Skip to content

Ecological EOH — where the numbers come from, and what is still missing

The ecological domain is the framework's largest open measurement debt: 41 of the 89 live placeholders sit in the land block. This page records what has been measured, what was checked and rejected, and what would close the rest.

Run it: eoh scenario run land_stewardship [--scope --allocation --amenity-weight]

The defect this addresses

ecological_eoh() used to take no area and no population — it returned base_rate / health and nothing scaled it. That made ecological the only domain with no extensive quantity behind it (personal scales with population, infrastructure with capital, knowledge with the corpus), which is why its share collapses as the system grows.

It is now keyed to area, via ecological_scale(). Spread over the land it is nominally the obligation for, the shipped anchor is 2.35 seconds per hectare per year — the defect in units that make it obvious. The form was fixed without moving the level: US_MAINLAND_HECTARES × ECOLOGICAL_INTENSITY_BASE reproduces ECOLOGICAL_BASE_RATE exactly.

Population deliberately stays out of the rate. Load per hectare does drive demand, but the model already carries that: ecosystem_health falls under load and the obligation divides by it.

The census, and what it reads

US, 2022 land use × 2025 workforce. Three measured inputs, one assumed mapping — which occupations steward which land class. That assumption is the weak link and is the first thing to attack.

land class area h/ha·yr
Forest-use land 250.1 Mha 0.182
Rural parks and wildlife — federal 71.6 Mha 0.161
Urban (at the declared amenity weight) 26.8 Mha 4.35
cropland, rangeland, state parks, misc 566.6 Mha unpriced

Coverage 38.1%; mean 0.498 h/ha·yr; 1.35× the anchor. Unpriced land is excluded, not costed at zero, so the mean is a lower bound over the part that has a costed path. Read coverage before comparing it to anything.

The one decision that moves the answer

Whether urban groundskeeping is ecological stewardship is worth 41×. Signed off at a declared weight rather than either corner: AMENITY_STEWARDSHIP_WEIGHT = 0.0468, from the amenity class's own occupational composition — canopy in (37-3013 maintains structure delivering three of the seven GUF services), turf out (37-3011, 1.24M workers, delivers none). The anchor is crossed at w* = −0.0229, so the weight sets the magnitude and not the sign — and more strongly than this passage used to say. It read w* = 0.0288, below the adopted weight, computed before the Phase-4b frame resolution; frame-consistently the crossing weight is NEGATIVE, meaning the census exceeds the anchor with the entire amenity class excluded. Both corners stay reachable via --scope.

Instruments checked and rejected

Four, and the failure mode moved each time — which is how the remaining gap got well-posed:

instrument verdict
NRCS EQIP payment schedules no time unit, no labour line in 2,691 rows; the dollar column mixes implementation cost with foregone income
Penn State extension budgets hours yes — but crop production hours for horticultural crops, not stewardship
Agency headcount, raw overstates 4.4× without the role mix
ASAE field capacity workswidth × speed × efficiency / 8.25, no price in the chain

The general rule these produced: a resolves_by that names a source without naming the field in it that carries the quantity has not been checked.

What is still missing

  • Cropland and rangeland, 416.9 Mha. Hours per treated hectare are now derivable from field capacity (cover crop: 0.303–0.844 h/ha·yr). The missing input is the adoption fraction — what share of cropland receives each practice — from the USDA Census of Agriculture. The unknown is bounded: even 100% adoption adds only 0.27–0.74× the current census, so cropland is not where the domain-balance gap is hiding.
  • State parks, 36.0 Mha. Needs NASPD state-park operating statistics.
  • Agency role mix band [0.2263, 0.4073]. Narrowed by a task decomposition inside series 0025 (park ranger) and 0456 (wildland fire); NPS budget justifications report FTE by activity.

Two cautions

These figures are US-only and do not transfer. Land-use composition, mechanisation and public land-management staffing all differ enough between jurisdictions that the intensities are jurisdiction-bound, the same way personal_basket's delivery productivity is bound to its agro-ecology.

NPS and FWS differ 5.3× in role mix (10.1% against 53.6%). NPS is a visitor-services organisation standing on land; FWS refuges are a land-management organisation. Do not quote a single federal stewardship rate.