This paper describes the physiological basis and validation of a generic
legume model as it applies to 4 species: chickpea
(Cicer arietinum L.), mungbean
(Vigna radiata (L.) Wilczek), peanut
(Arachis hypogaeaL.), and lucerne
(Medicago sativa L.). For each species, the key
physiological parameters were derived from the literature and our own
experimentation. The model was tested on an independent set of experiments,
predominantly from the tropics and subtropics of Australia, varying in
cultivar, sowing date, water regime (irrigated or dryland), row spacing, and
plant population density. The model is an attempt to simulate crop growth and
development with satisfactory comprehensiveness, without the necessity of
defining a large number of parameters. A generic approach was adopted in
recognition of the common underlying physiology and simulation approaches for
many legume species. Simulation of grain yield explained 77, 81, and
70% of the variance (RMSD = 31, 98, and 46
g/m2) for mungbean (n =
40, observed mean = 123 g/m2), peanut
(n = 30, 421 g/m2), and
chickpea (n = 31, 196
g/m2), respectively. Biomass at maturity was
simulated less accurately, explaining 64, 76, and 71% of the variance
(RMSD = 134, 236, and 125 g/m2) for mungbean,
peanut, and chickpea, respectively. RMSD for biomass in lucerne
(n = 24) was 85 g/m2
with an R2 of 0.55. Simulation
accuracy is similar to that achieved by single-crop models and suggests that
the generic approach offers promise for simulating diverse legume species
without loss of accuracy or physiological rigour.
