# External calibration checks

These checks compare selected synthetic summaries with published Kolkata evidence. They are **not** a formal goodness-of-fit test because source periods, stations, instruments and aggregation rules differ.

## PM2.5 seasonality

A 100-site evenly spaced synthetic-site sample gives monthly PM2.5 means (µg/m³):

| Month | Synthetic mean |
|---:|---:|
| Jan | 190.4 |
| Feb | 162.6 |
| Mar | 116.6 |
| Apr | 79.4 |
| May | 59.8 |
| Jun | 44.8 |
| Jul | 32.3 |
| Aug | 32.6 |
| Sep | 43.0 |
| Oct | 77.5 |
| Nov | 139.2 |
| Dec | 186.7 |

The Jan–Feb mean is approximately **176.5 µg/m³**. A CPCB committee report lists Kolkata winter PM2.5 at **175 µg/m³**. A published multi-city lockdown-era analysis reports Kolkata winter PM2.5 of **186 ± 123 µg/m³ (2018)**, **198 ± 82 (2019)** and **168 ± 51 (2020)**.

A six-year megacity study reports a Kolkata ratio of **6.98** between the highest and lowest monthly mean PM2.5 concentrations. The synthetic 100-site sample ratio is approximately **190.4 / 32.3 = 5.90**. Both exhibit the intended winter-high / monsoon-low pattern.

## PM10 site-type scale

The final synthetic full-history site-type means are:

| Synthetic site type | PM10 mean (µg/m³) |
|---|---:|
| background | 103.7 |
| residential | 124.1 |
| traffic | 158.9 |
| industrial | 169.8 |

A Kolkata study using 24-hour PM10 measurements during 2003–2004 reported **140.1 ± 43.1 µg/m³** at a residential site and **196.6 ± 88.2 µg/m³** at an industrial site, with the industrial site about 1.4× higher. The synthetic values are lower in absolute magnitude but preserve the same ordering and broad scale.

That same study reported inverse relationships between particulate pollution and meteorological variables including wind and rainfall. In the final synthetic station-0 diagnostics, PM2.5 correlates **-0.381 with wind speed** and **-0.141 with rainfall**.

## Interpretation

These comparisons support **broad plausibility**, not equivalence. The synthetic benchmark spans 2015–2026, uses invented sensor coordinates and source classes, and has controlled stochastic/event mechanisms. It should therefore remain labeled **literature-informed synthetic**.

## References

- Central Pollution Control Board committee report, seasonal PM2.5 variation in the Indo-Gangetic Plain, including Kolkata.
- Bera et al. (2021), *Variation and dispersal of PM10 and PM2.5 during COVID-19 lockdown over Kolkata metropolitan city, India investigated through HYSPLIT model*.
- Singh, V., Singh, S., & Biswal, A. (2021), *Exceedances and trends of particulate matter (PM2.5) in five Indian megacities*, Science of the Total Environment 750, 141461. DOI: 10.1016/j.scitotenv.2020.141461.
- *Seasonal variations and chemical characterization of ambient PM10 at residential and industrial sites of an urban region of Kolkata (Calcutta), India*, Atmospheric Research (2006).