing in stark contrast to an anomalously cool swathe in the central North Atlantic.

Tropical cyclones extract energy from the seas. With the Atlantic basin having been almost totally devoid of those so far this month, the tropical waters have become almost entirely warmer than average, along with most of the subtropical waters.

🔥 in the sky

September 2025 has so far seemed very wet in my local area, but historical stats reveal that it's been nowhere near the wettest in recorded history. Note the absence of 2024 from the top 5. It became the 2nd wettest September on record, yet was 'only' on 47.8 mm as of 17th.

Mixed fortunes when it comes to UK & Ireland rainfall through the rest of September. Especially if GFS' latest det. run is nearer the mark than ECMWF's. Either way a weekend soaking on the cards for part of Ireland, perhaps Northern Ireland, & around the England-Scotland border.

Next week is favoured to bring more settled (dry) conditions in the UK & Ireland. Chilly nights possible depending on how the high shapes up. A risk of a low affecting the south for a time - a scenario that the ECMWF model is a lot more interested in than the GFS & GEM models.

The forecast modelling of low pressure affecting the UK & Ireland this weekend has been rather inconsistent. The scenario of an exceptionally chilly day in central/southern England has all but vanished now, due to a north-westward adjustment to the position of the frontal system.

In the London & SE England region, South Newington had yesterday's highest max temp, lowest min temp, & highest rainfall. One of only a few times I've noticed a location achieving that combo in over a decade of keeping track.

'Gravity waves' have produced a cloud pattern somewhat resembling a ribcage over southern England today. My local area has been underneath one of the near-stationary cloud banks, with low light & faint drizzle making for one of the dreariest September afternoons I can recall.

The typically most active month of the North Atlantic hurricane season has only just managed its first named tropical cyclone, & it's a poorly organised mess. However, Gabrielle is at least expected to become a more respectable system by the weekend, perhaps a hurricane.

Signs are, September 2025's Hadley mean-to-date Central England Temperature will have spent most of the time gradually descending toward the 1991-2020 average. The next few days will interrupt that tendency, before a resumption starting Sunday that might continue to month's end.

Finally got a chance to asess the run of days reaching at least 20°C in my local area that culminated on 8th September, while I was abroad. At 91 days, this was by far the longest on record, the prior highest being 'only' 69 days in 2022. 1943-1995, not one year saw more than 40.

For the time of year, Sunday has the potential to bring some extraordinarily chilly conditions to parts of lowland central & southern England. A northerly flow undercutting a decaying weather front could limit daytime temperatures to 9-10°C, which challenges date record lows...

37°C in north-western France 19 days into September... seems like GFS still has a problem with temperature-soil moisture interaction. ECMWF's 00z deterministic has a peak near 33-34°C which is still high-end for so far into this month.

It's the typical peak weeks of the Atlantic Hurricane season & there's only one system with any clear potential to develop into a tropical cyclone within the basin during the next week. Forecast models aren't showing much interest in anything else in the next fortnight 🤨

This Friday's looking impressively warm or hot for the time of year across much of western & central Europe. It remains to be seen just how much so, with GFS predicting higher max temperatures than ECMWF. The exact progression of weather systems is still being resolved.

Just like last weekend, there's been a westward adjustment to the low pressure this weekend into early next week, which has reduced forecast rainfall totals for eastern parts of the UK. In fact, it's not looking all that wet for the next week there. Different story in the west!

So this is the view from my window right now…

Radar is showing rates of over 150 mm per hour for this cell that’s just crossed Ringwood. Wish I could have experienced that first hand - ideally under shelter of course 😄

Guess who's going to be in the Balearic Islands next week 😆 Full disclosure, the wet (& probably thundery) weather might only be for a night, maybe a day, with generally dry weather either side. Plus, it will be interesting to experience there!

Sparkling sunshine between torrential showers, it could be April if not for the fully in leaf trees.

The Met Office's in-depth summer 2025 analysis talks of how it doesn't stand out from recent experience as much as 1976. This is well supported when looking at the Hadley mean CET measured against the average of the preceding 20 summers.

Rainfall totals since the onset of unsettled weather on 26th August... wide swathes of at least 70 mm, & sizable patches of over 100 mm. For a southern location, I've escaped lightly with 52.8 mm. Totally down to chance, downpours often passing within a mile or two of here.

BlueSky v1.107... As someone who often posts threads, the simple addition of a link to a post you've just made is a monumental advancement 😄

Today's forecast model consensus max temperature here is 19°C, which would end a run of 86 days reaching at least 20°C. However, twice this past week, a forecast of 19°C has proved at least 1°C too low. If today reaches 20°C, forecasts suggest at least 2 more to follow. Tense!

27.1 mm in the first 3 days of September here in NE Dorset. Modest compared to some parts of the south, but still 580% of the long-term average for that period. Already more than seen last July (20.0 mm) & nearly as much as last August (28.5 mm) or June (29.9 mm).

9 days ago versus this afternoon here in NE Dorset.

Predicted rainfall totals from the overnight det. runs of ECMWF & GFS, versus observed up to 6 pm BST. Generally, ECMWF overestimated while GFS underestimated - typical biases. Some local differences, e.g. GFS was too high for the Bristol area. A soggy day widely in S. UK.

ECMWF's deterministic model is highly interested in pelting NW Europe with another succession of lows not just next week but into the week after as well. Resulting rainfall totals are frankly gargantuan. GFS continues to explore some drier options around the middle of the month.

These are my local area (NE Dorset / W Hampshire) monthly mean temperature anomalies versus a 1991-2020 average. See how well balanced they are during 2000-2015, with similar numbers of anomalously warm & cold months. Since 2016, anomalous warmth has instead been very dominant.

There was some drama involved with summer 2025 securing the highest Hadley mean CET on record. For almost all of August, GFS overestimated temps, but during the final few days the opposite occurred. Had reality instead been at or below GFS, 1976 would have held onto the record.

Local to me, despite August not being as emphatically above average, the mean daily max for the past 6 months has been well clear of the prior record set in 2022, which itself was well clear of the next warmest year (2018). Unprecedented in at least the past 82 years.

Here we see the true spatial scale of the developing La Niña event in the Pacific Ocean. Contrasting with the anomalous warmth further north & south, it makes for a striking visual. pulse.climate.copernicus.eu

Having an enjoyable breather outside watching towering cumulus roll past while soaking up some warm sunshine, but I think it’s time I headed in 🌦️

Meteorologically speaking, summer 2025 has concluded. The mean daily max temp here was on par with 2018, 2022, 1995, & 1976. Interesting how those are tightly clustered, yet well distanced from the rest. The mean daily min was 2nd warmest on record, well above 2022, 1976, & 1995.

The 00z ECMWF deterministic run basically chose all of the most unsettled options for the UK & Ireland, resulting in 60-80 mm widely, & over 100 mm in many southern & western parts, by the midpoint of September. GFS' 00z was a bit more balanced but still pretty soggy for many.

At face value, the anomalous warmth between 30-70 N in the Pacific Ocean seems to far outweigh the anomalously cool waters along the equatorial Pacific. However, this map's projection hides the fact that Earth's circumference is largest at the equator...

Without weather forecasting, I could have been fooled into setting off cycling after a shower cleared at 10:10 with seemingly no further ones inbound. As it is, I know a conducive atmosphere is predicted until ~mid-afternoon here, so I stayed indoors, which has proved very wise!

Usually unremarkable but such a rare sight here these past six months 🌦️