‘MaxClover’ grazing experiment: I. Annual yields, botanical composition and growth rates of six dryland pastures over nine years

Six dryland pastures were established at Lincoln University, Canterbury, New Zealand, in February 2002. Production and persistence of cocksfoot pastures established with subterranean, balansa, white or Caucasian clovers, and a perennial ryegrass‐white clover control and a lucerne monoculture were monitored for nine years. Total annual dry‐matter (10.0–18·5 t DM ha −1 ) and sown legume yields from the lucerne monoculture exceeded those from the grass‐based pastures in all but one year. The lowest lucerne yield (10 t ha −1 yr −1 ) occurred in Year 4, when spring snow caused ungrazed lucerne to lodge and senesce. Cocksfoot with subterranean clover was the most productive grass‐based pasture. Yields were 8·7–13·0 t DM ha −1 annually. Subterranean clover yields were 2·4–3·7 t ha −1 in six of the nine years which represented 26–32% of total annual production. In all cocksfoot‐based pastures, the contribution of sown pasture components decreased at a rate equivalent to 3·3 ± 0·05% per year ( R 2  = 0·83) and sown components accounted for 65% of total yield in Year 9. In contrast, sown components represented only 13% of total yield in the ryegrass‐white clover pastures in Year 9, and their contribution declined at 10·1 ± 0·9% per year ( R 2  = 0·94). By Year 9, 79% of the 6.6 t ha −1 produced from the ryegrass‐white clover pasture was from unsown species and 7% was dead material. For maximum production and persistence, dryland farmers on 450–780 mm yr −1 rainfall should grow lucerne or cocksfoot‐subterranean clover pastures in preference to ryegrass and white clover. Inclusion of white clover as a secondary legume component to sub clover would offer opportunities to respond to unpredictable summer rainfall after sub clover has set seed.