1 hour ago, Parabolic said:

I wanted to add the Oct. geomagnetic had quite a lot of negative By along side the strong southward Bz. I also included Dst and the 3hr Kp from that day.

Forgot to add the legend but Blue is By, Orange is Bz, and Green is Bt. Just realized the title for this topic and what I posted doesn't exactly pertain to the subject. Oct. 28th might be a good example of postive Bz and negative By. I'll put together something when the sun rises for me.

I think it's fine given the point about the By. I suppose one might note that any flux rope which hits us head on will primarily have variation in By and Bz due to the geometry of it; the less direct the hit, the more variation there will generally be in the Bx. So it's ultimately not that surprising that a good flux rope like here brings a solid By with it too. The sheath will of course be more variable, there I suppose one can get more lucky with some additional By.

Here I've plotted the coupling for both the October and the May storm for a comparison, using the hourly averages from OMNI (currently looking into the high-resolution data); the May storm is taken to start at 15Z and the October one at 14Z (should have added that to the legend, but too late):

As we can see, at its peak the October storm actually matched the coupling of the May one. In the legend I've added the area under the curve (AUC, thousandths of it, hence the "m") as per the trapezoid rule, whose physical interpretation would likely be some estimate of the total energy added to the system. Here I've cut the storms short at 20 hours to compare just the main body of the October storm to the same interval in the May one, and as we can see it actually beats it out in terms of AUC over that time:

It's more spread out though, which is likely part of why the May one had a significantly stronger peak earlier on too. And that is just for illustration obviously, neglecting the later peaks of the May storm wouldn't make much sense given how it had a second G5 during the peak from 20-28 hours in.

Here I've plotted what the storms would have looked like in terms of coupling had the By been zero all the time, Bz and v left unchanged:

Interestingly, but not unexpectedly, it's not very affected near the peaks where there is generally already a high ratio of Bz to By, and arguably these are the most important parts; but at the various troughs where the By is high and the Bz low we can see that it can make quite a difference, like e.g. the troughs of the May storm at 8 and 14 hours in where the coupling would have been virtually zero if it hadn't been for the By, raising it to levels that would have been sufficient to cause at least minor geomagnetic storming on its own, if not moderate. For the October storm we can see some of the parts seen in your plot above too, like e.g. the difference at 3 hours in (around 10T17Z in your plot).

Edited March 1Mar 1 by Philalethes
fewer hours for last plot, found high-resolution data

11 hours ago, Philalethes said:

I will assume you thus mean between 01:00 and 03:00 UTC on 02-27 in the data, as that seems to fit most of the parameters you describe (although looking at the RTSW it seems like the density is hovering around 2-4 rather than 10-20 during that time, so not entirely sure, feel free to clarify, if it's a different time it would of course change everything); to be clear, this is the data in question that I'm looking at, a 6-hour interval centered on the 2-hour interval you mention:

And for the substorm(s) I thus assume it was some of the smaller leftmost drops here, presumably that around 02:00:

Not a huge storm by the looks of it, but at a geomagnetic latitude of ~55° the location in question I'm guessing it would still be sufficient to produce some visible aurora there, though I probably wouldn't have guessed there would have been anything "very bright" there if you had just given me the data and nothing else. Do you perhaps have the footage to share? I'm not that well versed when it comes to photography, but I know there are ways to photograph/film aurora so it looks a lot brighter than it looks to human eyes directly, so maybe there was some of that going on too.

In this particular case it could certainly have played some role in sustaining some geomagnetic activity, although from the AL and AE it seems like the bigger storms that happened later that day didn't occur until after a foray into strongly negative Bz. But your analysis is sound, in principle a decently strong By can in fact make up for a lack of Bz, and the coupling is indeed ~0.4 of what it would be for a completely southwards clock angle. The higher speed also helps. That being said, typically you really need a stronger field, and even a faster speed too, before you start seeing solid activity that way. If you calculate the coupling for the given parameters, being generous and setting the By to 10 and the speed to 650 (and Bz to 0), the resulting coupling is 10,400, whereas more normal parameters like e.g. a By of 2, Bz of -2, and speed of 425, would yield a coupling of ~5200, so we can see that there's some enhancement, but not overly much. You'd e.g. match it with a By of 5, Bz of -5, and speed of 450 (which isn't all that bad when you think about it). For comparison the May storm of 2024 peaked at a coupling of ~73,700.

With that in mind, I'd also note that there had been some geomagnetic activity preceding this, with some prolonged periods of -Bz earlier on 02-26; it does look like it had calmed down before activity started after midnight, but it's not impossible that it could have primed the magnetosphere:

I personally wouldn't mind seeing it either, but that's something for the admins to consider I guess. Newell's paper is certainly physically sound, and the coupling shows good correlations with various magnetospheric indices, as mentioned previously in the thread. In the paper he even quips:

So at least considering coupling functions like these, especially the Newell coupling in question, is probably a good idea. One can also get some ideas about the relations from the plots and tables in the paper, like e.g. the ones comparing the coupling to AE or Kp. A rough estimate based on what it looks like to me (don't hold me to it) would be that you start seeing some minor activity with potential substorming when the coupling approaches 10,000 or so, and minor geomagnetic storm conditions around 15,000.

Also, an important note about the coupling function itself that I got confused myself previously due to a misleading statement in the abstract: the "T" in B_T likely stands for "transverse" or "tangential", because in the literature it's very explicitly used to refer to the field strength in the yz-plane exclusively, and even contrasted explicitly with just B (which is what we call Bt, but is not B_T; confusing!), the total field strength overall. In other words, you calculate the function entirely from the By and the Bz (and the speed, of course), where instead of B(t) = sqrt(Bx² + By² + Bz²) you just do B_T = sqrt(By² + Bz²). The Bx is ignored entirely.

Nope, the indices we generally look at are calculated from geomagnetic measurements (including but not limited to Kp/Ap, AE, and Dst), quite different from calculating a quantity from solar wind parameters; the idea of a coupling function is more to both be some physical representation of the reconnection process and to correlate as well as possible with those geomagnetic indices, which the paper does a lot of analysis on, showing that the Newell coupling correlates more strongly than any of the other coupling functions investigated for 9 of the 10 indices looked at (the remaining one interestingly being the Dst; section [41] is on that, with a very intriguing result).

As you take the sine of half the clock angle, the domain becomes [0°, 180°], where the sine ranges from 0 up to 1 and back down to 0 again without ever becoming negative regardless of what By and Bz are or whether they're positive or negative, avoiding any such problems! For example, in the cases where Bz is 0, the clock angle becomes 90° or 270° depending on the sign of By; in the two cases you end up with sin(45°) or sin(135°), which both yield the exact same value, 1/sqrt(2), due to symmetry. This is indeed a perfectly valid positive number that we can raise to any fractional exponent, and raising it to 8/3 we end up with ~0.4; all is well.

Thank you for your detailed response. It is much appreciated. Here are answers to your questions.

Visibility of the auroral display on 2/27 1-3UT: I do not have footage - I was watching a web cam. Based on the constellations I saw in the camera view, I estimated that the cam was registering stars of visual magnitude about 4. Since the N horizon there is very good and free of light pollution, I would assume the limit for naked eye limiting magnitude would be about 5.5-6. Based on my own experience with visual observations near the horizon and imaging, I am fairly certain that the greens were easily visible as well as the brightest red columns that sprung above during the peak of this storm (perhaps visible as white only). @Peogauuia @Jay could probably chip in as they were watching, too. As far as the size of the display, I am estimating the tops of the reds being about 15 degrees above the horizon. This angular estimate is based on the head of Draco that was visible within the field of view.

Good point about the geomagnetic latitude being much larger than 45.7 for this location. This plays a big role, indeed, when assessing whether this was a "mid latitude" event.

The data I got from SWPC shows density being 11 at 2am and 14 before 3am when the brightest display was going on. You said that the density would be a major factor if it was this high. Based on the SWPC data, it was. Here is a screen shot with the line centered at 2 am:

l

I did not realize that B_T was not Bt for that formula. Good catch. This makes the index smaller indeed. And, silly me with the sin. It can't be negative of course.

Regarding adding the index to SWL, someone could say that By is already there and thus we can take it into consideration already. However, what might be useful is a By vs. Bz graphics that would show the "clock hand" colored based on the formula perhaps in log scale. The formula does not contain density though. Some of the coupling functions do contain the density n (some via p) in Table 1. If density is a major player, shouldn't it be incorporated?

My final question is, given the positive By for this event and given the season, did the seasonal variation and perhaps the R-M effect play some role here? I need to make an effort to remember which way the Earth rotational axis tilts during the spring equinox (when viewed from x, the direction from the sun). Is it left or right?

12 hours ago, Philalethes said:

14 hours ago, JessicaF said:

This begs one question. Shall we add this quantity to the SWL board? It is computable from Bt, Bz, By, and speed. Should be easy to add. Having access to this could mean the difference between catching and not catching a display that is challenging to identify from established indices. Of course, the challenge would be how to color the values - e. g., the thresholds at which it would predict a storm.

I personally wouldn't mind seeing it either, but that's something for the admins to consider I guess. Newell's paper is certainly physically sound, and the coupling shows good correlations with various magnetospheric indices, as mentioned previously in the thread. In the paper he even quips:

After reading the main paragraph by JessicaF I had that same question, and obviously that's where she was headed with that paragraph.

Now, it would be awesome to have that quantity available on the SWL board, and we could all be "beta testing" it and making notes about how it correlates to what we're seeing (with or without a camera) in the sky. It could be another tool in the Aurora chaser's toolbox.

11 minutes ago, JessicaF said:

The data I got from SWPC shows density being 11 at 2am and 14 before 3am when the brightest display was going on. You said that the density would be a major factor if it was this high. Based on the SWPC data, it was. Here is a screen shot with the line centered at 2 am:

Hmm, I'm guessing the discrepancy is that the data in our images are from two different sources (ACE and DSCOVR respectively), since we're definitely looking at the same time. I have the series set to "Active spacecraft"; would you by any chance have it set to "DSCOVR only"? If I do that I reproduce what you're seeing perfectly, so I assume that's the case, especially since the active one at that point was ACE.

In that case it's hard to say what's really the true value, since there's such a discrepancy, but DSCOVR is known to exaggerate the number by as much as several multiples when the density is below 5/cc, and ACE tends to be a bit more accurate for the density as far as I'm aware. Hard to say for sure though.

In either case I'm not sure when I might have said that or in what context, although quite a while ago I was looking more at quantities like the dynamic pressure of the solar wind, which uses the density, but in this case it's actually a bit interesting how the Newell coupling function doesn't explicitly regard the density at all, as you can tell from the fact that it's not a parameter in it. This is likely because it gets subsumed under the other parameters, i.e. that the resulting magnetic field strength hinges on the density.

In the paper Newell does note that the Dst is the only index the function doesn't do best at, as I mentioned briefly in passing, and for that one specifically he notes that simply adding the square root of the dynamic pressure fixes that issue and makes it the best performing function there, which hints at the fact that the Dst works a bit differently than the other indices, which probably has something to do with the compression of the geomagnetic field; in that specific case the density is explicitly part of the function due to the dynamic pressure being a function of it.

25 minutes ago, JessicaF said:

Regarding adding the index to SWL, someone could say that By is already there and thus we can take it into consideration already. However, what might be useful is a By vs. Bz graphics that would show the "clock hand" colored based on the formula perhaps in log scale. The formula does not contain density though. Some of the coupling functions do contain the density n (some via p) in Table 1. If density is a major player, shouldn't it be incorporated?

Yeah, I certainly wouldn't mind it. Could perhaps be shown on this page. You could always open a request for it in the support section, or contact the admins directly, I'm guessing they don't always find time to read every single post on the forum, heh.

As for the density, hopefully what I mentioned above explains that; indeed the inclusion of p in some of the functions does rely on it, including the modification of the Newell function itself for Dst specifically. But for the general case it seems that not explicitly using it for anything, and just having it implictly as part of the magnetic field strength, seems to beat out everything; still important, of course.

30 minutes ago, JessicaF said:

My final question is, given the positive By for this event and given the season, did the seasonal variation and perhaps the R-M effect play some role here? I need to make an effort to remember which way the Earth rotational axis tilts during the spring equinox (when viewed from x, the direction from the sun). Is it left or right?

I will reply to this in the other thread (here).

Edited March 1Mar 1 by Philalethes
link to thread

20 hours ago, JessicaF said:

And another question: Do indices, such as AE or HP take into account this quantity? Are there indices that do take this into account?

Having been prompted to read a bit about how the OVATION auroral oval prediction model (the one being used by SWL) works, I'll have to update my answer to this, which was:

18 hours ago, Philalethes said:

Nope, the indices we generally look at are calculated from geomagnetic measurements (including but not limited to Kp/Ap, AE, and Dst), quite different from calculating a quantity from solar wind parameters

This was my mistake in assuming the HP (hemispheric power) was just another of these, but as it's actually part of the OVATION model, it does in fact use the Newell coupling function under the hood, because Newell was the one who developed OVATION in the first place. The HP isn't exclusively based on that, but it's probably close enough that it works well enough as a proxy for the coupling, possibly even better since the model is explicitly designed to forecast aurora. For the other indices it's still true, though.

Thus, while it's interesting to read about how the coupling function works, aurora enthusiasts would probably just do well to rely on that model for now, and use the HP as a proxy for the coupling. Turns out SWL has things under control after all!

Tagging @NightSky as well so you see this clarification.

Edited March 2Mar 2 by Philalethes
clarity

25 minutes ago, Philalethes said:

Having been prompted to read a bit about how the OVATION auroral oval prediction model (the one being used by SWL) works, I'll have to update my answer to this, which was:

This was my mistake in assuming the HP (hemispheric power) was just another of these, but as it's actually part of the OVATION model, it does in fact use the Newell coupling function under the hood, because Newell was the one who developed OVATION in the first place. The HP isn't exclusively based on that, but it's probably close enough that it works well enough as a proxy for the coupling, possibly even better since the model is explicitly designed to forecast aurora.

Thus, while it's interesting to read about how the coupling function works, aurora enthusiasts would probably just do well to rely on that model for now, and use the HP as a proxy for the coupling. Turns out SWL has things under control after all!

Tagging @NightSky as well so you see this clarification.

Except for February 26 when HPI was 25. On that day, people who relied on HPI missed a great event. The February 26 display remains a BIG mystery to me. By was > 0, so if anything it must have inhibited the seasonal and R-M effects. No clue.

Otherwise, I agree with you that HPI has the BEST performance for predicting visible aurora. It has the greatest weight for my decision making as an aurora hunter. But there are times when it does not catch a substorm, such as on 2/26. I guess there are many other factors related to a substorm that cannot be caught by DSCVR data or HPI.

Is it possible - and forgive me if I'm too dense to understand this stuff - but, can the Ovation model and HPI be like a very large ship that takes a long time to turn?

In other words, with all the data included in it, it doesn't respond to quick changes, and hence we may miss the finer details that could enable us to spot the sub storms with a "lighter weight", separate calculation that acts faster to show the changes?

Just asking, because I'm a dummy as far as fancy calculations go :)

15 minutes ago, NightSky said:

Is it possible - and forgive me if I'm too dense to understand this stuff - but, can the Ovation model and HPI be like a very large ship that takes a long time to turn?

In other words, with all the data included in it, it doesn't respond to quick changes, and hence we may miss the finer details that could enable us to spot the sub storms with a "lighter weight", separate calculation that acts faster to show the changes?

Just asking, because I'm a dummy as far as fancy calculations go :)

The ovation model uses real time solar wind data, computes it, and produces what would be an auroral oval. The lag time mostly depends on how accurate/available the data is from ACE and DSCOVER. When it's functioning as intended it's usually ahead 15-60 minutes, when there's tracking and data errors it can be behind minutes to several hours depending on the circumstances.

3 hours ago, NightSky said:

Is it possible - and forgive me if I'm too dense to understand this stuff - but, can the Ovation model and HPI be like a very large ship that takes a long time to turn?

In other words, with all the data included in it, it doesn't respond to quick changes, and hence we may miss the finer details that could enable us to spot the sub storms with a "lighter weight", separate calculation that acts faster to show the changes?

Just asking, because I'm a dummy as far as fancy calculations go :)

HPI is a wonderful measure. The best single scalar there for aurora hunters based on my own experience. Given the complexity of what we are discussing here, it cannot be completely contained in a single number. The substorm I watched on a web cam on 2/27 1-3UT was from N45.7 but here in US NE, this means geomagnetic latitude 55N, which is one of the reasons for the strong show. Second, if the magnetosphere is disturbed from previous storms, it takes less to trigger a storm. Say, HPI stays in teens and then rises to 45 for 30 minutes and goes back to teens. This will likely do nothing to mid latitudes but if, say HPI stays around 45 for 5 straight hours, then you may have a substorm at mid latitudes. HPI does take into account the past history but not this far back. Lastly, there is a mechanism for triggering a storm called Newell coupling numerically characterized by the formula discussed above. A strong By with high enough speed can trigger a storm even when Bz fluctuates around zero. I hope this helps.

Ten degrees of latitude?? Wow. Impressive

15 hours ago, hamateur 1953 said:

Ten degrees of latitude?? Wow. Impressive

Here is the May 10 outbreak as seen from above. You can see the brightest part crossing Sweden (high 50sN) while in the US NE we only needed mid 40sN. It is 10 degrees that we gain here in the NE, indeed. Dang, looking at this makes me feel worse and worse with time (I missed this Greatest Show on Earth).

I have been very sick in the flu and have mostly spent my time in bed in front of Youtube.

I have researched this a bunch the last month and I have a lot I want to add to this discussion about lobe reconnection and what it is. Magnetic reconnection in the lobe regions is often forgotten since the reconnection on the day-side is what's way more common and what's behind the big geomagnetic storms. Lobe reconnection however can give some nice aurora substorms with unusual auroral structures and shapes.

I will write some more during the weekend (if my brain is up for it) but in the meantime you can google "lobe reconnection east west imf" or something similar if you want to read papers on the subject.

On 3/2/2025 at 1:40 PM, JessicaF said:

Here is the May 10 outbreak as seen from above. You can see the brightest part crossing Sweden (high 50sN) while in the US NE we only needed mid 40sN. It is 10 degrees that we gain here in the NE, indeed. Dang, looking at this makes me feel worse and worse with time (I missed this Greatest Show on Earth).

Very cool overlay there Jessica!

I missed this somehow earlier btw

I agree, I like these overlays a lot!

This is called Theta Aurora where there is a auroral line in the middle of the auroral oval. This is a rare pattern that can happen as a consequence of lobe reconnection. Very cool!

Here's where I got the image from and they also write a bit about theta aurora: https://phys.org/news/2014-12-theta-auroralong-standing-space-mysteryrevealed.html

On 3/21/2025 at 9:30 PM, arjemma said:

I have been very sick in the flu and have mostly spent my time in bed in front of Youtube.

I have researched this a bunch the last month and I have a lot I want to add to this discussion about lobe reconnection and what it is. Magnetic reconnection in the lobe regions is often forgotten since the reconnection on the day-side is what's way more common and what's behind the big geomagnetic storms. Lobe reconnection however can give some nice aurora substorms with unusual auroral structures and shapes.

I will write some more during the weekend (if my brain is up for it) but in the meantime you can google "lobe reconnection east west imf" or something similar if you want to read papers on the subject.

Hi @arjemma ! Just wondering as I've reached the end of this thread, after @Samrau linked to it elsewhere, whether you have more gold nuggets to add(or if more information has been posted elsewhere on the forums)?

I'd love some more of this supposed 'pseudo' science you're offering, and I'm guessing I'm not alone in this. Always open to gaining a better understanding of this complex system in order to better predict its reaction to changing parameters.

PS: Did the proposed added graphic mentioned by @JessicaF ever "reach your desk"? @Vancanneyt Sander I'm not going to even pretend to understand half of what has been talked about here, but on the surface it looks like another tool to make aurora hunting more approachable and predictive for non-geeks.

1 hour ago, Rudolph said:

Hi @arjemma ! Just wondering as I've reached the end of this thread, after @Samrau linked to it elsewhere, whether you have more gold nuggets to add(or if more information has been posted elsewhere on the forums)?

I'd love some more of this supposed 'pseudo' science you're offering, and I'm guessing I'm not alone in this. Always open to gaining a better understanding of this complex system in order to better predict its reaction to changing parameters.

PS: Did the proposed added graphic mentioned by @JessicaF ever "reach your desk"? @Vancanneyt Sander I'm not going to even pretend to understand half of what has been talked about here, but on the surface it looks like another tool to make aurora hunting more approachable and predictive for non-geeks.

I agree with ya Rudolf! I grok the Newell coupling concepts but the interrelationships of Bt by bz etc my eyes glaze over for the most part! Sander explained how a high wind speed might allow a positive Bz to be geomagnetically effective once awhile back but that’s the present limit of my understanding tbh.

I was actually reading about that :

When the interplanetary magnetic field is northward, although the low-latitude magnetic reconnection on the dayside rarely occurs, it can occur at the high latitude of the magnetosphere and the tail lobe regions of the 2 hemispheres, thus allowing the solar wind and magnetosheath particles to enter into the magnetosphere. There are other possible mechanisms for the solar wind to enter into Earth’s magnetosphere when the interplanetary magnetic field is northward, such as Kelvin–Helmholtz instability, dynamic Alfvén wave resonance, pulse injection, and global gradient drift.

I can see what high-latitude / tail lobe reconnections are and also KH instabilities, but about the other ones, I don't know much and would be happy to learn more. As far as I have understood the high-latitude reconnection and the diffusion facilitated by KH instabilities would be the main mechanisms.

Anyway, interesting to see that there are various possible interactions allowing solar wind entry under northward IMF conditions.

Aurora at this moment from St. Petersburg with positive Bz.

On 9/22/2025 at 11:16 PM, Jhon Henry Osorio Orozco said:

Aurora at this moment from St. Petersburg with positive Bz.

It is from two days ago, so not exactly at this moment, though I am surprised they were so visible with such poor IMF conditions. Also, look a bit weird for auroras, from what I usually see in ASIs.