Figure 1: Relationship between a neighborhood's percentage of Black and white residents and its average PM2.5 concentrations.
Figure 1: Relationship between a neighborhood's percentage of Black and white residents and its average PM2.5 concentrations.
To determine whether racial composition affects whether a neighborhood is exposed to higher PM2.5 levels, a scatter plot was created with the percentage of residents in a specific neighborhood who identify as Black, white, Asian, or Latino as the independent variable and the averaged PM2.5 concentrations from the CalEnviroScreen 2.0 and 3.0 surveys as the dependent variable. PM2.5, or particulate matter with diameters of less than 2.5 micrometers, is an important air pollutant that poses dangerous effects to our respiratory and cardiovascular systems; therefore, we use PM2.5 throughout this study as a way to quantify air quality.
Since both variables are numeric ratios and each neighborhood is associated with a specific value for each variable, we represent each neighborhood as a dot. Thus, we can compare any trends we may see across neighborhoods on a 2D coordinate plane. Percentages (rather than raw population counts) were used to normalize to different census tracts’ populations, and a scatter plot was chosen because each data point is discrete. Furthermore, dot sizes were calculated by the number of measurements taken in each neighborhood, as certain neighborhoods had more measurements than others. Since a smaller measurement sample size could skew the average a lot more than larger datasets, they are visually weighted less than communities with more measurements. We plotted two races on each graph so that one could act as a point of comparison for the other: does an observed relationship appear across multiple racial groups or a select few? Are observed relationships observed regardless of racial composition?
Fig. 1 illustrates that once the percentage of Black residents reaches above 20%, the average PM2.5 concentration is consistently higher than the EPA’s acceptable limit (12 𝜇g/cubic meter). As a result, predominantly Black communities likely experience higher ambient air pollution regardless of other social factors. On the other hand, as the percentage of white residents increases, there is little change in average PM2.5 concentrations. Some predominantly white communities have PM2.5 levels much lower than the acceptable limit, while others are marginally higher. Nevertheless, the relationship between the percentage of white residents and PM2.5 concentrations is much less correlated than the percentage of Black residents, indicating that (1) the trend observed in Fig. 1a for predominantly Black communities is not seen as with white communities, and (2) it is likely that other confounding variables (e.g. household income, age) contribute to differences in air quality among predominantly white communities. Our finding corroborates with a Los Angeles Times investigation by Sammy Roth that discovered higher pollution in low-income and Black communities (Roth 2020). By also studying PM2.5 levels, a study by Su et al. discovered that low-income communities of color experience higher environmental burdens than white and affluent communities (Su et al. 2009).
Figure 2b: Relationship between a neighborhood's percentage of Latino and white residents and its average PM2.5 concentrations.
Figure 2a: Relationship between a neighborhood's percentage of Asian and white residents and its average PM2.5 concentrations.
Figs. 2a and 2b use similar analysis to investigate how the percentage of Asian and Latino residents in a community correlates with PM2.5 concentrations, with the percentage of white residents also serving as a point-of-comparison for both. In Fig. 2a, we observe that like predominantly Black communities, once the percentage of Asian residents in a neighborhood exceeds 20%, the data points cluster above the acceptable limit, although a select few data points (such as 60% Asian composition) lie below the limit. On the contrary, Fig. 2b demonstrates that the trend for predominantly Latino communities is more similar to white neighborhoods than the previous trends for Black or Asian communities. At both low and high Latino percentages, neighborhoods experience relatively similar air pollution, albeit with a couple of outliers between 20% and 60%.
Notably, the observed relationship between Latino racial make-up and pollution differs from the Sammy Roth's report, which found that Latino communities experience higher pollution burdens (Roth 2020). These outliers may be due to other confounding factors that influence a community’s exposure to particulate matter, as mentioned above.
Figure 3: Correlations between a community’s percentage of Black residents and how much average PM2.5 levels changed between the 2009-2011 CalEnviroScreen 2.0 and 2012-2014 CalEnviroScreen 3.0 campaigns.
Similar to Figures 1 and 2, Figure 3 consists of a scatter plot, with each neighborhood represented as a dot, and dot sizes representing the number of measurements taken at that location throughout the combined CalEnviroScreen 2.0 and 3.0 campaigns. Furthermore, the mean difference regardless of racial composition was found to be approximately -0.5 𝜇g/cubic meter; this difference was plotted on the graph as the goldenrod-colored line. Since the difference is negative, we know that PM2.5 concentrations were lower in CalEnviroScreen 3.0 than 2.0, indicating that as a whole, air quality improved between the two campaigns.
Figure 3 illustrates that mostly Black communities (percentage of Black residents above 20%) had decreased PM2.5 levels, with the differences being close to the overall mean difference (-0.5 𝜇g/cubic meter). Figure 4a shows that PM2.5 concentrations in predominantly Asian communities decreased more than than the mean difference, while 4b indicates that Latino communities experienced changes in PM2.5 independent of the percentage of Latino residents in that neighborhood. Similarly to Figs. 1 and 2, we also used the percentage of white residents as a point of comparison, and the percentage of white residents had little correlation to the change in air quality, a trend similar to the percentage of Latino residents.
Figure 4b: Correlations between a community’s percentage of Latino residents and how much average PM2.5 levels changed between the 2009-2011 CalEnviroScreen 2.0 and 2012-2014 CalEnviroScreen 3.0 campaigns.
Figure 4a: Correlations between a community’s percentage of Asian residents and how much average PM2.5 levels changed between the 2009-2011 CalEnviroScreen 2.0 and 2012-2014 CalEnviroScreen 3.0 campaigns.
As a result, we observed that as a whole, predominantly Black and Asian neighborhoods had greater decreases in PM2.5 levels and thus more improved air quality between the 2 campaigns than white or Latino, with Asian neighborhoods having a larger decrease in PM2.5 than Black neighborhoods.
It is notable that Latino and white communities did not have clear trends between the percentage of residents and their changes in air quality. This finding may be due to multiple factors: first, it is possible that Black and Asian neighborhoods have gained political capital through activism and political engagement, allowing them to advocate for clean-up efforts. Secondly, it could also be because Black and Asian communities already experience more severe air pollution in the first place. Since it is easier to improve air quality in heavily polluted areas versus areas that already have clean air in the first place, these findings could instead be a symptom of the existing racial disparity itself. Finally, variations can also be due to other physical or social factors. The prevalence of wildfires, for instance, can affect particulate matter concentrations, so if one of these time periods had worse fire seasons, this variable would not be accounted for in our visualizations.
This page was started with Mobirise