Mosaic agriculture in northern WA

Page last updated: Tuesday, 18 January 2022 - 12:50pm

Please note: This content may be out of date and is currently under review.

Risk of agricultural plants becoming environmental weeds

The viability and sustainability of agricultural systems, including pastoral enterprises, may be improved by the introduction of non-indigenous (non-native or exotic) plant species. However there is clear evidence that throughout Australia, some introduced non-indigenous species have established in non-target areas and have become environmental weeds. As a result, there is concern amongst the community that agricultural plants grown for irrigated or dryland fodder and pasture may become environmental weeds.

To directly address this issue, a series of field trials were established in the West Kimberley and Pilbara to evaluate the persistence and/or spread of a wide range of agricultural species. The results of these trials will inform policy on balancing agricultural potential with environmental risks. Since most of northern WA is under pastoral lease, a diversification permit from the Pastoral Lands Board (PLB) is required to grow non-indigenous species.


A focus of DPIRD’s Northern Beef Development project is to inform weed risk policy settings and strengthen the weed risk assessment process. Prior to this research, the weed risk assessments were based on a desk-top analysis as there was little or no data on many of the agricultural species in the Pilbara and Kimberley.

The research project collected field data from a series of ‘weed risk’ field nurseries in four key environments (climate, soils) across the west Kimberley and Pilbara (Table 1). The field nurseries were established under irrigation in 2015-16 to simulate the worst case scenarios such as plants establishing and setting seed following a tropical cyclone and, an extremely high rainfall year.

Table 1 Summary of site details for the established field nurseries


Location (Station)

Native vegetation

Soil Type

Long-term annual rainfall (mm)

Date established


Derby (Birdwood Downs)

Cleared pasture

Red-brown sand (Pindan)


April 2015


La Grange (Wallal Downs)

Native vegetation (Pindan)

Red-brown sand (Pindan)


April 2015


Woodie Woodie (Warrawagine)

Native vegetation (spinifex)

Red-brown loam


Sept 2015


Fitzroy Valley (Gogo)

Cleared – dryland cropping

Grey-black cracking clay (Vertisol)


April 2016

There were two replicated trials at each site. The grass field nursery trials had 23 entries that included a range of warm season (C4) annual and perennial grasses such as Rhodes, panic, buffel and a selection of hybrid sorghums and millets. Each entry had plus and minus (+/-) fertilizer sub-treatments and each combination was replicated three times.

The legume trials had 23 entries and included tropical legumes such as lablab, cowpea, siratro and butterfly pea, a range of stylos and the fodder shrub leucaena, plus the temperate legume lucerne. Each entry had +/- fertilizer and +/- rhizobia sub-treatments (i.e. the seed being inoculated with the correct Rhizobia).

The sites were cleared, cultivated and weed-free before the treatments were sown as 3m rows and irrigated, to assist seedling establishment. The sites were kept largely free of ‘weeds’ (i.e. non-sown plants) to provide the best possible conditions for growth and seed production. Once the plants were well established (six to 12 months), the irrigation was gradually turned off and there was no further control of non-sown species, either native or naturalised. The persistence and/or spread was regularly assessed from 2016 through to 2020.

Key results

There was generally good to excellent establishment across all sites. There has been a similar pattern of response across the four sites. After the irrigation was turned off, the original vegetation has re-colonised much of the site including many rows of sown species. For example, at the La Grange site (Wallal Downs) the native vegetation, in particular poverty bush (Acacia stellaticeps) and spinifex (Triodia spp.), rapidly re-colonised the site (see photos below).

Legume nursery under irrigation in November 2015
La Grange (Wallal Downs) legume nursery under irrigation in November 2015
General site view in May 2017 showing the re-colonisation by the native vegetation.
General view of La Grange site (Wallal Downs) in May 2017 showing the re-colonisation by the native vegetation
WW Grass FN 2 May 2017 IMG_6302.JPG
Spinifex re-colonising the grass field nursery in the east Pilbara (Woodie Woodie)


The annual warm season grasses such as hybrid sorghum, millet and maize were very competitive when sown, however there was little recruitment after they senesced, despite large amounts of seed being produced. There are no (zero) warm season annual grasses remaining at any of the sites.

In the medium-term the majority of the perennial grass treatments have failed to persist (dead) or are struggling to persist. The notable exceptions across all sites were buffel (Cenchrus ciliaris) and birdwood grass (Cenchrus setiger) which is not surprising as they are naturalised in the Pilbara and Kimberley. Rhodes grass (Chloris gayana) can persist and spread under favourable conditions through its stoloniferous growth habit (i.e. runners), but has persisted poorly under dry conditions.


At each site, only a small number of legumes have shown fair to good persistence, either as perennials or through seedling recruitment.

While some of the tropical annual legumes like lablab and cowpea established and grew well across a number of sites, there have been few to nil recruits and they have failed to persist despite setting a large amount of seed. Lucerne has failed to persist at all sites.

Across a range of sites, the stylos have been the most persistent legumes. On the loamy soils at Woodie Woodie, the shrubby stylos (Stylosanthes scabra) are showing good persistence, while Caribbean stylo (Stylosanthes hamata) which behaves as a long-season annual has persisted through recruits at a number of sites.

Next steps

The department is reviewing the weed risk assessment process in conjunction with the results of this project. The experimental data are being prepared for publication in a department research bulletin.

Contact information


Clinton Revell
Geoff Moore