Weimer vs. LFM FAC Line Plots

Weimer vs. LFM FAC Line Plots 

Investigating the Hiss Occurrence Rate as a function of IMF Clock Angle

Investigating the Hiss Occurrence Rate as a function of IMF Clock Angle:

We would like to explain a number of features of the hiss occurrence rate noticeable from the figures above. These six key features are listed as follows:

1. Prenoon hiss corresponds with negative IMF By. It is most apparent in the small to large negative Bz cases as seen from the IMF Clock Angle plot. This observation is also supported by the superposed epoch analysis.

2. Postnoon hiss is, instead, favored by positive IMF By and this is most apparent for small IMF Bz cases. 

3. There are double peaks in the hiss occurrence rate on the nightside of the Earth as seen from the IMF Clock Angle figure (i.e. postnoon hiss is apparently separated from the premidnight hiss by a small minima at dusk)

4. A relatively wider Null at noon separates the prenoon hiss from the postnoon hiss 

5. Premidnight hiss is associated with negative IMF By and also has a high correlation with the negative IMF Bz. 

6. There are very few occurrences of the postmidnight hiss but the superposed epoch analysis show that these events are favored by positive IMF By.

Weimer05FAC, Wing10FAC & Null at Noon

FAC and Auroral Hiss Occurrence Rate:

Above figure shows the integrated, around SP geomagnetic latitude, Ionospheric FAC, taken from the Weimer05 model, for +90 and -90 deg. IMF Clock Angle, showing IMF By dependence, from morning to midnight magnetic local time. It also shows, in red, auroral hiss occurrence  for the said IMF Clock angles.

Notes:
- As IMF By goes from -ve to +ve, we see that FAC in the post-noon sector become more negative / upward FAC increase, and we see a corresponding increase in the auroral hiss occurrence rate.
- We cannot explain from this figure the correlation of pre-midnight and pre-noon hiss on IMF By +ve. This is because upward currents increase in these two time sectors for IMF By -ve and Not for IMF By +ve.

Another important fact to note here about the Weimer 05 model is that, it shows dominating downward currents in the pre-noon sector for IMF By -ve at the South Pole latitude. This is in contrast to what Yan et al. (2012) point out from Wing et al. (2010) that at the South Pole latitude, we expect upward R2 FAC to dominate under the IMF By -ve conditions, as shown below:

First figure is for IMF By>0, which should logically correspond to IMF By<0 in the Southern hemisphere as Doyle et al. (1981) and others suggest about the conjugate FAC patterns.

Null at Noon in the Hiss Occurrence Rate:

I searched up that a list of authors (as discussed below) have shown evidence of and have tried to explain this gap in the occurrence of auroral phenomenon in the Southern Hemisphere:
1. Yan et al. (2012) say that "During a few hours around noon, the station is in the so-called throat
ow from dayside into the polar cap with little energetic electron precipitation."
2. Dandekar et al. (1978) say that "A  gap in the midday sector of the auroral oval is examined on the basis of Defense Meteorological Satellite Program (DMSP)  satellite auroral photographs"
3. Iijima and Potemra (1976) report a "confusion in the direction of the large-scale field-aligned currents
in  the  polar  cusp region around  1130-1230 MLT  during weakly disturbed conditions"
4. McDiarmid et al. (1976) note that electron spectra are harder in the postnoon (1400-1600  MLT)  sector than  in  the prenoon (1000-1200 MLT)  sector
5. Zaitzeva and Pudovkin (1976) show an existence of a midday auroral gap and hence a unique localized region of particle and field phenomena.

Weimer05FAC_Radial-LinePlots_IMFClockAngle

Below are the conjugate ionospheric field aligned currents (FAC) maps for -90 deg. and 90 deg. IMF Clock Angle respectively. 

From these maps, we are particularly interested in explaining the increase in post-noon/pre-dusk SP auroral hiss occurrence rate for increasing By. -90deg. IMF Clock Angle plots (top one) corresponds to the -ve By and the bottom one, for +90deg IMF Clock Angle, corresponds to +ve By.  Also note, that blue color represent the FAC coming out of the ionosphere (i.e. upwards and have -ve sign).

-90deg. IMF Clock Angle:

+90deg. IMF Clock Angle:

Note specifically that the FAC maps for the Southern Hemisphere, which show that for -ve By, upward FAC around SP latitude range increase as we move from Noon towards Dusk and then towards Midnight in MLT. 

But of particular interest to us are the comparative magnitudes of these upwards FAC in the pre-dusk sector for -ve and +ve By conditions. For this reason we draw line plots of integrated FAC against MLT (Noon to Midnight), with FAC integrated over +/- 3deg. or 4 deg. around SP geomagnetic latitude of -74 deg.

Below are these Integrated FAC line plots:

-90deg. IMF Clock Angle:

+90deg. IMF Clock Angle:

This shows that the upward FAC do increase in the post-noon/pre-dusk sector for an increase in By (as the FAC are more negative). This might indicate that the downward electron precipitation increases with increased By in the pre-dusk sector hence resulting in a higher pre-dusk hiss occurrence rate associated with +ve By.

 We also investigate this for different IMF Clock Angles, especially when the Bz is also -ve. This is shown below for IMF Clock Angles -135deg (Bz -ve and By -ve) and +135deg (Bz+ve and By -ve).

-135deg. IMF Clock Angle:

+135deg. IMF Clock Angle:

And Integrated FAC line plots are as follows:

 As can be readily seen upward FAC increase for post-noon sector as By goes from -ve to +ve but there is not much of an increase in upward FAC for times right before the dusk. This also correlates well with our observations from the Hiss occurrence rate plots against the IMF Clock Angle. We saw that pre-dusk hiss was more correlated with increasing By when Bz was kept small. Otherwise pre-dusk hiss doesn't show large enough occurrence rate.

IMF Clock Angle, Super Posed Epoch, Onset Analysis & More

IMF Clock Angle Auroral Hiss Radial Spectrograms:

Following up on our first attempt to visualize the variation in hiss occurrence rate as a function of IMF Clock Angle (variation in IMF By and IMF Bz), I plotted a second set of plots to better see certain features of hiss dependence on IMF. Below is a list of changes that I've made in the data analysis technique and dependencies that I've highlighted in each plot:

Zero Level for IMF By and Bz:
1. I've adjusted the 'zero level' for both By and Bz to ensure large (as much as possible) number of hiss samples going into the plotting of the radial spectrograms.
-> Since By and Bz are hardly ever exactly 0, so I had to set a bound around zero, s.t. if the absolute value of By/Bz for a hiss interval is below this bound, then term it as a "hiss for 0 (very small) By.
-> 'Zero level' for each of By and Bz was set, in most cases, to the half of the mean value of By and Bz respectively over the entire period of hiss data acquisition. This value was empirically chosen to ensure large number of hiss samples at each clock angle.

Averaging IMF By / Bz over the Past Half Hour from the Time of Hiss Occurrence:
2. I've plotted two plots for each category described below, one where I've used IMF By / Bz values representing a hiss event as the exact IMF By / Bz values found at the Time of Hiss Occurrence, and the other plot, where I've used the mean value of IMF By / Bz for past half hour from the time of hiss occurrence to represent the value of IMF By / Bz for a particular hiss event.

Total Hiss Occurrence Rate Plotted at the Center:
3. Total Hiss Occurrence Rate or Total Cumulative Hiss Spectrogram have been plotted at the center of the 3x3 matrix IMF Clock angle plots for comparison purposes.

Variation in By only, Variation in Bz only, and Variation in both By and Bz:
4. I have modified IMF clock angle radial spectrogram plots to depict the 'Variation in Hiss Occurrence Rate as a function of IMF By only, as a function of IMF Bz only, and finally Variation in Hiss Occurrence rate and Cumulative Spectrogram as a function of Both By and Bz.

Other Changes:
5. Colorbar has been kept uniform for all plots and radial frequency dependence has been corrected to show log. dependence.

Plots:
1. Dependence of Hiss Occurrence Rate on IMF By

Above plot makes prominent the two key observations that we made from the more comprehensive IMF Clock Angle plot which included the dependence on IMF Bz as well:
As By grows positive,
a. we see a gap forming in hiss occurrence rate at dusk
b. increase in the hiss occurrence rate is seen in late afternoon/ pre-dusk sector
c. hiss occurrence increases in near and post midnight sector, i.e. hiss is spread over a broader MLT range at higher By.

2. Same plot as above, but with values of IMF By for hiss events are the mean values over the past half hour from the time of hiss occurrence

shows a little difference, if any, from the previous plot. It may act to reinforce our observations.

3. Dependence of Hiss Occurrence on IMF Bz

and one with half hour averaged values of IMF Bz taken to represent Bz at hiss event time:

I am not sure if averaging Bz over past half hour is yielding some more insightful information about hiss occurrence here.

Anyways, we do observe, that as Bz decreases from positive value:

a. hiss occurrence increases in all sectors, dusk and pre-midnight.

b. hiss occurrence covers a wider range of MLT sector 

c. hiss occurrence increases at nearly all frequencies for the MLT sectors where hiss is observed

4. Dependence of Hiss Occurrence on IMF Clock Angle (By and Bz):

Again taking half hour average By and Bz for hiss events puts more events in lower Bz sector:

I believe looking at the individual By and Bz variations separately gives more insight into hiss' dependence on IMF.

Substorm Onset Time & Location Dependence on IMF By & Bz:

I used the substorm onset database made using the IMAGE-FUV data taken from Frey's 2004 paper on substorm observations between May 2001 and Dec. 2002. Again, I plotted two sets of plots to see if there is a change in a general trend, first set of plots used exact values of IMF By and Bz at the time of onset, while the second set used past half hour average of IMF By and Bz from the time of substorm onset.

Note the following for each figure below:
1. There are two plots showing the dependence of Substorm Onset Geomagnetic Latitude and MLT on IMF By (or Bz) respectively.
2. Each plot has three line plots. First just plots the raw data of ~2400 onset times in two and a half year period. Second line shows the running median of the given length to smooth out the data. While the third one is an nth degree polynomial fit through this data.
3. Each plot has x-axis centered at 0 Bz, and the y-axis centered at the mean MLT (or Geom. Lat.) of the entire data.
4. Morning times have been offset by 24 to fit on the graph nicely.

1. Below is the figure showing the dependence of Substorm Onset Geomagnetic Latitude and MLT on IMF Bz for exact Bz values:

2. While figure below is the same plot with half hour averaged Bz values at each onset time:

-> I believe we were expecting to see a downward trend in Geomagnetic Latitude with decreasing Bz, which is also reflected by the fact that auroras are seen at lower latitudes with more negative Bz.
-> MLT of Onset decreases increases a little bit from 23 to ~24 MLT for decreasing Bz. However, I can't fully explain this trend?
-> I also think that you mentioned about Substorm Onset occurring at roughly about 21MLT, which doesn't seem to be the case here?

3. I also plotted similar plots for By, and below figure is plotted using past half hour average By values at onset times:

-> Onset location increases in lat. with increasing By and onset hardly occurs above 72 degree latitude below -5nT By, but there are more samples with onset latitude above 72 deg. as By becomes positive
->  MLT of onset decreases from ~24 to little below 23 with increasing By

Thus, It is difficult to say decisively here but a case can be made that with By going from -ve to +ve, Onset MLT gets closer to the dusk side and Onset latitude also gets closer to SP's geom. lat. s.t. more hiss might be seen in lower MLTs for increasing By.


IMF SuperPosed Epoch Analysis with Short and Medium Time Intervals

Here, I show the superposed epoch analysis of IMF By and Bz with respect to hiss's occurrence in 2 and 3 hr MLT intervals respectively.

1. To compare, I've re-plotted the original superposed epoch analysis with 6 hr time intervals of hiss occurrence, starting at 6 MLT. I've kept the y and x axes uniform for all figures:

2. Following figure shows the plots for 3 hr long hiss occurrence intervals, starting at 2 MLT:

We note the following observations from this figure:

First in Bz:

a. All intervals from morning up to noon have a minimum in Bz that shifts towards the 0hr mark, i.e. they need less and less time after the reversal to southward Bz for hiss to occur. 

b. Also only the above times, also including the 11-14MLT interval, show a reversal from positive to negative Bz and for other intervals, Bz stays -ve throughout, getting more -ve close to the 0hr mark.

c. All the time intervals beyond noon have a minimum value in Bz that shifts away from the 0hr mark, i.e. they take more and more time before Bz sees a minima before hiss is observed.

Now, variation in By:

a. Early morning hiss, from 23-8MLT coincides with positive By, while Post-dusk to Pre-midnight hiss, from 17-23 MLT, coincides with negative By. 

b. Pre-noon hiss, 8-11 MLT, corresponds to negative By, while Post-noon hiss, 11-17MLT, occurs when By is positive.

3. Below is for 2 hr long hiss occurrence intervals:

Again, similar observations can be made here as well:

From Bz plot:

a. Morning to noon hiss shows a reversal in Bz, while Bz for post-noon to midnight hiss stays negative throughout, with minimum in Bz shifting leftwards from the 0hr mark. 

South Pole Magnetometer Database Read and Analysis:

I have inspected the data and have sent a set of questions to Andy regarding its formatting. Its proving to be a little difficult to read but I'll send you a few figures as soon as I read the database in. 

I should first find out the baseline of the H component from the data, i.e. the mean value, right?

Then I can treat del_H as an index similar to the AE and perform occurrence rate analysis. 

What other plots can I generate from this data?