Patterns in FONSIs: Deliverable 6
Categorical-exclusion opportunities from prior decarbonization Findings of No Significant Impact
Executive Summary
Key Findings
- Every bounded, low-impact decarbonization action has a text-similar existing CE candidate (pending verification). None looks net-new (0 to develop), and none exceeds a parsed limit in its matched CE candidate (no deterministic expand candidates under the current heuristic). The likely opportunity is to adopt / harmonize an existing CE — 4 candidate adopt opportunities, pending CE-coverage adjudication.
- The clearest case is transmission upgrades within existing right-of-way: 37 bounded, low-impact FONSIs across BLM, DOE, and the power-marketing administrations, for work TVA already covers under its CE #17.
- Most decarbonization FONSIs lean on committed mitigation — 310 of 452 (69%) are mitigated FONSIs, so a CE built from them must encode the recurring mitigations as design criteria, not rely on case-by-case commitments.
- Adopting a peer agency’s CE is well-precedented: of 2,105 existing federal CEs, 130 action-families are already shared across two or more agencies.
- Net-new CEs most likely lie outside this candidate set — in the ~159 decarbonization FONSIs not yet categorized (the recommended next step).
This report delivers:
Technical support for new regulatory categorical exclusion development: identifying patterns in FONSIs.
Methodology
The deliverable runs three analyses, each on a different data source. They are not a single funnel — only Analysis 1 starts from the full FONSI corpus; Analysis 2 works from the mitigated subset of it, and Analysis 3 works from the existing federal CE catalog (not FONSIs at all).
| Analysis | Data source | Scope |
|---|---|---|
| 1 · Scaling CEs (develop, expand, or adopt) | Decarb EAs with FONSIs | 53/452 projects (~12%) |
| 2 · Mitigated FONSIs | Decarb Mitigated FONSIs | 310 of 452 projects (69%) |
| 3 · Current federal CE catalog | Existing federal CE catalog | 2,105 CEs |
Steps for each analysis
Analysis 1 · Scaling CEs
- Data — the 452 decarbonization EA→FONSI projects. A FONSI concludes a full Environmental Assessment, so each is a documented federal “no significant impact” determination — the evidence base for a CE.
- Sort each FONSI into a candidate action type (transmission upgrades, solar, geothermal exploration) by keyword and technology.
- Keep the bounded, low-impact subset within each type — the small, routine projects that could plausibly be a CE (in-corridor upgrades, not major rebuilds). This is a yes/no judgment about scale, not a numeric threshold: a FONSI is kept whether or not it states an acreage or length.
- Match each candidate against the existing federal CE catalog by text similarity.
- Compare sizes — read the matched CE’s stated size limit (acres, miles, kV, MW, wells), where it has one, against the bounded FONSIs’ own sizes, to test whether ours exceed it. This is the only step where stated numbers matter.
- Classify as develop (no existing CE covers it), expand (our FONSIs exceed the CE’s stated limit), or adopt (a CE exists, but at a different agency than the ones doing these FONSIs).
Analysis 2 · Mitigated FONSIs
- Data — the mitigated FONSIs: the subset whose “no significant impact” finding depends on committed mitigation rather than the action being inherently low-impact (310 of 452, 69%).
- Flag each FONSI’s mitigation dependence from the LLM read — case-specific-dependent (the finding rests on committed mitigation) vs design-feature-only (inherently bounded) vs none.
- Measure the mitigated share and the resource areas where mitigation recurs, within each action type.
- Read for CE design — a mitigation that recurs consistently across a class is a codifiable CE design criterion; an idiosyncratic one is a disqualifier.
Analysis 3 · Review of existing federal Categorical Exclusions
- Data — the existing federal CE catalog (the CE Explorer export — 2,105 CEs across 78 agency units, each linking to its eCFR source).
- Compare every CE to every other by text similarity.
- Flag near-duplicates that recur at more than one agency.
- Group those into shared “action families.”
- Read each CE’s stated numeric limits, where it has any → what is already covered and how routinely agencies share the same CE (the precedent for adopt).
Analysis 1 — Scaling Categorical Exclusions (CEs)
This analysis investigates how we can scale Categorical Exclusions (CEs) to respond to patterns we find across Findings of Non-Significant Impact (FONSIs). It looks at FONSI findings and helps us understand whether there is evidence suggesting we can develop a completely new CE to categorically exclude reviews that are expected to have specific characteristics, expand the parameters of a current CE to encompass project characteristics currently not covered by an existing CE, or encourage agencies and departments to adopt CEs that already cover projects of specific characteristics that are not covered by a current CE.
Step 1: Data at a glance
The 452 decarbonization FONSIs split three ways (colors match the figure):
- Bounded, low-impact → adopt — the 53 projects in a recurring action type and small/routine enough to be CE-shaped. Every one resolves to adopt — none net-new to develop, none exceeding a CE limit to expand.
- Recurring type, broader — the 240 projects in a recurring type but too large to be a CE (major rebuilds, greenfield).
- Not a recurring type — the 159 uncategorized projects: the net-new search pool (Analysis 3).
Steps 2-3: Sorting the FONSIs into action types and keeping bounded subsets
Every decarbonization FONSI is sorted (Step 2) into one of five recurring action types (transmission upgrades, solar, geothermal exploration, onshore wind, temporary resource assessment), or a sixth catch-all — the 159 FONSIs that don’t fit a recurring pattern (the net-new search pool, examined in Analysis 3). Within each recurring type we then keep (Step 3) the bounded subset:
- Bounded, low-impact (teal) — the small, routine version that could plausibly be a CE: in-corridor work, short segments, previously disturbed land, no major new construction. These are what we carry forward.
- Broader (grey) — the larger or greenfield projects (major rebuilds, new corridors, utility-scale greenfield). Set aside; not CE-shaped.
Keeping is a yes/no judgment about scale — is the action small and routine? — not a numeric cutoff. A FONSI is kept whether or not it states an acreage or length; stated sizes don’t enter until Step 5.
The kept subset shares a consistent low-impact fingerprint — the traits that make these FONSIs CE-shaped regardless of any stated acreage or length. Most sit on land that has already been disturbed and run inside an existing right-of-way; smaller shares are explicitly temporary or state that they add no new permanent road. In plain terms: these are not greenfield builds — they are modifications within existing corridors, on ground that has already been developed, which is exactly the low-impact signature a CE is meant to capture.
One honest caveat on reading this chart: the “no new permanent access road” bar is a conservative flag — it fires only when a FONSI explicitly states no new road, so its low share reflects under-reporting (most FONSIs simply do not address roads), not that the rest build major new access. The two dominant, well-evidenced traits are previously disturbed land and within an existing right-of-way. (Beyond what the chart shows, the enrichment also records whether each action tiers from a programmatic review and its overall bounded, low-impact judgment — further signals behind the same conclusion.)
Step 4: Matching to the existing CE catalog
Each candidate action type — e.g. geothermal exploration, the cluster of bounded geothermal-drilling FONSIs, summarized by its defining terms — is compared against all 2,105 existing federal CEs to find the closest one. The similarity score runs 0 to 1 and blends two signals: semantic similarity (do the texts mean the same thing?) and word overlap (do they share the key terms?). It is not a percent of matching text — it’s a cosine-style similarity.
So what counts as a strong match? First set a baseline: how similar is a candidate to a random, unrelated CE? Almost always tiny — about 0.07, and rarely above 0.20 (the grey band in the figure). Measured against that floor, the best matches — 0.45 to 0.67 — are several times higher: genuinely close, on-topic matches, not coincidence. For example, geothermal exploration matches a DOE exploration-drilling CE at 0.50 (sharing geothermal, exploration, well, temperature gradient); solar matches a DOE solar CE at 0.67, the strongest of the four.
The score is a ranking aid, not the verdict — it surfaces the closest CE, but the actual develop-vs-adopt call is made by confirming that CE against its eCFR text. No candidate is anywhere near the unrelated-CE baseline, so none resolves to develop.
| Action | Closest existing CE (held by) | Similarity |
|---|---|---|
| Solar | DOE-1--5-87 (Department of Energy) | 0.67 |
| Temporary resource assessment / site investigation | DOE-1--3-43 (Department of Energy) | 0.54 |
| Geothermal exploration | DOE-1--3-43 (Department of Energy) | 0.50 |
| Transmission upgrades within existing ROW | TVA---1-16 (Tennessee Valley Authority) | 0.45 |
Step 5: Compare sizes against the existing CE limits (the expand test)
This is the only step where stated numbers matter. The expand test asks whether our bounded FONSIs exceed the matched CE’s stated size limit. They don’t — because none of the four matched CEs state a numeric limit at all (they bound scope with words), so there is nothing to exceed and nothing resolves to expand.
It is still informative to compare our projects to the limits CEs do state elsewhere. The three metrics split cleanly — only the miles panel below is shaded green, to flag the one place our FONSIs run past the limits CEs typically state:
- Line length: our FONSIs exceed the CE limits (green). The bounded transmission FONSIs run a ~29-mile median (up to ~187 mi) against a ~2-mile typical CE limit (max 25 mi). So “bounded by action type” is not “small by the numbers a CE would use”: a CE written for these would need a much longer length bound, or none. This is the one dimension where these projects break the existing mold.
- Voltage & acreage: existing CEs already cover them (blue). On acres our projects (median ~20 ac) sit well within the existing CE range (median 70 ac, up to 10,000); voltage is comparable, though CEs rarely set a kV limit at all (only 2 do). A conventional CE limit would fit.
(Sizes here are a comparison input, not a filter — a FONSI with no stated size is still kept; it just can’t trigger expand.)
Step 6: Classify, then rank
First, classify each candidate with a three-question rule:
- Is there a close existing CE? No → develop a new one. (None here — every match clears the baseline.)
- Does a close CE exist, but our FONSIs exceed its stated size limit? Yes → expand that CE. (None here — no matched CE states a limit our projects exceed.)
- Does a close CE exist, but at a different agency than the ones doing these FONSIs? Yes → adopt it. (All four land here.)
So all four candidates classify as adopt. The figure below is the second step — ranking the four adopts so you know which to pursue first. Each gets a transparent score (0–1) summing six factors; the bar length is the priority and the colors show why:
- Novelty — a brand-new CE (develop) outranks expanding or adopting one.
- Volume — how many bounded FONSIs back it (more evidence ranks higher).
- Agency/state spread — how widely the action recurs (broader = more agencies benefit).
- Has size limits — whether we can state a numeric bound (easier to write the CE).
- Low mitigation dependence — less reliance on case-by-case mitigation (a cleaner CE).
- Profile candidate — a small bonus for being a core recurring subtype.
So transmission ranks first not because it matches a CE most closely (it doesn’t — solar does), but because it has the most FONSIs across the most agencies and states.
Main Finding: Adopt existing CE opportunities
Each of the four bounded candidate actions already has a categorical exclusion at another agency — none looks net-new (0 to develop) and none exceeds a stated CE limit (0 to expand), so all 4 resolve to adopt. Listed in recommendation priority (the Step 6 rank), which puts transmission first given its scale and reach:
- Transmission upgrades — adopt the routine-maintenance CE held by the Tennessee Valley Authority (
TVA #17), so BLM, DOE, NNSA, and the power-marketing administrations can stop running full EAs for in-corridor work. This is the highest-impact case: the most bounded FONSIs, the broadest agency and state reach. - Solar (CE-shaped subset) — adopt the small-solar CE held by the Department of Energy (
DOE-1--5-87), for bounded projects on previously disturbed land. - Geothermal exploration — adopt the exploration-drilling CE held by the Department of Energy (
DOE-1--3-43), for slim-hole / temperature-gradient wells. - Temporary resource assessment — adopt the site-characterization CE held by the Department of Energy (
DOE-1--3-43), for surveys, stream gauges, and borings (thin evidence — only 2 FONSIs).
| Action | Low-impact FONSIs | Existing CE (held by) | Agencies that could adopt it | States |
|---|---|---|---|---|
| Transmission upgrades within existing ROW | 37 | TVA---1-16 — Tennessee Valley Authority |
BLM, DOE, NNSA, PMA | 11 |
| Solar | 8 | DOE-1--5-87 — Department of Energy |
BLM, NNSA, PMA, USFS | 4 |
| Geothermal exploration | 7 | DOE-1--3-43 — Department of Energy |
BLM | 3 |
| Temporary resource assessment / site investigation | 2 | DOE-1--3-43 — Department of Energy |
NNSA, PMA | 2 |
| Thin-evidence flag: rows with fewer than ~5 low-impact FONSIs rest on very few cases — temporary resource assessment (n = 2) in particular — and should be treated as illustrative, not robust. | ||||
The adoption gap below shows the evidence weight behind each opportunity — how many bounded FONSIs are being run as full EAs when an existing CE would cover them.
The transmission case is the most geographically broad — the same in-corridor work runs through full EAs across 11 states, concentrated in the West.
Timing: do these EAs predate the FRA?
The 2023 Fiscal Responsibility Act (FRA, June 2023) gave agencies explicit authority to adopt another agency’s categorical exclusion — the exact move this analysis recommends. So the age of these EAs matters. Using decision dates merged from the D4 timeline, all but one of the dated bounded FONSIs were decided before the FRA — these agencies ran full EAs without the adoption shortcut now available. (Dates are known for ~⅔ of the bounded set; the rest are undated in the D4 output.)
This cuts two ways. It is a real limitation: the corpus is almost entirely pre-FRA, so we cannot see whether agencies have already started adopting these CEs since June 2023 — the recent picture may differ, and that bounds how current these specific opportunities are. But it also sharpens the recommendation: these agencies ran full EAs precisely because, at the time, the streamlined adoption path did not exist. Now that the FRA provides it, adopting the existing CE is exactly the efficiency the record points to. The clear next step is to refresh the corpus with post-FRA EAs and check whether the shift is already underway.
Worked example: transmission upgrades within existing right-of-way
The strongest-evidenced case: 37 bounded, low-impact FONSIs for in-corridor transmission upgrades (reconductoring, component replacement, minor rebuilds), across BLM, DOE, NNSA, and the power-marketing administrations in 11 states. The Tennessee Valley Authority already holds a CE for this class of work — TVA CE #17, “Routine modification, repair, and maintenance of, and minor upgrade of and addition to, existing transmission infrastructure…”
The table below lists sample FONSIs from different agencies, each with a verbatim excerpt from the cited record that shows the same in-corridor, modify-existing- infrastructure work that TVA CE #17 already covers (reconductoring, replacing aging structures, upgrading within the existing right-of-way). Each row names the source document (EA or FONSI) and page, and the project title links to the full record in the document explorer.
| Agency | Project title | Source | Excerpt |
|---|---|---|---|
| Western Area Power Administration | EA, p. 30 | “Western proposes to rebuild the Davis–Kingman Tap 69-kV Transmission Line by: • Removing the existing wood pole H-frame structures and conductors • Excavating for new structure foundations; including augering, drilling, blasting or installing special rock anchors …” |
|
| Bonneville Power Administration | FONSI, p. 1 | “Under the Proposed Action, BPA would rebuild the 26-mile-long transmission line, improve the access road system and foot-trail network, and remove trees and other vegetation that pose a danger to safely and reliably operating the transmission line. BPA would remove and replace 224 wood-pole transmission structures; realign segments of line miles two and three; replace wood pole structures with steel monopole structures in line mile five; replace existing …” |
|
| Bonneville Power Administration | FONSI, p. 1 | “BPA is proposing to replace aging and deteriorating wood pole structures and associated structural components on the existing Albany-Burnt Woods 115-kV No. 1 transmission line and along a portion of the existing Santiam-Toledo 230-kV No. 1 transmission line. Wood pole structures would be replaced along the entire length of the 26-mile Albany-Burnt Woods transmission line between BPA's Albany and Burnt Woods substations …” |
|
| Bureau of Land Management | EA, p. 12 | “Under the Proposed Action, BLM will approve the amendment to expand APS's ROW width from 15 feet to 65 feet, and APS will upgrade and replace its existing infrastructure on BLM land in a similar alignment to the existing one. The existing infrastructure includes steel-reinforced aluminum conductors, which have a maximum electrical-current carrying capacity of 505 amps.” |
|
| Bureau of Land Management | EA, p. 5 | “The proposed action is to grant a right-of-way to PG&E, authorizing the operation, maintenance, and termination of an existing 12 kV distribution power line on BLM administered land located in and Sections 30 and 32, T. 12 N., R. 23 E., SBM, Kern County, CA. The approximate length of the existing line is 5,647.11 feet and 10 feet in width. In addition, PG&E is proposing to remove and replace four existing power poles within Section 30.” |
|
| Bureau of Land Management | EA, p. 2 | “It is my decision to approve Tri-State's request for a variance from the approved right-of-way (ROW) amendment atthorization (COC-66840 and COC-63427) for the Montrose-Nucla- Cahone (lvINC) Transmission Improvement Project. Variance Request 12 will enable Tri-State to continue work on the Montrose-Nucla segment of the MNC transmission line with Environmental Protection Measures (EPM) AR-l suspended and during the winter closure period for elk (Cervus canadensis) and mule deer (Odocoileus hemionus) crucial winter …” |
(Size limits — line length, voltage, acreage, well count — now come from the LLM enrichment pass, which separates the disturbance footprint from study-area figures and keeps voltage distinct from length. Each value is a structured field backing the verdicts above, with its quote verified against the source.)
Analysis 2 — Mitigated FONSIs
Mitigated FONSIs imply that the project will have “no significant impact” if the applicant commits to some form of mitigation. This analysis systematically examines the language surrounding mitigated FONSIs to determine if there are patterns of actions that are routinely mitigated and if we can capture those recurring mitigations as a new CE.
How “mitigated” was determined
This is not an embedding or similarity model — it is the per-FONSI reading from the one-pass LLM enrichment (Claude Sonnet), the same pass used throughout the deliverable. For every one of the 451 FONSIs the model read the finding and the committed conditions and set two fields:
is_mitigated_fonsi— does the “no significant impact” finding depend on committed mitigation, or is the action inherently low-impact?mitigation_dependence—none/design_feature_only/case_specific_dependent/permit_or_consultation_condition/monitoring_only/unclear.
The supporting mitigation text is captured verbatim and 97%-quote-verified, and each FONSI is tagged with the resource areas its mitigation addresses (biological, water, cultural…) from a fixed vocabulary, so we can measure which recur.
This spans the whole corpus, not just the candidates
The mitigated pattern is corpus-wide: 310 of 452 (69%) decarbonization FONSIs are mitigated, and 95 of those fall outside the five Analysis-1 candidate types — so this is not a candidate-only artifact. By action type (including the large Other pool, which is 61% mitigated):
The dependence is overwhelmingly case-specific
Of the 310 mitigated FONSIs, 309 are case-specific-dependent — the finding rests on project-specific mitigation, not on the action being inherently low-impact.
One honest caveat.
is_mitigated_fonsiandmitigation_dependenceare near-equivalent judgments: the model almost always pairs “the finding depends on mitigation” with “that mitigation is project-specific.” So this is one consistent signal, not two independent confirmations. The FONSIs it reads as inherently low-impact (design-feature-only or none) are exactly the ones it flags as not mitigated. The substantive finding stands: when a FONSI leans on mitigation, that mitigation is tailored to the project — not a standard, transferable design feature. The word cloud below is the direct evidence for that.
Examples: the committed mitigation
So you can verify the read, here are verbatim mitigation summaries the model extracted (each tagged case-specific-dependent):
| Project Title | Committed mitigation (verbatim summary) |
|---|---|
| A Mitigation Action Plan is attached to the FONSI listing all committed mitigation measures. Measures include: coordinating construction activities (including helicopter use and tree removal) with Ankeny NWR and USFWS during Section 7 ES... | |
| A Mitigation Action Plan was prepared committing BPA and contractors to measures including: minimizing construction footprint especially in Forest Service and Corps habitat restoration areas, wetlands, and waterbody crossings; locating s... | |
| Mitigation measures include restricting construction to the minimum necessary work area, ongoing weed control on the ROW, installing bird flight diverters on the new line where it crosses the South Fork Snake River (SFSR) for avian prote... | |
| A detailed Mitigation Action Plan (MAP, Appendix D of the EA) was developed and incorporated into the FONSI, obligating BPA to implement all mitigation measures. Key measures include: instream work timing restrictions per WDFW requiremen... | |
| Mitigation includes: no surface occupancy within 2 miles of known sage grouse leks; avoidance of sage grouse nesting/brood-rearing/winter habitat within 0.6–2 miles of leks; reclamation of disturbed areas with perennial weed-free seed mi... |
The mitigation language is project-specific
Why is the dependence “case-specific”? Because the measures themselves are written per project — SHPO consultation here, raptor-nest buffers there, site-specific erosion plans elsewhere. In the most-frequent words across all 310 mitigated FONSIs’ summaries, no single term dominates:
That spread is the CE-design challenge made visible: a CE cannot copy any one project’s measures. What does recur is the resource area the mitigation protects — consistently biological, soils, water, and cultural — and those are what a defensible CE would distill into standing design criteria.
| Action | FONSIs examined | Mitigated | Mitigated share | Recurring resource areas |
|---|---|---|---|---|
| Geothermal exploration | 7 | 7 | 100% | soils_geology(6); biological(5); cultural(4); public_health(3); air_quality(3) |
| Temporary resource assessment / site investigation | 2 | 2 | 100% | biological(2); water(1); soils_geology(1); cultural(1); air_quality(1) |
| Transmission upgrades within existing ROW | 37 | 33 | 89% | biological(32); soils_geology(26); water(20); cultural(16); noise(8) |
| Solar | 8 | 5 | 62% | biological(6); cultural(4); soils_geology(3); other(2); visual(2) |
Boundary language: where agencies drew the significance line
Alongside mitigation, the enrichment captures the agency’s explicit significance-threshold statements — “would be significant if it exceeded X,” “an EIS would be required unless Y.” Those sentences are a CE’s natural numeric bounds. Each is a verbatim, span-verified quote; real examples:
| Project Title | Significance-threshold statement (verbatim) |
|---|---|
| Visual impacts would be significant if: Views to the Project area resulted in major visual contrast in sensitive or visually unique areas in proximity to high sensitivity viewers. | |
| This EA was prepared to determine whether the Burns & McDonnell's Proposed Project could cause significant impacts, which would require the preparation of an Environmental Impact Statement (40 CFR 1508.9), or, whether a Finding of No Significant Impact can be issued for the Burns & McDonnell's Proposed Project. | |
| I conclude the approved action will not result in significant impacts to the environment under the criteria in Title 40 CFR Section 1508.18 and 1508.27. | |
| High impacts could be considered significant impacts, if not mitigated, while moderate and low impacts are not. | |
| Public schools are located greater than 0.5 mile from the proposed action at Severstal Dearborn and as discussed in this EA the impacts of the proposed action, except for the regional air quality impacts discussed in Section 3.5, do not extend to such distances. |
These are candidate CE bounds straight from the agencies’ own findings — the thresholds a new or adopted CE would be written to stay within.
Analysis 3 — Review of existing federal Categorical Exclusions
This analysis reviews all 2,105 existing federal Categorical Exclusions across 78 agency units for patterns and insights.
The CE landscape, by department and agency
The concentration is stark — four of the 35 departments hold half of all 2,105 CEs:
Within those departments, the individual agencies (the four big departments in teal):
Numeric bounds are rare — and scattered
The single biggest fact about the existing catalog: only 86 of 2,105 CEs state an explicit numeric limit. The other ~2,019 bound the action qualitatively.
And the few that do carry a number are themselves scattered — acres and miles dominate, but every CE picks a different value. The single most common acreage limit (10 acres) is used by only 14 of the 71, and the stated limits run all the way from 1 to 10,000 acres. There is no common threshold to expand against:
So an “expand the stated limit” argument has almost no precedent to lean on — part of why nothing resolved to expand in Analysis 1.
What the non-numeric limits look like
Most CEs bound scope with words, not numbers — consistent in spirit but disjointed in form, each agency phrasing it differently:
| Limiter type | Example CE language (verbatim) |
|---|---|
| Routine / minor | (h) Routine or minor facility maintenance, custodial, and groundskeeping activities such as window washing, lawn mowing, trash collecting, and snow removal that do not involve environmentally sensitive areas (such as eroded areas, wetlands, cultural sites, or areas with endangered/threatened species). |
| Small-scale / limited | A4. Approving and issuing grants for social services, education and training programs, including but not limited to support for Head Start, senior citizen programs, drug treatment programs, and funding internships, except for projects involving construction, renovation, or changes in land use. |
| Temporary / short-term | (f) Supportive services that include health care and housing services, permanent housing placement, day care, nutritional services, collection of payment for services, short-term payments for rent/mortgage/utility costs, and assistance in gaining access to local, State, and Federal government benefits and services. |
| Within existing footprint | B2 Transportation of personnel, detainees, equipment, and evidentiary materials in wheeled vehicles over existing roads or jeep trails established by Federal, Tribal, State, or local governments, including access to permanent and temporary observation posts. |
How related are the CEs?
The CE body is highly clustered. Below, each point is one CE laid out by text similarity (t-SNE of the embeddings), colored by its k-means family and labeled by the family’s top terms. Distinct families emerge — equipment replacement, property acquisition, personnel/administrative, roads — and many recur across different departments, which is the precedent for adopt.
Cross-agency duplication — the precedent for adopt
The point: the same low-impact action is frequently excluded by several agencies under near-identical language. 317 CEs have a near-twin at a different agency, forming 130 shared “action families.” Adopting a peer’s CE — the exact move Analysis 1 recommends — is not unusual; it is how the landscape already works.
Each row below is one shared family (not a single example): the action, how many agencies use a near-identical CE for it, and a few of those agencies.
| Shared action | Agencies sharing it | e.g. |
|---|---|---|
| 3. Routine procurement of goods and services. | 8 | AFRH, DOD - DA, DOD - DAF, DOD - DLA, DOD - DON, DOD - DT... |
| (g) Normal personnel, fiscal, and administrative activities involving civilian personnel (recruiting, processing, paying, and records keeping). | 7 | AFRH, DOD - DA, DOD - DAF, DOD - DLA, DOD - DTRA, DOD - M... |
| (USDA-34d-USFS) Post-fire rehabilitation activities, not to exceed 4,200 acres (such as tree planting, fence replacement, habitat restoration, heritage site restoration, repair of roads and trails, and repair of damage to minor facilitie... | 6 | DOI - BIA, DOI - BLM, DOI - BOR, DOI - NPS, DOI - USFWS, ... |
| 34. Demolition, disposal, or improvements involving buildings or structures when done in accordance with applicable regulations including those regulations applying to removal of asbestos, PCBs, and other hazardous materials. | 5 | AFRH, DOD - DON, DOD - DTRA, DOD - MDA, NEH |
| (NR4) Preparation of policies, procedures, manuals, and other guidance documents for which the environmental effects are too broad, speculative, or conjectural to lend themselves to meaningful analysis and for which the applicability of ... | 5 | DOC - NIST, DOC - NOAA, DOC - NTIA, DOI, DOJ - FBI |
| *E2 New construction upon or improvement of land where all of the following conditions are met: (a) The structure and proposed use are compatible with applicable Federal, Tribal, State, and local planning and zoning standards and consist... | 5 | DHS, DOD - MDA, DOJ - FBI, NEH, TREAS |
The largest shared families are mundane (procurement, personnel, routine maintenance) — which is the point: when an action is genuinely low-impact, agencies converge on a shared CE. The decarbonization actions in Analysis 1 are simply ones where that convergence hasn’t happened yet.
Where net-new CEs would come from
Net-new CEs are empty here by construction: the candidate action types we hand-picked are all well-trodden, so all are already CE’d. But of the 452 decarbonization FONSIs, only 293 fell into those types — the remaining ~159 are other decarbonization actions we have not yet categorized. Net-new CEs most likely live there. The recommended next step is to cluster those un-categorized FONSI action descriptions to surface recurring categories nobody has hand-picked.
Caveats & next steps
- LLM-extracted, verified facts. Action definitions, numeric limits, mitigation dependence, and significance thresholds come from a one-pass enrichment of all 451 decarbonization FONSIs (Claude Sonnet); every quoted value was checked verbatim against its cited source page (97% verified). Values that did not verify are flagged, not shown.
- CE matches are a ranking aid, not verified coverage. Confirm each candidate CE against its eCFR text before acting on it — this is the one remaining verification step.
- Recommended next steps: (1) verify the 4 adopt matches against the eCFR CE text; (2) cluster the ~159 un-categorized decarbonization FONSIs to search for net-new CEs.
For CATF: tell us which adopt opportunities to pursue, and whether to launch the net-new search — and we fold your direction into the next iteration.
Reproduction
Run from the repository root in the nepa conda environment:
conda run -n nepa python phase2/code/deliverable06/_run.py
quarto render phase2/reports/deliverable06.qmd_run.py executes the numbered chain 01→08 (the 08_analyze.R figures step needs Rscript). Outputs are written to:
phase2/data/analysis/deliverable06/— analysis parquets (verdicts, mitigation, CE landscape/clusters)phase2/output/deliverable06/—figures/, the slim comparison table, andreview/drill-down tables
Draft generated 2026-06-26 | NEPA Analysis — Phase 2, Deliverable 6