-General comments

The manuscript describes a PM10 source apportionment study in Belgrade. The methodology is briefly described and the factor/sources are presented in detail. It is not clear why the study focuses on PM10 while it is well known that PM2.5 is a better indicator of health impacts. More effort is needed to emphasize the scientific added value of this study compared to previous work and to place it in the broader context. 

-Introduction

The most appropriate citations for PMF are: 

Paatero P. and Tapper U.,1994 Envirometrics; 

Paatero P. and Hopke P., 2009 J. Chemometrics

The objective of the study should be clearly indicated in the introduction.

-Methodology

The authors should explain more clearly in the methodological section the procedure they followed to establishing the number of factors and attributing factors to sources including the model diagnostics and the use of independent external evidence (e.g. results from other studies) used to support their choices.

Also the way in which polar plots and backward trajectories is used should be explained in the methodological section.

-Results and discussion

Important publications regarding the study area should be considered in the discussion of results (e.g. Almeida et al., 2020 Environmental Pollution)

-Conclusions

The conclusion section should be revised to match ACP guidelines for authors

The authors are presenting a source apportionment study based on the PM10 chemical composition data obtained at an urban background site in Belgrade, Serbia. There were numerous source apportionment studies conducted for the city in the past covering the sampling period from 2003 to 2015. Therefore, the presented study aims to identify the PM10 using recent filter samples collected between June 2023 and May 2024, and a broader spectra of the tracers. However, it is a solid work, the used approach of collecting 24h PM10 filter samples even not every day for a year and analyse them for the standard compounds/tracers is not very innovative and up-to-date. Additionally, this approach doesn’t allow a precise differentiation of the PM sources. Therefore, further work for better understanding of PM sources in Belgrade will be needed.

2. Material and methods

The section is quite brief and some information/parts are missing e.g. detailed information about filter sampling as duration and frequency, meteorological measurements, FLEXPART and HYSPLIT model description and settings.

2.6. PMF model

The section has to be revised.

• First and only equation of the paper quoted with number “4”.
• Glojek et al., 2024 is not an original paper see Glojek et al., 2024.
• Table S2 not cited at all in the main text.
• Table S4 cite earlier than Table S3 in the main text.
• Sentence 151-151 is not easy to follow.

Questions:

What are the S/N ratios and LODs for Rb, Sr, Cd, Sb and Sn?

How was the PM10 uncertainty calculated?

Was the total variable (PM10) considered as “weak”?

3. Result and discussion

I recommend to present the average values with SD and use same number of decimal numbers in the whole manuscript.

Fig. 2 I recommend to change y axis to μg/m³. Please discuss the difference in “unaccounted” part during the seasons.

Fig. S1 is analogue to Tab. S4.

Fig. 3 and 4 used abbreviations, which should be clarified in the figure’s legend.

Questions:

How was “mineral dust” calculated?

What represent the whiskers in the Fig. 4?

4. Conclusion

Most of the text is rather summary than conclusion(s).