Federal land burns roughly 4-6x more than state/private land per acre owned — but the gap is mostly the arid West, not ownership itself

Summary

Federally managed land burns considerably more than state and private land, on every measure that adjusts for land ownership size — but a large share of that gap is really about geography, not ownership itself. Nationally (NICC/CRS data, 2018-2022): federal land accounts for only ~28% of total U.S. acreage, yet averaged about 64% of all acres burned annually over the last decade. Normalizing by ownership size, federal land burns roughly 4-5x more of its own acreage each year than nonfederal (state/local/private) land, and the average federal fire is roughly 6x larger than the average nonfederal fire. However, an AskAmerica state-level cross-check finds that most of this federal-land effect is explained by federal land being disproportionately concentrated in the arid West: once region (West vs. East) is controlled for, the state-by-state relationship between % federal land and burn rate is no longer statistically significant (partial r = -0.19, p = 0.20, vs. a raw r = 0.46, p<0.001 before controlling for region).

National comparison: federal vs. nonfederal acres burned

The Congressional Research Service's Wildfire Statistics report (IF10244), sourced from the National Interagency Coordination Center (NICC)'s official Wildland Fire Summary and Statistics annual reports, is the standard federal reference for this comparison. Its Table 1 (2018-2022):

Over 2018-2022: federal land burned an average of 24.0M acres across 63,500 fires (avg. fire size ≈ 378 acres); nonfederal land burned 14.3M acres across 232,000 fires (avg. fire size ≈ 62 acres) — federal fires average about 6.1x larger. CRS states the federal share of total acreage burned averaged 64% over the 2013-2022 decade, even though only 20-30% of fires by count start on federal land — nonfederal land accounts for the large majority of ignitions (70-80%) but federal fires, once they start, tend to be far larger, especially in the West.

Normalizing by how much land each group actually owns

Raw acreage-burned shares overstate the ownership effect unless divided by how much land each group holds. The federal government owns roughly 640 million acres (~28% of the ~2.27 billion acres of total U.S. land area, per CRS/BLM), leaving about 1.63 billion acres (~72%) as state, local, and private land.

Applying the CRS 10-year-average federal share of acres burned (64%) to the 10-year average total burned (7.2M acres/year, per CRS): federal land burns an estimated 4.6M acres/year against its 640M-acre base (a 0.72%/year burn rate), while nonfederal land burns an estimated 2.6M acres/year against its 1.63B-acre base (a 0.16%/year burn rate). That implies federal land burns at roughly 4.5x the rate of nonfederal land, measured as share of owned acreage burned per year.

AskAmerica corpus cross-check — and a real defect found along the way

AskAmerica's disasters.wildfire_perimeters table (NIFC WFIGS perimeters, ~10,000 fires loaded, 2018-2023) is documented as carrying per-fire ownership fields (landowner_kind, landowner_category, jurisdictional_agency) specifically so a federal-vs-nonfederal comparison could be run directly. On querying, none of these three columns actually exist in the live table — only the declared schema metadata lists them; a plain SELECT on any of the three fails with a binder error, while every other declared column works normally. This is reported as a data defect (see below) and blocks the direct per-fire national comparison from this corpus; the national comparison above instead rests on CRS/NICC's official tallies.

To still use the corpus for an independent check, we joined lands.county_federal_fee_land_share (a PAD-US-derived spatial estimate of % federal-fee land per county, rolled up to state, land-area-weighted) against disasters.wildfire_perimeters totals aggregated by state, normalized by each state's land area. Across the 50 states, the raw (zero-order) correlation between % federal land and the annual share of state land burned is positive and significant: r = 0.46, p < 0.001, n = 50. On its own, this appears to confirm the federal-land-burns-more finding at a finer grain.

But federal land is heavily concentrated in the 12 CRS-defined "Western" states (AK, AZ, CA, CO, HI, ID, MT, NV, NM, OR, UT, WA, WY), which are also the most fire-prone due to aridity and vegetation type — a classic confound. Controlling for West-vs-East region with a partial correlation collapses the relationship: partial r = -0.19, p = 0.20 — no longer statistically distinguishable from zero. In other words, at the state level, most of the apparent "federal ownership causes more burning" signal is actually "the arid West burns more, and the arid West happens to be where most federal land is" — the corpus data does not support a strong independent ownership effect once region is accounted for.

This does not overturn the national CRS finding (federal land does account for a disproportionate share of acres burned relative to its size) — federal land management practices, fuel loads, and fire suppression policy differences between agencies plausibly still matter. But it does mean the ~4-5x national ratio should be read as "federal land, which is concentrated in fire-prone terrain, burns more," not as clean evidence that federal management itself burns land at a higher rate than equivalent land under state or private ownership in the same region.

Exclusion disclosure: the 50-state correlation and the scatter chart above exclude the District of Columbia and the five inhabited territories (Guam, Puerto Rico, American Samoa, U.S. Virgin Islands, Northern Mariana Islands), which the underlying SQL filtered out with state_fips NOT IN ('11','60','66','69','72','78') and state_fips IS NOT NULL. Re-running the same query with DC included (n=51) changes essentially nothing: DC computes a federal-land share of 17.4% and zero recorded acres burned in the 2018-2023 window, so it sits at the origin of the burn-rate axis and would very slightly weaken, not strengthen, the raw federal-share/burn-rate correlation. No territory carries a comparable federal-fee-land estimate in this corpus, so they could not be added rather than being deliberately dropped for another reason.

Data quality issue reported

Filed as a defect: disasters.wildfire_perimeters's catalog description states the table's landowner_kind/landowner_category/jurisdictional_agency columns exist specifically to answer "federal-vs-nonfederal wildfire comparisons directly from this table," and describe_table lists them — but querying any of the three returns a binder error (column not found), while every other declared column (state_fips, county_fips, fire_cause, gis_acres, etc.) works. The WFIGS source fields (attr_POOLandownerKind, etc.) appear not to have actually been materialized into the live table despite being declared in the schema. Recommend re-running the ETL for this table to backfill these columns.

Every query behind this report

8 warehouse calls ran in this session, in order. Each is reproducible against the same snapshot.

SELECT incident_id, fire_year, state_fips, county_fips, poo_state, fire_cause, gis_acres, final_acres, type FROM disasters.wildfire_perimeters LIMIT 3

SELECT c.state_fips,
  SUM(f.federal_fee_land_share_pct/100.0 * c.land_area) / NULLIF(SUM(c.land_area),0) AS federal_share,
  SUM(c.land_area) AS land_area_sqm
FROM lands.county_federal_fee_land_share f
JOIN geo.counties c ON f.county_fips = c.county_fips AND c.year = '2025'
GROUP BY c.state_fips

SELECT state_fips, SUM(gis_acres) AS acres_burned, COUNT(*) n_fires
FROM disasters.wildfire_perimeters
WHERE state_fips IS NOT NULL
GROUP BY state_fips

SELECT MIN(fire_year) miny, MAX(fire_year) maxy, COUNT(DISTINCT fire_year) ny, COUNT(*) n FROM disasters.wildfire_perimeters

WITH federal AS (
  SELECT c.state_fips,
    SUM(f.federal_fee_land_share_pct/100.0 * c.land_area) / NULLIF(SUM(c.land_area),0) AS federal_share,
    SUM(c.land_area) AS land_area_sqm
  FROM lands.county_federal_fee_land_share f
  JOIN geo.counties c ON f.county_fips = c.county_fips AND c.year = '2025'
  GROUP BY c.state_fips
),
fires AS (
  SELECT state_fips, SUM(gis_acres) AS acres_burned, COUNT(*) n_fires
  FROM disasters.wildfire_perimeters
  WHERE state_fips IS NOT NULL
  GROUP BY state_fips
)
SELECT fe.state_fips, fe.federal_share, fe.land_area_sqm,
  COALESCE(fi.acres_burned,0) AS acres_burned,
  COALESCE(fi.acres_burned,0) / (fe.land_area_sqm * 0.000247105) AS burn_rate_pct_of_land
FROM federal fe
LEFT JOIN fires fi ON fe.state_fips = fi.state_fips
WHERE fe.state_fips NOT IN ('11','60','66','69','72','78')

WITH federal AS (
  SELECT c.state_fips,
    SUM(f.federal_fee_land_share_pct/100.0 * c.land_area) / NULLIF(SUM(c.land_area),0) AS federal_share,
    SUM(c.land_area) AS land_area_sqm
  FROM lands.county_federal_fee_land_share f
  JOIN geo.counties c ON f.county_fips = c.county_fips AND c.year = '2025'
  GROUP BY c.state_fips
),
fires AS (
  SELECT state_fips, SUM(gis_acres) AS acres_burned
  FROM disasters.wildfire_perimeters
  WHERE state_fips IS NOT NULL
  GROUP BY state_fips
)
SELECT fe.state_fips, fe.federal_share,
  COALESCE(fi.acres_burned,0) / (fe.land_area_sqm * 0.000247105) AS burn_rate_pct_of_land
FROM federal fe
LEFT JOIN fires fi ON fe.state_fips = fi.state_fips
WHERE fe.state_fips NOT IN ('11','60','66','69','72','78')

WITH federal AS (
  SELECT c.state_fips,
    SUM(f.federal_fee_land_share_pct/100.0 * c.land_area) / NULLIF(SUM(c.land_area),0) AS federal_share,
    SUM(c.land_area) AS land_area_sqm
  FROM lands.county_federal_fee_land_share f
  JOIN geo.counties c ON f.county_fips = c.county_fips AND c.year = '2025'
  GROUP BY c.state_fips
),
fires AS (
  SELECT state_fips, SUM(gis_acres) AS acres_burned
  FROM disasters.wildfire_perimeters
  WHERE state_fips IS NOT NULL
  GROUP BY state_fips
)
SELECT fe.state_fips, fe.federal_share,
  COALESCE(fi.acres_burned,0) / (fe.land_area_sqm * 0.000247105) AS burn_rate_pct_of_land,
  CASE WHEN fe.state_fips IN ('02','04','06','08','16','30','32','35','41','49','53','56') THEN 1 ELSE 0 END AS is_west
FROM federal fe
LEFT JOIN fires fi ON fe.state_fips = fi.state_fips
WHERE fe.state_fips NOT IN ('11','60','66','69','72','78')

WITH federal AS (
  SELECT c.state_fips,
    SUM(f.federal_fee_land_share_pct/100.0 * c.land_area) / NULLIF(SUM(c.land_area),0) AS federal_share,
    SUM(c.land_area) AS land_area_sqm
  FROM lands.county_federal_fee_land_share f
  JOIN geo.counties c ON f.county_fips = c.county_fips AND c.year = '2025'
  GROUP BY c.state_fips
),
fires AS (
  SELECT state_fips, SUM(gis_acres) AS acres_burned
  FROM disasters.wildfire_perimeters
  GROUP BY state_fips
)
SELECT fe.state_fips, fe.federal_share,
  COALESCE(fi.acres_burned,0) / (fe.land_area_sqm * 0.000247105) AS burn_rate_pct_of_land
FROM federal fe
LEFT JOIN fires fi ON fe.state_fips = fi.state_fips