Climate variability is a major constraint to farming in south-eastern Australia and one that is out of the farmers' control. However, a better understanding of long-term climate variability would be beneficial for on-farm management decisions. A series of long-term simulations were undertaken with the GrassGro decision support tool to determine the effect of climate variability on pasture and animal production at 6 locations in south-eastern Australia. The simulations ran from 89 to 119 years using daily weather records from each location. All simulations were for spring-lambing flocks of medium sized Merino ewes stocked at above-average district stocking rates, grazing well-fertilised, perennial grass–subterranean clover pastures. Annual rainfall total and, in particular, the distribution of rainfall during the year, were found to be more important than other weather variables in determining the amount of pasture grown in a year. The timing of the season opening rains (autumn break) was most important. The localities varied in their responses to climate variability, particularly in the timing of the autumn break; the pasture growth response to winter rainfall; and the relationship between rainfall and animal production.

Improved dryland pastures for sheep and beef cattle production in south-western Victoria are typically based on summer-dormant cultivars of perennial ryegrass (Lolium perenne L.) or phalaris (Phalaris aquatica L.). These are highly productive in spring but exhibit low accumulation rates over summer–autumn. Summer-active perennial pasture species could potentially alleviate this summer–autumn feed gap. Three pasture systems that used different pastures on each of the three landscape classes (crest, slope, and valley floor) were compared over 4 years. The perennial ryegrass system (henceforth Ryegrass) had a different ryegrass cultivar on each landscape class. The Triple system used lucerne (Medicago sativa L.) (crest), perennial ryegrass (slope), and summer-active tall fescue (Lolium arundinaceum (Schreb) Darbysh.) (valley floor). The Novel system used chicory (Cichorium intybus L.) (crest), Italian ryegrass (Lolium multiflorum Lam.) or hybrid ryegrass (L. × boucheanum Kunth.) (slope), and kikuyu (Pennisetum clandestinum Hochst. ex Chiov.) (valley floor). The pastures were grazed by either one (in the case of the Novel system) or three (in the case of the Ryegrass and Triple systems) animal systems that varied over the life of the experiment. Total annual herbage accumulation of the Ryegrass and Triple systems did not differ. The Novel system consistently had lower total annual dry matter accumulation than the other two systems. Lucerne pastures generally had the highest accumulation rates over summer, followed by the chicory pastures. The kikuyu pastures responded well to summer rainfall but otherwise had similar accumulation rates to the perennial ryegrass and tall fescue pastures over summer. Tall fescue pastures grew well in autumn following wet summers. In spring the perennial ryegrass pastures based on Fitzroy or Avalon were highly productive but seldom grew faster than other pastures. The results support the hypothesis that incorporating deep-rooted, summer-active perennial species will increase pasture production over summer–autumn compared with conventional pasture systems but not at the expense of winter–spring production.

Lucerne (alfalfa; Medicago sativa L.) is the key forage for grazing in dryland temperate regions around the world. While rotational grazing of lucerne is recommended, in southern Australia the intervals between grazing events are often chosen in an opportunistic manner, to meet livestock production targets and utilise excessive spring and summer growth. To assess whether the persistence of lucerne is sensitive to variations in rotational grazing management practice, we report on an experiment with four sheep grazing treatments that was conducted for 2.5 years, including three summers, in southern New South Wales. The grazing management treatments were a crash-grazing control, frequent grazing, feed-based rotational grazing and time-based rotational grazing, replicated four times. The number of grazing events, percentage of time under grazing, lucerne top dry matter (DM) at the beginning and end of grazing periods and plant density were measured. The results relating to number of grazing events, percentage of time grazing and DM removed during grazing indicated that four grazing practices had been achieved. The treatments all had significant periods of rest for at least 73% of time and were empirically different in their approach but resulted in similar grazing pressures, in terms of overall pasture removed during grazing. Nevertheless, there was little difference in lucerne densities between grazing treatments over the life of the experiment. We conclude that there is flexibility in the rotational management of grazed lucerne provided adequate rest periods are part of the management program.