The atmosphere,
under observation.
Investigating atmospheric optical depth, UV absorption, seasonal variability, and Earth system relationships across Punjab.
Aerosol index is increasing over the selected record. This is an area-averaged satellite signal, not a measurement of ground-level pollution or proof of a cause.
Aerosol index through time
Aura OMI · OMAERUVd v003 · Punjab study box; exact coordinates not supplied in CSV
Area-averaged monthly retrievals. Annual values use available months; coverage is listed below.
Season-adjusted trend and 95% confidence band
Supplied OLS fit, deseasonalized monthly values and pointwise 95% confidence band
Monsoon versus winter
Winter Dec–Feb · monsoon Jun–Sep · means of valid selected months
Two instruments, one winter signal
AOD (MODIS) and Aerosol Index (OMI) both rise significantly in winter, and neither changes significantly in the monsoon. Supplied monthly correlation: r = 0.34.
Consistent with growing smoke and pollution in winter; satellite data alone can't prove the source.
Monthly climatology
Annual means & data coverage
| Year | Mean aerosol index | Months | Coverage |
|---|---|---|---|
| 2005-01-01 | 0.584544 | 12/12 | Eligible annual mean |
| 2006-01-01 | 0.615672 | 12/12 | Eligible annual mean |
| 2007-01-01 | 0.489339 | 12/12 | Eligible annual mean |
| 2008-01-01 | 0.606058 | 12/12 | Eligible annual mean |
| 2009-01-01 | 0.573986 | 12/12 | Eligible annual mean |
| 2010-01-01 | 0.633793 | 12/12 | Eligible annual mean |
| 2011-01-01 | 0.490882 | 12/12 | Eligible annual mean |
| 2012-01-01 | 0.625224 | 12/12 | Eligible annual mean |
| 2013-01-01 | 0.519504 | 12/12 | Eligible annual mean |
| 2014-01-01 | 0.621219 | 12/12 | Eligible annual mean |
| 2015-01-01 | 0.571600 | 12/12 | Eligible annual mean |
| 2016-01-01 | 0.695566 | 12/12 | Eligible annual mean |
| 2017-01-01 | 0.750534 | 12/12 | Eligible annual mean |
| 2018-01-01 | 0.903813 | 12/12 | Eligible annual mean |
| 2019-01-01 | 0.596224 | 12/12 | Eligible annual mean |
| 2020-01-01 | 0.567030 | 12/12 | Eligible annual mean |
| 2021-01-01 | 0.787041 | 12/12 | Eligible annual mean |
| 2022-01-01 | 0.640657 | 12/12 | Eligible annual mean |
| 2023-01-01 | 0.552723 | 12/12 | Eligible annual mean |
| 2024-01-01 | 0.577537 | 12/12 | Eligible annual mean |
| 2025-01-01 | 0.617438 | 12/12 | Eligible annual mean |
UV Aerosol Index · supplied seasonal findings and cleaning notes
The UV Aerosol Index detects particles that absorb sunlight, mainly smoke and desert dust. Higher = more absorbing haze. It helps separate smoke and dust from other pollution.
OMI on NASA's Aura satellite · OMAERUVd v003 · daily, 1° · area-averaged over the supplied Punjab study box · NASA Giovanni. Exact box coordinates were not included in these CSVs.
Winter absorbing haze is rising.
Supplied findings cover January 2005–December 2025, regardless of chart date filters. Whole year: no significant trend.
OLS 95% slope CI: −0.003 to +0.093 per decade. Autocorrelation-adjusted OLS p = 0.248. Winter also passes the stricter four-season threshold, 0.05 ÷ 4 = 0.0125.
UV Aerosol Index through time
252 selected months · baseline 2005–2025
Highest month: Nov 2020 (1.622). Lowest: Mar 2005 (0.083). Rolling values require 12 consecutive months. Anomalies, fitted lines and confidence bounds are read directly from the supplied CSV.
Yearly averages
Seasonal cycle · 2005–2025 normal
Season by season
| Season | Mean | Sen / decade | p | Seasonal means | Evidence (four-season correction) |
|---|---|---|---|---|---|
| Winter (Dec-Feb) | 0.586 | +0.218 | 0.0004 | 20 | Significant |
| Pre-monsoon (Mar-May) | 0.514 | -0.052 | 0.450 | 21 | Not significant |
| Monsoon (Jun-Sep) | 0.626 | -0.088 | 0.057 | 21 | Not significant |
| Post-monsoon (Oct-Nov) | 0.819 | +0.127 | 0.043 | 21 | Weak evidence |
Post-monsoon: weak evidence (+0.127 per decade, p = 0.043); it does not pass the stricter threshold. Monsoon and pre-monsoon: no significant trend. Winter uses complete December–February seasons, 2006–2025.
Two instruments, one winter signal
AOD (MODIS) and Aerosol Index (OMI) both rise significantly in winter, and neither changes significantly in the monsoon. Supplied monthly correlation: r = 0.34.
Consistent with growing smoke and pollution in winter; satellite data alone can't prove the source.
How we cleaned the data
450 days were missing (NASA fill values). We removed 239 more days (3%) that were sensor artifacts: values up to 34, jumping on alternate days in a repeating pattern linked to OMI's known 'row anomaly'. We used a Hampel filter (more than 4 robust standard deviations from the 15-day median) plus a cap at 5. Every month keeps at least 18 valid days.
| Cleaning sensitivity | Whole-year p |
|---|---|
| None | 0.0003 |
| Cap only | 0.050 |
| Hampel 15-day (k=3) | 0.074 |
| Hampel 31-day (k=3) | 0.119 |
The winter rise is significant under every method; the whole-year trend is not robust.
Sensitivity runs use k=3; the supplied primary clean series uses k=4. Findings are supplied statistics, not recalculated here.
Download daily raw & clean data