Development of a hemp (Cannabis sativa L.) simulation model 4. Model description and validation

In studies assessing the prospects for a hemp industry, as well as in longe r term research activities, the use of a hemp simulation model to complemen t the more traditional agronomic field trials would offer a number of poten tial advantages. In addition to being cost and labour intensive, field tria ls with hemp have political, social and security implications. With these i mplications in mind, a simulation model that captures the growth and develo pment processes of hemp in response to management, genotypic, soil and clim ate factors, has the potential to increase research efficiency. The model c ould be used to assess the need, extent and nature of field trials, to help interpret field trial results, and to investigate temporal and spatial var iability in selected crop responses. This paper describes a hemp crop growth and development model (APSIM-Hemp) and its validation against an independent dataset. The model was developed as a crop module within the framework of the larger systems model, Agricult ural Production Systems sIMulator (APSIM), to extend the capability to enco mpass the agricultural system in which hemp is grown. APSIM-Hemp incorporat es relationships developed in the previous papers in this series relating t o pre- and post-emergent phenology and leaf area production. Other paramete rs relating to biomass partitioning, biomass production, water uptake and n itrogen uptake were derived from separate field studies and selected refere nces. APSIM-Hemp adequately predicted phenology, leaf area and biomass production for the cultivar Kompolti at Forthside in north-western Tasmania, for a da taset comprised of results from trials conducted over 3 seasons and includi ng treatments of sowing date, irrigation regime and plant density. Although performing well against this independent dataset, the performance of the m odel needs to be further validated over a range of other soil, climate and management conditions in order to assess its broader predictive capability. Notwithstanding these limitations, the sound basis of a model for simulati ng the growth and development of hemp has been developed.