his research is highly significant as it addresses the coupled challenges of salt-thermal-topography-driven groundwater flow systems. The discussion could be strengthened by connecting this research to related studies and highlighting potential implications. Here are some minor comments:

• In the Abstract, "water table topography" should be revised to "water table undulations."
• Line 35: Should "horizontal variations in the water table" be "vertical variations"?
• Line 38: Revise "thereof."
• Line 144: How do you account for the relationship between salinity and density?
• Line 173: How do you ensure that the model discretization (time step and spatial mesh size) meets the requirements for the Peclet number and Courant number?
• Line 184: The boundary condition for a constant salinity concentration seems unreasonable. While the heat source can originate from deep geologic units, where does the constant salt flux come from?
• Figure 10: The caption font is too small to read.

General Comments

 The manuscript by Galsa et al. presents a detailed numerical investigation into the interaction of topography-driven groundwater flow and thermohaline convection. Usually this process is referred to as “mixed convection”.

The authors use synthetic and real-world scenarios, specifically the Buda Thermal Karst (BTK) system.

I enjoyed reading the paper and the results are scientifically sound, and relevant considering the growing importance of geothermal exploration and understanding deep flow systems.

The use of COMSOL Multiphysics for fully coupled thermal, hydraulic, and solute transport is appropriate, and the modeling framework is clearly described.

There are a few minor issues that should be addressed before publication. These include clarifications on modeling assumptions (e.g., viscosity), more detailed justifications of boundary conditions, lack of fault implementation, and limited comparison to observed data in the BTK section.

Addressing/discussing these points would significantly strengthen the manuscript.

Specific Comments

• Temperature-Dependent Viscosity:

The modeling assumes a constant dynamic viscosity of water (μ = 5×10⁻⁴ Pa·s), despite the fact that the model includes temperature-sensitive density and diffusivity. Since viscosity is also temperature-dependent, especially in geothermal systems, the authors should justify this simplification or consider its impact on the results. It can strongly impact hydr. permeability and hence the resulting flow pattern.

• Boundary Conditions in Table 2:

The choice of a fixed salt concentration at the bottom boundary (e.g., 70 g/L in the base model) lacks justification. Is this value based on real geological conditions (e.g., evaporitic layers)? If not, some explanation is needed regarding its representativeness or role in the sensitivity analysis. This is particularly relevant due to the geological time-scale of the models.

• Boundary Condition Dominance in Figure 2:

Figure 2 and similar reveal that the salt boundary condition strongly influences the flow field at the discharge side (left end) of the model (boundary-controlled). The paper would benefit from a short discussion on how this artificial boundary might bias the results, and whether alternative configurations were tested.

• Implementation of Faults in BTK Model:

Faults are known to play a significant role in hydrothermal systems. I would assume BTK is not an exception. However, they do not appear to be implemented explicitly in the COMSOL model (only layered anisotropic media are used, am I right?). The authors should clarify whether fault zones were neglected, or implicitly represented, and what implications this might have for flow connectivity and thermal upwelling.

• BTK Model – Validation Against Observed Data:

While the authors discuss general agreement with previous studies (e.g., thermal spring temperatures and chemical compositions), there is no quantitative comparison of model outputs with field observations (e.g., temperature profiles, water age data, TDS distributions). Including even limited validation with observed data would enhance the credibility of the results.

Technical Corrections

• did not find anything critical.

Overall very nice study! Thank you for the considerations.

F. Magri