Kernel density estimation analyses based on a low power GPS for monitoring environmental issues of grazing cattle

Abstract

The use of wearable sensors that record animal activity in intensive livestock systems has become more and more frequent for both early detection of diseases and improving quality of production. Their application may be also significant in extensive livestock systems, where there is an infrequent farmer-to-animal contact. The aim of the present study was to prove the feasibility of a novel automatic system for locating and tracking cows in extensive livestock systems based on space-time data provided by a low power global positioning system (LP-GPS). The information was used to study the pasture exploitation by the herd for modelling the environmental impacts of extensive livestock systems, trough Geographical Information Systems. A customized device, placed within a rectangular PVC case compatible with the collar usually worn by animals, was equipped with a LP-GPS omnidirectional system, an integrated SigFox communication system and a power supply. The experimental trial was carried out in an existing semi-natural pasture characterized by good pasture allowance and cultivated grazing areas. Ten cows were embedded with LP-GPS collars and the data, i.e., geographical coordinates and the time intervals related to each cow detection, were recorded every 20 minutes. Data were collected through a specifically developed AppWeb to be further imported and elaborated by using a GIS software tool. In GIS environment, the daily distances travelled by each cow were linked with Heatmaps obtained by applying Kernel Density Estimation models from the points obtained from the LP-GPS collars. The results of the study made it possible to obtain information on some relevant aspects for livestock’s environmental issues. In detail, it was possible to acquire information on herd behaviour related to the use of the pasture, e.g., the area of the pasture most frequently used during the day, individual use of the pasture, possible animal interactions. These results represent a first step towards further insights and research activities because monitoring of animal locations could allow the reduction of several environmental issues such as soil degradation and greenhouses emissions.

Keywords

• Grazing cows
• IoT
• cow behaviour
• GIS
• spatial analysis.