Data Sources & Scoring Cheat Sheet
100% public data · scoring is transparent, auditable, tunable
How scoring works (quick reference)
Hard eligibility filters (project is OUT if any apply)
| Filter | Why |
|---|---|
| Not mechanical thinning / biomass removal | Charm doesn't handle prescribed burn or mastication |
| DATE_ACCOMPLISHED > 2 yr ago (no follow-up) OR DATE_PLANNED > 12 mo out | Out of actionable piles-on-ground window |
| Inside Wilderness area | No mechanized equipment allowed |
| <50 acres | Not worth Charm's mobilization |
| Activity is pile-burn disposal | Piles already gone |
| NEPA signed >3 yr ago, no accomplishment | Stalled / shelfware |
| Within economic draw of active biomass-power / pellet mill | Carbon-credit eligibility requires "destined for pile burn otherwise" |
Core scoring signals
Haul30%Truck-route miles × 1.3 fudge to Fort Lupton, converted to $/ton delivered at $0.14/loaded-mile.
Biomass25%Green tons = acres × tons/acre by forest type × 0.6 recovery fraction.
Urgency20%WHP raster + WUI overlap on project footprint. Falls back to forest-type heuristic if raster unavailable.
Timing15%Pile perishability. Full credit for piles completed in last 6mo or planned next 6mo; fades after.
Contractor10%Known Charm-friendly CO forestry contractor in service radius.
Bonus flags (surfaced on card, don't drive base score)
- Metro burn-ban zone — inside 7-county Front Range ban (Nov 1 – Mar 31). Charm's alternative is legally constrained there → higher value.
- Ozone NAA — Front Range Ozone Nonattainment Area; additional smoke restrictions.
- Stewardship contract — contractor retains biochar credit; aligned incentives with Charm.
- Beetle-kill overlap — formerly Eagle Valley / Confluence Energy feedstock; now unserved since 2024.
- Tribal proximity — Southern Ute / Ute Mountain Ute — separate sourcing channel.
- Bundleable — 3+ nearby projects for single-contractor conversation (500 acres alone often doesn't pencil).
Data source inventory
Core datasets (in use)
| Dataset | Source | Cadence | Status |
|---|---|---|---|
| USFS FACTS — Hazardous Fuel Treatments The atomic unit of the tool. 28,000+ CO polygons. |
USFS EDW catalog · Live REST (layer 15) | Live / continuous | live |
| USFS Activity Project Areas (NEPA) The PLANNED pipeline — NEPA-signed, not yet accomplished. |
EDW REST | Live / continuous | live |
| USFS Wildfire Hazard Potential (WHP) National 270m raster; structural fire hazard. |
Firelab | ~5yr refresh | downloaded, rasterio env issue |
| USFS Admin Forest + Wilderness | USFS EDW | Annual | live |
| TIGER/Line roads + counties | Census Bureau | Annual | 2024 loaded |
| Front Range 7-county open-burn ban | CDPHE | Regulatory | derived from TIGER counties |
| EPA Ozone Nonattainment Area | EPA Green Book | Regulatory | derived from TIGER counties |
| CSFS forestry contractor seed list | CSFS regional offices | Hand-curated | 12 seed entries · cousin-in-fire expanding |
Core datasets (planned / in progress)
| Dataset | Source | Value |
|---|---|---|
| SILVIS WUI polygons | SILVIS Lab, UW-Madison | Wildfire urgency — structures-at-risk component |
| R2 Aerial Detection Survey — beetle-kill | USFS R2 FHP | Beetle-kill overlap bonus flag — formerly Eagle Valley feedstock |
| TreeMap 2022 (30m biomass raster) | Earth Engine · RDA | Replaces crude forest-type tons/acre heuristic with real stand-level biomass |
| Wildfire Crisis Strategy priority landscapes (2022) | EDW REST | Federal funding pipeline = higher project confidence |
| CWDG (Community Wildfire Defense Grant) awardees | EDW REST | County-level political pressure & cost-share availability |
| Joint Chiefs Landscape Restoration Projects | EDW REST | USDA + NRCS co-funded = most reliable pipeline |
Deferred / evaluated but out of scope
- SOPA quarterly reports — machine-readable version currently 404 (fs.usda.gov mid-migration). Click-through from tool still works once USFS site is back.
- BLM ePlanning — no public API; would require HTML scraping with headless browser.
- Regrid parcel data — paywalled; landowner outreach is a Phase 2 channel for Charm anyway.
- Post-fire salvage / burn severity polygons — Charm doesn't do standing-dead removal.
- Sawmill proximity as positive signal — mills don't compete for slash; only Confluence Energy Kremmling has a 50mi buffer that matters, and only for clean roundwood.
Transparency
Every threshold in this tool lives in one file:
pipeline/refs/charm_policy.py.
Haul curve, scoring weights, forest-type tons-per-acre, burn-ban FIPS,
contractor seed CSV — all inspectable, all tunable.
Questions or additions? ian@itogeospatial.com ·
source repo