Review "High Resolution 3D Winds Derived from a Newly Developed WISSDOM Synthesis Scheme using Multiple Doppler Lidars and Observations"

Summary:

The study introduces a new version of the WISSDOM algorithm. In comparison to the previous version, multiple wind datasets can be incorporated as independent input sources. In addition, here, Doppler Lidar data is used instead of Doppler radar data. This allows the algorithm to operate in clear-sky conditions. With a case study, the WISSDOM results are compared to wind measurements and sensitivity tests are performed on key features of the algorithm.

General remarks:

The overall goal of the study could be presented more clearly. The authors present a new version of the WISSDOM algorithm; however do not compare it to the performance of the old version. There is also very little discussion about the change from Doppler radar to Doppler LIDAR data (issues, benefits, etc.). Some aspects of the algorithm could be explained more clearly – the goal should be that the algorithm can be reproduced independently with the information given.

The validation of the control run is performed with sounding data and LIDAR quasi-vertical profiles. As both sounding data and LIDAR data are used in WISSDOM, this is not an independent verification and rather reflects on the importance of these data sources in combination with the others.

 The sensitivity tests do not present a clear conclusion or interpretation of the results. They are only compared to the control run and not to independent data sources; to me it is not clear how this results in a validation of the parameter choices in the control run.

The algorithm looks like an elegant solution to merging multiple data sources into one cohesive 3-D wind dataset. However, the verification and sensitivity tests are not quite convincing yet. The discussion needs a clearer interpretation of the results and it would be nice to compare the control run to the initial version of WISSDOM.

Detailed remarks:

A line-by-line review is in the attached, annotated PDF.

General commnets:

This study presents the advanced WISSDOM synthesis scheme that can provide high-resolution three-dimensional wind fields in clear weather, which was only available in storms before.  The authors examined the performance of the wind retrieval from the advanced WISSDOM in a strong wind case over complex terrain.  The mean biases are generally small, demonstrating that the new WISSDOM is capable of providing reliable wind fields in fair weathers and topography.   Several sensitivity tests are also performed to determine the optimal settings of the new WISSDOM system.  

The manuscript is organized and well-written. The figures are well illustrated.  Yet, there are some issues that need to be clarified and addressed before publication.    

Specific comments 

1. The authors quantify the sensitivity tests and discuss these tests respectively in table 2 and Figs. 6 and 9. Yet, these only state the number of the biases but without further discussions.  Having some more discussion and a summarized recommended recipe in conclusion would be more helpful for the future users, e.g. what observations are the most critical in some cases, how to determine the related RI/VE, and why these suggestions may work (e.g. the height of the PBL, decorrelation of the atmospheric state based on observations, etc.).

• Why is the average bias higher between WISSDOM-lidarQVP than between WISSDOM-sounding?  
• Why is v-component correlation of WISSDOM-lidarQVP particularly lower (down to 0.38)?  

• In addition, what sensitivity tests do the authors suggest readers must do based on your experiences if they are interested in applying the WISSDOM system in different terrains and the weather systems (e.g. typhoons, land-sea breeze etc.)?   

2. Sec. 3.2.1 and 3.2.2   Why use the interpolated observation data in the Cartesian coordinate instead of the original high-resolution observations of scanning Doppler lidars and AWS?  Why use the lidar QVP instead of the high resolution lidar information?   

3. In the evaluation of the control run, the observations used for evaluation are also used in the wind retrieval by WISDOM.  Are there other independent observations that can be used for evaluations?   

4. Sec. 2.1:  For the methodology:

• In fair weather, how are the radar reflectivity and radial velocity constrained in new WISSDOM? Do you use clear echoes or insects? Do you use both Z and Vr? 
• Why do the authors include the vorticity equation in the cost function?  What are the advantages and disadvantages?  
• What are the time steps in the 4DVAR system of WISSDOM?  

Minor comments: 

1. L 60:  It will be helpful to specify what weather systems.

2. L 67-77:  I’m not sure how the VTD/GBVTD/EGBVTD are related to the developing of WISSDOM system or three-dimensional wind retrieval.  It’s not clear.  Please clarify it.  

3. Figure 16f: There is a very artificial band in Fig. 16f, the difference between the control run and the C3.   Please explain the artificial pattern.  

Technical issues: 

1. L 153: The acronym of  the time steps (t) and two time (t) steps are the same.  It’s confusing. 

2. L 290:  gird → grid 

3. Figure 3.  It will be helpful to indicate the location of the analyzed area on the synoptic surface map for readers who are not familiar with Asia .