In the response to earlier comments the authors used TIEGCM simulation to justify wind field differences between N and S viewing directions cannot be larger than certain values.  I don't think TIEGCM with resolution 100 km? can be a good reference.    TIEGCM electric field is based on empirical convection pattern does not have small structure you would expect to see near auroral zone.    Some additional clarifications are needed.   

When we consider the simple reflection path illustrated in (b) and (c) of Figure 1, it appears that the wind speed fluctuation might be expected to follow the order U_LOS_S > U_LOS_Z > U_LOS_N. As shown in Figures 3b, 3f, and 5b, the model calculations seem to support this expectation. However, it is important to note that the observed results do not reflect the same pattern. This discrepancy may benefit from further explanation in the text. While it is reasonable to assume that reflection has some influence on wind speed, the notable difference between the model calculations and the observed results suggests that it may be difficult to agree that the effect of reflection is substantial.

L70:  “During two geomagnetic storms with visible auroras” > “During two geomagnetic storms on May 10th and Oct. 13, 2024 with visible auroras”

L72-73: “The observations were unaffected by moonlight or clouds, and the interferometer retrieval errors were acceptable.” How did you confirm these?

L139: “The LOS speed is used directly instead of the Doppler shift, assuming a constant background temperature.” This sentence requires revision. I have two comments: (1) To my knowledge, the LOS speed is derived from the Doppler shift measured by the optical interferometer. In this process, the Doppler shift seems more like a "direct" measurement rather than the LOS speed itself. (2) I find it unclear why the assumption of a constant background temperature is necessary. This should be more explicitly addressed in the text.

L140: “After binning different LOS speeds and computing the corresponding scattered light intensity, contaminated LOS speeds are calculated via weighted average, simplifying the wind simulation.” The procedure for "binning different LOS speeds" is unclear to me. Is it similar to pixel binning, a technique used in digital camera sensors?