loading...
running...
The science
What the model solves. One square metre of greenhouse floor with its air, a crop canopy and the climate controller. Every simulated 6 minutes it updates (1) the air temperature from an energy balance, (2) the CO2 concentration from a mass balance, (3) leaf photosynthesis in eight canopy layers, and (4) the growth of leaves, stem, roots and fruit from the assimilates. The 2D and 3D pages run the same code, so they show identical numbers.
Leaf photosynthesis (Farquhar, von Caemmerer and Berry, C3). The net rate is the smaller of a Rubisco-limited and a light-limited (RuBP regeneration) rate, minus day respiration:
A = min(Wc, Wj) - Rd
Wc = Vcmax (Ci - Γ*) / (Ci + Kc (1 + O/Ko)), Wj = J (Ci - Γ*) / (4 Ci + 8 Γ*)
θ J² - (I2 + Jmax) J + I2 Jmax = 0, I2 = Q · 0.85 (1 - 0.15) / 2
Vcmax, Jmax, Kc, Ko, Γ* and Rd follow an Arrhenius law with the activation energies of Bernacchi et al. (e.g. 65.3 kJ/mol for Vcmax); Vcmax and Jmax also fall above about 35 C through a peaked (deactivation) term. The intercellular CO2 is a fixed fraction Ci = 0.75 Ca (slider); there is no stomatal model, so water stress and humidity are not simulated. That is why the A-Ci graph has a Rubisco branch that rises with CO2 and a light branch that saturates, and why extra CO2 helps most on bright days and hardly at all in dim light.
Canopy. Light falls through the leaf area index L as Beer's law with extinction K = 0.7: the leaf at depth x sees Q0 K e-Kx / (1 - 0.2) (the 0.2 is leaf scattering) and Vcmax falls by e-0.18x. The canopy rate is the sum over eight layers of equal leaf area. PAR is 2.1 µmol per W/m² of global radiation.
Greenhouse energy balance (per m² floor). Heat capacity C = 50 kJ/(m² K) (air plus soil, water, plants), cover area 1.25 m²/m², air volume 4 m³/m²:
C dT/dt = 0.85 Iin - LE + Plamps + Qheat - (Ac U + ρ cp V n/3600)(T - Tout)
Iin is sunlight through the cover (single glass 85 %, double 76 %, film 80 %) and the shade screen; LE is a latent-heat share (half of the absorbed light times 1 - e-0.6L) taken by transpiration; lamps give all of their electrical power as heat (2.5 µmol per J); the heater is a proportional controller (80 W/m² per K); the vents open linearly over 2 K above the threshold, from the leakage rate to the maximum air-change rate n. The temperature is advanced with the exact exponential solution of this linear equation, so even the fast ventilation term stays stable.
CO2 balance. In ppm: dC/dt = n/3600 (Cout - C) + (Fdose - Acanopy + Rsoil) / 163.6, where 163.6 µmol of CO2 is one ppm in 4 m³ of air at 25 C and Rsoil = 1 µmol/(m² s). The dosing controller is proportional and capped; with the vents open nearly all of the dosed CO2 is blown out (the model shows why growers stop enriching then).
Growth and degree-days. Net assimilate goes into a pool (30 µg CH2O per µmol CO2), which is converted to dry matter at 72 % efficiency (the rest is growth respiration) at a rate limited by temperature, fT = (T - Tbase)/(Topt - Tbase). Development follows thermal time, GDD = ∫ max(0, T - Tbase) dt (capped at 32 C). For tomato flowering starts at 350 and ripening at 700 C d; fruit then takes a growing share of the assimilates (up to 62 %). Leaf area is leaf mass times a specific leaf area; surplus leaf area above the crop limit is removed, like pruning.
Checks in the page. (1) the initial quantum yield dA/dQ against its closed form; (2) the Rubisco-limited rate against a hand calculation; (3) light absorbed by the eight layers against Beer-Lambert; (4) an unheated house converging to the outdoor temperature, and the mean temperature with sun against the analytic balance; (5) CO2 rise in a sealed house against the dosed amount; (6) degree-days at constant temperature; (7) carbon balance between assimilated and grown mass. The 3D page also checks that the plants drawn match the model.
Honest simplifications. A single lumped air and mass temperature, no humidity, no condensation, no stomatal or water stress, no nitrogen, no pests; the same shape of crop is reused for three species with different parameters; the sun is a smooth daily curve with a deterministic cloud factor; plant size is drawn from mass with simple rules and the visible plant count is smaller than a real house. Use it to see how light, CO2, temperature and ventilation trade off, not to plan a real crop.