In general this is an interesting contribution to the problem of black ice forecasting where where a system dynamics modelling is used, rather than a more traditional physcial based model that solves the energy and mass balances.

Generally, the paper is well-written, but the I recommend moving section 4.1 from the discussion to the method.

The use of the unit for the amount of black ice is confusing. The paper both uses g/m2 and g/m3. To me only g/m2 makes sense and this the most common way to express the amount of snow/ice contaminations on roads.

The amount of predicted black ice (presuming the correct unit is g/m2) seems unrealistically high, for the predictions made at 18^th and 19^th of December. (50 000 g/m2 equals 50 mm of black ice) Whether the predicted amounts are high, low, realistic/unrealistic are not sufficiently discussed.

Please find more detailed comments attached.

By mixing System Dynamics and GIS, this study overcomes the weakness of System Dynamics, which is challenging to use spatial information. Since Black Ice Prediction in this paper was performed on various scenarios using spatial and meteorological data, so it is judged to be differentiated from traditional studies.



The paper is well written but needs to be improved further. Additional considerations should be given to the following.



1) The units must be unified. For example, g/m3 and g/m2 are virtually the same.

2) The map in Figure 4 should also show the total length of the test road, and this is because it does not cover the entire Suncheon Wanju highway, so the scope should be seen more intuitively.

3) The map in Figure 10 should include a background illustration so that the reader can monitor the installation environment of the sensing system. In connection with the modeling results, it would be great if the reader could understand the environment in which the sensor is installed more intuitively.

That is all from my end.

Thanks.