Nice and very extensive analysis done on a set of low cost GNSS stations near Summit Station in Greenland. The authors compare their snow accumulation measurements derived from GNSS-IR to a set of snow stakes and find only small differences between the two datasets. The authors also attempt to quantify surface roughness within the sensing footprint and find a moderate correlation.

I don't have many comments though I wonder if additional details could be included in Figure 8--specifically, if for the precip panels the points could be colored by wind speed, and vice versa. Does that help at all with understanding some of the source for the scatter?

Also, do you have any ideas for why your analysis showed minimal differences between L1 and L2 signals, whereas (as you note) the Larson research group found more substantial differences?

Finally, do you have any comment for what might happen to the reflection data if another melt event were to occur at Summit, as I believe one did in 2019?

The authors have established and analyzed a network of low cost GNSS instruments to measure snow accumulation and surface roughness in the vicinity of the Summit Station in Greenland thought the GNSS-IR technique. The use of GNSS technology for this type of measurement is very interesting as well as the use of low cost instruments.  The experiment was well designed and conducted.

I have just few minor remarks, concerning the presentation of the results. 

1. No information about the type of low cost GNSS used
2. In the Abstract: “show a statistically significant (p < 0.01)”. What do you mean for “p”? To avoid any doubts it is better to define it
3. Line 192: (n=76), I suppose that “n” means “number of observations”, isn’t? So, I suggest to write (with 76 observations), later too.
4. Line 194: (p = 1.4 x10^-9), please write the meaning of “p”, that is used also later
5. Line 295: “(64–65 cm^-1)” ==> “(64–65 cm yr^-1)”