Learner question: how does an eruptive flare differ from an impulsive one?

[Ester89]

Hi! I've been following the forum for a while to try to learn (I'm learning from 0 🙈) and for me it's all still confusing.  I would be grateful if someone could explain to me in a simple way (or tell me where I can find information about it) how can I differentiate between an eruptive and impulsive flare?  Does it have to do with the duration of the flare?  Thank you!

[mozy]

5 minutes ago, Ester89 said:

Hi! I've been following the forum for a while to try to learn (I'm learning from 0 🙈) and for me it's all still confusing.  I would be grateful if someone could explain to me in a simple way (or tell me where I can find information about it) how can I differentiate between an eruptive and impulsive flare?  Does it have to do with the duration of the flare?  Thank you!

Basically yes, short duration flares are considered impulsive & rarely produce significant CME's or anything at all, you literally just see it spike on the X-ray chart and go down immediately, you also only see a quick bright flash on imagery, longer durations flares are the interesting ones as they tend to release CME's, either by a slow rise & slow fall or a sharp uptick & then a slow/drawn out decline.

Edited May 3, 2023 by mozy

[Ester89]

1 hour ago, mozy said:

Básicamente sí, los destellos de corta duración se consideran impulsivos y rara vez producen CME significativos o algo en absoluto, literalmente solo lo ve aumentar en el gráfico de rayos X y baja de inmediato, también solo ve un destello brillante rápido en imágenes, más tiempo las llamaradas de duración son las interesantes, ya que tienden a liberar CME , ya sea por un aumento lento y una caída lenta o un repunte brusco y luego una disminución lenta/prolongada.

I am glad to know that my conclusions were quite correct.  Thank you so much for your explanation!! 

[Philalethes]

On 5/3/2023 at 3:08 PM, Ester89 said:

I am glad to know that my conclusions were quite correct.  Thank you so much for your explanation!! 

Another interesting thing to note is that it has to do with the total amount of energy produced by the event. What the flare levels are measured in is X-ray flux, which is essentially a measure of power per area at those specific wavelengths; since power is a measure of energy per unit of time, then for a lot of energy to be released it's necessary to have both a high flux and a long duration. This is also why some events with less power, like high-level C-class flares, can end up being more energetic than certain M-flares, or even X-flares in some rare circumstances, as long as they last long enough.

The measure of the total energy of a given flare is thus done by approximating the area under the curve, as this is essentially multiplying the flux at those various times with each respective time period. That might be a bit too complex to think about, but mathematically this is done through a method called integration, hence why you might often see references to "integrated flux", which refers precisely to that area, and thus to the total energy produced by the flare at those wavelengths. On SWPC's site for the X-ray flux, here, this value is provided for the last flare:

Here you can see the unit is J instead of W, as the joule (J) is a unit of energy, while the watt (W) is a unit of power. So if you want to reference or make notes of the total energy of a flare to compare it to others or for whatever reason, that's definitely something to keep in mind.

Here I've tried to illustrate what is meant by the area under the curve by filling it in with yellow:

So essentially the larger that area is, the more energy a flare is releasing. It's not always that easy to compare areas since the scale is logarithmic (each successive step from A to B to C to M to X represents 10 times as much flux), but the general rule is indeed what you've observed yourself: the longer the duration of a flare, the better the chance of it being eruptive.

[Ester89]

First of all, i want to thank you @Philalethes for taking the time to write this interesting information to the thread of my question.  Your answers are really enriching and valuable for me, and I hope for other people as well.  Right now a new question is coming up for me: for example, a C8 flare coming up from a B background flux will probably be more powerful than an M1 flare coming from a C8 background flux? This assuming that both flares have the same duration. 

(First M flare in several days as I write this 😀. Almost X, in fact 😀😀😀)

Edited May 16, 2023 by Ester89

[Wolf star]

An example of such an impulsive flare-up right now just beyond the eastern edge.😉

[Philalethes]

Just now, Ester89 said:

First of all, i want to thank you @Philalethes for taking the time to write this interesting information to the thread of my question.  Your answers are really enriching and valuable for me, and I hope for other people as well.  Right now a new question is coming up for me: for example, a C8 flare coming up from a B background flux will probably be more powerful than an M1 flare coming from a C8 background flux? This assuming that both flares have the same duration. 

(First M flare in several days as I write this 😀)

That's a great question. As we speak I was actually downloading the 7-day data that NOAA presents on the site so that I could show in a plot how it's easy to get misled by the logarithmic scale.

But to deal with the question with some mathematical assumptions, let's assume that both flares in your hypothetical scenario immediately spike up instantly and then back down after an equal period of time; that's obviously not how real flares work, but it's just to illustrate the differences in the fluxes.

Given this assumption, the total energy difference for each flare will be given by the difference in power between the peak and the background. Let's assume the C8 flare is coming up from a B8 background. Now, it's easy to think that the rise a whole level up from B8 to C8 would be a bigger difference, but that's where the logarithmic nature of it really starts to do tricks. See, every single level of C represents the entire B-level, and so on with every successive scale. So from C1 to C2 is actually slightly more of a rise as from B1 to C1, but almost the same; without units, it's essentially like this:

B1 = 1 * 10^(-7) = 0.0000001
C1 = 1 * 10^(-6) = 0.000001
C2 = 2 * 10^(-6) = 0.000002

So the rise from B1 to C1 in this case would be 0.0000009, while the rise from C1 to C2 would be 0.000001, which is only very slightly more.

Given this, we can also look at the values in question:

B8 = 8 * 10^(-7) = 0.0000008
C8 = 8 * 10^(-6) = 0.000008
M1 = 1 * 10^(-5) = 0.00001

To get the differences, we simply subtract:

C8 - B8 = 0.000008 - 0.0000008 = 0.0000072
M1 - C8 = 0.00001 - 0.000008 = 0.000002

With all the zeros it might not be so easy to see, but the rise from B8 to C8 is actually 3.6 times as large as the rise from C8 to M1; so turns out that for these particular assumptions, it would indeed be correct.

But what happens if it's not exactly an M1-flare, but rather an M1.6 flare? That's where the logarithmic scale can trick you, because that extra 0.6 on the M-scale is equal to 6 entire levels on the C-scale. Thus you'd get this instead:

M1.6 = 1.6 * 10^(-5) = 0.000016
M1.6 - C8 = 0.000016 - 0.000008 = 0.000008

This is already slightly larger than the difference between C8 and B8 of 0.0000072.

So in conclusion: while it's true that a C8-flare from a B8-background will tend to represent a larger increase than an M1-flare from a C8-background, it will actually be smaller increase than an M1.6-flare from a C8-background; and needless to say, the more powerful the M-flare from that point on, the larger the difference will be.

[Ester89]

4 hours ago, Philalethes said:

It's not always that easy to compare areas since the scale is logarithmic (each successive step from A to B to C to M to X represents 10 times as much flux)

It's just what you mentioned here, I wasn't sure I understood it correctly

51 minutes ago, Philalethes said:

That's a great question. As we speak I was actually downloading the 7-day data that NOAA presents on the site so that I could show in a plot how it's easy to get misled by the logarithmic scale.

But to deal with the question with some mathematical assumptions, let's assume that both flares in your hypothetical scenario immediately spike up instantly and then back down after an equal period of time; that's obviously not how real flares work, but it's just to illustrate the differences in the fluxes.

Given this assumption, the total energy difference for each flare will be given by the difference in power between the peak and the background. Let's assume the C8 flare is coming up from a B8 background. Now, it's easy to think that the rise a whole level up from B8 to C8 would be a bigger difference, but that's where the logarithmic nature of it really starts to do tricks. See, every single level of C represents the entire B-level, and so on with every successive scale. So from C1 to C2 is actually slightly more of a rise as from B1 to C1, but almost the same; without units, it's essentially like this:

B1 = 1 * 10^(-7) = 0.0000001
C1 = 1 * 10^(-6) = 0.000001
C2 = 2 * 10^(-6) = 0.000002

So the rise from B1 to C1 in this case would be 0.0000009, while the rise from C1 to C2 would be 0.000001, which is only very slightly more.

Given this, we can also look at the values in question:

B8 = 8 * 10^(-7) = 0.0000008
C8 = 8 * 10^(-6) = 0.000008
M1 = 1 * 10^(-5) = 0.00001

To get the differences, we simply subtract:

C8 - B8 = 0.000008 - 0.0000008 = 0.0000072
M1 - C8 = 0.00001 - 0.000008 = 0.000002

With all the zeros it might not be so easy to see, but the rise from B8 to C8 is actually 3.6 times as large as the rise from C8 to M1; so turns out that for these particular assumptions, it would indeed be correct.

But what happens if it's not exactly an M1-flare, but rather an M1.6 flare? That's where the logarithmic scale can trick you, because that extra 0.6 on the M-scale is equal to 6 entire levels on the C-scale. Thus you'd get this instead:

M1.6 = 1.6 * 10^(-5) = 0.000016
M1.6 - C8 = 0.000016 - 0.000008 = 0.000008

This is already slightly larger than the difference between C8 and B8 of 0.0000072.

So in conclusion: while it's true that a C8-flare from a B8-background will tend to represent a larger increase than an M1-flare from a C8-background, it will actually be smaller increase than an M1.6-flare from a C8-background; and needless to say, the more powerful the M-flare from that point on, the larger the difference will be.

Great explanation. Very grateful to you 🙏

Edited May 16, 2023 by Ester89

[Philalethes]

10 minutes ago, Ester89 said:

It's just what you mentioned here, I wasn't sure I understood it correctly

Great explanation. Very grateful to you 🙏

Yeah, it can be hard to wrap your head around.

Think of it like this: instead of writing M1.6, you could write C16, or B160; that's perhaps the easiest way to think about it. So a C8 would be a B80, and an M1 would be a C10, or a B100. That perhaps makes it more clear that it's a bigger jump from B8 to B80 (C8), which is a jump of 72 units, than from B80 to B100 (M1), which is only a jump of 20 units; but also how it's smaller than the jump from B80 (C8) to B160 (M1.6), which is a jump of 80 units.

Thinking about it like this you don't have to use all those zeros, so it should at least be a bit clearer.

Here are the plots I was making to illustrate; first one just plotting roughly the last three hours using a logarithmic scale, just like on SWPC's site:

This should look familiar, as it's essentially the same. However, if we now plot it in absolute terms to see more clearly what the relative fluxes really look like:

Here the dashed horizontal lines are at the exact same values as in the first plot, and you can see more clearly just how much larger the top portion of the flare really is in absolute terms. The minor variation at the bottom is hardly even noticeable, it looks more or less like a straight line.

[Ester89]

39 minutes ago, Philalethes said:

Yeah, it can be hard to wrap your head around.

Think of it like this: instead of writing M1.6, you could write C16, or B160; that's perhaps the easiest way to think about it. So a C8 would be a B80, and an M1 would be a C10, or a B100. That perhaps makes it more clear that it's a bigger jump from B8 to B80 (C8), which is a jump of 72 units, than from B80 to B100 (M1), which is only a jump of 20 units; but also how it's smaller than the jump from B80 (C8) to B160 (M1.6), which is a jump of 80 units.

Thinking about it like this you don't have to use all those zeros, so it should at least be a bit clearer.

Here are the plots I was making to illustrate; first one just plotting roughly the last three hours using a logarithmic scale, just like on SWPC's site:

This should look familiar, as it's essentially the same. However, if we now plot it in absolute terms to see more clearly what the relative fluxes really look like:

Here the dashed horizontal lines are at the exact same values as in the first plot, and you can see more clearly just how much larger the top portion of the flare really is in absolute terms. The minor variation at the bottom is hardly even noticeable, it looks more or less like a straight line.

I understand it and love it ❤️ 

@Philalethes Using the example of our recent M9.6 flare we could say that it would correspond to a C96 😏

Last question for today 🙈 What about flares from X10 up? Also each unit would represent the entire level from X1 to X10? I mean, would a X11 correspond to a X20 or would it just be X11?

Edited May 16, 2023 by Ester89

[Philalethes]

1 hour ago, Ester89 said:

Using the example of our recent M9.6 flare we could say that it would correspond to a C96

Yep, that's exactly right; or B960, and so on, even X0.96.

1 hour ago, Ester89 said:

What about flares from X10 up? Also each unit would represent the entire level from X1 to X10? I mean, would a X11 correspond to a X20 or would it just be X11?

That's another good question; it is the latter, i.e. just X11. If there were another class above X, e.g. Y-class flares, then an X11 would essentially be a Y1.1, and an X20 be a Y2, and so on. Since such huge flares are so rare, it's simply not common to categorize that way, so you typically just count upward from there. X11 is just 11 times the flux of X1, and X20 just 20 times the flux of X1, and so on.

There was however this interesting thread that I remember from first frequenting the forums where discussion of such classes above X took place, so perhaps it might become more common at some point; maybe if we get some big cycles with a lot of activity in the coming decades.

[Ester89]

28 minutes ago, Philalethes said:

There was however this interesting thread that I remember from first frequenting the forums where discussion of such classes above X took place, so perhaps it might become more common at some point; maybe if we get some big cycles with a lot of activity in the coming decades.

As discussed in that other thread, I'm glad there isn't a Y class. Definitely Y2 sounds tiny compared to X20 🤣