Dear Tongjiang,

It looks like this is quickly becoming technical so this will be our last reply on the list:

1) The bottom panel of Fig.2 shows the blue-wing asymmentry from R-B analysis in range Y=60"-110" (which is consistent with that of my skewness analysis), however, the propagating disturbances in intensity and Doppler
shift span from Y=80" up to 200". How do you understand the blue wing
asymmetry is only limited at the footpoint of fan loops? In addition,
the B-R aysemmtry pattern does not show the propagating feature as seen
in intensity and Doppler shift, this suggests that the outflow present
at the loop footpoint and the propagating disturbances may be different
things.


As we mention in Section 3,  the Fe XIII line is biased towards the red because of a small, hot blend, and a slight gradient in the background emission. This means that Fe XIII cannot be used to exclude upflows in the region of y=80 to 200". In  fact Fe XIV shows blue asymmetries throughout that region (see, e.g., the RB signal for velocities between 65 and 130 km/s in the following movie: http://dl.dropbox.com/u/3057160/Wang/Wang-EIS-FeXIV264-RBRasterScan.mov 
The R-B asymmetry does show traces of the propagating features, but as we show in Figure 6, the modest S/N of the data does not allow good visibility of the 3rd moment (as opposed to the first two moments). It just becomes too noisy, except for some locations.
However, the double Gaussian fit results in our paper shows clear signatures of propagating signatures in the second component at 50 km/s.

2) You stated that "Most of the line width peaks are associated with stronger blueward asymmetries in Fe XIII" and "correlation of variations in intensity and Doppler shifts in Fe XIII and Fe XIV are significant".
I don't agree with you. The approximate accounting from your Fig.4 and
appendix Fig.1 and 2, show that the associations are less than 50%.

We have done a detailed analysis of the effects of S/N on the correlation between the oscillations between different moments, as shown in Figure 6, and for the S/N of the observations the observed correlation is what would be expected. So, I guess we'll have to agree to disagree on this topic until we get higher S/N data.

3) Unfortunately, you did not show results of single-Gaussian fitting for
Fe XIV as a comparison with Fe XIII.
In fact, there is a clear difference in intensity, Doppler shift, line width, and R-B analysis of rastering image and time sequence between
Fe XIII and Fe XIV. The FeXIV line has a very strong blue-wing asymmetry, so
you could make a double-Gfit to separate the flow component. Have you also made a double-Gfit for Fe XIII image and time sequence? If so, could you
show me your result? I'm very interested in it. Moreover, your double gaussian fit to Fe XIV does not directly explain the propagating disturbances seen in Fe XII or Fe XIII, because you even did not prove in
your paper that the intensity and Doppler shift in Fe XIV show the same propagating pattern as in Fe XIII (at least I did not find that in my analysis). In addition, from your Fig.9 (time sequence of 2nd component) I hardly see any propagating signature. If this is due to the
quanlity of the plot, I'm sorry about it.


The key words here are signal to noise. We are looking at very faint signals that are more difficult to discern especially as we go to higher order moments. However, we do believe that Fe XII and Fe XIII show decent correlation given the limitations of the data, and as we write "while the correlation with Fe XIV 274 ĚŠ is not as clean, it is significant." But, in our view it is somewhat beside the point whether or not Fe XIII and Fe XIV show exactly the same oscillations [after all, AIA data clearly shows that such propagating disturbances are sometimes similar for several lines, and sometimes not, I guess that depends on the thermal properties of the upflows/loops]. The point is that within the S/N and other limitations of the data, there is enough evidence to suggest that blueward asymmetries *and* linewidths are temporally correlated with the first two moments for several lines (including Fe XIII and Fe XIV). Double fits to the one line where it seems feasible (Fe XIV) for this data (not Fe XIII which has a blend, as mentioned), supports the scenario of a separate component that periodically recurs. Remember that given the S/N and resolution of this data, there is only so much "clear correlation" that you can expect to observe, as we show with Monte Carlo simulations in the paper. One of the issues we have tried to illustrate with the Monte Carlo simulations is that the signal of a faint upflowing component can easily hide in higher moments, and will be visible the most clearly in the intensity and velocity.
This is exactly what is observed. :-)

With respect to Fe XIV propagating disturbances -- they are clearly visible in the double fit plots, they occur in the same region as the Fe XII and Fe XIII signals and the detailed Fe XIV signals (first 4 moments) are reasonably well correlated with those of Fe XII and Fe XIII. We think it's a stretch to say that this means Fe XIV shows a completely different phenomenon from Fe XII/Fe XIII.

4) I think it is too early to make any conclusion, more detailed analysis of this and other data sets are needed.

We agree with you on the latter part and hope our paper will trigger such analysis. As we mentioned yesterday, we believe that flows and waves both occur on the Sun -- the question is which phenomenon dominates specific observations. As illustrated in our paper (and the above discussion), this is a trickier topic than has been assumed so far.

With respect to the conclusions, we think it is not too early to already draw one conclusion: it is premature to perform coronal seismology based on sound waves (emanating from plage regions) using intensity and velocity oscillations alone, which has been the norm for the past 10 years. We will contact you directly with a more technical reply tomorrow morning to keep the list uncluttered.

Cheers,
Bart & Scott
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Bart De Pontieu    --  Lockheed Martin Solar & Astrophysics Lab, Palo Alto
bdp@lmsal.com   --     3251 Hanover St., Org. ADBS, Bldg. 252, CA 94304
http://leffe.lmsal.com/bdp -- Phone 1-650-424-3094 / Fax 1-650-424-3994 
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

On Sep 20, 2010, at 3:50 PM, Tongjiang Wang wrote:

Dear Bart and Scott,

Thank you for your preprint and discussions. I have read your paper in detail and have some comments and disagreements as following:

1) The bottom panel of Fig.2 shows the blue-wing asymmentry from R-B analysis in range Y=60"-110" (which is consistent with that of my skewness analysis), however, the propagating disturbances in intensity and Doppler
shift span from Y=80" up to 200". How do you understand the blue wing
asymmetry is only limited at the footpoint of fan loops? In addition,
the B-R aysemmtry pattern does not show the propagating feature as seen
in intensity and Doppler shift, this suggests that the outflow present
at the loop footpoint and the propagating disturbances may be different
things.

2) You stated that "Most of the line width peaks are associated with stronger blueward asymmetries in Fe XIII" and "correlation of variations in intensity and Doppler shifts in Fe XIII and Fe XIV are significant".
I don't agree with you. The approximate accounting from your Fig.4 and
appendix Fig.1 and 2, show that the associations are less than 50%.

3) Unfortunately, you did not show results of single-Gaussian fitting for
Fe XIV as a comparison with Fe XIII.
In fact, there is a clear difference in intensity, Doppler shift, line width, and R-B analysis of rastering image and time sequence between
Fe XIII and Fe XIV. The FeXIV line has a very strong blue-wing asymmetry, so
you could make a double-Gfit to separate the flow component. Have you also made a double-Gfit for Fe XIII image and time sequence? If so, could you
show me your result? I'm very interested in it. Moreover, your double gaussian fit to Fe XIV does not directly explain the propagating disturbances seen in Fe XII or Fe XIII, because you even did not prove in
your paper that the intensity and Doppler shift in Fe XIV show the same propagating pattern as in Fe XIII (at least I did not find that in my analysis). In addition, from your Fig.9 (time sequence of 2nd component) I hardly see any propagating signature. If this is due to the
quanlity of the plot, I'm sorry about it.

4) I think it is too early to make any conclusion, more detailed analysis of this and other data sets are needed.

Best Regards

Tongjiang

-------------------
Tongjiang Wang

NASA GSFC - Code 671
Bldg 21 - RM C121
Greenbelt, MD 20771

Tel. 301-286-6575
Fax. 301-286-1617
tongjiang.wang@nasa.gov

On Sun, 19 Sep 2010, Bart De Pontieu wrote:

Dear all,

You may also be interested in our paper (De Pontieu & McIntosh, 2010), which was accepted to ApJ  in early August, and which uses detailed modeling and double component fits of EIS observations to show that some widely studied examples of "slow magnetoacoustic waves" are in fact most likely caused by quasi-periodic upflows of order 50 km/s.  Here is the link: http://arxiv.org/abs/1008.5300

It is clear that both flows and sound waves occur in the corona. The point of our paper is that disentangling which scenario dominates specific observations is not straightforward. We find that coronal seismology just based on quasi-periodic changes in intensity (e.g.., using TRACE, STEREO or AIA) is ambiguous at best. Instead, a detailed study of at least intensity, velocity, linewidth and asymmetry of the spectral line (as well as signal-to-noise) is required to distinguish between both scenarios. While Erwin's paper expands in general on one of the interpretations we offered in our paper, our results show that the devil is in the details of the observed amplitudes, periods and correlations of the oscillations of the first 4 moments of the spectral line profile, and their phase/amplitude relationships. Such a detailed comparison with specific observations remains to be done for the waves interpretation. At least for the observations we focused on, the evidence from our double Gaussian fits support the upflow scenario. We are preparing a follow-up paper that shows that several other popular "waves" datasets are compatible with the presence of quasi-periodic upflows.

It is possible that some observations end up being compatible with both flows and waves (although this remains to be shown directly). However it is clear that the difficulties of discriminating between both scenarios in specific observations presents significant challenges to coronal seismology based on sound waves. Especially because the presence of ubiquitous upflows at least are well supported by spectral imaging data from various instruments (Hinode, ground-based observations, SDO). Detailed analysis that looks into S/N, line of sight, the first 4 moments of the spectra and high quality imaging data will be necessary for each dataset to see if both scenarios co-exist (e.g., waves triggered by upflows), or whether one or the other dominates.

Cheers,
Bart & Scott

On Sep 19, 2010, at 9:25 AM, Erwin Verwichte wrote:

Dear all,

You may be interested in our paper, which has just been accepted for publication in ApJL and which adds to the current debate on periodic flows and waves.

Periodic spectral line asymmetries in solar coronal structures from slow magnetoacoustic waves
Verwichte, E., Marsh, M., Foullon, C., Van Doorsselaere, T., De Moortel, I., Hood, A.W. and Nakariakov, V.M.

Abstract: Recent spectral observations of upward moving quasi-periodic intensity perturbations in solar coronal structures have shown evidence of periodic line asymmetries near their footpoints. These observations challenge the established interpretation of the intensity perturbations in terms of propagating slow magnetoacoustic waves. We show that slow waves inherently have a bias towards enhancement of emission in the blue wing of the emission line due to in-phase behaviour of velocity and density perturbations. We demonstrate that slow waves cause line asymmetries when the emission line is averaged over an oscillation period or when a quasi-static plasma component in the line-of-sight is included. Therefore, we conclude that slow magnetoacoustic waves remain a valid explanation for the observed quasi-periodic intensity perturbations.

http://go.warwick.ac.uk/erwin_verwichte/publications/verwichte_apjl2010.pdf

Best regards,

Erwin

Dr Erwin Verwichte
Centre for Fusion, Space and Astrophysics
Department of Physics
University of Warwick
Coventry CV4 7AL, United Kingdom
Office PS1.09
tel: +44(0)2476524917
fax: +44(0)2476524887
http://www.warwick.ac.uk/go/erwin_verwichte

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Bart De Pontieu    --  Lockheed Martin Solar & Astrophysics Lab, Palo Alto
bdp@lmsal.com   --     3251 Hanover St., Org. ADBS, Bldg. 252, CA 94304
http://leffe.lmsal.com/bdp -- Phone 1-650-424-3094 / Fax 1-650-424-3994
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~



_______________________________________________
Loops mailing list
Loops@solar.physics.montana.edu
https://mithra.physics.montana.edu/mailman/listinfo/loops

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Bart De Pontieu    --  Lockheed Martin Solar & Astrophysics Lab, Palo Alto
bdp@lmsal.com   --     3251 Hanover St., Org. ADBS, Bldg. 252, CA 94304
http://leffe.lmsal.com/bdp -- Phone 1-650-424-3094 / Fax 1-650-424-3994 
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~