Reply to the 2nd referee report regarding MS# 95.0431:

"Observations of polar patches generated by solar-wind-driven Pc5 field
line resonances and pulsed magnetic reconnection at the dayside
magnetopause", by Prikryl et al.

I. Response to general comments

  We have been encouraged by the first paragraph of the referee's report
noting the improvements of the manuscript and we have appreciated
further comments. Also, we sincerely believe that the referee's
constructive criticism have made this revision a worthwhile effort.
  The referee has raised two major objections in the second report. First,
our conclusions that field line resonances (FLRs) were excited near the
magnetopause have been questioned again. However, this time because
they seem to occur too close to, or even within, the cusp (based on the
radar cusp evidence presented in the first revision of the manuscript).
Second, a lack of precision in our use of the term "patch" has been
pointed out. In retrospect, we agree with the referee that these two major
issues were not satisfactorily resolved in our first revision and that they
need more clarification. Also, we agree with the referee that a single case
of a rather complex event is hard to argue convincingly, particularly when
it involves 3 important phenomena (patches, FLRs and pulsed
reconnection) the origins of which have been intensely studied in the past
decade. The original manuscript addressed two complex issues - the origin
of patches and the solar wind-magnetosphere coupling. We believe that
this lack of focus is one of the reasons why the manuscript apparently still
did not reach the "maturity" and was not recommended by the referee
even after a significant first revision.
  Therefore, in order to efficiently respond to the new criticism that has
been raised by the referee, we decided to split the manuscript into two
parts which we now submit as a revision of what was the original
manuscript. We would like to emphasize that even though we added some
new material to further support our conclusions these two complementary
parts together address the same topics (see above) and reach the same
general conclusions as those stated in the original manuscript. We believe,
that this revision made the presentation of the results more focused and the
conclusions which have been further substantiated and refined. Learning
from our mistakes and following the critical advice by the referee we have
now clarified a number of points, particularly the use of the term "patch"
vs. "backscatter band", and have further checked the manuscript for the
consistency of statements. Also, we clarified the role of FLRs on "firmly"
closed field lines vs. the FLRs on the field lines that are about to be
opened (merged). New evidence is included to support our conclusions
about the link between FLRs and flow channel events (FCEs) and the
direct coupling (via FLRs) of the IMF oscillations to the magnetosphere
near the magnetopause. Manuscript A (paper 1 in the text) concentrates
on the coupling beween the solar wind IMF and the
magnetosphere/ionosphere system. On the other hand, manuscript B
(paper 2 in the text) mainly deals with the questions about the polar patch
formation. Because these two processes are related, in manuscript B we
also briefly discuss some observations pertinent to the solar wind-
magnetosphere-ionosphere coupling making references to the compaignon
paper (manuscript A) to avoid duplication. In manuscript B, the discussion
of phenomena like DPY currents and FLRs is minimized and restricted
mainly to the observations on the December 2, 1993 event in support of
our conclusions in manuscript A. The role of these phenomena in the
coupling of the IMF oscillations to the polar and auroral ionosphere is
discussed in more details in manuscript A. Dividing the manuscript into
two parts allowed us to include another, yet more "compact", event to
further illustrate the ideas that are proposed in these two manuscripts.
Using these new and rather clear-cut radar observations we argue what we
believe is a strong case of the FLR modulated reconnection at the dayside
magnetopause. Furthermore, by exploiting some previously published data
sets and the results presented in manuscript B we have built a stronger
case for pulsed reconnection without muddling and defocusing the
discussion by simultaneously addressing the question of patch formation
as was the case in the original manuscript.
  In these two manuscripts we have responded to both general and
specific comments. Below we briefly describe the major modifications and
clarify the points made in the referee's reports. In view of the new
evidence (presented in manuscript A) showing that FCEs are extensions
of the FLR structure to the outermost closed field lines that are about to
be reconnected with the IMF, we can now somewhat "relax" the common
assumption that FLRs occur on "firmly" closed field lines well inside the
magnetosphere. These observations show that if an FLR is excited near
the magnetopause the field lines can be opened in the midst of the
resonance leading to FCE signatures of FTEs as more tension is being
applied on these now opened field lines carried across the cusp into the
polar cap, while the IMF oscillations continue to be transmitted to the
polar ionosphere along the connecting (newly merged) field lines. Such
disturbed FLR/FCE signatures are shown in both manuscripts A and B. 
  Responding to the referee's comment suggesting some apparent
inconsistency in the interpretation of FLR signatures after the equatorward
edge of the "radar cusp" was delineated in Figure 5b (manuscript B) let
us further explain it for that particular case. Note that, at 1800-1830 UT,
the Nipawin radar data (Figures 5a and 7bc) clearly show such "disturbed"
FLR signature near BACK which then evolved into flow bursts (FCEs)
poleward of FCHU. At this time, the equatorward boundary of the "radar
cusp"  was near BACK (Figure 5b). However, a clear FLR signature was
observed by the Red Lake radar near GILL which was undoubtedly
equatorward of the cusp. Furthermore, it should be noted that large
spectral widths can also be observed in LLBL, i.e., the equatorward
boundary of the "radar cusp" is somewhat fuzzy. This issue of the cusp
footprint being determined from spectral widths is even further
complicated in the E region during the second pulsation event (after 1940
UT). We have new discussed and clarified this uncertainty in section 3.2
and Figure 5b in manuscript B. 
  Now we turn to the question of the correlation between the IMF and
ground data. We agree with the referee that the correlation between ISLL
X and IMP8 By (Figure 3a) is "not exactly eye-catching". The reason why
the visual impression and even some of the cross-correlation result do not
appear to be convincing has now been clarified. However, the correlation
that is shown in Figure 10 is now supported by new evidence of a strong
correlation in the frequency domain. The correlation between FFT power
spectra of the IMF and ground signals (By and Z, in particular) is found
to be the strongest at ISLL. This further evidence of a direct coupling of
the IMF oscillations to FLRs is now discussed in manuscripts A and B.
  The second general comment made by the referee was about our
inconsistent use of the term "patch" vs. "backscatter band". Freeing our
hands from some of the complex issues of solar wind-magnetosphere
coupling and its signatures in the ionosphere (which have been handed
over to manuscript A) we concentrate on patches in manuscript B. We
have carefully examined our use of the above terms and also corrected any
other apparent inconsistencies of our statements. We have included more
detail on the interpretation of the data and discussed the relative
importance of various ionization sources for patches. Also, we discuss the
importance of the IMF-By polarity but because By was mostly positive we
firmly stand by the cross-identification of patches vs. backscatter bands as
presented in manuscript B. We agree with the referee that the schematic
model (originally presented in Figure 19 but now presented in Figure 7,
manuscript A) will "reverse" the sequence of patches vs. FCEs. The By
polarity has been specified in the caption for Figure 7 (manuscript A) and
the case of By<0 is briefly discussed in the text. However, it should be
noted that even though By<0 will result in an eastward FCEs which is
more intimately associated with a "patch-to-be" structure equatorward of
such FCE (e.g., augmenting the ionization by precipitation associated with
the upward FACs), it is the fully formed patch poleward of an FCE that
is "detached" by the FCE related downward currents and O+
recombination blowing a "hole" in the ionosphere. Admittedly, an
alternating By will make patches more irregularly spaced and this was
observed (see, e.g., the gap at 1955 UT in Figure 5a to be compared with
a deeper and wider trough between patches at 2025 UT; middle panel of
Figure 16). Also, a more variable By will likely distort patches (see a
reference to optical observations with regard to the interval 1800-1830 UT
made in manuscript B).
  We now consider the general remarks made by the referee raising two
main issues successfully resolved and proceed with our response to the
specific comments.

II.  Response to specific comments.

#1.  The "CADI" acronym has been explained.

#2.  This "typo" has been fixed.

#3.  This sentence has been removed, because it appears earlier in the
first paragraph of section 3.2 in a more readable form.

#4.  This reference to patch O that can be seen in Plate 1 as a decrease
in the spectral widths has been removed. (Patch O has been mentioned in
the text and is labeled in Figures 5a and 11.)

#5.  The magnetometers that "observed a strong burst..." have now
been specified in the text.

#6.  This refers to both absorption and magnetic pulsations. The text
was reformulated.

#7.  The association between the backscatter bands and absorption
enhancements (1800-1830 UT) that was made in the original version of the
manuscript has been relaxed. Only the backscatter bands are now labeled
(including band a which was observed by the Red Lake radar although
this is not so clear from figures - the uncertainty of this band is indicated)
and labels pointing to absorption enhancements have been removed. The
complexity of the riometer and radar signatures during this period is
discussed in more detail (see, e.g., second paragraph in section 4.1). 

#8.  The statements about the correlations have been clarified.

#9.  Because narrow windows are required to restrict the spectral
analysis to a particular pulsation we have opted for MEM in Figure 8
(manuscript B). However, we have also computed the FFT power spectra
for approximately the same time intervals. The two methods gave
consistent results. Because of rather short FFT/MEM windows the MEM
spectra better resolved peaks which was the reason why we used them to
illustrate the spectral content of the time series. Naturally, the
corresponding FFT power spectra (now shown instead of MEM spectra)
show broader peaks. However, this does not mean the actual spectra
content was broadband. On the contrary, we show evidence that the time
series contain many discrete frequencies (this is now further discussed in
both manuscripts). Also, the FFT method confirmed some of the results
that may have looked suspicious in MEM spectra. For example, for GILL
magnetometer data, the 16-pts (1800-1832 UT) FFT spectra with
resolution of ~0.5 mHz showed broad peaks (~1-mHz half-power width
for all three components) centered at 1 mHz and 5.1 mHz which is in
reasonably good agreement (except for the width) with the MEM spectra
for GILL. It should be noted (no stated in manuscript B) that the time
series were detrended (by subtracting the 2nd order polynomial fit) which
removed very low frequency components.
     
However, we now admit that MEM have produced some "unwanted"
results. For example, some peaks were shifted and the spectral estimates
emphasized only the dominant components. However, in general, the
agreement between MEM and FFT (in spite of slightly different windows)
was reasonably good. While we significantly modified the discusion of
Figure 9ab (previous Figure 8) the essential conclusions have not been
changed.  

#10. This comment is related to the previous remarks, but we feel grateful
to the referee for bringing it up again and suggesting further checking by
using the FFT method. We have reconsidered our approach when
extracting the spectral information from these time series that are
apparently quite rich on spectral content. This impression that was
discussed in the manuscript has now been further supported by new
results. Because we were somewhat misled by seemingly short durations
of the oscillation bursts in the time series and tried to avoid the "zone of
confusion" (1835-1930 UT) in the BARS data, we initially aimed the
spectral analysis at very short periods opting for MEM and neglected
FFTs with a wider windows. We have now included ~130-min FFTs
presented in Figure 9cd in manuscript B (see also similar results in
Figures 5/8 in manuscripts A). In addition to all of the spectral
components discussed with Figure 9ab (previously Figure 8), more
discrete peaks are found in these spectra. Figure 9cd shows further
evidence in support of our conclusion that the FLRs were directly driven
at latitudes equatorward of the cusp. Namely, there was a strong
correlation between IMP8 By and ISLL Z. A similar case is discussed in
manuscript A.

Furthermore, these new results brought us back to the beginning when we
somewhat "overemphasized" the section (in the original version of the
manuscript) on a possible link to solar oscillations. We have now
expanded this discussion and presented new evidence for solar oscillation
frequencies that are found in the IMF. However, we have restricted this
discussion to some essential results (presented in both manuscripts) and
labeled as preliminary because we believe they should be reported in a
separate paper.

#11-13. After reconsidering our statements we agree with the referee that
we are on a somewhat shaky ground trying to explain possible differences
between the "sharper" and "wider" absorption bays in terms of differences
between hard and soft precipitation. This possibility is  still mentioned but
the argument suggesting different altitudes is removed from manuscript B.

#14. From previous comparisons between radar and optical data we
strongly believe that the BARS radar signatures discussed in the
manuscript "provide a strong evidence for the presence of auroral arcs".
There is nothing else that we could think of that could cause this type of
VHF spectra, temporal characteristics and spatial distribution of
backscatter. Also, we think that there is nothing wrong in saying that VHF
radars can "see" auroral arcs. In fact, there is quite striking resemblance
between the observations of optical transients (PMAFs) and the radar data
which fully support the "radar evidence" of auroral arcs (this is now
discussed in sections 4.1. in both manuscripts A and B). We may have
used (perhaps more than necessary) a careful language when interpreting
other phenomena studied in the manuscripts A and B but we firmly stand
by our conclusion that BARS data "provide a strong evidence for the
presence of auroral arcs".

#15. A reference to Sandholt et al. [1990] is made. Also, section 4.1 is
expanded to provide more evidence on the presence of dayside auroral
arcs based on the similarity of the optical and radar signatures of flux
transfer events.

  It should be noted that we have not restricted this revision to the
referee's list of comments and other parts of the original manuscript have
been reworded. We believe that the manuscript has been substantially
improved and that the conclusions are substantiated. Once again, let us
emphasize that the clarity of the manuscript(s) improved (as we hope)
mainly because of splitting this work into two complementary parts.
Naturally, this presupposes that one part can not be published without the
other and we may be running a risk of not having one manuscript
recommended/accepted for publication meaning that both are "doomed".
However, we hope that this will not be the case and that the two
manuscripts will be considered as two complementary parts into which the
original manuscript has evolved. 


                    ~~~~~~~~~~