APSIM-Lucerne Validation in the Temperate Climate of New Zealand

Abstract

Shoot yield data collected from an irrigated lucerne (Medicago sativa) crop at Lincoln University, Canterbury, New Zealand (43°38′S, 172°28′E, 11 m a.s.l.) from 1997–01 were used to validate the APSIM-lucerne crop model for a temperate climate. Adjustments to APSIM-lucerne were required to account for autumn shoot death through frosting, defoliation of leaf area as occurs from animal grazing, a change in low temperature response for thermal time accumulation, a reset of stem population as they reduced over time and inclusion of a radiation use efficiency function to account for low temperatures.

Changes were then validated with a separate dataset based on a dryland (670 mm average annual rainfall) crop dataset from the same period. The modified model increased the accuracy of prediction. Root mean squared deviation (RMSD) improved within each season (excluding the first spring regrowth cycle) with averages ranging from 15 to 35% of the mean. The exception was the first autumn regrowth cycle that was underestimated and the mean RMSD was 79%. Further validation was performed using a separate irrigated dataset (2001–2004). The modified model generated a distinct pattern of bias within each season, under predicting yield within the first three spring regrowth cycles and over predicting the final three autumn regrowth cycles. These results showed that targeted experimental data to account for storage and remobilization of root reserves in autumn and spring, respectively, are required to further improve accuracy of the model.