Local time extent of magnetopause reconnection using space–ground coordination

Zou, Ying; Walsh, Brian M.; Nishimura, Yukitoshi; Angelopoulos, Vassilis; Ruohoniemi, J. Michael; McWilliams, Kathryn A.; Nishitani, Nozomu

doi:https://doi.org/10.5194/angeo-37-215-2019

Articles | Volume 37, issue 2

https://doi.org/10.5194/angeo-37-215-2019

© Author(s) 2019. This work is distributed under
the Creative Commons Attribution 4.0 License.

https://doi.org/10.5194/angeo-37-215-2019

© Author(s) 2019. This work is distributed under
the Creative Commons Attribution 4.0 License.

Articles | Volume 37, issue 2

Regular paper

|

10 Apr 2019

Regular paper |

| 10 Apr 2019

Local time extent of magnetopause reconnection using space–ground coordination

Ying Zou, Brian M. Walsh, Yukitoshi Nishimura, Vassilis Angelopoulos, J. Michael Ruohoniemi, Kathryn A. McWilliams, and Nozomu Nishitani

Download

Final revised paper (published on 10 Apr 2019)
Supplement to the final revised paper
Preprint (discussion started on 03 Jul 2018)

Interactive discussion

Status: closed

AC: Author comment | RC: Referee comment | SC: Short comment | EC: Editor comment

- Printer-friendly version

- Supplement

RC1: 'Referee Report', Anonymous Referee #1, 15 Aug 2018
- AC1: 'Reply to reviewer #1', Ying Zou, 27 Sep 2018
- AC2: 'Reply to reviewer #2', Ying Zou, 27 Sep 2018
RC2: 'Referee Review', Anonymous Referee #2, 24 Aug 2018
- AC4: 'Reply to reviewer #2', Ying Zou, 27 Sep 2018
RC3: 'Reviewer report', Anonymous Referee #3, 28 Aug 2018
- AC3: 'Reply to reviewer #3', Ying Zou, 27 Sep 2018

Peer-review completion

AR: Author's response | RR: Referee report | ED: Editor decision

ED: Reconsider after major revisions (further review by editor and referees) (05 Oct 2018) by Minna Palmroth

AR by Ying Zou on behalf of the Authors (12 Oct 2018) Author's response

ED: Referee Nomination & Report Request started (16 Oct 2018) by Minna Palmroth

RR by Anonymous Referee #2 (02 Nov 2018)

Suggestions for revision or reasons for rejection

The author’s response has provided clarification, and the changes to the paper have largely helped to convey better the objectives of the study and the reasoning behind the analysis techniques used. Ideally, I would still recommend measuring the reconnection-driven ionospheric flows at the OCB if the data allow, as this is the most direct method of remotely-sensing reconnection (and its extent) from the ionosphere. However, I do accept that (in the absence of good measurements at the OCB) there is still merit to measuring the extent of reconnection bursts from strong ionospheric flows poleward of the OCB, as is the case here. This has been used in many previous studies, even though it represents a simplified approach. However, the authors still need to consider seriously the points made below before the paper is published.

Major Comments:

(A) Regarding the latitudes chosen to measure the longitudinal extent:

Obviously, the magnetic local time (MLT) extent and width of the flow channel will change as the flow proceeds further into the polar cap. In all three events, the latitude chosen to determine the longitudinal extent (for panels e and f in figures 2 and 4, and e-h in figure 6) is 2 degrees poleward of the estimated location of the OCB. The variation of the magenta lines encompassing the flow in these figures shows that the extent of the reconnection burst flows, and especially their MLT position can change significantly across these 2 degrees.

The authors have strongly defended their decision and provide comments to justify using data at the more poleward latitude, e.g., lines 384-386 – “While this latitude is 2 degrees poleward of the open-closed field line boundary, the shape of the flow did not change much over the 2 degrees displacement and thus still presents the reconnection extent.” My response would be, if this statement is true, then why not present the variation at the OCB? Or, at least at a point closer to the OCB. I am not totally convinced, from looking at the data shown, that this statement is necessarily true. If the authors are convinced that the higher latitude (e.g., 80 degrees) is to be used then showing a comparison of the longitudinal extent of the flows (preferably of the poleward component of the SECS flows and not the line-of-sight (LOS) values [see below]) at 78 and 80 degrees (and possibly at other locations in between) is needed to prove the point.

I am still of the opinion that the closer that you can get to actually using the flows at the estimated OCB location, the better.

(B) Regarding the use of LOS flows and the time series of the longitudinal extent:

The time series plots that show the evolution of the extent of the flows are very informative and help to interpret the evolution of each of the events. However, using LOS measurements for this purpose is not a particularly good idea. The beam directions across each radar field-of-view (FOV) vary by ~+/-26 degrees from the central look direction. Hence, comparing LOS measurements from radar beams looking in multiple directions can be seriously misleading.

For example, lines 511-512 – “The velocity at -74 to -30 degrees MLON dropped by 100-200 m/s during 1900-1910 UT, while the velocity at -88 to -74 degrees MLON did not change substantially” – Firstly, in LOS measurements, changes like this can happen just due to slight changes in the direction of the bulk plasma flow relative to the line-of-sight direction. Secondly, these are LOS measurements made by different radars with different look directions. At the join between the measurements from the two radars (between figs 6e and 6f), the look directions of the beams from the two radars differ by more than 90 degrees. Hence, these two radars will measure very different LOS flows at this location. Hence, it is difficult to match together the two figures, and it should not be attempted! It would be much better to show the temporal evolution of the poleward component of the SECS flow here.

In addition, MLON should not be used as one of the axes in these plots. Variations should be plotted against magnetic local time (MLT). This removes changes in the longitude of the flow that are related solely to the rotation of the Earth, which is not relevant to this study.

Hence, my recommendation is that panel e in figs 2, 4, and 6 would be much clearer, less ambiguous, and more easy to interpret if (i) the poleward component of the SECS flow was plotted instead of the LOS velocity, and (ii) MLT was used on the y-axis instead of MLON.

(C) Regarding the potential effects of IMF Bx and By on the reconnection burst extent:

I don’t think that enough events have been observed to allow there to be any significant comment on the effects of Bx and By on the reconnection burst extent. There is not enough evidence to support the conclusions presented in section 3.4. I would consider removing section 3.4.

Minor Comments:

(1) Lines 82-83 – Petrinec and Fuselier (2003) appears twice in this list of references.
(2) Line 116 – “FTEs have been observed to be > or < 2 Re wide in local time” – surely this relates to any size of FTE, it will either be smaller or greater than 2 Re. Hence, I don’t get the point of this statement.
(3) Line 198 – Remove ‘are’ after the [Broll et al. 2017] reference.
(4) Lines 327-328 – The phrase “…was confined within the utilized few radar FOVs” would be better written as “…was confined within the FOVs of the radars used”.
(5) Line 553 and Figure 7 – Figure 7 would benefit from the addition of the IMF clock angle variation. The predicted locations of anti-parallel reconnection vary significantly with clock angle, and it is easier to be able to see the clock angle variation without having to visualise the variation based on the variations in IMF By and Bz.
(6) Lines 569-570 – “Studies have found that small |By|/|Bz| relates to anti-parallel and large |By|/|Bz| to component reconnection” – Significant anti-parallel reconnection still occurs for large |By|/|Bz|, but it occurs at higher latitudes on the magnetopause, away from the equatorial plane.

Hide

RR by Anonymous Referee #1 (15 Nov 2018)

RR by Anonymous Referee #3 (03 Dec 2018)

Suggestions for revision or reasons for rejection

Second review of Zou et al 2018

As stated by the title, the goal of this study is to measure the local time extent of magnetopause reconnection bursts using space-ground coordination. This is a very worthwhile scientific goal because knowing the extent of reconnection is important for understanding the geometrical and other factors influencing the reconnection process, which in turn is fundamental to understanding so much of magnetosphere-ionosphere physics and space weather. However, in my opinion, the definition of a reconnection burst and the methodology used to estimate its extent remains too imprecise and inconsistent that the quoted extents of 3, 5, and 11 Re are of questionable scientific value. If this could be improved then I think this would become an excellent and valuable study.

There are really too many detailed points for me to go through so I shall focus on my major concerns:

1. Definition of a reconnection burst. In their response to my first review, the authors say that they are not interested in the extent of non-zero reconnection rate along the magnetic separator but rather the extent of reconnection bursts within it. However, I can find nowhere in the manuscript where a reconnection burst is objectively defined. The implication seems to be that it is a patch of los or poleward component of ionospheric flow above some threshold that is physically distinct from lower los or poleward flow. In practice, a reconnection burst is effectively defined in the paper as a continuous region with los or poleward flow component exceeding half of the peak value. Thus I see no evidence that a reconnection burst is a distinct physical phenomena but merely the highest reconnection rate region of a more extended reconnecting region.

In view of this, I strongly recommend that you do not use the term burst. Instead in the title and elsewhere you should say that you are measuring the local time extent of magnetopause reconnection (not local time extent of magnetopause reconnection bursts) and then clearly state what your definition of local time extent is. At minimum, this could be the definition that you have been using – the region exceeding half the peak reconnection rate value. However, note that for the a Gaussian spatial variation in reconnection rate (e.g., line 387), the FWHM points are at +/-1.18 standard deviations from the peak and thus about 30% of the total reconnection rate lies outside these bounds.

Thus, personally, I think it would be helpful to also quote the full width of non-zero reconnection rate. The full width of non-zero reconnection rate is arguably a better measure too because choosing the half-maximum rather than some other fraction (e.g., 1/e) is arbitrary (see point 3 below) whereas non-zero is not, and the relationship of the FWHM to the total reconnection rate contained within it depends on the shape of the reconnection rate spatial variation. That is, I recommend quoting the region over which the reconnection rate exceeds zero within uncertainties (i.e., the difference from zero is statistically significant). If the non-zero region extends beyond the observed region then the quoted value would be a lower bound.

2. Estimation of the reconnection rate. In my previous review, I recommended that the authors estimate the reconnection rate from the ionospheric electric field in the frame of the generally moving open-closed field line boundary, following the methodology of Chisham et al. (2008). I thank the authors for trying this for the Feb 2013 event. However, the authors relegate this to the supporting information and dismiss this approach in lines 393-401 of the main manuscript and in their response. I really must take issue with the reasoning for this and strongly recommend that the Chisham et al method is used:

Firstly, in the authors’ response, they reject the need for the method because they say “our approach is consistent with a number of past works cited above”, by which I assume that they mean the 17 references from Goertz et al (1985) through Zhang et al. (2008) that they cite in response to my point 1. However, it should be noted that these works are all 10 or more years old and pre-date the Chisham et al. (2008) method. Thus in my view the state of the art has changed since then and this should be reflected in the standard of data analysis used in this paper.

Secondly, the authors argue that the uncertainty in the estimation of the OCB velocity is large and thus it is reasonable to focus “on the velocity profile poleward of the open-closed field line boundary, which is less affected by the error associated with the boundary”. So what the authors are effectively saying is that in some way the los velocity several degrees poleward of the OCB is a better estimate of the magnetopause reconnection rate than estimating the electric field in the moving frame of the OCB. How can this be? No scientifically based arguments are given as to why this should be so. I fear that what the authors are really saying is that they don’t want to acknowledge and deal with the inconvenience of observational uncertainties when estimating the local time extent of reconnection (whether a burst or not). In my opinion this is not good science.

If one truly wants to estimate the local time of reconnection then one must be able to identify where the reconnection rate is non-zero. This inevitably requires identifying the OCB, its motion, and the ExB velocity component perpendicular to the OCB. As the authors correctly say, first order spatial differences of the OCB latitude from SuperDARN measurements introduce an uncertainty of 45 km in 2 min, corresponding to an OCB velocity uncertainty of 375 m/s or about 23 mV/m. However this can effectively be reduced somewhat using higher-order differences for the time derivative, or considering a longer sampling interval if this seems appropriate. Either way, this uncertainty has to be taken into account, as detailed in Chisham et al. (2008).

Applying the first-order uncertainty to Figure S3 one would conclude that the -83 to -94 MLON region has a non-zero reconnection rate at the 1 sigma level and this is thus the minimum extent of reconnection. Admittedly the 1 sigma level is not very compelling to a statistician but that is the reality and the scientific method. At least you have quantified the extent for a given confidence level even if that level is low.

The alternative is that one limits oneself to determining the extent of high-speed or non-zero los or poleward flows at given latitude, as you have done, but then one cannot really claim that this is the extent of reconnection in my view.

3. Patchy versus extended reconnection. To further emphasise what I believe is the questionable scientific value of the three quoted reconnection extents, I would like to compare the los and SECS velocity profiles shown in figures 2f and 6g. The former is inferred to have a reconnection extent of 13 deg MLON or 3 Re at the magnetopause and the latter 63 deg MLON or 11 Re. Yet I suspect from what I can see in figures 2a and 2b that if the velocity profile shown in figure 2f were extended over the full longitudinal extent of the SuperDARN measurements then it would be similar to that in figure 6g.

Specifically figure 6g has a los velocity maximum at -73 deg MLON. It is clearly above the half-maximum value over a 20 deg MLON region between -87 and -65 deg MLON, at or below the half-maximum value between -55 and -65 MLON and rises again to intermediate values between -65 and -20 deg MLON. (Incidentally -27 MLON shown by the black dotted vertical guideline is not the FWHM point). This is concluded to be an extended reconnection example. Figure 2f has a los velocity maximum at -82 deg MLON and is clearly above the half-maximum value over a 13 deg MLON region -92 and -79 deg MLON before dipping down to just below the half-maximum and likely increasing again above it over an extended region beyond the limit of the plot at -70 deg MLON. This is concluded to be a patchy reconnection example yet I believe the distinction between this and the extended reconnection example depends on a marginal difference in the dip below the half-maximum value in the two cases.

For example, if one had chosen 40% of the maximum rather than 50% then both might have been extended. Or if one chose a slightly lower latitude (closer to the OCB) then I suspect from figure 2b that the dip below the half-maximum in Figure 2f might not be as evident. If there is such a sensitivity in the ‘reconnection’ extent to the velocity threshold and/or latitude then this casts doubt on the scientific robustness and value of the quoted extents, even without the caveats of point 2 above. Apologies if I am wrong but I’d appreciate seeing los and SECS profiles at different latitudes and over the full MLON range to clear this up. Thanks.

4. Other points. Besides the above major points, I’d also like to mention:
a. It’s really difficult to relate the MLON profiles with the FOV maps when you don’t put MLON labels on the maps!

b. I think you might be getting your east and west the wrong way round in some places, such as lines 362-366. You say the eastern boundary is at -82 deg MLON and the western boundary is at -77 deg MLON. But isn’t westward in the sense of more negative MLON? As confirmed by THA being westward of THE in Figure 1 and Figure 2e.

c. I felt that the argument involving distinguishing between 200 m/s and 220 m/s spectral widths in lines 334-345 to be doubtful. Firstly, I’m not aware that a simple spectral width threshold corresponding to newly-reconnected field line precipitation has been calibrated (as opposed to the Chisham spectral width boundary method). Secondly, the eastward edge of the pre-noon flow region marked by the magenta line in Figure 2a actually lies through the eastern one of the two dark red spectral width regions in Figure 2d, whereas by your argument shouldn’t it lie between them or at the eastern edge of the western dark red region? I think this further supports my argument in point 3 above that this is an extended rather than patchy reconnection region.

d. Why do you not publish all 6 events that you have identified? For example in the supporting information at least as a brief description and summary figure like figure 2f, 4f, 6g for each case. It might help strengthen your conclusions.

Hide

ED: Reconsider after major revisions (further review by editor and referees) (03 Dec 2018) by Minna Palmroth

AR by Ying Zou on behalf of the Authors (14 Jan 2019) Author's response Manuscript

ED: Referee Nomination & Report Request started (15 Jan 2019) by Minna Palmroth

RR by Anonymous Referee #2 (04 Feb 2019)

RR by Anonymous Referee #3 (12 Mar 2019)

ED: Publish as is (13 Mar 2019) by Minna Palmroth

AR by Ying Zou on behalf of the Authors (22 Mar 2019) Manuscript

Short summary

Magnetopause reconnection is a process whereby the Sun explosively transfers energy to the Earth. Whether the process is spatially patchy or spatially continuous and extended has been under long debate. We use space–ground coordination to overcome the limitations of previous studies and reliably interpret spatial extent. Our result strongly indicates that both patchy and extended reconnection is possible and, interestingly, that extended reconnection grows from a localized patch via spreading.