Income-based exposure disparities across California counties, 2000 to 2023: a generalizable statistical framework

Abstract Environmental justice studies have long documented disparities in air pollution exposure, yet few provide a generalizable framework to quantify these disparities across time, pollutants, and spatial scales. We develop a reproducible, population-weighted framework to quantify income-based disparities in air pollution exposure across California counties from 2000 to 2023. Daily regulatory measurements of PM 2.5 , NO 2 , O 3 , SO 2 , and CO are aggregated to county-day means and then to group-level exposures using county-year population weights, yielding the estimated exposure of an average individual in each income group. Counties are assigned to low, middle, and high income strata using fixed statewide thresholds derived from 2019 to 2023 median household income (MHI) percentiles, and disparities are summarized through absolute gaps (low minus high), relative ratios, and the share of days with low-income exposure exceeding high-income exposure. Statewide, PM 2.5 and O 3 exposures generally experienced higher PM 2.5 and O 3 in low-income counties, while NO 2 and CO concentrations are consistently higher in high-income areas. SO 2 transitions from a high-income advantage in early years to a low-income advantage later. Trend analysis indicates a significant narrowing of disparities for PM 2.5 ( <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" overflow="scroll"> <mml:mrow> <mml:mo>−</mml:mo> <mml:mn>0.172</mml:mn> <mml:mstyle scriptlevel="0"/> <mml:mi>μ</mml:mi> <mml:msup> <mml:mtext>gm</mml:mtext> <mml:mrow> <mml:mo>−</mml:mo> <mml:mn>3</mml:mn> </mml:mrow> </mml:msup> </mml:mrow> </mml:math> yr −1 , 95% CI: −0.279, −0.065), O 3 ( <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" overflow="scroll"> <mml:mrow> <mml:mo>−</mml:mo> <mml:mn>3.06</mml:mn> <mml:mo>×</mml:mo> <mml:msup> <mml:mn>10</mml:mn> <mml:mrow> <mml:mo>−</mml:mo> <mml:mn>4</mml:mn> </mml:mrow> </mml:msup> <mml:mstyle scriptlevel="0"/> <mml:mtext>ppm</mml:mtext> </mml:mrow> </mml:math> yr −1 , 95% CI: <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" overflow="scroll"> <mml:mrow> <mml:mo>−</mml:mo> <mml:mn>3.76</mml:mn> <mml:mo>×</mml:mo> <mml:msup> <mml:mn>10</mml:mn> <mml:mrow> <mml:mo>−</mml:mo> <mml:mn>4</mml:mn> </mml:mrow> </mml:msup> </mml:mrow> </mml:math> , <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" overflow="scroll"> <mml:mrow> <mml:mo>−</mml:mo> <mml:mn>2.35</mml:mn> <mml:mo>×</mml:mo> <mml:msup> <mml:mn>10</mml:mn> <mml:mrow> <mml:mo>−</mml:mo> <mml:mn>4</mml:mn> </mml:mrow> </mml:msup> </mml:mrow> </mml:math> ), and NO 2 ( <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" overflow="scroll"> <mml:mrow> <mml:mo>+</mml:mo> <mml:mn>0.114</mml:mn> <mml:mstyle scriptlevel="0"/> <mml:mtext>ppb</mml:mtext> </mml:mrow> </mml:math> yr −1 , 95% CI: +0.076, +0.152). CO disparities intensify in favor of high-income counties ( <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" overflow="scroll"> <mml:mrow> <mml:mo>−</mml:mo> <mml:mn>0.0051</mml:mn> <mml:mstyle scriptlevel="0"/> <mml:mtext>ppm</mml:mtext> </mml:mrow> </mml:math> yr −1 , 95% CI: −0.0087, −0.0015), while SO 2 shows a positive trend ( <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" overflow="scroll"> <mml:mrow> <mml:mo>+</mml:mo> <mml:mn>0.083</mml:mn> <mml:mstyle scriptlevel="0"/> <mml:mtext>ppb</mml:mtext> </mml:mrow> </mml:math> yr −1 , 95% CI: +0.052, +0.115), consistent with a flipped gradient. Sensitivity analyses using an alternative income definition based on 2014–2018 MHI and 25th and 75th percentile cut points yield the same qualitative gradient patterns and trend signs, demonstrating robustness to income-group specification. All data, code, and a versioned one-command workflow are publicly available to facilitate reuse for other states or for alternative socioeconomic stratifiers such as race or education.