We thank the referee for many useful comments (including suggestions and questions) on the manuscript.  Below are our item-by-item response to the comments from the referee.

Regarding the 50-point moving average, this corresponds to temporal window of 50-second duration — given the 1-second time resolution of the DMSP ion density data.  However, since the DMSP orbital speed is significantly faster than the TID propagation velocity (i.e. from the perspective of the DMSP spacecraft, it is as if the TIDs were almost stationary), assessing the situation in spatial dimension would be more meaningful.  Orbiting at an altitude of ~840 km, DMSP has an orbital velocity of ~7.3 km/s.  Hence, a 50-second time window corresponds to 50 sec × 7.3 km/s = 365 km spatial interval.  The length of this spatial interval is greater than the conventional range for MSTID wavelengths (100-300 km), which means that we are in a comfortable zone for capturing MSTIDs with the 50-point moving average.  Although, as the referee mentioned, some MSTIDs at the far corner (with ~300 km wavelength) might be somewhat suppressed.  In the revised manuscript, we will include this additional specification in order to provide a clearer physical picture for the readers.

In order to avoid the merging between plot labels and numerical parameters, we will modify Figure 3 accordingly.  In the revised manuscript, we will modify Figure 3 to create an extra space between the MLAT label and its first numeric value, to fix the conjoining labels.

In this paper, we are characterizing the TID propagation based on their horizontal phase velocity away from the source region.  This is the case for both the data we analyzed and the comparison with the existing literature.  In the revised manuscript, we will specify this aspect in order to clarify it for the readers.

Regarding the decay of the AGW/TID amplitude (and hence their detectability) at further distances, we believe that there are two factors in play:  One is the spread of the AGW/TID wavefront over a wider area while conserving total energy; and the other is an actual energy dissipation.  These are perhaps the main reason why TIDs from the 4 August 2020 Beirut explosion were not detected in the TECP measurements by Kundu et al. (2021) and Jonah et al. (2021) beyond the immediate area of Lebanon and Israel/Palestine regions.  At further distances, we believe that the detectability of the TIDs from Beirut rests on the much higher sensitivity of HF diagnostics to bottomside ionospheric undulations (in the case of ionosonde observations) and amplitude amplification at the less dense topside ionospheric altitudes (in the case of DMSP observations).  At San Vito ionosonde (1731 km from Beirut), we were still able to identify the characteristic TID signatures.  Moving slightly further away, at Gibilmanna (1988 km from Beirut) and Rome (2199 km from Beirut) ionosondes, we were no longer able to clearly identify the TID signatures.  Thus, if we have to speculate or make a prediction, we believe that the bottomside ionospheric undulations were dissipated at a distance of 2000-2200 km from Beirut.  Meanwhile, for the electron density fluctuations at topside ionospheric altitudes, we believe that they would be dissipated as they crossed the auroral oval into the polar region.  Given the solar local time sector of the DMSP F17 orbit (~18:00 MLT) and Kp = 1+ condition, the auroral oval boundary can be estimated to be at 75° MLAT which was approximately 4800 km from Beirut.  Thus, if we have to make a prediction, the electron density fluctuations at topside ionospheric altitudes probably did not survive beyond 4800 km distance from Beirut.  In the revised manuscript, we will include some additional remarks on this matter.

We thank the referee for many useful comments (including suggestions and questions) on the manuscript.  Below are our item-by-item response to the referee’s comments.

In the revised manuscript, we will correct the mentioned typographical error (as the referee mentioned, it should have been Cervera and Harris, 2014).

Examination of foF2 and h’F variations with regard to this Port Beirut explosion event, as the referee suggested, can be carried out in order to complement the existing data analysis.  If the foF2 and h’F examination yields positive results (i.e. there are discernible patterns associated with the explosion event), we will include graphs/plots of the additional results in the revised manuscript.  On the other hand, if the data examination yields negative results (i.e. there are no discernible patterns found), we believe that the results will still be valuable to the readers — in that case, relevant remarks shall be made in the revised manuscript and we will make the results available as a supplementary material to accompany the paper.

Regarding the formations of the characteristic loop/twist and Y-forking features seen in the ionograms, their underlying mechanisms can be understood based on Munro’s (1953) and Heisler’s (1957) explanations, which we may summarize here. These features are the direct result of a concave indentation in the ionospheric isodensity contour overhead the ionosonde station.  This situation could materialize the when the ionosonde is located under a trough of the TIDs, with two opposing TID crests on either sides.  This setting can be illustrated using the diagram below, borrowed from Munro (1953) with some modifications:

In Munro’s (1953) paper, the indentation in the isodensity contour was constructed using a set of circular arcs and straight-line segments.  The middle part of the indentation may be called the canopy/roof (in Munro’s paper this was referred to as Y3).  The two sides of the indentation may be called the brims (in Munro’s paper these were referred to as Y1 and Y2).  The brims smoothly extend to the ambient parts of the isodensity contour (represented here by the straight-line segments).  Depending on the position of the ionosonde (at ground level) relative to the indentation, return signals may come from the left brim, the right brim, and/or the canopy.  The special geometrical construction of the isodensity contour based on circular arcs and straight-line segments allows a straightforward rule for determining whether or not the ionosonde will receive a return signal in this simple mirror reflection model.  A normal incidence on the isodensity contour would result in a return signal, where normal direction is pointing radially away from the center of the corresponding circle.  Otherwise, there would be no return signal.

The loop/twist feature can occur if the ionosonde position is non-symmetrical with respect to the center of the indentation, and the TID wavefront is tilted at an angle from the vertical.  On the other hand, the Y-forking feature will occur if the ionosonde position is symmetrical with respect to the center of the indentation, and the TID wavefront has little/no tilt (i.e. relatively close to a vertical orientation).  This geometrical configuration is illustrated in the following diagrams, adapted from Munro (1953) with some modifications.

We show five representative ionospheric isodensity contours at increasing heights, numbered accordingly.  Sample rays from the ionosonde are drawn, generally aimed at either the brims or the canopy of the indentation at each contour level.  Rays that are launched toward the left (right) brim are displayed with green (red) color; whereas rays that are launched toward the canopy are displayed with blue color.  Some of these sample rays do not result in a successful return signal for the ionosonde, representing situations where normal incidence is impossible given the configuration.  For clarity, if a ray results in a successful return signal, the location of the reflection is marked/highlighted on the corresponding isodensity contour line.

The diagrams on the left illustrate the configuration that gives rise to the loop/twist feature in an ionogram.  Due to the tilt angle of the TID structure, in this case direct returns from the right brim (marked with red color) can only occur from the lowest set of isodensity contour lines (No. 1-3).  At the higher contour lines (No. 4 and 5), the reflected rays cannot return because the right brim of the indentations is located too far toward the right to allow normal incidence.  This causes the echoes from the right brim (labeled Y1) to appear only at the lower frequencies of the synthetic ionogram shown at the bottom.  The opposite happens with the rays that are directed toward the left brim (marked with green color).  In this case, due to the tilt angle of the TID structure, direct reflections cannot happen at the lowest contour lines (No. 1) as the left brim is located too far toward the left.  Direct returns from the left brim can only happen at the higher contour lines (No. 2-5) as the overhead contour straightens out at higher altitudes.  This causes the echoes from the left brim (labeled Y2) to occur primarily at the higher frequencies in the synthetic ionogram.  Meanwhile, direct returns from the canopy of the indentation (marked with blue color) can only happen at some intermediate altitudes (contours No. 2 and 3) due to highly selective focusing properties of a concave mirror.  This causes the echoes from the canopy (labeled Y3) to appear only within a narrow frequency interval in the middle.  Conceptually, such a clustered/uneven separation of the three types of rays (the canopy and left/right brims) at different altitudes creates the loop/twist feature in the synthetic ionogram.

The diagrams on the right illustrate the configuration that gives rise to the Y-forking feature in an ionogram.  In this case, the ionosonde is simultaneously looking at two different cross-sections of the ionosphere:  one directly overhead (reflection from the roof of the indentation) and the other oblique from the sides (reflections from the two brims of the indentation which are symmetrically positioned).  Conceptually, this gives rise to the Y-forking feature in the synthetic ionogram.

This explanation, based on Munro (1953) and Heisler (1957), only involves a simple mirror model, which ignores ray bending.  However, conceptually this simple mirror model already captured most of the basic essence and it provides good analytical intuition.

Regarding why the Y-forking feature was only observed over San Vito (but not over Nicosia and Athens) after the 4 August 2020 Beirut explosion, our lines of thinking are as follows.  We believe that there are two potential factors: the strength of TID undulations at F1-layer height and the tilt angle of the TID wavefront.  It is possible that when reaching San Vito, the strength of TID undulations at F1-layer (or E-F valley region) height had weakened.  Moving further away from the explosion source, the local AGW/TID phase velocity might also have leveled out to become more horizontal, and the TID wavefront lost some of its tilt angle as a result.  Indentations around F1-layer height and tilt angle in the TID wavefront are important ingredients for the formation of loop/twist features in the ionogram trace.  As such, these two factors could have caused the change from loop/twist features (seen over Nicosia and Athens) into Y-forking feature (seen over San Vito).

In the revised manuscript, we will include additional explanation regarding the mechanisms behind the loop/twist and Y-forking features seen in the ionograms, as well as some remarks regarding plausible reason(s) why the Y-forking feature was seen only at San Vito.

In the revised manuscript, we will also be adding brief description(s) of the symbols/parameters in Figure 1 for a quick reference by the readers.  These additional descriptions will be added in the figure graphics itself and/or in the figure caption.