fresh-fuchs
fresh-fuchs (the FUCHS project) builds a stochastic, risk-aware forest
landscape planning prototype on the femic-tsa29mini-instance bundle: a
~100,000 ha square subset of Williams Lake TSA29 (90,499.8 ha total;
35,083.0 ha managed land base after the proportional retention split).
The research design is a nested decision problem:
Outer problem (policy / administration). Landscape-scale policy on (1) target species / development-type composition by area share and (2) harvest policy — AAC as a function of development type and rotation age. Evaluated risk-sensitively on the distribution of outcomes using downside-risk measures (CVaR).
Inner problem (enterprise / implementation). Harvest scheduling and replanting decisions (which species to plant) maximizing NPV subject to the outer policy constraints, formulated as a Model I linear program.
The outer problem runs as a full Monte-Carlo: disturbance (fire) — and, later, wood-price — realizations are sampled, the inner LP is solved once per scenario (full foresight within a scenario), and each policy is evaluated on the resulting NPV distribution.
The package reuses the UBC-FRESH ecosystem rather than re-implementing it:
ws3— wood-supply engine and LP machinery (inner solver).femic— tsa29mini instance bundle, Woodstock bridge, BTC log-grade and FAN$IER yield/revenue-curve orchestration.fhops— harvest-cost estimation.nemora— diameter distribution fit and sampling for stochastic drivers.freshforge— workflow and matrix orchestration plus evidence manifests.fresh-salvage— economic calibration anchors (reference only).
Status: v0.1.0a1 (Phase 5 release). The end-to-end pipeline — model
build -> full-MC fire scenarios -> per-scenario inner LP -> NPV
distribution -> policy grid search -> CVaR-based ranking — is implemented,
validated, and tested, with freshforge orchestration and evidence. The
detailed master plan lives in planning/v0.1.0a1-plan.md in the
repository; ROADMAP.md is the condensed issue-tracker view.
Documentation Map
Page |
What it answers |
|---|---|
How do I get a working environment (including the external ws3 and femic source dependencies) and verify it? |
|
How do I go from a clean checkout to a first pipeline result, and where do the outputs land? |
|
What exactly does the model compute — the nested inner/outer formulation, the MC scenario engine, and known limitations? |
|
What does each command read, write, and exit with? |
|
How are the modules organized and what are the design invariants? |
|
How do I test, lint, extend the model, and get a PR accepted? |