Simulating lucerne growth and water use on diverse soil types in a Mediterranean-type environment

Abstract

The performance of the Agricultural Production Systems Simulator (APSIM) lucerne (Medicago sativa) model was assessed by comparing calculations from APSIM with measured (or observed) data from 9 sites in Western Australia. This comparison was also to obtain new insights into lucerne production and the effect of lucerne on the water balance in this environment. APSIM accounted adequately for the temporal change in above-ground biomass production and the plant-available water (PAW) for most of the sites. The root mean square deviation (RMSD) for biomass was 1.3 t/ha for the mean observed biomass of 4.17 t/ha. The RMSD for PAW was 16 mm to a depth of 1.6–2.1 m, with the mean observed PAW of 50 mm. The good prediction of PAW was partly because critical soil parameters used in APSIM were derived from the soil water content measurements. APSIM also adequately (within the standard error) estimated evapotranspiration (Et) and drainage below the root-zone, which was measured at 2 of the sites. The analysis supports previous findings with lucerne that increased storage of carbohydrates in root reserves occurs in autumn and winter. Given that APSIM performed adequately when calibrated, it was used to simulate Et and drainage for the 7 sites in which measurements were not taken. Et in all 3 years of lucerne and drainage in Year 1 were related to the amount of rainfall. Fifty-one percent of rainfall above 230 mm was lost as water excess in Year 1 (R2 = 0.68). Drainage in Year 3 was less than drainage in Year 1, confirming previous studies that established lucerne can reduce drainage.