Utilization of the Maryland Environmental Justice Screening Tool: A Bladensburg, Maryland Case Study

Maryland residents’ knowledge of environmental hazards and their health effects is limited, partly due to the absence of tools to map and visualize distribution of risk factors across sociodemographic groups. This study discusses the development of the Maryland EJSCREEN (MD EJSCREEN) tool by the National Center for Smart Growth in partnership with faculty at the University of Maryland School of Public Health. The tool assesses environmental justice risks similarly to the U.S. Environmental Protection Agency’s (USEPA) EJSCREEN tool and California’s tool, CalEnviroScreen 3.0. We discuss the architecture and functionality of the tool, indicators of importance, and how it compares to USEPA’s EJSCREEN and CalEnviroScreen. We demonstrate the use of MD EJSCREEN through a case study on Bladensburg, Maryland, a town in Prince George’s County (PG) with several environmental justice concerns including air pollution from traffic and a concrete plant. Comparison reveals that environmental and demographic indicators in MD EJSCREEN most closely resemble those in EPA EJSCREEN, while the scoring is most similar to CalEnviroScreen. Case study results show that Bladensburg has a Prince George’s environmental justice score of 0.99, and that National Air Toxics Assessment (NATA) air toxics cancer risk is concentrated in communities of color.


Introduction
Environmental justice research since the 1970s has illustrated siting disparities of environmental hazards and locally unwanted land uses (LULUs) in communities of color [1][2][3][4][5]. Specifically, these neighborhoods are disproportionately burdened by noxious facilities that threaten the social, environmental, and physical health of local residents [5][6][7][8][9]. There are a number of environmental stressors that affect the physical and social well-being of a population such as: (1) proximity to hazardous waste sites, (2) exposure to air and water pollution, (3) residential crowding, (4) high levels of ambient noise, (5) the work environment, and (6) quality of local schools [8][9][10][11]. These harmful exposures have been examined using Geographic Information System (GIS) and Public Participatory Geographic Information System (PPGIS) mapping tools. The use of GIS and PPGIS in public health research has produced a wealth of information that can be used to improve the quality of life in overburdened neighborhoods.
The majority of the stakeholders included Prince George's County Environmental Action Council (EAC) members, Port Towns residents, the environmental justice legislative team (EJLT), and the statewide Commission on Environmental Justice and Sustainable Communities. In order to obtain feedback on the indicators that were deemed necessary and acceptable for MD EJSCREEN, posters containing lists of different indicators were displayed at the stakeholder meetings. Surveys were also distributed to all stakeholders to gather feedback regarding the decidedly necessary and acceptable indicators that were to be highlighted in MD EJSCREEN.
The first stakeholder meeting occurred with the Prince George's County EAC, an organization that presents feedback on environmental concerns to the Prince George's County Department of the Environment. The feedback concerning each indicator was documented on flashcards and shared with the EJSCREEN development team. This information was used to build MD EJSCREEN.
Posters highlighting the relevance of each specified indicator were displayed at other stakeholder meetings. For instance, posters were displayed on site at Bladensburg Waterfront Park, and attendees were instructed to rank the importance of each indicator in the community. Four stickers were used to capture feedback from the stakeholders on listed indicators: (1) green, (2) blue, (3) yellow, and (4) red symbolizing very important, important, somewhat important, and not important, respectively. Unfortunately, the significance of each colored sticker became inconsistent for each individual stakeholder. To account for this issue, we removed the important/not important value from each sticker, and instead counted the physical number of stickers placed on each indicator to properly reflect the significance of each indicator. In the development of MD EJSCREEN, community members advocated for the inclusion of seven indicators to highlight the aspects of economic, environmental, and exposure factors that can significantly alter community health. The seven indicators were: (1) asthma emergency room discharges; (2) percent non-White; (3) proximity to treatment, storage, and disposal facilities (TSDFs); (4) myocardial infarction discharges; (5) low birth weight infants; (6) particulate matter (PM 2.5 ); and (7) pathogenic infrastructure. Of these indicators, only pathogenic infrastructure was not mapped due to lack of sufficient data.
Additional feedback was obtained through surveys that asked community members to prioritize indicators taken from EPA EJSCREEN, CalEnviroScreen, and new indicators suggested by community members. A Likert scale was used to rank indicators by importance, in which a score of 1 indicated that said indicator was of low priority and a score of 5 indicated that it was of high priority. These surveys were distributed to several stakeholder groups including, the Association of Baltimore Area Grantmakers (ABAG), the Commission on Environmental Justice and Sustainable Communities, the Prince George's County EAC, and the EJLT.

Scoring Process
MD EJSCREEN illustrates the extent of environmental injustice in an area by assigning an EJ Score to each census tract. To calculate the MD EJ score, the tool applies the methodology used by CalEnviroScreen [8]. A MD EJ score is calculated by first converting the raw indicator scores within each of the four categories: (1) Pollution Burden: Exposure; (2) Pollution Burden: Effect; (3) Population Characteristics: Sensitive Populations; and (4) Population Characteristics: Socioeconomic Factors into percentiles. The raw indicator scores are ranked from lowest to highest at the censustract level, and then assigned percentile values from 0 to 1 based on how the indicator scores compare to other census tracts in the state. MD EJSCREEN also produces a PG EJ score, which applies the same methodology, but assigns percentile values from 0 to 1 based on how the indicator scores compare to other census tracts in Prince George's County.
These values are averaged to create two scores per census tract: (1) the Population Characteristics score, and (2) the Pollution Burden score. The Population Characteristics score is the average of all indicators in the Sensitive Populations and Socioeconomic Factors categories. The Pollution Burden score is the average of all indicators in the Environmental Effects and Exposures categories. In this calculation; however, the Environmental Effects component is half-weighted. This means that the Environmental Effects category has a weight equal to 1/3, while the Environmental Exposure category has a weight equal to 2/3. This is done because Environmental Effects make a smaller contribution to Pollution Burden than Environmental Exposures.
To calculate the EJ scores, the Population Characteristics score is multiplied by the Pollution Burden score [38]. This score is reported as a percentile value 0 to 1 based on how it compares to other census tracts. An EJ score illustrates the magnitude of environmental justice concern in an area; thus, areas with high concentrations of low-income and/or non-White populations and high exposures to environmental hazards will have scores closer to 1.

Comparison of MD EJSCREEN to CalEnviroScreen and EPA EJSCREEN
While MD EJCREEN, CalEnviroScreen, and EPA EJSCREEN all aid users in increasing environmental awareness, there are a number of differences in the usability, functionality, and type of information used in each tool. For instance, EPA EJSCREEN contributes useful demographic and environmental indicator data; however, the broad national scope of the tool limits the specificity and relevance of a given block group or census tract. CalEnviroScreen on the other hand, provides specific population characteristics and pollution burden data for the state for California and more importantly, has proved to be an apt model for MD EJSCREEN. However, CalEnviroScreen is missing pertinent information on Maryland specific indicators such as lead paint prevalence.

Scoring Process
MD EJSCREEN illustrates the extent of environmental injustice in an area by assigning an EJ Score to each census tract. To calculate the MD EJ score, the tool applies the methodology used by CalEnviroScreen [8]. A MD EJ score is calculated by first converting the raw indicator scores within each of the four categories: (1) Pollution Burden: Exposure; (2) Pollution Burden: Effect; (3) Population Characteristics: Sensitive Populations; and (4) Population Characteristics: Socioeconomic Factors into percentiles. The raw indicator scores are ranked from lowest to highest at the census-tract level, and then assigned percentile values from 0 to 1 based on how the indicator scores compare to other census tracts in the state. MD EJSCREEN also produces a PG EJ score, which applies the same methodology, but assigns percentile values from 0 to 1 based on how the indicator scores compare to other census tracts in Prince George's County.
These values are averaged to create two scores per census tract: (1) the Population Characteristics score, and (2) the Pollution Burden score. The Population Characteristics score is the average of all indicators in the Sensitive Populations and Socioeconomic Factors categories. The Pollution Burden score is the average of all indicators in the Environmental Effects and Exposures categories. In this calculation; however, the Environmental Effects component is half-weighted. This means that the Environmental Effects category has a weight equal to 1/3, while the Environmental Exposure category has a weight equal to 2/3. This is done because Environmental Effects make a smaller contribution to Pollution Burden than Environmental Exposures.
To calculate the EJ scores, the Population Characteristics score is multiplied by the Pollution Burden score [38]. This score is reported as a percentile value 0 to 1 based on how it compares to other census tracts. An EJ score illustrates the magnitude of environmental justice concern in an area; thus, areas with high concentrations of low-income and/or non-White populations and high exposures to environmental hazards will have scores closer to 1.  [35,36]. X X X Traffic Proximity and Volume Count of vehicles (average annual daily traffic) at major roads within 500 meters or close to 500 meters, divided by distance in meters [35,36]. X X X

Pesticide Use
Total pounds of selected active pesticide ingredients (filtered for hazard and volatility) used in production-agriculture per square mile, averaged over three years (2012 to 2014) [36]. X

Drinking Water Contaminants
Water tested to contain one or more contaminants listed in 'Update to California Communities Environmental Health Screening Tool'. Reported as yearly averages of chemical contaminant concentrations for each census tract [36].

X Toxic Releases from Facilities
Toxicity-weighted concentrations of modeled chemical releases to air from facility emissions and off-site incineration (averaged over 2011 to 2013) [36]. X

Pollution Burden: Environmental Effects
Lead Paint Indicator Percent of houses built before 1960, which likely contain lead paint [36]. X X

Proximity to Risk Management Plan (RMP) Sites
Count of RMP (potential chemical accident management plans) facilities within 5 kilometers or close to 5 kilometers, divided by distance in kilometers [36]. X X

Proximity to Treatment Storage and Disposal Facilities (TSDF)
Count of TSDF (hazardous waste management facilities) within 5 kilometers or closest to 5 kilometers, divided by distance in kilometers [36]. X X

Proximity to National Priorities List (NPL) Sites
Count of NPL/Superfund sites (polluted sites that pose a risk to human health and/or the environment) within 5 kilometers or close to 5 kilometers, divided by distance in kilometers [35,36].
X X X  [39]. X Groundwater Threat Nature and the magnitude of the threat and burden to groundwater safety posed by sites maintained in GeoTracker [35]. X Impaired Water Bodies Contamination of streams, rivers, and lakes by pollutants which compromise the ability to use a body of water for drinking, swimming, fishing, aquatic life protection, etc. [35]. X Solid Waste Sites and Facilities Solid waste landfills, composting, and recycling facilities [35]. X

Population Characteristics: Sensitive Populations
Asthma Emergency Discharges Count of patients released from the hospital after being admitted for asthma or asthma-related distress [40]. X

Myocardial Infarction Discharges
Patients released from the hospital after being admitted for a heart attack or heart attack symptoms [35]. X X Low Birth Weight Infants Babies born weighing less than 5.5 pounds [35]. X X Asthma Emergency Visits Patients admitted to the emergency room for asthma or asthma-related distress [35]. X

Population Characteristics: Socioeconomic Factors
Percent Non-White Percentage of individuals who define themselves as any race/ethnicity besides non-Hispanic White [35,36]. X X X Percent Low-Income Percentage of individuals whose household income in the past 12 months is less than two times below the federal poverty level [35,36]. X X X Less than high school education Percentage of individuals 25 and older who lack a high school diploma [35,36]. X X X Linguistic Isolation Percentage of households in which no one 14 years old and older speaks English "very well", or households which speak only English [35,36]. X X X Individuals under age 5 Percentage of people under the age of 5 [36]. X X Individuals over age 64 Percentage of people over the age of 64 [36]. X X

Unemployment
Percentage of the population over the age of 16 that is unemployed and eligible for the labor force. Excludes retirees, students, homemakers, institutionalized persons except prisoners, those not looking for work, and military personnel on active duty [35].

X X Housing Burdened Low Income Households
Percentage of households in a census tract that make less than 80% of the HUD Area Median Family Income and paying greater than 50% of their income to finance housing [35].

Comparison of MD EJSCREEN to CalEnviroScreen and EPA EJSCREEN
While MD EJCREEN, CalEnviroScreen, and EPA EJSCREEN all aid users in increasing environmental awareness, there are a number of differences in the usability, functionality, and type of information used in each tool. For instance, EPA EJSCREEN contributes useful demographic and environmental indicator data; however, the broad national scope of the tool limits the specificity and relevance of a given block group or census tract. CalEnviroScreen on the other hand, provides specific population characteristics and pollution burden data for the state for California and more importantly, has proved to be an apt model for MD EJSCREEN. However, CalEnviroScreen is missing pertinent information on Maryland specific indicators such as lead paint prevalence.
NATA respiratory hazard index, and Proximity to risk management plan (RMP) sites. A comparison of the usability and functionality of each tool is provided in Table 3.
In terms of specific indicators, the tools have a number of similarities and differences. Table 1  In the development of MD EJSCREEN, feedback was gathered on what specific indicators stakeholders felt should be included in the tool, including indicators that highlight the aspects of economic, environmental, and exposure factors that can significantly alter community health. Table 2 includes additional indicators that were advocated for by community members, some of which are already in the 'Additional Context Layers' category in MD EJSCREEN. Once sufficient data is obtained for the indicators not already included in the tool, they will be added to this category.
In addition to the differences in indicators mapped by the three tools, shown in Table 1 above, CalEnviroScreen, EPA EJSCREEN and MD EJSCREEN also differ in their usability and functionality as shown in Table 3. In our analysis of usability and functionality of the three tools, we identified the main features of GIS tools including searching locations, navigation (zoom in/out), printing maps, sharing maps, and creating maps [64]. We also compared additional features that we identified by searching through the tools. We then classified these features as assets of usability or functionality. We defined usability as, "the extent to which a product can be used by specified users to achieve specific goals with effectiveness, efficiency and satisfaction in a specified context of use" [65]. We expand on this definition by referring to aspects of the tool that can be controlled by the user as 'usability'. Table 2. Additional indicators in MD EJSCREEN tool promoted by stakeholders.

Indicator Description Data Source and Year
Pathogenic Infrastructure Features of the built environment that increase a population's vulnerability to chemical and non-chemical stressors leading to adverse health outcomes. For example, liquor stores, fast food restaurants, convenience stores, pawn shops, and payday lenders [41].
American Society of Civil Engineers (ASCE), 2017 Salutogenic Infrastructure Physical, economic, natural, social, and spiritual features of the environment that foster health and nourish wellness. For example, hospitals, primary care providers, grocery stores, parks, recreational facilities, and churches [41].

ASCE, 2017
Tree Canopy Coverage Layer of leaves, branches, and stems of trees that cover the ground when viewed from above. Refers to the quantity and quality of trees in a specific geographic area [42].
Maryland Department of Natural Resources (DNR), 2017

Brownfields
Refer to any area that is saturated by water, either permanently or seasonally (e.g., swamps, marshes, estuaries, lakes, rivers, etc.) Can include data from monitoring sites, the location and scope of restoration efforts from Maryland's Non-Tidal Wetland Mitigation Program and the distribution of funding provided through the Federal Clean Water Act Section 319 (h) [43].

USEPA, 2017
Wetlands and Waterways Refer to any area that is saturated by water, either permanently or seasonally (e.g., swamps, marshes, estuaries, lakes, rivers, etc.) [44]. Can include data from monitoring sites, the location and scope of restoration efforts from Maryland's Non-Tidal Wetland Mitigation Program and the distribution of funding provided through the Federal Clean Water Act Section 319(h) [45].

USEPA, 2017
Health & Environmental Advocacy Groups Studies have reported a strong, inverse correlation between social capital and health inequalities [46]. Local health and environmental advocacy groups promote access to social capital and give voice to marginalized members of the community [47].

Indicator Description Data Source and Year
Point Source Discharge Any identifiable source from which pollutants are discharged. For example, a pipe, ditch, channel, tunnel, conduit, well, discrete fissure, container, rolling stock, concentrated animal feeding operation, or vessel or other floating craft. Agricultural stormwater discharges and return flows from irrigated agriculture are not included in this definition [54].
EPA Clean Water Act (CWA), 1972 Major Air Pollution Sources Any source that emits 10 tons of any of the 187 toxic air pollutants listed in the Clean Air Act, or 25 tons of a mixture of air toxins, per year. There are multiple sources of pollution such as mobile (cars, trucks, and busses), stationary, (factories, refineries, and power plants), indoor (select building materials and cleaning solvents), and naturally-occurring sources (volcanic eruptions and forest fires) [55].
EPA Clean Air Act (CAA), 1992 Bladensburg Air Pollution Minor Any source which emits, or has the potential to emit, regulated New Source Review (NSR) pollutants in amounts less than the major source thresholds [56]. CAA, 1992 Hazardous Waste Site Any area which is contaminated by pollutants deemed dangerous or capable of having harmful effects on human health or the environment. More specifically, any area contaminated by chemicals on the F, K, P, and U lists, found in title 40, section 261, of the Code of Federal Regulations (CFR) [57].
Health Provider Shortage Area A shortage of primary care and dental providers for the entire population or specific population groups within a defined geographic area [58].

Health Resources and Services Administration (HRSA), 2016
Supermarkets Any establishment which retails a general line of food, such as canned, dry and frozen foods, fresh fruits and vegetables, fresh and prepared meats, fish, poultry, dairy products, baked products and snack foods. These establishments may also retail non-food products such as household paper products, toiletries, and non-prescription drugs [59].

NAICS, 2012
Limited Supermarket Access Areas One or more contiguous census block groups where residents must travel significantly further to reach a supermarket than the, "comparative acceptable" distance that residents in well-served areas must travel [60].
The Reinvestment Fund (TRF), 2013 EPA Superfund Sites Any land that has been identified by the USEPA as a candidate for cleanup because of hazardous waste contamination. These sites are placed on the National Priorities List (NPL) [61].
The Comprehensive Environmental Response, Compensation, and Liability Act (CERCLA), 1980 Public Schools Public schools including elementary, middle, and high schools; charter schools; special centers [62].

Prince George's County Public Schools (PGCPS), 2018
Railroads Main lines such as spur lines, and rail yards, mass transit rail lines such as carlines, streetcar track, monorail or other mass transit rail and special purpose rail lines such as cog rail lines, incline rail lines, and trams [63]. All three tools have user-defined search, user-defined location markers, and zoom in/out features. Only EPA EJSCREEN has user-defined base map options, and the ability to bookmark maps in the tool for future reference. Both EPA EJSCREEN and CalEnviroScreen have additional resources such as help manuals or explanatory videos that users may reference. Currently, MD EJSCREEN does not possess these features. A training guide and videos will be released once MD EJSCREEN is in its final stages of development.

US Census Bureau, Department of Commerce, 2015
Functionality is defined as the range of operations that can be performed by the tool. Currently, the features all three tools have in common are their ability to generate coordinates for a user-defined location and to create maps. EPA EJSCREEN can map a number of environmental and demographic indicators, and can generate side-by-side maps in place of overlaying different indicators on the same map. The tool can also map demographic data from the 2012 to 2016 ACS, and the 2000 and 2010 censuses. Additionally, EPA EJSCREEN gives users the option to search for already made maps using GeoPlatform. It also possesses a broad range of other functions such as the option to print custom maps, create custom reports and graphs, measure distances, and download raw data. Moreover, it has a mobile version of the tool that users can access on smartphones and tablets.
CalEnviroScreen allows users to map each individual indicator or CalEnviroScore in an individual window, with no option to overlay maps or make side by side maps. It also provides users with the options to print maps and download raw data, but also includes a 'share' feature that allows users to share information over social media, statistics, and create scores. While CalEPA does produce a general report for the state of California with the 2018 update of CalEnviroScreen, the maps and data presented in the report cannot be customized by the user [35]. As expected for a provisional tool, MD EJSCREEN has a relatively limited range of capabilities, and has no print or share feature. However, it is able to overlay indicators with one another, allowing for easy visual comparison. It also displays legislative districts and county lines, allowing users to easily identify state and local governments to which they can address concerns.

Bladensburg: A Case Study
Bladensburg is a town in Prince George's County that has experienced environmental injustice due to racial and economic stratification. The Bladensburg population is primarily Black (62.7%) and Latinx (33.0%), with small concentrations of White (13.9%) and Asian (1.6%) residents [66]. Studies have shown that proximity to hazards is directly related to race/ethnicity, with one study specific to Maryland, Louisiana, and West Virginia finding that African-Americans are more likely to be located in close proximity to Toxic Release Inventory (TRI) facilities [67][68][69][70][71]. In comparison to the national average of 14%, 20.1% of Bladensburg residents live below the federal poverty line [66]. Due to underlying social and economic vulnerabilities, residents in Bladensburg are more likely to experience low property values and the disproportionate siting of environmental hazards in their communities than their more affluent counterparts [72,73]. Bladensburg residents also endure a large pollution burden, with heavy commuter and industrial traffic and a concrete block plant acting as sources of particulate matter (PM), volatile organic compounds (VOCs), and polycyclic aromatic hydrocarbons (PAHs) [74]. Due to their exposure to air pollutants, Bladensburg residents are at increased risk for respiratory problems and cancer [72,[75][76][77]. An example analysis of environmental injustice in Bladensburg using MD EJSCREEN is presented below (Figures 2-6). For the purpose of the Bladensburg case study, the PG EJ score, a separate score created by the tool, is used, in addition to the MD EJ score (Figure 2). The PG EJ score compares the Bladensburg census tract to the rest the census tracts in Prince George's County, while the MD EJ score compares the Bladensburg census tract to all census tracts in Maryland. A comparison with EPA EJSCREEN is provided for Bladensburg in Figure 7.     For example, this map illustrates the PG EJ score for Bladensburg. The PG EJ score is represented as a percentile, meaning that Bladensburg has an EJ score higher than 99% of the census tracts in Prince George's County. Furthermore, the score is an overall indication of the prevalence of environmental hazards in the Bladensburg area.  This map visualizes the percent non-White population in Bladensburg. Bladensburg appears within the 0.75-0.90 percentile range, meaning that its percent non-White population is higher than 75-90% of the census tracts in the state or county.  This map visualizes the percent non-White population in Bladensburg. Bladensburg appears within the 0.75-0.90 percentile range, meaning that its percent non-White population is higher than 75-90% of the census tracts in the state or county.  This map visualizes NATA air toxics cancer risk in Bladensburg. Bladensburg appears within the 0.9-1 percentile range, meaning that its NATA air toxics cancer risk is higher than 90-100% of the census tracts in the state or county. This map visualizes the percent non-White population in Bladensburg in relation to NATA air toxics cancer risk. Bladensburg appears within the 0.9-1.0 percentile range, meaning that the calculated risk of developing cancer due to air pollution is higher than the risk in 90-100% of the census tracts in the state or county. Coupled with the high percentage non-White population, these findings illustrate the disproportionate exposure to human and environmental health hazards experienced by populations of color. This map visualizes the percent non-White population in Bladensburg in relation to NATA air toxics cancer risk. Bladensburg appears within the 0.9-1.0 percentile range, meaning that the calculated risk of developing cancer due to air pollution is higher than the risk in 90-100% of the census tracts in the state or county. Coupled with the high percentage non-White population, these findings illustrate the disproportionate exposure to human and environmental health hazards experienced by populations of color.  For example, this map illustrates the PG EJ score for Bladensburg. The PG EJ score is represented as a percentile, meaning that Bladensburg has an EJ score higher than 99% of the census tracts in Prince George's County. Furthermore, the score is an overall indication of the prevalence of environmental hazards in the Bladensburg area.
This map visualizes the percent non-White population in Bladensburg. Bladensburg appears within the 0.75-0.90 percentile range, meaning that its percent non-White population is higher than 75-90% of the census tracts in the state or county.
This map visualizes NATA air toxics cancer risk in Bladensburg. Bladensburg appears within the 0.9-1 percentile range, meaning that its NATA air toxics cancer risk is higher than 90-100% of the census tracts in the state or county.
This map visualizes the percent non-White population in Bladensburg in relation to NATA air toxics cancer risk. Bladensburg appears within the 0.9-1.0 percentile range, meaning that the calculated risk of developing cancer due to air pollution is higher than the risk in 90-100% of the census tracts in the state or county. Coupled with the high percentage non-White population, these findings illustrate the disproportionate exposure to human and environmental health hazards experienced by populations of color.
While MD EJSCREEN conducts analyses at a larger scale by organizing data by census tract, EPA EJSCREEN analyzes data on a smaller scale, by block group. According to Figure 7, this block group is 97% non-White and it is in the 95th percentile. This means that the percent non-White in this block group is greater than or equal to the percent non-White where 95% of the US population lives. The lifetime NATA cancer risk for this block group is 55 people per 1 million. This area is also in the 92nd percentile, meaning that the lifetime NATA cancer risk in this block group is greater than or equal to the risk where 92% of the US population lives.

Discussion
Despite the presence of tools such as EPA EJSCREEN and CalEnviroScreen, stakeholders expressed the need for a mapping tool which caters to specific environmental health concerns in Maryland. Other EJ mapping tools lack pertinent information on indicators specific to Maryland communities such as lead paint prevalence, NATA respiratory hazard index, and proximity to RMP sites. EPA EJSCREEN can map useful demographic and environmental indicators, but is hindered by its broad scope, and the use of block groups as the geographic unit of analysis.
EPA EJSCREEN was not developed with the intention of determining the presence or absence of environmental injustice in communities. All conclusions drawn from the tool must be substantiated with additional information at the local level in order to perform a true assessment of environmental justice. Due to the broad scope of the tool, it is unable to address all issues relevant at the local level. Therefore, many environmental concerns central to communities are not shown.
EPA EJSCREEN also identifies marginal uncertainty in EJ Index values when mapping block groups. Mapping an individual block group can identify pollution 'hot spots', but produces uncertainty because the tool must estimate the location of residences. Due to such uncertainties, the tool cannot confidently compare or rank indicators when only minor differences exist between percentile scores. In order to avoid such uncertainties, the US EPA recommends that users create a buffer by applying the tool to a larger geographic area that covers multiple block groups, but which may overlook 'hot spots' [36]. We seek to address this weakness in MD EJSCREEN by conducting analysis at a larger scale (using a census tract), rather than by block group.
As mentioned previously, only CalEnviroScreen and MD EJSCREEN are able to produce EJ scores for a user-specified census tract. The computation of an EJ score allows for clear and succinct analysis of the level of environmental injustice in an area of concern. EPA EJSCREEN cannot produce cumulative EJ scores but does produce EJ indices. An EJ index combines demographic indicators with a single environmental indicator, resulting in 11 EJ indexes that reflect the 11 environmental indicators in the tool. EJ indexes apply the idea of 'excess risk' by examining the disparities between block group averages and nation averages for environmental indicators across different demographic indicators. EJ indices can therefore be used to identify geographic locations which may burdened by environmental hazards. Despite this, EJ indices are limited in their applications to environmental justice because they do not consider the cumulative impacts of multiple environmental stressors on the health and well-being of a population.
Research has repeatedly shown that the concentration of environmental hazards elevates stress and other adverse health outcomes, while simultaneously decreasing quality of life and community sustainability [8,[78][79][80]. Environmental stressors include water contamination, hazardous waste, land used for incinerators and landfills, and lack of green space [78]. MD EJSCREEN provides a more accessible and interactive way for residents to grasp how their health is affected by the built environment, visualize trends in who is directly affected and to what degree, and create long-lasting change in their community.
The US EPA has set environmental health standards, yet, these programs have not specifically confronted the issue of cumulative impacts on non-White populations [81]. This makes mapping tools such as EJSCREEN, CalEnviroScreen, and MD EJSCREEN significant in determining how populations of color suffer an unjust burden of environmental risks. The scores available in MD EJSCREEN allow residents and other stakeholders to compare the rates of pollution, the effects of pollution, and the most vulnerable subpopulations to other areas of the state. For instance, users can easily visualize that areas with a higher percentage of low-income residents also tend to be located near areas with high traffic volume and PM 2.5 , as well as the potential health risks associated with those locations [81]. These scores can be used by residents to better advocate for their health, and government officials to determine specific changes that need be made to reduce environmental injustice.
The ability of MD EJSCREEN to identify areas of environmental justice concern has been mirrored by other tools such as CalEnviroScreen. Researchers utilizing this technology found that non-White individuals are more likely to live near pollution sources than non-Hispanic Whites, and that they also experience higher concentrations of poverty, ozone, DPM, pesticide use, solid waste sites, and gas-fired power plants [8,82]. Additional studies isolating the Latinx population found that they are especially vulnerable to environmental injustice due factors such as linguistic isolation, racial segregation, systematic housing market discrimination, real estate steering practices, and blocked channels of residential and economic mobility [83,84]. Additionally, socioeconomic indicators such as education and income were found to be strongly associated with levels of disease burden [10,85].
CalEnviroScreen has also been used to link vulnerable populations to visionary policies capable of transforming overburdened areas into healthy, thriving communities. For example, California Environmental Justice Alliance (CEJA)'s Green Zones project applies the CalEnviroScreen tool to identify concentrations of industrial pollution sources thereby streamlining resources, regulatory attention, and sustainable economic development to vulnerable communities [86]. The term 'Green Zone', refers to mostly low-income communities of color that have organized against discriminatory land use patterns and for neighborhood restoration that addresses economic, social, and environmental health concerns [86].
The EPA's Clean Power Plan employed EPA EJSCREEN to inform policy that addresses the disproportionate siting of environmental hazards in communities of color, low income, and indigenous communities by performing proximity-based analyses of power plants, taking into account demographics and cumulative impacts [87]. The analyses revealed that in comparison to nation averages, low income and people of color are more likely to live near power plants [83]. The proximity analysis outlined by the Clean Power Plan allows states to improve access to energy efficiency (EE), renewable energy (RE), and financial assistance programs [87].
Similarly, the North Carolina Department of Environmental Quality (NC DEQ) conducted an environmental justice assessment using EPA EJSCREEN for a proposed fly ash project at the Duke Energy Lee plant in Goldsboro, North Carolina. The EPA EJSCREEN analysis revealed that the per capita income of individuals living within two miles of the Lee plant is $17,847 per year, less than the Goldsboro average of $19,243, and the state average of $26,779. This information demonstrates that if the plant is to expand, the pollution burden would fall primarily on low-income communities [88].
EPA EJSCREEN has also been used to address the environmental health concerns of incarcerated populations. Currently, the tool is at the forefront of the 'prison ecology movement', which utilizes GIS mapping to examine the proximity of correctional facilities to known hazardous waste sites [89]. Studies performed by the Prison Ecology Project (PEP) found that among a range of federal and state prisons from Colorado to Indiana, 589 were located within three miles of a Superfund or hazardous waste site [90].
In conducting this study, we encountered multiple barriers to obtaining reliable feedback on the indicators. The amount of information we received for each indicator was not as complete as it could have been due to low response rates from stakeholders. Other limitations include the accuracy of the feedback received during the stakeholder meeting held in Bladensburg, Maryland. Due to these limitations, the utility of the additional indicators included within the 'Additional Context' category to the Prince George's community, may be lower than anticipated. Nevertheless, the findings of this study can be further implemented in various future EJ investigations.
Presently, we are working to increase the functionality and usability of MD EJSCREEN so it has similar capabilities to EPA EJSCREEN and CalEnviroScreen particularly geostatistical analysis. In the future, we will add the ability of users to upload their own qualitative or quantitative data collected via citizen science or another scientific approach. Additionally, MD EJSCREEN will be expanded to the entire state of Maryland, enabling more communities to engage with and take control of environmental justice in their neighborhoods. As the tool expands, more pertinent indicators specific to Maryland could be incorporated based on feedback from stakeholders across the state, strengthening the tool and its ability to combat environmental injustice. Although no stakeholder evaluations have been conducted yet, once the tool is in its final stage of development, we will perform an evaluation of stakeholder satisfaction with the tool. These evaluations will focus on the major metrics used to assess the functionality and usability of GIS mapping tools. This will allow us to understand if the tool has achieved its primary objective: to represent the environmental justice concerns specific to Maryland residents.

Conclusions
MD EJSCREEN can be used to make public health improvements for all communities, allowing residents to advocate for new policies and better enforcement of policies. In addition, government officials can use the information provided by this tool to identify pressing concerns of their constituents and implement more equitable policies. A comprehensive training guide and video showing how to navigate and use the tool will be developed and released for community members and stakeholders. A series of training sessions will serve to provide technical assistance on an as-needed basis, making it easier to apply MD EJSCREEN in policy decisions and community advocacy.
Funding: This research was funded by the National Center for Smart Growth's Partnership for Action Learning in Sustainability (PALS).