Seed rain
Mortality
Fire
Scenarios
Camera
Playback
Quality
Model checks
The science
What is simulated
A forest stand of 48 × 30 patches, each holding at most one tree. Two kinds of trees trade off dispersal against competitive strength: pioneers (birch-like: many seeds, fast growth to 20 m, short-lived) and climax trees (spruce-like: few seeds, slow growth to 35 m, shade-tolerant, long-lived). Each year every patch can be colonised, displaced or emptied at random, following the competition–colonisation model of Tilman (1994): dpC/dt = cC pC (1 − pC) − (mC + g) pC and dpP/dt = cP pP (1 − pC − pP) − (mP + g) pP − cC pC pP. Climax trees can take over pioneer patches (the last term); pioneers can only use empty ground. c are seed-rain rates, m mortalities and g the storm-gap rate. The equilibrium is pC* = 1 − (mC+g)/cC, pP* = 1 − pC* − (mP+g)/cP − (cC/cP) pC* (zero if negative).
Growth, seed dispersal and fire
Tree height follows the Chapman–Richards curve h(t) = Hmax (1 − e−rt)b with b = 1.6; individual trees get a random ±15 % size factor. The sliders mix well-mixed seed rain with local rain from the 24 nearest patches. A lightning ignition lands on a random patch and the fire spreads to each of the 8 neighbours with probability dryness × flammability (pioneers 0.75, climax 0.38, open ground 0.12); burned patches become open ground and are recolonised first by pioneers, then replaced by climax trees: fire-driven succession.
Independent numerical checks
(1) The closed-form Chapman–Richards height is compared with a separate fourth-order Runge–Kutta integration of its growth ODE, du/dt = r b u(b−1)/b (1 − u1/b) with u = h/Hmax. (2, 3) A well-mixed, fire-free 80 × 80 stand is simulated for 500 years and its average cover over the last 175 years is compared with the analytic equilibrium above, for the current sliders and for a storm-gap rate g = 0.03.
Honest simplifications
All rates, heights, flammabilities and the 20 m / 35 m size limits are illustrative, not calibrated to a real forest type. Real forests have many species, resource-limited growth and light competition between neighbours, seed banks, soil and climate, insects and disease, and fire that depends on wind, slope, fuel moisture and weather; none of that is modelled. Fire spread is not checked against an analytic result (it is a stochastic cascade), only its qualitative role is shown. A patch holds one tree, so crowns do not overlap in the model although they do in the drawing. The mean-field equilibrium matches only the well-mixed, fire-free case; local seed dispersal and fire change the outcome, which is intended.