Data
Raw data tables
Table 1. Collar data that is cleaned up but not transformed.
The collars sent regular reports every fifteen minutes, with information about location, the environment, battery life, animal activity, etc. The collars also sent a report every time an animal received an audio warning or electric shock, and when they escaped the virtual fence boundary or returned.
Table 2. Forage data that is cleaned up but not transformed.
At the beginning, in the middle, and at the end of each grazing period I took six stratified, random forage biomass samples. I used this data to estimate how forage supply changed over the course of the grazing period.
Table 3. Weight data that is cleaned up but not transformed.
I weighed the cattle every two weeks and used the data to monitor their wellness - a sudden decrease in weight can indicate an animal is sick or injured. I also used the weight data to calculate the exact forage demand of the animals on pasture.
Virtual fence interaction data
Figure 1. Scatterplot showing the number of audio warnings that occurred each day. This graph represents the total number of audio warnings that occurred each day of the trial. The black dotted lines represent rotation days where the cattle were moved to the next pasture. Overall it looks like the number of audio warnings might be increasing as the trial goes on.
Figure 1. Scatterplot showing the number of electrical shocks that occurred each day. This graph represents the total number of electrical shocks that the cattle experienced each day. The black dotted lines represent rotation days where the cattle were moved to a new pasture. Unlike the previous graph, you can clearly see that the number of audio warnings per day decreased throughout the trial.
Figure 2. Number of audio warnings experienced per day in at the beginning, middle, and end of the grazing periods. This graph suggests a trend where cattle experience fewer audio warnings per day at the beginning of a grazing period, and more at the end.
Figure 3. Number of electrical shocks experienced per day in at the beginning, middle, and end of the grazing periods. Unlike the audio warnings graph, this figure does not suggest that there is a significant difference in electrical shocks between the beginning, middle, and end of the grazing periods.
I chose not to remove any outliers in my collar data. In this case my dataset is not a sample of a population, it is the total number of virtual fence interactions that occurred. Rather than being sampling errors, the outliers were caused by behavioural anomalies and represent important behavioural anomalies. You can see there are two outliers during the middle grazing phase in Figure 4. July 2nd was the first day of the training phase that we fully challenged the animals with the virtual fence, and so it makes sense that the cattle received more shocks on this day. The July 7 outlier occurred during the middle of the first experimental period, when the cattle were still new to the virtual fence system. I cannot pinpoint an exact cause of the increase in shocks for this day, but it could have been caused by the presence of a predator, for example. This is still important behavioural information.
Forage Supply Data
Figure 4. Box plot of total forage (kg) supply available per grazing period.
Figure 5. Box plot of total forage (kg) supply available per grazing phase.
You can see how the forage supply decreases throughout a grazing period, as cattle remove the forage through grazing. This means that the ratio of forage supply to forage demand decreases as the cattle graze, and that causes an increase in grazing pressure.
Forage Demand Data
Figure 6. Line graph showing the forage demand, which is the amount of forage that would be consumed by the herd of cattle throughout any one grazing period. Forage demand is calculated using animal unit months, or the amount of forage that one 1000 lb cow consumes in a month. This number is adjusted based on the actual weight of an animal, and then the consumption can be calculated for the entire herd across any number of days.