While insightful, Liévanos [
18] also had important limitations. First, it did not assess the bivariate and multivariable relationship between population vulnerability and PM
2.5, the major environmental health hazard contributor to cumulative pollution burden in California. Second, it conceptualized neighborhood racial composition as an extrinsic dimension of environmental health vulnerability in accordance with prior environmental health scholarship [
28]. Accordingly, it undertheorized how neighborhood racial composition is a “structural,” community-level [
9] component of population vulnerability of environmental health hazard exposure that is arguably distinct from extrinsic factors of socioeconomic status and intrinsic factors of population sensitivity and susceptibility. Third, it did not analyze Asian environmental inequality outcomes because of its reliance on previous literature and the empirical details of the California case, which collectively indicated that tract-level Black and Latinx composition were the consistent race-based population vulnerability correlates of cumulative pollution burden.
This present study addressed the limitations in Liévanos [
18] and related statewide research on California [
19,
20,
21,
58] through the following. First, it drew on the perspectives of critical race, space, and environmental justice studies [
7,
8,
9,
29,
30,
31,
32,
33,
34,
35,
36,
37,
38,
39,
40,
41,
42,
43,
44,
45,
46,
47,
48,
49,
50,
51], and empirical research on select regions of California and throughout the United States [
7,
8,
9,
10,
11,
12,
13,
14,
15,
22,
23,
24,
25,
26,
27,
34,
38,
39,
52,
53,
54,
55,
56,
57,
59] to advance a “racialized structural vulnerability” approach to environmental health vulnerability, particularly for Asians, Blacks, and Latinxs, that is separate from extrinsic and intrinsic vulnerability. This framework elaborates on the community-level component of Gee and Payne-Sturges’s [
9] environmental health disparities framework. It bases the analysis of race-based environmental health vulnerabilities within a contemporary environmental racial formation that has historical roots in California [
34,
41,
42,
43,
44,
45,
47,
49,
52,
56]. This environmental racial formation upholds the elite status and environmental privilege of Whites (e.g., low PM
2.5 exposures) through a variety of institutional mechanisms that maintain racial segregation and lead to the relative subordination and heightened environmental health hazard (e.g., PM
2.5) exposures of Asian, Black, and Latinx bodies and spaces. The racialized structural vulnerability perspective thus argues against subsuming neighborhood racial composition within the realm of “extrinsic” vulnerability factors like socioeconomic status [
28] (p. 85). It especially argues against excluding neighborhood racial composition all together from the data as is done with colorblind tools like CalEnviroScreen [
20,
21,
58]. Such omissions grossly miss the “social facts” [
42] and political projects [
43] of race, racism, and environmental racial inequality, especially in California [
14,
18,
22,
23,
24,
25,
26,
27,
34,
38,
39,
41,
42,
43,
44,
45,
49,
51,
52,
53,
59].
Empirically, this study showed that tract-level Latinx, Black, and Asian composition were among the strongest positive multivariable correlates of PM
2.5 percentile rankings in California, net of other factors (see
Table 4, Model 1). A secondary finding of this study is that it demonstrated that tract-level Latinx and Black composition were significantly and differentially associated with concentrated socioeconomic and health disadvantage, respectively, as well as with elevated PM
2.5 exposures for Black and Latinx populations in California (see
Table 1 and
Table A1) in ways that have not been uncovered in previous research [
18].
This secondary finding was facilitated by the development of two new population vulnerability indicators of “concentrated disadvantage”—isolated Latinx economic disadvantage and Black health disadvantage—with a single PCA approach that illuminated how extrinsic and intrinsic vulnerability were differentially associated with the structural vulnerability factor of neighborhood racial composition. These new factors improve upon prior related research [
18,
20] by capturing how population vulnerabilities operate spatially for different racial groups in California. The isolated Latinx economic disadvantage factor variable is consistent with prior research on Latinx environmental inequalities, which shows how economic disadvantage and linguistic isolation (indicators of extrinsic vulnerability) are concentrated among segregated Latinx residents [
10,
11,
14,
18,
59]. The Black health disadvantage factor variable is consistent with California-based environmental and public health research, which shows that segregated Black residents have an elevated prevalence of asthma and low birth weight (indicators of intrinsic vulnerability) [
52,
53]. In addition, the present study found these race-based, concentrated disadvantage factor variables performed differently in the multivariable regression analyses when compared to models that included the tract racial composition variables alongside the other new factor variables developed in this study: isolated economic disadvantage (a measure of extrinsic vulnerability) and health disadvantage (a measure of intrinsic vulnerability) (see
Table 4, Models 1 and 2).
Limitations and Future Research and Policy Implications
Despite its contributions, future research should address important limitations of this study. First, the main analyses in this study used data from 2004 to 2012, and especially from CalEnviroScreen 2.0, to examine the effects of neighborhood racial composition, other indicators of population vulnerability, traffic and industrial emission sources, and spatial dynamics on PM
2.5 exposures in California. To what extent do the results differ when using more recent data from CalEnviroScreen 3.0? The
Appendix A presents additional bivariate correlation analyses that consider the main variables used in this study, and the percentile rankings of annual average PM
2.5 concentrations from 2012–2014 and the change in percentile rankings of annual average PM
2.5 concentrations from 2009–2011 to 2012–2014. The annual average PM
2.5 concentration estimates for 2012–2014 were derived from CalEnviroScreen 3.0. These were similarly measured as the 2009–2011 period in CalEnviroScreen 2.0 [
58,
67,
68]. In briefly discussing these additional analyses, this section seeks to draw out future research and policy implications from this study.
Table A2 shows that the percentile rankings of annual average PM
2.5 concentrations from 2009–2011 were positively, but not perfectly, associated with the percentile rankings of annual average PM
2.5 concentrations from 2012–2014 (Pearson
r = 0.902;
p < 0.001). This association as well as the minor negative correlation between the percentile rankings of the annual average PM
2.5 concentrations from 2009–2011 and change in the percentile rankings of annual average PM
2.5 concentrations from 2009–2011 to 2012–2014, indicates that there have been some minor reductions in PM
2.5 concentrations across California census tracts following the establishment of the new PM
2.5 NAAQS of 12 μm/m
3 in 2012. However,
Table A2 also suggests that those reductions in annual average PM
2.5 concentrations are more consistently associated with tract-level Asian composition, but not so for tracts with elevated levels of Blacks and Latinxs, wherein annual average PM
2.5 concentrations remained significantly high in 2012–2014. In addition, percentile rankings of annual average PM
2.5 concentrations from 2012–2014 were positively associated with this study’s factor variables of health disadvantage, isolated economic disadvantage, Black health disadvantage, and isolated Latinx economic disadvantage as well as with industrial zoning and traffic density in 2004 and tract containment in the South Coast and SJV air basins. Of these, the change in percentile rankings of annual average PM
2.5 concentrations from 2009–2011 to 2012–2014 was positively associated with the factor variables and percent tract containment in the SJV air basin. The change in percentile rankings of PM
2.5 were negatively associated with the tract containment in the South Coast air basin.
There are important policy implications from the main findings of the present study as well as the preliminary demographic and spatial relationships summarized in
Table A2. Namely, they speak to the urgent need for targeted PM
2.5 reduction strategies for the SJV and its vulnerable neighborhoods experiencing concentrated health and socioeconomic disadvantages. Decades of environmental justice organizing and research by allied environmental health scientists throughout the SJV have helped to bring increased state-level regulatory scrutiny to the inadequate and exclusionary development patterns, environmental policies, and decision making processes in the SJV. Advocates and allies have targeted these developments, policies, and processes because they have placed vulnerable residents and communities at increased risk of exposures to toxic air pollution and other hazardous substances in the SJV [
69,
70,
71,
72,
73,
74,
75,
76,
77,
78,
79].
Compelled by these environmental justice movement and allied efforts, the California Air Resources Board (CARB) and the San Joaquin Valley Air Pollution Control District (SJVAPCD) produced a new “2018 Valley PM
2.5 Plan” that was approved by the US EPA [
80,
81]. The plan moves the SJV to the “Serious” nonattainment classification for four major federal PM
2.5 standards, particularly the 2012 annual NAAQS of 12 μm/m
3, which mobilizes additional resources and institutional mechanisms to move the SJV into the attainment of those federal standards in the 2020s. In addition, it is claimed by officials to be a precedent-setting, integrated approach to aggressively reducing primary and secondary contributors to PM
2.5 in the SJV through combining old and new regulatory controls and voluntary incentives from the CARB that address mobile sources and from the SJVAPCD, which address stationary and area sources [
80,
81]. The SJVAPCD prepared a “Negative Declaration for the 2018 PM
2.5 Plan,” that found that the plan “would not have a significant impact on air quality and would have a less than significant impact on the environment” [
80] (p. 28).
In keeping with a continued focus on “participatory justice” within the Californian environmental protection agencies [
72,
74,
75,
76], considerable efforts were apparently made by the CARB to improve public participation in and deliberation about the plan [
80]. However, it remains to be seen how participating agencies in the plan assessed or plan to monitor the distributional effects and environmental health equity implications of the plan. Nor does the plan stipulate how it may achieve its goals in coordination with California’s broader effort to combat the dual problems of cumulative pollution burden and climate change vulnerabilities [
18,
75,
82], which have been previously identified as significantly concentrated in the SJV [
70].
There are significant political and organizational constraints on regulatory scientists and officials to pursue such questions in California [
18,
72,
74,
75,
76,
83] and under the Trump Administration’s US EPA [
11,
84]. Therefore, it is incumbent upon advocates and academic scientists to conduct critical analyses of the environmental health equity implications of any PM
2.5-related policies and proposed reduction strategies in California. This study’s analytical techniques and racialized structural vulnerability framework could guide future advocacy, environmental, public health, and social science research aimed at taking on this arduous task. In addition, there are a number of nationwide datasets that were not featured in this study that could inform such analyses at both the census tract and census block group levels. These other datasets could include updated population, housing, and land use estimates from the Decennial Census, the ACS, and the US Geological Survey’s National Land Cover Database (NLCD) [
10,
11,
59], respectively. They can also include more frequent estimates of traffic volume and PM
2.5 exposures from the US EPA’s Community Multi-Scale Air Quality Model [
12], the US EPA’s National Emissions Inventory (NEI) [
85,
86], and the US EPA’s EJSCREEN [
87,
88,
89,
90]. When merged together, particularly at the census block group level, these data can provide a higher-resolution analysis of race-based ambient PM
2.5 exposure vulnerabilities, and their relationship to recent estimates of industrial land use and traffic volume in California. Assessing such relationships with regularly-updated data and at a finer scale could overcome this study’s use of the 2004 industrial zoning and traffic volume estimates, which performed poorly in the regression analyses at the census tract level, despite the known relationship between industrial and traffic emissions and PM
2.5 [
2].
In addition, future research could pair the alternative nationwide datasets listed above with other data that can advance our understanding of race-based PM
2.5 exposure vulnerabilities and exposures to facility-level “co-pollutants” of PM
2.5 and CO
2 in California [
75,
85]. This additional data includes, in particular, facility-level carbon dioxide (CO
2) emissions estimates from the US EPA Greenhouse Gas Reporting Program [
85] and California’s Climate Investments (CCI) Project Map [
90]. The CCI Project Map is associated with California’s program to redistribute Greenhouse Gas Reduction Funds from its carbon cap-and-trade program to the most overburdened census tracts identified with CalEnviroScreen data [
18,
75,
82].
Furthermore, this study acknowledged the important historical and institutional forces that generally contribute to the existence of race-based environmental health vulnerabilities in California. However, it did not directly measure the effects of those forces on contemporary PM
2.5 exposure vulnerabilities nor on the health outcomes of such exposure. To do so, future research could link the racialized structural vulnerability framework, drawn in part from critical race, space, and environmental justice perspectives [
9,
29,
30,
31,
32,
33,
34,
35,
36,
37,
38,
39,
40,
41,
42,
43,
44,
45,
46,
47,
48,
49,
50,
51], with Sewell’s [
91] “racism-race reification process” model of racial health disparities. The resulting synthesis could examine, for example, how racial disparities in asthma hospitalizations in California [
92] are associated with contemporary racial inequalities in mortgage market access (as identified with Home Mortgage Disclosure Act data) and PM
2.5 exposures as well as historical patterns of racial and spatial exclusion from the home mortgage market (i.e., “redlining”) [
44,
48,
49,
50]. Illuminating the influence of such historical and institutional forces might help to identify causal factors in the formation of race-based environmental health vulnerabilities as well as strategic points of intervention in planning and governance processes to advance environmental justice and health equity [
93].