Annual and Monthly Mean TROPOMI NO2 Vertical Column Densities Aggregated to GHS-SMOD (R2023A) Urban Areas for 2019 to 2024

These datasets provide city-level tropospheric $NO_2$ vertical column densities derived from TROPOMI satellite observations for 2019–2024. Satellite $NO_2$ columns were spatially aggregated to urban clusters defined by the 2023 Global Human Settlement Layer Settlement Model (GHS-SMOD) urban boundaries. For each urban polygon, pixel-level $NO_2$ values were averaged to produce monthly and annual mean columns. This dataset is associated with a manuscript in ACP by Daniel Huber et al. In each of the .csv files, each row corresponds to a single urban polygon from the GHS-SMOD (R2023A) dataset; because GHS-SMOD does not provide city names, we assigned approximate names for cities with a GHS_SMOD population greater than 500,000 using a map API, but these labels may not always exactly match the full urban extent (some polygons include multiple cities), so users should cross-reference the original GHS-SMOD polygons for thorough geographic identification. Annual $NO_2$ .csv file: The column titled "ID_UC_G0" corresponds with the GHS-SMOD R2023A ID for each respective urban cluster. The column titled "POP_2020" is the 2020 population as provided by the GHS-SMOD dataset for each urban cluster. The column titled "NO2_VCD" is the mean TROPOMI $NO_2$ VCD for each urban cluster. Monthly $NO_2$ .csv file: Each column corresponds to a month between January 2019 and December 2024, and each row corresponds with the respective GHS_SMOD urban cluster ID ("ID_UC_G0" column in annual .csv file). $NO_2$ Change Per Year .csv file: The column titled "ID_UC_G0" corresponds with the GHS-SMOD R2023A ID for each respective urban cluster. The PercentChangePerYear column represents the annual percent change in $NO_2$ taken directly from the time-slope coefficient from a linear regression with time as the predictor and fixed effects for calendar month to control for seasonality. The column titled pValue represents the p-value from the linear regression of the de-seasonalized percent anomalies.