What this research found
Regulatory compliance status for fine particulate matter across the Los Angeles-Long Beach-Anaheim metropolitan area was reconstructed from EPA monitoring records for 2015 through 2023, applying the federal design-value procedure to all 14 qualifying monitors. The 2021-2023 area design value came out at 12.2 micrograms per cubic metre, above the 12.0 standard then in force and well above the 9.0 level adopted in 2024, placing the area in nonattainment by a margin of just 0.2 and on the strength of a single monitor at Compton. Across the nine-year record the controlling annual concentration shows no statistically detectable trend.
- The 2021-2023 area design value is 12.2 micrograms per cubic metre, set by the Compton monitor whose annual means of 13.41, 12.12 and 11.12 average to 12.22. Because the area figure is the maximum across qualifying sites rather than a network average, a single monitor decides the status of the whole metropolitan area, and the network median three-year average is 11.4.
- All seven monitors that qualified for the design-value window exceed the revised 2024 annual standard of 9.0, by 3.2 at the controlling site — a network-wide violation, unlike the marginal 0.2 exceedance of the 2012 standard. Concentrations vary widely across the basin: near-road and central sites top the annual means, while the high-desert Lancaster-Fairgrounds site recorded 3.37 in 2023.
- Short-term exceedance activity is episodic rather than trending. Days above 35.0 micrograms per cubic metre ranged from 4 area-wide in both 2016 and 2023 up to 32 in 2020, the severe California wildfire year, when the worst single monitor logged 19. Intermediate elevations appear in 2017 and 2018 at 14 and 15 days and in 2021 at 17.
- No monotonic trend is detectable in the area's controlling annual mean over 2015 to 2023. The Theil-Sen slope is −0.013 micrograms per cubic metre per year with a 95% confidence interval running from −0.435 to +0.452, and the Mann-Kendall test returns a Kendall tau of exactly zero and a p-value of 1.000. The series simply oscillates, between 10.99 in 2019 and 14.83 in 2020.
- Removing the 2020 wildfire year steepens the slope estimate nearly sixfold, to −0.085 micrograms per cubic metre per year, while leaving the verdict unchanged at a p-value of 0.902. In a record this short, one extreme year moves the estimated rate of change enough that any point estimate needs to be reported alongside how anomalous years were handled.
How it was done
Nine years of EPA daily fine-particle measurements from federal reference and equivalent method instruments were filtered to the metropolitan area, turning up 14 monitoring sites, 12 in Los Angeles County and two in Orange County. Where a site ran collocated instruments, a fixed preference order resolved exactly one authoritative value per site per day. Each site-year was screened against the federal completeness requirement of at least 75% of scheduled samples in every calendar quarter, with the sampling schedule inferred from observed sample density and the most lenient plausible schedule applied so that borderline sites were not failed spuriously; 95 of 105 candidate site-years passed. Annual means were then built from quarterly means, site design values as three-year averages of those means, and the area value as the maximum across qualifying sites. Trend was assessed on the series of annual maxima using the Mann-Kendall test for significance and the Theil-Sen estimator for magnitude, with both re-run on an eight-point series excluding 2020.
Data sources
- EPA Air Quality System daily summary PM2.5, parameter 88101 from federal reference and equivalent methods, 2015 to 2023, accessed 11 July 2026 — 14 monitoring sites and 105 candidate site-years
- 40 CFR Part 50, Appendix N — the federal design-value computation and data-completeness requirements
- EPA final rules revising the primary annual standard to 12.0 micrograms per cubic metre in January 2013 (78 FR 3086) and to 9.0 in March 2024 (89 FR 16202)
- Enayati Ahangar et al., Atmospheric Environment 255:118431 (2021) and McClure and Jaffe, PNAS 115:7901 (2018) — regional context on wildfire contributions to particulate trends
Limitations
Nine annual points give the trend test little statistical power, so failing to detect a trend is not positive evidence that concentrations are flat, and the maximum-site series is sensitive to which monitors happen to stay complete — network attrition cut the qualifying set to eight sites in 2022. Sampling schedules were inferred rather than read from authoritative metadata, no exceptional-event exclusions were applied so wildfire-influenced values enter at face value, and only the annual standard was computed rather than the companion 24-hour design value.
How this research was produced
K-Dense Web planned and ran this environment investigation end to end — gathering the sources, carrying out the analysis, producing the figures, and drafting the report. The full session transcript, including every intermediate step, is available to view.
