|The authors have improved this version of the paper somewhat, but there are still a number of problems to be addressed. The most severe problem is that the workflow presented in this paper in no way represents a generalised approach for producing paleobathymetry grids, as many ad hoc fixes used in the 38 Ma reconstructions are custom-designed just for reconstructing this time, especially when it comes to mixing different vintages of rotation models, paleo-age and paleo-depth grids. The current title “A Generalised Approach to Reconstructing geographical Boundary Conditions for Palaeoclimate Modelling” is misleading. The paper would be acceptable if solely focussed on the 38 Ma reconstruction. I therefore suggest a change of the title of the paper to:|
Reconstructing Geographical Boundary Conditions for Palaeoclimate Modelling at 38 Ma
Some specific comments:
There is a mismatch between using the present-day age-grid from Seton et al. 2012, and the paleobathymetry from Muller et al. 2008, since the present-day age grid for 2012 is based on the reconstruction history/rotations of that model, which differs significantly, especially for the Pacific (Manihiki/Hikurangi breakup).
This new workflow uses the Crosby et al. (2006) age-depth relationship, to convert the reconstructed age grid into depth. But the Muller 2008 paleobathymetry grid, which they use for filling gaps in their initial paleobathymetry reconstruction based on the Seton et al. (2012) present-day agegrid, is based on Stein and Stein’s (1992) age-depth relationship, and also has sediments added. So mixing different age-depth relationships and sediment cover (vs. no sediment cover) in the Baatsen workflow certainly does not represent a sensible generic workflow for constructing paleobathymetry grids., even if this works reasonably well for 38 Ma.
Some comments by line:
Line 42: Not sure what "poorly constrained and/or in agreement" means.
Line 62: "transition" in Eocene-Oligocene Transition is capitalised (earlier, line 34, it isn't).
Line 69: "require" is misspelt
Line 87: Seton et al., 2012 isn't a Phanerozoic model.
Lines 90-91: It is insufficiently justified to what extent a paleomagnetic reference frame is better suited for paleoclimate reconsrtructions, even if this is obvious theoretically. How different it is from other reference frames for 38 Ma in terms of paleolatitude? Is there a reference for this?
Lines 114-136: This entire paragraph is hard to follow.
Line 133: Is there a reference for the robustness for continental shelf margins through time?
Workflow – I couldn’t get this to work very well – it crashed after producing Figure 5. Exporting the files exactly as needed for the matlab script is extremely tedious, and not suitable for any sort of generic workflow.
Exporting files as a png. (step 1 and 2): I can now see why the authors didn't export a netcdf file directly – they are reconstructing polygons (gpml), not a grid, and GPlates can't convert polygons to grids. Nevertheless, this step would be much better completed using GMT for this to become generic and easier to use (e.g. exporting reconstructed polygons, and using a combination of grdmask and grdmath steps).
For part 2., there are numerous files needed, with specific colours (from the Apple palette, which Windows users will not have!). These files seem to be specific for 38 Ma (their reconstruction time) – again this is no generic recipe.
Step 5: I don’t really follow what is happening with rotating the grids here, and I found it convoluted to get this step to work.