Restoration foresters increasingly plant seed sources whose home climate already matches a site's future climate rather than its current one — a practice called assisted migration. Each plot's home temperature is set by its position on an elevation gradient across the terrain (warmer in the valley, cooler on the ridge). The site's future condition is:
T_future(plot) = T_home(plot) + warming_slider
The chosen seed provenance carries its own climate optimum, offset from the *local* optimum by the provenance-shift slider (negative = colder/equatorward source, positive = warmer/poleward-adapted source):
T_optimum(provenance) = T_home(plot) + provenance_shift
mismatch = T_future(plot) − T_optimum(provenance)
P(establish) = exp( −mismatch² / (2·σ²) ), σ ≈ 2.2 °C niche width
Each planted tree draws a Bernoulli outcome from P(establish): a "thriving" tree grows tall and green, a "stressed" one grows short and yellowed, a "failed" one shrinks and browns. Genetic diversity is modelled as rising with provenance distance (more source populations mixed in) but is penalised once mismatch exceeds the niche tolerance, since a maladapted cohort mostly dies rather than adding usable diversity. Years-to-canopy-closure is an inverse function of mean survival — a stand near 100% survival closes canopy in roughly a decade, a badly mismatched planting can take 40+ years or never close at all.
- Provenance shift — 0 km is the historically "correct" local seed source; sliding right simulates sourcing seed from further away/warmer sites, pre-adapting the planting to future warming.
- Warming slider — how much the site's climate is projected to shift by 2070; higher values punish a local-only provenance and reward a forward-shifted one.
- Plant restoration grid — resamples every plot's outcome and regrows the instanced stand from bare ground.