The need for adaptive management of high-yielding dairy cows in low-input mountain systems

Abstract The majority of dairy farms in mountainous European regions rely heavily on concentrates to supplement herbage-based diets for high-yielding cows. This puts their economic sustainability at risk. We need to design management systems that enable milk production from herbage with few inputs. To examine whether low input systems could be technically efficient with high-yielding breeds, we tested two experimental farming systems —a Very-Low-Input system (VLI) and a Low-Input System (LI)— with Holstein and Montbéliarde cows for five years (24 cows/system.year), with a short mating season in spring to align the milk and the grass production curves. The VLI system used almost no concentrates and no mineral fertilisers. The LI system used fewer concentrates and mineral fertiliser than conventional systems, while safeguarding milk production and the nitrogen balance of the land. During the first year, only 44% of the cows were pregnant after the mating season. Reproductive management needed to be adjusted to ensure herd renewal while keeping lactation and grass growth curves synchronised: Sexed semen was used on heifers, natural mating was applied at the end of the mating season, and the lactation of some cows was extended by 8–10 months while these cows were mated the next year. Depending of the year between 7 and 24 cows out of the 48 had extended lactations. The cows with extended lactations produced around 80% as much milk as cows with standard lactations during the same time (two years). Over the five years, the VLI system produced 486,000 kg of milk and consumed 3.12 t of concentrates (used only for female calves) while the LI system produced 550,000 kg of milk and consumed 130.6 t of concentrates, resulting in 156.8 vs. 4.2 kg of milk/kg of concentrates in VLI and LI, respectively. The VLI system used twice as much grassland as the LI system (59.9 vs. 29.6 ha) and is thus less than half as efficient in land use (1,630 vs. 3,720 kg of milk/ha/y). Compared to the LI system, the VLI system generates a higher milk income (+ 3,600 €/y) but its land productivity is lower (-371 €/ha/y). We conclude that the VLI system is well-adapted to a context of a drastic reduction of inputs because it is almost self-sufficient, while LI is a good option if land availability is low. However, if high-yielding cows are to be maintained in both systems, it is necessary to adopt a flexible approach to cow reproduction, such as extending lactations. In addition, the body condition of these cows is put at risk in such systems, which can raise welfare concerns.