Newsletter with the latest news. Problem viewing? View it in your browser.

Improving the NorFor forage intake model by registration of chewing activity

During his Master thesis, Knut Olav Øvreeide researched ways of improving the NorFor forage intake model for individual dairy cows. The goal of this master thesis was to investigate if the use of sensory data with focus on chewing and rumination time could improve prediction of daily individual forage intake.

NorFor team from Norway. From left to right: Jon Kristian Sommerseth (Researcher), Knut Olav Øvreeide (Master student), Tone Roalkvam (NorFor chaiman).

The trial was conducted during the autumn of 2018 at the Norwegian University of Lifesciences (Norway). The study involved 43 Norwegian red dairy cows from their calving date and lasted 60 days. The cows were divided into three groups with different feeding regime: low digestibility of the forage and medium concentrate level, medium digestibility of forage and medium concentrate level, and medium digestibility with high concentrate level.

The cows were fitted with sensors by Nedap CowControl and RumiWatch for recording chewing activity. Individual daily milk yield, concentrate intake, forage intake, body condition score (BCS) and body weight were recorded, as well as periodic chemical analysis of milk, forage and concentrate.

Table 1: Sumaary of models evaluated in the study.

From the observed recordings, prediction ability of the models was evaluated. Prediction of individual forage DMI by the existing NorFor-model resulted in a 13,1% mean prediction error (MPE) and r2 = 0,68. When chewing activity (NDE and NDR) were included in the NorFor model, into the error was reduced (11,5% MPE and r2=0,75).

Other included parameters in the model were BCS and calculated grams of AAT per MJ of NEL by NorFor (Prot). The inclusion of Prot improved the model, with the advantage that it is an existing calculation in NorFor and does not need additional investments. By combining NorFor with both recorded chewing time and Prot the prediction model with an MPE of just over 10% was obtained.

By combining all available information from this trial, the prediction error resulted in 8,8% or 900g DM on an individual level.

In summary, the use of commercially available sensors measuring chewing activity, in combination with NorFor, can predict individual forage dry matter intake in dairy cows in early lactation with satisfying results.


homepagefacebookyoutube

You are receiving this newsletter since your are a customer or you have signed up on our newsletter. If you would like to unsubscribe, click here.