Modelling perennial grass phenology

Perennial ryegrass is widely used for pasture and hay in sheep, dairy and beef production. Modelling is able to help guide farmers when the best time is to harvest and/or allow grazing to maximize digestibility, yield, and annual regrowth. These characteristics are dependent on the stage of development during biomass removal, which is regulated by climatic conditions and genetic variation.

Computer modelling can help us understand and predict perennial grass phenology for optimal management.

Simon Rouet, an agronomy PhD Student at the French National Institute for Agriculture, Food, and Environment (INRAE), and coauthors present a functional-structural plant model which represents the development of perennial grasses in a new article published in in silico Plants. The model, L-GrassF is an extension of the existing model L-Grass.

The new model includes the interactions between the reproductive and vegetative development for each individual tiller, rather than just vegetative development. The vegetative stage of perennial ryegrass is marked by the appearance and elongation of leaves and tillering. Reproductive development occurs after two floral induction phases, which trigger floral transition of the apex and ultimately, internode and spike production, and heading of individual tillers. One of the main features of grassland management is their frequent cutting, which was also included in the model.

The authors evaluated the model using a dataset which describes the heading date of 7 cultivars of perennial ryegrass for 6 locations over 14 years. Meteorological data for those locations and years was also fed into the model. From the interactions between the vegetative and reproductive development, L-GrassF was able to accurately simulate the heading date for the cultivars.

“Explicitly modeling the effects of temperature and photoperiod is crucial to consider the effects of climate change on the phenology of perennial grasses. Indeed, original combinations of these climate variables are foreseen, potentially invalidating current empirical prediction methods,” explains Rouet.

The authors performed a sensitivity analysis to determine which parameters most controlled heading date. These were the rate of appearance and elongation of leaves during the vegetative and reproductive phases, along with the rate of secondary induction.

Because vegetative growth of tillers stops at the time of heading, including reproductive development in the model allowed the authors to estimate harvestable leaf area throughout the season for multiple cultivars and under different seasonal precipitation and temperature.

Rouet concludes, “In addition to the potential uses of L-GrassF to predict the effects of climate change, the consideration of individual tillers morphogenesis and tillering by the model paves the way to understanding the seasonal dynamics of the tillers population in grasslands and evaluating their perenniality.”

READ THE ARTICLE:

Simon Rouet, Jean-Louis Durand, Denis Leclercq, Marie-Hélène Bernicot, Didier Combes, Abraham Escobar-Gutierrez, Romain Barillot, L-GrassF: a functional-structural and phenological model of Lolium perenne integrating plant morphogenesis and reproductive development, in silico Plants, 2022;, diac012, https://doi.org/10.1093/insilicoplants/diac012

The code for L-GrassF is freely available at https://github.com/openalea-incubator/lgrass, and released under a CecILL-C license. The archive of the code used to generate the present results is available at 10.5281/zenodo.5116065.