General comments

This is an interesting study of multilayer folds in the Ocnele salt mine, Romania, that combines field data with numerical modeling. The conclusions on effective viscosity ratio and development of folds on different scales are supported by the results. The paper is well-written and clear.

I have a few small specific and technical comments listed below.

Specific comments

Line 307: Please indicate the diffuse boundaries in Fig. 4e and 5b. You don’t refer to Fig. 4f and Fig 5d.

Line 392-393: I would put this in the method

Line 406: are these truncations indicated by one of the white arrows? If not, can you indicate them in the figure.

Line 411: Please add a figure reference for “subgrain boundaries appear white in blue halite cores”.

Line 420: where are the fluid inclusions in Fig. 11f?

Line 422: I don’t see fibrous crystals in Fig 11e.

Line 424: Please explain why you refer to Krabbendam et al., 2003 here.

Line 557: subgrain rich grains are uncommon but in line 437 you say abundant presence of subgrains, also visible in Fig. 10d and 11d?

Line 632: Why does this constraint show that the viscosity ratios were less than R=50?

Technical corrections

Line 49: delete one MPa

Line 91: The shape

Line 304: remove folds

Line 307: a locally diffuse boundary or locally diffuse boundaries

Fig. 10a) what are the green lines in the figure? c & d: what do the white arrows point to? There is no (f) at the beginning of the description of (f).

Table 1 – Change Fig. 12 to Fig 13, and Fig. 13 to Fig. 15 Perhaps good to somewhere indicate that Fig. 13d, 15b and 15f are the same.

Line 499: Moreover, the thickness of each layer decreases causing the thickness of the whole D1 domain to decrease

Line 506: represented by the thick layer

Line 506: Please refer to Fig. 13a &b for the distinct small folds.

Line 506: The thin layer in this domain

Figure 15 caption: A_pert/h_ref 0.03 should be 0.045, 0.1 should be 0.15 and 0.3 should be 0.45.

Line 577: In this study, we show that

Line 579: effectively as a thick

Line 607: Refer here to Fig. 9

Kind regards,

Jolien Linckens

The manuscript of Marta Adamuszek and coauthors "Rheological stratification in impure rock salt during long-term creep: morphology, microstructure and numerical models of multilayer folds in the Ocnele Mari salt mine, Romania" describes the fold patterns and microstructures in an underground salt mine in Romania. The study also presents original numerical models of multilayer folding, which aim to reproduce the folding patterns identified on the walls of the mine. 

The manuscript addresses an important process, which is relevant for the general structural geology audience as multilayer folding is not relevant only to salt successions. The authors use state-of-art methods (photogrammetry-based reconstruction of the fold patterns on the walls of the mine, G-irradiation of salt, numerical modeling). My general impression is that there are two themes in the manuscript that are not well linked - the microstructural analysis of the salt rocks and the numerical modeling of the multilayer folding.

The introduction to the microstructures and rheology of rocksalt is long and presents the general overview of the present knowledge about the rheology of rocksalt and methods that allow estimation of rocksalt's long-term flow properties. This introduction also mentions the viscosity differences in evaporitic sequences, in terms of enclosed layers of different compositions (e.g. low-viscosity, potassium salts, or anhydrite). However, the introduction neglects the rheological effect of solid second phases (impurities) in halite-dominated rock salt. This topic was studied for example by Závada et al. (2015). This latter study is particularly mentioned, because Závada et al., (2015) came to completely different conclusions than the authors of the present study. Závada et al., (2015) explained that rocksalt, which is rich in dispersed solid second phase particles, can become weaker (in terms of viscosity) than the pure salts because the pressure solution-precipitation creep rates are much faster on boundaries of the solid phases (impurities) with halite crystals than on halite-halite boundaries. Therefore, in this perspective, how is it possible that the fine-grained material of the darker layers in the folded sequence of the Ocnele Mari salt mine is stronger in deformation than the white (more pure) layers? 

Altogether, the microstructural part of the manuscript fails to deliver the relevant information from the detailed description and quantification of the microstructures. In the results part and in the discussion, the outcomes shrink to a couple of weakly based statements that aim to bridge the microstructural part to the numerical modeling of the multilayer folding: 

1) the dominant deformation mechanism is grain size-sensitive solution-precipitation creep (which justifies for the Newtonian viscous num. model), 

2) the solid-inclusion rich layers (dark layers) are more susceptible to the grain size-sensitive flow,

3) the deviatoric stresses were very low.

First, you clearly see that both dislocation creep and solution-precipitation (SP) creep were active at the same time (see Fig. 11d, where white parts of grains are full of subgrains and locally mark the "necked" domains of the long grains) - so how clearly dominant was the SP creep? Can we estimate the relative contribution of dislocation creep and SP creep? 

Second, small grain size and small differential stress supporting the SP creep should be linked to lower viscosity in contrast to the coarser-grained (white salt) equivalent. So - why is the dark salt layers more viscous than the white layers??? 

Third - the deviatoric stress estimates span from 0.4MPa to 4Mpa in the same thin section - so the low deviatoric stresses (below 1MPa) are simply unfounded and the high-stress estimates can not be discarded simply by attributing it to the older (detachment shear) deformation event.

There are other problems with the structural and microstructural description;

1) The text explains how grey, dark, white, and yellow-brown layers are folded together, however, only white and dark layers are visible in the black and white images of the fold patterns (Figs. 4 and 5). 

2) The total amount of insoluble second solid phases is not indicated anywhere in the manuscript (only the relative contents of phases in the residuum and in the Conclusions that is somewhat below 10 wt.%).

3) Since two phases of deformation are recorded in the folded sequence, it is difficult to judge on the meaning of the subgrain size piezometry. Nevertheless, since the results of the piezometry is not used for implementation of the boundary conditions of the numerical model, it could be simply removed.

The microstructural part should try to address and discuss the microscale effects of solid second phases in the developing folds and rates of material exchange between the white and dark layers. For example - I see that in Fig. 11e, a single crystal of halite is able to distort a thin layer of dark salt (formed by flakes of clay and halite) - therefore, locally, it is clear that halite is mechanically stronger than the dark particle bearing salt. This is to say that the microscale effects need to be properly discussed, even though it seems (from macroscopic fold geometry) that the dark layers are stronger. 

The numerical modeling part is presented in a somewhat more coherent way and requires another lengthy introduction (section 4 - presented after "geological setting"), which can be reduced in places (see the annotated pdf for suggestions). It seems that the results of the numerical modeling are supported by the validity of the thin single layer fold estimates using FGT toolbox (Figure 9) and the complex multilayer modeling. By presenting viscosity estimates by analysis of fold geometry of the two different models (Fletcher and Sherwin (1978) and Schalholz and Podladchikov (2001)), one should also explain, how are they defined.

Could the numerical modeling part try to provide clues for how would the microscale material transfers (local velocities of solution-precipitation controlled by grain size and solid inclusions) affect the geometry of folds and the viscosities? 

Since the viscosity contrast of 10-15 is quite high, would it be possible to estimate the bulk viscosity of the multilayer in contrast to pure salt on the basis of the modeling? This would be potentially interesting for the salt tectonics community, where people are modeling the development of salt structures (e.g. by numerical or analog modeling).

I suggest for the reorganization of the manuscript, sections 1 and 4 can be merged and shortened, the microstructural part needs to be complemented and related discussion improved. I provide a list of comments and suggestions in the annotated pdf that I hope will help the authors to improve the manuscript.

Prokop Závada

References:

Bruthans, J. et al., 2006. Holocene marine terraces on two salt diapirs in the Persian Gulf, Iran: Age, depositional history and uplift rates. Journal of Quaternary Science, 21(8): 843-857. https://doi.org/ 10.1002/jqs.1007

Závada, P. et al., 2015. Impact of solid second phases on deformation mechanisms of naturally deformed salt rocks (Kuh-e-Namak, Dashti, Iran) and rheological stratification of the Hormuz Salt Formation. Journal of Structural Geology, 74: 117-144. https://doi.org/ 10.1016/j.jsg.2015.02.009