This is a charming paper with very useful messages.  The ability for stakeholders to quickly identify the air pollution situation in different cities is beneficial, as is the ability to see the chronology of air pollution in the cities and how this chronology can rapidly change when successful clean air action is implemented.  The paper is novel in its approach and should be published once the following relatively small comments have been addressed.

The paper should acknowledge more fully that there are very few measurements of air pollution anywhere in the world before the 1950s and hence the model outputs are very unconstrained, especially before the satellite period.

In lines 33-38, the role of production and loss terms of PM should be tightened up. Like all pollutants the concentration is governed by the relative importance of source and loss terms.  The examples given are a little simplistic.  The possible sources of PM should be expanded upon. The one example of the importance of meteorology – “while other urban locations are affected by strong onshore winds which can influence concentrations” should also be further expanded. All locations are going to be influenced by meteorology.

Line 36 “…atmospheric lifetime of a few weeks,…” should be changed to “…atmospheric lifetime of approximately a few weeks,…”  

The manuscript “Visualising historical changes in air pollution with the Air Quality Stripes” describes a data visualization project aimed at making “complex information about air quality trends understandable and engaging for a broad audience”. Overall, the manuscript provides an interesting approach that stems from the well-known “Warming Stripes”, which should help resonating with different audiences given their recognisability.

The manuscript is well structured, as it presents title, authors, abstract, introduction, information about the data, how the images are developed, discussion and conclusion. References seem overall appropriate. All figures are mentioned and described in the text.

The impact these “Air Quality Stripes” may have on the communication with different stakeholders is evident. However, I believe that a "formal testing" phase would benefit the project (and the paper). Although the feedback provided by "colleagues, friends and family" (line 97) certainly provided interesting and constructive suggestions, the involvement of other testers (with different backgrounds and roles) could help making this data visualisation tool even more "accessible and engaging" (line 96) for different audiences, as the authors themselves underline (lines 135-137).



Overall, the work presented in the paper can certainly have an impact on air quality communication, considering also the annotated images available also on the website, that guide the users.