student profile: Miss Jaime Manning


Thesis work

Thesis title: Heterogeneity in extensive pasture systems: the effect on beef cattle behaviour, selection, paddock utilisation and production

Supervisors: Gregory CRONIN , Lachlan INGRAM

Thesis abstract:

Managing grazing livestock can be a complex process. Cattle producers require a range of capabilities, from understanding cattle behaviour to ensuring sufficient pasture resources are available to meet the demands of the grazing animal. A key objective of beef cattle producers is to provide animals with access to sufficient quantities of their “preferred diet”, to achieve profitable animal production, whilst also ensuring the animals are maintained at high standards of health and welfare. Future expansion of the beef industry is likely, as the increasing demand for animal-based protein is driven by a combination of the growing world population, and increasing middle-class wealth in developing countries. However, animal welfare concerns have been expressed over the low frequency of livestock monitoring in extensive / rangeland management systems. This is especially relevant as herd sizes increase and farm labour inputs decline. There is a need therefore to improve on the traditional methods of managing and monitoring extensively produced livestock, and on how management strategies are implemented. In this global market, the livestock sector needs to increase productivity and production efficiency, for example through better utilisation of available pasture resources whilst also meeting consumer animal welfare concerns. The use of technology offers one solution, supplying producers with new techniques to manage livestock and implement strategies on farm. The majority of extensive / rangeland beef enterprises graze livestock in paddocks (pasture based systems), which are considered heterogeneous (non-uniform) in the quality and quantity of available pasture, both temporally and spatially. Cattle actively search their environment in order to select pasture based on quality and quantity attributes. Thus, cattle are referred to as selective grazers. Selective grazing however, can lead to adverse environmental implications if not managed appropriately. For example, cattle may overgraze desired areas and avoid other areas, resulting in overall poor utilisation of paddock resources. Additionally, there is limited information on the pasture quality factors that influence livestock site selection (time spent at a site or location). Improved understanding of pasture – livestock interactions are potentially the key to further improve pasture management and livestock production. Both of which have associated implications for farm profitability.

Chapter 1 highlighted the importance of understanding cattle behaviour, factors affecting animal – environment interactions and the quantification of site selection decision making for improved management and allocation of pasture resources. To investigate cattle site selectivity, Global Navigation Satellite System (GNSS) tracking collars (commonly referred to as Global Positioning System (GPS)) were placed around the necks of beef cattle, enabling the interaction between animals and their environment to be explored spatially and temporally. However, it is recognised that the attachment of a device to an animal could impede their ability to behave “normally”, potentially influencing research outcomes relevant to livestock production and welfare. Chapter 2 therefore examined the effect of GNSS collars on cattle behaviour and whether an habituation period to wearing a collar is required. That is, how quickly do beef cattle become accustomed to wearing a neck collar with an attached GNSS tracking device, or the duration before the animal’s time budget of behaviour returns to “normal” and collected data can be processed and interpreted. To determine if there were any behavioural time budget changes due to the presence of GNSS collars, collared (CD; n = 10) and non-collared (NC; n = 10) Charolais cows were compared. Welfare was assessed on the basis that if no behavioural differences were apparent between CD and NC cows, animals were therefore unrestricted and able to perform ‘normal’ activities such as graze, rest etc. Our findings indicated that GNSS collars weighing 0.61 kg or <0.1 % of liveweight had no negative effects on behaviour (P > 0.05) between CD and NC cows, with the exception of Stand stationary (P = 0.03). Although the days when Stand stationary was significant between CD and NC cows was prior and after the addition of a GNSS collar (Days 1 and 12). Hence, these differences cannot be attributed to the presence of a GNSS collar. Grazing is the behaviour of production importance, and no difference between CD and NC cows was found, emphasising that there should be no impact on enterprise production and profitability. Additionally, as the presence of a GNSS collar had no effect on behaviour in Chapter 2, also highlights that a high welfare standard was maintained. Furthermore, an habituation period to the light-weight collars used in these and future studies is not necessary, as highlighted by no significant behavioural differences during the first hour post collar deployment (P > 0.05). Therefore, data generated from GNSS collars can be reliably submitted for analysis straight after deployment.

The literature suggests that numerous pasture quantity and quality attributes influence livestock behaviour, selectivity and paddock utilisation. However, there is a large knowledge gap regarding how the different pasture attributes interact to affect site selection and paddock utilisation by grazing cattle. Cattle behaviour was examined in Chapter 3 using visual observations in response to changing pasture biomass, estimated via Normalised Difference Vegetation Index (NDVI). Additionally, GNSS collars enabled the determination of site selection choices and distances travelled by Charolais cows. As NDVI declined over the study (r2 = 1.00), distance travelled increased (P < 0.001; r2 = 0.88), and time spent grazing per day increased from 31 to 69% (P < 0.001; r2 = 0.71). Hence, highlighting the ability of cattle to adjust the duration of particular behaviours in order to meet nutritional requirements. Livestock tracking and pasture sensor technologies therefore, are potentially useful for providing bio-indicators reflecting the amount of pasture currently available to livestock. Such bio-indicators could also be refined to assist producers better manage pasture resources.

Whilst Chapter 3 identified the role of pasture biomass on livestock behaviour, it did not identify the influence of pasture quality attributes. Pasture quality analysis was conducted on a range of sown, non-sown and weed species, and is reported in Chapter 4. Variables analysed included; biomass, non-fibre carbohydrates (Fructose, Sucrose, Glucose), fibre carbohydrates (Acid Detergent Fibre (ADF), amylase and sodium sulfite treated Neutral Detergent Fibre (aNDF), Hemicellulose, Cellulose, Lignin, Total Digestible Nutrients (TDN), Non-fibrous Carbohydrates (NFC), Starch, Crude fat (EE)), organic acids (Malic acid, Citric acid), alcohols (myo-Inositol, Pinitol), protein (Nitrogen, Crude Protein) and minerals (Ca, Cu, Fe, K, Mg, Mn, Na, P, S, Se, Si, Zn). Species sampled for pasture quality analysis included sown species; Cocksfoot (Dactylis glomerata L.), Perennial ryegrass (Lolium perenne L.), Phalaris (Phalaris aquatica L.), White clover (Trifolium repens L.) and Subterranean clover (Trifolium subterraneum L.). In conjunction with non-sown species; Silver grass (Vulpia spp.) and Barley grass (Hordeum leporinum Link), and weed species; Shepherd’s purse (Capsella bursa-pastoris (L.) Medik) and Wireweed (Polygonum aviculare L.). There were significant differences between species for all pasture quality variables (P ≤ 0.05), apart from Starch (P = 0.47), Cu (P = 0.56) and Se (P > 0.05). Furthermore, the variogram output highlighted large variability across the paddock (spatial heterogeneity) for a number of pasture quality variables and species. Spatial variation highlights the importance of implementing site-specific strategies on-farm to manage areas that differ in performance (e.g., high and low quality) and sensitive regions (streams, dams etc.) across the paddock. Additionally, these findings reinforce the need to understand how spatial variation in pasture attributes influence livestock behaviour and utilisation patterns.

Previous studies of paddock production have focussed generally on singular aspects of pastures, such as biomass or quality variables, and have thus failed to take into account the complex interaction between paddock and pasture factors in influencing where grazing livestock spend time (selection). As such, the aim of Chapter 5 was to investigate herd site selection in relation to paddock factors (distance to water, shelter, fenceline and elevation) coupled with pasture biomass and quality attributes that were previously analysed in Chapter 4. The addition of GNSS collars enabled Angus heifers (n = 11) to be tracked over one month and the determination of sites selected. Factors that had the largest influence on site selection by the herd were paddock variables (close proximity to water and shelter) and NDVI. Cattle were predicted to be within 25 m of water and the nearest tree (shelter), followed by NDVI. Sites with low (<0.3) and high (>0.55) NDVI were selected by the herd. Yet, selection of low NDVI sites is related to the large role water and shelter had on the results, which inherently have a low NDVI. The selection of high NDVI reinforces the selective nature of grazing cattle, and their ability to seek out higher quality and actively growing regions. Interestingly, the study found that a large number of pasture quality variables did not influence site selection by the herd. Hence, such detailed analysis of pasture quality attributes is probably not required. However, a key variable for predicting site selection by the herd was NDVI, which is measured using remote sensing technologies. The findings support the use of pasture sensors (including NDVI) as an invaluable, relatively cheap tool to provide close to real-time and frequent information at a paddock level. The assessment of paddocks using NDVI can also be used to identify low and high performing regions, prior to cattle grazing, thus making pasture and livestock management more precise. Furthermore, by improving how pasture resources are allocated, profitability and productivity can potentially be improved.

Finally, stocking rate (SR; the number of animals per given area for a period of time) is the standard means by which producers allocate livestock depending on available pasture (feed) in extensively grazed systems. However, little is known about how cattle utilise their environment (paddock utilisation) under different stocking rates, in combination with potential effects on production variables (e.g., weight gain) and site selection differences. Hence in Chapter 6, three stocking rates (Light; n = 15, 0.12 steers/ha, Moderate; n = 22, 0.17 steers/ha and Heavy; n = 31, 0.24 steers/ha) were investigated at the end of a grazing season in a semi-arid ecosystem. There were no production differences between SR for liveweight (P = 0.23) or average daily gain (P = 0.54). The main driver of patch selection for all SR was daily change in NDVI, with cattle selecting sites of little or no change in NDVI. But differences in paddock utilisation were apparent between SR. Regardless of the paddock utilisation analysis undertaken (95% Minimum Convex Polygon (MCP), Utilization Distribution (UD) and 95% Kernel Utilization Distribution (KUD)), the Heavy SR utilised a significantly smaller area of the paddock (P < 0.001). In terms of MCP, the Heavy SR occupied 122 ha compared to 126 and 131 ha for the Light and Moderate SR respectively (paddock size = 128 ± 4.0 ha). Furthermore, the Heavy SR spent more time within close proximity to water (P = 0.005), implying that they were spending less time searching for and consuming available pasture. In order to make paddock utilisation and management improvements on farm, producers need to carefully consider the SR to ensure sufficient pasture resources are available and to minimise any potential negative environmental implications. Through the collation of near real-time information on animal behaviour and paddock utilisation, producers will have more accurate, lead indicators to assist decision-making and the development / refinement of future management strategies, rather than relying on lag information (e.g., production, liveweight). While remote sensing technologies have the ability to improve how we have traditionally managed livestock, future focus needs to be directed more at obtaining near real-time information or lead indicators rather than production or lag tools.

In summary, this thesis investigated the underlying pasture factors (quality and quantity) affecting cattle site selection, animal – pasture interactions, paddock utilisation, and the applicability of GNSS collars for livestock studies. The adoption of remote-sensing technologies to autonomously measure pasture and livestock variables also has the potential to improve animal welfare standards via more frequent livestock monitoring. Simultaneously, the acquisition of near real-time data should enable producers to improve management practices, for example by modifying livestock access to underperforming or sensitive regions of the paddock, and facilitating producers to make closer to real-time strategic decisions. The information reported in this thesis should also assist researchers in the process of applying remote sensing technologies for investigations on pasture and livestock interactions. Moreover, this thesis proposes a range of bio- or lead indicators/tools that could be developed for use by producers to assist management decisions at a paddock (pasture) and animal level.

Note: This profile is for a student at the University of Sydney. Views presented here are not necessarily those of the University.