2017 Global Model Runs Discussion (Out to day 16)

[Hurricaneman]

The 0zGFS is showing something trying to form around 228hrs and is close to the Euro in terms of lowering pressures in the western Caribbean around the same timeframe which leads me to believe development around the first of next month

[LarryWx]

Like clockwork, the 0Z GEFS has much less (though not none) W Caribbean TC action in the means vs the 18Z GEFS. This was almost 100% predictable based on what we learned this evening about the 18Z. For the same reason, I predict that the next very active GEFS will be the next 18Z. Let's see what it will show.

[toad strangler]

Like clockwork, the 0Z GEFS has much less (though not none) W Caribbean TC action in the means vs the 18Z GEFS. This was almost 100% predictable based on what we learned this evening about the 18Z. For the same reason, I predict that the next very active GEFS will be the next 18Z. Let's see what it will show.

You must have a different 0z GFS run than I do.

[LarryWx]

Like clockwork, the 0Z GEFS has much less (though not none) W Caribbean TC action in the means vs the 18Z GEFS. This was almost 100% predictable based on what we learned this evening about the 18Z. For the same reason, I predict that the next very active GEFS will be the next 18Z. Let's see what it will show.

You must have a different 0z GFS run than I do.

0Z GEFS mean, not 0Z GFS. Check out Tropical Tidbits, compare the 0Z GEFS mean vs 18Z GEFS mean for the first 9 days of October overall in and near the western Caribbean and you'll clearly see what I'm referring to: stronger activity on 18Z vs 0Z.

[otowntiger]

Like clockwork, the 0Z GEFS has much less (though not none) W Caribbean TC action in the means vs the 18Z GEFS. This was almost 100% predictable based on what we learned this evening about the 18Z. For the same reason, I predict that the next very active GEFS will be the next 18Z. Let's see what it will show.

You must have a different 0z GFS run than I do.

0Z GEFS mean, not 0Z GFS. Check out Tropical Tidbits, compare the 0Z GEFS mean vs 18Z GEFS mean for the first 9 days of October overall in and near the western Caribbean and you'll clearly see what I'm referring to: stronger activity on 18Z vs 0Z.

what'd the 0z Euro show? Anything?

[DESTRUCTION5]

You must have a different 0z GFS run than I do.

0Z GEFS mean, not 0Z GFS. Check out Tropical Tidbits, compare the 0Z GEFS mean vs 18Z GEFS mean for the first 9 days of October overall in and near the western Caribbean and you'll clearly see what I'm referring to: stronger activity on 18Z vs 0Z.

what'd the 0z Euro show? Anything?

EURO showing lowering pressures starting around the 200 hr marker and has the last several runs. I expect the coming runs to start showing a closed low in the 7-10 day range.

[Hurricaneman]

Day 10 shows a low on the Euro

[LarryWx]

Like clockwork, the 0Z GEFS has much less (though not none) W Caribbean TC action in the means vs the 18Z GEFS. This was almost 100% predictable based on what we learned this evening about the 18Z. For the same reason, I predict that the next very active GEFS will be the next 18Z. Let's see what it will show.

The 6Z GEFS mean, also like clockwork, shows very few sub 1,000 mb TCs in/near the W Caribbean in early Oct. The quiet contrast to last evening's (and the last week or so of 18Z GEFS runs) is almost laughable. I'm still looking for another active 18Z GEFS run this evening as silly as that sounds. I wish there were a way this could be fixed. I've seen a "we toss" phrase associated with the 18Z. I used to think of that as silly. Now I know it actually makes sense.

Meanwhile, the 0Z EPS has about 30% of its members with sub-1,000 mb TCs form in/near the W Caribbean during early Oct., a respectable signal, an increase over the 20% of yesterday's 12Z EPS, way more active than the 6Z GEFS run, and more active than the 0Z GEFS run on a % basis. However, it is nowhere near as active %wise as the 18Z GEFS run, of course.

[MJGarrison]

I'm a bit skeptical that there's a known initialization bias in the 18Z gfs that's unaccounted for. Any solar noon bias would happen all over the world and could be adjusted prior to use. The sensors likely have operating range and when the sensor is out of range the data can be adjusted or discarded.

Also, bad sensor data can manifest itself in many ways. It could be that fewer TCs indicate bad data. How do we know it's not the 0z 12z and 6z that are out of range?

Can anyone comment on who uses what's data? It's been implied in other forums that the recon data makes it into at least GFS and Euro. I'd be surprised if they both didn't use the best data available recon or otherwise. Common data source = everyone might be wrong.

[LarryWx]

I'm a bit skeptical that there's a known initialization bias in the 18Z gfs that's unaccounted for. Any solar noon bias would happen all over the world and could be adjusted prior to use. The sensors likely have operating range and when the sensor is out of range the data can be adjusted or discarded.

Also, bad sensor data can manifest itself in many ways. It could be that fewer TCs indicate bad data. How do we know it's not the 0z 12z and 6z that are out of range?

Can anyone comment on who uses what's data? It's been implied in other forums that the recon data makes it into at least GFS and Euro. I'd be surprised if they both didn't use the best data available recon or otherwise. Common data source = everyone might be wrong.

Good post. We don't know that the 12Z, 0Z, and 6Z aren't underdone, especially the 6Z. I could see a case made for the 18Z being overdone, the 6Z being underdone, and the 0Z/12Z being the least biased. I hope RL3AO responds.

I had a somewhat related post last night: "Taking the above further, would the 6Z tend to underdo TC genesis chances due to it being in the middle of the night/zero solar insolation? Also, would the 12Z runs tend to show less activity than the 0Z runs due to it being a cooler time of day than 0Z in the western basin even though solar insolation is similar? Finally, do other models show these same biases?"

[DESTRUCTION5]

Day 10 shows a low on the Euro

Your right...missed the 240hr frame.

[LarryWx]

^ I just looked at the Tropical Tidbits GEFS runs again for 18Z on 10/2 and sure enough the 6Z GEFS was always the quietest run!! So, the 18Z GEFS was by far the most active, the 6Z was by far the least active, and the 0Z/12Z were in between every single day for the week's worth of runs available at the site.

So, my hypothesis is now that the 18Z GEFS is biased too active, the 6Z GEFS is biased too inactive, and the 0Z/12Z are closest to accurate/least biased.

[TheStormExpert]

 https://twitter.com/WeatherdotUS/status/911951551870185472

[TheStormExpert]

I truly don’t think the GFS has a good handle at all on the potential Western Caribbean monsoon tropical cyclone development that the Euro has been persistently consistent with showing gradual and slow development just after the beginning of October.

[psyclone]

These systems often take their time coming together and the models are often intermittent with development in the extended range. I'm not making much of the run to run variability at this point as it's too far out. The signals are there (climo, enso state, the ferocity of the season thus far). At this juncture I'd guess odds are above even that something pops in the first third of October. We just need to keep an eye on it. As is often the case with the tropics it's a "hurry up and wait" situation. But extra vigilance is warranted since things can escalate quickly and a storm can threaten a region in short order. we won't have the luxury of watching a system for 10 days as we get with the long trackers from the east.

[LarryWx]

The 12Z GEFS is much more active than the quiet biased 6Z GEFS and even a good bit more active than the 0Z GEFS in/near the western Caribbean in early Oct. The only recent runs more active are the last few 18Z GEFS runs (no surprise there). So, I consider this a legitimately somewhat stronger signal for development. Looking forward to seeing the Happy Hour (18Z) GEFS, which I fully expect to be very active due to high activity bias and even a good bit more active than this relatively active 12Z GEFS.

[LarryWx]

Fwiw, the 12Z Euro has a weak (1006 mb), broad low in the NW Caribbean about 150 miles south of west-central Cuba moving slowly northward. It appears to then be slowly strengthening amidst low shear.
This is just something to keep in the back of your mind for potential in early October.

[TheStormExpert]

12z EPS (Euro) Ensembles fresh off the press even more bullish!

[LarryWx]

Following up to the above post: Looking at the 11-15 day 12Z EPS in/near the W Caribbean/Gulf/just off the SE US coast, it is somewhat more active than the already fairly active 0Z EPS for the 50 members as a whole with 10 (20%) potential Hs or sub 990 mb vs 9 (18%) on the 0Z EPS; and about 22/44% potential TS+ (sub 1000 mb) vs ~15 (30%) on the 0Z EPS and ~10 (20%) on yesterday's 12Z EPS.

Regarding the 10 apparent Hs, 1 hits N MX, 1 hits TX/LA line, and 2 hit E LA/MS. Regarding the other 6 as of day 15: 1 is about to hit LA, 3 are about to hit W FL, 1 is still well out in the W GOM moving slowly N, and 1 is moving NE well out in the open Atlantic.

Next up: the often over-dramatic 18Z GEFS (and 18Z GFS). I expect to see strong activity on this considering the rather significant signal on the 12Z GEFS, the pretty strong signal on the 12Z EPS, and the 18Z GEFS bias toward activity.

[LarryWx]

^Well, the 18Z GEFS didn't disappoint at all with the streak of being by far the most tropically active of the GEFS runs in the western basin in early Oct. continuing. The predictability of it is kind of funny. Practically every member becomes a TS+ and 50%+ become a H with several monster Hs. Contrast that to the typical quietest GEFS, the 6Z, which had only one H and a few TSs. But it is even much more active than the 12Z, 0Z, and all non-18Z GEFS runs of any for the last week+.

So, there will practically definitely be a big downtick from this 18Z GEFS on tonight's 0Z GEFS and likely a further downtick on the 6Z GEFS. But don't let these expected downticks fool you into thinking they mean much of anything.