Buffel grass (Cenchrus ciliaris L.) has invaded extensive areas of arid and semi-arid Australia following its introduction as a pasture species and for erosion control. It has been suggested that buffel grass has initiated a positive fire-invasion feedback in central Australia, disrupting existing fire regimes, encouraging further buffel grass invasion, and disadvantaging the native woody flora in particular, but this hypothesis has not been tested quantitatively. This study investigated recently burnt woodland areas near Alice Springs for evidence of a fire-invasion feedback, including the impact of changing fire behaviour (intensity) on the native woodland overstorey flora. Despite the limitations inherent in a short study of ecological processes in a highly heterogeneous environment, substantial field evidence was found to support the existence of a buffel grass-initiated fire-invasion feedback. Buffel grass invasion was significantly correlated with increased fuel loads. Increased fuel loads were significantly correlated with increased burn severity, although the direct relationship between the proportion of buffel grass and increased burn severity was marginally non-significant. High field variance resulted in inadequate power to test whether or not the relative abundance of buffel grass had increased in the post-fire community. Burn severity was significantly correlated with the mortality of woodland overstorey species, and with the proportion of fire survivors that were reduced to basal resprouts. Seedling density of canopy species was low. It appears likely that future recruitment of canopy species will be hindered by the dense post-fire reestablishment of buffel grass cover at some sites. The overstorey flora is thus likely to be adversely affected by increased severity of fire associated with buffel grass invasion. As a result, there may be major change in the structure and composition of some woodlands.

No abstract

Range condition assessment procedures that rely on field—collected botanical data face major problems in nonequilibrium rangelands, which are spatially variable and extensively grazed. These problems include the difficulty of interpreting changes in plant species composition and the logistics of obtaining representative data for large areas. Consideration of ecosystem behavior through time and in space shows that certain spatial and temporal patterns exist that may be used to isolate the impact of grazing from other processes. The patterns also make it possible to distinguish between temporary changes and those that are more long term. All relevant patterns may be expressed in terms of total plant cover and may be monitored from remote sensing satellites. It is therefore possible to derive a set of range condition indicators that may be measured and monitored from space. These indicators use trends in average vegetation cover with distance from water at the end of very wet periods, trends in cover variance with distance from water, and the magnitude of observed vegetation response at individual points in the landscape compared with that which is expected when vegetation recovers fully from grazing. It is also possible to use spatial variability in the rate of cover depletion after rainfall to infer relative differences in the amount of forage present. When used in combination, the methods offer a realistic alternative to field—based assessment and are capable of detecting many types of rangeland degradation. They are also considerably cheaper to use.

No abstract

No abstract