LAUSR

Abstract Ruminant production is predominant in mountainous areas, where grassland plots are often difficult to reach or non-machinable. Many mountain-area livestock systems combine a dairy-cow herd with a sucker-cattle herd. Here we posit that mixed herds allow better utilization of grazed grass than specialized herds on plots with constraints. To determine the influence of herd composition on grassland valorization, animal production and economic performance, we studied the functioning of two farms derived from two farm types representative of real mixed cattle mountain systems in the south of the Auvergne (France). Farm_1 is a medium-sized operation with a summer mountain pasture, and Farm_2 is a large operation with fragmented plots of grassland and cereal crops. We used modeling to vary the proportion of herd livestock units (LU) between the dairy herd and the suckler herd. We used the bioeconomic optimization model ‘Orfee’ (Optimization of Ruminant Farm for Economic and Environmental Assessment) modified for mountain conditions by integrating geographical constraints and the heterogeneity of grassland plots based on the French permanent grassland typology. The experimental design used 5 simulations defined by percentage of LUs found in the dairy (D) herd: 100D, 75D, 50D, 25D, and 0D. The results show better utilization of grazed grass with mixed herds than specialized herds. Based on total amount of pasture grass available in each farm, Farm_1 had the maximum grazed grass utilization rate at 85%, while Farm_2 had total valorization by grazing (rate of 100%). This better utilization of grazed grass increases average profit per Annual Work Units, especially when the difference in profitability between dairy and suckler herds is small. In addition, mixing herd reduces profit variability, taking into account market and policy variations over the period 2000–2015. This study shows that understanding the farm structure is key for getting the right herds mix to improve grazed grass utilization with good economic performance.