Temporal Analysis
Separate long-term patterns from seasonal cycles and interannual variability.
Research summary
2005-01-01 — 2025-12-31 · NASA Giovanni / MERRA-2 / NASA Giovanni / MODIS Aqua / NASA Giovanni / Aura OMI OMAERUVd v003
Of the 3 variables studied, 2 show a statistically significant trend between 2005 and 2025: Surface albedo decreasing, Aerosol optical depth increasing.
Surface albedo
NASA Giovanni / MERRA-2 · Dimensionless fraction (0–1) · 2005-01 to 2025-01 · 241 valid months
Between 2005-01 and 2025-01, how much sunlight the land reflects decreased by about -0.000796 Dimensionless fraction (0–1) (-0.373%), roughly -0.000398 every ten years.
Winter readings are higher on average (0.218 versus 0.214 Dimensionless fraction (0–1)).
Largest seasonal anomalies: 2006-02: -0.00986 Dimensionless fraction (0–1); 2020-01: 0.00966 Dimensionless fraction (0–1); 2022-01: 0.00839 Dimensionless fraction (0–1).
Highest annual mean: 2020 (0.216); lowest: 2018 (0.213). Annual means require at least 9 months.
Aerosol optical depth
NASA Giovanni / MODIS Aqua · Dimensionless at 0.55 μm · 2005-01 to 2025-12 · 252 valid months
Between 2005-01 and 2025-12, how hazy the air is (how much sunlight particles block) increased by about 0.0826 Dimensionless at 0.55 μm (13.9%), roughly 0.0395 every ten years.
Monsoon readings are higher on average (0.855 versus 0.476 Dimensionless at 0.55 μm).
Largest seasonal anomalies: 2008-06: 0.423 Dimensionless at 0.55 μm; 2022-05: 0.254 Dimensionless at 0.55 μm; 2011-07: 0.246 Dimensionless at 0.55 μm.
Highest annual mean: 2018 (0.719); lowest: 2005 (0.542). Annual means require at least 9 months.
UV Aerosol Index
NASA Giovanni / Aura OMI OMAERUVd v003 · Dimensionless UV Aerosol Index · 2005-01 to 2025-12 · 252 valid months
UV Aerosol Index: supplied 2005–2025 findings show no significant whole-year trend (+0.047 per decade, p = 0.089); annual p = 0.291. Winter absorbing haze is rising (+0.218 per decade, p = 0.0004, passes p < 0.0125). Post-monsoon +0.127 per decade, p = 0.043: weak evidence, not significant after four-season correction. Monsoon and pre-monsoon have no significant trend. OLS anomalies +0.045 per decade, p = 0.065, 95% CI −0.003 to +0.093.
Monsoon readings are higher on average (0.626 versus 0.586 Dimensionless UV Aerosol Index).
Largest seasonal anomalies: 2012-03: 0.749 Dimensionless UV Aerosol Index; 2018-01: 0.653 Dimensionless UV Aerosol Index; 2006-05: 0.652 Dimensionless UV Aerosol Index.
Highest annual mean: 2018 (0.904); lowest: 2007 (0.489). Annual means require at least 9 months.
How they relate
| Pair | Pearson r | Spearman ρ | Strength | Months used | Best lag, exploratory |
|---|---|---|---|---|---|
| Surface albedo / Aerosol optical depth | -0.133 | -0.191 | very weak | 241 | 0 months, albedo leading aod |
| Surface albedo / UV Aerosol Index | 0.0228 | 0.121 | very weak | 241 | 1 months, albedo leading aerosol_index |
| Aerosol optical depth / UV Aerosol Index | 0.544 | 0.539 | moderate | 252 | 0 months, aod leading aerosol_index |
Surface albedo / Aerosol optical depth move in opposite directions to a very weak degree (-0.133, 241 overlapping months), after removing the seasonal cycle.
Surface albedo / UV Aerosol Index move together to a very weak degree (0.0228, 241 overlapping months), after removing the seasonal cycle.
Aerosol optical depth / UV Aerosol Index move together to a moderate degree (0.544, 252 overlapping months), after removing the seasonal cycle.
A link between two measurements does not prove one causes the other.
Bottom line
Between 2005-01 and 2025-01, how much sunlight the land reflects decreased by about -0.000796 Dimensionless fraction (0–1) (-0.373%), roughly -0.000398 every ten years. Between 2005-01 and 2025-12, how hazy the air is (how much sunlight particles block) increased by about 0.0826 Dimensionless at 0.55 μm (13.9%), roughly 0.0395 every ten years. UV Aerosol Index: supplied 2005–2025 findings show no significant whole-year trend (+0.047 per decade, p = 0.089); annual p = 0.291. Winter absorbing haze is rising (+0.218 per decade, p = 0.0004, passes p < 0.0125). Post-monsoon +0.127 per decade, p = 0.043: weak evidence, not significant after four-season correction. Monsoon and pre-monsoon have no significant trend. OLS anomalies +0.045 per decade, p = 0.065, 95% CI −0.003 to +0.093. These area averages describe relationships, not causes.
ORBIT · 2026-11-15 · Albedo/AOD calculated from observations; UVAI uses supplied 2005–2025 findings
Trend summary
| Aerosol optical depth | Monthly | 0.0395 Dimensionless at 0.55 μm | 0.0826 | 13.9% | 0.00016 | ● Significant | 0.021 to 0.0589 | 252 | ✓ |
| Aerosol optical depth | Annual | 0.0261 Dimensionless at 0.55 μm | 0.0522 | 8.78% | 0.032 | ● Significant | 0.00232 to 0.0582 | 21 | ✓ |
| Aerosol optical depth | Monsoon | -0.00276 Dimensionless at 0.55 μm | -0.00552 | -0.664% | 0.928 | ○ Not significant | -0.0532 to 0.0461 | 21 | ✓ |
| Aerosol optical depth | Winter | 0.0623 Dimensionless at 0.55 μm | 0.118 | 28.4% | 0.00476 | ● Significant | 0.0225 to 0.0952 | 20 | ✓ |
| Surface albedo | Monthly | -0.000398 Dimensionless fraction (0–1) | -0.000796 | -0.373% | 0.00539 | ● Significant | -0.000628 to -0.000139 | 241 | Methods differ |
| Surface albedo | Annual | 0.00015 Dimensionless fraction (0–1) | 0.000284 | 0.133% | 0.721 | ○ Not significant | -0.000582 to 0.00083 | 20 | Methods differ |
| Surface albedo | Monsoon | -0.000575 Dimensionless fraction (0–1) | -0.00109 | -0.51% | 0.0179 | ● Significant | -0.00113 to -0.000107 | 20 | Methods differ |
| Surface albedo | Winter | 0.00206 Dimensionless fraction (0–1) | 0.00371 | 1.72% | 0.0931 | ○ Not significant | -0.000566 to 0.00438 | 19 | Methods differ |
| UV Aerosol Index | Monthly | 0.0474 Dimensionless UV Aerosol Index | n/a | n/a% | 0.0892 | ○ Not significant | n/a to n/a | 252 | ✓ |
| UV Aerosol Index | Annual | 0.0301 Dimensionless UV Aerosol Index | n/a | n/a% | 0.291 | ○ Not significant | n/a to n/a | 21 | ✓ |
| UV Aerosol Index | Monsoon | -0.0881 Dimensionless UV Aerosol Index | n/a | n/a% | 0.0571 | ○ Not significant | n/a to n/a | 21 | ✓ |
| UV Aerosol Index | Winter | 0.218 Dimensionless UV Aerosol Index | n/a | n/a% | 0.000406 | ● Significant | n/a to n/a | 20 | ✓ |
How we calculated this
Subtract each calendar month’s long-term mean to remove the seasonal cycle. Monthly Mann–Kendall scores and tie-corrected variances are summed across calendar months; Sen’s slope is the median within-month pairwise rate. Annual means require 9 of 12 valid months. Monsoon is June–September; winter is December–February, with December assigned to the next winter year. Seasons require all their months. Annual and seasonal series use ordinary Mann–Kendall and Sen’s slope. Trends require 10 valid points. Significance is p < 0.05. Percent change uses the first five years’ mean. Nearby monthly observations are not fully independent, so an annual check and OLS cross-check are shown.
Season-adjusted trend evidence
Sen slope and seasonal Mann–Kendall · 2005-01-01 — 2025-01-01
Methodology About this trend test
Monthly slopes and Mann–Kendall scores are calculated within calendar months, then combined. The confidence band is a pointwise 95% large-sample interval for the OLS mean fit through deseasonalized observations. Serial correlation is not corrected.
Moving-average analysis
A trailing window over available monthly observations
Linear regression
Ordinary least squares fits the deseasonalized observations and provides the displayed pointwise 95% confidence band.
For monthly records, the median rate is calculated from observation pairs within the same calendar month.
A two-sided seasonal, tie-corrected test sums calendar-month scores and variances with continuity correction. Serial-dependence adjustment is not applied.
Monsoon versus winter
Means of valid selected observations · winter Dec–Feb · monsoon Jun–Sep
Annual variability & coverage
| Year | Mean | Months | Year-to-year change | Coverage |
|---|---|---|---|---|
| 2005 | 0.214084 | 12 | — | Complete year |
| 2006 | 0.213037 | 12 | -0.489% | Complete year |
| 2007 | 0.213875 | 12 | 0.393% | Complete year |
| 2008 | 0.214103 | 12 | 0.106% | Complete year |
| 2009 | 0.213273 | 12 | -0.388% | Complete year |
| 2010 | 0.213641 | 12 | 0.172% | Complete year |
| 2011 | 0.213532 | 12 | -0.051% | Complete year |
| 2012 | 0.215346 | 12 | 0.849% | Complete year |
| 2013 | 0.213747 | 12 | -0.742% | Complete year |
| 2014 | 0.214056 | 12 | 0.145% | Complete year |
| 2015 | 0.213684 | 12 | -0.174% | Complete year |
| 2016 | 0.213324 | 12 | -0.168% | Complete year |
| 2017 | 0.215372 | 12 | 0.960% | Complete year |
| 2018 | 0.212668 | 12 | -1.256% | Complete year |
| 2019 | 0.214468 | 12 | 0.847% | Complete year |
| 2020 | 0.215607 | 12 | 0.531% | Complete year |
| 2021 | 0.213553 | 12 | -0.953% | Complete year |
| 2022 | 0.214445 | 12 | 0.418% | Complete year |
| 2023 | 0.212944 | 12 | -0.700% | Complete year |
| 2024 | 0.213981 | 12 | 0.487% | Complete year |
| 2025 | n/a | 1 | — | Partial year |