By Marie-Noëlle Thivierge1 , Xiawei Shi2, and Philippe Seguin2
1Agriculture and Agri-Food Canada; 2McGill University
To compare the performance of various forage mixtures intended for pastures in organic
dairy production, a field trial was established as part of the Organic Science Cluster 4 led
by Caroline Halde and Marie-Noëlle Thivierge. This study used mixtures of six forage
species adapted to trampling and intensive grazing: two legumes (white clover and
birdsfoot trefoil), three grasses (meadow fescue, meadow bromegrass, and orchardgrass),
and one herb (chicory). At two sites (Saint-Augustin-de-Desmaures and Sainte-Anne-de-
Bellevue), 48 mixtures were tested, including combinations of all six species together,
and different mixtures of two, three, four, or five species. Pure stands of the six species
were also grown, as a reference.
In the first production year of 2025, plots were fertilized with dairy cattle slurry (annual
application of 50 kg/ha of total nitrogen) and harvested each time the vegetation reached
approximately 25 cm in height (five harvests at Saint-Augustin and four harvests at
Sainte-Anne in 2025) to simulate intensive grazing. At each harvest, plants were
separated by species to quantify biomass yield, and weeds were separated from sown
forage species.
Experimental plots of the project in Saint-Augustin-de-Desmaures in 2025. Photo: Chantal Lachance © AAFC.
At the Saint-Augustin site, the average yield in the first production year was 6.8 t of dry matter (D.M.) per hectare. The most productive species in pure stands were white clover, birdsfoot trefoil, meadow bromegrass, and orchardgrass. At the Sainte-Anne site, forage yield was very low (1.7 t D.M./ha on average) due to dry conditions that persisted throughout the 2025 season. Birdsfoot trefoil offered the best forage yield (2.2 t D.M./ha) among the six pure species: it is possible that the dry conditions favored this legume, which roots more deeply and is more drought-tolerant than white clover.
Generally, increasing the number of species in the mixture led to higher forage yields (Figure 1a, b), and this was more apparent at Saint-Augustin. The proportion of weeds, meanwhile, decreased sharply with more species sown in a mixture, falling from an average of 23% (pure species) to 1% (six-species mixture) at the two sites (Figure 1c, d).
Figure 1. Total forage yield in the first production year (kg D.M./ha) and proportion of weeds (%) as a function of the number of species in the mixture and regardless of the botanical family of these species.
Although increasing the number of sown species tended to favor higher forage yields, some of the 47 mixtures comprising 2, 3, 4, or 5 species still exceeded the yield of the six-species mixture. It is interesting to note that different mixtures stood out at both sites, likely due to the very different soil and climatic conditions encountered. The six-species mixture proved to be a safe bet at both sites by maximizing interactions between botanical families and offering species adapted to all conditions, which unfortunately no one can predict in advance!
Combinations of botanical families greatly contributed to the generally better performance of the species mixtures. In general, regardless of the number of species involved, the combination of legume(s) and grass(es) (LG and LGH mixtures; Figure 2a, b) favored better forage yields at both sites, compared with the average yield of pure species from these same families. The complementarity between these families in terms of nitrogen sharing largely explains this effect. The proportion of weeds decreased greatly in legume-grass or legume-herb mixtures at both sites (LG, LH, and LGH), as well as in grass-herb (GH) mixtures at Sainte-Anne (Figure 2c, d). Grasses and herbs can be complementary in several respects, including their growth pattern during the season and the distribution of their roots in the soil profile, which limits the resources available for weeds. Chicory, the herb used in this study, is particularly well-adapted to the dry conditions that prevailed at Sainte-Anne due to its long taproot.
Figure 2. Total forage yield in the first production year (kg D.M./ha) and proportion of weeds (%) as a function of the botanical families composing the mixtures (L=legume, G=grass, H=herb) and regardless of the number of species.
In conclusion, this study shows promising results: the most diversified mixtures comprising a combination of legume(s) and grass(es), with or without herbs, offered the best yields at both sites, while being the most effective at reducing the presence of weeds. Some mixtures with two, three, four, or five species performed better than the six-species mixture. However, the mixtures that stood out were not the same at both sites. The six-species mixture proved to be a safe bet at both sites by offering species adapted to all conditions. Despite difficult conditions at establishment and a lower yield at the Sainte-Anne site due to dry conditions, the interactions between botanical families proved positive and even stronger than at the Saint-Augustin site, suggesting that it is in non-optimal conditions that the resilience of diversified systems is most expressed. The study continues in order to compare the persistence of species and the nutritional value of the mixtures over the long term and according to the real climatic conditions.
This study is part of a project co-led by Caroline Halde (Laval University) and Marie-Noëlle Thivierge (Agriculture and Agri-Food Canada) and funded by the Organic Science Cluster 4. It is a joint initiative led by the Organic Federation of Canada in collaboration with the Organic Agriculture Centre of Canada at Dalhousie University, to meet the needs of the Canadian organic sector. This Organic Science Cluster 4 project is funded by Agriculture and Agri-Food Canada (70%) and by industry partners (30%), primarily Novalait and the Dairy Farmers of Ontario, with contributions from the Organic Council of Ontario, AgroEnviroLab, the Syndicat des producteurs de lait biologique du Québec, and Lactanet. Collaborators in this study: Julie Lajeunesse (AAFC), Rafael De Andrade Moral (Maynooth University, Ireland), Jonathan Lafond (Laval University), Jean-Thomas Denault (AAFC), Mary-Cathrine Leewis (AAFC), Fadi Hassanat (AAFC), Édith Charbonneau (Laval University) and Nicolas Devillers (AAFC). Thanks to all the farmers and advisors who participated in the discussions leading to the development of the protocol for this study, as well as to the research assistants and summer students who made the realization of the project possible.