# EJS to EBS Crosswalk -- Canopulse UCRI v1.0 to v2.0

Status: Locked before compute.
Date: 2026-07-10.

---

## Summary

The environmental justice / environmental burden component was renamed and
redefined between UCRI v1.0 and v2.0.

| Attribute              | v1.0 (LA)                                  | v2.0 (California)                                |
|------------------------|--------------------------------------------|-------------------------------------------------|
| Component abbreviation | EJS                                        | EBS                                              |
| Full name              | Environmental Justice Score                | Environmental Burden Score                       |
| Source dataset         | CalEnviroScreen 4.0                        | CalEnviroScreen 4.0                              |
| Source metric          | CES 4.0 overall percentile                | CES 4.0 **Pollution Burden percentile only**     |
| Weight in UCRI         | 0.15                                       | 0.15 (unchanged)                                 |
| Percentile scope       | County-wide (LA County)                    | Statewide (all CA tracts)                        |

---

## Rationale for the change

The CES 4.0 overall score combines two major sub-indices:

1. **Pollution Burden** -- environmental exposures and effects (ozone,
   PM2.5, diesel PM, pesticides, toxic releases, traffic, cleanup sites,
   groundwater threats, hazardous waste, impaired water, solid waste).

2. **Population Characteristics** -- sensitive populations and
   socioeconomic factors (asthma, cardiovascular disease, low birth weight,
   educational attainment, housing burden, linguistic isolation, poverty,
   unemployment).

The Population Characteristics sub-index substantially overlaps with the
CDC Social Vulnerability Index (SVI), which is the input to the SVS
component. Using the full CES score for EJS in v1.0 therefore introduced
partial double-counting of population vulnerability.

In v2.0, EBS uses only the Pollution Burden percentile from CES 4.0. This
isolates the environmental exposure signal and avoids redundancy with SVS.

---

## Impact on scores

Because EBS uses a narrower (pollution-only) input than EJS:
- Tracts with high pollution burden but low population vulnerability will
  see their EBS similar to or slightly higher than their v1.0 EJS.
- Tracts where the high CES score was driven primarily by population
  characteristics will see EBS lower than their v1.0 EJS.
- The net effect on UCRI is modulated by the 0.15 weight.

Additionally, the change from county-wide to statewide percentiles means
that a tract's relative position is measured against all CA tracts, not
just its county peers. This further contributes to v1.0 vs v2.0 score
differences.

---

## Field mapping

| v1.0 field    | v2.0 field    | Notes                              |
|---------------|---------------|------------------------------------|
| comp.EJS      | comp.EBS      | Renamed; different input metric    |
| ctx.cesPctl   | ctx.cesPollutionPctl | Renamed to clarify source   |

---

## LA back-compatibility note

LA v1.0 remains the frozen, citable release for the submitted UCRI paper.
California v2.0 values for LA tracts will differ from v1.0 by design
because:
1. The percentile reference population changes from countywide to statewide.
2. EBS replaces the broader EJS definition.

This is expected and is NOT an erratum. See MODEL_REGISTRY.json for the
full back-compatibility statement and outputs/tables/la_v1_v2_correlation.csv
for the correlation report.