From controlled environments to field simulations: Developing a growth model for the novel perennial pasture legume Cullen australasicum

To allow development of a dynamic plant growth model for the simulation of dry matter (DM) production in the field based on meteorological data, in the absence of detailed information on phenology and growth, a model was developed solely using data from plants grown in large pots under glasshouse conditions. The little studied Australian native perennial legume Cullen australasicum, which has potential as a pasture species, was used. Phenological development (e.g. rooting depth, leaf appearance, leaf growth and shedding) and physiological characteristics (e.g. photosynthetic capacity) were measured in the glasshouse for model development and parameterisation. Whenever possible and appropriate, established models were used to explain individual processes (e.g. canopy photosynthesis and evapotranspiration); if unavailable, models were developed. A diverse independent data set was used to validate the model, including published and unpublished results from the Western Australian wheatbelt under non-irrigated conditions and different plant densities (1, 2, 4, 8 and 16 plants m−2), and from other locations as well. Main stem node appearance was simulated fairly well, with a root mean square of deviations (RMSD) of 1.86 nodes day−1 (r = 0.996, n = 11). The model predicts the DM production well for non-irrigated conditions in the Western Australian wheatbelt under densities of 1, 2, 4 and 8 plants m−2, and at Wagga Wagga and Barmedman, New South Wales, but not for a density of 16 plants m−2 in the Western Australian wheatbelt and not under irrigated, fertilised conditions. Overall, the model explained over 78% of the observed variation in DM with a RMSD of 54 g m−2 (r = 0.89, n = 31). The high prediction accuracy suggests that this generic approach to model development offers promise for the simulation of the growth of C. australasicum for the tested locations, particularly under non-fertilised and non-irrigated conditions. This approach should prove valuable for quick evaluation of the diverse array of novel crop and pasture species now under evaluation for low-rainfall areas, as it requires only limited information and minimal parameterisation.

Research highlightsThe attached manuscript reports on the development of a new and interesting growth model for simulating the growth of the novel perennial pasture legume Cullen australasicum under moisture-limited field conditions. Data for model development (e.g. phenological development and physiological characteristics) were collected from plants grown in a glasshouse and the model was subsequently validated using a diverse array of data collected from the field, both published and unpublished data. Due to the fact that the growth and phenology of this novel pasture legume has not been investigated in detail, the information available was relatively limited. Therefore, we used the novel approach of growing plants in glasshouses in close proximity, allowing us to collect data regularly to develop the growth model. We then used the limited data available from field trials for the purpose of validation. Our results indicate that the model performed well, as compared with well-established models, thus supporting our approach to model development as a valid and valuable approach when first constructing growth models for novel agricultural species. This finding will be of general interest to crop modellers, agronomists and plant ecophysiologists.