A MODEL DESCRIBING THE INFLUENCE OF HARD SEEDEDNESS ON THE PERSISTENCE OF AN ANNUAL FORAGE LEGUME, IN A LEY FARMING SYSTEM, IN A MEDITERRANEAN-TYPE ENVIRONMENT

A deterministic, life-cycle model is described which predicts the long-term effects of hard seed levels on the density of annual subterranean clover (Trifolium subterraneum) plants established in various farming systems that rely on self-regeneration of the pasture after crop years. A stochastic form of the model incorporates the effects of variations between growing seasons, both on seed production and on the proportion of the soft seeds which produce established Plants. Model output is presented using parameters appropriate for a 350-400 mm rainfall environment in the southwest of Western Australia where rotations of 1-3 years of pasture followed by 1 year of crop (1:1 to 3:1) generally prevail. Although of value in indicating basic time trends, the deterministic form of the model can be seriously misleading, as it is unable to reflect effects such as poor pasture years in which few, if any, seeds are produced, The model shows that a clover is needed with hard seededness greater than that available in present cultivars for effective clover Persistence at useful plant densities in 1:1 rotations in this environment. Such a variety, for which about 40% of the seeds are soft at the end of the first summer, would also improve clover production, particularly in the pasture re-establishment year, in rotations with an extended pasture phase. The model also shows that rotations employing no-till cropping systems, which are not yet in general use in Western Australia, produce better clover densities than those from tilled systems in the early years of a newly established rotation. However, at higher levels of hard seededness, this apparent disadvantage of tilled systems eventually disappears through the development of a well-buffered soil seed reserve and this results in more stable clover densities between years. An assessment of the sensitivity of the model output to changes in the various model parameters is described, and strategies for improving clover performance in addition to increased hard seededness are discussed.