This manuscript presents a valuable effort to model global-scale trace metal PM₁₀ concentrations and quantify their changes and the associated health impacts during the COVID-19 period. My primary concerns pertain to the emission inventory compilation and the modelling framework of this study. While the statistical indicators (R, RMSE, and the slopes) in Figure 1 suggest the robustness of the modelling approach, I believe the manuscript could still be strengthened in methodology and the clarity of presentation in related sections to enhance its comprehensiveness and utility to the broader research community:

1) The model–observation comparison appears to rely predominantly on data from industrialized regions (e.g., China, the United States, Europe). However, as this is a global-scale assessment, and since the manuscript discusses PM trends in regions such as South America, Sub-Saharan Africa, Russia, and Australia, it would be beneficial to incorporate observational sites from these less-represented regions into the evaluation to see how the model-obs statistics could be affected. Many of these areas are less industrialized compared to China/U.S./Europe- it would be interesting to see how the model performs in such regions where emissions are dominated by natural instead of anthropogenic sources. Doing so would help address potential biases stemming from a mid-latitude Northern Hemisphere focus, making the model more convincing. While I understand the scarcity of observational sites in these regions, incorporating even a small number of additional locations might offer valuable insights and enhance the credibility of the study’s global-scale conclusions.

2) While the primary focus of the manuscript is on quantifying changes in ambient trace metal PM levels before and during the COVID-19 period, the spatial distribution and source attribution of trace metals on a global scale is itself a critical contribution. I feel like that the current presentation, which primarily includes global mean concentrations and spatial maps, could be expanded to provide more in-depth diagnostics and benefit the broader atmospheric chemistry & biogeochemistry community. For example, what percentage of total emissions for each trace metal is attributable to anthropogenic sources versus natural sources? How do these proportions relate to the observed changes between the two study years? Such analysis would offer a valuable complement to the discussion of meteorological drivers. Including a breakdown of anthropogenic vs. natural contributions could also clarify the degree to which observed changes are emission-driven.

3) Natural emissions, especially from soil dust, may constitute a significant fraction of total emissions for some trace metals. In relation to Supplementary Text 2, could the authors clarify the basis of the “average mass concentration of each element in soil”? Was this value derived from global crustal averages, or did it incorporate land-use and soil type variability? This is important, as trace metal concentrations (e.g., Pb, As) can vary substantially depending on local conditions, particularly in urban and industrialized settings where historical contamination is present. To increase transparency and robustness, I suggest: a) Listing the specific mass concentrations used for each element in the inventory; b) performing a sensitivity analysis to test how variation in these values (e.g., within plausible upper/lower bounds for different soil types) affects modelled concentrations. This would provide confidence that the results are not overly sensitive to potentially uncertain parameters.

Attached PDF