See attached.

Subsurface floats in the Filchner Trough provide first direct under-ice tracks of eddies and circulation on shelf

by Jean-Baptiste Sallee, Lucie Vignes, Audrey Miniere, Nadine Steiger, Etienne Pauthenet, Antonio Lourenco, Kevin Speer, Peter Lazarevich, and Keith W. Nicholls

--- Summary ---

Sallee et al present a fascinating study of trajectories and watermass transport implications from 7 RAFOS-enabled profiling floats on the eastern Weddell Sea continental shelf of Antarctica (Filchner Trough region). The floats operated in a mode similar to floats in the global Argo array but with a 5-day profiling interval (daily profiling in the summer), drifting at 250m or 400m between profiles and receiving  RAFOS positions up to 4 times daily, with total float lifetime ranging from 1 to 5 years. Acquiring this dataset is in itself an impressive accomplishment.

--- General Comments ---

The manuscript's presentation of the scientific context, the float experiment and results, and the discussion of implications is well done and informative. Figures are generally clear. In particular, the floats are simultaneously able to illustrate flow pathways and the evolution and variability of layer temperature and thickness along those pathways. Overall, not a lot of changes are needed before publication, although I do have a number of suggestions and questions which the authors may choose to address.

Although the discussion of pathways and transport mechanisms is interesting and informative, there has not been an attempt to make quantitative estimates of volume or heat transport or diffusive watermass transformations. Clearly any attempt to do this would be subject to some guesses at unmeasured quantities (such as flow width or duration), but even very rough values would be helpful for comparing these observations to numerical models or shipboard or moored measurements.

In one especially interesting and curious feature of Fig.8 (the comparison of a float's measured trajectory and temperature with the progressive-vector trajectory and temperature from a mooring), the temperature at the float made a sudden change immediately before reaching the mooring. Is there any suggestion of how this might have happened (either through diffusive heating or non-Lagrangian movement of the float)? The change brought the float measurement up to the mooring's temperature just before the closest approach, but the warm temperature had been present at the mooring considerably earlier. Later, the warm patch seems to leave the mooring but the float continues to follow it.

The authors make the intriguing suggestion that ISW could block the heat flux of mWDW toward the FRIS. Is this energetically possible? (Meaning, what is the source of energy maintaining the potential energy barrier of the dense water and accompanying geostrophic front?) Or is this really a statement about a different cause? For example, that the atmospheric cooling that produces the ISW (say close to the ice face) removes all heat before it can reach the cavity? Or simply that the amount of on-shelf mWDW transport and heat flux here are small relative to the West Antarctic shelf?

--- Specific Comments ---

The bathymetry in Fig 1 is difficult to make out against the temperature shading. Would a light color work better? Also some bathymetric contour labels would be helpful. The bathymetry is presented in a more readable form in other panels, but there is no definitive label of any isobath and the colorbar is too finely graded to help. How about making one isobath (e.g. 300m or 400m) thicker than the others to provide at least one clear reference (along with the fixed 100m contour interval).

Also on the subject of bathymetry, I'm curious whether the floats drifting at 400m ever got stuck on the bottom during their drift phase. From counting the number of 100m-interval contours on the shelf east of Filchner Trough (e.g., in Fig.3), most of the shelf appears to be shallower than 300m. However, the depth-time plots in Fig.4 show water mostly between 400m and 500m on the shelf. Does the float-inferred bathymetry actually match the mapped contours? Or is the shallowest contour shown not the 100m one?

I suspect that a T-S diagram would help clarify some of the discussion of watermasses and layers. And coloring by or otherwise indicating variations in latitude or water depth might be a good way to illustrate the location, density range, and T-S characteristics of the ISW and mWDW watermasses and the front between the two.

In addition, it would be useful to mark some candidate isopycnals (e.g., important ones for mWDW and/or ISW) on the depth-time plots (figs 4,5,7).

Ending the Fig.4 trajectories on the shelf but continuing the timeseries off the shelf is confusing (and it is difficult to tell whether this is a choice that has been made deliberately or due to a lack of GPS data). Is there at least an approximate trajectory or direction of the floats that left the shelf (12682 and 12703) that could be indicated? They must have reached the surface at some point to send the data back.

--- Minor Comments/Typos ---

l.9 typo: Pobservations

l.202-210. The description of the southern float trajectories is interesting in it's comparison to the behavior of the eastern floats but should also re-iterate the fact that these floats parked at a different depth (250m) than the others (400m).

Fig.3 caption lists different colors for the floats than the legend in the figure. Which is correct? (I'm guessing it's the legend.)

Fig.4 mentions float 12672 but probably means 12682

"Hovmoller diagram" is usually not what a timeseries of vertical profiles is called (at least in oceanography). Not sure whether there's a definitive definition, though. "Depth vs. time" or "profile timeseries" seems more common for the types of plots shown in Figs 4, 5, and 7. Usually Hovmoller is reserved for horizontal distance vs. time (and in particular as a way to emphasize wave propagation that might be seen in an animation of 2D structures changing in time).

Fig.6 shows temperature on the 27.75 isopycnal (mWDW). Can this isopycnal be added to Figs 4+5?

Fig.7 shows ISW layer thickness. Mention the definition of this layer as water colder than -1.9C in the caption.