st australian and new zealand spatially enabled livestock ......palmerston north new zealand email:...

1st Australian and New Zealand

Spatially Enabled

Livestock Management

Symposium

University of New England

Armidale NSW Australia

15th July 2010

Proceedings of the 1st Australian and New Zealand

Spatially Enabled Livestock

Management Symposium

University of New England

Armidale NSW Australia

15th July 2010

ISBN 978-1-921597-23-7

Editors: M.G. Trotter, D.W. Lamb and T.F. Trotter

Published by:

Precision Agriculture Research Group, University of New England, Armidale, Australia.

Printed:

July 2010.

Example Citation:

Donald DD and Daffy DD (2010) “GPS-enabled duck plucking”. In: Proceedings of the 1st

Australian and New Zealand Spatially Enabled Livestock Management Symposium. M.G.

Trotter, D.W. Lamb and T.F. Trotter (Editors), Precision Agriculture Research Group,

University of New England, Australia.

1st Australian and New Zealand Spatially Enabled Livestock Management Symposium

i

Foreword

Welcome to the 1st Australian and New Zealand Spatially Enabled Livestock Management

Symposium and welcome to the New England region and the University of New England

(UNE)! The university prides itself on its regional and global impact and, with a long history

of education and research in rural and environmental science, is particularly proud to host

this first symposium.

Spatial monitoring of livestock and their environment has been a rapidly expanding area

of research in Australia, New Zealand and around the world over the past few years.

Furthermore, systems for commercial deployment are currently being developed to

enable producers to make use of this technology. It’s an exciting time to be a researcher;

however we need to be aware of the many challenges we face if these technologies are

to be successfully adopted by industry.

Communication between technology developers, researchers, funding bodies and

producers is fundamental to ensuring a coordinated approach to research and to ensure

practical outcomes for industry. This is one of the primary functions of this symposium,

bringing together all these parties to share knowledge and build working relationships!

This symposium has a diverse range of speakers and, as a direct consequence of

discussions from last year’s GPS Livestock Tracking Forum, keynote speaker Toby Patterson

will provide insights into animal movement and behavioural modelling in ecology, an

allied research area from which the livestock industry can learn a great deal. So whether

you are a researcher intent on using the technology to understand the mysteries of plant

and animal interactions, or a producer just wanting to know where your cows are, we

hope you will glean something useful from the event!

Mark Trotter

Symposium Organiser

Acknowledgements

This symposium has been enabled by the support and resources of the CRC for Spatial

Information II (CRCSI2), established and supported under the Australian Governments

Cooperative Research Centres Program.

Funding for the Symposium has also been generously provided by Meat and Livestock

Australia and CRCSI-2.


ii

Program

11:00am Introduction & opening address

Jim Barber (UNE Vice Chancellor)

11:10am Why does it matter where animals urinate?

Keith Betteridge (AgResearch NZ)

11:20am Active Optical Sensors for grazing systems research

David Lamb (UNE PARG)

11:30am GNSS monitoring of temperament variations in cattle

Lauren Williams (University of Sydney – Masters candidate)

11:35am Spatially enabled livestock management: increasing biomass utilisation in

rotational systems

Jessica Roberts (UNE PARG – PhD candidate)

11:40am Lunch (ASAP close)

1:00pm Keynote Address: The state of the art in movement and behavioural modelling

in ecology

Toby Paterson (CSIRO)

1:45pm Research directions in Northern Australia – the development of CSIRO’s

Lansdown Research Station

Greg Bishop-Hurley (CSIRO)

1:55pm On saleyards and surveillance – using livestock movement records for risk-

based surveillance planning

Kerryn Graham (CSIRO)

2:05pm GNSS livestock tracking at the University of Sydney

Greg Cronin (University of Sydney)

2:15pm Linking GPS and satellite remote sensing to monitor animal behaviour and

environmental Interactions

Rebecca Handcock (CSIRO)

2:25pm New tools in spatio-temporal grazing systems research

Mark Trotter (UNE PARG)

2:30pm Understanding sheep grazing in complex pastures to better manage natural

resources and production outcomes

Felicity Cox (Charles Sturt University - PhD candidate)

2:35pm Spatio-temporal monitoring of sheep to investigate shelter and shade use

Geoff Hinch (UNE PARG)

2:40pm Afternoon tea

3:10pm The business case for investment in development of precision livestock

management technologies and applications

Rod Dyer (Meat and Livestock Australia)


iii

3:30pm Does peer training facilitate adaption in young cattle relocated from the

rangelands to a temperate agricultural grazing system

Dean Thomas (CSIRO)

3:40pm Precision livestock production and environmental influences

Graham Donald (CSIRO & UNE PARG)

3:50pm Translating industry research to farm profit improvement - a commercially

viable approach to precision livestock and remote monitoring in Australia

Chris Andrews (Taggle Systems)

4:00pm Bayesian change-point analysis of grazing sheep behaviour to identify lambing

Robin Dobos (Industry and Investment NSW & UNE PARG)

4:10pm Symposium close

6:30pm Informal dinner – Wicklow Hotel


iv

Table of Contents

Foreword i

Acknowledgements i

Program ii

Table of Contents iv

Why does it matter where animals urinate? .....................................................................................1

Active Optical Sensors for grazing systems research.......................................................................2

GNSS monitoring of temperament variations in cattle ...................................................................3

Spatially enabled livestock management: increasing biomass utilisation in rotational

systems ......................................................................................................................................................4

The state of the art in movement and behavioural modelling in ecology .................................5

Research directions in Northern Australia – the development of CSIRO’s Lansdown

Research Station .....................................................................................................................................6

On saleyards and surveillance – using livestock movement records for risk-based

surveillance planning .............................................................................................................................7

GNSS livestock tracking at the University of Sydney ........................................................................8

Linking GPS and satellite remote sensing to monitor animal behaviour and environmental

Interactions ..............................................................................................................................................9

New tools in spatio-temporal grazing systems research .............................................................. 10

Understanding sheep grazing of complex native grasslands to better manage production

and natural resource outcomes ....................................................................................................... 11

Spatio-temporal monitoring of sheep to investigate shelter and shade use .......................... 12

The business case for investment in development of precision livestock management

technologies and applications ......................................................................................................... 13

Does peer training facilitate adaptation in young cattle relocated from the rangelands to

a temperate agricultural grazing system? ..................................................................................... 14

Precision livestock production and environmental influences ................................................... 16

Translating industry research into farm profit: a commercially viable approach to precision

livestock and remote monitoring in Australia. ............................................................................... 17

Bayesian change-point analysis of grazing sheep behaviour to identify lambing ................ 18


1

Why does it matter where animals urinate?

Keith Betteridge and Des Costall

AgResearch Grasslands

Palmerston North New Zealand

Email: [email protected]

Mitigating N losses from agricultural land in NZ is of paramount importance for maintaining the clean, green image. Livestock urine provides most of the leached N and ~25% of our GHG emissions as N2O. But how does this vary spatially and temporally within paddocks, and does it matter anyway?

Previous research has shown hugely variable N concentrations in urination events of individual cattle within and between days and between cows. Thus, we believe there is no such thing as “average urine” within or between species.

As part of a larger programme we compared nitrogen leaching from sheep, cattle and deer grazed pastures. On those 0.4 ha, flat plots, N leaching from cow grazed pastures was nearly double that from ewes and hinds. But is this applicable to commercial hill country farms where stock are known to camp?

We developed urine sensors for cows and ewes which detected each urination event by the rise in temperature at a thermistor hanging below the vulva. This temperature is logged along with the time stamp. With continuous GPS logging, we match the urination time with GPS time to determine where urination events occurred.

Urine distribution maps overlaid on contour maps have enabled us to quantify the proportion of urination events excreted in all areas of a paddock. With cows, 50% of urination events were on 5-10% of the paddock areas which were typically low and flat, and/or near the water troughs. Sheep however distributed 50% of their urination events over 25-30% of the paddock area and typically camped on steeper land at a high elevation within the paddocks. We also found in one sheep trial, in autumn (warm days, frosty nights), that sheep were predominantly in the eastern sector of a hill paddock in the morning (sunrise warms the animal) when urine is most concentrated. However they resided in the western sector at night (warmer soils) when urine is less concentrated after a hot day with high water intake.

We are now testing a modified urine sensor that estimates urine volume and urinary nitrogen concentration ‘on-the-hoof’. This will quantify the variation in daily urinary N output amongst animals and amongst individual urination events of each animal. Importantly, the N modellers say that increasing the load of N in a urine patch increases N leaching exponentially. They are unsure if this also applies to nitrous oxide emissions.

By knowing where the most concentrated urine patches are in a hill country paddock, we can then target these small areas with nitrification inhibitors such as dicyandiamide (DCD) to mitigate N losses. This will allow farmers working under an ‘N Cap’ to increase stocking rates and productivity without increasing N emissions. So yes, it does matter where animals urinate, especially cows.


2

Active Optical Sensors for grazing systems research

David Lamb and Mark Trotter

CRC for Spatial Information and Precision Agriculture Research Group University of New England Armidale NSW Australia

Email: [email protected] Web: www.une.edu.au/parg

Efficiently measuring and mapping green herbage mass using remote sensing devices offers substantial potential benefits for improved management of grazed pastures over space and time. In particular, the integration of data from the spatial monitoring of livestock with spatio-temporal measures of biomass will offer considerable insights into pasture utilisation, one of the key drivers of profitability in grazing systems.

Several techniques and instruments have been developed for estimating herbage mass, however, they face similar limitations in terms of their ability to distinguish green and senescent material and their use over large areas. The Precision Agriculture Research Group has been exploring the application of an Active Optical Sensor (AOS) to quantify and map pasture biomass mass using a range of derived spectral indices (Trotter et al. 2010). In order to map pasture biomass, the AOS has been integrated with a Global Positioning System on a 4-wheel motor bike (Figure 1). The systems offers obvious benefits in terms of quantifying pasture biomass in situations where tree cover precludes the application of remote sensing (airborne or satellite) or where spatial variability in ground cover exceeds the spatial resolution of remote sensing instruments. There is also the opportunity to develop rapid assessment systems based on ‘representative transects’, thereby reducing the requirement to cover all ground in surveys.

Figure 1. Active Optical Sensor platform integrating a Crop CircleTM sensor with Trimble DGPS. (Inset Crop CircleTM showing optical emission foot print.)

References

Trotter MG, Lamb DW, Donald GE, Schneider DA (2010) Evaluating an active optical sensor for quantifying and mapping green herbage mass and growth in a perennial grass pasture. Crop and Pasture Science 61, 389-398.


3

GNSS monitoring of temperament variations in cattle

Lauren Williams1, Russell Bush1, Mark Trotter2 and Greg Cronin1

1 Faculty of Veterinary Science, University of Sydney, Camden NSW

2 Precision Agriculture Research Group, University of New England, Armidale NSW


Research on the measurement of temperament in beef cattle is becoming increasingly popular. Phillips (2002) described temperament as “a major parameter in the ‘personality’ or mood of cattle in relation to their reaction to man”. Some studies have shown that animals with ‘poor’ temperament have lower weight gain than calmer animals under both pasture and feedlot conditions. Furthermore, there is some evidence that ‘good’ temperament in cattle is genetically correlated with better meat quality (improved tenderness) and less carcass bruising. Temperament therefore, may have value as a predictor of growth and meat quality in beef cattle and could be integrated in the selection criteria for superior animals for breeding. However, while yard tests may predict temperament in cattle, a criticism of the tests is that they measure agitation and fear in response to human handling and yard experiences. One concern therefore, is that these temperament characteristics may not be relevant to production capability of cattle in the paddock. This experiment aimed to compare the effects of temperament, determined in yard tests, on cattle behaviour in the paddock. Behaviour was measured using two remote sensing devices (GNSS neck collars and a motion sensor attached to the rear leg) and direct observation. The effect of wearing the sensing devices on behaviour was assessed by comparing behaviour of animals with and without the devices over 6 weeks.

To investigate the relationship between temperament measured in yard tests and cattle behaviour in the paddock, the temperament of 64 Angus heifers was assessed at about 7 months of age. Three common behavioural yard tests were performed on the heifers during the two days after weaning in a yard: Test 1) the crush test, Test 2) the flight time test and Test 3) the human test. The tests were performed on heifers individually and testing occurred in the same sequence of testing, viz. Test 1, 2 and 3, respectively. Heifers were then ranked on their performance in each test and the three scores were combined to form a single rank score per animal. The 12 heifers at each end of the distribution were selected for inclusion in the experiment. The two treatments therefore were Low Temperament Score (Low; N=12) and High Temperament Score (High; N=12). Within each treatment, three heifers were selected at random to wear a GNSS unit attached to a neck collar (UNEtracker, Armidale, Australia). In addition, these animals wore an IceTag3D motion sensor attached to the left rear leg (IceRobotics Sensor Systems Ltd, Edinburgh, UK). The six heifers wearing GNSS collars and IceTag3D units (three Low and three High treatment heifers) and six heifers not wearing any measurement device (three Low and three High treatment heifers) were numbered using fluorescent spray paint for individual identification. The 24 heifers were weighed and maintained as a group in the yards for 8 days before being relocated to a 40 ha paddock. Behaviour observations were conducted on the 12 numbered heifers in the paddock over the next 6 weeks, and after 7 weeks, the heifers were returned to the yards, weighed and the GNSS collars and IceTag3D units were removed from the animals.

The location and behaviour (activity and posture) of heifers in the two treatments were measured using three methods – GNSS coordinates, IceTag3D activity and direct behaviour observation. The effect of wearing the remote sensing devices on heifer behaviour was measured by comparing the behaviour of the 3 (numbered) heifers per treatment wearing the devices with the 3 numbered heifers per treatment not wearing devices.

References

Phillips C (2002) Cattle Behaviour and Welfare. Second Edition. Blackwell Publishing Company, Oxford UK.


4

Spatially enabled livestock management: increasing biomass utilisation in rotational

systems

Jessica Roberts, Mark Trotter, David Lamb, and Geoff Hinch

Precision Agriculture Research Group, University of New England, Armidale NSW Australia

Email: [email protected]. Web: www.une.edu.au/parg

Increasing pasture utilisation is one of the primary means of increasing the productivity of livestock systems. The precise control of grazing pressure across a paddock can increase pasture utilisation above 50%, much higher than the current industry average of 30-40% (MLA 2009). In a recent review of new approaches to grazing management, Laca (2009) highlights the importance of understanding the spatial variability of pastures, grazing activity and the potential for information technology and spatial monitoring in livestock systems to inform management decisions and lift pasture use efficiency. Indeed, there are several real-time spatial monitoring systems currently in development for deployment in commercial situations (Stassen 2009).

Considerable effort has been put into developing systems that can determine animal behaviour from ‘on-animal’ motion sensors (Reed and Solie 2007) and recent research has combined motion sensors with spatial information to predict behaviour (Ungar et al. 2005). However, limited research has been undertaken into the potential for spatial data alone to predict the behaviour of livestock (Schwager et al. 2007). This is an important deficiency in the current knowledge as many commercial systems in development are limited to position sensors only and their application in grazing systems will be limited unless further research is undertaken. This project combines GPS tracking with visual observations of animal behaviour in order to determine whether spatial livestock data alone can be used to predict the biomass characteristics (quantity and quality) of the pasture. Moreover we aim to see whether such information can be used to as a trigger to moving stock in rotational systems.

References

Laca E (2009) New approaches and tools for grazing management. Rangeland Ecology and Managament 62, 407-417.

Reed S and Solie J (2007) Foraging detection of free-grazing cattle using a wireless motion sensing device and micro-GPS. In 'ASABE Annual International Meeting'. Minneapolis Convention Center, Minneapolis, Minnesota. (American Society of Agricultural and Biological Engineers).

Schwager M, Anderson DM, Butler Z and Rus D (2007) Robust classification of animal tracking data. Computers and Electronics in Agriculture 56, 46-59.

Stassen G (2009) Sirion, the new generation in global satellite communications: livestock GPS tracking and traceback. In '13th Symposium on Precision Agriculture in Australasia: GPS Livestock Tracking Workshop'. Armidale, Australia. (Eds MG Trotter, EB Garraway, DW Lamb) pp. 68-70. (Precision Agriculture Research Group The University of New England).

Ungar ED, Henkin Z, Gutman M, Dolev A, Genizi A and Ganskopp D (2005) Inference of animal activity from GPS collar data on free-ranging cattle. Rangeland ecology and management 58, 256-266.


5

Keynote Address:

The state of the art in movement and behavioural modelling in ecology

Toby Patterson

CSIRO Wealth from Oceans Flagship / Marine and Atmospheric Research, Hobart, Tasmania


Despite being fundamental to the dynamics of populations and communities, understanding the movements of individual animals has remained elusive.

However, the last few decades have seen an explosion of interest in questions surrounding animal movement. This has been driven largely by technological innovation which has lead to collection of high resolution data from many species in both marine and terrestrial settings. Advances in telemetry have been accompanied by the ability to record fine scale data from animal-borne sensors. These parallel data streams have lead to new interest in applying statistical techniques, developed in disciplines such as engineering and signal processing, to animal movement and allied data. Accordingly, movement ecology is beginning to mature as a discipline and is now starting to define a clear set of questions. A set of empirical and analysis techniques that can address these questions is beginning to emerge. Inevitably, such progress is not without setbacks. The complexity of the data on the one hand, and infatuation with fashionable, yet ultimately unhelpful theories and methods, has provided some distractions along the way. In this talk I will give an overview of several approaches to animal movement analysis from individual tracks and will follow the recent applications of techniques such as state space modelling.


6

Research directions in Northern Australia – the development of CSIRO’s Lansdown

Research Station

Greg J. Bishop-Hurley, Chris O’Neill, Luciano Gonzalez, Nigel Tomkins, Carlos Ramirez and

Ed Charmley

CSIRO Livestock Industries


Livestock are an integral part of the rangelands ecology in northern Australia. They account for over 45% of the agricultural economy in the north, but livestock production has to contribute to the restoration of the rangelands and not be responsible for their ongoing degradation. Responsible management through the development and adoption of precision livestock management solutions can play a significant role in optimizing productivity and environmental outcomes. CSIRO is developing a wireless sensor network (WSN) on its Lansdown research property, outside of Townsville, Queensland. Sensor networks can provide researchers with a level of information about the system not previously possible. The site is 638 ha of native and improved pastures with an average annual rainfall of 810 mm and a carrying capacity of approximately 200 head. The CSIRO Sustainable Agriculture Flagship goal of increasing productivity by 50% and reducing net carbon emissions intensity by at least 50% between now and 2030 aligns well with our research agenda. Our research will be focussed as much around environmental outcomes as around productivity and will focus on the following:

� refining the tools and science to monitor and control cattle in real-time in extensive systems;

� understanding the impact of cattle on carbon balance in a northern rangelands community; and,

� understanding animal behaviour and using this information to improve health and welfare.


7

On saleyards and surveillance – using livestock movement records for risk-based

surveillance planning

Kerynn Graham1, Peter Durr1 and Sandra Eady2

1Australian Animal Health Laboratory, Geelong, Australia

2CSIRO Livestock Industries, Armidale, Australia


Saleyards (or livestock markets) are critical points in the transmission of infectious animal disease. Not only is there mixing of animals from different farms, but the stress of travel and intermingling can lead to increased shedding and /or susceptibility to infection. In addition, where saleyards receive and dispatch animals over long distances, the possibility for rapid spread over wide areas becomes possible. This was the case of the foot-and-mouth disease incursion in the UK in 2001, where sheep markets were a critical determinant in allowing the initial outbreak to become an epidemic (Mansley et al. 2001).

One of the lessons learnt from the 2001 epidemic in the UK was the need for animal disease managers to understand the scale and extent of movements to and from saleyards as part of exotic animal disease (EAD) incursion preparedness. Traditionally this could only be approximated through surveys or expert opinion exercises, but the existence of national livestock tracing systems has made complete data analyses possible.

Following closely the spatial approaches used by Durr et al. (2006) in characterizing the abattoir catchment areas for cattle in the UK, we undertook an analysis of Australian cattle saleyards. By combining the data from the national cattle movement register (the National Livestock Identification System) with State and Territory farm property registers, it was possible to identify 38.7 million individual movements to and from 286 saleyards during the period 2006-07. Using Oracle Spatial 11g, we were then able to assign for each 100-km2 grid square of mainland Australia and Tasmania a figure for it’s relative importance for selling and buying cattle, and a probability of receiving or transporting cattle distances less than or greater than thresholds of 100-km, 200-km and 500-km. The resulting maps thus permit the identification areas of greater risk for transmitting disease and potentially become a keystone for constructing sophisticated risk-based surveillance systems.

References

Durr PA, Tait N, Greene J, McDonald R and Sinclair D (2006) Determining abattoir catchment areas to permit enhanced surveillance and welfare monitoring. Proc. ISVEE, Cairns, Australia, 6 - 11 August 2006.

Robinson SE and Christley RM (2007) Exploring the role of auction markets in cattle movements within Great Britain. Preventative Veterinary Medicine. 81, 21-37.

Mansley LM, Dunlop PJ, Whiteside SM and Smith RGH (2001) Early dissemination of foot-and-mouth disease virus through sheep marketing in February. Veterinary Record. 153, 43–50.


8

GNSS livestock tracking at the University of Sydney

Greg Cronin1, Russell Bush1 and Mark Trotter2

1 Faculty of Veterinary Science, University of Sydney, Camden NSW

2 University of New England, Precision Agriculture Research Group, Armidale NSW


The extensive livestock industries are becoming increasingly interested in the application of remote sensing technologies for monitoring livestock and measuring biometric data without the need to interfere with the animal by mustering and yarding. Universities have the important role in facilitating the education and training of young animal scientists of the future in the application of such remote sensing technologies for livestock management. To help achieve this objective, the Faculty of Veterinary Science at the University of Sydney is working with the Precision Agriculture Research Group at the University of New England to develop scientific and technical expertise to incorporate the application of remote sensing technology in the curriculum of animal science students. At the University of Sydney in 2010, a lecture / tutorial and a practical class were presented in the Third Year Animal and Veterinary Bioscience Unit of Study ‘Animal Behaviour and Welfare Science’ on the use of remote sensing devices such as GNSS tracking. In the lecture / tutorial, the students were informed on how the technology is currently being used by the livestock industry and how researchers at the Precision Agriculture Research Group are using the technology to improve understanding of the interaction between livestock and their environment, contributing to improved management of both livestock and pasture. In the practical class, students worked with GPS data from experiments on sheep, to introduce students to the possibilities for using remote sensing technology to aid the study of animal behaviour and to contribute to improved management of livestock and the resources in their environment.

In addition, and contributing to the overall objective of providing education and training in the use of remote sensing technology, two student projects (a Master of Animal Science and a 4th year Honours project) were conducted in 2010 at the University of Sydney using GNSS technology to track cattle and sheep, respectively.


9

Linking GPS and satellite remote sensing to monitor animal behaviour and

environmental Interactions

Rebecca Handcock1, Dave Swain2, Greg Bishop-Hurley1, Tim Wark3, Philip Valencia3 and Peter Corke3

1Commonwealth Scientific and Industrial Research Organisation (CSIRO) – Livestock Industries, Australia

2CSIRO - Livestock Industries, now at Central Queensland University, Australia

3CSIRO - ICT Centre, Australia

There has been increasing interest in how remote monitoring of animal behaviour in the environment can assist in managing both the animal and its environmental impact. GPS collars which record animal locations with high temporal frequency, with rates at rates of up to 4 Hz, allow researchers to monitor both an animal’s behaviour and its interactions with the environment. This GPS positional data on the cattle locations can be combined with high-resolution remotely-sensed satellite images to understand animal-landscape interactions. The key to combining these technologies is communication methods such as wireless sensor networks. We explore this concept using a case-study from an extensive cattle enterprise in northern Australia (Queensland) and demonstrate the potential for combining GPS collars and satellite images in a wireless sensor network to monitor behavioural preferences of cattle.


10

New tools in spatio-temporal grazing systems research

Mark Trotter CRC for Spatial Information and Precision Agriculture Research Group University of New

England Armidale NSW Australia Email: [email protected] Web: www.une.edu.au/parg

Behavioural observation tools

Traditional tools for recording animal behaviour are expensive (e.g. Observer XT, Noldus Information Technology) and require the purchase of separate mobile computers to enable field observation. An alternative system has recently become available as a cheap application (app) for Apple I-Phone, I-pod and I-pad devices. The behavioural observation app “WhatIsee” is available for less than $20 and can be used on a the I-pod touch device which costs around $250 substantially cheaper than traditional alternatives. The ethogram is simply entered into the touch interface and monitoring can be undertaken as a continuous or discreet process. Data is collected as a text string (.csv file). Data output is achieved by connection to wifi network for I-pod touch. Using WhatIsee on the apple I-phone, which has integrated GPS enabled geo-referencing of the records. This has potential for confirming the distance between observer and the subject being observed if this is critical. Furthermore the GPS enables the I-phone to function as a simple geo-referencing field device for other data for example pasture species mapping.

Data sharing tools

Collecting spatial data from animals is expensive and this has probably been one of the primary limiting factors in the development of tools and techniques for handling and analyzing this type of data. Ecologists have for many years realized the value in sharing data sets through spatial database management systems (Urbano et al.2010) we propose that data sharing may provide a solution to increase the possible to share data sets amongst researchers. Sharing of data sets would enable those researchers already involved in spatio-temporal animal research to test their processing techniques and modeling on different data sets. Of more potential benefit to this field of research is the making available of data to researchers who are not able to collect their own data effectively introducing a whole new segment of scientists to the spatio-temporal analysis of livestock movement data.

Wild animal studies are commonly undertaken within greater constraints than livestock researchers. Making high temporal resolution data and spatial data linked with behavioral observation available to those researchers in ecology may enable them to test models for which wild animal data can simply not be collected. Movebank (www.movebank.org/) is one example of a web based spatial database that can be used by researchers to enable sharing of spatio-temporal animal movement data. Whilst it has almost exclusively been used for wild animal studies, Movebank offers the potential to share livestock data in a controlled manner with the access designated by the person uploading the data as either publicly available or restricted. The level of acknowledgement for use of data in publications is also set by the owner of the data.

References

Urbano F, Cagnacci F, Calenge Cm, Dettki H, Cameron A and Neteler M (2010) Wildlife tracking data management: a new vision. Philosophical Transactions of the Royal Society: Biological Sciences 365, 2177-2185.


11

Understanding sheep grazing of complex native grasslands to better manage

production and natural resource outcomes

Felicity Cox1, David Kemp1, Warwick Badgery2 and Graye Krebs1

1Charles Sturt University, School of Agricultural and Wine Sciences, Orange NSW 2800

2Industry and Innovation NSW


This study aims to build on and contribute to current work in assessing the effect of grazing systems on animal production. To date little work has investigated an animal’s foraging behaviour, diet selection and diet quality within grazing systems. An understanding of these plant-animal interactions within grazing systems is a major knowledge gap limiting animal production, in particular within native grasslands and pastures. The study will investigate the relationship between grassland structure in terms of quantity and quality and animal behaviour including grazing location, plant species and plant part selection, diet quality and energy intake of sheep and how this varies over time according to seasonal variation, grazing management practices and landscape factors. This will be achieved by detailed pasture monitoring in conjunction with GPS technology and faecal sampling of grazing animals. Ultimately the study will build knowledge and better advise producers of what animals are actually doing in the field and how to use this information to improve management within south-eastern Australia.


12

Spatio-temporal monitoring of sheep to investigate shelter and shade use

Donnalee Taylor1,3, Wendy Brown1, Ian Price2, Mark Trotter3, David Lamb3 Derek Schneider3 and Geoff Hinch1,3

1 Animal Science, 2 Psychology, 3 Precision Agriculture Research Group

University of New England, Armidale, NSW Australia


In Australia inclement weather contributes to losses of new-born lambs and recently shorn sheep. Provision of forced shelter has been observed to reduce lamb losses by up to 10 percent and when given a choice, ewes preferentially seek shelter on offer, but for a limited period of around two weeks post shearing (Alexander et al. 1980). Given significant sheep losses can occur during adverse weather conditions a better understanding of sheep use of shelter and/or alternative ways of attracting sheep to shelter are needed.

This study reports on the results of deploying GPS collars on sheep on a commercial property in

the Northern Tablelands region of NSW Australia, (latitude: 30.99°S, longitude: 151.59°E and an

altitude of 1060-1151 m) with the aim of understanding the relationship between local climate and topography and sheep preference for shelter during pregnancy. In this work, two 20+ ha field designs were evaluated. Field A comprised of perimeter shelter belts (3-4 rows of native trees) and individual, free-standing trees (lone trees) within the field. Field B comprised of perimeter shelter belts, a single, internal shelter belt (‘boomerang’ shape) and a number of free-standing trees. Over two shearing and lambing seasons a random sample of 5 ewes from two flocks of 200-300 ewes (ranging from 2 to 5 years of age) were fitted with GPS collars providing continuous (43-51 days in Spring, September and October) observations of the ewes’ movement and proximity to shelter. GPS collars obtained and recorded individual animal location every 10 minutes. Weather stations and temperature loggers were strategically located throughout the fields to provide localized hourly measures of temperature, wind speed and precipitation during the two observation periods. Daily

minimum and maximum temperature ranged between -6 and 27°C respectively. Nights were

generally still and frost was common, days were often sunny and windy; hourly wind speed reached a mean maximum of 49.6 km per hour. Strong westerly winds prevailed; northerly and southerly winds were unusual. The mean annual rainfall for the two years was 724.8 to 795.4 mm mostly falling in the summer and winter months. The region experiences warm summers, rain and sleet are not uncommon in early spring and winters occasionally have light snow and are cool enough to inhibit plant growth markedly for about 4 months.

During night camping and where the choice of an internal shelter belt was provided (Field B), sheep were observed to spend more time in the vicinity (0-25m) of interior shelter belts than free-standing trees. In Field A where only internal trees or perimeter shelter belts were available, a slight preference for the free-standing trees was observed. During daytime, shade-seeking behaviour indicated an increase in preference for individual, internal trees in both fields compared to night-time camping. Night camping also used interior shelter or free-standing trees which could be associated with predation prevention, however, the tree canopies reduction of night radiation loss maybe of more importance to the sheep than wind protection provided by the exterior shelter belts. These results suggest sheep prefer to manoeuvre in and around shelter and freestanding trees provided within a field rather than huddle against exterior shelter belts provided along fence lines. The effects of local climatic temperature ‘extremes’, wind direction, altitude and diurnal movements on daytime and night time preferences are also expected to play a role in the observed shelter-seeking behaviour and are in the process of analysis. Alexander G, Lynch JJ, Mottershead BE & Donnelly JB. (1980). Proceedings of the Australia Society Animal Production 13, 329-332.


13

The business case for investment in development of precision livestock management

technologies and applications

Rodd Dyer

Meat & Livestock Australia

Email: [email protected]. Web: www.mla.com.au

Ongoing decline in terms of trade for agriculture requires beef producers in northern Australia to continually improve production efficiency and reduce costs of production to remain profitable and economically sustainable.

Meat & Livestock Australia (MLA) recently commissioned a study to identify the economic benefits and highlight investment opportunities of Precision Livestock Management (PLM). In this context PLM refers to a range of existing and emerging technology applications with potential to improve the efficiency and cost effectiveness of individual animal or herd measurement, monitoring, movement and management.

The scope of PLM applications considered included: walk-over-weighing, automatic drafting; individual animal recording, and decision making; machine vision applications; unmanned aerial vehicle (UAV) mounted sensor platforms for infrastructure monitoring, livestock location and mustering; and virtual fencing. Many were considered as bundled applications.

Potential on-farm PLM applications were identified and described for representative enterprises in four regions. The enterprise capital costs, operating costs and gross benefits associated with implementing each PLM were estimated using an economic herd model. The net present value (NPV) and internal rate of return (IRR) was calculated for each enterprise and used to determine the net benefit and rank of applications.

The net economic benefits of PLM applications for the northern industry were calculated from the estimated costs of research, development and commercialization needed to get each PLM application to market and the total net benefits from business thought likely to adopt each PLM application in each region.

The study suggested that PLM applications produced only poor to moderate net industry benefits. Unmanned aerial surveillance and UAV assisted mustering had benefit cost ratios (BCR) of 4.7 and 1.9 respectively. They were identified as the PLM application with most potential for positive industry returns and candidates for MLA investment in the short-term. Investment in remote livestock management systems that combine walk-over-weighing, auto drafting, telemetry and individual animal identification produced a BCR of 1.1 suggesting only limited scope as a longer term investment. Virtual fencing, applications utilizing artificial recognition technology (machine vision) and use of NLIS data to improve herd fertility did not produce a positive net present value.

These results are not conclusive. The study has highlighted serious analytical challenges associated with determining the net economic benefits of complex investment options.


14

Does peer training facilitate adaptation in young cattle relocated from the rangelands

to a temperate agricultural grazing system?

Dean T. Thomas, Matt Wilmot and Dean Revell

1CSIRO Livestock Industries, Private Bag 5, Wembley, WA 6913, Australia.


This study compared the growth and behaviour of young cattle that were relocated from the Western Australian rangelands (Wiluna) to a pasture in a more temperate agricultural region (Dongara), with and without the presence of local ‘peer trainer’ cattle present in their new paddocks. One hundred and eighty four rangeland-raised Brahman-cross heifers were transported to the experimental site. The cattle were split into 6 groups, comprised of 2 treatments (with peer trainers (peer trained) or without (control)) x 3 replicates and each group was grazed on a pasture (predominantly annual and sub-tropical perennial grasses) plot for 6 weeks. Paddocks were stocked at 1 cow per hectare, with peer trained groups comprised of 25% peer trainer animals. Liveweight gain and condition score were measured at days 0, 14, 28 and 36 and grazing behaviour was recorded continuously (at 5 minute intervals) using Global Positioning System (GPS) tracking collars with activity (head movement) sensors installed. The activity sensors measured minimum and maximum tilt angle of the vertical (front-to-back) and horizontal (side-to-side) rotation arc of the collar at 5 minute intervals. Kernel density distributions (ArcMap 9.3, ESRI, Australia) were plotted for GPS positions of both treatment groups during the first week after relocation.

During the 6 weeks following relocation, the peer trained cattle had 34% lower horizontal head movements (suggesting less grazing activity) compared to the control group (30.1 v 40.3 δ pitch angleº/5 minutes; P<0.01; Figure 1). There was a significant time of day x treatment interaction for horizontal head movement (P<0.05), indicating that the diurnal pattern of grazing activity differed between the control and peer trained groups. Head movement activity tended to be more variable in the control group (Figure 1) and there was a greater range in mean weekly head movement in the control (20.1 – 51.1 δ pitch angleº/5 minutes) compared with the peer trained cattle (25.3 – 34.3 δ pitch angleº/5 minutes). There was no effect of peer training on the daily distance travelled by cattle (P>0.05). Kernel density distributions indicated that relocated cattle that were combined with peer trainers explored a greater proportion of the paddock during the first week after relocation. The changes in grazing behaviour associated with peer training did not lead to differences in the growth of cattle following their relocation. All cattle gained weight over the 6-week period following relocation, however, liveweight gain during weeks 5 and 6 were more than double that during the first 4 weeks of the study (P<0.01). There was a trend toward lower growth in the peer trained cattle compared with the control group (0.70 v 0.86 kg/day; SEM = 0.052; P = 0.097) which may have been associated with relocated cattle receiving antagonistic behaviour from the local cattle, but this was not observed or quantified. Under other circumstances these behavioural effects might affect cattle performance positively, such as when cattle are moved to a more complex grazing environment. However, in this experiment feed, water and shelter were readily available there appeared to be no production benefit from peer training.


15

0

20

40

60

80

0 4 8 12 16 20 24

Hours

Head movement

(δ pitch angleo/5

minutes)

Figure 1. Mean daily horizontal (side-to-side) head movement (δ pitch angleº/5 minutes) of pastoral

cattle either with (□, Peer Trainer) or without (■, Control) local cattle present during six weeks

after relocation to a new agricultural pasture paddock.


16

Precision livestock production and environmental influences

Graham Donald1, Ian Purvis1, Mark Trotter2 and David Lamb2

CSIRO, Chiswick Research Station, Armidale, NSW 2350, Australia

Precision Agriculture Research Group, University of New England, Armidale, NSW 2351, Australia.


Precision livestock production (PLP) involves the manipulation of livestock activity within the environment to improve production and best utilize spatial variability. The livestock enterprise consists of animal, plant, soil, topography, climate, and human subsystems and is complicated by the various interactions between these. Precision agriculture offers the ability to measure many aspects of the livestock system enabling more precise and timely management decisions to be implemented. Precision farming has become a major driver within the cropping and horticultural industries. In contrast, the livestock industry has only just begun to investigate using these sensor tools to improve production efficiency. Based on the implementation of accurate timing and geo-locational tracking devices it may become possible to identify and quantify the individual animal’s pasture intake, nutritive value of selected forage, digestibility, grazing intervals, feeding pauses and rate of consumption. PLP also use automated weighing systems which provide a real-time means of relating the plant to livestock production by measuring live-weight change and the rate of change.

Linking the spatial monitoring of livestock with autonomous weighing technology may enable the calculation of feed conversion efficiency (residual feed intake). This will obviously benefit the livestock industry seeking to increase production from reduced inputs whilst minimising environmental impact of grazing (Johnson et al. 2003). It is proposed that these measurements will provide the information to identify individual livestock that have the propensity for higher production. In addition to this core research the application of other precision agriculture tools will be examined including remote sensing mechanisms such as active proximal sensors, airborne and high and low resolution satellite data for quantification of feed on offer and pasture growth rates. Other issues such as methane production and animal welfare (e.g. internal parasite infestations) will also be investigated by integrating spatial and allied technologies to record physiological attributes.

The proposed research would identify the interactions between livestock’s social behaviour, herd/flock structure and the relationships between key environmental factors and the resource utilisation of an animal as a means to improve livestock production.

References

Johnson DE, Ferrell CL and Jenkins TG (2003) The history of energetic efficiency research: Where have we been and where are we going? Journal of Animal Science 81, 27-38.


17

Translating industry research into farm profit: a commercially viable approach to

precision livestock and remote monitoring in Australia.

Chris Andrews

Taggle Systems


Taggle Systems has developed a low cost method of tracking and monitoring remote objects (e.g. livestock). The system is made up of small, long life tags (small enough to be used as an ear tag and lasting greater than 3 years) that are capable of transmitting over a long range - in excess of 7 km, with 20+ km pure line of sight. The tag transmissions are picked up by low cost base stations, developed using radio telescope technology. A single base station is capable of picking up telemetry data over areas as large as 300 km2.

These tags can be connected to a range of sensors to provide remote telemetry data readings such as: Water flow rate and depth, soil moisture content, wind speed and temperature, open/closed state for gates, irrigation controls etc.

If the tag transmission is picked by three base stations, a location for the tag can be calculated to approximately ± 15 meters allowing for significant improvements in areas such as Improved muster rates, improved land utilisation (grazing density mapping), behaviour alerts –wandering, theft, calving and feral pest control/tracking.

The system has been in development for 2 years and is now in commercial trials with networks in Tweed Heads and Mackay (Koumala).

Figure 1. A producer with a Taggle ear tag applied to a cow. (Inset: Taggle tag.)


18

Bayesian change-point analysis of grazing sheep behaviour to identify lambing

Robin Dobos1, Lee Taylor2 and Geoff Hinch2

1Industry & Investment NSW Beef Industry Centre of Excellence, Division of Primary Industries, Armidale, Australia, 2351

2School of Environmental & Rural Science, University of New England, Armidale, NSW, Australia, 2351

The main aim in analysing data for animal movement is to reveal behavioural mechanisms by which the animal utilises complex and variable environments. Movement data also reflects behaviours that are heterogeneous. Statistical analysis of multidimensional, auto-correlated and irregular interval movement data is difficult. Animal tracking devices such as collars with global positioning system (GPS) capabilities enable continuous and automatic tracking of an animal’s position and the value of such spatial–temporal information is improved if the corresponding activity of the animal is known. GPS animal tracking devices (collars) were used in the spring (Sep – Nov) of 2008 and 2009 to monitor movement of 20 pregnant grazing fine wool Merino ewes. Mean daily velocities (m/s) were calculated for three time periods within a day: 0500 - 1100h, 1200 – 1800h and 1900 – 0400h and subjected to Bayesian change-point analysis (BCP) in an attempt to identify when a change in behaviour (reduction in velocity) occurred and if this was associated with lambing. The approximate day of lambing, within 48h, was known to have occurred between 16 and 29d into the data collection period for both years. In 2009, 5 ewes were closely monitored for lambing date and BCP successfully identified a change-point that could be associated with lambing in the majority of ewes monitored. This method was then applied to the movement data collected in 2008 to identify lambing for that season. Based on the analysis, the time period 0500 – 1100h appeared to be the best in which to determine lambing using BCP. Posterior probabilities and means calculated using BCP appears to be a useful and robust methodology that could be incorporated into decision support tools for farmers to help in decision making based on movement data from GPS collars.

st australian and new zealand spatially enabled livestock ......palmerston north new zealand email:...

Documents