The use of a lucerne phase in crop rotations can reduce water lost as drainage past the root zone under dryland agriculture in southern Australia. During the lucerne phase the perenniality of lucerne and its deep rooting ability allow extraction of soil water from below the root zone of annual crops and the creation of a soil water storage buffer against deep water loss. The longevity of the soil water storage buffer depends on rainfall patterns, management of the crops and summer fallows, as well as the magnitude of the buffer created during the lucerne phase. Results from a previously reported field experiment in north-eastern Victoria (average annual rainfall 600 mm) suggested that a 2-year lucerne phase could be insufficient to prevent drainage under subsequent crops for more than 1 year.

Computer simulations were used to explore the implications of climatic variability on the creation and refilling of the soil water storage buffer. After first testing that the simulations described the experimental data satisfactorily, they were then used to extend the results and conclusions of the field experiment. These showed that the outcome of the experimental evaluation was affected by the climatic conditions experienced during the experiment and that a lucerne phase duration of 2 years was not appreciably less effective than a 3-year lucerne phase in reducing drainage past 1.8 m (the depth evaluated in the experiment). This conclusion was, however, sensitive to the depth at which drainage was evaluated and also depended on management factors such as the timing of lucerne removal and weed control during the summer fallows. For example, when drainage was evaluated to the maximum depth of lucerne rooting (3.6 m), lucerne was removed in December rather than April, and weeds were permitted, a third year of lucerne allowed a longer cropping phase without refilling of the profile in 47% of years. As a general recommendation a 3-year lucerne phase might, therefore, be an appropriate option for maximising the prevention of drainage. The large variability in the longevity of the soil water storage buffer (from 3 to > 45 months) and its sensitivity to management suggest, however, that it may be more beneficial to link phase changes to local assessment of the status of soil water storage buffer.