Preliminary indications that Merino sheep graze different areas on cooler days in the Southern Rangelands of Western Australia Dean T. Thomas A,C , Matt G. Wilmot A , Mark Alchin B and David G. Masters A A CSIRO Livestock Industries, Private Bag 5, Wembley, WA 6913, Australia. B Department of Agriculture and Food Western Australia, PO Box 108, Meekatharra, WA 6642, Australia. C Corresponding author. Email: dean.thomas@csiro.au Abstract. High ambient temperature can decrease weight gain in livestock, but domestic livestock in extensive grazing systems may have some capacity to adjust their behaviour to reduce the impact of periods of high temperature. We hypothesised that sheep grazing in the semiarid Southern Rangelands of Western Australia would reduce distance travelled on days with high mean daily temperature. Eight Merino ewes were fitted with GPS collars for 3 weeks in a 5575-ha paddock on Carlaminda station (28  20 0 S, 116  41 0 E). Mean daily temperature was used to separate the 18 days of the study into three temperature classes, cool (23.2  C), warm (23.3 to 25.9  C) and hot (26.0  C). Sheep travelled more quickly (P < 0.05) and further from water on cool days, compared with warm and hot days (3.74 v. 2.93 and 2.73 km from water, respectively; P < 0.001). On cool days, sheep spent most of their time grazing in the western area of the paddock. This area was rarely visited on warm or hot days. Mapping livestock distribution may assist in strategic relocation of existing water points and/or justify the development of additional watering points. However, our results suggest that sheep adjust their behaviour during hot weather, which may be a strategy to conserve energy, manage higher water requirements and/or reduce thermal load. Although grazing range decreases with higher temperatures, overall utilisation of a paddock may not be severely affected unless cooler days were too infrequent to facilitate regular access to these areas. Introduction Global positioning system (GPS) monitoring of livestock could reduce costs substantially in extensive grazing systems, by providing a labour-efficient means of collecting real-time information about grazing patterns, animal performance and watering frequency. Long-term, uninterrupted data acquisition is important to understand livestock activities in extensive grazing systems (Scheibe et al. 1998). It is likely that GPS collars with remote data access integrated with biological monitors, such as liveweight change and body condition score, will facilitate the use of animal behaviour as an indicator of pasture utilisation and allow timely management interventions. This study presents an example of the use of GPS collars to monitor grazing at different temperatures and demonstrates how this could be used in paddock design. High summer temperatures and few permanent natural water sources in the Southern Rangelands of Western Australia necessitate the provision of water points for grazing livestock. Therefore, the distribution of water points within paddocks has a key influence in how rangeland pastures are utilised, particularly over the summer period. Weight gain in livestock may also be depressed as temperature increases. For example, ambient temperatures above 26  C (range 19.5--33.1  C) can reduce liveweight gain in sheep (Murthy et al. 2005). Sheep expend more energy as body temperature increases [338 kJ/kg body mass 0.75 per  C body temperature] (Shinde et al. 1998) and this produces an energetic penalty to grazing in conditions that raise body temperature. Since livestock adapt their behaviour in response to weather events (Conradt et al. 2000) and to improve foraging efficiency (Stephens and Krebs 1986), we hypothesised that sheep grazing in the Southern Rangelands of Western Australia would restrict their travel range on days with high mean daily temperature. Materials and methods Experimental site The experiment was conducted on Carlaminda station, 10 km west of Yalgoo (28  20 0 S, 116  41 0 E), Western Australia. Ten-year (1997--2007) mean annual rainfall at the site is 247 mm, and for the 12 months before the study there was 156 mm of rain (mostly during winter--spring). Ten-year (1997--2007) mean average daily temperature recorded during October at the site is 20.2  C, with less than 10% of days above 25  C. Historical rainfall and temperature data reported in this study are based on monthly rainfall gridded data provided by the Australian Bureau of Meteorology. The paddock selected for the grazing study was 5575 ha and was stocked with 394 adult Merino ewes and 130 Merino lambs. The paddock contained four water points (north-central, east-central, south-central and south-east in the paddock, located on the boundary) that were fitted with total grazing management (TGM) yards. All livestock were removed from the experimental site immediately after this study when feed supply was low due to low seasonal rain. The paddock mainly comprised alluvial saline wash-plains, which have very scattered to moderately close low shrublands CSIRO PUBLISHING www.publish.csiro.au/journals/ajea Australian Journal of Experimental Agriculture, 2008, 48, 889--892 Ó CSIRO 2008 10.1071/EA08061 0816-1089/08/070889