Loss of patch-scale heterogeneity on secondary productivity in the arid shrubland of Western Australia

General models of degradation suggest soil and nutrients are lost and conversion of rainfall into primary productivity is diminished when rangeland is degraded. These models are supported by studies on non-resilient landscapes, where loss of primary productivity also translated into loss of secondary productivity, but have not been tested on resilient landscapes. Elsewhere we showed that loss of chenopod shrubs from a landscape characterized as resilient was associated with declines in plant productivity and efficiency of conversion of rainfall into plant mass. To explore whether these differences in primary productivity translated into differences in secondary productivity, we grazed sheep at five rates of stocking for 10 years on 2000 ha of this landscape. The experiment was necessarily replicated in time not space (i.e. pseudo-replicated), which limits confident extrapolation of results to other landscapes. Productivity of sheep at all except highest rates of stocking varied little between sites where shrubs were abundant or scarce. From an animal production point of view, greater rates of stocking were unsustainable through dry years on the degraded site, but animal performance was generally unaffected on the non-degraded site, where shrubs were abundant. While these results provide evidence of economic penalties associated with degrading a resilient landscape, important ecological penalties were only partially explored.