This is a very nice manuscript describing the TEMPERA temperature measurements, the retrieval scheme, and the ability of this instrument to provide measure temperature variations (primarily diurnal) from tides.

The instrument apparently no longer does tipping curves.  Are these not necessary to establish the level of tropospheric attenuation of the stratospheric measurement?

The diurnal temperature tides do not appear unreasonable, but, given that the measurements need to be taken from an instrument at a diurnally varying surface through a diurnally varying troposphere, are there any steps taken to ensure that these tropospheric variations are not mapped into the small (<1%) stratospheric variation shown in Figure 6?  Perhaps the errors have been estimated and are much smaller than 1%, but, in any case, some short discussion of this would be appropriate.

Please present Figure 7 and Figure 8 in the same aspect ratio so that they can be more easily be compared.

The result that stands out in these figures is the consistency of the 18 LST phase in the upper stratosphere and lower mesosphere.  This recent article by Leroy and Gleisner seems to agree: https://doi.org/10.1029/2021EA002011

Line 241- “Mai” should be “May”

Paragraph beginning on line 248 and Figure 11- According to the text Figure 11 is a correlation between ozone mixing ratio and tidal amplitudes.  If that is correct then please state it clearly in the caption as well.  I don’t understand the relevance of the ozone diurnal cycle mentioned in the first line of the paragraph.  If ozone is driving a diurnal temperature variation I would think that this is related to the presence of solar irradiation during the day and has nothing to do with the small diurnal ozone variation.  The final sentence of this paragraph is also confusing in this regard.

Minor comments and typos:

Line 9 “was located partially” should be “was 

Line 110 “frequency stretch”?  What does this mean?

Figures 2 and 3 are referred to before Figure 1.

Figure 5 – Any comment on why the there is such a very low altitude peak in the earliest data (Feb. 2014?)

Thermal tides in the middle atmosphere are mostly examined on global scales with a temporal resolution of several weeks. Higher resolved studies require local measurements with a cadence of (at most) a few hours. The paper describes the utilization of a microwave radiometer for that purpose. After a short introduction to the measurement principle, the results of overall 5 years of observations are shown. The results focus on seasonal mean tidal amplitudes and their variation between years.

The authors demonstrate successfully that microwave radiometers have great potential for studies of local thermal tides. Nevertheless, the manuscript lacks consistency and coherence. Some of the presented results are not conclusive or should be interpreted in more depth.

General remarks:

1. The authors emphasize the advantage of local quasi-continuous temperature measurements against satellite data that typically need several weeks for full diurnal coverage. On the other hand, they only present monthly or seasonally averaged data. All variabilities are “hidden” in the error bars that include both instrumental errors as well as natural variabilities. I recommend showing at least one representative multi-day case that displays the need for the study of tides with high temporal resolution.
2. The authors claim in the abstract and the main text, an altitude coverage of 25 km to 55 km. a) I assume some kind of weather dependence. Is all data discarded that does not cover that range? b) All plots cover the range between 20 km and 60 km and features below 25 km are described as well without discussion (e.g., line 190). Please make clear which data can be used for tidal analysis. All other data should be omitted or clearly marked in the plots.

Specific remarks:

• l. 55: I recommend providing the coordinates in degrees and decimals (as for the remaining text), but not in degrees and minutes.
• l. 114 and Figure 4: The MR>0.6 is only fulfilled below ~49 km. This is in contradiction to the “used” data coverage of 25-55 km and to the display of data up to 60 km. Please clarify.
• l. 119-125: This paragraph partly repeats Section 2. I recommend shortening it and moving the reference to Figure 3 into Section 2.
• l. 126: Obviously, there is some weather dependence on the data coverage, even if much smaller than, e.g., with lidars. Please, provide some numbers for the typical data coverage (resolved by season or month). What is the minimal accepted data coverage per day?
• Figure 8: If the phase plots belong to the 24-hour period, I recommend aligning them with the amplitude plots in the first row.
• Fig. 7+8 and l. 175: I wonder whether the decrease in amplitude above 45 km is partly instrumental. Fig. 4 shows a FWHM of the AVK of > 15 km for most altitudes. Additionally, AVK and MR values decrease above ~40 km, indicating an increasing contribution by the (tide-free) a-priori. Please discuss.
• Fig. 7+8 etc. and line 178: Line 178 suggests that the error bars show the standard deviations, i.e. they include natural variability and the (presumably much smaller) error from the ASF (line 153). Please mention explicitly what is shown. The natural variability should not be described as “error”. On the other hand, the error of the mean would be a good indicator of the error of the seasonal amplitude or phase values. Furthermore, above 40 km the amplitude error and the phase error are quite high. This makes the amplitude decrease above 40/45 km as well as the constant phase questionable (if not instrumentally induced anyway). Please discuss.
• l. 179: You should not describe features below 25 km if the usable range starts above.
• Fig. 9: Again, the error of the mean may be a better indicator here to differentiate true peaks from variability.
• l. 188: If the (phase) results are contaminated by long-period gravity waves, I would assume the same for the amplitudes. Please discuss. The mean amplitudes of 12-h and 8-h variations are about half of the diurnal amplitude. I wonder why this has such a strong effect on the accuracy of the phase calculation. Please explain in more detail.
• l. 204-208: MERRA-2 is not used in the main section of the manuscript. I suggest either removing these lines or incorporating Appendix A into the main text.
• l. 225-227: I am sorry, but I get confused about which part of the data is trustworthy. I assume that the limitations of data quality are reflected in the error bars. I do not see a difference between 2014/15 and 2016/17. On the other hand, a deviation from the climatological mean would be an interesting result.
• l. 231-242: I am sorry, but I am still confused. You argue that only the year 2015 confirms the Lindzen theory. But a few lines above you mention that the data quality in this period is reduced. Then this data is not a good proof for the theory. I do not understand the further arguments. a) Why is the large variability an indicator of the agreement with the Lindzen theory? I assume that linear theory does not include much variability. b) What is the conclusion from the convective gravity waves? Do they perturb/obscure the tidal signal (technical effect) or do they influence the propagation conditions for the tidal waves (geophysical effect)? My last point: I do not see enhanced variability in summer. Fig. 9 is dominated by year-to-year variability. “Amplitude errors” in Fig. 7 do not vary with season. The yellow line (JJA) is right behind the other ones. Please explain.
• l. 243: A small amplitude by itself does not explain the noise in the phases. The signal-noise ratio matters more, which is influenced by, e.g., natural variability or other waves with similar periods.
• l. 246/247: Do you suggest that the upward propagating tides are filtered at 35 km and new ones are formed above? Please comment. How does this (constant) phase progression compare to other observations? How reliable is the constant phase, given the large phase error shown in Fig. 7?
• l. 250/251: I understand that a detailed comparison (correlation) between ozone variation and temperature tides will be outside the scope of this paper. However, please briefly sketch how the ozone variation (and not e.g. the absorption of solar radiation) affects the temperature.
• l. 274: Please check the height range. For readers focusing on the conclusion, it should be noted that the new capabilities are only available since 2022.
• l. 287: You should not bring up new aspects in the conclusions. I suggest adding some text in the discussion with the pros and cons of radiometer and lidar for observation of tides.
• l. 305: I would interpret this as a slight change in phase by 1-2 h between 40 and 55 km, rather than a "constant phase".