Since no one wants to, I'll take it upon myself to create a post about the upcoming possible G1 storm on October 7th.

So, we have two weak CMEs, estimated arrival time is noon on October 7th.

3 minutes ago, Samrau said:

Since no one wants to, I'll take it upon myself to create a post about the upcoming possible G1 storm on October 7th.

I enjoy the fact that Jester seems to anticipate the evolutions somehow. I have zero idea what markers in data he uses or whether “ mysticmegan” is his better half Tnx fer stepping up dude. 😊

The current solar wind speed is about 450 km/s and is gradually decreasing. I'd bet on G2 and isolated G3 on October 8th if the solar wind speed drops to 300-350 km/s by then.

2 minutes ago, hamateur 1953 said:

I enjoy the fact that Jester seems to anticipate the evolutions somehow. I have zero idea what markers in data he uses or whether “ mysticmegan” is his better half Tnx fer stepping up dude. 😊

Are you saying that we don't have to wait for anything?

2 minutes ago, Samrau said:

The current solar wind speed is about 450 km/s and is gradually decreasing. I'd bet on G2 and isolated G3 on October 8th if the solar wind speed drops to 300-350 km/s by then.

Are you saying that we don't have to wait for anything?

Nup. What I meant was him and Jessica and probably all of y’all anticipate substorms better than I ever could imagine doing. Also enjoy his eyeballing the magnetometers after the shock hits. And his patience. Having trouble with my reaction rail Cat wants brekkie edit: fed the cat Let’s hope that the silly CH compacts and reorganises the combined transiting plasma instead of scattering it everywhere this time Need a G3 here 😊 especially with the shiny thing being full

Edited Sunday at 04:40 PM2 days by hamateur 1953
🐈‍⬛ interference

Model HUXt

Guys, what is the standard deviation or percentile intervals for the predicted impact? The entire US northeast is going to be overcast on the 7th :( This is ironic since we have had clear skies for a week by now and they will last till tomorrow, then on the 7th it will be overcast, and clear right after. This "weather line filter" is going to block out the best part of this unless the prediction is off, which is why I am asking. Agree @hamateur 1953 that the full Moon is yet another factor that could "square root" this experience. Fingers crossed.

3 hours ago, JessicaF said:

Guys, what is the standard deviation or percentile intervals for the predicted impact? The entire US northeast is going to be overcast on the 7th :( This is ironic since we have had clear skies for a week by now and they will last till tomorrow, then on the 7th it will be overcast, and clear right after. This "weather line filter" is going to block out the best part of this unless the prediction is off, which is why I am asking. Agree @hamateur 1953 that the full Moon is yet another factor that could "square root" this experience. Fingers crossed.

Sounds like my luck and @pnw and @GeomagneticGreenhorn and @MajorCatAura up here. I almost forgot @Apophenia_Savant Bummah.

3 hours ago, JessicaF said:

The entire US northeast is going to be overcast on the 7th :( This is ironic since we have had clear skies for a week by now and they will last till tomorrow, then on the 7th it will be overcast, and clear right after.

Add me to the pile! Clear every night since Wednesday last week, and clear this coming week.... EXCEPT for Tuesday night.

That's just absurd! LOL

Mercy!!

Heads-up, since we've now got SolO between us and the Sun. Noticeable impact earlier today around 11Z, presumably from the front of the slower structure(s); not all that much to speak of in terms of any variables, but will be elevated relative to regular slow wind at least:

From the MAG it also looks like there's something happening around 16Z, which can also be seen above to some extent, but hard to say if it's a separate structure behind it or part of the same:

There was also what appears to have been an impact yesterday at around 19Z, presumably the earlier (and also relatively weak by the looks of it) CME in front, i.e. the one from the M1.5 on Oct 3:

SolO isn't exactly in line with the Sun-Earth line, so there could of course be differences in impact here, but overall it's so far not looking like we'll see much over 10 nT of Bt based on SolO's MAG; but again, might be different here.

2 hours ago, JessicaF said:

Guys, what is the standard deviation or percentile intervals for the predicted impact?

I'm not actually sure what the CI is for the ranges given by HUXt or listed by CCMC, but the latter ones sure look a lot less stringent, and are from different models too.

The HUXt one looks more like what I'd expect for a standard 95% CI, but someone should definitely inquire about it (or find information about it, I couldn't). If so then roughly speaking there'd be a 95% chance of the CME arriving within that time (granted that it hits at all, but could potentially miss too, naturally). I don't really know if that actually is the case, though; but if we assume it for now, then it gives a 95% chance of the first CME arriving within approx. 10-06T12Z and 10-07T18Z, and the second one within approx. 10-07T13Z and 10-09T12Z.

There may be parts of these structures that aren't fully modeled though, as it's a bit complex of an event; I've e.g. seen EUHFORIA runs posted that have the brunt of the slower blowout structure coming in even later than both of those hits, not until Oct 10 or so. Will be interesting to see this slow train come through.

Earthbound magnetometers not showing anything yet. Need to wake up Jester. His profile shows him at L1 from memory. 😊. Oops. Wrong day. Its still Sunday. 🤣🤣Hagrids tail is up.  Good chance for G2

Edited Monday at 02:17 AM2 days by hamateur 1953
Checked with Hagrid.

2 hours ago, hamateur 1953 said:

Sounds like my luck and @pnw and @GeomagneticGreenhorn and @MajorCatAura up here. I almost forgot @Apophenia_Savant Bummah.

The earlier posts from, I think, 'Unspecified Geomagnetic Activity' suggested an October 8th impact. I will see if I can find and link them...

Its cool @Stella my error. Spaced a day. We are still Sunday in USA.🤣🤣being retired one day much like another!!

2 minutes ago, Stella said:

The earlier posts from, I think, 'Unspecified Geomagnetic Activity' suggested an October 8th impact. I will see if I can find and link them...

Edited Monday at 02:11 AM2 days by hamateur 1953

38 minutes ago, Philalethes said:

95% chance of the first CME arriving within approx. 10-06T12Z and 10-07T18Z, and the second one within approx. 10-07T13Z and 10-09T12Z.

These are very wide CIs. What are the likelihoods within these intervals? Is it Gaussian or something skewed? I have no idea what kind of distribution the arrival would follow. Has anyone ever done a larger scale analysis of CME arrival prediction errors and their distributions?

Thanks for the analysis. You are amazing.

5 hours ago, Philalethes said:

Heads-up, since we've now got SolO between us and the Sun. Noticeable impact earlier today around 11Z, presumably from the front of the slower structure(s); not all that much to speak of in terms of any variables, but will be elevated relative to regular slow wind at least:

From the MAG it also looks like there's something happening around 16Z, which can also be seen above to some extent, but hard to say if it's a separate structure behind it or part of the same:

There was also what appears to have been an impact yesterday at around 19Z, presumably the earlier (and also relatively weak by the looks of it) CME in front, i.e. the one from the M1.5 on Oct 3:

SolO isn't exactly in line with the Sun-Earth line, so there could of course be differences in impact here, but overall it's so far not looking like we'll see much over 10 nT of Bt based on SolO's MAG; but again, might be different here.

I'm not actually sure what the CI is for the ranges given by HUXt or listed by CCMC, but the latter ones sure look a lot less stringent, and are from different models too.

The HUXt one looks more like what I'd expect for a standard 95% CI, but someone should definitely inquire about it (or find information about it, I couldn't). If so then roughly speaking there'd be a 95% chance of the CME arriving within that time (granted that it hits at all, but could potentially miss too, naturally). I don't really know if that actually is the case, though; but if we assume it for now, then it gives a 95% chance of the first CME arriving within approx. 10-06T12Z and 10-07T18Z, and the second one within approx. 10-07T13Z and 10-09T12Z.

There may be parts of these structures that aren't fully modeled though, as it's a bit complex of an event; I've e.g. seen EUHFORIA runs posted that have the brunt of the slower blowout structure coming in even later than both of those hits, not until Oct 10 or so. Will be interesting to see this slow train come through.

I roughly modeled the angle

As @Philalethes noted, in reality, everything may turn out differently than what is shown in the SoLo data - Earth is in a more advantageous position.

The SWPC timing model is quite accurate, given that SoLO is located at a distance of 0.49 AU from the Sun and 0.52 AU from Earth.

At the moment, the solar wind speed has dropped below 400, so I think everything is going well. I still expect better.

Perhaps forecasters will recalculate based on the data received from SoLO, but I think the difference in the new model will not be noticeable.

Another one was probably just produced by 14241 👀

EPAM is slowly growing. Nothing bothers him, I think he can be informative.

My desktop is till out for cleaning, so I can only attempt to go off of memory. I think my notes had 3 days of travel time to maximum solar wind velocity of ~425km/s, 3.5 days at ~375km/s, and 4 days at ~325km/s.

7 minutes ago, Jesterface23 said:

My desktop is till out for cleaning, so I can only attempt to go off of memory. I think my notes had 3 days of travel time to maximum solar wind velocity of ~425km/s, 3.5 days at ~375km/s, and 4 days at ~325km/s.

Now the density and speed have increased

This is either the first CME or CH HSS 84?

Finally, I got the Stereo shots! Here is the first culprit - trinity 4233-4236-4232, more precisely, I have no way to determine, because I am not at home right now, but at my work computer.

9 minutes ago, Samrau said:

Now the density and speed have increased

This is either the first CME or CH HSS 84?

The IMF didn't really change at all, so I don't believe it is a shock arrival. It could just be a weak influence of old CH 84 or just a brief solar wind bump.

11 hours ago, JessicaF said:

These are very wide CIs. What are the likelihoods within these intervals? Is it Gaussian or something skewed? I have no idea what kind of distribution the arrival would follow. Has anyone ever done a larger scale analysis of CME arrival prediction errors and their distributions?

Yeah, it's an imperfect science, heh. But like I mentioned I also don't know for sure if it's a standard 95% CI or not. I'll see if I can make an inquiry about it to the good Dr. Owens and pry out some details about it.

But in any case, from looking at the stated range it seems clear that it's a right-skewed distribution; intuitively this is very likely due to the nonlinear effects of drag. Thus there should be equal probability for the arrival time to be on each side, with the time range on the late arrival side being notably longer. Here's the table in question for reference, from the HUXt site:

But yeah, should really get it fully clarified by those behind the model before making any definitive statements, this is just assuming fairly standard statistical practices of course.

And yeah, there's a lot of literature on CME arrival time predictions and their distribution and errors, so probably something that could be worth digging into at some point. This is a fairly recent paper that also references CCMC explicitly (I see at least one author whose name regularly shows up as a modeler there), so maybe it could shed some light on that too.

1 hour ago, Philalethes said:

Yeah, it's an imperfect science, heh. But like I mentioned I also don't know for sure if it's a standard 95% CI or not. I'll see if I can make an inquiry about it to the good Dr. Owens and pry out some details about it.

But in any case, from looking at the stated range it seems clear that it's a right-skewed distribution; intuitively this is very likely due to the nonlinear effects of drag. Thus there should be equal probability for the arrival time to be on each side, with the time range on the late arrival side being notably longer. Here's the table in question for reference, from the HUXt site:

But yeah, should really get it fully clarified by those behind the model before making any definitive statements, this is just assuming fairly standard statistical practices of course.

And yeah, there's a lot of literature on CME arrival time predictions and their distribution and errors, so probably something that could be worth digging into at some point. This is a fairly recent paper that also references CCMC explicitly (I see at least one author whose name regularly shows up as a modeler there), so maybe it could shed some light on that too.

Thanks. So, a late arrival outlier is more likely than an early arrival outlier, right? In other words, the right tail is thicker. This is funny because based on my own memory of the past significant events, the shockwaves came earlier (April 14, June 1, and Sep 1). This must be the proverbial "selection effect" that proves nothing and may give wrong impressions due to limited human memory and most of all a small statistical sample.

Interesting paper you linked. Figure 2 shows that there seems to be a small bias towards negative "delta t" (late arrival). It is also interesting to see the performance of individual models as well as evolution over time. I will need to read it through later.

Thanks again for the feedback and info!

16 minutes ago, JessicaF said:

So, a late arrival outlier is more likely than an early arrival outlier, right? In other words, the right tail is thicker. This is funny because based on my own memory of the past significant events, the shockwaves came earlier (April 14, June 1, and Sep 1). This must be the proverbial "selection effect" that proves nothing and may give wrong impressions due to limited human memory and most of all a small statistical sample.

In the ideal case, not quite; as the median is defined in such a way as to split even skewed distributions evenly, then ideally there is equal probability for arrival on each side of it. See e.g. this illustration:

Whether or not this actually holds true for the model is another question, but that's at least what the median time aims at. What it rather means is that late arrivals will be spread out more over a longer interval further from the median, while early arrivals will tend to cluster more closely together and not as far from the median.

(edit: I see you referred to outliers specifically; in that case yes, that would be right, I assume you have a good grasp of what this means)

And yeah, various psychological biases can for sure easily give a distorted view of things that you don't get when looking more objectively at the data; intuitively there's at least no surprise that early arrivals would tend to be more salient, especially for stronger events.

41 minutes ago, JessicaF said:

Interesting paper you linked. Figure 2 shows that there seems to be a small bias towards negative "delta t" (late arrival). It is also interesting to see the performance of individual models as well as evolution over time. I will need to read it through later.

Yeah, I just skimmed through it a bit myself. In Fig. 3 they transform the data from Fig. 2 to give an easier view of things:

Heat map timeline of the errors in arrival time for all predictions. The top panel shows the signed error and the bottom panel shows the absolute (unsigned) error. The lines show a rolling mean (blue) or median (cyan) over ±6 months.

As they mention in the corresponding section the largest bias is observed during the time of minimal activity and minimum number of predictions being made, which isn't too surprising:

Figure 3 highlights the obvious increase in the number of predictions since about 2021. We lose any information about the individual models but can better see the trends over time. The bias (signed error) remains fairly small over the full duration, typically remaining with ±10 hr. In 2018, we see a brief dip in the mean signed error toward −20 hr, but this is also a low activity period with only a few observed events. Beyond about 2022 the bias remains very low, essentially consistent with the zero error line.

The bias towards later arrivals there could also have to do with the overall correlation they found between longer transit times being associated with early predictions (and vice versa, they mention the opposite case); this essentially means that the fastest CMEs tend to arrive even earlier than predicted (so perhaps it's not all psychological bias after all, hehe), and the slowest CMEs even later than predicted, on average:

We find little evidence of any correlations between the arrival time errors and any other properties. The one noticeable exception is a tendency for late predictions for short transit times and vice versa.

I guess this should have a separate topic of its own for discussion at this point, since we've strayed a bit from that of the current arrivals. Very interesting, though.

Edited Monday at 05:13 PM1 day by Philalethes
outliers

Is the first weak CME here ?

1 hour ago, Jay said:

Is the first weak CME here ?

Was wondering that also. Looking at Canmos. Hard to say. But looks as though something arrived recently. Edit: looking at Hemispheric power it’s almost at 50 gw And appears to be climbing. Wait and see. Density still low and to be expected. Bz -3. At least that much is good news😊. First high latitude heads up on Fairbanks AK. 😊

Edited yesterday at 08:01 AM1 day by hamateur 1953
Checked mags etc.