&EPA
EPA/600/R-14/364| September 2014| www.epa.gov/ord
United States
Environmental Protection
Agency
R9 RARE Final Project Report Title:
Partnering with Environmental Agencies
and Communities to Pilot Use of the
Environmental Justice Screening Method (EJSM)
Cumulative Impacts Tool
Prepared by Eric S. Hall, James Sadd,
Rachel Morello-Frosch, and Manuel Pastor
Office of Research and Development
National Exposure Research Laboratory
-------
-------
EPA/600/R-14/364| September 2014| www.epa.gov/ord
United States
Environmental Protection
Agency
R9 RARE Final Project Report Title:
Partnering with Environmental Agencies
and Communities to Pilot Use of the
Environmental Justice Screening Method (EJSM)
Cumulative Impacts Tool
Prepared by Eric S. Hall, James Sadd,
Rachel Morello-Frosch, and Manuel Pastor
I Office of Research and Development
I National Exposure Research Laboratory
-------
R9 RARE Final Project Report
Title: Partnering with Environmental Agencies and Communities to Pilot Use of the
Environmental Justice Screening Method (EJSM) Cumulative Impacts Tool
US EPA ORD Project Officer:
Eric S. Hall
US EPA R9 Technical Advisors:
Debbie Lowe Liang. Jacquelyn Hayes, and Charles Swanson
EPA Region 9 QA Program Manager:
Eugenia McNaughton
Prepared by:
Eric S. Hall, James Sadd, Rachel Morello-Frosch, Manuel Pastor
Principal Investigators:
James Sadd, Manuel Pastor. Rachel Morello-Frosch
Project QA Manager:
James Sadd
University Collaborators/Organizations:
Occidental College. University of Southern California (USC), University of California at Berkeley (UC Berkeley)
Cooperative Agreement Number:
AE 8352800-01
Project Period:
September 1, 2012 to September 30, 2014
Disclaimer:
The information in this document has been funded wholly or in part by the U. S.
Environmental Protection Agency under cooperative agreement number AE 8352800-
01 to Occidental College under the R9 RARE Program. It has been subjected to review-
by the National Exposure Research Laboratory and approved for publication. Approval
does not signify that the contents reflect the views of the Agency, nor does mention of
trade names or commercial products constitute endorsement or recommendation for use.
-------
This is the final report of a 24-month research project. This
project was awarded as a $75,000.00 cooperative agreement
to Occidental College. The project began on 1 September
2012, following final approval of the funding request package
by the EPA Grants Office (US EPA/GIAMD). This project is
a component of EPA's Sustainable and Health Communities
Research Program (SHCRP) and is a deliverable FY2014
research product.
-------
-------
This research project was designed to assist EPA Region 9
(R9) and EPA ORD in identifying and acquiring information,
tools, and collaborative community-based processes to
help address the concerns and goals expressed by certain
California communities overburdened with the cumulative
impacts of exposures to multiple air pollutants. The EPA
R9 Regional Science Plan describes the need for better
methodologies to help EPA evaluate risks and impacts in
overburdened communities. This project also responds to
requests from a number of California communities wanting
to address their concerns with respect to the impact of
community cumulative impact effects on local, regional, and
statewide regulatory decision-making.
Specifically, this project explored the application of the
Environmental Justice Screening Method (EJSM), including
its data sources, analytical procedures and design, and
validation procedures, to help regulators and policy makers
efficiently target their efforts to remediate cumulative
impacts, environmental inequities, and re-focus regulatory
action at the neighborhood level. The EJSM was developed
and applied in a previous R9 RARE Project, which ended
in July 2010 (Hall, E.S., Morello-Frosch, R.A., et al., 2011).
The initial use of EJSM was in determining if certain socio-
economic (vulnerability) indicators had an impact on the
association between ambient air pollution exposure and
certain adverse birth outcomes (e.g., low birth weight [< 2500
grams] and pre-term birth [<37 weeks]).
Currently, the burden of proof is placed on communities to
demonstrate the cumulative impacts of environmental and
social stressors. Cumulative Impact (CI) screening tools such
as the EJSM provide environmental policy and program
decision makers with a more proactive approach towards
removing this burden from vulnerable communities so that
those communities, usually without the history of or capacity
for civic engagement, can obtain regulatory relief for the
environmental exposures they experience. The EJSM can
also advance regulatory evaluation and the implementation of
environmental policies. As a proof-of-concept of the EJSM
capabilities, and to develop guidelines about how it can/
should be used, this project applied the EJSM to two pilot
applications in two different California pilot communities/
regions. Each region was chosen so that the EJSM could act
in a supporting role to address and resolve policy relevant
questions. The goal was to influence projects of local
importance facing cumulative impacts and vulnerability
considerations due to community exposures to environmental
hazards and pollution.
In this project, there were two completely different
approaches used when applying EJSM to each pilot study
region in California. In the Pilot Application 1, EJSM was
used in the City of Commerce. In this pilot study, EJSM
was the sole Environmental Justice (EJ) Cumulative Impacts
(CI) tool used to assist the City of Commerce Green Zone
Working Group in developing policy recommendations for:
a) creating buffer zones between environmental hazards and
sensitive populations, and; b) developing land use strategies
to create economic development zones for attracting new
businesses and 'green'jobs for community residents.
In the Pilot Application 2, EJSM was used in the eight county
region of the San Joaquin Valley (SJV). Ultimately, three
cities were chosen within the SJV as the locations where
ESJM would be applied: Arvin, Huron, and Stockton. The
SJV is a focus area for environmental justice activity and it
is a geographically diverse region that can benefit from the
application of a variety of CI screening approaches to inform
policymakers on: a) the local and regional patterns of where
high cumulative impact "hot spots" are located and which
communities are disproportionately impacted; b) determining
which policy questions different CI screening approaches are
best suited to answer (i.e., map each approach to the specific
policy[ies] it was designed to address).
In the SJV, EJSM analysis results were compared with two
additional CI screening methods that were developed by the
Center for Regional Change at the University of California
Davis (Cumulative Environmental Vulnerability Assessment
or "CEVA") and the California Environmental Protection
Agency (CalEPA) Office of Environmental Health and
Hazard Assessment (OEHHA) (California Communities
Environmental Health Screening Tool [CalEnviro Screen or
"CES"]). This allowed the pilot communities to target areas
where the three methods agreed on where the cumulative
impacts were the most 'intense'.
VII
-------
-------
AB2588 Air Toxics "Hot Spots" Information and
Assessment Act - Note: the objective of the
AB2588 legislation is to collect emission
data from air toxics sources, identify facilities
with localized impacts, assess health risks and
notify affected individuals
AFSCME American Federation of State, County, and
Municipal Employees
AQMD Air Quality Management District
AVSS Automated Vital Statistics System
CalEPA California Environmental Protection Agency
CARB California Air Resources Board
CASIL California Spatial Information Laboratory
CBPR Community-Based Participatory Research
CCA Coalition for Clean Air
CCEJN Central California Environmental Justice
Network
CDPH California Department of Public Health
CEIDARS California Emission Inventory Development
and Reporting System
CEJA California Environmental Justice Alliance
CES California Communities Environmental Health
Screening Tool (aka: CalEnviroScreen)
CEVA Cumulative Environmental Vulnerability
Assessment
CHAPIS Community Health Air Pollution Information
System
CI Cumulative Impact(s)
CUGU Clean Up Green Up
CVAQ Central Valley Air Quality Coalition
DTSC California Department of Toxic Substance
Control
EJ Environmental Justice
EJA Environmental Justice Analytics
EJAF Environmental Justice Analytics/Analytical
Framework
EJSM Environmental Justice Screening Method
EPA Environmental Protection Agency
ESRI Environmental Systems Research Institute'
EYCEJ East Yard Communities for Environmental
Justice
FOI Facilities of Interest
FTP File Transfer Protocol
GEP Google Earth Pro
GIAMD Grants and Interagency Agreement
Management Division (EPA)
GIS Geographic Information System
GPS Global Positioning System
KML Keyhole Markup Language
NAICS North American Industrial Classification
System
NATA National Scale Air Toxics Assessment
NERL National Exposure Research Laboratory
NTAD National Transportation Atlas Database
OEHHA Office of Environmental Health and Hazard
Assessment (CalEPA)
OPR (the Governor of California's) Office of
Planning and Research
ORD Office of Research and Development
PERE Program for Environmental and Regional
Equity (USC)
QA Quality Assurance
QAPP Quality Assurance Project Plan
QC Quality Control
R9 Region 9 (EPA)
-------
RARE Regionally Applied Research Effort
RSEI Risk Screening Environmental Indicators
SCAG Southern California Association of
Governments
SCAQMD South Coast Air Quality Management District
SHCRP Sustainable and Health Communities Research
Program
SJV San Joaquin Valley
SJV-APCD San Joaquin Valley Unified Air Pollution
Control District
SJV-CHIP San Joaquin Valley Cumulative Health Impacts
Project
TIGER Topologically Integrated Geographic Encoding
and Referencing system
TRI Toxic Release Inventory
UC Berkeley University of California at Berkeley
USC University of Southern California
USGS United States Geological Survey
ZCTA Zip Code Tabulation Area(s)
-------
1.0 Introduction 1
2.0 Project Details 5
3.0 Pilot Application 1: City of Commerce 7
4.0 Pilot Application 2: San Joaquin Valley (SJV) Pilot Community 9
5.0 Pilot Application 1: City of Commerce Working Group Meetings (Recap) 19
6.0 Pilot Application 2: San Joaquin Valley (SJV) Pilot Community Interactions 25
7.0 Pilot Application 1: City of Commerce Working Group Recommendations 29
8.0 Pilot Application 2: San Joaquin Valley (SJV) Site Selection Process 31
9.0 Pilot Application 2: San Joaquin Valley (SJV) Detailed Region/Site Maps 35
9.5 Summary 53
10.0 References 55
APPENDIX A: SJV (Pilot 2 Application) Webinar A-l
APPENDIX B:City of Commerce Planning Commission Green Zone Policy Report B-l
APPENDIX C: Response to Comments - R9 RARE Project Comparing Three Screening Methods
win the San Joaquin Valley C-l
APPENDIX D: City of Commerce Green Zones Working Group Opportunity Areas (Map) D-l
Appendix E: Conference Agenda - Central California Environmental Justice Network:
Roots of Resilience, 2014 E-l
XI
-------
Figure 1. Distribution of screening CI scores for the SJV region by population and area: EJSM 11
Figure 2. Distribution of screening CI scores for the SJV region by population and area: CES 11
Figure 3. Distribution of screening CI scores for the SJV region by population and area: CEVA 11
Figure 4. CES Score (SJV and Statewide) Versus Population (ZCTA) 26
Figure 5. CEVA Score (SJV) versus Population (Census Block Group) and EJSM Score Versus
Population (Census Tract) 27
XII
-------
Map 1. Land Use in the City of Commerce classified by aggregating standard Anderson Land Use urban
classes into groups that reflect the pilot application 7
Map 2. Proposed buffers surrounding sensitive land uses as a land use planning tool to ensure separation
from air pollution point and area sources. Area inside inset rectangle shown in Map 3, below. 8
Map 3. Proposed buffers surrounding sensitive land uses in the Washington/Atlantic Blvd. corridors, a
special focus area recognized by the Working Group both for protection/separation of residential
and sensitive land uses from air pollution sources, but also for business development and "green
design" amenities 8
Map 4a. EJSM scores for the SJV: Mapped using census tracts to allow comparison with other screening
methods 10
Map 4b. CEVA scores for the SJV: Mapped using census block groups for comparison with other
screening methods 10
Map 4c. CES group scores for the SJV: Mapped using 2010 zip code tabulation areas (ZCTAs) for
comparison with other screening methods. ZCTAs that extend outside the eight SJV counties
have been clipped so that colored polygons do not extend outside the boundaries of the defined
comparison area 10
Map 5a. Top Population Quantile - Kern Co 12
Map 5b. Top Population Quantile Central SJV 13
Map 5c. Top Pupulation Quantile Northern SJV 13
Map 6a "High Decile" - Kern Co 15
Map 6b. "High Decile" - Central SJV 16
Map 6c. "High Decile" - Northern SJV 16
Map 7a_l. Fresno and Selma, CA 17
Map 7b_l. Stockton, CA 17
Map 7a_2. Bakersfield 17
Map 7b_2. Modesto, CA 17
Map 7c. Merced (all 3 methods overlap/agree) 18
Map 8. City of Commerce (2013) 21
Map 9. EFFECTIVENESS OF CURRENT CITY ZONING PRACTICE IN SEPARATING AIR
QUALITY HAZARDS FROM SENSITIVE LAND USES [Facilities of Concern - 300 foot
Buffer in the City of Commerce] 22
Map 10. EFFECTIVENESS OF CURRENT CITY ZONING PRACTICE IN SEPARATING AIR
QUALITY HAZARDS FROM SENSITIVE LAND USES [Facilities of Concern - 1000 foot
Buffer in the City of Commerce] 23
Map 11. EJSM Ground-Truthing Locations in the SJV 31
Map 12. Map of San Joaquin Valley Region 32
XIII
-------
Map 13. Map of San Joaquin Valley showing three areas where ground-truth validation was completed. 33
Map 14. Map of San Joaquin Valley Region 35
Map 15. Map of San Joaquin Valley showing three areas where ground-truth validation was completed. 37
Map 16. Arvin study area showing location and types of facilities identified from standard public
databases 38
Map 17. Arvin study area showing location and types of facilities validated using ground-truthing. ... 39
Map 18. Huron study area showing location and types of facilities identified from standard public
databases 40
Map 19. Huron study area showing location and types of facilities validated using ground-truthing.
Locational error shown as black lines connecting original and corrected positions 42
Map 20. Huron study area showing location and types of new facilities (not part of original standard
public databases) discovered during ground-truthing 42
Map 21. Stockton study area showing location and types of facilities identified from standard public
databases. Note locations of duplicate and missing facility records 43
Map 22. Stockton study area showing location and types of facilities validated using ground-truthing. New
facilities are shown in yellow 43
Map 23. EJSM total score for San Joaquin Valley prior to ground-truth correction 45
Map 24. Locational error for California DTSC permitted hazardous waste handling facilities and
generators located in the San Joaquin Valley 46
Map 25. California DTSC permitted hazardous waste handling facilities and generators in San Joaquin
County showing locational error in reported positions. Facility boundaries are shown as violet-
colored polygons 47
Map 26. California Air Resources Board "Facilities of Interest" (FOI) locations in the San Joaquin Valley,
showing location corrections 47
Map 27. EJSM total CI score for San Joaquin Valley prior to ground-truth correction 48
Map 28. EJSM Hazard Proximity CI Scores (Arvin, Huron, and Stockton) before and after ground-truth
correction of facility location and status 48
Map 29. EJSM Total CI Scores (Arvin, Huron, and Stockton) before and after ground-truth
correction of facility location and status 49
XIV
-------
Table 1. Land Use and Hazard Proximity Indicators 2
Table 2. Air Pollution Exposure and Health Risk Indicators (all at Census tract levels) 3
Table 3. Social and Health Vulnerability Indicators (all at Census tract levels) 3
Table 4. Project Schedule 5
Table 5. Project Budget 5
Table 6. Participants in November 1, 2012 Meeting of Project Partners 10
Table 7. Screening Scores by Aggregate Population 12
Table 8. Screening Scores by Aggregate Population 14
Table 9. Facilities in Huron study area with location correction greater than 100 meters 41
Table 10. CARB - FOI facilities in the San Joaquin Valley found by ground-truthing to have been
mislocated by at least 10 kilometers 44
Table 11. CARB - FOI facilities in the San Joaquin Valley found by ground-truthing to have been
mislocated by at least 10 kilometers 49
Table 12. CARB - FOI facilities in the San Joaquin Valley found by ground-truthing to have been
mislocated by at least 10 kilometers (continued) 50
Table 13. PowerPoint Presentation - Webinar Agenda A-2
Table 14. PowerPoint Presentation - Webinar Goals A-2
Table 15. PowerPoint Presentation - Project Goals A-2
Table 16. PowerPoint Presentation - Project Team and Partners A-3
Table 17. PowerPoint Presentation - Project Timeline A-3
Table 18. PowerPoint Presentation - Project Timeline (continued) A-4
Table 19. PowerPoint Presentation - Title A-4
Table 20. PowerPoint Presentation - Summery of Presentation A-4
Table 21. PowerPoint Presentation - Abbreviations A-4
Table 22. PowerPoint Presentation - EJSM A-5
Table 23. PowerPoint Presentation - CEVA A-5
Table 24. PowerPoint Presentation - CES A-6
Table 25. PowerPoint Presentation - Difference Among Methods A-6
XV
-------
Table 26. PowerPoint Presentation - Comparison Table of Indicators A-7
Table 27. PowerPoint Presentation - How comparison was done A-7
Table 28. PowerPoint Presentation - How comparison was done A-8
Table 29. PowerPoint Presentation - How comparison was done A-8
Table 30. PowerPoint Presentation - Q/A Process used A-9
Table 31. PowerPoint Presentation - Preliminary Result A-9
Table 32. PowerPoint Presentation - EJSM Scores A-9
Table 33. PowerPoint Presentation - Distribution of scores A-9
Table 34. PowerPoint Presentation - CES SJV Distribution of Group by area and population A-10
Table 35. PowerPoint Presentation - CEVA Scores - San Joaquin Valley A-10
Table 36. PowerPoint Presentation - Summary - Distribution of scores A-ll
Table 37. PowerPoint Presentation - Location of "hot spots" A-ll
Table 38. PowerPoint Presentation - Screening Scores by Aggregate Population " Top Quantile" . . . . A-12
Table 39. PowerPoint Presentation - Map 6a "High Decile" - Kern Co A-12
Table 40. PowerPoint Presentation - Map 6b "High Decile" - Central SJV A-13
Table 41. PowerPoint Presentation - Map 6c A-13
Table 42. PowerPoint Presentation Summary - High Quantile A-14
Table 43. PowerPoint Presentation - Areas Identifies by all Three Methods A-14
Table 44. PowerPoint Presentation - Areas Identifies by all Three Methods A-15
Table 45. PowerPoint Presentation Areas - Identifies by all Three Methods A-15
Table 46. PowerPoint Presentation - Areas Identifies by all Three Methods A-16
Table 47. PowerPoint Presentation - Areas Identifies by all Three Methods A-16
Table 48. PowerPoint Presentation - Areas Another way to examine areas identified by
all three methods A-17
Table 49. PowerPoint Presentation - Areas Northern SJV, Stockton Area A-17
Table 50. PowerPoint Presentation - Northern SJV, Modesto Area A-18
Table 51. PowerPoint Presentation - Central SJV, Fresno A-18
Table 52. PowerPoint Presentation - Central SJV, Tulare Area A-19
Table 53. PowerPoint Presentation - Southern SJV, Bakersfield A-19
XVI
-------
In this final report, we discuss, describe, present and/or
explain the following key elements of this research project:
1. Pilot Application Areas (California)
a. City of Commerce (and Green Zones Working
Group)
b. San Joaquin Valley (SJV)
i. Arvin
ii. Huron
iii. Stockton
2. Cumulative Index (CI) Screening Methods
a. Environmental Justice Screening Method (EJSM)
b. California Communities Environmental Health
Screening Tool (CES)
c. Cumulative Environmental Vulnerability Assessment
(CEVA)
3. Comparison of the three CI Screening Methods
a. Distribution of screening CI scores (focusing on
population and area)
b. Determining how to define 'extreme' CI scores
c. Assess statistical regions (decile, quintile) where
'extreme' CI scores are comparable
d. Determining the geographic areas where 'extreme'
CI scores are comparable
4. Impact of ZCTA/Census Tracts/Census Block Groups
on CI Scoring
5. Community Interactions and Organizational
Partnerships
6. Webinar to Explain the three CI Screening Methods to
the SJV Community
This research project represents first major step forward in
the development of an approach towards the assessment
and characterization of Environmental Justice (EJ)
communities known as Environmental Justice Analytics or
Environmental Justice Analysis (EJA). EJA is defined as
the application of models, tools, databases, and information
resources in a coordinated, organized, well-defined process
to analyze the multiple risks from air pollutants and hazards
(including those caused by land use/land use policies)
and the resultant cumulative impacts (CI) that influence
the environmental exposures experienced by vulnerable
individuals and communities. The EJA approach provides
a procedure and guidance on how to apply a single CI
screening methodology/tool/approach to assess the impact
of environmental (air pollution) hazards on what is termed
'sensitive receptors' (CARB, 2005). There are a number of
data and information resources from local, regional, state and
national sources that can be used in an EJA assessment, but
the most useful (and common) ones include as a minimum:
a) land use and land use databases (e.g., SCAG [derived
from Anderson], USGS Land Cover Database, California
Department of Education, etc.); b) facility databases (e.g.,
CASIL); c) commercial {business} databases (e.g., Dunn and
Bradstreet [NAICS Codes], ESRI Business Analyst, etc.);
d) toxicity databases/models; e) hazard proximity/exposure
information; f) traffic proximity/exposure information; g) air
pollution exposure and health risk information (e.g., from
models and air pollution monitors [air pollution concentration
data]); h) social and health vulnerability information (e.g.
from US Census data); i) birth outcome data (e.g., California
Department of Public Health AVSS); j) aerial imagery; k)
street maps; 1) GIS spatial/geospatial analysis (e.g., ESRI
ArcGIS); n) Google Earth resources (e.g., Google Earth
Pro, *.kml files, etc.); m) GPS (i.e., GPS receivers used
to validate location [GPS coordinates] of facilities during
"ground truthing" [verification of facility location] activity;
n) webpage (i.e., to view/compare areas of low, moderate,
and high cumulative impacts). The use of mathematical and
statistical analysis tools and techniques to quantify CI on
various communities, along with the resultant analysis and
assessment, is the final and most important component of
EJA (Hall, E.S., Morello-Frosch, R. A., et al., 2011).
Communities are exposed simultaneously to physical
environmental hazards or risks (e.g., air pollution),
socioeconomic influences (e.g., education, family income),
and psychosocial risks (e.g., linguistic isolation), resulting in
cumulative impacts which can exacerbate health outcomes
for vulnerable individuals and subpopulations (CARB, 2005).
Application of an EJA approach is designed to provide an
objective way to quantify and characterize the CI effects
and potential vulnerabilities experienced by communities
and to provide information, which can be used to inform
local, regional, and state-level policymakers on the potential
impacts of land use changes and decisions on affected
communities. This is accomplished through the calculation
of the CI score for a specified geographic region(s) where
a vulnerable community is located. The size/extent of
geographic regions that can be assessed through EJA
includes but is not limited to: a) ZCTA (Zip Code Tabulation
Areas); b) US Census Blocks; c) US Census Block Groups;
d) US Census Tracts; e) real estate tax parcel, etc. The
TIGER system is used to define the boundary of US Census
Tract and US Census Block files. ESJM was developed to
-------
Table 1. Land Use and Hazard Proximity Indicators
Indicator Indicator Subtopic
Sensitive Land Uses
Land Use Air Quality Hazards
and Hazard
Proximity
Hazardous Land
Uses
Indicator Subtopic
Childcare Facilities
Heath Care Facilities
Schools
Urban Playgrounds
CHAPIS Facilities
Chrome-Plating Facilities
Hazardous Waste
Railroad Facilities
Ports
Airports
(Petrochemical) Refineries
Intermodal Distribution
GIS Spatial Unit
Land Use Polygons
Buffered Points
Land Use Polygons
Land Use Polygons
Buffered Points
Land Use Polygons
Point Locations
Point Locations
Point Locations
Land Use Polygons
Line Features
Land Use Polygons
Land Use Polygons
Line Features
Land Use Polygons
Land Use Polygons
Line Features
Information Source
SCAG 2008
Dunn and Bradstreet
(by NAICS code)
SCAG 2008; CASIL
SCAG 2008
CA Dept of Education
SCAG 2008
GARB
GARB
DISC
SCAG 2008
NTAD
SCAG 2008
SCAG 2008
NTAD
SCAG 2008
SCAG 2008
NTAD
implement the EJA approach and accomplishes this through
the procedure used to apply ESJM in specific CI community
assessment scenarios.
EJSM is applied using the following general steps when
conducting a CI community assessment: i) data resources
a) through m) listed above [excluding 1)] are used as inputs
to the GIS spatial assessment; ii) ESRI Arc GIS is used to
conduct the GIS spatial assessment; iii) the outputs of i)
and ii) are incorporated into a GIS map (process: land use
is provided [in m2] and the land use polygons [e.g., facility,
school, etc., location and extent] are 'intersected' with the
Census polygons [blocks, block groups, tracts], and the
Census polygons are used to obtain the CI score, which
determines the degree of adverse exposure [s] experienced
by a community). The CI score is obtained through the CI
scoring algorithm, which is implemented as follows: 1) [Land
Use and Hazard Proximity Indicators] points are allocated
to Census polygons and sensitive land use areas (1 point
for each); 2) a proximity analysis (distance-weighted) is
performed by counting the number of hazards within each
distance band/buffer (determined by 2005 CARD Handbook),
where the hazards are weighted, based on which buffer
they reside in, as follows: 0-1000 feet (100% - [multiple:
1.0]); 1000 - 2000 feet (50% - [multiple: 0.5]); 2000 - 3000
feet (10% - [multiple: 0.1]); 3) hazard proximity indicators
are calculated by inserting the distance-weighted hazards
into the appropriate Census polygon; 4) an estimate of the
population in each CI polygon (by Census block: for eventual
CI score) is aggregated to the Census tract level (Note:
there are approximately 3500 Census tracts in California,
[Hall, E.S., Morello-Frosch, R. A., et al., 2011]); 5) each CI
polygon receives a score of 1 [low] to 5 [high], which is area-
weighted to the population and then population-weighted to
the appropriate Census tract; 6) the same algorithm described
in 1) through 5) above is applied to: a) [Air Pollution
Exposure and Health Risk Indicators] and, b) [Social and
Health Vulnerability Indicators]; 7) the Land Use and Hazard
Proximity Indicator Scores, Air Pollution Exposure and
Health Indicator Scores and Social and Health Vulnerability
Indicator Scores are added together (each CI polygon
receives a score of 1 [low] to 5 [high]), meaning that a CI
polygon can obtain a minimum CI score of 3 (1 for each of
the three indicators) to a maximum CI score of 15 (5 for each
of the three indicators). The Land Use and Hazard Proximity
Indicators are shown in Table 1. The Air Pollution Exposure
and Health Risk Indicators are shown in Table 2. The Social
and Health Vulnerability Indicators are shown in Table 3.
An EJ cumulative risk framework was developed during
the implementation phase of the Pilot Application 2 (SJV
Community) to facilitate the comparison of two or more
different CI screening methods being applied to the same
geographic area(s). The Environmental Justice Analysis
Framework (EJAF) is used in a scenario where different
CI screening methods are normalized (harmonized) in
their implementation details to: a) determine geographic
areas containing locations where there is a high level of
mutual agreement or correlation of CI scores (especially
where 'high-end' cumulative impacts [e.g., "hot-spots"]
are indicated by multiple CI screening methodologies);
b) indicate the specific geographic regions where each CI
screening method is optimal or provides the information best
suited to the community and/or policy issue; c) apply the
-------
Table 2. Air Pollution Exposure and Health Risk Indicators (all at Census tract levels)
Indicator
Air Pollution
Exposure and Health
Risk
Indicator Subtopic
RSEI toxic concentration hazard source (from TRI
information)
NATA respiratory hazard, air toxics, mobile/stationary
GARB estimated cancer risk, air toxics, mobile/stationary
(from CEIDARS information)
Information Source
US EPA 2007
US EPA 2002
GARB 2001
PM25 estimated concentration from monitoring
Ozone (O ) estimated concentration from monitoring
GARB 2009
GARB 2009
Table 3. Social and Health Vulnerability Indicators (all at Census tract levels)
Indicator Indicator Subtopic
Information Source
% minorities (in the total population of non-Hispanic whites)
% below 2X national poverty level
% living in rented households (Home Ownership)
US Census 2010
US Census 2010
US Census 2010
Social and % > age 24 with < high school diploma (Educational Attainment) US Census 2010
Health
Vulnerability % < age 5 (Age of Residents) US Census 2010
% > age 60 (Age of Residents)
% > age 4 years when no one in the household > age 15 speaks
English well (Linguistic Isolation)
% votes cast in 2008 general election (Voter Turnout)
US Census 2010
US Census 2010
UC Berkeley Statewide
Database
-------
output of multiple CI screening methods in a coordinated
fashion to inform and guide specific policy decision scenarios
(e.g., land use planning/zoning, new residential, commercial,
and/or transportation planning, environmental remediation,
etc.); d) develop a standard process for implementing an
EJ analysis using multiple CI screening methodologies to
ensure that the CI scores (for the same geographic region)
obtained from each method can be compared and provide
the same general indications (i.e., yield an 'apples-to-apples'
comparison). The complete details of the implementation of
the EJAF in this research project are provided in the section
describing Pilot Application 2 (SJV Community).
-------
2.0
Project Details
The programmatic details of this project are presented in
this section. The project was funded and managed by EPA's
Office of Research and Development (ORD), through its
National Exposure Research Laboratory (NERL) located
in Research Triangle Park North Carolina. This research
is implemented under EPA's RARE (Regionally-Applied
Research Effort) Project initiative, where ORD works in
collaboration with technical experts in the EPA Regions
(1 through 10) to develop and implement research projects
focused on providing scientific assistance and solutions for
high-priority regional issues. This project was designed as a
cooperative agreement (awarded to Occidental College with
UC Berkeley and USC as additional university collaborators)
where EPA personnel contribute substantive effort towards
the implementation of the research in conjunction with the
university collaborators, unlike a research grant where the
grantees do not work jointly on research with government
personnel. The research tasks as outlined in the project
workplan (research plan) are provided below.
Task 1. Develop work plan
Task 2. Scoping Meeting; Select pilot application sites
Task 3. Pilot Application Planning and Design meetings and
training workshops
Task 4. Ground truth field work and data collection, location,
verification
Task 5. Analyze Data
Task 6. Hold Community meetings to review results and
policy implications
Task 7 Prepare revisions based upon feedback and review;
draft journal article preparation; draft report
preparation
Task 8. Submit journal article; Submit final report
Project Workplan and Schedule
Development of the project (research) work plan was
completed, and the work plan was approved in its final
form on 23 October 2012, with a signed EPA Cooperative
Agreement issued on that date. The schedule for this project
is provided in Table 4 below.
Summary of Project Budget
The approved budget for this research project, as listed on the
SF424 A budget summary form, is shown in Table 5 below.
The process of determining pilot applications and
environmental questions was initiated with a project research
team "kickoff" conference call in mid-February (2013).
During this call, the community attributes and logistics for a
variety of possible pilot communities were discussed. This
Table 5. Project Budget
Budget Categories
Personnel
Fringe Benefits
Travel
Equipment
Supplies
Contractual
Total Direct Charges
Indirect Charges
Totals
Approved Budget
$25,000.00
$6,930.00
$5,750.00
$0.00
$702.00
$29,800.00
$68,182.00
$6,818.00
$75,000.00
Table 4. Project Schedule - Note: Month 1 = Oct 2012, and subsequent months follow to Month 24 = Sept 2014
Month
Taskl
2
3
4
5
6
7
9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
-------
project was carried out in two different "focus areas" within
Region 9, specified in the research work plan as: the San
Joaquin Valley region, and; the Los Angeles area along the
Interstate 710 corridor linking the ports of Los Angeles/Long
Beach with inland markets. The process of determining
the environmental question for each pilot application was
handled differently for the two focus areas. Each focus area
had its own pilot application and associated activities. The
project team communicated during the project to review
progress, troubleshoot analytical and logistical questions,
discuss a variety of analytical and data issues, and to
solicit feedback.
-------
3.0
Pilot Application 1: City of Commerce
The pilot location for Los Angeles is the City of Commerce,
with a pilot application to apply EJSM tools, data and
techniques to an ongoing process of creating policy actions
that reduce cumulative air pollution exposure among the
highly impacted residential communities in the City. In
2005, responding to the leadership of East Yard Communities
for Environmental Justice (EYCEJ), the City of Commerce
formed an environmental justice task force to examine
and improve non-occupational environmental hazards and
exposure citywide. The most prominently identified issues
by the task force were the connection between cumulative
impacts from air pollution the juxtaposition of incompatible
land uses, and also declines in business and job opportunities
for local residents.
Following a task force report, the City Council called for
a Working Group to be established that would draft policy
recommendations to (a) create a buffer zones to provide a
safe distance between hazards and sensitive receptors, using
buffer and land use recommendations in the California Air
Resources Board's 2005 Air Quality and Land Use Handbook
for guidance, and (b) to develop economic development zone
strategies as an overlay to attract new developments, using
the lever of the green economy.
This Working Group was tasked to complete the policy
recommendations by summer, 2013 for consideration by
the City Council. Subsequent decisions of the City Council
may lead to changes in City policy documents such as a
General Plan Amendment, Overlay Zone, or other policy
and/or planning tools. This pilot project seemed particularly
well suited to this project because of its clear and focused
environmental question and anticipated use of the EJSM,
as well as the fact that Co-Pi James Sadd served as one of
the original 2005 task force members and introduced the
EJSM to the task force as a tool for both data exploration and
strategy design.
The Working Group, labeled the City of Commerce Green
Zone Working Group, was formally constituted in September
2012 to include a variety of local stakeholders representing
the local residential, organized labor and business
Land Use (SCAG)
City of Commerce
• Municipal Facilities
• High Density Residential
D Schools
• Hospital
• Heavy Industrial
• Parks Open Space
O Transportation
O Light Industrial
• Utilities
• Utilities
K3 Vacant
• Warehousing
• Commercial
• Mixed Residential
O Low Density Residential
Map 1. Land Use in the City of Commerce classified by aggregating standard Anderson Land Use urban classes into
groups that reflect the pilot application.
-------
communities. In addition, non-voting advisors to the
Working Group include Carlin Hafiz and Deldi Reyes of EPA
R9, and Alex Hamilton and Mathew Martinez of the City of
Commerce planning staff. The Working Group met six times
at roughly monthly intervals beginning 3 October 2012 at the
City of Commerce City Hall Emergency Operations Center,
with meetings facilitated by a third party service contracted
by EPAR9. The Working Group made significant progress
and completed its work as described below.
Phase 1: Identify Issues, Opportunities, and
Existing Conditions
During this phase, a process for how the Group operates
and makes decisions, as well as timeline and framework for
developing a set of policy recommendations were established.
In addition, the key issues to be addressed through the
recommendations were identified during this phase.
Phase 2: Strategy and Policy Framework Development
The second phase of the project included presentations and
review of the results provided by outside experts, in order
to better understand the entire landscape of problems and
wopportunities that bear on the two goals of the Working
Group policy recommendations, and to identify strategies and
opportunities based upon this understanding.
During this phase, the Working Group engaged in a process
to identify the specific hazards and pollution sources that
would be the subject of the eventual policy recommendations
and its associated products and tools. After studying different
ways in which this question has been addressed by other
groups with a similar purpose, the Working Group agreed
to specific types of facilities and land uses that, for the
purposes of the policy, represented both hazards and sensitive
receptors. The EJSM methodology was largely used as the
blueprint for this final agreement.
The EJSM methodology was also used to explore the spatial
distribution of land uses as pollution sources, and the location
and adjacency of residential and sensitive (schools, parks
and playgrounds, daycare and childcare centers, healthcare
facilities, senior housing facilities) to these sources, as well as
to test the impact of various buffer options as suggested in the
CARB Handbook (CARB, 2005). The EJSM was also used
in conjunction with business information service databases
to examine these patterns with respect to the individual
businesses present on a real estate tax parcel level.
Examples of the cartographic output of this process
(Land Use: Southern California Association of Governments
[SCAG] and City of Commerce) are shown, below:
The geography of the EJSM and other mapping results were
exported to .kml format (keyhole markup language files)
readable by the Google Earth web application, to better allow
Working Group members and other constituencies to examine
these patterns and datasets in the context of Google Earth's
high-resolution aerial imagery at their leisure and outside of
Working Group meetings. As part of this part of the project,
the research team made an offer to provide the City of
Commerce with a library of GIS spatial data that they can use
for any initiatives that arise from this process, and for other
City erf Commerce - proposed Buffers
Overview
Map 2. Proposed buffers surrounding sensitive land uses
as a land use planning tool to ensure separation from
air pollution point and area sources. Area inside inset
rectangle shown in Map 3, below.
Cily oi Commerce - Proposed Suiters
Wasti.nglorwAtiomic Area
I "';'•:!..r i--^i '
xo tact e-jfcH
SOGtoctkufcr
Map 3. Proposed buffers surrounding sensitive land uses
in the Washington/Atlantic Blvd. corridors, a special
focus area recognized by the Working Group both for
protection/separation of residential and sensitive land
uses from air pollution sources, but also for business
development and "green design" amenities.
work that the City Planning Department needs. The Planning
Department does not currently use GIS, but is interested in
leveraging this opportunity to do so.
Phase 3: Policy Recommendations Development
This phase was dedicated to creation of detailed and specific
(a) Draft Proposed Amendment to City of Commerce
Zoning Ordinance, which included specific land use
recommendations and buffers derived from the EJSM
methodology, and (b) Draft Proposals for a Voluntary
Business Retrofit Program. These drafts were rewritten
and reviewed during the next four scheduled meetings. Key
stakeholders used them to further, support, and/or validate
the work of this pilot application. This phase began in March
2013, and was completed in July 2013.
-------
In contrast to pilot application 1, which was well denned
and progressed to an advanced stage fairly quickly, Pilot
Application 2 used a different process, and proceeded more
slowly. This pilot application was more difficult to define,
both in geographic and policy terms, because of its greater
complexity, wider variety of stakeholders, wider array
of policy options, and the challenge of communications
among stakeholders who are located from Los Angeles
to San Francisco and Sacramento. During the February
2013 project kick off conference call, a March 2013 SJV
(Pilot Application 2) Webinar was planned to introduce
the project to various EPA R9 and State agency partners,
including the EPA R9 Air and Waste Programs, and the
Office of Environmental Review. Also present for the SJV
(Pilot Application 2) Webinar were representatives from the
California Environmental Protection Agency (CalEPA) Office
of Environmental Health and Hazard Assessment (OEHHA)
and California Air Resources Board (CARD), the San
Joaquin Valley Air Pollution Control District (SJV-APCD)
as well as San Joaquin Valley Cumulative Health Impacts
Project (SJV-CHIP), a community organization collaborative,
and the Center for Regional Change at the University of
California Davis.
The San Joaquin Valley has become a key focus region for
environmental justice screening approaches in California.
The state of California, through CARB and CalEPA, has
funded the development of three different environmental
justice-screening tools to assess the cumulative impacts of
multiple air pollutants on vulnerable communities at the
neighborhood, community, and regional scale. This provides
California with a range of environmental justice screening
tools that can be applied in situations where decisions must
be made to ensure that the (negative) cumulative impacts
of local planning and land use changes on vulnerable
communities is minimized. In addition to EJSM, which
was jointly funded by CARB and the US EPA (through
a previous EPA R9 RARE Research Project), two other
screening tools were used to produce a detailed study and
screening products in 2011 with funding from CalEPA. The
two additional screening methods were developed by the
Center for Regional Change at the University of California
Davis (Cumulative Environmental Vulnerability Assessment
or "CEVA"), and OEHHA (California Communities
Environmental Health Screening Tool (CalEnviro Screen
or "CES"). The research team is very familiar with both
of these methods, and has served as technical advisors and
reviewers for both screening groups as they developed their
methods using the EJSM as a template.
Among SJV stakeholders, there has been a particular interest
in comparing the three methods to better understand their
similarities and differences, the strengths and weaknesses
and various trade-offs inherent to each method, requirements
for updating each as new data becomes available and new
methods are developed.
The initial pilot application consisted of several steps. First,
a detailed comparison of the three CI screening methods
became an integral and defining part of this project, and
had the following goals: (a) explore the regional pattern of
CI score by the three methods; (b) compare results of the
three CI screening methods for the eight county SJV region
to identify areas of agreement/disagreement in relative
CI scores as a means to select areas for ground truthing
under this cooperative agreement; (c) define and identify
impact/vulnerability CI "hot spots" for each method and
determine the degree to which the methods agree/disagree;
(d) summarize differences in data and metrics used, and CI
scoring methods for each of the three methods.
In order to use the entire San Joaquin Valley (SJV) for a
study region for this comparison, the research team expanded
EJSM mapping and CI scoring to include an additional three
SJV counties (Merced, Stanislaus, San Joaquin) to add to the
five southern SJV counties (Madera south to Kern County).
The resulting region is shown in Maps 4a through 4c, below.
This approach allowed us to prepare a series of comparison
metrics among the three methods, and a series of comparison
scoring and mapping procedures that allowed us to address
the concerns stated above. This comparison effort was
slightly delayed due to the delayed release of the CES data,
which underwent public comment and OEHHA refinement.
The next step in the pilot application called for agency
partners to develop a series of policy relevant questions,
which the screening methods would address along with
suggestions on how each method and its maps could be
made more useful in informing the policy questions. These
elements were discussed at a Nov 1, 2012 meeting of project
partners. Participants at that meeting, whether in-person or
by teleconference or videoconference are shown in Table 3
below.
A variety of policy relevant questions were presented and
discussed at this meeting. It was decided that once the
operative policy relevant questions that could be addressed
by EJSM were finalized, we would develop a case study
report done in collaboration with the agency, which had
proposed that policy relevant question.
At the November 1, 2012 meeting of project partners, the
research team presented a first draft comparison of screening
methods, focusing only on EJSM and CEVA owing to the
-------
Table 6. Participants in November 1,2012 Meeting of
Project Partners
EJSM
Organization
EPAR9
EPA ORD
SJV-APCD
SJV-CHIP
CCA
CalEPA/OEHHA
CARS
UC Davis
DISC
OPR
Occidental College
use
Participant
Debbie Lowe Liang
Mike Bandrowski
Charles Swanson
Jim Polek
EricS. Hall
David Lighthall
Sarah Sharpe
Shankar Prasad
Arsenic Mataka
Malinda Dumisani
John Faust
Laura August
Alvaro Alvarado
Johnathan London
Tara Zagofsky
Ignacio Dominguez
Brian Johnson
Debbie Davis
James Sadd
Manuel Pastor
fact that CES was still under development at the time. It was
agreed that the comparison would be revised, with the CES
data incorporated into the next comparison, and the final
comparison of all three methods that would take place when
the complete version of the CES data was available.
In collaboration with EPA R9 and ORD and using the
meeting discussion along with input from project partners, we
developed a flexible structure for this pilot project. An initial
draft summary of efforts to complete data was accomplished
using a draft release of the CES data, and it is summarized
briefly below.
Comparison of Screening Methods
in the SJV Region
The three screening methods (EJSM, CEVA, and CES) have
similar goals, anticipated uses, and use many of the same data
sets. However, they also differ, in some cases substantially,
in many ways, including:
1. Spatial unit of analysis (or spatial resolution) of
results: Each Method uses a different spatial unit for
both analysis and mapping. EJSM uses real estate tax
parcels that are edited to reflect that land use (termed
"CI polygons") and CI score are done at the census tract
level. CEVA uses census block groups for both analysis
and scoring, and CES uses zip code tabulation areas
(ZCTA) from the US Census.
2. Different "base maps": Differences in spatial units used,
as described above, results in a different level of spatial
resolution on maps of results that, in turn, reflects the
Regionally scored at
tract level; tax parcels
used to calculate
hazard proximity and
land use
Tracts do not receive a
score if population is
very small or too few
variables are available
(show in white)
Final maps mask out
land uses that are not
classed as residential
or "sensitive" byCARB
(not shown)
Map 4a. EJSM CI scores for the SJV: Mapped using
census tracts to allow comparison with other screening
methods.
CEVA
Regional scoring of
block groups
All block groups in
the region are given
a score
Map 4b. CEVA CI scores for the SJV: Mapped using
census block groups for comparison with other screening
methods.
CES
10
Scores calculated at
the ZCTA level, for
entire state
Incomplete Coverage - p
some ZCTAs unscored
(shown in white)
Scored areas are very
large compared to
census tracts or blocks
= lack of granularity
Because of statewide
scoring, SJV has a
disproportionate
percentage of high CES
scores compared to
other regions of the
State
Map 4c. CES group CIscores for the SJV: Mapped using
2010 zip code tabulation areas (ZCTAs) for comparison
with other screening methods. ZCTAs that extend outside
the eight SJV counties have been clipped so that colored
polygons do not extend outside the boundaries of the
denned comparison area.
-------
map pattern differently, significantly complicating
comparison of results in an objective "apples to apples"
manner. There are also differences in geographic
extent, making different map patterns that do not
overlay one another well. It also complicates estimating
the distributional impacts of the various CI scores, as
different spatial units represent different populations
These differences are most manageable when
comparing EJSM census tracts with CEVA census
block groups, as there is a logical manner in which
block groups aggregate to tracts. It is more difficult
when comparing to CES scores at the ZCTA level.
ZCTAs are much larger spatial units than either blocks
or tracts, and do not share boundaries with either; they
also cross county boundaries. As for distributional
impacts, tract or block group level metrics and
demographics must be imputed and aggregated to the
ZCTA level. The larger ZCTAs can mask small areas
of concentrated impact and vulnerability, as they are
"averaged in" to the larger ZCTA area
3. Different metrics: The three methods share many
metrics in common, but also use different metrics to
represent a specific indicator. Some common metrics
include: estimated cancer risk from air pollution vs.
reported cancer mortality; RSEI metrics vs. TRI site
location, use of % poverty vs. 200 % poverty as a
threshold definition.
4. How CI scores are calculated: The three screening
methods use different ways of grouping indicators
together for scoring, resulting in different implicit
weighting of certain metrics. Also, the methods each
use a different numerical range of CI scores; however,
for each method, a higher score indicates a greater
cumulative impact of air pollutants on a vulnerable
community. The EJSM CI scores range from 3-15,
where 3 represents the minimum cumulative impact
score and 15 indicates the maximum cumulative
impact score for a given community, but the EJSM
scoring algorithm is open-ended and does not limit
the value of the high end in order to accommodate
additional indicators as future data or tools allow
for improvements. CES CI score values are
continuous, ranging from 0.86 to 72.80 statewide.
These continuous CI scores are aggregated into nine
classes (1-9) called "CES group CI scores". CEVA
CI scores also range from 1-9, but this method uses a
two-dimensional scoring matrix resulting in some low
CI score numbers indicating higher impact/vulnerability
and eliminating a one-to-one comparison of CI scores
among these two methods (CES and CEVA).
5. Region over which CI scores are calculated: Both
EJSM and CEVA score regionally, so that the CI scores
compared, below reflect only the 8-county SJV region
as defined. However, CES scoring is currently done
statewide. Maps 4a-c, abovet show the CI scores for
the three screening methods using their respective
mapping units.
EJSM Scores - San Joaquin Valley
Figure 1. Distribution of CI screening scores for the SJV
region by population and area: EJSM.
CES Scores - San Joaquin Valley
5111S 1270*5 41J1N) 5JTJ1T IW5* MSJ» WSJ60 TMUN
Figure 2. Distribution of CI screening scores for the SJV
region by population and area: CES.
CEVA Scores - San Joaquin Valley
Figure 3. Distribution of CI screening scores for the SJV
region by population and area: CEVA.
11
-------
Table 7. Screening Scores by Aggregate Population
Table 1. Screening
Scores by Aggregate
Population
• The highest scores for the
three methods represent
census polygons hosting
8.7% to 9.7% of the total
SJV residential population
• EJSM 14 and 15 = 9%
• CEVA 9 = 9.7%
• CES 10 = 8.7%
EJSM Score Population Cumulative Population Cumulative % Population
0
3
4
5
6
7
3
9
10
11
12
13
14
15
CEVA Score
1
2
3
4
5
6
7
8
9
CES Score
1
J
3
4
S
B
7
a
9
10
39,547
126,262
230,628
290,873
359,131
298,139
406,567
378,511
470,204
400,750
255,481
260,607
205,052
69.781
291,017
1,162,596
325,386
84,579
746,720
232,036
17,945
563,780
369,338
28605
53115
127045
413170
527337
394756
805390
&J.n3AA
o**OJQU
769109
379304
3,791,533
3,751,986
3,625,724
3,395,096
3,104,223
2,745,092
2,446,953
2,040,386
1,661,875
1,191,671
790,921
535,440
274.833
69,781
3,793397
3,502,380
2,339,784
2,014,398
1,929,819
1,183,099
951.063
933.118
369,338
4,344,191
4,315,586
4,262,471
4,135,426
3,722,256
3,194,919
2,800,163
1,994,773
1,148,413
379,304
100.0%
99.0%
95.6%
89.596
81.9%
72.4%
64.5%
53.8%
43.8%
31.4%
20.9%
14.1%
7.2%
1.8%
100.0%
92.3%
61.7%
53.1%
50.9%
31.2%
25.1%
24.6%
9.7%
100.0%
99.3%
98.1%
95.2%
85.7%
73.5%
64.5%
45.9%
26.4%
8.7%
Map 5a. Top Population Quantile - Kern Co.
• All three methods identify Bakersfield
- CES only identifies greater Bakersfield (one ZCTA), and is not as area-specific as
CEVA and EJSM
- EJSM and CEVA agree on North Bakersfield, Wasco (see overlap color)
• EJSM solely identifies Delano, Oildale, Lamont, Arvin
• CEVA solely identifies Taft
CES store 9
CEVAscoreS
EJSM stores l« ana 15
Ovettac cev* ana EJSM
Map 5a. Top Population Quantile - Kern Co.
12
-------
Map 5b. Top Population Quantile
Central SJV
All three methods identify
Fresno
- CES identifies greater Fresno
(three ZCTAs)
— CEVAand EJSM have more
specificity and show significant
overlap
CEVAand EJSM also agree on
Selma, Tulare, Lindsay and
Porterville
EJSM solely identifies
Hanford, Madera, Visalia
CEVA solely identifies
Woodlake, Corcoran, Sanger,
Parlier Reedley, Dinuba,
Coalinga
CES store 9
CEVA score 9
EJSM scores UandlS
Overlie CEVA «i<) EJSM
~
Map 5b. Top Population Quantile Central SJV
Map 5c. Top Population Quantile
Northern SJV
CES identifies three cities -
greater Stockton (three
ZCTAs), Modesto, and
Merced; CEVA agrees but
with much more specificity
No EJSM high scores in
these three counties
CEVA solely identifies
several other small towns -
Lodi, Turlock, Tracy,
Livingston, Atwater, Gustine,
Patterson.
Map 5c. Top Pupulation Quantile Northern SJV
13
-------
We next compared the three methods in terms of the
distribution of cumulative impact scores region wide. The
goal was to determine how common and where "high"
cumulative impact scores are located for each method, and
to examine the distributional impact, by population and by
area, of each cumulative impact score for each method. This
overall pattern of scores is summarized in Figures 1-3 on
Page 11. Note that the pattern of EJSM and CES scores are
reasonably regular and near Gaussian in their distribution,
with EJSM showing a symmetrical pattern. In contrast,
CES shows a pattern clearly skewed toward high CI score,
probably due to the fact that CES scoring is done statewide
and much of the impact and vulnerability in California is
concentrated in the SJV. CEVA cumulative impact scores
display a non-standard (multi-modal) statistical distribution.
When the distribution of cumulative impact scores are
evaluated by area and by population, other significant
patterns emerge. Most of SJV area has low CI scores in all
three methods, and this pattern is clearest for CEVA and
EJSM. The greatest variation of CI scores amongst the three
methods can be observed when each CI score is mapped to
the total number of people in each CI 'score/scoring range'
Table 8. Screening Scores by Aggregate Population
Table 2. Screening Scores
by Aggregate Population
• A larger sample of high scores
for the three methods
represent census polygons
hosting 26.4% to 31.4% of the
total SJV residential
population
* EJSM 11-15 =31.4%
* CEVA 6*, 8, 9 = 30.7%
* CES 9. 10 = 26.4%
* *Note: because of the way
CEVA scores are calculated,
census blocks with a score of
6 have a higher impact than
score of 7.
(i.e., experiencing each level of cumulative impact) for each
method. Most of the SJV area population is subject to middle
value EJSM scores. This pattern is similar but weaker for
CES CI scores. CEVA CI scores have no regular pattern.
Overall, in spite of their differences, EJSM and CES scores
appear more similar to one another, and less similar to CEVA
scores, in the SJV region.
In order to use these distributional patterns as a means of
comparing of the three screening methods, each method
was evaluated to identify CI "hot spots" based on the
aggregate population represented by each scoring class.
This allowed us to explore the questions "What defines a
CI "hot spot" in each method?" and, "What is a "high" CI
score for each method?" The first attempt used the highest
CI scores: EJSM CI scores 14 and 15 represent an aggregate
population of 9% of SJV, and for a CEVA score of 9 this
value is 8.7%. Because the 1-9 range of CES group CI
scores were calculated statewide, we instead used the CES
continuous CI score for this part of the analysis; all ZCTAs
with CES CI scores >55.75 represent an aggregate 9.8% of
the SJV population. Distribution of CI scores by aggregate
population and these results are shown in Table 7 on Page
12. The yellow bands in Table 7 are used to highlight the
EJSM_Scora Population Cumulative Population Cumulative % Population
0
3
4
5
6
7
8
9
10
11
12
13
14
15
CEVA_Scor«
1
2
3
A
S
7
6
8
9
CES Score
1
2
3
4
S
6
7
S
9
10
39,547
126,262
230.628
290,873
359,131
298,139
406,567
378,511
470,204
400,750
255.481
260,607
205,052
69.781
291,017
1,162,596
325,386
84.579
746,720
17,945
232,036
563.780
369,338
28605
53115
127045
413170
527337
394756
805390
846360
769109
379304
3,791,533
3,751,986
3,625,724
3,395,096
3,104,223
2,745,092
2,446,953
2,040,386
1,661.875
1,191,671
790,921
535,440
274,833
69.781
3,793,397
3,502,380
2,339,784
2,014,398
1,929,819
1,183,099
1,165.154
933.118
369,338
4,344,191
4,315,586
4,262,471
4,135,426
3,722,256
3,194,919
2,800,163
1,994.773
1,148,413
379,304
100.054
99.0%
95.6%
89.5%
81.9%
72.4%
64.5%
53.8%
43.8%
31.4%
20.9%
14.1%
7.2%
1.8%
100.0%
92.3%
61.7%
53.1%
50.9%
31.2%
30.7%
24.6%
9.7%
100.0%
99.3%
98.1%
95.2%
85.7%
73.5%
64.5%
45.9%
26.4%
8.7%
14
-------
percentages of the total SJV population with the highest CI
scores as calculated by each of the three methods. Using
ESJM, approximately 9% of the total SJV population has the
highest CI scores. With CEVA and CES, approximately 9.7%
and 8.7% of the total SJV population respectively has the
highest CI scores. Despite the differences between the three
EJ screening methods, they each reliably and consistently
indicate that approximately somewhere between 9% (8.7%)
to 10% (9.7%) of the SJV population experiences the
maximum CI from air pollutants in their communities (only
a 1% difference). This demonstrates that there is a degree
of harmonization and correlation between the three methods
with respect todetermining high-end cumulative impacts for
vulnerable SJV communities.
Using these CI scores as a basis to define screening CI "hot
spots" is a reasonable first approximation; these areas are
mapped for the three methods across the SJV study area, as
shown in Maps 6a-c below.
All three methods agree only on CI "hot spots" in Bakersfield
and Fresno. EJSM and CEVA agree with far greater
geographic specificity, and are in overall agreement. CEVA
and EJSM identify a number of smaller, relatively isolated
towns that are not detected by CES because of the averaging
effects of using the large ZCTAs, described above. However,
although they both identify smaller towns missed by CES,
differences in metrics and scoring between CEVA and EJSM
result in more disagreement than agreement in which towns
are identified.
Map 6a "High Decile" - Kern Co.
All three methods identify greater Bakersfield
- CES char^cterizajon of Bakersfield is unchanged from the Top Decile example
— CEVA and EJSM both extend identified area in north Bakersfield, and highlight Larnont and Arvin. as well;
they are in general agreement
— EJSM and CEVA agree on North Bakersfield, Wasco (see overlap color)
CEVA now agrees with EJSM in identifying Delano, Oildale, Lamont, Arvin
EJSM now agrees with CEVA in indentifyingTaft
CEVA again solely identifies some small towns - Lost Hills, Wasco,
CES score 9
CEVA score 9
EJSM scores 14 and 15
Overlap CEVA and EJSM
Map 6a "High Decile" - Kern Co.
15
-------
Map 6b "High Decile" - Central SJV
CES expands identification
of the entire Hwy 99
corridor, but still does not
pick up north Fresno or any
areas along Interstate 5
The problem of large
census polygon size making
those identified by CEVA
and EJSM appear more
prominent on the map is
even more pronounced.
EJSM and CEVA generally
agree, but a few locations
are only identified by one
method
- CEVA: Cutler, Wood lake, and
Coalinga as before; also
Woodville, Ivanhoe, London
— EJSM: large, sparsely
populated areas along 15,
Parley, Hanford, Kerman
_
I CES score 9
CEVA store 9
EJSM scores 14 and 15
Overlap CEVA and EJSM
Map 6b. "High Decile" - Central SJV
Map 6c "High Decile" - Northern SJV
As before, CES identifies
large population centers
also identified by CEVA and
EJSM, but they do so with
much greater specificity
CEVA again solely identifies
several small towns not
identified by EJSM -Tracy,
greater Manteca, Ripon,
Oakdale, Patterson,
Livingston, Atwater, Gustine,
Patterson.
EJSM identifies fewer areas,
and more geographically
focused areas, than both
other screening methods
CES score 9
CEVA score 9
EJSM scores 14 ana 15
Overlap CEVA ana EJSM
.'
Map 6c. "High Decile" - Northern SJV
16
-------
Map 7a_l. Fresno and Selma, CA
Map 7b_l. Stockton, CA
Map 7b_2. Modesto, CA
17
-------
If CI "hot spots" are defined more broadly, is agreement
among methods better? In examining the distribution of
high CI scores for a broader population sample, the range of
high CI scores for each screening method apply to a similar
proportion of the total SJV population as follows: EJSM CI
scores 14 and 15 (31.4% of SJV population), CES CI scores
>45.56 (26.4% of SJV population), and CEVA CI scores 6,
8, and 9 (30.7%). Distribution of CI scores by aggregate
population and these results are shown in Table 8 on Page
14. The yellow bands in Table 8 are used to highlight the
percentages of the total SJV population with the highest CI
scores as calculated by each of the three methods when the
definition of CI "hot spots" is more broadly defined (i.e.,
lower CI scores are included [widening the CI score range
at the 'high-end'] when determining the percentages of the
population experiencing high-end CI {potential exposures}).
Using ESJM under this scenario, approximately 31.4% of
the total SJV population has the highest CI scores. With
CEVA and CES, approximately 30.7% and 26.4% of the
total SJV population respectively has the highest CI scores.
Even when range of high-end CI scoring criteria is expanded,
the three EJ screening methods, still agree within a narrow
(population percentage) range. The three methods indicate
that approximately 26.4% to 31.4% of the SJV population
experiences the maximum CI from air pollutants in their
communities (only a 5% difference). There is a tremendous
amount of agreement between these methods, and they
could be used either individually or in combination when
overburdened SJV communities need to determine areas that
experience high-end cumulative impacts.
A final way to compare these three methods is to map the
areas where all agree on a CI "hot spot" location. These
areas are shown in Maps 6a-c. In these maps, areas that
score in the "top decile" of population, in Maps 4 a-c, those
areas are shown in red, and areas that overlap in the "High
Decile" region, shown in Maps 7 a-c, are displayed in pink.
Additional areas nearby with high EJSM CI scores are shown
in yellow.
The results of the comparison of CI screening methods
summarized, above, are consistent with the considerable
differences among these three CI screening approaches.
However, there is agreement among the methods that is
somewhat surprising given the degree of difference in
methods and data used. We believe that this suggest two
things. First, in spite of their differences, the process of
CI screening is robust and meaningful as a technique for
identifying areas that represent the "high end" of a spectrum
of metrics. Second, the existence of a pattern of inequitable
exposure to environmental hazards and their attendant risks,
and vulnerability to those risks, are real and quite stark
in SJV
As this comparison process continued, we repeated these
analysis steps and developed additional approaches to the
comparison, as better data was made available for CES, such
as regional scoring at the tract level. We also aggregated
and rescored using the CEVA method by tract (from block
groups). Comparing tract level CI scores for the three
Map 7c. Merced (all 3 methods overlap/agree)
methods significantly reduces the methodological differences
that hampered this effort. This comparison, combined with
filtering the map results using land use information from
EJSM was expected to be far more useful in defining the
specific areas for ground truthing.
The results of the final comparison will be shared with
agency partners, and include .pdf maps and Google Earth
.kml files to allow partners to provide the best review of
the results and input based upon their experience with
various parts of the SJV. A webinar/conference call was
held to answer any questions about the comparison and
revised maps. Partner agencies and SJV CHIP were
invited to share their observations and questions about the
strengths, weaknesses and data gaps associated with each
of the three methods as well as provide suggestions for a
location(s) where ground truthing would be helpful in better
understanding the strengths and weaknesses of each method.
That information is contained in this report.
Based on the input provided, EPA worked with the three
screening groups (EJSM, CEVA, and CES) and SJV CHIP
to select one location for ground truthing. The location was
selected based on the following criteria: (a) community
capacity to engage in ground truthing, (b) will help answer
questions about whether the three screening methods are able
to inform the policy relevant questions, (c) will help inform
future development of the three screening methods, and (d)
includes a range of CI scores, from low to high, to help us
get an idea how ground truth validation relates to highly
impacted areas vs. areas with lower impact and vulnerability.
18
-------
The pilot location for the Los Angeles area was the City of
Commerce, where EJSM tools, data and techniques were
applied to an ongoing process of creating policy actions
that reduce cumulative air pollution exposure among the
highly impacted residential communities in the City. From
the beginning of this project, the project team was engaged
in the process of providing policy advice as part of the City
of Commerce Green Zones Working Group. This group
has been meeting monthly since July 2012 at the City of
Commerce headquarters.
The Working Group began as a partnership between
the community organization East Yards Community for
Environmental Justice (EYCEJ), EPA Region 9, and the City
of Commerce in early 2012. At the first meeting, the project
team joined representatives from the City of Commerce
Environmental Justice Advisory Task Force, City Planning
staff in an organizational meeting facilitated by Esmeralda
Garcia, who was contracted separately by EPA R9 for this
role. As directed by the City Council, the Working Group
identified additional stakeholders to include in the process.
This process continued over several subsequent meetings due
to the difficulty of obtaining commitments from individuals
representing the stakeholder group. The final composition of
the organizations participating in the Working Group is as
follows:
• Business Community - small business representative
• Business Community - large business representative
• City of Commerce Industrial Council - the "chamber of
commerce" for this city
• Environmental Justice Advisory Task Force
• A representative for organized labor
• Member of the City of Commerce Planning Commission
• A community resident
• The USC Program for Environmental and Regional
Equity (PERE)
The representatives (voting members) of the Working Group
participating organizations included:
• Jose Bojorquez, (Officer of 990 Only Stores, Inc. and
representing large business)
• Doug Ashmore (Officer of Amvac, Inc and representing
small business)
• Eddie Tafoya, Executive Director of the City of
Commerce Industrial Council
• Angelo Logan, East Yard Communities for
Environmental Justice representing the City of
Commerce Environmental Justice Advisory Task Force
• Jason Stinnette, (American Federation of State, County
and Municipal Employees (AFSCME)
• Nancy Barragan, (Planning Commission)
• Kristina Santana (resident)
• James Sadd, (representing USC PERE)
The decision to have James Sadd serve as the voting member
from the project research team was made because of his
geographic proximity, and because he was one of the original
2005 task force members and introduced the EJSM to the
task force as a tool for both data exploration and strategy/
policy design.
At various meetings, there were a number of non-voting
attendees that attended either regularly or occasionally. Most
were City of Commerce business owners and community
residents, but others were there to provide technical guidance
and opinion. Regular non-voting attendees included:
• Alex Hamilton and Mathew Martinez of the City of
Commerce planning staff Donald Spivack, a planning
consultant and retired professional planner with the City
of Los Angeles.
• CarlinHafiz and Deldi Reyes of EPA R9
The goals specified for the Working Group by the City
Council were to draft policy recommendations to: (a)
create separation to provide a safe distance between
hazards and sensitive receptors, using buffer and land use
recommendations in the California Air Resources Board's
2005 Air Quality and Land Use Handbook for guidance,
and; (b) to develop economic development zone strategies
as an overlay to attract new developments, using the general
philosophy of the "green economy".
This Working Group completed its work and decisions on
recommendations at the September 23, 2013 meeting. The
Working Group final report and recommendations with
documentation were delivered to the City of Commerce
Planning Commission in October 2013 and to the City
Council in early November 2013.
19
-------
The Working Group continued its work on defining the
framework of the recommendations that were planned for
completion, and refining the details of the policy options. In
terms of "framing", the Working Group agreed on several
guiding principles:
1. Balance - Recognizing the fact that the City of
Commerce has a large industrial base with a relatively
small residential population and area, it was important
to maintain a balance between the needs and quality
of life of the residential community and successful
economic development.
2. Community Health - A commitment to practices that
that result in improved quality of life and health for
residents and local workforce.
3. Improving distance relationships between Industrial
Uses that represent air toxics hazards and Sensitive
Receptors - The goal is to understand and monitor
the present pattern, and make improvements that will
improve this pattern over time. These improvements
will be guided by the recommendations in the 2005
CARB Air Quality and Land Use Handbook, and the
Working Group agreed by vote to adopt the definition
of "sensitive land uses" used in that document.
4. "Green" Practices - Look for ways to bring ideas
associated with what are generally identified as "green"
practices of urban planning and development.
5. Image - Design actions that would maintain, and even
enhance, the City image as a business friendly city, with
strong community amenities.
6. Infrastructure - Design with appropriate and sufficient
infrastructure to meet the current and future needs of
the community and business.
7. Local Workforce - Design with ways in which to
improve workforce opportunities for local residents
in light of anticipated changes in types business and
industry within the City
8. Connectivity/Mobility - Design improved
transportation-related elements that consider all transit
modes, capitalize on existing successful transit issue,
and improve opportunities for the City of Commerce
community and workforce.
The framing piece of this process also included analysis of
challenges^arriers and opportunities that are relevant to
the anticipated policy recommendations. They included the
following considerations:
1. Although the City has a successful tax base generated
by local business, the City of Commerce is relatively
small compared to other cities in the region, resulting in
certain limitations. For example, the City relies heavily
on outside contracting for services.
2. The permitting process is not particularly efficient
or streamlined, complicating implementation of new
practices or requirements needed to address the goals
20
of the Working Group. For example, the City does not
use modern geospatial data and analytical practices,
and although the City is a member of the Southern
California Association of Governments (SCAG),
they do not take advantage of the broad and highly
accurate geospatial resources available to member cities
and, in some cases, appear to be unaware of what is
available. City planning staff also reported that they are
understaffed and lack certain types of training.
3. The City planning staff reports that the City is in some
respects relatively isolated from some resources that
would ease or enhance progress toward Working Group
goals. For example, utility companies and other permit
issuing agencies do not have offices close to the City of
Commerce, reducing accessibility for the City planning
staff, and the City has no control over utility rates and
permitting processes.
4. Rail is concentrated within the City of Commerce,
more so than any other city in the region (with the
exception of Vernon, located adjacent to the City of
Commerce), and the City has relatively high exposure
to high volume traffic corridors. Both of these land uses
are a major source of air quality hazard, but the City
has no does not have jurisdictional oversight or control
over either.
5. Existing commercial land uses need improvement
and/or redevelopment to improve both air quality and
economic vitality.
6. Although the City lacks control of highways or rail,
they are aware of local land use planning tools that
can address some impacts on sensitive land uses. For
example, adaptive re-use might be useful in transition
areas to address problems of proximity of harmful land
uses near sensitive receptors, as well as to attract new
businesses.
7. The City of Commerce wants to attract artisanal
and cottage industry as new business clusters, and
to encourage entertainment and other attractions to
provide a greater sense of place for residents and as a
means of institutionalizing improved amenities. The
City also favors beautification projects as a means to
reinforce a stronger community image.
The Working Group agreed on a set of recommendations that
were organized into four issue areas, with considerable detail
in terms of strategies and option for each area. The areas are:
• Prevention - Provide separation of facilities and land
uses of concern that represent air toxics hazards from
sensitive land uses
• Reduction - Ways to reduce pollution from existing
businesses and industry
• Revitalization - Opportunities that could contribute to a
"green economy"
• Reinvestment - Provide infrastructure to support
economic growth and protection of community health
-------
The primary focus of this R9 RARE project on the Working
Group activities has been the first goal of the original City
Council directive - separation to provide a safe distance
between hazards and sensitive receptors, using buffer and
land use recommendations in the California Air Resources
Board's 2005 Air Quality and Land Use Handbook for
guidance. The Working Group developed two different
approaches to this goal.
The first was initiated by Angelo Logan, EYCEJ, who
chaired the original City of Commerce Environmental Justice
Advisory Task Force. This approach was a set of proposed
amendments to City of Commerce Zoning Ordinance,
which includes specific land use recommendations derived
from the CARD Handbook. Each recommendation takes
an existing City land use ordinance and amends it to fit the
requirements of the Working Group's charge, and all were
developed in close consultation with Donald Spivack, acting
as a consultant to EYCEJ. A former professional planner with
the City of Los Angeles, he has the longest and most relevant
experience of anyone in the Working Group meetings and
is working in a similar capacity on the City of Los Angeles
Green Zones ordinance (a project referred to as "Clean Up,
Green Up" [CUGU]). This approach is referred to below
as the "zoning amendments". The zoning amendments
were introduced very early in the Working Group process,
discussed and amended during (and between) several
Working Group meetings as members suggested changes
and posed challenges. The research team performed several
types of analyses to provide analytical evidence in support
of the proposed amendments and to help the Working Group
in envisioning how and where the proposed changes would
be applied, and to help them answer specific questions posed
during meetings about the proposed changes. There were
concerns about the number of business that would be affected
by the proposed zoning amendments, and we developed a
series of three metrics to assess the impact of the proposed
zoning amendments on City of Commerce businesses as
shown below:
1. Parcels affected by the proposed buffers: Source:
LA County Assessor's Office Parcel database, 2012
TOTAL 4050 TAX PARCELS LOCATED WITHIN
THE CITY OF COMMERCE BOUNDARY AS PER
ASSESSOR'S RECORDS
2481 PARCELS ARE COMPLETELY WITHIN
THE 300 FOOT BUFFER (11.9 %by area) 2593
PARCELS ARE COMPLETELY WITHIN THE 500
FOOT BUFFER (19.8% by area) 2951 PARCELS ARE
COMPLETELY WITHIN THE 1000 FOOT BUFFER
(26.4% by area)
AN ADDITIONAL 203 PARCELS ARE PARTIALLY
WITHIN THE 300 FOOT BUFFER
AN ADDITIONAL 264 PARCELS PARTIALLY
WITHIN 500 FOOT BUFFER
AN ADDITIONAL 152 PARCELS PARTIALLY
WITHIN 1000 FOOT BUFFER.
Hsi Wl AM1;-NUM KMTi
CITY OF COMMERCE (2013)
O IM1SIM--SSI-S Of COWKRN A-S t}fU-"lNLn IN 'HIE 1»ROH
• OMU.KUrSJSl-.SSES
| ^ [OOOI--]. FtlH-l-KM:RRCH.:NDIE'«illiENSlTtV(- l.AMirs
Map 8. City of Commerce (2013)
2. Businesses affected by the proposed buffers:
Source: Dun and Bradstreet Business database, 2013
TOTAL BUSINESSES LOCATED WITHIN CITY
BOUNDARY: 1797
NUMBER OF BUSINESSES INSIDE 300 FT
BUFFER: 116 (6.5% OF TOTAL FOR CITY)
PARTIALLY INSIDE 300 FT - 76
NUMBER OF BUSINESSES INSIDE 500 FT
BUFFER. 175 (9.7%)
PARTIALLY INSIDE 500 FT - 76
NUMBER OF BUSINESSES INSIDE 1000 FT
BUFFER. 284 (15.8%)
PARTIALLY INSIDE 1000 FT - 58
3. Businesses of Concern: THE BUSINESSES
AFFECTED BY THE PROPOSED ZONING
AMENDMENTS ARE SHOWN AS OPEN CIRCLES
ON THE MAP BELOW. ALL OTHER BUSINESSES
ARE SHOWN AS BLACK DOTS.
Development of and changes to the zoning amendments were
accomplished over several months and a major object of the
Working Group meeting objectives to ensure a transparent
and "good faith" process. However, some throughout this
process, several Working Group members (and non-voting
members) regularly expressed their opposition to any land
use or zoning changes on the basis of their fear that land use
or zoning changes would "hurt business", "limit reinvestment
and expansion of existing businesses", and "result in
businesses not wanting to locate in the City of Commerce".
This 'opposition bloc' did not support their position with fact,
analysis, statistics, or anecdotal evidence. In fact, during one
meeting, a Working Group member of the 'opposition bloc'
21
-------
was critical of the use of "science" in our policy deliberations
because, "Scientists always change their mind, like with
global warming". It is difficult for professional scientists to
defuse an intense and vocal opposition to the use of science.
AMI discussion of this impasse did not occur and, a number
of Working Group members who stated their opposition
to the proposed zoning amendments seldom attended the
Working Group meetings.
It is also worth noting that one voting member who
aggressively supported the 'opposition bloc' position is a
real estate broker with a private company and regularly
explained his personal financial interest in marketing the
City of Commerce to prospective business and industrial
customers as a place with business-friendly land use
practices. Naturally, each stakeholder has a different "stake"
in the process, but a person working on behalf of the City
of Commerce government has a responsibility to do what
is best for the overall good of the City of Commerce, over
and above his own individual interest. This was an obvious
conflict of interest situation.
The Working Group faction opposed to the proposed zoning
amendments retained a land use attorney to evaluate the
zoning amendments and during the 12th (August 7, 2013)
of the 14 monthly meetings of the Working Group and
presented a letter from their attorney. At that meeting and the
two subsequent meetings, no time was provided to discuss
this letter, despite requests to do so. The attorney's letter
presented a series of legal theories describing their clients'
opposition to the zoning amendments, and the letter did
not contain any factual information or scientific analysis
to demonstrate that the zoning amendments would be
Facilities of
Concern
300 ft Buffer
in Commerce
unreasonably burdensome to existing businesses, or would be
otherwise injurious to the City of Commerce. The letter also
contained this concluding clause (italics added):
"In rendering this letter, as to the relevant factual matters
we have examined the reference materials described
herein and such other documents as we have deemed
necessary including, where we have deemed appropriate,
representations and certifications of industry leaders and
public officials. We have made no inquiry, have conducted
no investigation, and assume no responsibility with respect
to, the accuracy of statements made by industry leaders
and public officials, or factual matters contained in any
reference materials."
The opinions and statements in this letter raised certain
questions among Working Group members, and there
were two questions that were neither aired nor discussed
because the facilitator would not allow time for it or allow
a discussion to take place during the August 7th meeting
when requested. The two most noteworthy issues revealed
by the letter are: (a) the opinion that the proposed zoning
amendments would result in large and burdensome negative
impacts (delays, costs and uncertainty) on existing and
new businesses in the City of Commerce, and; (b) the fact
that the letter contains no supporting evidence that current
zoning regulations are inadequately protective. The research
team attempted to address these two issues using the
analytical means at our disposal (EJ screening approaches).
The first issue noted in the letter is an opinion and cannot
be objectively tested or quantified. To address the second
issue, the research team presented maps to the Working
Group and calculated CI scores using EJSM and CES. This
Unregulated Facilities
A Businesses ot Concern
Otticr businesses
Regulated Facilities
* AB2588 'Hot SpO
• Ha* .'.aste Large Quantify Generators
* SCAQMD Title V
g Solid A'asl* Disposal
• CARS CHAPIS Air Tonics (acuities
• HS: .'. asie Treatment Storage Disposal
-^- Railroatfs
I Sensitive Land Uses
. iMtLmAmriM « »
*
Santa
Oftwwy 4 * 'i
•: • • ••' *">•
Map 9. EFFECTIVENESS OF CURRENT CITY ZONING PRACTICE IN SEPARATING AIR QUALITY HAZARDS
FROM SENSITIVE LAND USES [Facilities of Concern - 300 foot Buffer in the City of Commerce]
22
-------
demonstrated that hazard exposure and overall cumulative
impacts in the entire City of Commerce rank among the
highest both regionally and statewide, and fall well above
the CalEPA action level. We also conducted an industrial
analysis of existing businesses in the City of Commerce (as
shown in the 2013 City of Commerce map on page 21 of this
report), which shows that there are a significant number of
business that pose a hazard to the surrounding community
because of toxic air emissions located within the two buffer
distances recommended by CARB for separation of sensitive
land uses from air quality hazards. Some of these facilities
are regulated by State regulatory agencies but most are not
regulated. The maps on page 22 (Facilities of Concern - 300
foot Buffer in the City of Commerce) and page 23 (Facilities
of Concern - 1000 foot Buffer in the City of Commerce) of
this report summarize this analysis. The City of Commerce
Green Zones Working Group Opportunity Areas (Map),
which displays the areas in the city that are available for
specific types of land use/development, is provided in
Appendix D.
There were no alternatives to the proposed zoning
amendments offered by any Working Group members until
the last of the 14 monthly meetings of the Working Group
on September 23, 2013. This new alternative, offered by
the Industrial Council and Jon Reno, was in the form of a
letter proposing a specific plan with the goal of ensuring
that no new sensitive land uses were allowed to encroach
upon or be sited too close to industrial and commercial land
and facilities. This alternative simply proposed an idea on
how the Working Group might recommend a solution to the
problem of separation of these two land use types, with no
Facilities of
Concern
1000 ft
Buffer in
Commerce
detail, analysis, or supporting evidence or documentation. It
also lacked any definition of "encroachment" or of specific
industrial/commercial land uses or facilities that would
be subject to this plan. In fact, this letter also contained a
critique of the CARB Handbook recommendations, and the
methodology that was used to develop the recommended
distance of separation buffers.
After a very brief discussion of this new alternative,
there was an amendment suggested by the small business
representative, DougAshmore (Officer of Amvac, Inc), to
have the proposed specific plan also operate to not allow new
facilities with toxic air emissions to encroach upon existing
sensitive land uses. At that point, the facilitator suspended
the meeting briefly at the request of several Working Group
members who asked to have a private discussion so that
they could "decide how to vote". At the conclusion of the
meeting, the vote of the eight Working Group members was
5-3 in favor of the amended specific plan alternative. There
were no plans or process for building this idea into a detailed
recommendation that is useful to the City, nor to evaluate it.
Although we can describe the Working Group process, it
is difficult to understand the motivations that explain the
actions of some of the Working Group members. The original
proposed zoning amendments are geographically limited to
the areas inside CARB-recommended buffers surrounding
existing sensitive land uses in the City of Commerce. This
area constitutes about a tenth to a quarter of the area of the
City of Commerce (see above maps). In our view, Working
Group members who did not support the proposed changes
in the zoning ordinance, may have taken this position
•* • • M^Htriy Park
**' * \>" ' »
'/ ^
4. . V-
Aipdn To -•» - ,'•••»
"-«.'* ** V-"" . »
*V ). « *. * • *
'" * * f f- - * F^%* " * ** - MonW&»llo *
.>*V^ '» •'•+„....
- r-t-s-t-^—
^ ***?* ^
Unregulated Facilities
» Businesses ol Concern
Other businesses
Regulated Facilities
. AS25S3 Hoi Spots (acuities
•j Max Waste Large Quantity Generators
• SCAQMDTiDeV
•J Solid Waste Disposal
• CARB CHAP1S Air Tonics facilities
^ issa^^srr
^ti'*\^fa/&A& * *
* - I?' *J ^v •/ /<'*»..
HK Aaste Treatment Storage Discosal fcz^J
* **
-r—r- Railroads
_ Sensitive Land Uses
' WM • , *
ij^^:-:^^4^:.§:^ ••.
^v-^.^^/^^-f^lW1' -•... ..
11 Vi* 'k,l * ' f\' *•/. *» "*• '•"•*;"••:•• *:
***" j • «.,» »» ' B« ^ w'(»twpmi»c.uj>Ni«i?i
t • " 4tMc4r'U *»Spr>ni«
*> t«*"^* -*j*\i^.* • y •
!<*^*U4.,j
*««» . * . J^f^4^- ' *f
•i^* £ * f- *
V •• 0*»n«y ' \
r»#V^i J »"
$3F^~. VA. ••>•••• •• r
S»nu f« Spr'ingj
Map 10. EFFECTIVENESS OF CURRENT CITY ZONING PRACTICE IN SEPARATING AIR QUALITY HAZARDS
FROM SENSITIVE LAND USES [Facilities of Concern - 1000 foot Buffer in the City of Commerce]
23
-------
because the ordinance is a citywide document, and all
property owners would have to be notified of the changes.
Those Working Group members probably feared that some
community members would not understand the geographic
limits to the proposed ordinance changes, so the response to
the proposed changes would come from a larger audience. A
specific ordinance/zoning modification plan can be written to
be applicable to a limited geographical area, and to add (land)
'uses' or 'options' not generally available in the general land
use plan and zoning. This can be accomplished, in effect, by
writing modified "zoning" rules that apply only to parcels
within the specified boundary. Normal municipal practice,
however, usually changes zoning rules to add restrictions,
while plans for specific areas are more liberalizing. In this
case, the normal practice is reversed.
In our view, it is not at all clear how, in this case, a specific
plan differs in practical terms from the geographically limited
zoning amendments, unless the intent is to use less protective
buffers or to mandate changes that are not in accordance
with separation of uses as recommended by CARD. What
we suspect is that some of the Working Group members only
want to direct their focus onto two selected areas inside the
City of Commerce - the Washington-Atlantic corridors, and
perhaps the area near the MetroLink station. This could leave
some sensitive land use areas, land uses that were identified
and agreed upon by vote, without any protection.
Unlike the situation in the second pilot application region
(SJV), there has been significant ground truth data collection
and field work of various types (including air pollution
monitoring) performed over the past several years in the
City of Commerce, so typical ground truthing activities
such as verification of facilities, community identification
of undocumented hazards, etc., was not required for the SJV
pilot.
During our collaboration with the Working Group, we
noticed that the City Planning Department did not maintain a
geospatial data library or use geographic information systems
(GIS) as a tool. The Working Group found that GIS data and
analysis was valuable in answering questions and validating
data requirements that were encountered during our
meetings. Representatives of the City of Commerce Planning
Department (Alex Hamilton and Matt Marquez) confirmed
the utility of geospatial data and training for their uses, and
a desire to use this tool. Discussions with Angelo Logan
and Isella Ramirez confirmed that EYCEJ also found this
data and technology useful and would like to have in-house
geospatial capabilities.
After discussions with EPA R9, ORD and a recommendation
from Deldi Reyes, the "ground truth" aspect of the City
of Commerce pilot application was redesigned to help the
City of Commerce stakeholders improve their access to and
analytical capabilities with geospatial information to assist
them in understanding land use and zoning, environmental
risk and exposure, and other issues of local importance to
both the municipality and the community. We developed a
data library to be shared with both the City of Commerce and
EYCEJ, and offered training in data exploration, cartography,
and spatial analysis of this data using both ArcGIS and
Google Earth. ArcGIS is the professional standard for
working with this type of information and is a standard
geospatial analysis tool, and is available to both groups on
their internal computers. Google Earth is useful primarily for
data exploration and cartography, but it also has tremendous
advantages in cartography, adding geographic context by
using the aerial imagery of the chosen location(s) and the
integrated search capabilities of the Google applications
suite. It also has the obvious advantage of requiring less
training to make it usable by a wide variety of users. We
worked to develop a comprehensive geospatial data library
of the City of Commerce and surrounding neighborhoods,
and also to provide relevant training in both ArcGIS (for
the City of Commerce and higher-level EYCEJ staff) and
Google Earth (for other users). Included in the training
package developed for the City of Commerce are techniques
for automation and error checking for new data, and specific
procedures for updating the datasets to keep them current.
The university collaborators plan to populate and maintain
updated data on an ftp site for a reasonable period after
expiration of this cooperative agreement. Geospatial data in
the data library includes:
1. Land use and land cover information from various
sources
2. Automated land use and zoning information from the
City of Commerce files
3. Real estate tax parcel information from the LA County
Assessor
4. Facility location and information on environmental
hazards from various government databases
5. The Topologically Integrated Geographic Encoding
and Referencing system (TIGER) and American
Community Survey information from the US Census
6. Various types of information on boundary files (County
and State government administrative districts, Air
Quality Management District (AQMD) designations,
street files with address-matching capacity, mass transit
and other transportation infrastructure, etc.)
7. Information and location of businesses and
non-business (sensitive land use) facilities (schools,
healthcare facilities, childcare facilities, parks, managed
care facilities, etc.) by tax parcel.
8. Aerial imagery
9. Other data sets as identified during work on this part of
the project
24
-------
San Joaquin Valley stakeholders were and are interested in
understanding the similarities and differences, and various
trade-offs among the three environmental justice screening
methods used in the SJV - our EJSM, CEVA (Cumulative
Environmental Vulnerability Assessment) developed by the
Center for Regional Change at the University of California
Davis, and CES (California Communities Environmental
Health Screening Tool or CalEnviroScreen) developed by
CalEPAOEHHA.
We completed a detailed comparison of the three
CI screening methods. This comparison was done to:
(a) explore the regional pattern of CI scores by the three
CI screening methods for the eight county SJV region;
(b) compare results of the three CI screening methods
to identify areas of agreement/disagreement in relative
CI scores as a means to select areas for ground truthing
under this cooperative agreement; (c) define and identify
impact/vulnerability CI "hot spots" for each method and
determine the degree to which the methods agree/disagree;
(d) summarize differences in data and metrics used, and CI
scoring algorithms for each of the three methods.
Due to the complex nature of the inter-comparison between
ESJM, CEVA, and CES, along with the need to fully explain
to the SJV community partners the implications inherent
in the analysis results, a decision was made to develop
and present two separate webinars to provide information
on the analysis results to the SJV community. The two
webinar presentations were identical and allowed everyone
in the large SJV community stakeholder group to review
the analysis and to pose questions for the research team to
answer. Some questions from the SJV community could
be answered during the two webinars, but more in-depth
questions were answered by the research team in writing and
sent to the entire SJV community stakeholder group after the
second webinar. The research team reviewed and answered
each question in a PDF document that was distributed to the
entire SJV community stakeholder group.
The comparison effort between ESJM, CEVA, and CES was
completed, and the results of the subsequent analysis reported
to the SJV stakeholder group in webinars on June 27 and
July 11. The PowerPoint used in that SJV project partner
webinar is included in Appendix A of this report. Following
the webinar sessions with the SJV community partners, the
geospatial data layers, detailing the comparison of the three
CI screening methods, was posted on an EPA R9 website
to allow stakeholders to examine and explore the map
patterns and CI screening scores during the comment period.
Successful posting of the data was completed on August 30,
2013 with the comment period extending from that date until
September 13, 2013. There were only two responses. One
respondent suggested Tulare as a desirable ground truthing
location, but provided no justification or connection to a
policy relevant question.
Following the webinars, the geospatial data layers detailing
the comparison of the three methods was posted on an EPA
R9 website to allow stakeholders to examine and explore
the map patterns and screening scores during the comment
period. Successful posting of the data was completed on
August 30, 2013 with the comment period extending from
that date until September 13, 2013. There were only two
responses. One respondent suggested Tulare as a desirable
ground truthing location, but provided no justification or
connection to a policy relevant question.
The web viewer for examining the maps was made available
to stakeholders in an improved version, with side-by-side
viewing capability, and a comment period that extended
to September 30, 2013. During the extended comment
period, we received two additional comments. We agreed
on a schedule that called for all comments to be delivered
to the project team on October 4, 2013 and a conference
call on October 8, 2013 to discuss all comments. Editorial
suggestions were then collected from the project team and
compiled into a final version, which was discussed in a
research team conference call on October 23, 2013 and
finalized October 25, 2013.
The detailed comparison of the three screening methods that
was presented during the June 2013 and July 2013 webinars
addressed the following questions:
1. What is the distribution of final CI screening scores
throughout the SJV area in terms of population and
area?
2. A principal purpose of CI screening is to find the areas
of highest impact and vulnerability - the areas with the
extreme CI scores. How should this extreme be defined
using these three methods, and how does the map
pattern of these extreme CI scores compare among the
methods when viewed region-wide?
3. Comparing these maps of extreme CI scores among the
three methods, where is there agreement?
4. Because ground truthing validates and explores the
location and distribution of hazards, an element that
is only one part of each screening method, where are
the locations with high extremes for hazard/pollution
exposure, using each method, and how do they
compare?
25
-------
After an introduction to how the comparison was done, the
webinar PowerPoint was organized to answer these four
questions:
1. What is the distribution affinal screening scores
throughout the SJV area in terms of population
and area [refer to slides 21-25 (Appendix A pages
A-9-A-10)]
All three methods characterize the majority of SJV
area with low CI scores. The EJSM shows high CI
scores for a very small portion of the SJV, and there is a
smooth pattern of variation in area vs. score. The CEVA
pattern is similar, but far more irregular with a distinct
"peak" in middle of scoring range. CES has an irregular
(multi-modal distribution) pattern, with several "peaks"
throughout the range of CI scores.
Differences in CI screening scores vs. population are
affected by variations in the CI scoring procedure
among the methods, but also by the size of the area
scored. The EJSM, which scores at the tract level,
shows most of the population with middle range scores,
a pattern that approximates a "bell curve". CES shows a
pattern of increasing population with higher CI scores,
while CEVA has a very irregular distribution of CI
scores.
The Gaussian ("bell curve") distribution best describes
the nature of the data. The low population density and
highly clustered nature of population in SJV suggests
this distribution, and any CI screening method that
identifies extremes (CI "hot spots") should score most
areas in the middle of the range. The EJSM has this
distribution in other areas of the state where it has been
applied.
2. A principal purpose of screening is to find the areas of
highest impact and vulnerability - the areas with the
extreme scores. How should this extreme be defined
using these three methods, and how does the map
pattern of these extreme scores compare among the
methods when viewed region-wide? [Refer to slides
26-37 (Appendix A pages A-12 -A-13)]
We evaluated the range of CI scores that define the high
extreme, or CI "hot spot", in each method, and found
that the distribution of CI scores was best examined
in terms of population represented by the "high score"
class for each method to mitigate the complication of
differences in scoring methods. We did two different
comparisons, each defining the high extreme differently.
The first comparison used the very highest CI scores
for each method (referred to as "top quantile"), that
represents about one tenth of SJV population (from 7.2
(EJSM) to 11.24% (CES)). A second comparison used a
more broad definition ("high quantile"), that represents
about one fifth of SJV population (20.7 (CES) to 24.6%
(CEVA)). This second definition has greater policy
relevance as OEHHA designates the top 10% of all
CES_SJV Score vs. population (ZCTA)
#j£€$ilP
IQOOO 20000 30000 40000 HK300 60000 /OOOO JiDOOO 900QO
CES_Statewide Score vs. Population (ZCTA)
,- *
'
Figure 4. CES Score (SJV and Statewide) Versus
Population (ZCTA)
ZCTAs as above the action threshold for application
to legislation; and example is its role in SB535, which
determines allocation of money to disadvantaged
communities from the greenhouse gas reduction fund.
Generally, all three methods agree on large, dense
population centers, including Bakersfield, the Fresno/
Selma area, Stockton, Modesto, Tulare and Madera.
EJSM and CEVA do so with greater geographic
specificity than does CES because of the limits of
ZCTAs, and they are in general agreement on these
more geographically specific parts of those large
populated areas. EJSM and CEVA also identify
numerous small population centers and towns, but
usually not the same ones. Smaller spatial unit of
analysis (EJSM and CEVA) makes identification of
small towns more likely. The greater focus on census-
based metrics in CEVA scoring increases likelihood for
small towns to receive high scores, when compared to
EJSM and CES. Overall, EJSM identifies fewer areas,
and more geographically focused areas, than the other
two screening methods.
3. Comparing these maps of extreme scores among the
three methods, where is there agreement? (Appendix A
pages A-14 andA-16) are a series of maps, which zoom
into the areas where all three methods agree. The areas
of best agreement are:
• East Bakersfield, particularly the area east of Interstate
99 and between Highways 58 and 204 (Appendix A
page A-14])
26
-------
CEVA_SJV Score vs Population (block group)
2000 4000 6000 8000
12000 14000 16000 18000
EJSM Cl Score vs. Population (tract)
Figure 5. CEVA Score (SJV) versus Population (Census
Block Group) and EJSM Score Versus Population
(Census Tract)
South central Bakersfield in the area east of Interstate
99 and west of the Bakersfield Municipal Airport
([Appendix A page A-15])
• Downtown Fresno, and extending southwest across
Interstate 99 to Hyde Park ([Appendix A page A-16])
• A large area of southeast Fresno between Interstate 99
and Highway 180 ([Appendix A page A-16])
Selma, a town about 15 miles south of Fresno
downtown, along Interstate 99 ([Appendix A page
A-16])
4. Because ground truthing will be validating and
exploring the location and distribution of hazards,
an element that is only one part of each screening
method, where are the locations with high extremes
for hazard/pollution exposure, using each method, and
how do they compare? [Refer to (Appendix A pages
A-17-A-19)]
A number of areas in the SJV showed agreement in
hazard/pollution scores used by the three methods.
Most of these are located in the areas as those
mentioned above for agreement regarding high total
scores. When considered in terms this comparison
for agreement among the three methods, there are
several areas where the methods agree that hazard/
pollution scores are high, but where the overall score
is not in the high extreme. The most prominent of
these are Stockton ([Appendix A page A-17]), Modesto
([Appendix A page A-18]), Clovis, northeast of Fresno
([Appendix A page A-18]), and portions of greater
Bakersfield ([Appendix A page A-19])
The results of the comparison of CI screening methods
and map patterns of CI scores, summarized, above, show
reasonable agreement among the methods in spite of the
many differences in methodology and data sources. This
gives us additional confidence in the process of CI screening
as a robust and meaningful technique for identifying areas
that represent the "high end" extreme of vulnerability
and impact/exposure. It also underscores the prevailing
conclusions by many working in the SJV of a pattern of
inequitable exposure to environmental hazards and their
attendant risks, and vulnerability to those risks.
It is worth mentioning that during the webinar, while trying
to explain some of the reasons for the patterns among the
three screening methods, there was a discussion about a
disagreement in interpretation that deserves a thoughtful
examination. CES identifies ZCTAs mostly associated
with large population centers, while CEVA and (to a lesser
extent) EJSM identify small towns as isolated population
concentrations not picked up by CES - at least using the
"extreme" CI scores values. In the webinar, we suggested
that CES uses "population weighting" which, strictly
speaking, is incorrect because it seems to refer to the CES
CI scoring method itself. However, this comment was meant
to try to make sense of the mapping results. Looking at the
data closely, it is clear that CES CI scores are skewed by
population in a way that appears like weighting, as the graphs
on page 26 and page 27 of this report demonstrate.
As can be seen in the graphs, on page 27 for both SJV and
Statewide, CES tends to give higher CI scores to more
populated areas, while EJSM and CEVA clearly do not. See
the graphs on page 27 displaying the same relationships for
CEVA and EJSM.
One possible reason is that ZCTAs have very different sizes
and populations, ranging up to over 110,000 statewide. This
is also true for the full US sample and of SJV. Tracts and
block groups, on the other hand, are defined specifically to
sample population and efforts are made to ensure they are
not too large or small. For example, tracts have a population
range where the least populated tracts are about 500 persons,
but they quickly go to 5,000 and stay at that level (with some
much larger at the extreme end). ZCTAs are not established
to ensure a fair amount of uniformity in population size so,
while CES does not population weight per se, scoring using
ZCTAs has the potential for some non-standard statistical
distributions, which is what we may be observing here.
In addition, since ZCTAs are physically larger in terms of
their area, they can make outlier tracts and/or small towns
that lie within them literally "disappear". We found in our
SJV mapping analysis and comparisons that there were
several ZCTAs that were not high scorers on CES, but within
27
-------
some of the lower scoring ZCTAs (by overlaying EJSM
results) that there were tracts that were "hot spots" according
to EJSM.
EPA R9 and ORD worked with the university collaborators
(Occidental College, UC Berkeley, and USC) and others
to incorporate stakeholder comments into the decision for
the location for ground truthing. This decision addressed
the following criteria: (a) community capacity to engage in
ground truthing, (b) anticipated value in determining whether
and how screening CI methods can inform the policy relevant
questions, (c) capacity to improve future development of
the CI screening methods, and (d) selection of an area that
includes a range of CI scores, from low to high, to help us
get an idea how ground truth validation relates to highly
impacted areas vs. areas with lower impact and vulnerability.
The complete PowerPoint slide presentation comparing
the CI methods (CEVA, CES, and EJSM) presented to the
stakeholder group in webinars on June 27, 2013 and July 11,
2013 is provided in Appendix A Pages A-2 -A-19.
28
-------
The Working Group authored and finalized its final report,
and delivered that report to the City of Commerce City
Council for review and further action. The Executive
Summary of that report, a Planning Commission Staff report
on the proposed Green Zones policy recommended by the
Working Group, and City Council agenda for the meetings
where the decisions were reached is included in Appendix B
of this report.
In mid-November 2013, the City approved three of the
four major recommendations of the Working Group in
its Green Zone Policy. Below is a news media report
summarizing the City of Commerce City Council
decisions on this policy (http://egpnews.com/2013/ll/
commerce-approves-three-prongs-of-green-zone-policy/).
A polity that would protect residents from toxic exposures
and will create job opportunities.
By Jacqueline Garcia, EGP Staff Writer
A local city with a large industrial base, crossed by freeways
and railyards that is often cited as having some of the most
polluted air in the region, has adopted a policy environmental
advocates hope will make the city healthier for its residents.
At the Nov. 5th City of Commerce council meeting, elected
officials voted to approve a Green Zones Policy supported by
a city task force and local activists. Supporters said not only
will the policy bring a healthier community; it will also create
more local job opportunities in manufacturing, specifically in
food production and artisan businesses.
Railyards, freeways and other industrial lands uses cause
highly concentrated levels of pollution that affect the health
of Commerce residents, workers and visitors, according to
East Yard Communities for Environmental Justice (EYCEJ),
an environmental health and justice organization backing
the policy. Ongoing exposure to these toxins can aggravate
asthma, cause pre-term births, low birth-weight babies, lung
disease, heart attacks, cancer and premature death, according
to East Yards, according to the Green Zones executive
summary.
In a study by the California Environmental Protection
Agency, Commerce was identified in the top 5 percent of
communities in California with the highest pollution burdens
and vulnerabilities.
Members of East Yard and Commerce residents were at the
Nov. 5 council meeting to push for passage of the Green
Zones Policy, a four-pillar plan that according to East Yard's
website will prevent toxic exposure to residents from new
land uses; reduce the level of existing impacts through
voluntary business collaborations, allowing participants to
utilize less polluting equipment; revitalize local economic
opportunities to contribute in a vibrant economy and increase
of jobs, and reinvest in key boulevards to bolster business
and quality of life opportunities."
At the meeting, Commerce resident and East Yard member
Tofia Lupercio presented a petition with hundreds of
signatures from community residents supporting the policy
to the council. "This policy is critical and we need your
leadership," she told council members.
Three of the Green Zone policy initiatives were developed
and sponsored by the city's Green Zones Policy Working
Group. The fourth, dealing with the issue of "prevention,"
was added by East Yard, stating it is needed in order to stop
the exposure of residents to toxic and harmful pollutants.
"The reason that the council did not include the [fourth]
element that amends the zoning ordinance to create buffer
zones, from my understanding, is that they did not fully
understand the recommendation," said Angelo Logan,
member of East Yards.
The new green policy will amend city-zoning law to restrict
new toxic land-uses close to homes, schools, churches and
senior centers.
After much deliberation, the council voted unanimously to
approve the working group's three original recommendations,
but decided to conduct study sessions on East Yards
recommended ordinance. This element is aimed to prevent
new hazards that range from truck idling at truck stops and
warehouses to chemical handling facilities.
"[East Yards] members are very excited that the council has
decided to adopt three of the four elements and will consider
the 4th one in the coming months." Said Logan.
The approval of the Green Zone Policy will help reduce
environmental dangers in the community, prevent pollution
and revitalize neighborhoods through targeted economic
development strategies, according to East Yard.
As part of its goal, the policy aims to create a protected zone
around sensitive land use areas such as schools, playgrounds,
homes, and daycare and senior centers to improve public
health.
The policy has been years in the making. In June 2011,
Commerce's Environmental Justice Advisory Task Force
urged city officials to hold workshops to explore ways to
maintain the city's focus on businesses and industry, while
also protecting the health of its nearly 13,000 residents.
29
-------
-------
8.0
Pilot Application 2: San Joaquin Valley (SJV) Site
Selection Process
We defined the goals of the ground truthing part of this
project by selecting locations for ground truth data collection
in the SJV, and completing the community-based data
collection and analysis phase of the project. The conceptual
framework for the ground truthing selection reflected the
results of a detailed comparison of the three screening
methods, and their agreement/disagreement in identifying
areas of environmental justice concern. This comparison
focused on defining and identifying specific areas for
each method where impact/vulnerability was highest, and
to compare these locations among the three methods to
determine the degree to which the methods agree/disagree.
An initial comparison revealed that the most significant
reason for differences in scoring between CES and the
other two screening methods was the fact that CES CI
metrics are scored only on a statewide basis, unlike EJSM
and CEVA, which score for a predefined region. It was
important to remove this difference for the purposes of
our comparison study in order to make comparisons of the
results meaningful.
We reviewed the numerous responses from project partners.
The responses varied significantly in their interpretation of
the comparison results, and in suggestions on how to proceed
with ground truthing. EPA R9 and ORD thought it important
that the core group develop responses to the comments, and
distribute them to the project partners. They also asked the
university collaborators to add some additional analysis and
maps to the comparison in order to clarify some important
conclusions, and to directly address specific commenter's
concerns.
Telephone meetings by the research team on October 23,
2013 and October 25, 2013 accomplished a review of
the responses to comments and an agreement to do some
additional analysis to clarify some questions posed by some
of the commenters. A summary of these responses, and the
project team's comments on the responses, were distributed
to the project partners in mid-November 2013, and is located
in Appendix C of this report.
Because of concern about how the slow pace of decision-
making on ground truthing goals and locations was
affecting the project, the research team took the initiative
Map 11. EJSM Ground-Truthing Locations in the SJV
31
-------
Map 12. Map of San Joaquin Valley Region.
to systematically detail the options available, and created
a set of specific recommendations on goals and ground
truth locations, including a ranking of the options. The late
November 2013 holiday period and travel by some of the
core group delayed a meeting on these recommendations until
early December 2013. The decision was made to use the
ground truthing results to (a) test for "false positives" among
areas with the highest total CI screening scores, as well as
those areas with high CI scores for pollution/hazard exposure,
and (b) determine how data corrected by the ground truthing
affected EJSM CI scores.
Three CI analysis sites were chosen as the best candidates
to address these goals, and to provide some reasonable
geographic variation within the very large and diverse S JV
area, which spans eight counties. Although the original
project workplan specified ground truthing in only one
location, we had the resources to be a bit more ambitious as
a result of our efficient use of the budget on fieldwork for the
City of Commerce Pilot Application. These locations - the
town of Arvin located southeast of population and commerce
center, Bakersfield; Huron, a somewhat isolated community
almost completely dependent on agriculture, and a
historically persistent environmental justice community; and
central Stockton, where the EJSM finds very high pollution/
hazard exposure cumulative impacts where the other two
tools do not - are shown in the map location above: Map 11.
EJSM Ground-Truthing Locations in the SJV.
Following the December 2013 holiday period, the project was
affected by change of EPAR9 personnel. Jacqueline Hayes
transitioned off of the project as EPAR9 lead Technical
Advisor. Jacqueline originally took on the role of technical
advisor for EPA R9 during the second six-month period
of the project. Charles Swanson, who had been assisting
Jacqueline Hayes for several months, became the EPA R9
lead Technical Advisor, and Jacqueline Hayes continued to
assist the project for several weeks into mid-February 2014
before formally leaving the project.
We received the anticipated comments from OEHHA in
January 2014, which were valuable and substantive, and
prepared a detailed response. These comments informed the
project significantly, but did not change the decisions made
previously on ground truthing goals and CI analysis sites.
32
-------
\
Map 13. Map of San Joaquin Valley showing three areas where ground-truth validation was completed.
With the goals and locations for ground truthing finalized,
and informed by the thoughtful OEHHA input, EPA R9
began the process of working with community leaders from
the project partners to identify community groups who would
receive the training for data collection, and accomplish the
field portion of the ground truthing effort. The decision
to ground truth in three locations depends on community
capacity to engage in ground truthing, and we originally
anticipated including SJV Cumulative Health Impacts
Project and their community partners in this process. The
two project partner members representing community groups
- Sarah Sharpe, Co- Director of Fresno Metro Ministry,
and Cesar Campos, Coordinator for the Central California
Environmental Justice Network (CCEJN) were identified by
EPA R9 to aid in identifying and enlisting community groups
for the ground truthing training and fieldwork.
Both of these community leaders are based in Fresno, where
the patterns of environmental justice and vulnerability are
clear and unambiguous, and where all three screening method
largely agree. For this reason, no ground truthing was
planned for that part of the San Joaquin Valley. However,
this made it difficult to find community members to engage
the fieldwork in the three CI analysis sites identified for this
project, and further progress on ground truthing was again
delayed. Eventually, Sarah Sharpe was unable to continue
her supportive role in this process, and Cesar Campos
worked hard to find community partners in Arvin, Huron and
Stockton to work with us.
At this time, Cesar had tentatively identified groups
interested in helping us with the ground truthing effort,
but had not obtained a final commitment or dates for the
group training necessary to proceed. The ground truthing
activity for the Arvin, Huron, and Stockton communities was
ultimately completed on July 19, 2014.
In preparation for the field data collection, we collected and
accomplished initial error checking on all geospatial and
facility data for the three ground truthing areas in preparation
for our field technician to do the data collection outlined
in the QA/QC plan for this cooperative agreement. This
fieldwork occurred May 7-10, 2014 and July 17-19, 2014.
33
-------
-------
9.0
Pilot Application 2: San Joaquin Valley (SJV)
Detailed Region/Site Maps
The study area for Pilot Application 2 is the eight county
southern San Joaquin Valley (SJV) region, as shown in Map
14 below.
During the Planning and Design process for this pilot
application, the Project Partners clearly identified their desire
to better understand the similarities, differences, and the
various trade-offs among the three environmental justice
screening methods used in the SJV: a) our Environmental
Justice Screening Method (EJSM); b) the Cumulative
Environmental Vulnerability Assessment (CEVA), developed
by the Center for Regional Change at the University of
California Davis, and; c) the California Communities
Environmental Health Screening Tool or CalEnviroScreen
(CES), currently under development by CalEPA OEHHA.
The possibility of making a comparison between these three
methods was anticipated in the Planning and Design section
of the research workplan and integrated into the project, to
apply ground-truthing to validate and correct the facility-
level data used by these three screening methods.
One key element of this pilot application is using ground-
truthing to validate (and correct, as necessary) the location
accuracy and activity of hazardous facilities and sensitive
land uses used in Environmental Justice (EJ) screening.
This process began with an activity to identify the specific
areas for ground-truthing. The conceptual framework for
the ground-truthing was designed to reflect the results of a
detailed comparison of the three CI screening methods, and
their agreement/disagreement in identifying specific areas
of high EJ cumulative impact scores. This comparison
focused on defining and identifying areas for each method
where impact/vulnerability was highest, and comparing these
locations among the three methods to determine the degree
of agreement/disagreement between them. This information
was then used to select the specific areas for ground-truthing.
Our initial comparison revealed that the most significant
reason for differences in the CES CI scores, as compared to
the other two screening methods (EJSM and CEVA), was
Map 14. Map of San Joaquin Valley Region.
35
-------
the fact that CES CI metrics are scored only on a statewide
basis. Both EJSM and CEVA score for a pre-defined
region determined by common data availability and quality,
common pattern of impact and vulnerability factors, and
the geography of decision-making. It makes sense to score
the SJV as a region, as these eight counties share a broad
concern about pesticide exposure, water quality, ozone and
paniculate matter pollution, impacts from agriculture and
the petroleum industry, and social factors that act as effect
modifiers. There is no transportation planning agency or
regional authority that monitors land use, so this information
was extracted from real estate parcel data from each county's
tax assessor's office.
It was important to remove this difference to make the
comparison results meaningful, so final comparison and
ground-truth site selection was delayed to allow time for
the California Office of Environmental Health Hazard
Assessment (OEHHA) to calculate scoring metrics using
only data for the San Joaquin Valley region. Although it took
time to receive this data, we felt it was very important to
compare the most current CI screening scores possible, and
to ensure that the comparison was meaningful by using the
CI scores for the three methods in the same (SJV) region.
After the webinars were held in 2013, the review of the
project partner comments was completed in 2014 and the
results of the updated comparison were shared with the
project partners. At this time, it was decided that in addition
to distributing the PowerPoint presentation summarizing the
comparison, project partners would also be able to examine
the details of the comparison of the three methods using
geospatial data layers in an online GIS mapping
application. These data were posted on to an ESRI-Online
map viewer imbedded in the EPA R9 website (during the
comment period).
The numerous responses and suggestions from project
partners were collected and reviewed, prior to being
incorporated into the decision-making aspect of the project.
The responses varied significantly in their interpretation
of the compared results, and in suggestions on how to
proceed with ground-truthing. We spent considerable time
developing responses to all comments. The responses
were helpful in guiding the research team to add additional
analyses and maps to the comparison activity to directly
address specific commenter concerns and clarify the
meaning of some important conclusions. A summary of these
responses, and the project team's comments on the responses,
were distributed to the project partners and are included in
Appendix C.
The research team took the initiative to systematically
detail the options available, and created a set of specific
recommendations on goals and ground-truth locations,
including a ranking of the options. In consultation with
EPA R9 and ORE), the decision was made to use the ground-
truthing results to (a) test for "false positives" among areas
with the highest total screening scores, as well as those
areas with high scores for pollution/hazard exposure, and (b)
determine how data corrected by the ground-truthing affected
EJSM scores.
Three CI analysis sites were chosen as the best candidates
to address these goals, and to provide some reasonable
geographic variation within the very large and diverse
San Joaquin Valley area, which spans eight counties.
The SJV locations selected were: a) the town of Arvin,
located southeast of the population and commerce center,
Bakersfield; b) Huron, a somewhat isolated community
almost completely dependent on agriculture, and a
historically persistent environmental justice community,
and; c) central Stockton, where the EJSM finds very high
pollution/hazard exposure where the other two tools (CES
and CEVA) do not.
Once the goals and locations for ground-truthing were
finalized, and the OEHHA input incorporated, we worked
with EPA R9 and community leaders from the project
partners to identify community groups who would receive
the training for the ground-truthing data collection effort.
The two project partner members representing community
groups, Sarah Sharpe, Co- Director of Fresno Metro Ministry,
and Cesar Campos, Coordinator for the Central California
Environmental Justice Network (CCEJN) were identified
by EPA R9 to assist in identifying and enlisting community
group members.
In preparation for the community based participatory
research (CBPR) portion of this pilot application via public
participation in ground-truth data collection, we collected
and accomplished initial error checking on all geospatial
and facility data for the three ground-truthing areas. This
preparatory work was done so that the field technician
could perform the data collection as outlined in the QA/QC
portion of the research workplan. A CBPR project depends
on community capacity to engage in the field work, and we
originally anticipated including SJV Cumulative Health
Impacts Project and their community partners in this process.
Both of these community leaders are based in Fresno, where
the patterns of environmental justice and vulnerability
are clear and unambiguous, and where all three screening
method largely agree. For this reason, no ground-truthing
was planned for that part of the San Joaquin Valley, but the
time and distance involved in training and fieldwork made
it difficult to find community members to implement the
fieldwork in the three CI analysis sites identified for this
project. Cesar Campos worked hard to find community
partners in Arvin, Huron and Stockton to work with us,
but with the project timeline approaching deadlines, we
were ultimately unable to solidify community participation
as originally planned. The SJV community groups were
supportive and enthusiastic about this project, but they just
could not make arrangements to complete the training and
fieldwork within the time limits of the project.
Given the time limitations of the SJV community groups, we
devised an alternative plan to complete the requirements of
this pilot application and still accomplish the stated goals in
36
-------
Map 15. Map of San Joaquin Valley showing three areas where ground-truth validation was completed.
the approved research workplan. The revised approach also
responded to the changes and improvements in the CalEPA
screening tool CalEnviroScreen (CES). These changes in
CES occurred during the second year of this project, and had
an obvious impact on the comparison between methods. This
revised approach was also endorsed by our SJV stakeholder
partner, Cesar Campos, who consulted with the community
groups on the proposed changes.
The ground-truthing part of this project was designed to:
a) validate accuracy of established facilities and land uses
from professional sources and available databases as a
way to check the accuracy of their use in screening tools;
b) determine the impact on CI screening scores using
unchecked (with location and other errors) hazard and facility
data as a test of the EJSM's susceptibility to identifying
false positives, and; c) involve community members in the
process of evaluating what should be included in a screening
method or tool, and how screening results affects their
role, positively and/or negatively, in decision-making and
policy development.
This plan was accomplished by: 1) completing field-based
ground-truthing validation of all facility information for the
three selected test areas by Occidental College/UC Berkeley/
University of Southern California staff; 2) reviewing and
correcting all facility data for the entire SJV region using
Google Earth Pro, and; 3) rescoring the SJV region using
validated and corrected facility data to look for differences
resulting from using unchecked (error filled) vs. validated
(errors corrected) information. This approach is in some
ways superior to the original approach, in that it is a more
comprehensive test of false positives/negatives involving
the entire eight (8) county area and not just the three
neighborhoods/localities identified for ground-truthing.
The results of this work was presented at a regional EJ
meeting (the Central California Environmental Justice
Network [CCEJN] conference in Fresno, CA on Saturday
September 6, 2014) to: a) explain to the community members
the results of our work; b) engage them in a conversation
about the way it was done and why; c) solicit their ideas and
input on which types of hazards should be included, what is
missing, and what does their experience suggest we should
include that we may have excluded. This provided a means
to incorporate the community perspective into the process,
including their ideas on hazards and EJ issues, and their
reactions to our attempts to create and improve screening
tools to guide decision makers.
37
-------
Community Engagement - Central California
Environmental Justice Network (CCEJN) Conference:
The community engagement component for this research
project culminated in our participation in the annual
conference of San Joaquin Valley (SJV) environmental justice
community organizations, "Roots of Resilience, 2014",
sponsored by the Central California Environmental Justice
Network (CCEJN) on September 6, 2014 in Fresno, CA. The
agenda of the meeting is attached as an addendum, which is
provided in Appendix E. The conference attendance included
more than 100 attendees from community organizations
throughout the eight county SJV region. Also in attendance
were our two community partners associated with this (SJV)
ESJM pilot application, Cesar Campos, Coordinator for the
Central California Environmental Justice Network and Sarah
Sharpe, Co- Director of Fresno Metro Ministry, as well as
other project partners and EPA R9 staff who took part in the
planning and design of this pilot application.
We presented our work in an afternoon workshop
titled "Partnering with Environmental Agencies and
Communities to evaluate the Environmental Justice
Screening Method". We estimated attendance at this
presentation and following discussion, which took place over
a two-hour period, at approximately seventy five persons.
The following description of the event was included in the
conference documentation:
We will summarize work on a project funded by EPA R9
to apply the Environmental Justice Screening Method
(EJSM) to help regulators and policy makers more
efficiently target their efforts to remediate cumulative
impacts, environmental inequities, and focus regulatory
action at the neighborhood level. Currently, the burden
of proof is usually placed on communities to demonstrate
the cumulative impacts of environmental stressors. CI
screening such as the EJSM provides a more proactive
approach, removing this burden from vulnerable
communities so that those without the history or capacity
for civic engagement can also receive regulatory attention,
and has potential to advance regulatory evaluation and
the implementation of environmental policies. We will also
report on how the accuracy of agency databases affects
the reliability of screening by comparing it with "ground
truthing" verification study in three key areas of the
Central Valley - Arvin, Huron and Stockton.
Dr. Manuel Pastor, USC PERE, presented a summary of the
EJSM to familiarize everyone with the general process of
environmental justice screening, with particular emphasis
on the indicators used and reasoning associated with the
choice of those indicators. This was followed by Dr. James
Sadd, Occidental College, who presented an overview of
the ground truthing and location validation process, and
results for both the entire San Joaquin Valley, and the three
communities where more intensive work was accomplished
(Arvin, Huron, and Stockton). Based on the questions and
discussion that followed, we felt that the presentation was
clear and understood by the participants, and that they came
to the meeting with some familiarity of EJ screening and its
influence on policy and decision-making. The presentation
FACILITIES FROM PUBLIC DATABASE!
A AB2586
• Autotody
9 CARB_FOI
+ Gas Station
tt Sensitive Landuses
Map 16. Arvin study area showing location and types of facilities identified from standard public databases.
38
-------
Arvin
Ground
Truth
Results
GROUND TRUTH RESULTS
9 ChiMcars
0 FOl
+ Q» saw
Map 17. Arvin study area showing location and types of facilities validated using ground-truthing.
was aided by a language interpreter who translated the
content into Spanish, and it seemed to us that the Spanish-
language speakers also understood the presentation quite
well.
Following the presentation, and some questions to clarify
some of its content, we led a discussion on screening
indicators and the role of screening in addressing community
concerns. The responses and discussion primarily focused on
the following topics:
1. Indicators appropriate for urban vs. rural/agricultural
communities:
Participants brought up questions about specific
indicators that are more important to residents of
these rural/agricultural communities than urban areas.
Examples include pesticide use and water quality.
There was considerable discussion about how the
sources of these data are obtained, how the associated
metrics are developed, and whether they represent
a fair and accurate measure of the issue of exposure
(e.g., does the estimated exposure impact match that
experienced by [the 'lived experience' of] people in the
SJV).
An example that was discussed in some detail is
pesticide use. The fact that pesticide use comes from
a State (California) regulatory database of self-reports
by users, as per State environmental regulations, they
are subject to error by estimation, under-reporting, etc.
However, the penalties for purposeful misreporting are
substantial, and the negative impact of over-reporting
is very low, so we consider the data to be reasonably
accurate. There was concern that the pesticide data is
aggregated by census tract, which does not take into
account movement of airborne pesticide by wind, or
the fact that some classes of people (e.g., agricultural
field workers) suffer greater exposure. Participants also
raised the concern that pesticides are, to them, a more
serious problem and health threat than some of the
other indicator metrics in the screening methods that
are weighted equally.
The EJSM does allow for differential weighting of
indicators in scoring, but we do not apply weights
because there is no scientific basis to determine how
the various measures should be weighted relative to
one another. However, it is certainly appropriate for
different groups to weight indicators in using the EJSM
for various purposes of data exploration, or to address
specific environmental or health/vulnerability concerns,
which is that capability was built into the methodology.
2. Water quality:
Participants were intensely interested in, and passionate
in their discussions about, water quality. We discussed
at length how the water quality indicator metrics are
39
-------
Huron Study area
FACILITIES FROM PUBLIC DATABASES
A AB2588
| Autobody
• CARB_FOI
4> Gas Station
• Sensitive Landuses
A A.
A
t
„ _-... . _
*
Map 18. Huron study area showing location and types of facilities identified from standard public databases.
developed, the data types used, and the fact that the
EJSM indicators were developed in coordination with
CalEPA OEHHA as part of the CalEnvironScreen
(CES) process, and that both CI screening tools
(EJSM and CES) use the same indicators. Participants
noted that the ways in which the data is reported can
contribute considerable error to the original measures
in the State database. For example, when water
wells are tested for contaminant levels, the reported
levels are often averaged for some number of wells to
characterize contaminant concentrations for an entire
aquifer. This averaging process can hide the presence
of one well with a very high concentrations in a group
of cleaner wells (with much lower concentrations),
allowing a small but significant plume of contaminated
groundwater to "appear" cleaner in the reporting than
in reality. This would "undercount" water quality as in
indicator in the screening scores, and decrease the level
of apparent impact for that area.
Some community members also shared significant and
detailed personal accounts of the water quality in their
particular neighborhood, reporting that in some cases
use of purchased and bottled water is required. This is
an economic and convenience burden, but also deeply
disturbing to them in terms of perceived fairness of
environmental regulatory practice, and it obviously
affects their trust in government and in any process of
making improvements. This part of the community
engagement session underscores two important
points. First, water quality, like pesticide use, is a
much more important problem to SJV residents and
careful consideration of weighting them appropriately
should be considered in using CI screening for analysis
and informing decision-making. Second, qualitative
information of this type is valuable and validating, but
tremendously difficult to collect and incorporate into
CI screening tools. In its efforts to listen and respond
to community input, and engage communities in the
process of their work, EPA might consider ways in
which to solicit and collect this type of information in a
systematic manner. Environmental justice CI screening
might help in that effort, by allowing people to see
and understand how their communities are scored, and
then by asking them to respond to note our error or
inaccuracy in the metrics used (using their own local
knowledge and experience). Some sort of web-based
presentation of CI screening results and maps with the
ability for user to provide comments is one way such an
effort might be attempted.
40
-------
Table 9. Facilities in Huron study area with location correction greater than 100 meters.
Facility Name
WESTFIELD GINNING CO.
VERIZION WIRELESS (VANGUARD)
VERIZION WIRELESS - GUIJARREL HILLS
LOS GATOS TOMATO PRODUCTS
WOOLF ENTERPRISES
HARRIS WOOLF CALIF ALMONDS
PACIFIC GAS & ELECTRIC CO
DOLE FRESH VEGETABLES, INC.
AT&T MOBILITY- EH&S COMPLIANCE
WESTERN FARM SERVICE
ANGKOR AUTO BODY AND PAINT
CANEPAS CAR WASH
COMCAST CABLE COMMUNICATIONS
GURU GAS & MARKET
PACIFIC BELL TELEPHONE (DBA AT&T
CA)
MONSANTO COMPANY
TELETECH CUSTOMER CARE
CALIFORNIA
CAL-STATE UPHOLSTERY
Type
AB2588
AB2588
AB2588
GARB FOI
GARB FOI
AB2588
AB2588
AB2588
AB2588
AB2588
AB2588
Gas Station
AB2588
AB2588
AB2588
AB2588
AB2588
AB2588
Result
Location error corrected
Location error corrected
Location error corrected
Location error corrected; now
"Antonini Tomato Company"
Location error corrected
Location error corrected
Location error corrected
Location error corrected
Location error corrected;
duplicate facility
Unable to locate; closest "Crop
Production Services"
Location error corrected
Location error corrected; new
name "Pacific Car Wash
Location error corrected
Location error corrected
Location error corrected
Location error corrected
Location verified; currently vacant
Location error corrected
Error (m)
19,969
10,036
6,386
6,364
3,331
3,327
2,234
836
718
631
423
391
350
222
185
162
142
126
3. Inaccuracy discovered in ground truthing:
There was general agreement that the degree and
type of error and inaccuracy we found in the standard
databases used in CI screening was significant,
mystifying, and disheartening. As we demonstrated,
correcting those errors has some effect on regional CI
scoring, but on a local level the impact of corrected CI
scores can be much greater. Correcting the data will
change CI scores in some census tracts, and which
tracts are the "winners" and "losers" matters greatly to
the residents of those tracts. All agreed that correcting
the data is vital to good process and practice but they
were, frankly, incredulous at the degree of error in
many cases. One particular example is the mislocation
of the Lawrence Livermore National Laboratories
hazardous waste facility by nearly 13 kilometers, and
the Forward Landfill in the Stockton area by nearly
12 kilometers.
Errors like this can usually be traced to a reporter using
an administrative office address as a location, rather
than the location of the actual facility. As bad as this
practice is, it is equally problematic that regulatory
agencies do not have data quality assurance procedures
in place to prevent or correct such errors. For many
41
-------
Huron Ground Truthing Results
Sites with >100 m location error
OIF c*ui« Ji»o*o»
Map 19. Huron study area showing location and types of facilities validated using ground-truthing. Locational error
shown as black lines connecting original and corrected positions.
ID
GT_2
GT 5
GT 6
ST_11
GT.li
GT 13
GT 19
GT 20
GT 25
GT 31
GT 34
Facility Name
NEW • Shell gas station (formerly Chase, Inc.)
MEW -76 gas station
NEW - Hill Crest Travel Plaza Diesel Truck Stop
"4EW- Solar power plant: S.Mnertor company.
NEW • FCC comm tower reg number 1014495
NEW - PG&E Huron Solar Station
NEW - BufartJ Star Mart station
NEW - Harris Fresh Onion Division
NEW - Terra Unda Onion Division
MEW • Five Stars Market station
NEW - Antenna with power generator FCC 005904337 Verizon
Wireless
Type
A82588
Gas Station
Gas Station
A62588
AB25S8
AB2588
Gas Station
AB2S88
AB2588
AB2S88
AB258S
Map 20. Huron study area showing location and types of new facilities (not part of original standard public databases)
discovered during ground-truthing.
42
-------
years, the US EPA was criticized for this type of error
in the Toxic Release Inventory (TRI) program and
it responded with a well-executed effort to find and
correct such errors, and to regularly report its progress
to the public. The current TRI data is very reliable
and accurate, and state agencies should consider
adopting EPA's practice. In most cases, these errors
only have to be corrected once, because facilities move
or close down relatively infrequently and the number
of new facilities each year that must be checked and
verified is not large. This type of error checking could
be accomplished internally by the State agency, or
contracted out, but community confidence and trust in
the State agency would be bolstered if they were to take
on this responsibility themselves. Ground truthing by
the public in the form of community based participatory
research (CBPR), is another way in which to contribute
to this error checking effort.
Overall, we were very pleased by the community response
to our EJSM workshop, and pleased with the quality of the
input we received from the community members. Many of
the participants have since contacted us by email or phone to
add additional information or suggestions, which we greatly
appreciate, as we work to improve the EJSM and contribute
even more to the use of CI screening methodology.
\
„ ~"\ — -Jr_
*
Stockton
Study Area
FACILITIES FROM PUBLIC DATABASES
A AS2M8
• CARB_FOI
+ GasSUtsn
f Sensitive Landuses
Map 21. Stockton study area showing location
and types of facilities identified from standard
public databases. Note locations of duplicate
and missing facility records.
\ *
—A
\
Map 22. Stockton study area showing location
and types of facilities validated using ground-
truthing. New facilities are shown in yellow.
43
-------
Field-based ground-truthing validation of all facility
information for the three selected test areas, Huron, Arvin
and Stockton (see Map 15), were accomplished in two
phases, on May 7-10 and July 17-19, 2014 by Occidental
College/UC Berkeley/USC staff members. These study
areas differ from one another in various ways, including size.
Note: Stockton has an area of 3.4 km2; Arvin has an area of
24 km2, and; Huron has an area of 816 km2.
The standard public databases that are used for EJ screening
tools, and to characterize hazard proximity were queried for
all environmental hazard facilities and sensitive land uses
located within 3000 ft of the boundaries of the three study
areas. This (3000 ft) buffer distance was used to account for
the EJSM hazard proximity analysis, which evaluates hazards
within three annular 1000 ft buffers.
These locations were first examined for locational accuracy
using Google Earth Pro (GEP), and all were geocoded to
further check facility location. These included:
California Air Resources Board (CARB) "Facilities of
Interest" (FOI) - consists of a subset of facilities from
the California Emission Inventory Development and
Reporting System (CEIDARS) statewide air toxics
emissions inventory of greatest concern to regulators
because of amounts, toxicity, possible impacts of
emissions
California Department of Toxic Substances Control
(DTSC) permitted hazardous waste handling facilities
and generators
Auto paint and body shops from the Dun and Bradstreet
Business Locator Service
Facilities reporting to the AB2588 Air Toxics "Hot
Spots" Information and Assessment Act Program -
Note: the objective of the AB2588 legislation is to
collect emission data from air toxics sources, identify
facilities with localized impacts, assess health risks and
notify affected individuals
Table 10. CARB - FOI facilities in the San Joaquin Valley found by ground-truthing to have been mislocated by at
least 10 kilometers.
EPAJD
CA2890090002
CAD990794133
CA1 570024504
CA41 7002441 4
CAD980813950
CAT000646117
CAL000190816
CAL000282598
CA21 700231 52
CAD982446882
CAD066113465
CAD981429715
PROJECT NAME
LAWRENCE LIVERMORE
NATIONAL LAB - SITE 300
FORWARD LANDFILL
EDWARDS AIR FORCE BASE
OCCIDENTAL OF ELK HILLS
INC
CRANE'S WASTE OIL INC
CHEMICAL WASTE
MANAGEMENT INC
KETTLEMAN
RIVERBANK OIL TRANSFER,
LLC
BAKERSFIELD TRANSFER INC
NAVAL AIR WEAPONS STATION
CHINA LAKE
EVERGREEN OIL INC FRESNO
SAFETY-KLEEN
KEARNEY-KPF
ADDRESS
CORRAL HOLLOW RD
9999 S AUSTIN RD
5 E POPSON AVE
28590 HIGHWAY 11 9
16095 HIGHWAY 178
KETTLEMAN HILLS
LDFLHWY41
5300 GLAUS RD
1620EBRUNDAGELN
1 ADMINISTRATION
CIR
41 39 N VALENTINE
AVE
3561 S MAPLE AVE
1 624 E ALPINE AVE
CITY
TRACY
STOCKTON
EDWARDS
TUPMAN
WELDON
KETTLEMAN CITY
RIVERBANK
BAKERSFIELD
RIDGECREST
FRESNO
FRESNO
STOCKTON
Error (m)
12,764
11,705
1,519
1,500
614
478
238
231
188
144
115
107
44
-------
EPAJD
CAL0001 02751
CAT0800 10606
CAD982435026
CAT0800 10283
CAD980675276
PROJECT NAME
WORLD OIL - SAN JOAQUIN
LLC
BIG BLUE HILLS PESTICIDE
CONT DISPOSAL
KW PLASTICS OF CALIFORNIA
EPC WESTSIDE DISPOSAL
FACILITY
CLEAN HARBORS
BUTTON WILLOW LLC
ADDRESS
1 4287 E MANN ING AVE
10 MILES NORTH OF
COALINGA
1861 SUNNYSIDECT
26251 HIGHWAY 33
2500 WEST LOKERN
RD
CITY
PARLIER
COAUNGA
BAKERSFIELD
FELLOWS
BUTTON WILLOW
Error (m)
99
76
34
33
21
These facilities were mapped using the 'best-known'
location, either geographic coordinates as reported in a
standard public database, or the geocoded address of the
facility. This data was then taken into the field in the form
of geospatial data layers loaded into ArcMap GIS software,
running on a laptop computer in a vehicle. The laptop was
attached to an external GPS receiver (Garmin Montana 600),
with software allowing the GPS locator to position the cursor
in the ArcMap session so that observer location could be
tracked on the map containing the facility locations in real-
time. With this system configuration, the GPS position could
be used to correct these facility locations or add new features
(i.e., new facilities), as needed.
Each study area was systematically searched in the field by
driving the public roadway network, to locate and validate
facilities. In each case, locational accuracy was verified and
corrected if necessary, as was facility name and whether
it appeared active/inactive or vacant. In some cases, new
facilities were found that were of the same type as those
previously recorded in an agency/regulatory or public
database. These "new" facilities were mapped, as well. For
example, the field researcher would use the road network
to confirm presence and activity of an AB2588 "Hot Spot"
facility or childcare facility, and compare its "real-world"
location to the reported location, correcting the location if
necessary. If similar facilities were found, their locations and
^^^m \ • i —. VI;
Map 23. EJSM total score for San Joaquin Valley prior to ground-truth correction.
13
14
5
6
1
S
9
10
111
112
113
114
115
45
-------
Map 24. Locational error for California DTSC permitted hazardous waste handling facilities and generators located in
the San Joaquin Valley.
attribute information were added to the geospatial data layer.
The results of field-based ground-truthing were revealing and
significant, but differed among the three study areas. These
differences are probably caused by various latent factors
that affect accuracy of the facility databases, and provide
us insight on how to think about and use this type of data in
screening and other analysis of exposure and vulnerability.
An/in:
Arvin is a small city of 19,304 residents located about 15
miles southeast of Bakersfield. Historically, Arvin has long
been a destination for immigrants, including Basques and
"okies" in the early 1900s, Mexicans and Central Americans
in the late 20th century, and several waves of Asian and
European immigrants throughout the 20th century. The
city currently has a very high proportion of Latino residents
(92.7% reporting "Hispanic or Latino of any race" in 2010
Census). In 2007, the EPA listed Arvin as having the highest
levels of air pollution of any community in the United States.
The city's level of ozone exceeded the EPA acceptable limits
an average of 73 days per year between 2004 and 2006.
The study area chosen is the census tract includes the entire
town area, as well as some surrounding agricultural land.
Standard public databases showed this study area to contain
31 facilities of interest. This equates to about one hazardous
facility per square kilometer, although most are concentrated
in the most densely populated "town" portion of the study
area (see Map 16).
• 18 AB2588 "Hot Spots" facilities
1 auto paint/body shop
5 gas stations
• 7 sensitive land uses (schools, childcare, healthcare)
Ground-truth validation showed that all auto facilities and gas
stations were accurately located and active. However five
(5) of the 18 AB2588 facilities are either vacant or inactive
and another 5 were incorrectly located from their nominal
locations by large distances (see Map 17):
Huron:
Huron is a small city located about 15 miles east of
Coalinga, with a population of 6754 (2010 Census). The
population swells to over 15,000 during the harvest season
due to an influx if migrant farm workers who work the
farmlands surrounding Huron, which is primarily devoted
to the production of lettuce, onions and tomatoes. Huron
has the highest proportion of Latino residents of any city
in California (with 96% reporting "Hispanic or Latino of
any race"). Over the past few decades, it has also been
characterized by high levels of poverty, but very low levels of
46
-------
Rn'fRSAKK OIL TfWNSFBt LLC
Map 25. California DTSC permitted hazardous waste handling facilities and generators in San Joaquin County
showing locational error in reported positions. Facility boundaries are shown as violet-colored polygons
• CARB-FOI site as reported
O Corrected location
Map 26. California Air Resources Board "Facilities of Interest" (FOI) locations in the San Joaquin Valley, showing
location corrections.
47
-------
EJSM Hazard Proximity Scores - Arvin
Before Correction After Correction
'
Map 27. EJSM total CI score for San Joaquin Valley prior
to ground-truth correction.
unemployment. Standard public databases showed this study
area to contain 47 facilities of interest. This equates to a very
low relative density of hazardous facilities (see Map 18).
• 31 AB2588 "Hot Spots" facilities
• 5 CARB - FOI facilities
• 7 gas stations
• 4 sensitive land uses (3 schools and one childcare
facility)
Ground-truth validation showed that 13 sites were
significantly mislocated (13 AB2588 and 2 CARB-FOI
facilities) as shown in Map 19 and Table 5. In addition,
seven of the AB2588 facility entries listed/reported in the
facility database are actually associated with only three
facilities, the rest of the entries (4) are duplicate records in the
facility database. Ten new facilities that are not included in
standard public databases were discovered by field validators
(3 gas stations and 7 AB2588 facilities) as shown in Map 20.
All sensitive land uses are present, but all were significantly
mislocated, with the recorded distance errors ranging from
100 meters to 15 kilometers
Stockton:
Stockton is a major city in the San Joaquin Valley. It is the
county seat for San Joaquin County, and is the 13th largest
city in California by population (291,707; 2010 Census).
The economy of Stockton is firmly grounded in Central
Valley agriculture as well as the city's inland seaport, and
an intricate network of canals, waterways and rivers, which
comprise the California Delta. Stockton has also suffered
from very significant economic shocks due to the 2007
subprime mortgage financial crisis, its violent crime rate,
and its public financing and subsequent 2012 bankruptcy
EJSM Hazard Proximity Scores - Huron
Before Correction After Correction
EJSM Hazard Proximity Scores - Stockton
Before Correction After Correction
Map 28. EJSM Hazard Proximity CI Scores (Arvin,
Huron, and Stockton) before and after ground-truth
correction of facility location and status
48
-------
EJSM Total Score - Arvin
Before Correction After Correction
EJSM Total Score - Stockton
Before Correction After Correction
J
EJSM Total Score - Huron
Before Correction After Correction
Map 29. EJSM Total CI Scores (Arvin, Huron, and
Stockton) before and after ground-truth correction of
facility location and status
Table 11. CARB - FOI facilities in the San Joaquin Valley found by ground-truthing to have been mislocated by at least
10 kilometers.
Name
SHELL WESTERN E & P INC.
CHEVRON USAINC
VINTAGE PRODUCTION
CALIFORNIA LLC
SENECA RESOURCES
AERA ENERGY LLC
PHILLIPS 66 PIPELINE LLC
MCKITTRICK LIMITED
BERRY PETROLEUM COMPANY
KAWEAH RIVER ROCK CO.
GRANITE CONSTRUCTION
COMPANY
HILMAR CHEESE COMPANY
CRES INC DBA DINUBA ENERGY
CALIFORNIA CORRECTIONAL INST
H PANOCHE PUMP STATION
Address
P.O. BOX 11164
WEST OF LOST HILLS GAS
PLANT
LIGHT OIL WESTERN
LIGHT OIL WESTERN
MAIN CAMP ROAD
JUNCTION PUMP STATION, 14
4905 REWARD RD, HEAVY OIL
WESTERN
HEAVY OIL WESTERN
P.O. BOX 515
ARVIN
9001 NORTH LANDER AVE
6929 AVENUE 430
PO BOX 1031
SEC. 18-T14S/R/12E
BAKERSFIELD
LOST HILLS
BAKERSFIELD
COALINGA
BAKERSFIELD
BAKERSFIELD
WOODLAKE
BAKERSFIELD
HILMAR
REEDLEY
TEHACHAPI
FRESNO COUNTY
Error m
148,490
79,994
73,874
71,932
66,831
65,786
58,750
48,263
37,558
35,757
35,316
29,644
29,381
26,828
49
-------
Table 12. CARB - FOI facilities in the San Joaquin Valley found by ground-truthing to have been mislocated by at least
10 kilometers (continued).
Name
NAVAL AIR WEAPONS STATION GB
THREE BRAND CATTLE CO
EXXON MOBIL CORPORATION
GOLDEN STATE VINTNERS
LIVE OAK LIMITED
WEST KERN WATER DISTRICT
CHEVRON RIO BRAVO STATION
MACPHERSON OIL COMPANY
NAVAL PETROLEUM RESERVE #1
NAVAL PETROLEUM RESERVE #1
HAZEL H HEUSSER TRUST
CONOCO PHILLIPS PIPE LINE CO.
MERCED POWER, LLC
WASCO HARDFACING
LIVE OAK LIMITED
CALIFORNIA DAIRIES, INC.
CAMBRIAN ENERGY WOODVILLE
ENERGY LLC
GRANITE CONSTRUCTION CO
BADGER CREEK LIMITED
MEADOWLAKE WEST DAIRY
GUSMER ENTERPRISES INC
ATAPCO OFFICE PRODUCTS
GROUP
WEST COAST CHROME
SIERRA SUMMIT
J G BOSWELL COMPANY
BRITZ GIN PARTNERSHIP II
KINGS RIVER COMMODITIES
FOSTER POULTRY FARMS-KOPRO
NEO TULARE LLC/TULARE
COUNTY
SHELL OIL WASCO PUMP STATION
Address
1 ADMINISTRATION CIRCLE
34377 LERDO HWY
18271 HWY. 33
7409WCENTRAL
7001 GRANITE ROAD
HWY 119 & CAAQUEDUCT
ENOS LANE 2 Ml SO OF
STOCKDALE
HEAVY OIL CENTRAL
ELK HILLS FIELD-GAS PLANT
ELK HILLS FIELD-PRDTN
FACILITY
41990 RADIO LN
34960 AMADORAVE
30 W SANDY MUSH ROAD
2660 S EAST
7001 GRANITE RD., HEAVY OIL
755 F ST
WOODVILLE LANDFILL
6950 OLD STAGE RD
535 FANO RD., HEAVY OIL
CENTRAL
6802 AVENUE 120
124MSTREET
2851 E FLORENCE ST
451 SONORAAVE, #J & D
59265 HIGHWAY 168
31500 SOUTH LAKE ROAD
25500 WMT WHITNEY
27498 HIGHWAY 198
12997 W HWY 140
VISALIA LANDFILL
MERCED AND WILDWOOD
CHINA LAKE
BAKERSFIELD
MCKITTRICK
FRESNO
BAKERSFIELD
TAFT
BAKERSFIELD
BAKERSFIELD
TUPMAN
TUPMAN
AUBERRY
COALINGA
ELNIDO
FRESNO
BAKERSFIELD
FRESNO
DUCOR
BAKERSFIELD
TIPTON
FRESNO
FRESNO
MODESTO
LAKESHORE
BAKERSFIELD
FIVE POINTS
LEMOORE
LIVINGSTON
VISALIA
WASCO
Error m
26,690
23,014
22,604
19,897
19,815
19,015
18,618
16,747
16,239
16,239
16,012
15,128
14,609
14,021
13,650
13,235
13,091
12,495
12,337
12,007
11,930
11,638
11,532
11,015
10,915
10,785
10,535
10,365
10,327
10,088
50
-------
of Stockton's city government. The Stockton census
tract study area is the smallest of those receiving ground-
truth validations. Standard public databases showed this
study area to contain 61 facilities of interest, including 56
hazardous facilities (see Map 21).
• 40 AB2588 "Hot Spots" facilities
• 3 CARB - FOI facilities
• 4 Auto paint and body shops
9 gas stations
5 sensitive land uses (3 childcare, and 2 healthcare
facilities)
Ground-truth validation showed that the majority of sites
were very close to the location recorded in standard public
databases, with only two facilities significantly mislocated
(see Map 22). Twelve reported facilities either were
duplicates or not found during field checking (8 AB2588,
1 auto paint/body, and 3 gas stations) as shown in Map 21.
Five facilities have business names different from those in
the standard public database records. Ground-truthing shows
6 total sensitive land uses to be present, four are childcare
facilities, one senior residential facility, and one major
healthcare facility.
Ground-truth validation also addressed location accuracy of
four facility types used in EJ screening for the entire eight-
county SJV region. This was done by comparing the location
reported in the standard public database with information
available with subscription-based Google Earth Pro. This
tool allows geographic searches and validation information
by facility name or address, provides parcel-based geocoding,
and allows verification of locations using high quality aerial
imagery.
These locations were first examined for locational accuracy
using Google Earth Pro (GEP), and all were geocoded to
further check facility location. These included:
California Air Resources Board (CARB) "Facilities of
Interest" (FOI) - consists of a subset of facilities from
the California Emission Inventory Development and
Reporting System (CEIDARS) statewide air toxics
emissions inventory of greatest concern to regulators
because of amounts, toxicity, possible impacts of
emissions
• California Department of Toxic Substances Control
(DTSC) permitted hazardous waste handling facilities
and generators
• Auto paint and body shops form the Dun and Bradstreet
Business Locator Service
Next, each facility location was compared visually with the
aerial imagery and parcel boundaries available in Google
Earth Pro to verify accurate location, or to correct the
location if necessary. These three datasets showed that
significant locational error is a common characteristic of this
type of information.
The California DTSC lists 17 permitted hazardous waste
handling facilities and generators located in the San Joaquin
Valley (see Map 23). Ground-truth validation demonstrated
significant locational error for most of these sites, with most
locations off by well over 100 meters (see Table 6). Table 6.
Locational error for California DTSC permitted hazardous
waste handling facilities and generators located in the San
Joaquin Valley.
Visual comparison also shows that many sites occupy a
large area. This is important to environmental screening
and hazard proximity calculations because using a point
location may not be an adequate way to characterize a
hazardous facility, and can introduce considerable error in the
hazard proximity calculation used in the EJSM. The largest
locational/positional errors were concentrated in San Joaquin
County (see Map 24), although most facilities are located in
Kern County.
Significant locational error was also found in ground-truth
validation of California Air Resources Board "Facilities
of Interest" (FOI). Of the 730 total CARB - FOI facilities
located in the study area, nearly half (n=343 or 47%) were
in error by at least 100 meters, and 151 were in error by at
least one kilometer. The facilities with the highest degree of
locational error (at least 10 km) are listed in Table 7. There
is no clear geographic pattern to the locational errors, but
sites in the sparsely populated regions of west Kern, Kings
and Fresno counties contained many of these very poorly
located facilities (see Map 25) CARB - FOI facilities in the
San Joaquin Valley found by ground-truthing to have been
mislocated by at least 10 kilometers.
We also determined by ground-truthing that 60 of total 317
auto paint and body shops in the study area were improperly
located (mislocated) by at least 100 meters. The positional
locations for 16 of these facilities were in error by more than
a kilometer. These facilities tend to be concentrated in more
densely population areas of the SJV.
As a test of the fundamental goal of this ground-truth
validation work, we rescored the SJV study area with the
EJSM methodology, using the location corrected facility
information to look for differences resulting from using
unchecked (error filled) vs. validated (errors corrected)
information to assess the degree to which CI score metrics
changed both regionally, as well as in the three CI analysis
sites where field work was done. Municipalities and regional
governments do not always have the resources to update
facility databases on a regular schedule, so ground truthing is
required to ensure that correct facility location data is used in
the EJSM to properly assign CI scores.
Comparison of total EJSM CI scores prior to, and after the
ground-truth corrections are shown in the two maps below
(Figures Map 26 and Map 27). The pattern of CI scores on
the two maps is similar, but with some visible differences
between the two regions. In the first map, higher CI scores
are displayed for the very large census tracts located along
Interstate 5, west of Delano. This area is sparsely populated
51
-------
and mostly agricultural, but with substantial oil and gas
production. The other obvious difference is the higher CI
scores for the area surrounding Fresno.
The three ground-truth CI analysis sites where field
observations were used had been originally identified, in part,
because of their high EJSM hazard proximity scores. Map
28 and 29 show comparisons of the EJSM hazard proximity
scores and total CI screening scores for those three CI
analysis sites prior to and after ground-truth correction. If
the EJSM was subject to false positives, these results would
be significantly different. The hazard proximity scores for
both Huron and Stockton are identical, and the total EJSM
CI scores for these two CI analysis sites are nearly the
same. However, both CI scores appear quite different for
Arvin. The explanation for a greater difference in Arvin after
ground-truthing is probably related to two causes. First, a
larger percentage of the hazards recorded in standard public
databases turned out to be vacant or inactive, than was the
case for the other two study CI analysis sites. In addition,
the hazards in the Arvin study CI analysis sites are mostly
concentrated in the southeast portion, where the town of
Arvin is located. That part of the study CI analysis sites
shows no change in either hazard proximity or total score.
52
-------
The completion of this research project has resulted in
a number of accomplishments including: a) successful
application of EJSM, developed in a previous R9 RARE
Research Project, in two pilot communities in California
- the City of Commerce and the SJV (three cities - Arvin,
Huron, and Stockton); b) maturation of the EJA approach as
an objective way to numerically quantify and characterize
the CI of air pollution exposures, psycho-social and health
vulnerabilities, along with land use/hazard proximity
experienced by vulnerable communities; c) successful
implementation of EJAF EJ cumulative risk framework in
the SJV in comparing 3 different CI screening methods (in
the same geographic area) in a systematic way to provide
correlated information on the impact of policy decisions on
vulnerable communities; d) development of an EJ cumulative
risk framework and CI screening method that can serve as an
adjunct to (or an initial input to) local, regional, or statewide
planning, including land use planning/zoning, transportation
planning, or environmental remediation. Although the
EJSM was developed to perform CI screening for California
communities, the tool is not strictly limited to application
in California. Most of the input data sources used by EJSM
are national in scope, and so would be applicable outside
of California by using the local values from those sources.
Other states would need to have the following state-specific
data sources in order to apply EJSM outside of California:
a) vital statistics database (birth outcomes); b) land use
databases; c) hazardous waste/hazard proximity databases;
d) air pollution monitor data/database; e) voter participation
database; f) inhalable cancer risk information/database; g)
state emission inventory; h) 'facility database[s] (e.g., locate
businesses, schools, 'polluters' of interest, etc.). Since a
number of states may have these state-specific information
resources, EJSM would be a tool that could be used in
various locations.
53
-------
-------
Adler, N. K, Rehkopf, D. H., (2008). "U.S. disparities in
health: descriptions, causes, and mechanisms." Annual
Review of Public Health, 29: 235-252
Boer, T. I, Pastor, M, et al., (1997). "Is there environmental
racism? The demographics of hazardous waste in Los
Angeles County." Social Science Quarterly, 78(4): 793-810.
Bullard, R. (1983). "Solid Waste Sites and the Black
Community." Sociological Inquiry 53: 273-288.
Burke, L. (1993). "Race and Environmental Equity: A
Geographic Analysis in Los Angeles." Geo-Info Systems
October: 44-50.
CARB (2005). "Air Quality and Land Use Handbook: A
Community Health Perspective." Available at: http://www.
arb.ca.gov/ch/handbook.pdf; Last Accessed: September 26,
2014
Available at: http://www.arb.ca.gov/ch/handbook.pdf; Last
Accessed: September 16, 2014
Centers for Disease Control and Prevention (2003). Second
National Report on Human Exposure to Environmental
Chemicals. Atlanta, Georgia, US Department of Health
Service, Centers for Disease Control and Prevention:
257 Pages.
Commission for Racial Justice, U. C. o. C. (1987). Toxic
Wastes and Race in the U. S.: A National Report on the
Racial and Socio-economic Characteristics of Communities
with Hazardous Waste Sites, UCC.
DeFur, P. L., Evans, G. W, et al. (2007). "Vulnerability as
a function of individual and group resources in cumulative
risk assessment." Environmental Health Perspectives 115(5):
817-824.
EPA's Risk Screening Environmental Indicators (RSEI)
Methodology, Office of Pollution Prevention and Toxics,
Washington, DC 20460, (RSEI Version 2.3.0), June 2010
Gee, G. C., Payne-Sturges, D. C., (2004). "Environmental
health disparities: a framework integrating psychosocial and
environmental concepts." Environmental Health Perspectives
112(17): 1645-1653.
Gelobter, M. (1986). The Distribution of Air Pollution, By
Income and Race. Energy and Resources Group, University
of California, Berkeley.
Gelobter, M. (1992). Toward a Model of Environmental
Discrimination. Race and the Incidence of Environmental
Hazards: A Time for Discourse. B. Bryant and P. Mohai.
Boulder, CO, Westview Press: 164-235.
Hall, E.S., Morello-Frosch, R.A., Pastor, M., Sadd, J.L.
(2011). "R9 RARE Final Project Report: How Vulnerability
Indicators Affect the Association Between Adverse Birth
Outcomes, Air Pollution and Traffic Burden.", EPA Office of
Research and Development, EPA/600/R-12/014, October 13,
2011, 80 Pages
Hersh, R. (1995). Race and Industrial Hazards: An Historical
Geography of the Pittsburgh Region, 1900-1990. Washington,
DC, Resources for the Future.
Huynh, M., J. Parker, et al. (2005). "Contextual effect of
income inequality on birth outcomes." International Journal
of Epidemiology, 34(4): 888-895.
Krieger, N., Chen, J., et al., (2003). "Choosing area based
socioeconomic measures to monitor social inequalities in
low birth weight and childhood lead poisoning: The Public
Health Disparities Geocoding Project (US)." Journal of
Epidemiology and Community Health, 57(3): 186-199.
Kraft, M., Scheberle, D., (1995). "Environmental Justice and
the Allocation of Risk - The Case of Lead and Public Health."
Policy Studies Journal 23(1): 113-122
LaVeist, T. A. (1992). "The Political Empowerment and
Health Status of African-Americans: Mapping a New
Territory." The American Journal of Sociology 97(4):
1080-1095.
Lavelle, M., Coyle, M., (1992). "Unequal Protection
- The Racial Divide in Environmental Law: A Special
Investigation." The National Law Journal, Volume 15,
Number 3, (7 pages); Copyright © 1992, The New York Law
Publishing Company; Copyright © 1999, The New York Law
Publishing Company
Lazarus, R. (1993). "Pursuing 'Environmental Justice':
The Distributional Effects of Environmental Protection."
Northwestern University Law Review 87(3): 787-845.
Mohai, P., Bryant, B., (1992). Environmental Racism:
Reviewing the Evidence. Race and the Incidence of
Environmental Hazards: A Time for Discourse. B. Bryant
and P. Mohai. Boulder, CO, Westview: 164-175.
Morello-Frosch RA, Pastor M., Porras C, Sadd, J (2002).
"Environmental Justice and Regional Inequality in Southern
California: Implications for Future Research." Environmental
Health Perspectives, 110 (Supplement 2): 149-154.
Morello-Frosch, R. and E. Shenassa (2006). "The
Environmental 'Riskscape' and Social Inequality:
Implications for Explaining Maternal and Child Health
Disparities.", Environmental Health Perspectives 114(8):
1150-1153.
Morello-Frosch, R. A. (2002). "Discrimination and the
political economy of environmental inequality." Environment
and Planning C: Government and Policy 20: 477-496.
55
-------
Morello-Frosch RA, Pastor M, Porras C, Sadd J (2002).
"Environmental Justice and Regional Inequality in
Southern California: Implications for Future Research.",
Environmental Health Perspectives 110 ((Supplement 2).):
149-154.
Morello-Frosch, R., Woodruff, T., et al., (2000). "Air
Toxics and Health Risks in California: The Public Health
Implications of Outdoor Concentrations." Risk Analysis,
20(2): 273-291.
Moses, M., Johnson, E., et al. (1993). "Environmental Equity
and Pesticide Exposure." Toxicology and Industrial Health
9(5): 913-959.
National-Scale Air Toxics Assessment (NATA), Issues in the
1999 Inventory/Changes in Assessment Results: http://www.
epa. gov/ttn/atw/nata 1999/99pdfs/changelistingfeb06 .pdf;
(Table of Changes: Last Accessed: September 16, 2014)
O'Campo, P., Xue, X., et al., (1997). "Neighborhood Risk
Factors for Low Birthweight in Baltimore: A Multilevel
Analysis." American Journal of Epidemiology, 87(7):
1113-1118.
Olshan, A. K, Anderson, L., et al., (2000). "Workshop to
identify critical windows of exposure for children's health:
cancer work group summary." Environmental Health
Perspectives, 108: 595-597.
Osmond, C., Barker, D., (2000). "Fetal, infant, and childhood
growth are predictors of coronary heart disease, diabetes,
and hypertension in adult men and women." Environmental
Health Perspectives 108(Suppl 3): 545-553.
Pastor, M., Sadd J., et al. (2001). "Which came first? Toxic
facilities, minority move-in, and environmental justice."
Journal of Urban Affairs 23(1): 1-21.
Pollock, P., Vittas, M. (1995). "Who Bears the Burden of
Environmental Pollution? Race, Ethnicity, and Environmental
Equity in Florida." Social Science Quarterly, 76(2): 294-310.
Ponce, N. A., Hoggatt, K. J., et al., (2005). "Preterm birth:
the interaction of traffic-related air pollution with economic
hardship in Los Angeles neighborhoods." American Journal
of Epidemiology, 162(2): 140-148.
Pulido, L., Sidawi, S., et al., (1996). "An Archeology of
Environmental Racism in Los Angeles." Urban Geography,
17(5): 419-439.
Sadd, J.L., Morello-Frosch, R.A., Pastor, M., Matsuoka,
M., Prichard, M., Carter, V, (2013). "The Truth, the Whole
Truth, and Nothing but the Ground-Truth: Methods to
Advance Environmental Justice and Researcher-Community
Partnerships", Health Education and Behavior, Dec 16 2013,
41 (3), pp 281-290
Sadd, J. L., Pastor, M., et al., (1999). '"Every Breath You
Take...': The Demographics of Toxic Air Releases in outhern
California." Economic Development Quarterly, 13(2):
107-123..
Sadd, J. L., Pastor M., Morello-Frosch, R., Hall, E. S., Lowe,
D. L., McNaughton, E., (2012). "Workplan and QAPP for
EPA R9 RARE Cooperative Agreement: Partnering with
Environmental Agencies and Communities to Pilot Use
of the Environmental Justice Screening Method (EJSM)
Cumulative Impacts Tool", 37 Pages
Sadd, J. L., Pastor M., Morello-Frosch, R., Scoggins,
J., Jesdale, W M., (2011). "Playing It Safe: Assessing
Cumulative Impact and Social Vulnerability through an
Environmental Justice Screening Method in the South Coast
Air Basin, California."International Journal of Environmental
Research and Public Health, 8: 1441 - 1459.
Wernette, D., Nieves, L., (1991). Minorities and Air Quality
Non-Attainment Areas: A Preliminary Geo-Demographic
Analysis. Socioeconomic Energy and Research Conference,
Baltimore, MD, United States Department of Energy. Wilcox,
A.J., (2001). "On the importance-and the nimportance-of
birthweight". International Journal of Epidemiology 30 (6):
1233 - 1241.
Woodruff, T, Parker J., et al., (2009). "Methodological
issues in studies of air pollution and reproductive health."
Environmental Research, 109(3): 311-320.
Zimmerman, R. (1993). "Social Equity and Environmental
Risk." Risk Analysis 13(6): 649-666.
56
-------
A-l
-------
Table 13. PowerPoint Presentation - Webinar Agenda
3:00- 3:10 p.m.
3:10- 3:20 p.m.
3:20- 3:35 p.m.
3:35-4:10 p.m.
4:10-4:25 p.m.
4:25-4:30 p.m.
Introduction
- Goals of the project
- Goals of Today's webinar
- Project team and partners
Project timeline
Process used to compare the three
screening tools
Preliminary maps that compare the
three screening tools
Feedback on the comparison and
Q&A
Next steps
Debbie Lowe Liang, EPA
Jacquelyn Hayes, EPA
James Sadd, Occidental College
James Sadd, Occidental College
All
Jacquelyn Hayes, EPA
Table 14. PowerPoint Presentation - Webinar Goals
Webinar Goals
Provide an overview of the project goals, project partners and project team.
Discuss the process used to compare the three tools.
Share preliminary comparison maps/tables with project partners.
Keeping in mind the policy relevant questions, receive feedback from project partners on the process used to
make comparisons and the comparison maps
Table 15. PowerPoint Presentation - Project Goals
reject Goals
Compare three cumulative impacts
screen tools.
Enironmental Justice screening
method (EJSM)
Cumulative Environmental
Vulnerabilities Assessment
(CEVA)
California Communities
Enveronmental
Health Screening Tool
(CalEnviroScreen) Version 1.0
Conduct ground truthing at
two location in partnership with
community partners.
• Ground truthing will help
elucidate the strengths and
weaknesses of the methods
with respect to the policy
relevant questions posed by
the project partners.
Develop a report discussing lesson
learned, how the tool can be applied
, and how they can be improved.
• What lessons were learned
from the ground truthing
efforts?
• Do the three screening
methods inform the policy
relevant questions? If so, how?
• How can the methods be
furthered developed to better
inform the policy relevant
questions?
A-2
-------
Table 16. PowerPoint Presentation - Project Team and Partners
'reject Team and Partners
Community Partners Agency Partners
SJV CHIP members
CVAQ members
CEJA members
Project Team
Eric Hall, EPA
Jacquelyn Hayes, EPA
Charles Swanson, EPA
Debbie Lowe Liang, EPA
James Sadd, Occidental College
Rachel Morello Frosch, DC Berkeley
Manuel Pastor, USC
OPR - Debbie Davis
DISC - Brian Johnson, Ignacio Dominguez
Cal/EPA-Arsenic Mataka, Malinda Dumisani
ARB - Alvaro Alvarado
SIV Air District - David Lighthall
Research Partners
UC Davis - Jonathan London, Tara Zagofsky
OEHHA-John Faust, Laura August, Shankar Prasad
Table 17. PowerPoint Presentation - Project Timeline
'reject Timeline
Step 1 - Review EJSM, CEVA, CalEniroScreen information
• Provide basic information about the EJSM, CEVA, and CalEnviroScreen
to project partners
• Status: Completed in November 2012
Step 2 - Develop and Share Preliminary Comparison Maps
• Develop preliminary comparison maps/tables. Share with project partners.
• Status: Will share with project partners via webinars on June 26 and July 11, 2013
Step 3 - Provide Feedback on Methods
• Project partners provide feedback detailing the ways that the results from each
method were useful or less useful in informing their policy questions. Project partners
provide feedback on potential improvements to data choices, metrics, analysis, and
scoring methodology, and/or the form of presentation will be included.
• Status: Please privide feedback to Debbie and Jackie by July 26, 2013.
Step 4 - Discuss Feedback, Revise Methods
• The research teams will consider the suggestions from the project partners and
identify possible ways the methods could be revised. A conference call or webinar
will be held to discuss the comments and possible revisions. The EJSM maps will be
revised, if necessary, and provided to the project partners. OEHHA and CEVA maps
will only be revised if funding and timing allows.
• Status: Aim to complete within two months of Step 3
A-3
-------
Table 18. PowerPoint Presentation - Project Timeline (continued)
Step 5 - Disuss Potential Ground Truthing Locations
• Project partners discuss their observations, questions and comcerns about the three
methods and associated maps. Project partners suggest one or two locations for
ground truthing that would help provide a better understanding of the strengths and
weaknesses of each method with regards to answering the policy questions.
• Status: Please provide suggestions to Debbie and Jackie by July 26, 2013.
Step 6 - Select Ground Truthing Location(s)
• Select the location(s) for ground truthing activities based on the feedback from project
partners and community capacity to engage in ground truthing.
• Status: Aim to complete within one month of Step 5
Step 7 - Conduct Ground Truthing
• Inpartnership with SJV CHIP and community partners, conduct community ground
truthing at the selected location(s).
• Status: Aim to complete within three months of Step 6
Step 8 - Develop Report on Lessons Learned, Use of Screening Tools, and Suggested
Improvements
• Discuss lessons learned from ground truthing efforts.
• Project partners work with the project team to develop case studies that demonstrate
how the tools were useful in answering policy questions. For questions where the
tools are inadequate, the project partners will provide suggestions for improving
the tools.
• Status: Aim to complete with ing two months of Step 7
Table 19. PowerPoint Presentation - Title
Table 21. PowerPoint Presentation -Abbreviations
Comparison of EJ Screening Methods
CEVA, CES AND EJSM
Data and Metrics, Mapping and Scores
San Joaquin Valley
Table 20. PowerPoint Presentation - Summery
of Presentation
Summary of Presentation
Background on the three screening methods
How comparison was done
Break for Q/A
Comparison results
- Distribution of scores by population and area
- Mapping the extremes ("top quantile")
- Mapping the extremes more broadly denned ("high")
- Areas where all three methods agree
- Mapping high hazard/pollution exposure scores
Abbreviations
EJSM - Environmental Justice Screening Method
- developed by James Sadd (Occidental College),
Manuel Pastor (USC), and Rachel Morello-Frosch
(UC Berkeley) under CARB, CEC and CalEPA
research contracts.
CES - California Communities Environmental Health
Screening Tool - developed by Cal/EPA Office of
Environmental Health Hazard Assessment (OEEHA)
to identify disproportionately burdened communities
CEVA - Cumulative Environmental Vulnerability
Assessment; developed by Jonathan London and
colleagues at the Center for Regional Change at
UC Davis
SJV - San Joaquin Valley Stady area (Kern, Kings,
Tulare, Fresno, Madera, Merced, Stanislaus,
San Joaquin)
ZCTA - Zip Code Tabulation Area from the US
Census (used by CES)
A-4
-------
Table 22. PowerPoint Presentation - EJSM
EJSM
Regionally scored at
tract level; tax parcels
used to calculate
hazard proximity and
land use
Tracts do not receive a
score if population is
very small or too few
variables are available
(show in white)
Final maps maskout
land uses that are not
classed as residential
or "sensitive" by CARB
(not shown)
Table 23. PowerPoint Presentation - CEVA
CEVA
Regional scoring of
block groups
All block groups in
the region are given
a score
~'^B
A-5
-------
Table 24. PowerPoint Presentation - CES
CES
Scores calculated at
the ZCTA level, for
entire state
Incomplete Coverage -
some ZCTAs unscpred
(shown in white)
Scored areas are very
large compared to
census tracts or blocks
= lack of granularity
Because of statewide
scoring, SJV has a
disproportionate
percentage of high CES
scores compared to
other regions of the
State
Table 25. PowerPoint Presentation - Difference Among Methods
Difference Among Methods
Different "base maps"
- Each methods uses a different spatial unit for both analysis and mapping.
° Tax parcels/tracts (EJSM)
° Census block groups (CEVA)
° ZCTAs (CES)
Each method also has a different way to merge their "base map" spatial units with information reported at the census
tract level.
- Different levels of aggregation and spatial resolution results in map pattern differences controlled by the method,
not the data.
- Distributional effects: different spatial units represent different populations.
The methods use many of the same datasets, but with different metrics to represent a specific indicator.
- NATA: estimated cancer risk, respiratory hazard, diesel PM2 5 estimates
- RSEI hazard-weighted emissions vs. TRI site location
- % poverty vs. 2X% federal poverty level
How scores are calsulated
- Different ways indicators are grouped together for scoring results in different implicit "weighting" of certain
metrics.
- Different range of scores among methods
° EJSM: linear ranks, open-ended to accommodate additional indicators (3-15)
° CES: continuous linear, with scores grouped by percentile (1-20)
° CEVA: two-dimensional (3X3) scoring matrix with separate axes for impact and vulnerability (1-9)
A-6
-------
Table 26. PowerPoint Presentation - Comparison Table of Indicators
Comparison Table
of Indicators
• EJSM is built on a
base map of land use
• CESandCEVAuse
Census polygons
INDICATOR
Seiislthr Laud l srs Populations
Clnldcaie facilities
Healthcare facilities
Schools
I'lb.iii Pniks Pla\ mounds
Senioi Residential
Hazard ProNlmitv
("luomephitins lacthtie-
Au IONICS facilities ' AB25X8)
Haz.iidoiis ,,.||,1 -.vaste f.icihties. cleanup sites
Railroad facilities
Foil facilities
Retiuei ies
Intel modal Di-tiil>u!iun
Tiaiticexposiiie
TRI tacilitie*
Health Risk and Exposure
Toxic i daises - TRJ RSEI
National Air Toxics Assessment I respiratory hazard
cancel n--k. die^elPMi
PM, it ARBiuoiutoisi
i.»;oiie 'CARS moniU'iM
['CstKHlcllse
\Vaiei iKilliition >mi[>.mcil \v;ircr Inulics. sioiuidu'.iici i
Vnlnei abilin
Race etlmicirv
Povem level
Eiliiciiiioiial .itinimnciii
Aae cvotina chihlieu and eUleilvi
Linguistic isolation
"oRenteis
Median house value
Voter participation
Bnth outcomes
Astluua ho>i>i!:ilizarK'ii
LJSM
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
CEVA
X
X
X
X
X
X
i
X
X
X
X
X
X .-
GES
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Table 27. PowerPoint Presentation - How comparison was done
How the comparison was done
1. Comparison of distribution of scores
- By population
- By area
EJSM Scores - San Joaquin Valley
Distribution of total scores by area and population
A-7
-------
Table 28. PowerPoint Presentation - How comparison was done
How the comparison was done
2. Comparing map location of "hot spots"
- Define "hot spot" for each method
- "extreme" vs. a broader definition of "hot spot"
CESGraupKJ
CEVAncoteS
EJSMscores Hand 15
OwrtapCEVAondEJSM
Example: Fresno area
Table 29. PowerPoint Presentation - How comparison was done
How the comparison was done
3. Identify overlap by all three methods
- Locations where all three methods agree
Example: Fresno area
i jT • . -L
i , —
A-8
-------
Table 30. PowerPoint Presentation - Q/A Process used
Q/A on Process used to compare the
three screening tools
Table 31. PowerPoint Presentation - Preliminary Result
Preliminary Results
Comparison of the tools
Table 33. PowerPoint Presentation - Distribution
of scores
Distribution of scores
How does the pattern of high scores compare
between these three methods ?
- By area
- By population
How does the pattern of hazard and pollution
proximity/exposure compare among the methods?
Table 32. PowerPoint Presentation - EJSM Scores
EJSM Scores - San Joaquin Valley
Distribution of total scores by area and population
25.000
20,000
15.000
10.000
5,000
• •
• Population
500.000
450,000
400.000
23203 10295 5811 8243 6463 2834 4171 2634 5484 1466 248
181
50
39547 126262 230628 290873 359131 298139 406567 378511 470204 400750 255481 260607 205052 69781
A-9
-------
Table 34. PowerPoint Presentation - CES SJV Distribution of Group by area and population
CES SJV
Distribution of Group by area and population
Regional Scoring
6,000
5,000
II
150,000
100,000
50,000
I
1) 14 15 16 17 18 19 20
Table 35. PowerPoint Presentation - CEVA Scores - San Joaquin Valley
I
!
25,000
20,000
15,000
10.000
5,000
CEVA Scores - San Joaquin Valley
Distribution of Total Score by area and population
12345
• Area 25792 23127 899 3482 15653
• Population 291017 11625% 325)86 84579 746720
6
1515
232036
7
35
479
563780
] .UKI.IIIIK
1,200,000
] i >
-------
Table 36. PowerPoint Presentation - Summary - Distribution of scores
Summary - Distribution of Scores
Scores vs. Area
- All three methods characterize the majority of SJV area with low scores.
° EJSM - high scores account for a very small portion of the SJV; smooth pattern of variation in area vs. score.
° CEVA pattern is similiar, but far more irregular with a distinct "peak" in middle of scoring range.
° CES does not show the same pattern;
• irregular, with several "peaks" throughout the range of scores
Scores vs. Population
° EJSM -most of the population has middle range scores; pattern nearly approximates a "bell curve."
° CES shows pattern of increasing population with higher scores.
° CEVA has a very irregular distribution.
We feel the "bell curve" distribution makes sense.
- The low population density and highly clustered nature of population in SJV suggests this distribution.
- A screening method that indentifies extremes ("hot spots") should place most areas in the middle of the range.
- The EJSM has this distribution in other areas of the state where it has been applied,
Table 37. PowerPoint Presentation - Location of "hot spots"
Location of "hot spots"
What range of scores define a "hot spot" in each method?
Distribution of scores was examined in terms of population represented by the "high score" class for each method.
- This was complicated by differences in scoring method.
Two comparisons:
- Very highest scores ("top quantile")
° Represents about one tenth of SJV population (7.2 -11.24%)
- More broad definition ("high quantile")
° About one fifth of SJV population (20.7 - 24.6%)
A-ll
-------
Table 38. PowerPoint Presentation - Screening Scores by Aggregate Population " Top Quantile''
Screening Scores by
Aggregate Population
"Top Quantile"
• Selected the highest scores for
the three methods that
represent a similar portion of
total SJV residential population
- EJSM scores 14 and 15
• 7. 2% of population
- CEVA score 9
• 9.7% of population
- CES Group 20
• highest5%ofZCTAsinSFV
• 11. 24% of population
EJSM Score
n
10
11
12
n
14
15
CEVA Scor*
1
2
4
i
'
7
6
8
9
CES Group
1
2
•
4
:
•
'
-
•-
10
U
H
is
16
17
18
19
:
Population
,26:262
230.628
290,873
359.131
298,139
406,567
378,511
470.204
400,750
255.481
260,607
205.052
69.781
291.017
1.102,596
3»:a£
746,720
17,945
232,036
563,780
36933*
11 i '
10.492
lliTW
37,979
205.232
238,842
209,194
125,152
207.948
Iw'ssf
278,987
363.760
212,803
311.356
194.139
327,951
380,283
4?0 ~! 1 '.
Cumulative Pop
3,791.533
3,751.986
3.625,724
3,395,096
3,104.223
2,745.092
2,446,953
2,040,386
1.661,875
1,191,671
790.921
535,440
274,833
69,781
3,793,397
3.502.380
2.339.784
2,255.205
1,929.819
1.183,099
1,165,154
933,118
369.338
4,009,611
3.998.474
3,987,982
3,961,273
3,847,485
3.809.506
3,604.274
3,365,432
3,156,238
3.031.086
2,823.138
2,693.720
2,524,135
2.245.148
1,881.388
1,664.442
1,353,086
1,158,947
830.996
Cumulative -°.i Pop
1
99.0%
95.6%
89.5%
81.9%
72.4%
645%
W».J7*
59.*%
43.8%
31.4%
20.9%
14 1*»
7.2%
1.8%
lim.i
61.7%
59.5%
50.9%
31.2%
30.7%
246'.
1IH.MI..
99.72%
9946%
9879%
95.96%
95.01%
89.89%
83.93%
78.72%
75.60%
67.18%
62.95%
55.99%
46.92%
41.51%
3,3.75%
28.90%
37 10,73%
] i :-'•
Table 39. PowerPoint Presentation - Map 6a "High Decile" - Kern Co.
Map 6a "High Decile" - Kern Co.
All three methods identify greater Bakersfield
— CES characteriigtpn of Bakersfield is unchanged from the Top OerJJe example
- CEVA and EJSM both extend identified area in north Bakersfield, and highlight Lamont and Arvin. as well
they are in general agreement
- EJSM and CEVA agree on North Bakersfield, Wasco (see overlap color)
CEVA now agrees with E1SM in identifying Delano, Oildale, Lamont, Arvin
EJSM now agrees with CEVA in indentifyingTaft
CEVA again solely identifies some small towns - Lost Hills, Wasco,Tfe.ha_c.Mei
CES score 9
CEVA score 9
EJSM scores 14 and 15
Overlap CEVA and EJSM
A-12
-------
Table 40. PowerPoint Presentation - Map 6b "High Decile" - Central SJV
Map 6b "High Decile" - Central SJV
ES expands identification
fthe entire Hwy99
Drridor, but still does not
ick up north Fresno or any
reas along Interstate 5
he problem of large
=nsus polygon size making
lose identified by CEVA
nd EJSM appear more
rominenton the map is
i/en more pronounced.
ISM and CEVA generally
*ree, but a few locations
re only identified by one
lethod
- CEVA: Cutler, Wood lake, and
Coalinga as before; also
Woodville, Ivanhoe, London
- EJSM: large, sparsely
populated areas along 15,
Parley, Hanford, Kerman
CES SCOTS 9
CEVA score 9
EJSM scores 14 and 15
oven-39 CEVA and EJSM
Table 41. PowerPoint Presentation - Map 6c
Map 6c "High Decile" - Northern SJV
.
As before, CES identifies
large population centers
also identified by CEVA and
EJSM, but they do so with
much greater specificity
CEVA again solely identifies
several small towns not
identified by EJSM -Tracy,
greater Manteca, Ripon,
Oakdale, Patterson,
Livingston, Atwater, Gustine,
Patterson.
EJSM identifies fewer areas,
and more geographically
focused areas, than both
other screening methods
CES scores
CEVA score 9
EJSM scores 14 arid 16
Overlap CEVA and EJSM
A-13
-------
Table 42. PowerPoint Presentation Summary - High Quantile
Summary - High Quantile
All three methods again agrain agree on Bakersfield and Fresno/Selma
- The broader definition of "high" scores results in EJSM and CEVA highlighting additional portions of ZCTAs
identified by CES
All three methods also agree on Stockton, Modesto, Tulare and Madera
EJSM and CEVA identify numerous small population centers and towns, but usually not the same ones
Table 43. PowerPoint Presentation - Areas Identifies by all Three Methods
Areas Identified by all Three Methods
All methods agree: Top Population Quantile in RED
All methods agree: High Population Quantile in Pink
Example from Bakersfield, CA
A-14
-------
Table 44. PowerPoint Presentation - Areas Identifies by all Three Methods
Areas identified by all three methods
— All methods agree: Top Population Quantile In RED
— All methods agree: High Population Quantile in Pink
Table 45. PowerPoint Presentation Areas - Identifies by all Three Methods
Areas identified by all three methods
(zoom in maps on next slide)
Top High
Fresno Tula re
Selma
A-15
-------
Table 46. PowerPoint Presentation - Areas Identifies by all Three Methods
Areas identified by all three methods
- Top Population Quantile in RED
— High Population Quantile in Pink
Table 47. PowerPoint Presentation - Areas Identifies by all Three Methods
Areas identified by all
three methods
A-16
-------
Table 48. PowerPoint Presentation - Areas Another way to examine areas identified by all three methods
Another way to
examine areas
identified by all
three methods:
High Hazard/Pollution
Exposure Scores
All three methods assess
hazards/pollution exposure
separately from social
vulnerability
• Because ground truthing will only
evaluate hazard location and pollution
exposure, another way to compare
methods is to only use this part of the
scoring.
Areas shown in purple are where all
three methods agree using only hazard
and pollution exposure score.
• Zoom in maps follow...
m
Arvin
,
x
o
«
Table 49. PowerPoint Presentation - Areas Northern SJV, Stockton Area
Northern SJV
Stockton Area
High Hazard - all methods
I High Population Quantile
Overlap
A-17
-------
Table 50. PowerPoint Presentation - Northern SJV, Modesto Area
Northern SJV
Modesto Area
High Hazard - all methods
High Population Quantile
Overlap
Table 51. PowerPoint Presentation - Central SJV, Fresno
Central SJV
Fresno
High Hazard - all methods
High Population Quantile
Overlap
A-18
-------
Table 52. PowerPoint Presentation - Central SJV, Tulare Area
Central SJV
Tulare Area
High H»»rd • ill rr.ithodi
Hi(h Population Quintlll
OnrUp
Table 53. PowerPoint Presentation - Southern SJV, Bakersfield
High Hazard - all methods
High Population Quantile
Overlap
Southern SJV
Bakersfield
A-19
-------
-------
APPENDIX B:
City of Commerce Planning Commission
Green Zone Policy Report
B-l
-------
TO:
FROM:
DATE:
CASE NO.:
Planning Commission
Public Works and Development Services Department
October 23, 2013
Green Zones Working Group
APPLICANT REQUEST:
Receive and file a report updating the Planning Commission on the work of the
Green Zones Working Group.
LOCATION:
APPLICANT:
Citywide
Commerce, CA 90040
City of Commerce
2535 Commerce Way
Commerce, CA 90040
ATTACHMENTS: 1) Green Zones Working Group Report
2) Documents Reviewed By Green Zones Working Group
The Green Zones Working Group was initiated because of concerns raised by the City's
Environmental Justice Advisory Task Force to the Commerce City Council regarding
proximity of hazardous sources to sensitive land uses such as homes, schools and
churches. In June 2011 the Commerce City Council directed city staff to convene a
workshop between the City Council, Planning Commission, Environmental Justice
Advisory Task Force, and Commerce Industrial Council Chamber of Commerce to
"discuss land use recommendations on Buffer Zones/Sensitive Receptors and Green
Zones". The workshop was intended to allow participants to discuss innovative
approaches to create separation of hazardous sources and sensitive uses as well as
economic development strategies with a focus on "green" practices and objectives". This
convening would be a work session between representatives from these four groups and
would not require the existing formal structure applied to City of Commerce commissions
-------
Planning Commission Staff Report
Green Zones Working Group
October 23,2013
Page 2
and task forces. Resources secured through partnerships between the Commerce
Environmental Task Force and organizations such as the University of California and the
United States Environmental Protection Agency would be used to inform and facilitate the
process. The Commerce City Council further directed staff to work with the Commerce
Environmental Task Force to further refine the approach and process to facilitate the
workshop(s).
Upon review of the scale and scope of work that would be addressed in the
workshop(s)the Commerce Environmental Justice Advisory Task Force and city staff
determined that a more comprehensive process would be required. The process would
now include a series of meetings between participants that would make up the Commerce
Green Zones Working Group. A consultant would be retained to facilitate the meetings
through resources provided by the United States Environmental Protection Agency. In
February 2012 city staff presented this approach to the Commerce City Council The City
Council directed staff to convene a meeting of an initial group of stakeholders that
included representatives from the Environmental Justice Advisory Task Force, Commerce
Planning Commission, Commerce Industrial Council Chamber of Commerce Board and
membership, and University of Southern California Program for Environmental and
Regional Equity (PERE). In July 2012 the process was launched with an initial meeting.
The consulting firm of MIG was hired to assist the group, with staff member Esmeralda
Garcia facilitating the Group meetings. As directed by the City Council, at its first meeting
the Working Group identified additional stakeholders to involve in the process. The final
composition of the Working Group included representatives from the residential and
business community, advocacy organizations and technical experts. These include:
* Business Community - Small Business
» Business Community - Large Business
« Commerce Industrial Council Chamber of Commerce
« Environmental Justice Advisory Task Force
» Labor/Jobs
• Commerce Planning Commission
« Commerce Residential Community
» USC PERE
**Please note that the majority of information provided in this report was taken from a
larger report which is attached to this document**
Overview of Topics Discussed
Over the course of 14 months, the Commerce Green Zones Working Group collectively
dedicated many hours to reviewing data and discussing technical information related to
economic factors, regulatory tools, and policy for the City of Commerce. The purpose for
this review was to establish a draft policy framework that will guide the Commerce City
Council in establishing land use planning policies and processes that address the
proximity of hazardous sources and sensitive land uses while encouraging green
economic development.
-------
Planning Commission Staff Report
Green Zones Working Group
October 23,2013
PageS
The Green Zones Working Group defined a set of guiding principles that provided
direction the design of recommendations and crafting of policy options:
1. Balance -The City of Commerce has a large industrial base with a relatively small
residential population and area. Seek to maintain a balance between the needs
and quality of life of the residential community while still fostering economic
development.
2. Community Health -Support practices that enhance the quality of life and health
for residents and the local workforce.
3. Uses that pose a harmful threat to health in close proximity to sensitive receptors -
Promote an environment that safeguards the wellbeing of sensitive land uses and
viability of businesses through careful planning and phasing of future improvement
activities guided by applicable laws and regulations (i.e. 2005 CARB Air Quality
and Land Use Handbook).
4. "Green" Practices - Incorporate "green" practices of urban planning and
development when formulating recommendations on land use and policy
directions.
5. Image - Advocate the City as business-friendly with strong community amenities,
6. Infrastructure - Ensure adequate infrastructure to meet the current and future
needs of the community and business.
7. Local Workforce - Support activities that improve workforce opportunities for
local residents in light of new industry clusters locating within the City
8. Connectivity/Mobility - Enhance the existing transportation system to consider
all transit modes, capitalize on existing successful transit, and improve
opportunities for the Commerce community and workforce.
Issues and Opportunities
Analysis of challenges/barriers and opportunities relevant to the anticipated policy
recommendations also framed this process. They include:
1. Although the City has a successful tax base generated by local business,
Commerce is relatively small compared to other cities in the region, resulting in
certain limitations. For example, the City relies heavily on outside contracting for
services.
2. The City's existing permitting process sometimes poses challenges for new and
existing business. The City is knowledgeable of numerous resources that can
improve the process. However, due to lack of staff and other City resources, they
have not been implemented.
3. The City is in some respects relatively isolated from some resources that would
ease or enhance progress toward Working Group goals. For example, utility
companies and other permit issuing agencies do not have offices close to the City
of Commerce, so agency staff is less accessible, and the City has no control over
utility rates and permitting processes.
4. The City of Commerce has within its boundaries existing rail yards and two (2)
major State highways, both land uses considered to be a major source of air
quality concerns. However, the City does not have jurisdictional oversight or
control over either.
5. Existing commercial buildings need redevelopment to improve both air quality and
economic vitality.
-------
Planning Commission Staff Report
Green Zones Working Group
October 23, 2013
Page 4
6. Adaptive re-use might be useful in transition areas to address problems of
proximity of harmful land uses near sensitive receptors, as well as to attract new
businesses.
7. Attracting artisanal and cottage industry as new business clusters provides a
unique opportunity to encourage entertainment and other attractions to provide a
greater sense of place for residents and as a means of institutionalizing improved
amenities. Beautification projects are also a means to reinforce a stronger
community image.
General Areas of Agreement
As early in its formation, and as early as its second working group meeting on October 3,
2012, the Green Zones Working Group discussed, confirmed, and ultimately established
the key decision-making process that included building consensus during the process and
required a fifty-percent (50%) plus one (1) requirement for decision-making on final
recommendations transmitted to the Commerce City Council.
Other methods for consensus building, included providing the post-meeting summary by
email to all Working Group members following the meeting for their review. This allowed
group members who could not attend the meeting to provide feedback on direction,
perspectives, ideas or concurrence presented by the Working Group. Comments were
provided to MIG via email and documented in a revised meeting summary. Meeting
summaries were reviewed during the following meeting. All Working Group meeting
summaries are included as appendices to this report.
The Working Group reached numerous points of agreement before making its final
recommendations.
» Establish a community identity for the City of Commerce
» Place special focus on Atlantic Boulevard and Washington Boulevard corridors
« Develop a City of Commerce marketing strategy to attract new business
» Incentives and business attraction are key to creating a new green economy
• The Working Group will use a framework to develop its recommendations:
o Prevention - provide separation of harmful uses from sensitive receptors.
o Reduction - apply methods to reduce pollution from businesses and
industry
o Revitalization - pursue opportunities that could contribute to a "green
economy"
o Reinvestment - provide infrastructure to support economic growth and
protection of community health
• Any recommendation on uses includes siting of any "new" uses
« Definition for sensitive receptors from the CARS Air Quality and Land Use
Handbook (Residences, schools, childcare and daycare centers, urban parks and
playgrounds, or medical facilities, senior residential facilities.)
» Exclude freeways and high traffic roads from land use discussion
Areas of Dissent
Potential recommendations affecting land use policy require careful review and thoughtful
discussion. The Green Zones Working Group dedicated many hours to review topics and
-------
Planning Commission Staff Report
Green Zones Working Group
October 23,2013
Page 5
considerations that would be included in a draft land use policy framework. This task
required that the Group, with representation from different stakeholder groups with
differing perspectives and opinions, arrive at recommendations agreed to by more than
half of the group. While the majority of the strategies described in the Recommendations
Matrix represent agreement from the group based on the decision-making process
developed by the group, there were some areas where there was dissention.
Zoning
As stated in the Overview of Topics Discussed item #3, there was agreement by the group
that the issue of uses that pose a harmful threat to health in close proximity to sensitive
receptors is extremely important and should be a priority. However, the Working Group
had divergent opinions about the tools that it should recommend to the City Council to
address this issue. Over the course of several meetings the Working Group discussed
changes to the existing City of Commerce Zoning Code. As directed by the City Council in
its motion to proceed with the Working Group process with additional resources provided
by community partners, the Working Group leveraged resources made available through
East Yards Communities for Environmental Justice to review proposed draft changes to
the existing Zoning Code. After careful review of these proposed changes to the Zoning
Code by all stakeholders represented on the Working Group the participants considered
trade-offs and implications of implementation. During this phase of the process the
Working Group's discussion informed other strategy recommendations. However, there
were some areas that the Group continued to deliberate.
Those in agreement that revising the city's Zoning Code would be a good tool to address
the issue of proximity between hazardous source and sensitive uses primarily believed
that the proposed changes would provide greater certainty for business and the
community that which/certain uses are permitted near sensitive receptors. Those with the
perspective that the existing Zoning Code should not be revised believed that existing
Federal and State regulations and the City's Zoning Ordinance include policies and
regulations that address provide issues resulting from uses that pose a harmful threat to
health in close proximity to sensitive receptors. The focus should not be on creating more
regulation but on identifying and attracting business that will not pose future hazards to
the community.
At the last meeting of the Commerce Working Group, four recommendations were
proposed to address the issue of zoning.
« Update the City's Zoning Code to prevent the intrusion of sensitive land uses into
industrial areas and prevent intrusion of new harmful uses into sensitive uses. Use
the proposed language developed for the Commerce Green Zones Working
Group.
• Develop a Specific Plan in designated area (or areas) to prevent the intrusion of
sensitive land uses into industrial areas.
» Develop a Specific Plan in designated area (or areas) to prevent the intrusion of
sensitive land uses into industrial areas and industrial uses into sensitive uses.
• Do not recommend any of the proposed zoning tools listed above. There are
potentially other tools beyond those listed that the Group did not discuss.
-------
Planning Commission Staff Report
Green Zones Working Group
October 23,2013
Page 6
Since there was dissention on these recommendations the Working Group agreed to
designate a preference for each of the proposals. Each designated stakeholder
representative voted on its preference for each of the alternatives. The results of the vote
are noted in the chart below.
Alternative
Update
Zoning Code
Develop
Specific Plan
vl
Develop
Specific Plan
v2
None
EJTask
Force
Yes
Planning
Commission
Yes
Industrial
Council
Yes
use
PERE
Yes
Resident
Yes
Small
Business
Yes
Large
Business
Yes
Jobs
Labor
Yes
USt?
'
In November of 2013, the City Council will be reviewing the subject matter.
£WFF RECOMMENDATION:
Staff recommends that the Planning Commission 1) Receive and file the subject report.
Prepared by:
Matt Marquez
City Planner
Reviewed by;
Alex Hamilton
Assistant Director of Development Services
Reviewed by:
Eduardo Olivo
City Attorney
-------
-------
C-l
-------
Response to Comments - RARE Project Comparing Three Screening Methods in the San Joaquin Valley
November 15, 2013
This document provides responses to comments received by project partners for the EPA RARE project comparing three cumulative impacts screening methods in
the San Joaquin Valley. Comparison maps were provided to project partners, and they were asked to provide feedback on the following:
1) The process used to make comparisons and the comparison maps;
2) Whether the results from each method are useful or less useful in informing/answering your policy questions;
3) Any potential improvements to tools' data choices, metrics, analysis, and scoring methodology; and
4) Potential locations for ground truthing that would help provide a better understanding of the strengths and weaknesses of each method with regards to
answering the policy questions.
General Response to Comments
We appreciate the time project partners have taken to provide guidance and feedback on this research project. This project has evolved based on feedback we
received early on in the project. Originally, the intent of this project was to apply the Environmental Justice Screening Method (EJSM) to the San Joaquin Valley.
After considering feedback from our partners, we decided to broaden the scope of this project to examine how existing screening tools can help inform
stakeholders' questions about cumulative impacts or environmental justice in the San Joaquin Valley. The purpose of the comparison is not to identify which tool or
methodology is "better." Rather, the purpose is to identify the strengths of the tools in hopes of informing the development of new or refinement of existing
environmental justice or cumulative impacts screening tools. The main objective of the project is to develop case studies demonstrating how existing screening
tools can answer policy relevant questions.
Screening tools cannot fully measure or capture all the burdens and vulnerabilities of communities. Field validation (commonly referred to as ground truthing) can
help us better understand the limitations of the tools and/or the datasets that support them. For example, field validation can help verify the location of sensitive
receptors or facilities, and can also help identify places that communities are concerned about but are not identified by the screening tools. Community partners
will be heavily involved in field validation for this project.
Specific Comments and Responses
Commenter: Jonathan London, DC Davis Center for Regional Change
Verbatim Comments with Responses Italicized:
Thank you for this opportunity to contribute to the US EPA/ Region 9 RARE pilot program on environmental justice mapping approaches. The DC Davis
Center for Regional Change is pleased to have our Cumulative Environmental Vulnerability Assessment (CEVA) included in this project. I am very
encouraged to see US EPA engaging in enhancing the development and application of socio-spatial environmental justice analysis in such a thoughtful
and inclusionary way. I offer these comments in the interests of mutual learning and on-going improvements of the field.
Before addressing the specific questions posed by the RARE team, it is important to make several observations about the intended applications of the
web-based platform. While the two websites provide a useful means to visually compare the three methods, there is little guidance to a visitor, not
Page 1 of 12
-------
Response to Comments - RARE Project Comparing Three Screening Methods in the San Joaquin Valley
November 15, 2013
Specific Comments and Responses
associated with the RARE project, about how to use the site or why they might chose one tool or another. Since a tool is only "good" to the extent that it
does the job intended, without specifying this job, there is little basis to evaluate the relative merits of the tools in question. This makes it difficult to
comment on the RARE review questions "whether the results from each method are useful or less useful in informing/answering your policy questions." I
would imagine that any of the tools could be useful for identifying the people and places confronting the highest degrees of environmental hazards with the
lowest level of social, economic and political resources to avoid, mitigate, or adapt to these hazards. In general, users looking for a state-wide comparison
would clearly gravitate to the CES. Those with a very broad range of target issues would be best served by the EJSM, which has an extensive set of
indicators. Users interested specifically in the San Joaquin Valley and committed to using a tool developed with extensive community engagement would
likely benefit most from the CEVA.
The original intention of this project was to apply the EJSM in the San Joaquin Valley. After hearing from project partners in 2012, however, we
decided to broaden the scope of the project even though the research contract was already put in place. Currently, the intent of this project is to
elucidate how currently available screening tools can help inform stakeholder questions about cumulative impacts in the San Joaquin Valley.
Because different stakeholders may have different questions or concerns, the research project looks to the project partners, who represent various
stakeholder groups in the Valley, to provide policy relevant questions that they hope screening tools in general can answer. If any of the tools, as
they exist, are able to inform the policy questions, we will ask the project partner to work with the research team to develop a case study of how
the tool(s) informed the policy relevant question(s). For some of the policy relevant questions, the project partners might find that these tools are
not sufficient, as they currently exist, to inform the policy questions. In this case, we would ask the project partners for specifics about how the
tools could be improved to be useful for their policy relevant purposes. The overall outcome is to inform the development of new or refinement of
existing environmental justice or cumulative impacts screening tools. Funding for this project can only support improvements/changes to the
EJSM.
Designing a "welcome page" with an orientation for potential users would be a critical improvement if these are to be a public-access sites. Even setting
aside the issue of public access, there is not a sufficient framework for your technical advisors to make their own judgments about the tools. While it is
interesting to review how the methods vary in their identification of environmental justice communities, it is not clear about the implications to be drawn
from this analysis. On a more practical basis, since the CalEnviroScreen is now the "law of the land" one approach is to use the comparisons with the
EJSM and the CEVA as a means to improve the CES. Another approach would be to differentiate circumstances in which the CEVA and/or the EJSM
would be more appropriate than the CES (and vice versa). Either approach is reasonable, but without some guidance on the intended use of this side by
side analysis, it is not clear what practical value it has.
The web-based platforms were developed solely for this research project and the intended audience is the project partners. Our intention was to
provide easy access to the maps so that project partners could compare the results of the tools. We currently do not have plans to maintain the
web-based platforms after this project is completed. One of the web viewers currently provides a brief description of the access and use
constraints. This is contained in the metadata and is pasted below. We are currently checking whether we can add this information to the second
web viewer.
Page 2 of 12
-------
Response to Comments - RARE Project Comparing Three Screening Methods in the San Joaquin Valley
November 15, 2013
%x5fJjjlto ^aujxuaiiv;: "ife-. if j^xilia^
"This web viewer has been made available to visually compare the results of three environmental justice or cumulative Impacts screening
tools applied to the San Joaquin Valley- This web viewer was developed as part of an EPA research project to pilot the application of the
Environmental Justice Screening Method- The results of the screening tools displayed in this web viewer and the web viewer itself do not
represent EPA's official views or positions on environmental justice screening. This web viewer has been developed solely for research
purposes under the EPA Regional Applied Research Effort (RARE) Program- More Information about the RARE Program Is available
at http://www.epa,gov/osp/reglons/rare,htm-"
Taking on the content of the websites themselves, there are several points to be made about the treatment of the CEVA. First, a point of terminology: in
some places the website refer to our work incorrectly as the California Environmental Vulnerability Assessment, instead of the Cumulative Environmental
Vulnerability Assessment. More significantly, I am concerned that your team may have misapplied the CEVA categories. Unlike the EJSM and CES, which
use a linear ranking from low to high, because the CEVA uses a 3x3 matrix, the "top" three categories that represent the Cumulative Environmental
Vulnerability Action Zones are categories 6, 8, and 9. These correspond to the census tracts that are ranked highest in both Cumulative Environmental
Hazards and Social Vulnerability (category 9) or high in one and medium in the other (categories 6 and 8). Category 7, in contrast, is ranked as high in
environmental hazards and low in social vulnerability and is not considered one of our Cumulative Environmental Vulnerability Action Zones (CEVAZ).
7=High CtHI/
T
Low S\|l
4=MedCEHl/
Low SVI
l=LowCEHI,
Low SVI
8= High CEHI/
Med SVI
5= Med CE;
Med SV
HI/
2=LowCEHI/
Med SVI
9= High C
HighS
6= Med C
High 5
EHI/
/I
EHI/
VI
3=LowCEHI/
High SVI
If the comparison included category 7 in the set of highest EJ scores, it would significantly misrepresent the CEVA and problematize the comparison. This
was a point made by the CRC Director of Informatics during the RARE webinar in July but it is not clear whether this was incorporated into the final
analysis. If the cross-tool analysis does indeed use category 7 as a "high" quantile, I would strongly recommend that US EPA rerun the comparison
between the three tools before going public with the website.
We apologize for incorrectly referring to CEVA as the California Environmental Vulnerability Assessment. We are working to correct this.
PageS of 12
-------
Response to Comments - RARE Project Comparing Three Screening Methods in the San Joaquin Valley
November 15, 2013
Specific Comments and Responses
The analysis and comparison reported in the June and July 2013 webinars did not include CEVA scores 6 or 7 in either definition of total
cumulative vulnerability score "hot spot". This can be seen if you refer to the webinar PowerPoint, slides 27 and 33 define the "top quantile" and
"high quantile," respectively.
During the webinar, three different ways of comparing the results of the tools were provided. Two of the methods were based on what percentage
of the San Joaquin Valley population is located in the "high score" areas of each screening tool. The first comparison method compares the
highest scores from each method that represent about one tenth (7.2 -11.24%) of the San Joaquin Valley population. For CEVA, we included
areas that have a score of 9. (Please see slide 27 of the presentation for a breakdown of the scores and population.) The second comparison
method compares the highest scores from each tool that capture about one-fifth to one-quarter (20.7 - 24.6%) of the Valley's population. For
CEVA, we included areas that have a score of 8 or 9. (Please see slide 33 of the presentation for a breakdown of the scores and population.) The
third comparison method addresses a different question, that of agreement among screening methods for identifying high hazard/pollution
exposure. It focuses on the hazard/pollution exposure scores and does not consider social vulnerability. Therefore, CEVA scores 7, 8, and 9,
those scores which CEVA identifies with the highest hazard/pollution exposure, were used in this comparison.
As our primary response to the question of "potential improvements to tools' data choices, metrics, analysis, and scoring methodology", I would direct you
to the CRC's new version of CEVA, as developed for an EJ mapping project in the Coachella Valley. In this version of the CEVA, we added a number of
new data sets (including drinking water contamination) improved the use of several indicators (including selecting regionally-relevant and high-concern
pesticide) corrected many of the index construction shortcomings, and drew on some of the more recent innovations from the EJSM and the CES. Like the
San Joaquin Valley CEVA project, this tool was developed in a collaborative partnership with key local and regional stakeholders, resulting in a regionally-
specific tool with numerous practical applications. Please see: http://reqionalchanqe.ucdavis.edu/ourwork/proiects/ceva-coachella-valley
Comment noted.
I would make the more general point about improvements to the EJ tools, that this RARE pilot project is an excellent affirmation of the value of developing
regionally-specific approaches. While the CES has clear benefits in being state-wide and ideally adopted by a wide range of state agencies, a regional
approach allows for selection of indicators that are relevant to the issues, concerns, and policy initiatives in that region. As I have recommended to
OEHHA, I believe that California would be better served by a state-wide framework that was built up from a set of regionally-specific tools. This could still
be used to allocate state funds and other resources, but on a stratified basis that identifies the highest vulnerability communities in each region. Such an
approach will also encourage greater community participation and buy in, especially among the stakeholders that are most affected by environmental
injustices and under-represented in public policy.
Comment noted.
Page 4 of 12
-------
Response to Comments - RARE Project Comparing Three Screening Methods in the San Joaquin Valley
November 15, 2013
Specific Comments and Responses
The question of "ground-truthing" is an important one, both as a means to better engage environmental justice stakeholders and to improve the
methodologies themselves. However, without a clearer definition of what is meant by ground-truthing in this context, it is difficult to recommend locations to
carry this out. Ground truthing could include fact-checking the specific data sets (e.g., verifying the location and operations of specific TRI sites); the
collection of primary data in target locations (e.g., drinking water contamination testing); the implementation of community-based workshops to compare
the tool outputs with local knowledge and experiences, among others. All of these are worthwhile activities, but each call for a very different level of
investments and organizational capacity, and result in very different outcomes. My recommendation to US EPA is to convene a group of key stakeholders
from the research, agency, and advocacy sectors to develop a strategic framework for ground truthing before doing any site identification. This stakeholder
engagement process can then begin to generate specific policy initiatives that one or more of the EJ tools can address, and which can become the basis
of case studies. Once this is accomplished, US EPA will be well positioned to initiative and document a set of implementation pilot projects.
Field validation for this project will include verifying locations of sensitive land uses and facilities - those that may or may not be included in
publicly available databases. In addition, community members will have the opportunity to identify additional hazards in their communities that are
not captured in the databases that the screening tools draw data from. Field validation and feedback from community members will be used to
enhance the utility of the EJSM, and hopefully inform other methodologies. (Funding for this project can only support improvements to the EJSM.)
Given the limited resources and timeframe for this project, we are unable to convene a large meeting of key stakeholders to develop a strategic
framework for field validation. Project partners were provided with an opportunity to inform field validation during the comment period following the
webinars. Some comments we received included recommendations for field validation, and we are considering these comments as we develop
the field validation strategy. Community members are an essential part of field validation because of their knowledge of their community.
Community partners will help finalize locations for field validation.
Thank you again for this opportunity to comment on the RARE project. It has been a pleasure interacting with US EPA staff and I am very enthusiastic
about deepening this partnership overtime. Please let me know if you have any questions about my comments or if we can be of any other assistance.
Commenter: David Lighthall, San Joaquin Valley Air Pollution Control District
Verbatim Comments with Responses Italicized:
I would like to add some comments that I believe are in consonance with Kevin's. Clearly the CES lacks sufficient spatial resolution. Regarding the other
two models, after looking at some sample locations in Fresno that I am familiar with, my sense is that CEVA has the best resolution for demographic
factors (and perhaps a scoring bias that overemphasizes demographic factors) with EJSM having a superior capacity to factor in land uses into the
scoring. As a result, you see EJSM giving higher scores in parts of Fresno that are middle class but adjacent to high emission sources such as SH 41.
PageS of 12
-------
Response to Comments - RARE Project Comparing Three Screening Methods in the San Joaquin Valley
November 15, 2013
Comment noted. Compared to the EJSM, CEVA places a higher weight on demographic factors in the final score. CEVA uses Census block
groups for scoring. EJSM uses block group demographic information, but aggregates up to the Census tract level because other datasets used in
the tool are available at the tract level or a larger level.
Commenter: Amy Vanderwarker
Verbatim Comments with Responses Italicized:
Hi Debbie - this viewing version is great! Really helpful and much easier to use than the other one. Is there capacity/plans to do a similar comparison for
EJSM and CES throughout the state? Obviously CEVA is only for the Central Valley but it would be so nice to have this for other areas to look at EJSM vs
CES.
Given the limited resources for this project, EPA is unable to develop a web viewer for the entire state.
Commenter: Kevin Hamilton, Clinica Sierra Vista, Inc.
Verbatim Comments with Responses Italicized:
With regard to the CES. Unfortunately ZCTA's make targeting specific populations and communities in the SJV difficult as many are quite large. An
example of this if Fresno's 93706. While the economic profile for the area is fairly homogenous, its social/demographic makeup is not. It contains densely
populated core urban centers, fairly dense suburban, light/heavy commercial and industrial, and rural/agricultural. Part of it is within the City of Fresno but
there are three unincorporated cities. This would make risk assessment, planning, policy making and resource allocation very challenging. It also limits
individual neighborhoods ability to be specific about the hazards and challenges they face on a daily basis.
We agree that using spatial areas smaller than the zip code is preferable. OEHHA has previously stated that CalEnviroScreen will eventually move
from using zip codes to Census tracts. However, the spatial resolution is limited by the indicator data used, and some data are only available at a
relatively coarse level of spatial resolution, limiting the granularity of any method dependant on that data.
CEVA. This tool has the highest level of resolution for valley communities where health and social vulnerability are the main considerations. It is also the
only tool to integrate "ground truthing" allowing residents to feel the results truly reflect the challenges they are experiencing day-to-day. However, the
data background is missing water related health and social impact data and so would need to be updated with that information.
CEVA maps or reports at the Census block group level, but it also conflates data that is only available at a coarser spatial resolution to the smaller
Page 6 of 12
-------
Response to Comments - RARE Project Comparing Three Screening Methods in the San Joaquin Valley
November 15, 2013
Specific Comments and Responses
Census blocks. EJSM uses Census blocks for its demographic information, but in all cases aggregates up to the level of spatial resolution that is
coarsest for all data used (e.g., Census tracts). EJSM development also included significant and sustained input from community partners
throughout the State to ensure community concerns are reflected in the method and its results, which is why land use and hazard proximity figure
so prominently in the EJSM. Community members will have the opportunity to provide feedback on the EJSM during field validation.
Cl. This tool has the design methodology that would allow it to be most useful. In fact, combining the data pool from the first two into this one-including
ground truthing-would probably make it the most useful to the community, agencies and policy makers.
Comment noted.
Commenter: Cesar Campos, Central California Environmental Justice Network
Verbatim Comments with Responses Italicized:
1) First of all, I think that once CalEnviroScreen moves to census tracts it will make it so that some of the smaller isolated towns are picked up, much like
CEVA and EJSM agreed on. This was a problem with CES to begin with because zip codes are very large and sometimes the differences between two
areas in one zip code are enormous. I find that the process of overlaying all three maps together and finding "hot spots" is geographically sound. The
problem is the scoring metrics and the fact that EJSM used many more variables than CEVA or CES. I know that all of those variable only factor to 1
score, but I'd be interested in seeing how different the CES and CEVA maps would look if all of those variables were taken into account. For the time
being, I understand the use of the overlaying method and I think that it does provide good information in terms of finding areas that are specially affected
and vulnerable. Moving forward I think that it will be important to protect the variables that were originally used and continue to be effective in telling a
story. I know that CES recently dropped race as a variable and feel that that neglecting variables that were originally used in data acquisition will
ultimately be more hurtful to the models than effective. I'd like to see CES and the other screening tools implement a variable that accounts for how far a
person has to travel to the nearest clinic, pediatrician, emergency room, etc.
Comment noted. Access to healthcare could be used as a vulnerability metric, and the EJSM researchers are considering that suggestion. The
best immediately available dataset for this is available from the California Geospatial Data Library (CalAtlas). In considering whether and/or how
to use this data, we first need to understand how the data was automated and how "medically underserved" is defined.
2) I find that all of these screening methods are very useful as they provide concrete examples of areas in the state that obviously need more attention. In
talking with policymakers, and forming positions on legislature, I find that CES and CEVA have helped me a lot to understand the situation in a place
based way. I am not as familiar with EJSM but will give it a try because I like it's extensive use of variables. Moving forward, I think that if you are able to
identify "hot spots" or "top quantile" of overlap, this will help CCEJN advocate for more resources and attention to overburdened communities.
Page? of 12
-------
Response to Comments - RARE Project Comparing Three Screening Methods in the San Joaquin Valley
November 15, 2013
Specific Comments and Responses
One of the web viewers (http://bit. ly/143UGkq) maps areas of overlap. As part of this project, we are looking for case studies of how any of the
tools assist stakeholders with answering their policy relevant questions.
3) I would like for this comparison efforts to bring about a true map of hot spots that takes into account the differences in variables used for each method,
but still finds the top quantile areas of overlap. I imagine that in order to ground truth one should look at areas in which two of the methods coincide in
respect to areas in which they all coincide. If an area around a "top quantile" shares similar demographics to the "top quantile" area and was identified by
at least two of the screening tool-l imagine those would be the best places to ground-truth.
We will consider this feedback as we develop the field validation strategy.
Thank you for letting us participate in this process. I imagine your team has put in a vast amount of work...
Commenter: Alvaro Alvarado, California Air Resources Board
Verbatim Comments with Responses Italicized:
I don't have many comments. I like the way the map shows areas where the three methods agree. My concern with the mapping is figuring out a way to
minimize false positives and false negatives. False positives leave us with too many areas to investigate, while false negatives means we are focusing our
efforts in the wrong areas. Where multiple mapping methods agree seems to me to give areas with no false positives, but probably many false negatives.
It seems a like focusing on these areas and comparing to areas that show up on only one map is good place to think about ground trothing. I am still not
sure the best way to do it, but at least this mapping narrows down the possible areas.
The false positives/negatives problem is why ground truthing is done. We will consider this feedback as we develop the field validation strategy.
Commenter: Randy Segawa, California Department of Pesticide Regulation
Verbatim Comments with Responses Italicized:
Thanks for the opportunity to comment on the cumulative impacts methods comparison. I think the methods and presentation of the results are fine, but I
have a few suggestions about the ground truthing. I think someone mentioned that the candidate communities are the five with the highest hazard scores.
If so, I suggest that you include Tulare as one of the communities for the ground truthing. The other four communities have the highest populations and are
the most urbanized of the SJV communities. Tulare is the only one that might be considered similar to other SJV rural communities.
PageS of 12
-------
Response to Comments - RARE Project Comparing Three Screening Methods in the San Joaquin Valley
November 15, 2013
We are still in the process of identifying candidate communities for ground truthing. We will consider the suggestion of including Tulare as we
develop the ground truthing strategy.
One concern that I have with these screening methods is the need or desire to assign scores to small areas, such as census blocks. Unfortunately, much
of the hazard and exposure data do not have this resolution, requiring estimates or extrapolations. While it would be prohibitively expensive to measure
ozone or PM in each of several census blocks to check their estimation methods, the methods to estimate pesticide use can be checked. The pesticide
scores rely on the Department of Pesticide Regulation's database of pesticide use reports. The finest spatial resolution in this database is section (1 x 1
mile area). County agricultural commissioners have the same data, but they can identify the location of application by field. More accurate pesticide scores
can be determined using the ag commissioner data, and compared to the scores determined with the three screening methods.
The methods do not include scores for Census blocks. Some of the spatial resolution of the data metrics used in the EJSM are not sufficiently
granular to allow scoring or mapping at the block level, and we follow good geospatial practice to aggregate up to tracts, the level of the coarsest
data we use. It is possible to express the ozone and PM data at the block level; however it would be less accurate given the characteristics of the
original data (CARS air monitor network locations). The EJSM uses data from the Department of Pesticide Regulation's Pesticide Use Reports. It
is aggregated at the census tract level, and we currently have data from 1991 through 2009. It was processed and made available from the
California Environmental Health Tracking Program (CA Dept of Public Health).
Processing field level pesticide use data would require more time and resources than this project can support. Although we cannot utilize these
data for this specific project, we are interested to better understand pesticide use at the field level. Please let us know if the Department of
Pesticide Regulation could assist us with obtaining field level data.
This may outside the scope of your ground truthing, but at least two of the three methods use static data. This is probably fine for certain pollution sources
such as waste sites that don't change location or emissions much overtime. However, ozone, pesticide use, and other hazards may change significantly.
You may want to evaluate year to year changes and how these changes impact the scores of some of the hazard parameters.
The tools use multi-year averages for some of the hazard/pollution burden indicators from static monitors. This is the only data available for these
indicators.
Commenter: Catherine Garoupa White, DC Davis
Verbatim Comments with Responses Italicized:
Hi Debbie and Jacquelyn, I haven't heard whether the web viewer is up yet, and wanted to submit my general comments while there is still time. I would be
Page 9 of 12
-------
Response to Comments - RARE Project Comparing Three Screening Methods in the San Joaquin Valley
November 15, 2013
Specific Comments and Responses
interested in looking more closely at the maps presented on the web viewer whenever it's available. In the meantime, here are my thoughts on the July
webinar I participated in:
Thank you for the time and resources put into reviewing these three methodologies for analyzing/capturing issues in the San Joaquin Valley. As a brief
background (which Debbie knows well already!), I was born and raised in the valley and have worked there on social and environmental justice issues for
over a decade now. I'm currently a doctoral student in Geography at DC Davis, and I enjoy checking out all these maps! In full disclosure I work on a
mapping project related to promoting equity in the valley at UCD's Center for Regional Change currently, and I was an active participant in the group that
helped develop the CEVAZ index. However, I am submitting these comments on my own behalf, not with any organizational affiliation. These methods, as
you've noted, have serious policy implications for addressing environmental justice in the Valley, so I'm glad to be able to participate in this conversation.
With all due respect to Jim Sadd, it felt skewed to have one presenter, who is a collaborator in developing 1 of the 3 methods, representing all three
different methodologies. Several times errors were raised that the presenter made in representing these methodologies, and these issues were not fully
addressed during the meeting. Overall, this could have been a more productive dialogue if each team presented their own methodology, and then
representatives of those methodologies were able to have a constructive discussion with stakeholders about the pros and cons of each, and their
limitations. These methodologies were tailored to specific needs, based on a variety of factors such as who the teams were that developed them, who the
collaborators and/or funders were, etc. who had a specific use in mind. Discussing these tools without the context they arose from also felt disorienting.
The closing conversation seem to center on which method was superior, rather than which tool is appropriate for which context. Not fully being a part of
this process, I'm not sure what additional communication may have taken place before or after, so perhaps some of these side conversations and shared
knowledge already exist.
The research contract for this project was granted to Dr. James Sadd and his team, which is why Dr. Sadd was the main presenter during the
webinar. We will ask the OEM HA and UC Davis research teams to participate as presenters in future webinars/presentations, if necessary.
Corrections were made during the webinar and the presentation was subsequently modified to correct any errors. The modified presentation was
provided to the project partners.
The purpose of the webinars was to compare the results of the three methods and not examine the methodologies in detail. We held a meeting in
November 2012 with project partners and a discussion of the methodologies was provided then. Reports that discuss the tools' methodologies and
limitations are publicly available. We have asked project partners to familiarize themselves with these reports.
It has never been our goal to show one method is better than another, and this was not discussed in the webinar. The main objective of the project
is to develop case studies demonstrating how existing screening tools can answer policy relevant questions. The overall outcome of the project is
to inform the development of new or refinement of existing environmental justice or cumulative impacts screening tools. This project can only
support improvements/changes to the EJSM.
Page 10 of 12
-------
Response to Comments - RARE Project Comparing Three Screening Methods in the San Joaquin Valley
November 15, 2013
Specific Comments and Responses
In addition to what was presented during the webinar, there are several standard issues that come up with capturing and manipulating data spatially that I
feel should also be a part of the discussion. These include issues such as edge effects, meaning that because of a political boundary such as county lines,
related issues just across the border may not be captured yet may have a significant influence on the factors being examined. A related issue is often
referred to as the modifiable areal unit problem, meaning that the unit of analysis that data is available in, such as census tracts, are often arbitrary units
which do not directly correlate to the communities being looked at. As was lightly touched on during the call, there are also limitations that must be
considered in terms of weaknesses and/or gaps in the data being used, whether it is census data or parcel data etc., especially in a predominantly rural
region like the Valley, which I've found often does not lend itself to easy or fully accurate analysis when working with aggregated data like the census. The
statistical methods used can also contribute to oversimplification or error. Of course many researchers are already well versed in these constraints but in a
conversation that includes advocates, I find it important to at least briefly touch on potential limitations to applying ANY of the methods presented.
Comment noted. The limitations of the methodologies are reviewed in the papers and reports that the different research teams have published.
The idea of "groundtruthing" is laudable; however I'm concerned that it might set the tone of the conversation to be about which method is better overall
rather than which method is appropriate for what use. Before going directly to communities to ground truth, I would recommend spending more time
looking at the sources of data, etc. for explanations as to why there may be different results between the different methodologies. Ground truthing also
implies that there is a truth to find, when differences may be related to other factors already mentioned such as data sources, unit of analysis, etc.
The purpose of ground truthing is not to demonstrate which method is better than the other. Field validation can help us better understand the
limitations of the tools and/or the datasets that support them. Field validation for this project will include verifying locations of sensitive land uses
and facilities - those that may or may not be included in publicly available databases. In addition, community members will have the opportunity to
identify additional hazards in their communities that are not captured in the databases that the screening tools draw data from. Field validation and
feedback from community members will be used to enhance the utility of the EJSM, and hopefully inform other methodologies. (Funding for this
project can only support improvements to the EJSM.)
From a general process perspective, offering an in person location might have made engaging with such detailed subject matter more feasible. These
tools do have important policy implications and applications, which there was not much room to discuss with a focus on methodology - is this going to be a
future conversation or is it outside the scope of this effort? Perhaps it's because I haven't been a part of the entire process, but I'm still unclear on what the
ultimate goal of these conversations is, and what EPA is hoping to achieve?
Compared to webinars and conference calls, in-person meetings are generally better for many reasons. Given the funding constraints, we were
unable to provide an in-person meeting to discuss the comparison maps and methodology.
Thanks for the opportunity to provide feedback, and don't hesitate to contact me with questions or if there are future meetings.
Page 11 of 12
-------
Response to Comments - RARE Project Comparing Three Screening Methods in the San Joaquin Valley
November 15, 2013
Specific Comments and Responses
Commenter: Catherine Garoupa White, DC Davis
Verbatim Comments with Responses Italicized:
Hello, Thanks for the detailed response and background info, and for letting me weigh in though I was not formally a part of the ongoing process. I can
definitely relate to timing, funding, and capacity constraints. It has been a crazy busy month preparing for the start of school and I have not been able to
use the web viewers yet, but I did want to reply to your feedback, and so will limit my comments now to the topic of ground truthing. If the funding is
available I think it's a worthwhile component, but given the shifting nature of the project and ultimate outcomes it seems like calling it something other than
ground truthing would honor the validity of each method and the process each underwent in development and focus more toward the question of which
method for what purpose, which seems like the central focus. As mentioned previously, my concern with the term ground truthing is that it implies there is
one truth to be teased out.
Ground truthing was a part of the CEVA and EJSM processes. We agree that the term ground truthing can be confusing, and thus we will be using
the term field validation instead. Field validation for this project will heavily rely on community participation.
Will the web viewers be up beyond the comment period? I would still like to take a look at the maps when I have time.
The web viewers will be available online until the end of the project period (September 2014). Due to limited funding for this project, the web
viewers will not be updated to reflect any changes to CEVA or CalEnviroScreen.
Page 12 of 12
-------
-------
APPENDIX D:
City of Commerce Green Zones Working Group
Opportunity Areas (Map)
MonttWIo
.Commtrct
Mttrolink
Strtor
GREEN ZONES
WORKING GROUP
OPPORTUNITY AREAS
- -. .
LEGEND
J.i nr*«»tf(WNMr*i
* * «>ts
Q Chlcunlrr
D-l
-------
-------
E-l
-------
Central California Environmental Justice Network
Roots of Resilience, 2014
AGENDA
Welcome 10:00-10:15 AM
Current State of Environmental Justice 10:00-10:45 AM
(Panel Discussion)
^^— ^ Break —
10:45-11:00 AM
11:00-12:45 PM — Workshops—Block A
11:00-12:45PM
The Future of
Tracking Activism
-Madeline Stano (CRPE),
Tia Lebherz (FWW)
11:00-11:45PM
Rural Opportunity Index I Environmental Racism, Mass Incarceration I Using Arts & Culture for Organizing
-Jonathan London (CRC) I and Immigration - Mie feyes (CPMPj I -Isabel Arrollo (EQS)
11:45-12:00 PM
^—^— Break
12:00-12:45PM
,11:45-12:00 PM
Community Monitoring Networks
--Luis Olmedo (CCV), Jessica Hendrid
Air Pollution Regulator Accountability
-Dolores Welter (CVAQ), Tom Frantz (Air)
Lunch
12:45 - 2:00PM
Workshops—Block B
2:00-3:45PM
2:00-3:45 PM
2:00-2:45PM
2:45-3:00 PM
,2:45-3:00 PM
Shifting Gears Fresno: Increasing Bike I Current State of Kettleman City Ivideo Voices & Youth Organizing
PartneringwithEnvironmentalAgendes Ridershipfromi%to3o% |--Mar!ce/aMares-A/atorre(£/PueWojl"Yoi'f/I'Green/ieWWa'/c''n9
and Communities to Evaluate the I • 'Group
Environmental
Justice Screening Method (EJSM)
-Manuel Pastor (USC),
James Sadd (Occi)
Break
3: 00-3:45PM
Community Resiliency in the Face • Land Use and Environmental Justice
of a Drought H -Leadership Counsel for Justice &
-RyanJensen (CWC) H Accountability
Closing 3:45-4:00 PM
-------
------- |