An exploratory study of the relation of pH data from pens boluses with othe sensor data

In the Dutch Smart Dairy Farming project, automated measurements of the pH in the rumen of dairy cows were used to analyze processes during the transition period. The pH data can be used to monitor the developments during the end of the dry period and the start of the lactation. If these pH values are related with other sensor data (like activity and feed intake), then monitoring these other data might results in possibilities to detect metabolic problems. On two practical dairy farms (farm 1: 300 cows, automatic milking system; farm 2: conventional milking, 110 cows), in total 70 cows were fitted with Smaxtec pH bolus a few weeks before calving. Valid measurements were available per bolus over two months. Other sensors were used to measure milk yield, activity, rumination, concentrates intake, visits to concentrate feeder (and milk robot on farm 1) and body weight. This resulted in a data set to be used to explore the relation between pH value and other traits. The first results show: (1) no clear relation between pH and concentrates intake; (2) great fluctuations in pH on the day of calving (caused by changes in feeding?); (3) that an increase in rumination might be correlated with a decrease in pH The pH data were aggregated by logistic regression; making it possible to convert the pH data of one day in an average level and a slope describing the variance within a day. More results will be included in the presentation.