October 2026 general discussion

[weatherwiz]

3 minutes ago, CoastalWx said:

I honestly don’t even look at the GFS anymore.

I kind of have too...or at least don't put much stock into it.

[SouthCoastMA]

next two weeks look mild, overall. 

[MJOatleast7]

10 hours ago, tunafish said:

Aim higher.  We are past that and onto drying out entire springs!

So ARE Maine wells drying up or IS Wells, Maine drying up? Or both?

[tamarack]

17 hours ago, dendrite said:

Depends on snowpack. I rarely get much frozen ground here. Mostly areas where I clear snow. 

I maintain a cleared trail from driveway to the toolshed, about 50 feet, and when the snow goes away in April that path is 4-6" higher than the surrounding lawn.  Takes a couple months for that elevated track to (mostly) settle back.

Coldest October morning here is 11, on the 31st in 2020.  In Fort Kent we reached 7 on the 30th in 1978.  That year had our coldest readings in August (28), October and November (-12)

[Typhoon Tip]

13 hours ago, FXWX said:

Depressing to see most, if not all, of the NWS hourly surface data maps are no longer plotting Canadian data sites... I frequently use the GIS site and it is no longer showing any real time Canadian metar data. 

I'm sensing an increasing privatization of general weather mappage is becoming more and more problematic to general usership. 

Pay-up or no forecast graphics for you.   The NWS doesn't really provide graphics worth a shit ... I mean, we don't need or really ask for all the bells and whistles, either. Hell, I'd take 2-meter temp progs in black and white numerology, and the old black and white difax charts.  Give me those and telecon and I'd smoke the industry

The FOUS/NAM is going away ... I don't know of any source that provides raw modeled grid/altitude numbers.  They may be out there but the web these days is too cryptic and hard to find anything.   Us "meteorologists" prefer grids some times.  These strings of numbers like this random ex for the mid atlantic, 

TTPTTR1R2R3     VVVLI   PSDDFF  HHT1T3T5
DFW//969272     03902   121110	65151409
06089979167	06500   121803	64161408
12020936119	-0602	113411	60151207
18000925106	-3605	183119	50110704
24000906705     -4806	193116	47090600
30000622206	-4114	223213	52100603
36000501320     -2316	193009	59120806

create a mental picture that to the educated is often way more correct than some X'er pimped out chart, or PIVOTAL this or Weatherbell that... anyway. 

But society can't seem to function at all without passing their next breath through the filter of capitalistic ambition and money aspirations... 

Please, bring humanity a corrective comet!  

It's like the only working "comet" we have, is CC, but it's only making it worse because storms and event ferocity are now a fad, so graphical enterprises are getting greed wet dreams. Combining that with the collapse of government .. it's all creating an exclusionary kind of access.  I think it's a matter of time before PIVOTAL, and Tropical Tiddly winks or Golden Gate ... start asking for cash for any of it.

I can't wait until I bring my storm generator on line and turn on the lightning bolt targeting system. ha

[dryslot]

28 minutes ago, tamarack said:

I maintain a cleared trail from driveway to the toolshed, about 50 feet, and when the snow goes away in April that path is 4-6" higher than the surrounding lawn.  Takes a couple months for that elevated track to (mostly) settle back.

Coldest October morning here is 11, on the 31st in 2020.  In Fort Kent we reached 7 on the 30th in 1978.  That year had our coldest readings in August (28), October and November (-12)

Same here Tom, Wherever i clear the snow on 2 paths to the shed, Ground is a good 6" plus higher with the frost getting down quite deep, And i have a lot of clay so its even worse. 

[dryslot]

Just came back from a walk and its quite raw out there with the temps in the mid 40's and winds from the NW gusting in the 20's.

[kdxken]

Another beauty. The beat goes on. 

 

[HIPPYVALLEY]

1 hour ago, SouthCoastMA said:

next two weeks look mild, overall. 

Warm days and fake cold nights. It will be delightful if it stays dry.

[WinterWolf]

Exactly what you want in late October.  Today isn’t warm though…but it feels like a good solid autumn day.  Mid 50’s with a cooling breeze…perfect. 

[WxWatcher007]

First big AI fail it seems with Champi hitting Taiwan.

[kdxken]

45 minutes ago, WxWatcher007 said:

First big AI fail it seems with Champi hitting Taiwan.

 

[WinterWolf]

47 minutes ago, WxWatcher007 said:

First big AI fail it seems with Champi hitting Taiwan.

Now will they self learn from this? If so…that necessarily isn’t a bad thing that they failed. 

[CT Valley Snowman]

It's great out there no doubt, perfect for outdoor activities.  Also a snooze fest in terms of any real excitement.  

 

[dendrite]

5 minutes ago, kdxken said:

 

Chinese algorithms trying to keep the Taiwanese offguard.

[WinterWolf]

Has it weakened at all from its super typhoon status? 

[WinterWolf]

19 minutes ago, CT Valley Snowman said:

It's great out there no doubt, perfect for outdoor activities.  Also a snooze fest in terms of any real excitement.  

 

Ya, but other than a tropical system getting up here this time of year….or a bonafide winter weather event(which are both rare in October), I’ll take this great autumn weather over a breezy rainy day. 

[weatherwiz]

33 minutes ago, WinterWolf said:

Now will they self learn from this? If so…that necessarily isn’t a bad thing that they failed. 

ehh I don't think that's how that really works. The overall scheme is much more complex than just "learning" from mistakes and misses. 

While its become apparent AI models are going to provide some tremendous value, at least in medium range forecasting, if we truly want to make a substantial leap with modeling we need two things

1) Superior computing (such as quantum computing) that is able to perform complex calculations at a much faster rate. Another key aspect about quantum computing is they will be able to handle and perform even calculations that are way more complex than what supercomputers can currently handle 

2) The ability to initialize current state better (the Euro has been worlds ahead in this) and part of this goes into quantum computing and ability to handle and perform challenging complex equations. 

3) Another challenge is putting knowledge, or our understanding on physical processes into mathematical terms to be computed. There is always going to be some sort of disconnect because we don't have a true 100% knowledge of every process which occurs during storm evolution (example: why some supercells with hooks produce tornadoes while others don't) so we can't 100% model that accurately. And models just can't learn that without us being able to feed it that information first. 

[SnoSki14]

The widespread warmth across the US is insane. Just shocking 

[weatherwiz]

I can dig 30s at night and then 50s and 60s during the day...preferably 60s but keep this going through November.