The authors constructed a monthly mean sea level and dynamic ocean topography dataset using data from CryoSat-2 observations. The main differences from previous studies (but not all) are that the data construction focused on intra-month variability, the handling of data bias in the sea ice region and open water, and the validation using time series data from mooring system observations.

Correlation coefficients and RMSDs are calculated using the CPOM DOT as a comparison, but the differences with the CPOM DOT do not lead to the conclusion that the authors' product is truly better. In particular, the sensors in the mooring system that acquire in-situ data are discrete in the vertical direction, and the steric height based on the linearly interpolated vertical profile is questionable. Also, the baroclinic Rossby radius is shorter than the L=300 km used in the first step of DIVA, which may be too much smoothing. Other comments are presented below.

------------------

Major comments:

1)  Mooring data1

The authors state that they are generating sea level data for the entire Arctic Ocean, but the only mooring system data used as validation are those installed in the Fram Strait and Laptev Sea. For example, WHOI has deployed  BGOS mooring systems with MMP in the Canadian Basin. If time series data are important, then all available mooring system data should be used.

2)  Mooring data2

The authors calculate steric height using vertically discrete water temperature and salinity from mooring system observations. In the Arctic Ocean, where tilt pressure structures dominate, it is questionable whether the authors' method can correctly determine Steric height. Do the linearly interpolated vertical profiles of temperature and salinity reproduce the results of CTD observations made at the same time?

3) Reference ellipsoid (line 82  "e.g., WGS84".)

  From Figure 6d, I see that you used WGS84 instead of TP ellipsoid. Please specify somewhere that WGS84 was used in this study, not "e.g."

4) Sea ice concentration (line 107)

Isn't the 15% sea ice concentration a threshold to avoid so-called pseudo sea ice that misinterprets water vapor as sea ice, for example in the Bering Sea in summer, and doesn't it need to be 15% in the Arctic Ocean? Wouldn't the results be the same if this threshold were set to, say, 5%, 10%, or 20%? Isn't it usually the Waveform, for example Pulse Peakiness, that determines if it is sea ice or sea surface?

5) ADCP (line 180)

ADCP velocity data is averaged in the upper 50 m. Shouldn't the velocities within the surface Ekman layer be excluded? Also, there must be a momentum flux due to sea ice movement, so shouldn't the surface still be excluded?

6) High-pss filter(line 205)

Is there evidence that the high-pass filtered data is valid? If no, should it be excluded from the validation data?

7) Specific volume anomay

The reference depth the authors used is 400 dbar. Why? It is the depth of upper Atlantic Water. R. Kwok and other scientists used about 750 dbar (deepest depth of ITP). 

8) Figure 2b

The INSET PANEL in Figure 2b is difficult to understand. How about color-coding the grid points by AWI and RADS?

9) FIgure 4 inset panel

The color of the inset panel in Figure 4 indicates the number of crossovers, which basically increases as you go to higher latitudes since these are polar-orbiting satellites.

If this inset panel is independent and the color of each season indicates the difference of η at the crossover point, it is easy to understand where and in which season there is a difference.

10) Spatial resolution

Is the resolution setting just following the CPOM DOT, or if you want to differentiate yourself from CPOM, is there some strategy to change the resolution? At the moment, it is no different from CPOM except in the Siberian Sea.

11) lines 358 & 360

I don't understand it because there is no detailed explanation of where the 4.2 cm and 8.2 cm came from.

12) Local gradients between 7W and 4W (line 462)

The authors used L=300 km, so they just applied too much smoothing. Why not take into account the bathymetry and reduce the value of L if there are steep velocity changes in the horizontal direction, for example?

13) Figure 9a

I guess relatively low correlation is due to incorrect steric height based on linear interpolation.

------------------

Minor comments:

Line 82 : e.g, --> e.g.,

Line 129: CPM --> CPOM

Table 3: The table shows FES2014, but the caption describes it as FES2004.

In this work, the authors assess a new Arctic-wide gridded dataset of sea surface height and geostrophic velocity at monthly resolution during the period 2011 to 2020. This dataset was generated using Cryosat-2 observations from two products (RADS and AWI). The authors describe how the gridded altimetry-derived variables are produced and show results from comparisons against available in situ measurements (moorings) and an independent state-of-the-art data (altimetry). The seasonal cycle emerging from the final monthly maps is finally discussed.

Overall the paper is certainly of interest to the Arctic community. There is a need to have a unified data set for the ice-covered region and for the ice-free region that is validated properly. The authors provide a description of RADS and AWI products as well as how the two products are made consistent and homogenous before gridding. The new gridded data set is validated, but only in the Fram Strait and Laptev Sea.

Overall the paper is well written with a clear rationale. My feeling is that the processing and validation part is rather limited. Altimeter data retrieval in the ice-covered region is very challenging due to presence of ice that perturbs radar echoes. It is stated that AWI product takes care using a customized processing. The reader expects more convincing analyses of RADS and AWI before their merging, with some statistics about their quality in the whole Arctic area. Authors only show one profile at certain date (Figure 3) that cannot be representative of a ten year period.

I have also a remark about the velocity geostrophic computation at the surface. It is well known altimetry data are noisy and the accuracy of the slope estimate from along-track SLA is strongly impacted. It might dominate the errors in estimating ocean currents because a simple finite-difference of the along-track SLA acts as a high-pass filter. This effect can be mitigated, see Liu, Y., Weisberg, R.H., Vignudelli, S., Roblou, L. and Merz, C.R., 2012. Comparison of the X-TRACK altimetry estimated currents with moored ADCP and HF radar observations on the West Florida Shelf. Advances in Space Research, 50(8), pp.1085-1098.

Also, the direct comparison of the altimeter-derived geostrophic current velocities with the mooring real current velocities does not account for a wind-driven Ekman velocity component. Why ?  is the wind contribution supposed negligible ?

Having said that, I think the paper deserves publication after the authors reinforce the statistics of the two data sources (AWI and RADS) that are used to generate the new data set. If possible I also recommend to extend the validation to other sites in order to provide the reader with a more complete picture about the accuracy of the altimeter-derived sea level and velocities against in situ measurements.