See attachment.

This manuscript provides a very thorough technical description of the data collected by the DOE buoy lidars off the California coast.  It will serve as an important reference for anyone using both past and future data sets collected by these instruments.  It also provides some basic analysis of the observations, including an evaluation of M-O similarity theory, which is interesting.  Other than some minor comments below, the paper is ready for publication.

Line 23.  What does “statistically averaged data” mean?  Is “statistically” necessary?

Lines 41-42:  The use of “1-year” and “annual” in this sentence seem redundant.

Line 48:  Do the publicly available NYSERDA data pre-date the DOE buoy lidar data?

Lines 223-226.  Given all of the problems with the Windcube IMU described here, I don’t see the point of even showing any of the wind data derived using that IMU.  Unless perhaps the authors are implying that data from all Windcube lidars would have these same problems.  But if this is a one-off bad instrument, it doesn’t make sense to me to show the data (e.g. figures 14, 15, 16).  At a minimum, the authors should state their reason for showing the data if it is from a broken instrument.

Lines 351-354: “During the Morro Bay deployment, the maximum and average Hmax/Hs were 2.5 and 1.6, respectively, when including questionable data, and were 2.2 and 1.6 when considering good data only. Based on theory, the expected values are 1.7 and 1.6 when including questionable data, and 1.7 and 1.6 when considering only good data. This indicates that the data follows the expected theory.”  I’m not sure I follow this.  The difference between 1.7 and 2.2 seems substantial (a 30% difference), so how does this indicate that the data follows expected theory.  How far off would it need to be to be considered not to follow theory?

Figure 9b.  The y-axis label says COD, should it be COT?

Line 662. “Obukhov length”