Gypsum blocks were used to monitor soil water under 5 rotationally grazed pastures [Eragrostis curvula cvv. 4650, 4660 and Consol; Medicago sativa cv. Aurora and Phalaris aquatica (a 50/50 mixture of cvv. Uneta and Sirolan)] at depths of 15, 30, 60 and 120 cm at Wagga Wagga in southern NSW, Australia from 18 December 1993 to 23 November 1998.

At 120 cm depth, soil matric potential (ψ) varied from approximately –14 kPa in winter and spring, to less than –1500 kPa in autumn. During 2 droughts the soil remained drier than its water holding capacity limit, but when rainfall was average or above average, there was no evidence that deficits were maintained through winter. The nature of the rainfall (rainfall/day and numbers of consecutive wet days), which varies through the seasons, impacted markedly on ψ, with considerable short-term variation observed in response to isolated heavy rainfall events in summer and autumn.

Medicago sativa cv. Aurora generated lower ψ than E. curvula in mid- to late-spring; whereas except for 4650, which was not persistent at the site, E. curvula reduced ψ in early to mid-summer, and to generally more negative extremes than by M. sativa. There was no indication that the roots of P. aquatica did not extend to the same depths as the other pastures, but it seemed unable to dry the soil to the same extent. Summer-active weeds, which were prominent in the P. aquatica pasture in years of above average rainfall, reduced the deficit difference between it and the other pastures.

Gypsum block ψ was converted to volumetric soil water content (θ_v) using soil water retention curves and estimates of soil bulk density (ρ_b). Relative to P. aquatica, E. curvua Consol and M. sativa maintained a deficit of between 25 and 75 mm for most non-winter periods. There was no evidence that the deficit created by either pasture was consistently greater than the other.

There was considerable dynamism within the datasets for all the pastures, which was ascribed to the frequent occurrence of non-matrix flow, mainly in response to heavy rainfall events (>25 mm in <5 days) that occurred when the soil was dry in summer and autumn. In the long-term it was estimated that such events could be associated with up to 25% of the rainfall.