This paper attempts to show that:

1. Open GGCM used in the paper does a good job of predicting the location of the open-closed magnetic field boundary (OCB)
2. The minimum latitude reached by the polar cap expansion during a storm follows the IMF clock angle during periods of rotation of the IMF
3. During times of strongest southward IMF, the polar shifts towards the dayside.

Disappointingly, in my opinion, the paper does not demonstrate these points.

The claim that that the model does a good job of predicting the OCB is not supported by the data shown in the paper. The paper shows histograms of the error of the predicted OCB compared to the observations for each of four storms, and it is readily seen that the histograms and standard deviation are roughly what would be expected for a uniform distribution of error over the range +/-5 degrees in latitude. This indicates to me that the model is essentially giving a random location for the OCB over a 10 degree range of latitudes, a range that is larger than the typical width of the auroral oval. There is some trend for both the model and the data to show the well-known tendency for the OCB to move to lower latitudes with increasingly southward IMF Bz.

The two conclusions about the polar cap location and shape are obtained solely from the model and no attempt is made to consider whether these trends are also seen in the data.  It would be interesting if such trends could also be discerned from the data, and perhaps the authors would consider whether they are able make such a determination.

A further point is that the paper does not adequately describe how the OCB was determined from the DMSP data. Many times the DMSP data shows a clean transition from the plasma sheet to polar rain, but this is far from always the case.  For example, low-energy (<~1 keV) can extend roughly continuously from the plasma sheet to the mantle on the morning side, and, in the vicinity of the cusp, the OCB can lie at an equatorward boundary of precipitation because cusp precipitation is on open field lines.  Enough information needs to be presented so that a knowledgeable person could reproduce the results, Simple saying “spectrograms of ion and electron differential fluxes in a range from 30 eV to 30 keV were inspected to identify the polar cap boundary crossings of the satellites” is not sufficient.  

A minor point:   I recommend not including statements of fact in a paper’s Introduction without a reference.  Examples in the current paper are:

“Convection can also change the shape of the OCB without changing the flux contained in the polar cap.”

“When the polar cape opens up, that plasma leaves the plasmasphere and convects away. Thus, the OCB shape also controls the shape of the plasmasphere.”

“During times of high geomagnetic activity these methods can fail because the precipitation is very intense, clobbering the radars’ return signal.”

Review of " Storm time polar cap expansion: IMF clock angle dependence" by Beket Tulegenov et al.

This paper presents the locations and dynamics of polar cap boundary during the five magnetic storm periods by using OpenGGCM-CTIM-RCM simulation together with the DMSP observations. The polar cap boundary and area are crucial parameters for the energy features in the solar wind-magnetosphere coupling, which are strongly related to solar wind and interplanetary magnetic field conditions. Therefore, the topic is very important. This papar shows the simulation and observation results of IMF clock angle dependence of the polar cap boundary mainly are consistent with each other, eventhough the simulation overestimates the latitude of boundary. Moreover, it shows the MLT of the largest polar cap expansion closely correlates with the IMF clock angle and the simulation can remedies observation limitation in local time. These results give new insights into the dynamics of polar cap expansion during storm time, and this paper is of course suitable to be publish in Annales Geophysicae. However, before it is published, I recommend the authors address the following points:

Major comments:

1. In Lines 91-92, the authors mentioned “the coupled numerical model is driven by observed solar wind and IMF data at Lagrangian 1 (L1)” and “…are obtained from OMNIWeb”. These seem to confuse the readers where the interplanetary data you used in your model are, L1 or nose of bow shock? Please make it clear in the text. Here I assume your input data have been shift to the nose of bow shock from OMNIWeb. If so, have you considered any time delay for the solar wind and IMF data propagating from the bow shock to the dayside magnetopause and the polar ionosphere? As we know the solar wind and IMF need several minutes to reach the dayside magnetopause by crossing the magnetosheath, and need 1-2 minutes to affect the dayside polar ionosphere when the solar wind-magnetosphere coupling happened at the dayside magnetopause. Thus, the time delay is very important when you compare the IMF with your observation and simulation results. I suggest the authors to consider these time delays at least for the delay for the solar wind and IMF crossing through the magnetosheath.
2. I remember that there is a debate for a long time: whether the whole polar ionosphere (convection pattern) immediately responses the IMF variation or need some time for the IMF effect propagating from dayside cusp region to nightside auroral oval through the polar cap. Fear and Milan (2012) argued the convection of magnetic field lines should take a number of hours from the dayside to the nightside after statistically analyzed the formation of the transpolar arcs. Zhang et al. (2015) suggested that cross-cap transit time of the field line is about 1-2 h from noon to midnight in MLT and is 3 h for the convection of the full Dungey cycle by tracing the polar cap patches. Browett et al. (2017) suggested that the timescale varied from 1 to 5 h for the penetration of IMF By into the magnetotail depending on the IMF orientation and solar speed. I think the OCB in different MLT may need different time to response the IMF variations. Thus, we may need to be careful when we compare the OCB at all MLT with the IMF clock angle, such as in Figure 6. I suggest the authors make any clarification of these in the text, because this is related to your main results.

Small suggestions and typos:

1. How do the authors identify the OCB from the DMSP spectrograms? Although the authors cited some references, I suggest to make a briefly interpretation about it for the readers easily following, because the OCB is the main topic of this manuscript.
2. Line 130: "in Fig. 1 through ??" Please check.
3. Page 6, the caption of Figure 1 line 1: There is an extra space in DMSP.
4. Lines 186-190: These sentences are not clear to me. Could you please explain how the northward IMF influence on the behavior of OCB?
5. Line 215: PC-> polar cap (PC). When an abbreviation of term first appears in the manuscript, the term should have a full name first.
6. Lines 219-220: please tone down of the express “five strong geomagnetic storm events”, because there is an event with a minima Dst of -39 nT, and an event with a minima Dst of -65 nT in your events list.

References mentioned above:

Browett, S. D., R. C. Fear, A. Grocott, and S. E. Milan (2017), Timescales for the penetration of IMF By into the Earth’s magnetotail, J. Geophys. Res. Space Physics, 122, 579–593, doi:10.1002/2016JA023198.

Fear, R. C., and S. E. Milan (2012), The IMF dependence of the local time of transpolar arcs: Implications for formation mechanism, J. Geophys. Res., 117, A03213, doi:10.1029/2011JA017209.

Zhang, Q.-H., M. Lockwood, J. C. Foster, S.-R. Zhang, B.-C. Zhang, I. W. McCrea, J. Moen, M. Lester, and J. M. Ruohoniemi (2015), Direct observations of the full Dungey convection cycle in the polar ionosphere for southward interplanetary magnetic field conditions, J. Geophys. Res. Space Physics, 120, 4519–4530, doi:10.1002/2015JA021172.

The end.