Onderwerp scriptie

Dairy cow behaviour as a possible predictor of individual grass intake - Djura Hoeksma

The aim of the present study was to examine whether cow behaviours continuously recorded with the use of sensors, in combination with other potential predictors related to the pasture and to (production) characteristics of the cow, could be used to estimate individual grass intake by grazing dairy cows and to develop a novel method to predict individual grass intake of grazing dairy cows.

Currently, there are no practically feasible methods to reliably estimate the individual grass intake of grazing dairy cows and the use of sensors is increasing on dairy farms in Western Europe. The aim of the present study was to examine whether cow behaviours continuously recorded with the use of sensors, in combination with other potential predictors related to the pasture and to (production) characteristics of the cow, could be used to estimate individual grass intake by grazing dairy cows and to develop a novel method to predict individual grass intake of grazing dairy cows. The data of grass- and cow characteristics, cow behaviour and individual grass intake were collected during a one-week experimental period with 60 dairy cows in a grazing trial. All cows were equipped with three sensors; (I) an ‘IceQube’ leg sensor, (II) ‘CowManager SensOor’ ear sensor and (III) a ‘Smarttag Neck’ neck sensor to record standing/lying/number of steps, grazing and ruminating behaviour, respectively. Grass height was determined with a JenQuip grass altimeter and bite rate was scored by observers. Grass intake (kg DM/day) was determined per individual cow with the use of the n-alkane method. Grass intake was considered as response variable (y), and stepwise multiple regression analyses were performed to identify the best fitted models based on percentage of adjusted variance explained (Adj. R Square). Three practical models with a significance below 0.001 were found to predict individual grass intake: (Model 1) days in milk, milk yield, grazing intensity, number of steps taken on pasture and grazing area per cow (Adj. R2 = 71,2%), (Model 2) days in milk, milk yield, grazing time and grazing area per cow (Adj. R2 = 70,6%), (Model 3) days in milk, milk yield, grazing time and grass height (Adj. R2= 71,8%). The findings demonstrate that individual grass intake of grazing dairy cows can be reliably estimated from a number of behavioural and other predictors. The models may present potential management tools for grazing management on dairy farms.

Student: DL Hoeksma

Supervisor: dr ir K van Reenen

36 Ects