Priorities for narrowing the yield gap and increasing farming systems resilience in the Fiji sugar industry: APSIM model, outputs and weather data

Full description

APSIM model, outputs and weather data used in Meier et al. (2023) published in Farming System.
Lineage: The model was configured with modules for sugarcane (Keating et al., 1999), soil water and soil nitrogen (Probert et al., 1998), surface residue (Probert et al., 1998; Thorburn et al., 2001), and management practices. The model focuses on carbon and nitrogen cycling whereby soil carbon and nitrogen respond to temperature, soil water, soil pH and application of organic amendments, which in turn interact with plant growth and plant development. Nitrogen that is not taken up by plants or immobilised in soil microbial biomass may leave the soil-plant system as nitrate leached below the root zone or in denitrification pathways (Weier et al., 1996). APSIM was not set up to simulate losses of N in runoff because this requires calibration (Thorburn et al., 2011; Vogeler et al., 2023), and tends to be a small portion of applied N (Ng Kee Kwong et al., 2002;
Thorburn et al., 2011). Nitrogen in runoff is reduced where there is a time lag between application and occurrence of the first runoff event (Ng Kee Kwong et al., 2002; Melland et al., 2017, Melland et al., 2022), which was likely given application during the Fiji dry season (Fig. 1; Table 2). APSIM has been well validated and tested for sugarcane systems under diverse management and in differing environments (e.g., Australia, Hawaii, South Africa and Swaziland, Keating et al., 1999; China, Peng et al., 2020; and Brazil, Dias et al., 2021). For this study, testing of APSIM was limited to a calibration step informed by reference to local soil properties (acidity, occasional seasonal waterlogging; Section 2.1),
and typical N fertiliser rates and crop yields at the location (Section 2.2). Further calibration and validation were not possible due to limited availability of detailed soil, management, crop cultivar characterisation and weather data for the site.
All scenarios included the same ‘run-in’ period consisting of two crop cycles of 11 years each (i.e., two crop cycles consisting of one plant crop followed by nine ratoon crops) to assist simulated soil water and mineral nitrogen values to approach equilibrium values. The duration of the run-in period was sufficient because there was little difference (~1 Mg ha) between the average yield of crops simulated with one or two crop cycles during the run-in period (data not presented, but available by request from the authors). The practices simulated during the run-in period were those described for the first scenario (Table 1), and the results from the run-in period were excluded from analyses. The
management practices associated with the scenarios (Table 1) were introduced after the run-in period and the scenarios were then simulated for the years 2013 to 2022, inclusive (the limit of available climate data). One nine-ratoon crop cycle and two three-ratoon crop cycles were simulated within this period.