Comparing the performance of nutritive value predictions in three timothy models

Abstract

Grasslands are the main source of energy and nutrients in ruminant production systems. Nutritive value of grasslands is in most feeding systems described based on energy, i.e. digestibility and cell wall content, and crude protein content of the feed and plays a significant role in the profitability of these production systems. Timothy (Phleum pratense L.) is a widely used forage grass grown either in pure stands or in mixtures with other forage grasses and legumes in cold-temperate regions of the world. Timothy management practices, including cultivar selection, cutting frequency, and fertilization are adapted to the climate and soil conditions as well as to the animal production system this grass is part of. Models exist that can simulate phenological development, dry matter growth, digestibility and nutritive value of timothy as a function of the weather, soil, and management practices. These models differ in how they represent plant processes related to nutritive value. An analysis of these differences is needed to identify the correct process representation, and requires comparing model outputs against data from experiments conducted under different climate, soil, and management conditions. The overall goal of this study was to compare the ability of three simulation models, BASGRA, CATIMO and STICS, to predict fibre and crude protein concentrations along with digestibility. Datasets covering a wide range of climate and soil conditions, cultivars, and management practices in major timothy grass production regions of Canada, Finland, Norway, and Sweden were used for model calibration and validation. Simulations results were then analysed to better understand the strengths and the weaknesses of the modeling approaches used in the evaluated models.