Effect of subtropical perennial grass pastures on nutrients and carbon in coarse-textured soils in a Mediterranean climate

In the northern agricultural region of Western Australia, some farmers have integrated C4 grass pastures into their farming system. This switch away from a farming system historically dominated by annual species could alter soil nutrient and carbon (C) levels. To explore this issue, 16 'over-the-fence', pair-wise comparisons were conducted between a field in an annual crop-pasture rotation and a field sown to perennial pasture. The perennial pastures were 2-8 years of age, and comprised a mix of perennial species and volunteer annuals. Soils were deep yellow sand (Tenosols), and perennial root systems extended to beyond 150 cm. Averaged over all sites, organic C (OC) in the top 90 cm of the profile was 38.7 t/ha, nitrate 34.5 kg/ha, Colwell phosphorus (P) 5.9 kg/ha, and Colwell potassium (K) 518 kg/ha. In general, there were no differences between annuals and perennials for any of these attributes. For OC, greater differences were observed between sites than between annual and perennial species. Site differences in OC ranged from 18.7 to 62.7 t/ha. There were some differences between annuals and perennials at different sites for OC, P, nitrate, and K, but these differences were not systematic. For example, at three sites, more OC was found in the annual system; at another site, more OC was found in the perennial system, and differences occurred in the surface layers at two sites and at depth in the other two. Overall, clear differences between annual and perennial farming systems were not evident and were highly variable. The survey was augmented with a simulation study using the APSIM crop model, where a continuous perennial pasture was compared with a wheat-wheat-lupin rotation. After 10 years, there was only 1.27 t/ha more OC in the perennial system than the annual. The simulation study and on-farm survey suggest the differences in the accumulation of soil C under a perennial or an annual system in this landscape are small and variable, and if differences do exist, they will take considerable time to accrue. In addition, management and local site effects were more important than the perenniality of the system per se in influencing C and nutrient levels.