This is very interesting study demonstrating the challenges of simulating the interaction between meteorology and plume emission conditions and should be of considerable interest to air quality modellers.

Thank you to the author's for their detailed responses and the revised manuscript, which is acceptable for publication after technical corrections.

A couple of areas that the authors may consider in their revisions:
(1) In one of the sensitivity analysis the results showed that: "Nearly half (48%) of the predicted plume rise values are less than half the observed values and a large fraction (34%) of the predicted plume rise values are more than double the observed values." - This seems like a rather large dichotomy, which could potentially be explained with available information leading to some insight. The authors should consider exploring this result and briefly sharing what may be the reason(s).
(2) While a strength of this study is the use of observed/measured inputs in the plume rise formulations, which tended to show an underestimate of plume rise and little skill in predicting rise for individual plumes, it could be useful to readers to also see plume rise results for these cases based upon using the modeled (GEM-MACH) inputs. Perhaps this information is in the identified companion paper, however, it would be of value to briefly present these results in this paper so that all the information is in one place. Could an additional line be included in Figure 3 and/or the statistics in Table 3? If these results are added then some text would be important to describe the result. However, ideally, further linking to the companion paper would avoid the need for too much additional revision in this paper to explain what was done.