UCRI Method
A documented screening index for California tracts.
The Urban Canopy Risk Index (UCRI) is Canopulse's current statewide model. It combines canopy deficit, surface heat, impervious surface, social vulnerability and pollution burden into one relative score for each California census tract. CalEnviroScreen 5.0 supplies the pollution burden component.
9,129
California 2020 census tracts
9,041
fully scored and ranked
66
provisional estimates, not ranked
22
water-only tracts, not scored
Decode any score
Reading a UCRI score
Drag across the statewide distribution to read any score in context: its percentile, how many tracts have similar scores, and the nearest California tract.
Score decoder | California distribution
Linking dataset
Drag the chart or tap a percentile chip. Every readout is computed in your browser from the released dataset.
Formula
The current UCRI weighting
Each component is a statewide percentile from 0 to 100, where higher means a stronger contribution to screening priority. The released score is the weighted sum, rounded to 0.1.
UCRI = 0.25 × CDS + 0.25 × HES + 0.15 × ISS + 0.20 × SVS + 0.15 × EBS
CDS
Canopy deficit
How little tree canopy the tract has (100 minus canopy percent)
25%
HES
Surface heat
Summer daytime land-surface temperature, not air temperature
25%
ISS
Impervious surface
Share of paved and built surface
15%
SVS
Social vulnerability
CDC/ATSDR SVI 2022 overall percentile, ranked statewide
20%
EBS
Pollution burden
CalEnviroScreen 5.0 Pollution Burden percentile, used as published
15%
Method version UCRI-CA v2026.07 (formula and weights unchanged since July 2026); data release CA_2026.10, frozen October 8, 2026.
Read with care
What the inputs measure, and what they do not
These limits apply to every tract value, map and report. They are part of the method, not footnotes to it.
Surface heat is not air temperature
Surface heat is the daytime temperature of roofs, pavement and ground measured by Landsat satellites on summer mid-mornings (about 10 to 10:30 a.m. local time). It runs much hotter than air temperature and does not measure air temperature or nighttime heat. A tract can read 120 °F or more on this measure on a day when the air is far cooler.
Satellite canopy undercounts street trees
Tree canopy comes from a 30 m satellite model (NLCD Tree Canopy Cover 2025). It misses many street trees and small canopy in dense urban areas, so values are usually lower than a city's own 1 m canopy assessment. Use it to compare tracts, not as a local canopy inventory.
Statewide percentiles reflect climate and biome
Canopy deficit answers how little canopy a tract has, not how far it falls below the canopy its climate can support. Desert and Central Valley tracts are naturally low in canopy and hot in summer, so they rank high on canopy deficit and surface heat by construction, and humid coastal tracts rank low on surface heat. Compare tracts within a region with that in mind.
Tract averages describe places, not people
Every value is an average over a whole census tract. It does not describe any one resident, household, block or parcel, and a tract average can hide large differences inside the tract (the ecological fallacy). Social vulnerability and pollution burden are themselves tract-level composite indices.
Tracts are uneven units
Tracts range from well under one square mile to thousands of square miles, and a few have very few residents. Canopulse flags tracts with fewer than 500 residents and special land-use tracts (Census tract codes 9800 to 9899, such as industrial areas, airports or parks). Flags never change a score or rank, but read those scores with care.
Screening, not site selection
Tract-scale screening cannot identify plantable parcels, ownership, underground utilities, irrigation feasibility or project cost. Use it to decide where to look closer, then check local canopy, site and community information.
Screening classes
Five classes, each a fifth of California tracts
Screening priority classes are fifths of the 9,041 fully scored California tracts in this release, ranked by UCRI. "Very high" means the top fifth statewide (UCRI 70.3 and above). The classes are relative screening categories, not hazard levels and not official designations.
| Class | UCRI range | Fully scored tracts | Stored data value |
|---|---|---|---|
| Very high screening priority | 70.3 and above | 1,814 | Extreme |
| High screening priority | 56.9 to below 70.3 | 1,816 | High |
| Moderate screening priority | 44 to below 56.9 | 1,808 | Moderate |
| Low screening priority | 30 to below 44 | 1,804 | Low |
| Very low screening priority | 0 to below 30 | 1,799 | Very Low |
The cut points are the 20th, 40th, 60th and 80th percentiles of the 9,041 released scores. The data files, CSV downloads and API keep the stored value "Extreme" for the top class; the website and reports display it as "Very high screening priority". The two names mean the same class: the top fifth of fully scored California tracts.
Not an official designation
UCRI classes are not official designations. They do not show SB 535 disadvantaged community status, AB 1550 low-income community status or eligibility for any program. Check the official CalEPA and California Climate Investments maps for those.
Statewide comparison
Why 36 of 58 counties have no top-class tract
UCRI classes answer one question: which tracts rank highest across all of California. They do not answer which tracts rank highest within a county or city.
What the statewide classes show
Because every class is a fifth of all fully scored tracts statewide, the top class holds about one fifth of tracts by construction, and it concentrates where hot, paved, low-canopy and burdened tracts cluster. Los Angeles County alone holds 1,030 of the 1,814 top-class tracts (56.8%). In Imperial County, 37 of 40 scored tracts are in the top class. 36 counties, including many coastal and mountain counties, have none. That pattern is a result of ranking against the whole state, not a sign that those counties have no tracts needing attention.
A second lens: rank within the county and place
Every fully scored tract also carries its rank and percentile within its county and within its city or census-designated place. Tract reports show them under the statewide score. Use the statewide score and class to compare across California, and the county or place rank to find the highest-ranked tracts in your own area. County and place ranks are released values; they are not separate classes.
Overlap patterns
The pattern label on each tract report
Tract reports also name the overlap pattern of a fully scored tract. Patterns describe which component percentiles are high; they are not screening classes, priorities or official designations. The first rule that matches is used, with no hidden weighting.
| Pattern | Rule (statewide component percentiles) |
|---|---|
| Canopy, heat and vulnerability overlap | CDS 70 or more, HES 70 or more and SVS 70 or more |
| Canopy and vulnerability overlap | CDS 70 or more and SVS 70 or more |
| Heat and paved-surface overlap | HES 70 or more and ISS 70 or more |
| Pollution burden and vulnerability overlap | EBS 70 or more and SVS 70 or more |
| Paved surface with elevated heat | ISS 70 or more and HES 55 or more |
| Canopy and heat overlap | CDS 70 or more and HES 70 or more |
| Near the High class threshold | No rule above matches and UCRI is within 5 points below the High cut point |
| General screening | None of the rules above applies |
Provisional tracts are labeled data-limited and water-only tracts not scored; neither receives a pattern.
Current source
One California model across every workflow
The Explorer, reports, project-area calculations, APIs, downloads and CanoProof all use this same pollution burden source.
| Component | Pollution burden (EBS) |
|---|---|
| Current source | CalEnviroScreen 5.0 Pollution Burden percentile (OEHHA) |
| Source release | July 1, 2026 |
| Status | Current and used in UCRI |
Earlier Canopulse releases (for example CA_2026.07_CES5) are archived in the technical archive for reproducibility and are not used in this release.
Method details
How the score is produced
The same specification is published as METHOD.md in the reproducibility package. The method supports screening and comparison; it does not provide parcel-level engineering certainty.
Statewide percentiles
Raw tract values are joined to 2020 census tracts by GEOID and converted to statewide percentile ranks so unlike units can be combined. CDS, HES and ISS are ranked among the 9,107 tracts with land area (the 22 water-only tracts are excluded). SVS is the CDC/ATSDR SVI 2022 overall percentile (RPL_THEMES) ranked among the 9,041 land tracts with a published SVI value. EBS is the CalEnviroScreen 5.0 Pollution Burden percentile (PollutionP) as published by OEHHA, not re-ranked; 9,106 of the 9,129 tracts have a CalEnviroScreen 5.0 row. Ties share the average of the ranks they span, so the lowest values round to 0.0 and tied maxima sit below 100 (the highest published CDS is 99.8 because 34 tracts have no canopy). Each component is rounded to 0.1 before weighting.
Inputs and observation periods
Canopy: NLCD Tree Canopy Cover 2025 (USFS v2025-6), 30 m. Impervious surface: Annual NLCD Collection 1.2 Fractional Impervious Surface 2025, 30 m. Surface heat: Landsat 8 and 9 Collection 2 Level-2 surface temperature, the per-pixel mean of every clear pass in 424 scenes from June 1 to September 15, 2026 (scenes under 20% cloud; cloud, cirrus, shadow and snow masked), averaged over each tract in °F; Landsat thermal data are natively 100 m. Social vulnerability: CDC/ATSDR SVI 2022, built from American Community Survey 2018 to 2022 estimates; RPL_THEMES combines all four SVI themes (socioeconomic status, household characteristics, racial and ethnic minority status, and housing type and transportation). Pollution burden: CalEnviroScreen 5.0, released July 1, 2026; each of its indicators has its own observation years, documented by OEHHA.
How stable the heat layer is
Canopulse alternated the passes in each Landsat path and row by date into two halves, built a composite from each half and compared the tract values. The split-half Spearman correlation is 0.90 statewide. The Spearman-Brown estimate for the full composite (the "rank reliability 0.95" quoted elsewhere) is 0.95: it estimates how closely another composite of the same number of comparable passes would order the tracts. Counties with at least 200 tracts below 0.90 on the split-half check: Alameda 0.80, Contra Costa 0.86, Los Angeles 0.88. This measures how repeatable the ranking is across observation days, not whether surface temperature represents air temperature or heat risk.
Missing data and provisional estimates
A tract is fully scored only when all five components are present. When a component is missing, the provisional estimate is the weighted sum of the available components divided by the sum of their weights; for a tract without SVS that is (0.25 × CDS + 0.25 × HES + 0.15 × ISS + 0.15 × EBS) / 0.80. Provisional estimates are labeled, excluded from the 9,041-tract ranking and given no class. Water-only tracts are not scored.
Ranks and ties
The statewide rank orders fully scored tracts by released UCRI, highest first. Tracts with the same published score are ordered by GEOID, so a rank difference between tied tracts carries no meaning. County and place ranks use the same rule within each county and place.
What is public and what is kept internal
Public in the reproducibility package: Every one of the 9,129 tracts with its five component scores, released UCRI or provisional estimate, class and statewide rank; The raw tract values behind the components, rounded to 0.1 (canopy percent, summer surface temperature in °F, impervious percent, SVI overall percentile, CalEnviroScreen Pollution Burden percentile); The method specification, a standard-library Python script that recomputes the release from those columns, and Canopulse's own weighting checks. County and place ranks and percentiles, confidence labels and tract populations are in the per-tract reports, the Explorer and the API, not in the package CSV. Kept internal: the raster processing pipeline and its intermediate files. The Explorer serves official 2020 tract geometry; simplified geometry is used only for small exports, report thumbnails and locator maps.
Reproduce UCRI yourself
The package lists every tract with its five component scores, released score, class and rank, plus the method specification and a short Python script that recomputes the release from those columns. It also holds Canopulse's own weighting checks, which are not an independent review.
Open the reproducibility packageUsing UCRI in a paper, grant or map? Cite the dataset.
Score status and confidence
What the labels mean
Two separate labels travel with every tract. Score status says whether a score was released. Confidence is a data-quality label based on the resolution and type of the input layers.
Score status
9,041 tracts. All five components are present. These tracts receive a released UCRI, a screening class and statewide, county and place ranks.
66 tracts. At least one component is missing (65 lack social vulnerability; 1 lacks social vulnerability and pollution burden). An estimate is shown and labeled, but it is not ranked and has no class.
22 water-only tracts with no land area. No score is published for them.
Confidence (data-quality label)
Every contributing layer at 30 m native resolution or finer, with a vintage within two years. No tract reaches High in this release.
All five components are present, but at least one layer is coarser than 30 m or is a tract-level composite index. Every fully scored tract is Moderate.
Reserved for scores built on layers with known quality problems. No tract is Low in this release.
A component is missing (provisional estimate) or the tract is water only (not scored).
Every fully scored tract is Moderate, so confidence does not separate one scored tract from another in this release. High requires every contributing layer at a native resolution of 30 m or finer with a vintage within two years. Landsat land-surface temperature is natively 100 m and was resampled to 30 m for alignment, which does not increase its information resolution, and both CDC/ATSDR SVI and CalEnviroScreen Pollution Burden are tract-level composite indices. Each of those three triggers the Moderate rule, so no tract in this release reaches High. A thermal source at 30 m native resolution, and component inputs that are measured rather than composite, would make High reachable. Both are data availability questions rather than changes to the index.
Validation status
Released checks, version boundaries and limits
Canopulse runs its own release checks and weighting-sensitivity checks. No independent review of the method has been completed.
California UCRI release
9,129 statewide tract records are indexed; 9,041 are fully scored, 66 are provisional and 22 are not scored.
CalEnviroScreen 5.0 provenance
The official July 1, 2026 PollutionP field is joined by tract GEOID, used as EBS and bound to the release by a source checksum. Tracts without an official row stay missing; no value is filled in.
What has not been done
No outcome validation (for example against heat illness or measured air temperature) and no independent methodology review. Reviewers are invited through the feedback page; none has been commissioned or completed.
Related tools
How Canopulse relates to existing tools
The index components are familiar and other tools cover parts of this ground. Each tool below has its own documented scope.
Tree Equity Score (American Forests)
Scores urban neighborhoods (Census block groups) across the United States on tree canopy against a canopy goal adjusted for density and biome, combined with surface temperature and socioeconomic and health factors. It answers how far canopy falls below a locally adjusted target.
i-Tree (USDA Forest Service)
A suite of tools for tree inventories and canopy assessments that estimates ecosystem services and benefits, such as air quality, stormwater and carbon, and produces reports for a site, city or region.
CalEnviroScreen (OEHHA)
California's versioned tract-level screening of pollution burden and population characteristics. CalEPA uses it to identify SB 535 disadvantaged communities. UCRI uses only its Pollution Burden percentile, as one of five components.
Canopulse UCRI
California census tracts only. Combines canopy deficit, surface heat, impervious surface, social vulnerability and CalEnviroScreen pollution burden as statewide percentiles, with versioned, citable tract, county and place reports. It is a relative screening index, not a canopy goal, a benefits model or an official designation.
Use the method on a real tract.
Open the Canopulse Explorer or read a current California UCRI tract report.