Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2022 May 1.
Published in final edited form as: J Expo Sci Environ Epidemiol. 2021 Apr 19;31(3):454–460. doi: 10.1038/s41370-021-00324-6

Socioeconomic Disparities in Incidents at Toxic Sites during Hurricane Harvey

Wil Lieberman-Cribbin 1,2, Bian Liu 1,2, Perry Sheffield 3, Rebecca Schwartz 1,2,4, Emanuela Taioli 1,2
PMCID: PMC8992005  NIHMSID: NIHMS1789586  PMID: 33875771

Abstract

Background:

Hurricane Harvey facilitated exposure to various toxic substances and floodwater throughout the greater Houston metropolitan area. Although disparities exist in this exposure and vulnerable populations can bear a disproportionate impact, no research has integrated disparities in exposure to toxic incidents following Hurricane Harvey.

Objective:

The objective of this study was to analyze the relationship between flooding, socioeconomic status (SES), and toxic site incidents.

Methods:

Data on toxic site locations, reported releases, and flood water depths during Hurricane Harvey in the greater Houston area were compiled from multiple sources. A multivariable logistic regression was performed to predict the odds of a toxic site release by flooding at the site, SES and racial composition of the census tract.

Results:

83 out of 1403 toxic sites (5.9%) had reported releases during Hurricane Harvey. The proportion of toxic sites with reported incidents across increasing SES index quintiles were 8.35, 7.67, 5.14, 4.55, and 0.51, respectively. The odds of an incident was higher in the lowest SES quintiles areas (OR=17.26, 95%CI: 2.39–124.84) compared to the highest quintile. Flooding was similar at toxic sites with and without incidents, and was distributed similarly and highest at toxic sites among lower SES quintiles.

Significance:

Despite similar flooding across toxic sites during Hurricane Harvey, areas with lower SES were more likely to have a toxic release during the storm, even accounting for number of toxic sites. Improving quality of maintenance, safety protocols, number of storm-resilient facilities may minimize this disproportionate exposure and its subsequent adverse outcomes among socioeconomically vulnerable populations.

Keywords: Environmental justice, chemical exposure, flooding, socioeconomic disparities

Introduction

Hurricane Harvey made landfall in August 2017 and resulted in record setting flooding, rainfall and high winds across the greater Houston metropolitan area. Outages at power and electricity transmission plants disrupted critical infrastructure1 and hundreds of thousands of individuals lost power in homes2. Defined by its extensive industry, manufacturing, and oil and gas refineries, the greater Houston metropolitan area experienced numerous potential environmental health risks as a result of Hurricane Harvey3,4. Numerous Environmental Protection Agency (EPA) Superfund sites are located in the area, some of which were flooded and potentially contaminated the flood waters in their surroundings5. During the storm, confirmed incidents of toxic substances from chemical, waste, and petroleum facilities were disseminated into both floodwaters and outdoor air3,6. These substances included carcinogens, such as butadiene and benzene, as well as volatile organic compounds and sulfur dioxide. Further, anecdotal exposure to mold and sewage bacteria in flooded streets was reported7,8.

In studying environmental exposures, an environmental justice lens has been employed to determine if specific disadvantaged populations bear a disproportionate impact of exposure9–11. In past natural disasters, research findings have suggested that racial/ethnic minorities and lower-income populations are more likely to live in close proximity to environmental hazards and that this is likely to result in a disproportionate impact of the exposure in these populations11–13. After Hurricane Katrina, study results identified that Black and low-income residents of New Orleans were exposed to a disproportionate amount of flooding, displacement from their homes, and long-term mental health concerns14–17, especially in children18,19. These disparities in exposure translate into inequities in the health effects of the storm, with vulnerable populations bearing a disproportionate effect20. In addition to the short- and long-term mental health effects, these populations were exposed to poor outdoor air quality and mold19,21, as well as lack of access to clean drinking water and exposure to contaminated water22.

Geospatial analyses have emerged as a central tool to study environmental exposures23, and have been utilized after Hurricane Harvey to assess flood exposure and sources of environmental pollution10,11,,24–26. One study emphasized that rainfall and flooding spread millions of gallons of hazardous chemicals throughout the Houston area and that parks and recreational spaces were at risk of contamination27. Hispanic, Black and other racial/ethnic minority households experienced more extensive flooding than white households, and lower SES households faced more extensive flooding than higher SES households24. Similarly, another study reported that non-Hispanic Black and socioeconomically deprived residents faced a higher burden of flood exposure, suggesting that racial, ethnic, and socioeconomic disparities were present in exposure and burden of the storm10. Additionally, research has reported that flooding was greater in areas with higher proportions of disabled residents, and that those with ambulatory and cognitive disabilities were more likely to reside in areas with the highest proportion of flooding in their neighborhood25. Disparities in flood exposure may be the basis for disparities in flood impacts, health outcomes, and post-disaster needs. For example, residents engaged in cleanup efforts in homes and water-damaged areas were also exposed to mold7,8, and this burden could be unequally distributed. Additionally, Hurricane Harvey also served as a mechanism to redistribute pre-existing environmental contaminants in the area, such as polycyclic aromatic hydrocarbons28 thus further modifying the individual risk of exposure. Despite the research studying disparities in flooding following Hurricane Harvey using an environmental justice context, research has not integrated the toxic releases that were prevalent in this heavily industrialized area. The objective of the this analysis was thus threefold: 1) to study the distribution of toxic sites and reported incidents at toxic sites during Hurricane Harvey, 2) examine the relationship between socioeconomic status and incidents at toxic sites, 3) determine if flooding was associated with SES and/or toxic site locations and incident locations. We hypothesized that the incidents at toxic sites occurred disproportionately among socioeconomically vulnerable populations, and that lower SES areas experienced increased flooding. The conceptual framework at the basis of this analysis and the interplay between toxic site abundance, socioeconomic status, flooding, and the toxic incidents occurring in an area are displayed in Figure 1.

Figure 1:

Figure 1:

Open in a new tab

Conceptual model of the interplay of toxic site abundance, socioeconomic status, flooding, and the toxic incidents occurring in an area.

Methods

Toxic Sites

The location of toxic sites and of those with reported incidents during Hurricane Harvey in the greater Houston area were retrieved from the Sierra Club Environmental organization website6. This dataset provided the facility name and location (latitude/longitude coordinates), type of facility (chemical, petroleum, wastewater treatment plant, etc.), as well as a description of the known incident(s) that occurred at that site. Data was sourced from the EPA Toxic Release Inventory, the Energy Information Administration, the US Coast Guard National Response Center, the Texas Commission on Environmental Quality, and the Railroad Commission of Texas. This data provided the frequency of total toxic sites and the number of toxic sites with incidents according to type of facility.

Flooding

Floodwater depth and extent information were obtained from the Federal Emergency Management Agency (FEMA) Harvey Flood Depth Grids product29. This product utilized Triangulated Irregular Network interpolation from high water marks, inspection data, and elevation data to create a 3-meter horizontal resolution water depth above ground product for Hurricane Harvey. This product was downloaded and imported into our licensed copy of ArcGIS (version 10.7.1; ESRI, Redlands, CA) for analyses. The flood height at each toxic site location was extracted in ArcGIS and log transformed

Socioeconomic Status Index

An index of socioeconomic parameters was created from census tract level data downloaded from 2013–2017 American Community Survey (ACS) 5-year estimates modified from Yost et al. 200130. Specifically, census tract level data on Median household income, Median House Value, Median Gross Rent, Percent Below 150% Poverty, Education Index, Working Class, and Percent Unemployed was downloaded and each variable was ranked, standardized, and applied factor loadings according to Yu et al. 201431. A linear combination was then created to obtain a socioeconomic status index score assigned to each census tract, according to SES score = (Median household income*0.191) + (Median home value*0.166) + (Median gross rent*0.169) + (Percent Below 150% Poverty*-0.179) + (Education Index*0.185) + (Working Class*-0.190) + (Percent Unemployed*-0.111). Census tracts were classified into quintiles of this SES score, with a score of 1 corresponding to the lowest SES quintile (i.e. lowest resourced) and a score of 5 representing the highest SES quintile. In addition to the elements included in the Yost et al. 2001 index, the racial composition of each census tract was also obtained from the 2013–2017 ACS 5-year estimates and included in the analysis.

Analysis

The absolute frequency of total toxic sites and number of sites with incidents were computed according to SES quintiles, and mapped using ArcGIS. To account for the fact that the distribution of toxic sites among SES quintiles may not be uniform, and that more toxic sites could exist in lower SES areas, a ratio of the number of toxic site incidents per the number of retoxic sites was computed according to different SES quintiles. Wilcoxon rank-sum tests were performed to assess differences in flood heights between toxic sites with incidents and without incidents. A 1-mile buffer around toxic sites was created in ArcGIS and the number of census tracts that fell within this area was summed to provide a measure of potential risk of exposure to a toxic site incident, similar to Chakraborty et al 201123. A multivariable logistic regression was proposed to predict the odds of incident occurrence at a toxic site by SES quintiles, the flood height at the toxic site, and the racial composition at the census tract level as predictors. The presence of spatial autocorrelation was tested using Moran’s I (p < 0.0006) and a Gaussian spatial covariant structure was incorporated into the logistic model to account for such autocorrelation. All analyses were performed in SAS v9.4 and ArcGIS v10.7.1.

Results

Facility Distribution and SES

The majority of toxic sites recorded in the dataset were wastewater facilities (n = 808; 7.6%), followed by chemical facilities (n = 238; 17.0%), and Oil & Gas Waste Facilities (n = 99; 7.06%). The distribution of Oil & Gas Waste Facilities, Petroleum & Natural Gas Facilities, and Petroleum Bulk Terminals were most abundant in lower SES quintiles. Specifically, the lowest two SES quintiles contained 52.5%, 57.5%, 59.6%, 66.7 % of Oil & Gas Waste Facilities, Petroleum & Natural Gas Facilities, Petroleum Bulk Terminals, and Superfund sites, while the top two SES quintiles contained 20.2%, 19.7%, 16.6%, and 25.0%, respectively (Figure 2). The majority of incidents at toxic sites during Hurricane Harvey (n = 83) occurred in Petroleum & Natural Gas Facilities (n = 29; 34.9%), followed by Chemical Facilities (n = 29; 34.9%), and Superfund Sites (n = 13; 16.0%).

Figure 2:

Figure 2:

Open in a new tab

Distribution of Toxic Facility Type according to Socioeconomic Status Index

Note: A score of 1 corresponds to the lowest SES quintile and a score of 5 represents the highest SES quintile. The black bar represents the median SES index quintile for each facility type.

SES and Toxic Site Incidents

When we overlaid the location of toxic sites with incidents overlaid on the SES quintiles -level (Figure 3), we observed that toxic site incidents seemed to occur in areas of lower SES. Lower SES quintiles had the largest ratio of incidents per toxic site locations, with the trend decreasing with increasing SES; SES quintile 1: 8.35 incidents per 100 toxic site locations, SES quintile 2: 7.67 incidents per 100 toxic site locations, SES quintile 3: 5.14 incidents per 100 toxic site locations; SES quintile 4: 4.55 incidents per 100 toxic site locations SES quintile 5: 0.51 incidents per 100 toxic site locations (Figure 4).

Figure 3:

Figure 3:

Open in a new tab

Distribution of incidents at toxic sites, overlaid on quintiles of socioeconomic status

Note: A score of 1 corresponds to the lowest SES quintile and a score of 5 represents the highest SES quintile. Each point represents a toxic site incident location.

Figure 4:

Figure 4:

Open in a new tab

Toxic Site incident per toxic site abundance according to quintiles of socioeconomic status index

Note: A score of 1 corresponds to the lowest SES quintile and a score of 5 represents the highest SES quintile.

An additional analysis was performed to assess if there was a particular component of the SES index that was driving this trend using Wilcoxon rank-sum tests to compare these variables in census tracts with toxic site incidents compared to those without incidents (Table 1). In census tracts with toxic site incidents, median household income, median home value, median gross rent, education index, and the percentage of the population in the working class were statistically significantly lower, while the percentage of the population living < 150% poverty and the percent unemployed were statistically significantly higher. The proportion of white only residents in census tracts with toxic site incidents was higher compared to those without incidents (p = 0.0060).

Table 1:

Socioeconomic status index component values in the areas with and without toxic site incidents.

SES Index Components Mean (SD)  Toxic Site Incident (n=83)  No Toxic Site Incident (n=1356)
Median Household Income (USD)* 49672 (16811) 59689 (24733)
Median Home Value (USD)* 105708 (43945) 135249 (80248)
Median Gross Rent (USD)* 867.89 (174.21) 985.20 (310.85)
Percent < 150% Poverty* 32.16 (12.95) 26.71 (13.10)
Education Index* 1296.77 (89.96) 1345.71 (97.22)
Percent Working Class* 74.66 (9.35) 67.87 (12.76)
Percent Unemployed* 8.11 (4.80) 7.10 (3.94)
Open in a new tab
*

p < 0.01

An additional buffer analysis determined how many census tracts fall within a 1 mile buffer distance around toxic site incidents. On average, each toxic site incident buffer intersected with 3.10 census tracts (standard deviation 1.81). A total of 258 census tracts overlapped with this 1-mile buffer, with at maximum of 10 census tracts overlapping with one toxic site incident. The average number of census tracts within this one mile distance of toxic site incidents (3.30 census tracts) was highest among the lowest SES quintile.

Flooding

Flooding at toxic sites with incidents (average flood height: 3.56±4.81 feet) was similar compared to toxic sites without incidents (3.56±6.99 feet), and this difference was not statistically significant (p = 0.0867). A statistically significant difference was found across all SES quintiles when assessing flood height at the toxic site according to SES quintile (p = 0.0043), with flooding highest in the second lowest quintile of SES (mean: 5.22 ft, standard deviation 8.54 ft) and lowest in the highest SES quintile (mean 2.79 ft, standard deviation 6.82) (Figure 5).

Figure 5:

Figure 5:

Open in a new tab

Distribution of flood height at toxic sites according to quintiles of socioeconomic status

Note: A score of 1 corresponds to the lowest SES quintile and a score of 5 represents the highest SES quintile. p = 0.0043.

Likelihood of a Toxic Site Incident

In assessing the factors associated with an incident at toxic sites (Table 2), increasing SES quintile of the census tract that the toxic site was located in was associated with decreasing likelihood of an incident with the highest two quintiles (SES quintile 4 vs 1; ORadj 15.38, 95% CI 2.11–112.33; SES quintile 5 vs 1; ORadj 17.26, 95% CI 2.39–124.84) positively associated with increased likelihood of a toxic site incident. Log-adjusted flooding at the toxic site location was not statistically significantly associated with the likelihood of an incident at a toxic site (ORadj 1.20, 95% CI0.74–1.94), while increasing proportion of white alone residents was positively associated (ORadj 1.02, 95% CI 1.01–1.04) (Table 2).

Table 2:

Predictors associated with the odds of an incident at a toxic site

Predicting the Likelihood of an Incident at a Toxic Site (yes vs no)
Reference: SES Index Quintile 5 Adjusted odds ratio* (95% confidence interval)
SES Index Quintile 4 0.55 (0.27 – 1.10)
SES Index Quintile 3 0.57 (0.27 – 1.18)
SES Index Quintile 2 0.29 (0.13 – 0.65)
SES Index Quintile 1 0.27 (0.13 – 0.59)
Proportion White Race 1.02 (1.01–1.04)
Log+1 transformed Flood Height at Toxic Site 1.20 (0.74 – 1.94)
Open in a new tab
*

Model was adjusted for quintile of SES status, flooding at the toxic site location, and proportion of White residents in the census tract. Models incorporated random effects using a spatial variance Gaussian structure.

Note: A score of 1 corresponds to the lowest SES quintile and a score of 5 represents the highest SES quintile.

Discussion

This analysis shows that more incidents at toxic sites occurred in low SES areas even after accounting for the fact that there are more toxic sites in lower SES areas. As flooding was not found to be a statistically significant predictor of an incident at a toxic site, the implication is that there are other unmeasured variables that contribute to incidents occurring in lower SES areas. These unmeasured factors could include lower maintenance or upkeep of facilities, gaps in safety measures, encompassing an overall absence of resiliency to natural disasters32.

Failures in the maintenance of infrastructure or in creating appropriate evacuation plans disproportionately influences the poor and racial minorities in the face of natural disasters, as articulated by Nixon33, who discusses how discrimination in neglected communities influences the outcomes of disasters. The effect of such neglect manifested after Hurricane Katrina, where flood levees were poorly maintained in low SES areas, and wealthy and predominately white groups were able to be safer contrasted to poorer and non-white peoples33. An analogous situation is evidence from the findings of this study where lower SES areas not only had more toxic sites, but also had more toxic site releases given widespread flooding during Hurricane Harvey. Additionally, this analysis reports that the count of census tracts within 1 mile of toxic site incidents was highest among lower SES areas, with only one toxic site incident occurring in the highest SES quintile. Thus the advent of Hurricane Harvey shed light on the environmental racism that existed prior to disastrous releases of toxic chemicals.

Structural differences could also exist among facilities. It was reported after Hurricane Harvey that failures in floating roof tanks were responsible for millions of pounds of pollutants released into the atmosphere and into the floodwater through spills34–35. These tanks and their drainage system were not equipped to handle the amount of rainfall that occurred during Hurricane Harvey. It is important to note that there are no state or federal regulations dictating design codes for floating roofs and floating roof drainage systems, instead minimum safety standards exist while following these standards is voluntary.

To our knowledge this is the first study to assess the relationship between socioeconomic status, flooding, and confirmed incidents at toxic exposure sites after a natural disaster in a heavily industrialized area. While prior analyses have utilized geographic information science to study environmental exposures in recreational areas as a result of flooding27, our current work relies on confirmed sources of exposure to emphasize disproportionate chemical exposures in lower SES areas.

This study integrates with recent literature describing Hurricane Harvey flooding; our finding of the equal distribution of flooding across SES quintiles stood in contrast with this and other previous results reporting that flooding from Hurricane Harvey was distributed unequally according to socioeconomic status and ethnicity10,24. One explanation for the difference in results could be the different definition of SES, and the discrepancy in how flooding was calculated, despite flood data originating from the same FEMA product. These studies relied on the proportion of a defined area, either census tracts or a buffer distance from participants, to define flooding within the Woodlands-Sugar Metropolitan Statistical Area. Instead, this present study examined flooding over a larger geographical area. This emphasizes the impact that different geographic areas and units of analysis can have on the interpretation of disparities in flood exposure. Another consideration in determining exposure is the presence of flood water acting as a transporter of certain pollutants. While the raw flooding height is an important component of exposure, the extent and presence of flooding should additionally be incorporated to determine individual exposure, especially in proximity to toxic sites.

One limitation of this analysis is that individual data was not incorporated to determine exposure. Despite having specific toxic site exposure information, widespread sampling measurements were not available throughout the greater Houston area, and the impact on human health could not be specifically addressed. One promising avenue to provide individual exposure information exists in the use of wearables, which have been added to other individual measurements of exposure following Hurricane Harvey, to indicate potential dermal exposure to chemicals36. Additionally, our research group has tested metal concentrations in saliva of a sample of Houston residents to attempt to assess individual exposure from toxic sites.

Conclusions

Disaster preparedness responses need to consider toxic site exposure from an environmental justice perspective and realize that exposure is more likely to occur in lower SES areas both due to the abundance of toxic sites in these areas and the potential for them to be less resilient in hurricane settings. Future environmental justice and disaster preparedness efforts must work to minimize this disproportionate exposure and its subsequent adverse outcomes among socioeconomically vulnerable populations.

Funding

No funding was received for this work.

List of Abbreviations:

ACS

American Community Survey

EPA

Environmental Protection Agency

FEMA

Federal Emergency Management Agency

SES

Socioeconomic status

Footnotes

Conflicts of Interest

The authors declare that they have no competing interests.

Declaration of Interests: We declare no competing interests.

References

  • 1.U.S. Energy Information Administration. Hurricane Harvey caused electric system outages and affected wind generation in Texas. https://www.eia.gov/todayinenergy/detail.php?id=32892 September 13, 2017
  • 2.CBS News. Hurricane Harvey: Teas power outages affect more than quarter-million. https://www.cbsnews.com/news/hurricane-harvey-texas-power-outages-affect-more-than-255000/ August 26, 2017. Accessed August 6, 2019.
  • 3.Glenn S Hurricane Harvey’s Environmental Impacts: A Summary of Regional Data. Houston Advanced Research Center. Accessed August 12, 2019. [Google Scholar]
  • 4.Associated Press. Hurricane Harvey’s toxic impact deeper than public told. March 23. 2018. Accessed September 2, 2019. [Google Scholar]
  • 5.Environmental Protection Agency. Status of Superfund Sites in Areas Affected by Harvey. September 2, 2017. https://www.epa.gov/newsreleases/status-superfund-sites-areas-affected-harvey. Accessed August 6, 2019.
  • 6.Sierra Club. Hurricane Harvey: Toxic Sites https://www.sierraclub.org/environmental-justice/hurricane-harvey-toxic-sites. Accessed August 6, 2019.
  • 7.Kontoyiannis DP, Shah EC, Wurster S, Aitken SL, Graviss L, Raad II, Chemaly RF. Culture-Documented Invasive Mold Infections at MD Anderson Cancer Center in Houston, Texas, Pre–and Post–Hurricane Harvey. InOpen forum infectious diseases 2019. Apr 2 (Vol. 6, No. 4, p. ofz138). US: Oxford University Press. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.Chow NA, Toda M, Pennington AF, Anassi E, Atmar RL, Cox-Ganser JM, Da Silva J, Garcia B, Kontoyiannis DP, Ostrosky-Zeichner L, Leining LM. Hurricane-Associated Mold Exposures Among Patients at Risk for Invasive Mold Infections After Hurricane Harvey—Houston, Texas, 2017. Morbidity and Mortality Weekly Report. 2019. May 31;68(21):469. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.Burger J, Gochfeld M. Conceptual environmental justice model for evaluating chemical pathways of exposure in low-income, minority, Native American, and other unique exposure populations. American journal of public health. 2011. Dec;101(S1):S64–73. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Chakraborty J, Collins TW, Grineski SE. Exploring the environmental justice implications of Hurricane Harvey flooding in Greater Houston, Texas. American journal of public health. 2019. Feb;109(2):244–50. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Flores AB, Castor A, Grineski SE, Collins TW, Mullen C. Petrochemical releases disproportionately affected socially vulnerable populations along the Texas Gulf Coast after Hurricane Harvey. Population and Environment. 2021. Mar;42(3):279–301. [Google Scholar]
  • 12.Walker G (2009). Beyond distribution and proximity: exploring the multiple spatialities of environmental justice. Antipode, 41(4), 614–636.. [Google Scholar]
  • 13.Mohai P, Lantz PM, Morenoff J, House JS, Mero RP. Racial and socioeconomic disparities in residential proximity to polluting industrial facilities: evidence from the Americans’ Changing Lives Study. American journal of public health. 2009. Nov;99(S3):S649–56. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Alexander AC, Ali J, McDevitt-Murphy ME, Forde DR, Stockton M, Read M, Ward KD. Racial differences in posttraumatic stress disorder vulnerability following Hurricane Katrina among a sample of adult cigarette smokers from New Orleans. Journal of racial and ethnic health disparities. 2017. Feb 1;4(1):94–103. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Rhodes J, Chan C, Paxson C, Rouse CE, Waters M, Fussell E. The impact of Hurricane Katrina on the mental and physical health of low-income parents in New Orleans. American journal of orthopsychiatry. 2010. Apr;80(2):237. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Fussell E, Lowe SR. The impact of housing displacement on the mental health of low-income parents after Hurricane Katrina. Social Science & Medicine. 2014. Jul 1;113:137–44. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Galea S, Brewin CR, Gruber M, Jones RT, King DW, King LA, McNally RJ, Ursano RJ, Petukhova M, Kessler RC. Exposure to hurricane-related stressors and mental illness after Hurricane Katrina. Archives of general psychiatry. 2007. Dec 1;64(12):1427–34. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Burton A Health disparities: Crisis not over for hurricane victims. 2006. [DOI] [PMC free article] [PubMed]
  • 19.Rath B, Young EA, Harris A, Perrin K, Bronfin DR, Ratard R, VanDyke R, Goldshore M, Magnus M. Adverse respiratory symptoms and environmental exposures among children and adolescents following Hurricane Katrina. Public Health Reports. 2011. Nov;126(6):853–60. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Lowe SR, Willis M, & Rhodes JE (2014). Health problems among low-income parents in the aftermath of Hurricane Katrina. Health Psychology, 33(8), 774. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Chew GL, Wilson J, Rabito FA, Grimsley F, Iqbal S, Reponen T, Muilenberg ML, Thorne PS, Dearborn DG, Morley RL. Mold and endotoxin levels in the aftermath of Hurricane Katrina: a pilot project of homes in New Orleans undergoing renovation. Environmental Health Perspectives. 2006. Aug 24;114(12):1883–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Institute of Medicine (US) Roundtable on Environmental Health Sciences, Research, and Medicine. Environmental Public Health Impacts of Disasters: Hurricane Katrina: Workshop Summary. [PubMed]
  • 23.Chakraborty J, Maantay JA, Brender JD. Disproportionate proximity to environmental health hazards: methods, models, and measurement. American Journal of Public Health. 2011. Dec;101(S1):S27–36. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Collins TW, Grineski SE, Chakraborty J, Flores AB. Environmental injustice and Hurricane Harvey: a household-level study of socially disparate flood exposures in Greater Houston, Texas, USA. Environmental research. 2019. Dec 1;179:108772. [DOI] [PubMed] [Google Scholar]
  • 25.Chakraborty J, Grineski SE, Collins TW. Hurricane Harvey and people with disabilities: disproportionate exposure to flooding in Houston, Texas. Social Science & Medicine. 2019. Apr 1;226:176–81. [DOI] [PubMed] [Google Scholar]
  • 26.Collins TW, Grineski SE, Chakraborty J, Montgomery MC, Hernandez M. Downscaling environmental justice analysis: Determinants of household-level hazardous air pollutant exposure in Greater Houston. Annals of the Association of American Geographers. 2015. Jul 4;105(4):684–703. [Google Scholar]
  • 27.Karaye I, Stone KW, Casillas GA, Newman G, Horney JA. A Spatial Analysis of Possible Environmental Exposures in Recreational Areas Impacted by Hurricane Harvey Flooding, Harris County, Texas. Environmental management. 2019. Oct 1;64(4):381–90. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28.Horney JA, Casillas GA, Baker E, Stone KW, Kirsch KR, Camargo K, Wade TL, McDonald TJ. Comparing residential contamination in a Houston environmental justice neighborhood before and after Hurricane Harvey. PloS one. 2018. Feb 8;13(2):e0192660. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Federal Emergency Management Administration (FEMA) (2018). FEMA - Harvey Flood Depths Grid, HydroShare, 10.4211/hs.165e2c3e335d40949dbf501c97827837. Accessed August 6, 2019. [DOI]
  • 30.Yost K, Perkins C, Cohen R, Morris C, Wright W. Socioeconomic status and breast cancer incidence in California for different race/ethnic groups. Cancer Causes & Control. 2001. Oct 1;12(8):703–11. [DOI] [PubMed] [Google Scholar]
  • 31.Yu M, Tatalovich Z, Gibson JT, Cronin KA. Using a composite index of socioeconomic status to investigate health disparities while protecting the confidentiality of cancer registry data. Cancer Causes & Control. 2014. Jan 1;25(1):81–92. [DOI] [PubMed] [Google Scholar]
  • 32.Ari P 2018. Preparing for the next storm: Learning from the man-made environmental disasters that followed Hurricane Harvey. https://www.environmentalintegrity.org/wp-content/uploads/2018/08/Hurricane-Harvey-Report-Final.pdf. Accessed September 2, 2019.
  • 33.Nixon R Slow Violence and the Environmentalism of the Poor. Harvard University Press; 2011. [Google Scholar]
  • 34.Blum J 2017. Failures of floating-roof oil tanks during Harvey rase concerns. October 11 2017. https://www.houstonchronicle.com/business/energy/article/Failures-of-floating-roof-tanks-during-Harvey-12269513.php. Accessed September 2, 2019.
  • 35.Misuri A, Moreno VC, Quddus N, Cozzani V. Lessons learnt from the impact of hurricane Harvey on the chemical and process industry. Reliability Engineering & System Safety. 2019. Oct 1;190:106521. [Google Scholar]
  • 36.Nicole W Wristbands for Research: Using Wearable Sensors to Collect Exposure Data after Hurricane Harvey. 2018. [DOI] [PMC free article] [PubMed] [Google Scholar]

RESOURCES