The challenges of integrating geospatial technologies into livestock industries

Abstract

The demographic shift in farm labour in Australia and reduced recruitment of younger people has created an aging farming population and increasing labour shortage. With the advent and commercial availability of geospatial technologies, a sensor network has the potential to improve productivity by increasing situational awareness of the state of the pasture and animals. However, the deployment of autonomous recording sensors on free-ranging animals so that many variables can be monitored at rates of many times a second guarantees large data sets. The challenge for the livestock industries is how to manage the data so that livestock managers can make timely and informed decisions. For scientists there are also challenges with the availability and use of these technologies. They include appropriate experimental design by combining field observations in ways that will act synergistically with these tools and to produce improved understanding on how the large data sets generated by these tools can be used by livestock managers. Various sensor devices and networks are being used by producers across livestock industries. For example, sensors are being used to determine onset of puberty in dairy cows and farrowing in sows. They are also being placed inside the rumen of cattle to identify health issues such as acidosis. Other uses include the identification of foot problems in dairy cows. Many of these devices can interrogate the data and identify any changes outside 'normal' parameters. An on-animal sensor that has become popular with scientists and developers is the satellite-based global positioning system (GPS). This sensor is being used by scientists to track animals to determine their position and behaviour within the landscape. With the development of Real Time Location Systems (RTLS) in the non-agricultural sectors that incorporate GPS and accelerometers, their use in grazing livestock management is being tested within the University of New England's Precision Agriculture Research Group (PARG). Investigations within PARG have been focussing on how to interpret output from GPS and accelerometer sensors attached to free-ranging animals. One important aspect of this research is to identify data mining methodologies that will be suitable to "make sense of the sensor data". To determine if GPS collar data from grazing pregnant Merino ewes could be used to identify lambing, Dobos (2010) used change point analysis of average daily ewe speed. Other potential data reduction methodologies include identifying behavioural states (grazing, resting, travelling) in free-ranging cattle with GPS and accelerometers (Trotter et al. 2012). Current on-animal sensor research involves assessing jaw movements using accelerometers to help identify grazing and ruminating activity. Sensor technologies offer the scientist a data rich future but it is crucial that methodologies are identified to allow producers to make timely and informed decisions from these technologies. Technology has progressed to a point which allows endless amounts of information to be collected about agricultural systems, what is lacking is the framework to allow seamless cost effective meaningful data presentation to the farm manager.