The paper deals with SABER-derived long-term trends of mesospheric temperatures. This is very comprehensive and carefully done analysis. Authors use real (not constant) Yaw cycle lengths. They are making corrections of several factors, which can affect resulted trends, including reduction of seasonal variation effect using climatological NRLMSISE2.1 mean temperature. So many corrections may introduce some imperfection (e.g. couple of years ago NRLMSISE00 was used, now we are using version 2.1 and after some years again a new and more accurate version appears). However, authors made best corrections at the level of present-day knowledge. I require only very minor essentially technical corrections, so after making these corrections I recommend the paper to publication.

Comments:

Line 370: Cooling trend is becoming larger with increasing height, so “decrease” should be replaced by “increase”

Wording and misprints:

• Line 33: delete “and”
• Line 71: “of a cooling trend” should be “cooling trends”
• Line 72: “mesosphere, were” should be “mesosphere were”
• Line 323: “7(b)” should be “5(b)”
• Lines 403 and 480: “The peaks” should be better “The heights of peaks”
• Line 464: “exclulding” should be “excluding”

The paper highlights the MLT temperature trend in high latitudes through a new innovated SABER data processing algorithm that solves the previous issue (fixed 60-day window) regarding the forward drifting of SABER local time coverage, while mitigating properly the embedded bias due to seasonal variations through the assistant of MSIS2.0. The results show mostly cooling trends around the globe with sporadic spots of warming, which is consistent with the numerical studies. The author states that the revealed large cooling at high latitudes MLT demonstrates the high sensitivity to the global climate change in this area. Note that similar statement has been raised by the climate studies focusing on the troposphere and stratosphere. The manuscript and figures are mostly clean, and I just have a few minor questions that need the author to address.

Based on equation 1, removing kt_ut from the mean Td should give mean T_d0, instead of the residual term. Please clarify.

Some of the trend profiles at high altitudes (geometric height) in figure 6 showing near or more than 10 K/decade near 100 km and above, even considering the fitting uncertainty. I feel these cooling trends are a little excessive. The author might want to double check the data quality or the algorithm for this altitude range.

1. Line 42, “highest latitudes” why not just say high latitudes?
2. Line 95, “lower heights”, please be more specific, troposphere or stratosphere? Or just say lower altitudes.
3. Line 121, 60-day
4. Figure 4 caption. I do not see red and green dots, but purple and blue ones. Also, it is very difficult to tell the “+” signs. May want to change the symbol.
5. Line 370, “higher heights”, again please be specific. And replace all the “heights” with “altitudes”.