|I am satisfied with the authors' responses and revisions to most of my comments. There are, however, one major comment and a few specific requests for clarification left that need to be addressed before the final version of the manuscript is published.|
All line numbers in the following refer to the annotated version of the manuscript (the one that was included in the author response pdf file).
line 548: It is puzzling that there is no QBO in the MIPAS wavenumber-4. This is inconsistent with SABER-based diagnostics of the DE3. The cited Li et al. paper does not quantify the QBO effect but uses vague language such as "slightly stronger" etc. Looking into Figure 4 of the Li et al. paper, I believe there's quite a bit of QBO like amplitude variability going on. Overall, Li et al., 2015, Figure 4 looks consistent with Oberheide et al., 2009, Wan et al., 2010, Zhou et al., 2015, who consistently found a 15-20% amplitude modulation by the QBO. This needs to be discussed more closely in the manuscript. At the very least, it needs to be stated that the MIPAS k=4 result (no QBO) is inconsistent with the the SABER results (which are internally consistent with TIDI tides) and QBO signals in the ionosphere (Wan paper, also recent Chang et al., COSMIC TEC diagnostics).
In response to my original comment 2 to credit earlier work by Huang et al. re migrating tide-QBO studies, the authors now refer to Huang et al., 2006. I grant that some tidal discussion is included there, but the paper is more about the mean temperature. The more appropriate reference would be Huang et al., 2010 (10.1029/2009JD013698).
I am a little confused about the response to my original comment 3 (the solar flux dependence). The manuscript states (line 543) that for the inter-annual variability a solar flux component had to be included to get better results. Why is it then not "adequate" to perform an analysis for individual years from MIPAS?
Response to my original comment 5 (the part about the purpose of the GSWM/MSIS comparisons). line 269: GSWM-09 and CTMT do NOT include any in-situ tidal forcing in the thermosphere. Both models are designed to look into upward propagating tides and thus produce a (1,1) mode structure even at 110 km. However, at this altitude in-situ forcing kicks in. It needs to be clarified that both GSWM-09 and CTMT do not include the in-situ DW1. For GSWM-09, see for example the statement towards the end of section 1.3 in Zhang et al., 2010a. As such, the difference to MIPAS at 110 km is further support for the in situ forcing mentioned in the previous sentence in the manuscript. In the same context: line 283. I also believe that using tides from NRLMSISE below 110 km or so is not very meaningful, due to the data entering MSIS at these altitudes. Doug Drob and the NRL team work on a new MSIS version that will do a much better job. I urge the authors to omit the NRLMSIS data in the Figs between 80-100 km!!!! There is no use in comparing a model with real world data in a height region where the model is known to have deficiencies.
line 44: Oberheide et al. 2009 only discuss the DE3 QBO which most likely comes from the mesospheric QBO through the Ekanayake et al. mechanism.
line 473: This is a very (too) strong statement. There is indeed a decent understanding of what's the DE3 is doing to the ionosphere, see for example recent papers by Loren Chang using COSMIC TEC. There are some unknowns, e.g., the SE2 contribution and possibly some uncertainty about how important F-region meridional winds are. I suggest to add a qualifying sentence.
line 485: An additional contribution to the higher peak altitude could be a the interference from D0 and TE1. I suggest to mention this as an additional possibility.
Figure 1 & 3, right column: figure titles seem to be printed over each other
line 500: "There oscillation"