Abstract :
This study evaluates how harvesting regimes (HRs) of grass silage influences economic
return and optimal use of inputs, in particular fertilisers and concentrates, in dairy goat
farming in mountainous areas of Norway. Goats in such areas are fed indoors with conserved
silages (mostly grass silage) and concentrate supplements for up to 9 months a year. A
whole-farm linear programming model was developed to examine three HRs: very early
(HR1), early (HR2) and normal (HR3), producing silage containing different concentrations
of net energy. Linear input/output response relations incorporated into the model were
derived from a field experiment with two levels of fertilisation and an animal experiment
with two levels of concentrate feeding to supplement silage from each HR. The model
maximises total gross margin of a dairy goat farm with 70,000 l milk quota, and stalling
capacity for 100 goats. Farmland availability varied from 4 to 10 ha with 6.5 ha as the basis.
The study demonstrates that farmland availability profoundly influences the choice of input
intensity and the profitability of producing and feeding high quality grass silage to dairy
goats. At 6.5 ha, optimal input of fertilisers and concentrates as well as output of milk per
goat was highest when fed HR1 and HR2 silage. However, HR3 was most profitable as
benefits such as higher milk yield and better milk price due to higher total solids content
for HR1 and HR2, did not offset the additional costs and increased shortage of silage at lower
yields achieved and the higher consumption of silage. HR1 was particularly unfavourable
as it was impossible to fully produce the milk quota. More land than 6.7 ha was necessary
for HR2 to outperform HR3. Inputs of fertilisers and concentrates could be reduced as more
land became available, however, at the lowest land constraint for HR3. By removing the
milk quota, the profitability of HR3 was strengthened at limited land availabilities, at 7 ha
HR2 surpassed HR3 in profit and above 8.5 ha HR1 was most competitive
