Grazemore DSS för att optimera utnyttjandet av bete i mjölkproduktionen

Abstract: The aim of the study was to investigate if the Grazemore Decision Support System (DSS) is able to provide a grazing management strategy that gives a high utilisation of grazed grass in milk production in the north of Scandinavia. To do this, a grazing experiment was planed and performed during the summer 2005. Simulations in the DSS were run to get a suggestion of how the cows should graze, grazing calendar 1. Deviations and updates during the season resulted in the simulated grazing calendar 2. During the experiment, the actual milk yield was recorded twice weekly. The difference between actual and predicted milk yield by Grazemore DSS was analysed statistically with regression analysis and the mean square prediction error (MSPE) was estimated. Plots of grass were cut in order to get an idea of the herbage mass during the experiment. The herbage mass during the experiment was higher than predicted by the model and a surplus of grass in the paddocks was not utilised. The average milk yield during the experiment was 29,9 kg/cow/day with a standard deviation of 1,4 kg/cow/day. The DSS predicted a milk yield of 30,8 kg/cow/day in grazing calendar 1 and 31,5 kg/cow/day in grazing calendar 2. The standard deviation was 0,7 kg/cow/day in both cases. The statistical analysis showed that the model had a prediction error of 5 and 6 percent respectively. The R2 values were 0,25 for grazing calendar 1 and 0,40 for grazing calendar 2. Grazemore DSS ability to provide a grazing management strategy that gives a high utilisation of grazed grass in milk production in the north of Scandinavia was insufficient. The milk yield observed in the experiment and the programs ability to predict this value were however satisfactory. After further research and evaluation, mainly concerning the herbage growth model and its ability to predict herbage mass, Grazmore DSS has potential to be a helpful tool for optimising grass utilisation.