Yields from annual and perennial pastures adjacent to tree belts are given, and the report recommends several perennial pasture species for use in agroforestry systems.

Who is the report targeted at?
The report is for use by farmers, agricultural advisors, farm forestry extension officers and researchers.

The results of this study should also inform agricultural policy and research and development (R&D) funding agencies.

Background
Research has shown that there can be considerable losses in the productivity of annual crops and pastures alongside trees in alley and windbreak systems. While these losses can be managed in some situations they remain a disincentive to producers considering adopting agroforestry systems.

However, work in Western Australia has shown that kikuyu pasture can be grown in the root zone of trees with no reduction in pasture productivity. This opens up the possibility of developing agroforestry systems based on perennial pasture species which incur no loss of forage production.

Aims/Objectives
The aim of the project was to identify perennial forage species that compete successfully with trees for water, nutrients and light in the medium to low rainfall zone in south-west Western Australia, and to assess what effect perennial pastures have on tree root morphology and water-use compared to annual pastures.

Methods used
Ten on-farm sites were located within the medium to low rainfall zone of the south coast of Western Australia. At each site, tree-pasture systems comprised of tree belts adjoining perennial or annual pasture were compared over a period of 12 months. Measurements included pasture growth, composition and root morphology, tree water use and root morphology, soil fertility and soil water content.

Results/Key findings
Pasture growth and composition
This study found that competition between trees and pastures consistently resulted in reduced annual pasture yield within the tree root zone, by up to 53% on average, confirming the results of a number of studies undertaken in southern Australia. There was no strong evidence that the reduction could be attributed to competition for soil nutrients. However, soil water content was less in the tree root zone compared to annual pasture alone suggesting that the principle mechanism reducing pasture growth was competition for water.

In contrast to annual pasture, the average decline in perennial grass pasture yield within the tree root zone was at most 10%. Furthermore, in comparison to the open annual pasture yield, perennial grass yield was similar or greater even within 0.5 tree heights of the tree belt, suggesting perennial grasses are more successful in competing with trees for moisture, or were less affected by periodic moisture stress. These results confirm that kikuyu growth is not reduced by tree competition and that tall wheat grass, veldt grass, rhodes grass and consol lovegrass are similarly competitive with trees in medium rainfall environments.

Previous research concluded that perennial forages were generally more competitive with trees because of their deeper and denser root systems. That does not appear to be the case for lucerne in this study where yields were reduced by tree competition, on average up to 63%. This may have been due to poor establishment when lucerne was sown into the dry soil alongside established trees or the reduced availability of phosphorus in close proximity to trees.

In this study it was difficult to draw any conclusions about the impact of tree competition on pasture nutritive value or botanical composition.

Tree water use and root morphology
Pines with a kikuyu understorey generally transpired less water compared to those with an annual pasture understorey. While the water use of kikuyu wasn’t measured, it is reasonable to assume that this was due to competition between the pines and kikuyu. Measurement showed tree roots were displaced from the upper 1.8m of the soil profile by kikuyu roots. While pines with a kikuyu understorey used less water than pines with annual pasture, other studies suggest that the additional water used by the perennial pasture would more than offset reduced tree water use. Consequently, perennial pasture based agroforestry systems would be more effective than annual pasture based systems in reducing groundwater recharge.

 Implications for relevant stakeholders

Perennial grasses that are drought tolerant and adapted to low fertility (eg. kikuyu, tall wheat grass, Rhodes grass, veldt grass and consol lovegrass) are better able to compete with trees than annual pastures and lucerne in the high to medium rainfall zone of southwest Australia. Other than the area directly under tree crown, agroforestry systems that include perennial grasses will produce at least similar amounts of forage to annual pastures alone. Lucerne competes poorly with trees, however pasture losses may be overcome by maintaining soil available P and establishing the trees and lucerne in the same growing season. This requires further investigation. It is likely that there will be less leakage under perennial pasture based agroforestry systems compared to annual based systems.

Undoubtedly “competitive” perennial pasture species will affect tree growth. Tree growth was not measured during this one year study, but at the 70% of sites where pasture biomass production in the root zone of trees was greater for perennial pasture than annual pasture, it is safe to assume that the perennial pasture was providing more competition with the trees for resources and so may have been reducing tree growth. If so, this will have economic implications where trees are being grown to produce commercial products but is of less importance if trees have been planted for shelter, aesthetic or landcare benefits. It is necessary to have a sound understanding of the economic consequences of both improved pasture production and any decrease in tree growth before such systems can be widely recommended to farmers.

Recommendations
The design of agroforestry systems in the south west of Australia should include drought tolerant perennial grasses if the objective is to minimise losses of forage production. Species proven to date include kikuyu, tall wheat grass, veldt grass (though there is limited seed available), Rhodes grass and consol lovegrass. It is likely that setaria, panic, digit grass and native grasses would also be competitive with trees.

Further research is required to determine the effect perennial pasture species have on the growth of commercial tree species in silvopastoral systems across a range of environments. These data should be used to assess the economics of various silvopastoral systems in a range of environments.

Further research is required to determine the best methodologies for establishing and managing perennial pastures in existing and new agroforestry systems.