Long‐term changes in semi‐arid vegetation: Invasion of an exotic perennial grass has larger effects than rainfall variability

Abstract. Questions: This paper examines the long‐term change in the herbaceous layer of semi‐arid vegetation since grazing ceased. We asked whether (1) there were differences in the temporal trends of abundance among growth forms of plants; (2) season of rainfall affected the growth form response; (3) the presence of an invasive species influenced the abundance and species richness of native plants relative to non‐invaded plots, and (4) abundance of native plants and/or species richness was related to the time it took for an invasive species to invade a plot. Location: Alice Springs, Central Australia. Methods: Long‐term changes in the semi‐arid vegetation of Central Australia were measured over 28 years (1976–2004) to partition the effects of rainfall and an invasive perennial grass. The relative abundance (biomass) of all species was assessed 25 times in each of 24 plots (8 m × 1 m) across two sites that traversed floodplains and adjacent foot slopes. Photo‐points, starting in 1972, were also used to provide a broader overview of a landscape that had been intensively grazed by cattle and rabbits prior to the 1970s. Species’abundance data were amalgamated into growth forms to examine their relationship with environmental variation in space and time. Environmental variables included season and amount of rainfall, fire history, soil variability and the colonization of the plots by the exotic perennial grass Cenchrus ciliaris (Buffel grass). Results: Constrained ordination showed that season of rainfall and landscape variables relating to soil depth strongly influenced vegetation composition when Cenchrus was used as a covariate. When Cenchrus was included in constrained ordination, it was strongly related to the decline of all native growth forms over time. Univariate comparisons of non‐invaded vs impacted plots over time revealed unequivocal evidence that Cenchrus had caused the decline of all native growth form groups and species richness. They also revealed a contrasting response of native plants to season of rainfall, with a strong response of native grasses to summer rainfall and forbs to winter rainfall. In the presence of Cenchrus these responses were strongly attenuated. Discussion: Pronounced changes in the composition of vegetation were interpreted as a response to removal of grazing pressure, fluctuations in rainfall and, most importantly, invasion of an exotic grass. Declines in herbaceous species abundance and richness in the presence of Cenchrus appear to be directly related to competition for resources. Indirect effects may also be causing the declines of some woody species from changed fire regimes as a result of increased fuel loads. We predict that Cenchrus will begin to alter landscape level processes as a result of the direct and indirect effects of Cenchrus on the demography of native plants when there is a switch from resource limited (rainfall) establishment of native plants to seed limited recruitment.