Unlike the 3D version (where the core–surface ΔT is a slider you set directly), here ΔT is computed, not chosen: you set the colony's metabolic heat output and the outside air temperature, and the sim solves a coupled heat-balance / buoyancy-flow system every frame — the same self-regulating loop a real mound runs on.
ρ = P / (R·T) (ideal gas: hot nest air is less dense)
ΔP = (ρ_amb − ρ_in)·g·H (stack/buoyancy pressure)
Q = Cd·A_eff·√(2ΔP / ρ_avg) (duct-flow rate through the mound's channels)
Q_loss = ṁ·cp·(T_in − T_amb), ṁ = ρ_in·Q
dT_in/dt = (Q_gen − Q_loss) / (ρ_in·V_nest·cp)
Warm air rises up the central chimney and exits near the apex; that draft draws replacement air in through the peripheral channels near the base, which is what actually keeps the nest cool. Turn up metabolic heat generation and T_in rises — but a hotter core also means a bigger ΔP, so Q rises too, hauling more heat out and capping how far T_in can climb: a real negative-feedback thermostat with no moving parts. H = 3 m mound height, ρ ≈ 1.2 kg/m³, g = 9.81 m/s², cp = 1005 J/(kg·K), Cd = 0.62, A_eff the narrower of the chimney throat or the combined peripheral channels (they sit in series). This is a simplified single-loop thermosiphon model (Turner, 2001); real mounds add wind-driven mixing and diurnal wall heating/cooling, left out here to isolate the buoyancy-feedback term. The sim runs the heat balance at roughly 300× real time so the thermostat settles in seconds instead of tens of minutes; particle speed is further exaggerated for visibility.
- Metabolic heat generation — how much heat the colony + fungus garden produce; the true driver of the draft.
- Ambient temperature — outside air temperature; a smaller gap to the (regulated) core temperature weakens the draft for the same heat output.