AN exciting new piece of technology has been developed to prototype stage, aiming to allow feed intake, and by default, feed conversion efficiency, to be accurately monitored in cattle grazing in pasture situations.
The eGrazor project relies on a non-invasive collar-based sensor system fitted to cattle, using inertia measurement units and other sensors.
The “gee-whiz” part of the technology relies on how these sensors detect certain changes in the way the animal is behaving, and the way those ‘behaviours’ are classified into groupings, like grazing, standing, walking or resting. Each of those outputs is correlated against benchmark measurements of feed intake, to derive a prediction algorithm that estimates feed intake.
Data is captured constantly while the solar-powered collar is fitted (see picture below), and downloaded in real time for analysis.
While it’s been possible to measure cattle for feed intake on an individual animal basis in feedlot environments for some years using testing stations, never before has the industry had access to a technology that could do so on pasture.
Researchers with a range of highly-specialised skills are working on the project, including people with advanced knowledge in data analytics, who can tease-out the different characteristics of each animal’s movement recorded on the collars, and associate that with a particular behaviour, such as grazing. This process is called behavioural annotation and classification.
“It has been a ‘holy grail’ for a long time to find a way to measure fed intake in a paddock situation”
“We know there is considerable variation in the efficiency of an animal to convert pasture into growth, or kilogram of calf weaned, but it has been a ‘holy grail’ for a long time to find a way to measure that,” CSIRO’s Dr Drewe Ferguson said.
Dr Ferguson, research director for CSIRO’s Productive and Adaptive Livestock Systems program, said while the industry had always been able to measure ‘output’ in the form of kilograms of weightgain, or kilograms of calf weaned per cow – it has never been possible to measure how much pasture is eaten by each animal to achieve that.
The collars promise to give industry for the first time, data on what sort of variability exists in feed intake between pasture-based animals, and hence feed conversion efficiency. It will also provide useful guidance on whether animals on a grain-based ration rank the same as those on a pasture-based diet, for feed intake and efficiency.
The eGrazor project is a collaboration between NSW DPI and CSIRO’s Agriculture and Food and Data 61 program.
The two project leaders are NSW DPI’s Dr Paul Greenwood, and CSIRO’s Dr Aaron Ingham.
Dr Greenwood’s focus has been on experimental work generating benchmark data on intake of individual animals, on which to gauge the collars’ performance.
Where is the project up to?
A prototype of the collar recording unit, pictured above, has been developed and patented, using the working title, eGrazor. It is yet to be determined if that name continues through to final commercial product stage.
While there is still ongoing development work happening, the R&D team is now starting to look at deployment and applications across industry.
“Once we can capture intake on pasture, it is relatively easy to determine output – either in terms of growth rate or in the case of maternal performance – to get at the individual animal’s feed efficiency on pasture,” Dr Ferguson said.
“We want to get to a point where we can identify the superior performing animals, regardless of whether feed is at its peak, or haying off and looking sparse.”
Dr Ferguson said at this stage the project was focused primarily on feed intake, but over time other important traits such as feed selectively (already known to be quite variable between animals) would potentially come into focus.
“In pasture-based systems, there are individual animals that will extract much more out of their available feed base than others, simply because they are more selective grazers,” he said.
Determining accuracy important step
Determining accuracy of the eGrazor assessment is obviously another important step.
“The accuracy is being measured against the proxy ‘gold standards’ that we have for measurement of intake,” Dr Greenwood said.
What’s being seen at this stage of development is a highly encouraging degree of correlation, based on predictions delivered by the sensor outputs.
Dr Greenwood said the initial algorithms showing “very promising results”, would only improve as the current process and the way the information is interpreted is further refined.
“We need to make sure those initial results stand up to scrutiny,” he said.
Every bit as important to be able to identify the ‘duds’ as it was to pick the ‘stars.’
Dr Ingham said there was a belief that in terms of ranking animals on feed intake for genetic selection purposes, it was not going to be necessary to have 100pc accuracy in the eGrazor data, however.
“If that assumption proves to be correct (still to be determined), we think we are very close to delivering a product that is ready to go,” he said.
He agreed that it in such genetic selection, it was every bit as important to be able to identify the ‘duds’ as it was to pick the ‘stars.’
Part of the R&D process from here, apart from fine-tuning, is in determining things like how long a collar needs to be fitted to an animal to deliver a reliable reading; best test protocols; and methods to measure and compare feed intake in different pasture conditions and regions.
There will obviously have to be adjustments for different pasture types, or pasture availability – paddocks with say 2200kg of DM per hectare, versus 1400kg. These issues are currently being addressed, researchers said.
Dr Greenwood said the initial trial work had been done within a specific type of grazing system (improved temperate type pastures on the NSW northern tablelands), to provide ‘proof of principle.’
“The objective in the first instance is to run experiments under those conditions, as to whether the approach allows prediction, with some degree of reliability, of the amount of pasture that an animal is eating. The next phase will be looking at transportability of that work, across different grazing environments, and variations in the pasture sward within a region,” he said.
The developers are now in early discussion with commercial partners about the manufacture of the technology for the commercialisation phase, Dr Ingham said.
“If these devices start to be produced at scale, as always, the cost compared with the prototype, is likely to drop dramatically,” he said.
Applications in seedstock
While the eGrazor technology will also have valuable application in the research field, its greatest impact is likely to be seen in adoption in the seedstock industry, for genetic selection based on feed intake/conversion performance on grass.
It’s not unrealistic to expect the emergence over time of Breedplan EBVs for feed intake on paddock-fed bulls – equally important for growth rate as in maternal performance, the research team said.
Discussions are already underway with several co-operator seedstock producers looking at the practicalities of delivering and using the eGrazor devices in herd measurement, including issues in scaling-up from small numbers of animals to larger commercial applications.
“If we can generate a herd that is more efficient, then we are going to be way ahead of the game,” Dr Ferguson said.
Asked to estimate the likely timeframe from here to commercial delivery, Dr Ferguson said it was probably still another 12-18 months away from full launch, but there was a lot of work already in the pipeline to deliver on that.