Fig 1.
Data preparation pipeline.
Fig 2.
Geographic distribution of air pollution monitoring stations included in the study (January 2015 – December 2023) – PM10 (dark blue), PM2.5 (orange), (light blue).
The background map is based on OpenStreetMap data [19]. (hypsometric map from WMTS: [20]; Ref. System: EPSG 2180).
Fig 3.
Data imputation and clustering pipeline.
Fig 4.
Dendogram of clusters from lesser Poland Voivodeship.
Fig 5.
Within cluster sum of squares metric for cluster number validation for PM10 (blue line), PM2.5 (orange line), PM2.5/PM10 (green line).
Fig 6.
Calinski-Harabasz Index for cluster number validation for PM10 (blue line), PM2.5 (orange line), PM2.5/PM10 (green line).
Fig 7.
Silhouette for cluster number validation for PM10 (blue line), PM2.5 (orange line), PM2.5/PM10 (green line).
Fig 8.
Davies-Bouldin Index for cluster number validation for PM10 (blue line), PM2.5 (orange line), PM2.5/PM10 (green line).
Fig 9.
PM10 station cluster distribution (hypsometric map from WMTS: [20]; Ref. System: EPSG 2180).
a) PM10, daily, k = 7. b) PM10, yearly, k = 7. c) PM10, daily, k = 16. d) PM10, yearly, k = 16.
Fig 10.
PM2.5 station cluster distribution (hypsometric map from WMTS: [20]; Ref. System: EPSG 2180).
a) PM2.5, daily, k = 7. b) PM2.5, yearly, k = 7. c) PM2.5, daily, k = 16. d) PM2.5, yearly, k = 16.
Fig 11.
PM2.5/PM10 station cluster distribution (hypsometric map from WMTS: [20]; Ref. System: EPSG 2180).
a) PM2.5/PM10, daily, k = 7. b) PM2.5/PM10, yearly, k = 7. c) PM2.5/PM10, daily, k = 16. d) PM2.5/PM10, yearly, k = 16.
Fig 12.
PM10 daily, k = 7 data distribution within clusters.