Low-vegetation growth
The P4.2 growth model is a fast, deterministic, segment-level saturating model. It grows height, cover, and density from the current state distribution, applies site, climate, and competition modifiers, and preserves uncertainty and age. It does not apply a treatment transition.
Model form
For each functional group:
next = asymptote - (asymptote - current) * exp(-rate * modifier * years)
Height and cover have separate asymptotes, rates, and uncertainty. Density is grown only when the current distribution has a density and the group has a density asymptote. The modifier is:
site_factor * climate_factor * competition_factor
Site factor uses
site_index_m / referenceand is clamped.Climate factor is bounded by the configured domain.
Competition factor is
1 - competition * competition_sensitivity.
Height and cover are monotone during growth and never exceed their asymptotes. A state already at the asymptote does not shrink.
from brushline.vegetation.growth import (
GrowthConfig,
GrowthEnvironment,
GrowthRequest,
grow,
)
from brushline.vegetation.state_space import StateDistribution
current = StateDistribution(
class_probabilities={"shrub": 0.8, "unknown": 0.2},
height_m=2.0,
height_sd=0.2,
cover_pct=30.0,
cover_sd=5.0,
density_per_ha=1000.0,
age_years=3,
confidence=0.8,
)
result = grow(
GrowthRequest(
segment_id="s1",
functional_group="shrub",
current=current,
years=3,
environment=GrowthEnvironment(site_index_m=24, climate_factor=1.1, competition=0.2),
),
GrowthConfig(),
)
assert result.status == "grown"
Functional groups
The parameters are group-specific priors:
Group |
Height max |
Cover max |
Height rate |
Cover rate |
Notes |
|---|---|---|---|---|---|
|
4 m |
90% |
0.35 |
0.35 |
Cutting can increase density. |
|
10 m |
95% |
0.25 |
0.30 |
Resprouting hardwood. |
|
20 m |
90% |
0.12 |
0.20 |
Young conifer or mature edge. |
|
0.6 m |
100% |
0.90 |
0.80 |
Herbaceous, not woody cover. |
|
n/a |
n/a |
n/a |
n/a |
Blocked. |
Domain and result states
A growth result is:
grownwhen the environment is in domain and the group is known;out_of_domainwhen the environment is explicitly flagged or site index/climate lies outside the configured domain;unknownwhen the functional group is unknown.
An out-of-domain or unknown result has no next distribution and a reason. It does not extrapolate or silently hold the state.
Output
GrowthResult carries:
previous and next distributions;
the next state class;
height and cover change;
environment context;
model card id and reason.
to_vegetation_state(state_year) converts a grown result to a valid brushline.state.v1 state.
Named errors
Prefix |
Meaning |
|---|---|
|
Unknown growth status. |
|
A grown result has no next distribution, or a blocked result has one. |
|
Height or cover decreased during growth. |
|
Height or cover exceeds its asymptote. |
|
The next class has no probability. |
|
Age did not advance by the step. |
|
A blocked result has no reason. |
Limits
The parameters are priors, not calibrated values; P4.6 validates them.
The model is not a physiological water/carbon model.
It does not apply a method-specific treatment transition; P4.3 owns that.
The model card’s applicability domain is enforced here and published in P4.7.