Aim: Exotic perennial grasses are significant invaders of native grassy communities and frequently multiple species invade communities, some from nearby agricultural areas.  There is little understanding of the landscape distribution of many species, making prioritisation for control a difficult decision.

Location: New South Wales, Eastern Australia

Methods: We undertook field surveys of exotic perennial grasses at 139 sites from nine grassy threatened ecological communities across four regions and assessed whether the profiles of exotic species varied amongst regions and communities.  We used a ranking of invasion risk based on plant characteristics to identify exotic perennial grasses that were likely to be the most invasive and then tested whether this ranking predicted the level of invasion measured in the survey.

Results: Using multivariate analysis we found that the threatened grassy communities surveyed were significantly invaded by exotic perennial grasses and that these assemblages were regionally distinct and distinct for most plant communities. Five widespread invaders were particularly established in all regions and communities, but regions also had distinct sets of invaders contributing significantly to degradation. Invasion by trade-off species was the most significant threat to grassy communities in all regions.  We showed that species with higher risk rankings based on plant characteristics were recorded in more sites but there were a few grasses that were more invasive than their ranking predicted.

Main conclusions: Our findings indicate that management of grassy plant communities for exotic perennial grasses should be undertaken at the community level although there are a suite of species that are important invaders in the whole landscape where improved understanding of pathways of invasion are needed for management across regions.  We identified a set of species which are important invaders but are not a focus in management currently, largely because many of these are species used in pastures.  Our study illustrates that higher levels of invasion were associated with species that were ranked more invasive on plant characteristics and this ranking could be used to initially allocate priorities for management of threatened plant communities. Trade-off species remain the major cause of degradation and must be included in discussions of regional conservation.

Nine ecological communities with prominent grassy understories were surveyed across the four climatic regions (Table 1, Fig. 1). Two threatened ecological communities, Themeda Grasslands on Seacliffs, and White Box Yellow Box Grassy Woodlands occurred across two regions, and we split the surveys for each of these communities into each region (Table 1, Fig. 1). All other communities occurred in a single region. These communities are of conservation significance as they are listed as threatened under state or federal legislation (Environment Protection and Biodiversity Conservation Act, 1999; Biodiversity Conservation Act No 63, 2016). As each ecological community varies in size and geographic range the number of surveys per community differed to ensure each was adequately represented (Table 1). The sites chosen were a mix of reserve land, travelling stock reserves, cemeteries, significant roadside vegetation, other uncleared crown land (owned and managed by state government) and agricultural land. We used government data, government officials who look after these communities, and other organisations (Friends of Grasslands, 2021; Grassy Ecosystems Conservation Management Networks, 2010) to find sites. Survey data from external projects in the north coast (J Hunter, personal communication, 9 sites) and south coast (J Miles, personal communication, 3 sites) within Themeda grassland sites were included in our analysis. In total, data was collected for 139 sites across the nine communities. The number of sites for each community varied due to availability and access to these communities and their rarity.

Surveys were completed over 2018-2020 during drought which may have decreased grass cover in habitats, particularly away from the coast. Thus, measurements of percentage cover of grasses were deemed inaccurate and were replaced by measurements of occupancy. To ensure coverage of the site we measured the occupancy of EPGs in plots placed along multiple transects through the site. The number of transects at a site increased as site size increased until a threshold size was reached (20 ha) with number of plots varying from 3 – 90 (average 22) according to site size (See Appendix S2 in Supporting Information). Linear transects varied between 50-800 m long depending on site size and shape. In each plot the presence of every grass species, native or non-native, was recorded in a10m x 2m area. The number of plots where a species was present divided by the total number of plots surveyed at a site was calculated as the proportion of occurrence. Shared survey data at both locations gathered species cover (%) in plots within a site, allowing equivalent proportion of occurrence to be calculated (Miles J, personal communication, Hunter J, personal communication). Survey data was taken from the most recent year i.e., 2017/18 for north coast sites, and 2016 in the south coast sites.