General comments

This paper addresses an important topic:  how to improve the evidence base to allow the effective targeting of air pollution interventions to improve public health. It includes some useful, relevant information and some interesting examples. However, it reads rather as a series of disjointed sections which aren’t drawn together in a clear narrative, and sometimes appear inconsistent. I think the paper would benefit from some redrafting. In particular, it would be helpful to include an introductory section explaining the main purpose of the paper, what it covers, and the authors’ view of what it adds to the literature already published on this topic. Some information on how the literature cited was selected would be useful, too.

The abstract gives the impression that the paper focuses on the differential toxicity of particulate air pollution; in fact much of the detailed information relates to aspects such as monitoring, modelling, emissions sources, atmospheric chemistry etc. Pulling out some of the conclusions, or specific recommendations for future research, from these sections and including them in the abstract might be useful to the audience. Perhaps the intended focus of the paper is to comment on how improved monitoring and modelling of components / metrics of particulate air pollution could contribute to informing policies and interventions to maximise health improvements?  If so, then some of the information included is perhaps not really relevant.

Specific comments

Controllable vs noncontrollable /anthropogenic vs natural sources: There is some inconsistency in the discussion in different sections of the paper regarding PM from natural/uncontrollable sources. The inclusion in the paper of Table 1, outlining evidence gaps and needs related to the health effects of natural PM, suggests that some of the authors consider that these are priorities for research.  Other parts of the paper seem to regard these as sources to be dismissed. For example, the paper calls for the exemption from guidelines of components from uncontrollable sources, and recommends collaboration with WHO to achieve this. WHO air quality guidelines are health-based, and do not reflect achievability. Instead, the extent to which sources can be controlled through policy or operational interventions is one of the factors taken into account by legislators when developing national (or regional) regulation or legislation. Whether the lack of control over a source necessarily means that it should be exempted from compliance assessments is a topic of debate - there are health-based reasons that might suggest that it should not – and some discussion of these issues could be included in the paper.

As the paper illustrates in Figure 6, categorisation of sources of PM as controllable or not controllable is not straightforward. Land-use and human activities can influence the emissions of biogenic VOCs and the likelihood of wildfires, for example. And the WHO good practice statement on particles originating from sand and dust storms (SDS) (in the WHO 2021 AQG document) includes measures that can be implemented to mitigate exposure. This distinction between natural/uncontrollable and anthropogenic/controllable emissions could therefore be discussed in a more nuanced way in the paper. 

Figure 4: The source of the data underpinning the illustration (in Figure 4) of the contribution of natural components to total PM, and how it varies with PM mass concentration, is not clear. The source of this data should be given, so that readers can access it. (Some locations / regions with high total PM also have high contributions of “natural” PM – arising from sources such as wind-blown desert dust, or wildfires – this doesn’t seem to be reflected in the discussion, or in the figure.)

Targeting interventions: based on toxicity or source contribution?: it is unclear whether the authors’ overall focus is on identifying PM components that are most detrimental to health or identifying local sources that are major contributors to PM mass concentration and should therefore be targeted. Both are important, and both are discussed – but separately. A summary of the different ways in which atmospheric science, monitoring and modelling can inform policy-making and operational decisions would be useful to tie these different aspects together.

Indoor air: There is inconsistency in different sections of the paper in the way that indoor pollutants are addressed. Early in the paper, the authors suggest that indoor air pollution “should be treated as a separate risk factor distinct from outdoor air pollution, akin to contaminated water”. The reasoning which led the authors to this view is not clear:  is it because different policies are needed to address emissions from indoor and outdoor sources, for example? Conversely, later in the paper, considerable emphasis is put on volatile chemical products (VCPs - including cleaning agents and personal care products, which are used indoors) as sources of outdoor organic aerosol. This inconsistency should be addressed.

Differential toxicity: Section 5.1 “Health effects of anthropogenic PM emissions” includes “Our review reveals mixed results regarding the differential health effects associated with different anthropogenic PM components”. How was this review undertaken? What search terms and literature sources were used?  Were recent reports which have reviewed the evidence related to the differential toxicity of ambient PM consulted?  [Examples include (USEPA  PM ISA, 2019; ANSES, 2019; COMEAP, 2022) and the HEI NPACT initiative.] 

As the paper notes, epidemiology using chemical speciation data will be key to investigating which components of PM might be most health-relevant. However, there will be limitations to how far epidemiology, alone, can address this question. If differential toxicity is to be a main focus of the paper (as suggested by the abstract) it would benefit from more discussion of these limitations. The authors note that confounding might occur because of the strong correlation between various PM components. Confounding by other co-emitted or co-located pollutants (eg NO₂, VOCs, SO₂) is likely an equally important issue, which should be mentioned. Such limitations suggest a need for experimental toxicological data (in vitro and/or in vivo) to inform considerations of differential toxicity.

Attributable mortality:  I would recommend making it clearer that all air pollution mortality burden figures are estimates, and are dependent upon the underpinning assumptions and data used (estimated pollution concentrations, exposure-response functions, counterfactuals etc).  The approaches may differ between the different estimates quoted. I would also suggest use of a term such as “attributable deaths” or “an effect equivalent to x deaths” or similar, rather than “premature” deaths:  in public health, “premature deaths” is often used refer to deaths in those aged less than 75 years old.

 Dosimetry of PM within the lung, translocation and causation of health effects:  discussion of these aspects could be more nuanced (for example, only a very small proportion of even nano-sized particles are understood to enter the blood stream).  But I don’t think that this is a main focus of the paper, so an alternative might be to scale these sections back. 

Technical corrections

Line 32:  “about 400 before our era” is unclear. “400 BCE” is more commonly used

Line 154: “To quantify the health impacts of PM, we currently rely on dose-response relationships that link cause-specific mortality” : many authoritative organisations use all-cause/natural cause mortality as the basis of estimates, rather than cause-specific mortality

Line 183 “insoluble particles, such as asbestos or elemental carbon, can bioaccumulate and lead to chronic inflammation”. The more correct term is “biopersistence” or similar (bioaccumulation is more usually used for accumulation of chemicals within food chains, for example bioaccumulation of dioxins in fish species such as salmon)

Line 869 “WHO has set new guidelines to limit PM concentrations to below 5 μg m^-3.” This should specify PM_2.5

Footnote 1: For this audience, I think the formal definition of PM₁₀ and PM_2.5 should also be included .

Some of the referencing needs to be checked.  For example:

• Line 102 “WHO, has recently updated its air quality guidelines to propose a much more stringent limit value of 5 μg m^-3 (Who)” – this reference is not listed
• Line 246 (Pope Iii et al., 2002)

This long manuscript proposed opinions about how are advances in aerosol science informing understanding of the health impacts of outdoor particulate matter pollution. It’s a very detailed review paper systematically introduced the development of aerosol pollution and corresponding control demands, summarized the globally monitoring or modelling and discussed the roles of PM components on human health. Some improving comments are suggested for considering as follows:

General Comments:

Key objective of this article was to introduce the concept of using specific PM components as metrics for health assessments in addition to total PM mass, and for reevaluation of air quality guidelines. However, the health effects of PM constituents had been widely cognized and investigated either by toxicology or epidemiology studies, so the known and unknown of this issue might be key contents and suggested in this review. There is still road for connecting specific PM components independently with health effects by reliable epidemiological evidences and toxicological mechanisms clearly, qualitatively and quantitatively. Moreover, PMs are chemical and biological mixture, their combined health effects result in the total PM mass exposure effects, surely not only additive by individual components. Furthermore, the aerosol pollution varied spatially and temporally, the moving personal exposure also varied spatial-temporally, how could the PM monitoring serve the health risk assessments more helpful? Since there are both primary and secondary aerosols from both natural and anthropogenic sources, not all components are known and could be monitored in the complicated atmospheric PMs. All these facts should be considered in current discussion.

In the key section of PM components, the list of known PM components to be monitored could be showed. In the section of PM sources, not all sources were covered and suggested to complete.

Particle size is a very important parameter influencing PM health risks, which is also related to components and sources, but was not considered much in this paper.

Finally, a question is suggested for consideration: How to control the specific harmful components in the particles selectively?

Minor comments:

Line 102: Who should be WHO. Check overall similar typos.

More figures than words are suggested to show the opinions.