*Present address and correspondence: Karen L. Viggers, 19 Bindaga St, Aranda, ACT, 2614, Australia (fax +2 6125 0654; e-mail email@example.com).
1The movement of native herbivores onto agricultural land is a key management issue world-wide as they may compete with domestic livestock for pasture and contribute to overgrazing and soil erosion. Eastern grey kangaroos Macropus giganteus are viewed as a problem species in temperate south-eastern Australia, where high-density populations in reserves encroach on nearby farmland.
2This study examined home range size and use in M. giganteus across different types of land use and in relation to population density and pasture availability.
3Farmland adjacent to either radiata pine plantations or reserves supporting high-density populations of M. giganteus was subject to frequent incursions by kangaroos moving onto farmland to rest or graze. However, animals from reserves moved on average only 135 m onto farms.
4Home ranges of M. giganteus were significantly smaller in the reserves than in farm study sites where population densities were lower. At reserve sites, home range size was limited by higher population densities and limited opportunity for dispersal across surrounding open farmland because of a lack of cover. Home range size was not affected by resource availability.
5Where suitable vegetation cover occurred on farmland (e.g. woodland remnants or scrub), M. giganteus occurred as resident or roving small mobs. This may be seen by farmers as a disincentive to preserve remnant vegetation as it can provide habitat for unwanted native wildlife.
6Home range attributes of M. giganteus suggest the species could be controlled by culling. However, recolonization occurs quickly and little is known of dispersal.
7Synthesisandapplications. Population density, presence of cover and reluctance to disperse across cleared landscapes are key factors influencing kangaroo home range size and use of adjacent farmland. Currently, little incentive exists for farmers to preserve remnant vegetation, as it may be regarded as providing habitat for unwanted or ‘pest’ kangaroos. Given the potential importance of remnant vegetation on private land for the conservation of plants and other species of wildlife, government incentives and compensation programmes may be required to limit land clearing on farms and to encourage improved pasture management.
In Australia, large kangaroos (macropods) can occur in great numbers on agricultural land. Most studies of ranging behaviour in kangaroos, and the level of competition with domestic livestock, have focused on red kangaroos Macropus rufus Desmarest and western grey kangaroos Macropus fuliginosus Desmarest in the pastoral rangelands (Caughley, Shepherd & Short 1987; Edwards, Dawson & Croft 1995) and the western Australian wheat belt (Arnold & Steven 1988). Although there have been some previous studies examining the home range of eastern grey kangaroos Macropus giganteus Shaw in temperate south-eastern Australia (Jarman & Taylor 1983; Jaremovic & Croft 1987, 1991; Coulson et al. 2000; Moore, Coulson & Way 2002), none has examined home range across different types of land use (e.g. reserves, forestry plantations and agricultural land). However, the highest densities of M. giganteus have been recorded in the Australian Capital Territory (ACT) in south-eastern Australia: 233 km−2 in reserves and 50 km−2 on farmland (Kangaroo Advisory Committee 1996). Given the size and density of populations of M. giganteus, and the potential for impacts on farm productivity and land degradation, information on how these macropods interact with the different landscape patterns is urgently needed.
Macropus giganteus is a medium-sized crepuscular and nocturnal grazing marsupial, with females weighing up to 40 kg and males up to 90 kg (Kaufmann 1975; Dawson 1995). The species occurs in open forests, woodlands and grasslands, and prefers areas close to trees and shrubs, which it utilizes for resting and as refuge from predators and high daytime temperatures (Caughley 1964; Taylor 1980; Hill 1981; Terpstra & Wilson 1989; McAlpine et al. 1999). Macropus giganteus can move quickly and can travel long distances (Kaufmann 1975; Dawson 1995). However, it can also be sedentary, with relatively small home ranges (Jaremovic & Croft 1991) and slow dispersal (Jarman & Taylor 1983). Most female M. giganteus give birth in the summer months of each year (Poole 1983; Banks 1997) and, if conditions are favourable, populations in productive temperate landscapes can reach very high densities. This is important in rural landscapes where kangaroos move between reserve systems and agricultural land (as observed in the closely related western grey kangaroo M. fuliginosus; Arnold & Steven 1988; Arnold, Steven & Weeldenburg 1989; Arnold et al. 1992). In such circumstances, large and apparently mobile populations of kangaroos have the potential to compete with domestic livestock for food and, in turn, contribute to overgrazing of pastures, land degradation and soil erosion.
The aims of our study were to: (i) estimate home range size of M. giganteus in areas subject to different land uses, kangaroo population densities and vegetation types; (ii) compare home range size, location, shape and activity centres of M. giganteus populations utilizing reserves and farms or moving between both; and (iii) examine the relationships between home range size, home range use, site, season and pasture availability.
We outline some of the key management implications of our findings and relate these to wildlife grazing management issues elsewhere around the world.
Three broad types of land use were selected for study, representing a gradient of intensity of land management and vegetation cover conditions. These were: farmland used for grazing domestic livestock (sheep, cattle and horses), a water catchment reserve and a wildlife conservation reserve (Table 1). Although replication at the site- or land-use level was not possible because of logistical and resource limitations, we selected sites that not only covered a range of land uses but also different vegetation types: native and exotic grassland, eucalypt woodland, remnant woodland patches, radiata pine Pinus radiata D. Don plantation and non-native pasture.
Table 1. Details of study sites in south-eastern Australia
Farm managed for horses, sheep and cattle
Grasslands (occasional paddock trees)
Remnant woodland patch
Googong Foreshores Reserve
Water catchment reserve
Tidbinbilla Nature Reserve
Wildlife conservation reserve
On the basis of previous population density estimates across ACT from walked line transects (Kangaroo Advisory Committee 1996), M. giganteus populations at our study sites were classified as high density (reserves) or low density (farms).
The Cotter property (149°01′E, 35°21′S; Fig. 1a) covers 610 ha and is bounded on the west and north by radiata pine plantations and to the south by the Murrumbidgee River Corridor (MRC), and contains a substantial (115 ha) patch of remnant open woodland. The remainder of the property largely comprises native and introduced grasslands with occasional scattered paddock eucalypt trees Eucalyptus sp. The steeper southern end of the property has small patches of sparse scrub that are loosely connected to the relatively heavily wooded MRC reserve. The property is managed to provide grazing for 70 horses, 660 sheep and 50 cattle. Some paddocks are periodically rested from domestic livestock grazing. An annual cull of M. giganteus is carried out between April and July. In 2001, all kangaroos that could be detected were culled (c. 180) apart from radio-collared animals in this study that were exempt from being shot.
Googong Foreshores Reserve
The Googong Foreshores Reserve is located in New South Wales (NSW) close to the ACT/NSW border (149°15′E, 35°10′S; Fig. 1b). The site is a water catchment for the Googong Reservoir and is managed by the ACT government for water production. Open woodland and grassland habitats supporting substantial populations of M. giganteus occur on the western shores of the reservoir (Muranyi 2000). Adjacent to this is freehold farmland where sheep are grazed. No M. giganteus is culled in the Googong Foreshores Reserve.
Tidbinbilla Nature Reserve
The Tidbinbilla Nature Reserve (148°55′E, 35°25′S; Fig. 1c) covers 5515 ha and is located approximately 30 km south-west of the city of Canberra. The valley is bounded on the east by steep heavily wooded and rocky slopes (up to 20°) and to the west by a radiata pine plantation and eucalypt woodland. The valley is approximately 4–5 km long and 1·5–2 km wide. Vegetation consists of patches of open grassland interspersed between patches of open woodland and more heavily wooded areas. Currently no M. giganteus is culled at Tidbinbilla Nature Reserve, but there have been major population reductions by culling in the past (Naeve & Tanton 1989).
capture and handling
Macropus giganteus was captured using a Pneudart® rifle fired from a vehicle (Pneudart®, Williamsport PA, USA). For each animal, a 1-mL dart containing the sedative drug tiletamine hydrochloride and zolazepam hydrochloride (Zoletil®; Virbac, Sydney, Australia) was fired into the hindquarters. Sedation with lateral recumbency was generally achieved within 5 min. While sedated, each animal was fitted with a VHF radio-collar (Sirtrack, Havelock North, New Zealand) emitting an individual signal at 40 pulses min−1 with a battery life of approximately 20 months. A mortality switch was incorporated into each transmitter whereby a rate of 80 pulses min−1 was emitted if there was no movement for 12 h. This ensured that lost collars were located quickly to maximize the amount of relevant data collected. Each animal was also fitted with coloured ear tags with an individual colour combination to facilitate direct identification by observation. White reflective tape was fixed to the ear tags and the collar to allow detection of animals at night.
Directional signals from transmitters fitted to individual M. giganteus were detected using a hand-held collapsible three-element Yagi antenna and a portable TR 2 receiver (Telonics, Mesa, Arizona, USA). Bearings were recorded from a hand-held compass. Animal locations were determined by triangulation of bearings from three to four different receiving locations per animal. Animals at each study site were radio-tracked in each season of the year (2001) at four different times of day (dawn, midday, dusk and after dark) for 6 days within a period of 2 weeks. Data were collected between March 2001 and February 2002. Error in location estimates was quantified by taking repeated bearings (n = 30) on 10 transmitters placed at known locations within each study site. Mean errors associated with triangulation were: Tidbinbilla, 59 m (range 25–125 m); Googong, 67 m (0–150 m); Cotter, 55 m (15–120 m).
Home range data were analysed using the Animal Movement Extension (Version 2·04) (Hooge, Solomon & Eichenlaub 1999) of the ARCView Spatial Analyst 2.1a Geographic Information System (ESRI®, Redlands, CA, USA). The fixed kernel model using bivariate normal density kernels (Worton 1989) was used to estimate the size, shape and location of kangaroo home ranges, using isopleths of 95% and 50% to represent home and core ranges, respectively. Data were analysed at two levels: seasonal and annual (i.e. pooled across seasons). A uniform smoothing factor was applied to all home range calculations. The smoothing factor was selected by visually assessing the home ranges of several animals from each study site using a range of smoothing factor values.
Range size, estimated using minimum convex polygons (MCP) (with 5% and 50% of outliers removed) and the ellipse model (Jennrich & Turner 1969), was reported to facilitate comparison with other studies.
Spatial and temporal autocorrelation of data
Animals radio-tracked within the same study site are often not spatially or temporally independent. Macropus giganteus is a gregarious species, with individuals generally forming open membership groups (Jarman & Coulson 1989). The high densities of M. giganteus at Googong and Tidbinbilla meant that home ranges of some animals overlapped. Although sightings of collared animals were common during radio-tracking, they were rarely sighted in the same social group. Hence we assumed that the use of different animals within the same study site was a valid approach. In addition, location fixes for any given individual were always 3 or more hours apart, and hence were sufficiently independent to avoid problems with autocorrelation between sightings.
Sample size and estimates of home range
Minimum sample sizes required for home range estimation were determined by examining the asymptotes of graphs plotting home range size against number of location estimates for several individuals (Harris et al. 1990). The minimum size samples for the seasonal and annual data sets were 15 and 30, respectively. Animals with fewer locational observations were omitted from subsequent analyses.
Pasture biomass was estimated in each season: (i) in paddocks on farms; and (ii) in natural grazing areas within reserves. Estimates were made on 120 randomly located plots at each of six sample sites (paddocks or grazing areas) within each study site (farm or reserve) using the comparative yield technique of Haydock & Shaw (1975).
Home range data were examined using generalized linear mixed models (GLMM) for unbalanced data (restricted maximum likelihood; Engel 1990) to assess the effects of site, season and pasture biomass on home range size. Seasonal variation in pasture biomass between study sites also was examined using GLMM. Seasonal variation in pasture biomass within the different study sites was examined using anova (Zar 1989).
Twenty-five animals were captured and radio-collared in this study. There were too few observations for five animals to permit inclusion in analyses of data at the annual level. Where sufficient data were collected for any of these animals, their annual and/or seasonal home ranges were estimated and mean values for each site are given in Table 2. All female M. giganteus in the study were reproductively active and had young in the pouch (PY), young at foot (YAF) or both. Only three animals moved away from the study sites. These animals were young adult males and may have been dispersing.
Table 2. Annual and seasonal home range and core range sizes of M. giganteus in ACT, Australia
Home (and core) range size (ha)
n = 9
n = 6
n = 10
n = 9
n = 8
n = 5
n = 5
n = 5
n = 5
n = 5
n = 6
n = 4
n = 7
n = 7
n = 7
Annual kernel home ranges varied from 36·6 to 129·3 ha and core ranges from 6 to 26·3 ha (Table 2). The shape of home and core ranges for most animals was approximately elliptical or circular (Fig. 1), and home ranges of almost all animals had one main centre of activity. Annual core ranges constituted 16–20% of home range area. Sample sizes were insufficient to examine differences in home and core range sizes between male and female M. giganteus. Neither the presence of young nor the stage of development had any significant effect (t-test = 0·69, d.f. = 14, P= 0·5) on range size of female M. giganteus.
Annual home range size was significantly larger (Wald statistic = 19·82, d.f. = 2, P= 0·001) for animals on Cotter Farm than in either Tidbinbilla or Googong reserve (Tables 2 and 3). Animals with home ranges located within the remnant woodland area on Cotter Farm appeared to be residents (Fig. 1a), and they rarely moved beyond the margins of the woodland vegetation. Two other animals at this property had home ranges that were largely located in the adjacent radiata pine plantation, but also included parts of the adjoining farm (Fig. 1a). At Googong, the home ranges of some animals extended a short distance onto adjacent farmlands (Fig. 1b).
Table 3. Mean values and standard errors for annual home range and core range areas (ha) of M. giganteus in ACT, Australia, estimated by the kernel model, minimum convex polygons (MCP) and ellipse model. HR, home range; CR, core range
Technique for home range estimation
Home ranges at all sites were significantly larger in autumn and winter than at other times of year (Wald statistic = 11·9, d.f. = 1, P < 0·001). There was little seasonal change in the locations of the core and home ranges for most animals. At Tidbinbilla, two animals showed seasonal shifts of up to 500 m in the location their core ranges. These animals largely used the steeper slopes towards the more heavily wooded areas of the reserve.
At all study sites where two or more individuals were captured in a similar area, there was substantial overlap of annual and seasonal home and core ranges (Fig. 1). For some individuals in high-density populations, there was complete overlap of core and home ranges, even though animals were not consistently members of the same social group. Most animals appeared to restrict their core ranges to major gully systems/catchments and rarely moved outside this area. Where animals were originally captured in different gully systems, there was no overlap of home ranges, even with less than 500 m separating the 95% isopleths of their respective home ranges. Some animals showed a minor shift in home range in different seasons but still remained within the same general area.
Estimates of home and core range areas using different analytical techniques (kernel, minimum convex polygon and the ellipse model) are given in Table 3.
Pasture biomass differed significantly between sites (Wald statistic = 405·36, d.f. = 2, P < 0·001) and seasons (Wald statistic = 27·90, d.f. = 3, P < 0·001) (Table 4). The amount of pasture available at all study sites was lower in winter than during the other seasons. At Tidbinbilla, pasture biomass was similar between grazing areas. At Googong, there was a trend for pooled pasture biomass to be lower on the adjacent farm than the reserve. One paddock, which was steep and rocky and contained a 5-ha patch of scrub, had lower pasture biomass in all seasons. This was the main farm paddock in which observations of kangaroos were consistently made.
Table 4. Mean seasonal pasture biomass (kg DM ha−1) at three study sites in ACT, Australia
At Cotter Farm substantial variation in pasture biomass existed between different paddocks as a result of differing grazing regimes. Rested paddocks consistently had higher measured biomass levels.
There was no significant effect of pasture biomass on annual home (Wald statistic = 0·44, d.f. = 1, P= 0·5) and core range sizes (Wald statistic = 1·49, d.f. = 1, P= 0·2) of M. giganteus. The larger seasonal home range sizes observed in autumn and winter may have been a consequence of changes in pasture biomass. However, even when pasture data for these seasons were pooled effects remained non-significant (Wald statistic = 4·11, d.f. = 3, P= 0·7).
The key processes influencing home range size of M. giganteus in our study were population density, presence of cover and reluctance to disperse across cleared landscapes. Resource availability was not important. Proximity to and social interactions with neighbouring animals may place limitations on space use (Adams 2003) and dispersal. Such effects will be largest in high-density populations. Jarman & Taylor (1983) suggested that dispersal of M. giganteus may be slow where animals have to disperse through areas already occupied by kangaroos. In high-density populations and in habitats surrounded by areas unsuitable for dispersal (i.e. cleared land), kangaroos maximize food intake by increasing the amount of time spent feeding, rather than by increasing range size (Jaremovic & Croft 1987). In our study, home range sizes for high-density M. giganteus populations in reserves were probably limited by population density, given that dispersal did not take place across surrounding open farmland lacking cover. Home range sizes were significantly larger on Cotter Farm, where population density was lower.
Macropus giganteus living within the remnant woodland area on Cotter Farm rarely moved onto more open treeless paddocks. Hence, where woodland remnants of sufficient size occur on farmland M. giganteus may occur as a resident. Although kangaroos did move onto open farmland along reserve and plantation boundaries, they remained close to cover.
Pine plantations may provide some areas suitable for grazing by M. giganteus, especially in areas of young regrowth vegetation, clearings and along forestry tracks. However, older stands yield limited pasture because of canopy closure and limited light penetration to the forest floor (Rishworth, McIIroy & Tanton 1995). Thus, M. giganteus utilizing the pine plantation adjacent to Cotter Farm may have required larger home ranges to access adequate forage.
Resident and roving mobs of M. giganteus may make extensive use of areas of remnant vegetation on farmland, but are reluctant to move far from cover onto open grazing areas. Generally, use of open areas is infrequent, for short periods only, and occurs on the periphery of home ranges. However, the extensive use of open woodland remnants, which may also be used by domestic livestock, will increase grazing pressure on these areas as well as on adjacent paddocks.
Farmers believe that M. giganteus competes with domestic livestock for grazing resources on their farms. The findings of this study provide little incentive for farmers to preserve remnant vegetation, as it may be regarded as providing habitat for unwanted or ‘pest’ native wildlife. Given the potential importance of remnant vegetation on private land for the conservation of other species of wildlife (Saunders et al. 1987; Lindenmayer et al. 2003), government incentives may be required to limit land clearing on farms (Kleijn & Sutherland 2003).
Cleared farmland adjoining reserves supporting high-density populations of M. giganteus will be subject to incursions by mobs of kangaroos. Where pasture biomass on a farm is similar to a nearby reserve, M. giganteus is unlikely to make extensive use of the farm, especially if there is little or no shelter available. However, M. giganteus may make more frequent use of paddocks rested from grazing by domestic livestock, thereby representing a disincentive for good pasture management by farmers. More heavily grazed areas may not be used as much by M. giganteus.
Farms adjacent to radiata pine plantations will be used by M. giganteus. However, these animals may be locally resident with high home range affinity and are unlikely to move far onto farmland. Farmers have two main choices in this situation: to either accept the limited but frequent use of land near to plantations or reserves, or consider culling to reduce M. giganteus populations.
Given that M. giganteus is sedentary and slow to disperse, culling should be an effective method of managing populations of kangaroos where they encroach on farmland (Hill 1979; Arnold et al. 1992). However, control of local populations of kangaroos can be unrewarding because of influxes of kangaroos from other areas (Priddel 1987). Control may therefore not be realistic and farmers may have to live with the presence of kangaroos on their land and/or recurrent culls of limited long-term efficacy.
The central issue is the perception by farmers that M. giganteus competes with domestic livestock for pasture resources. An overlap in the resources used by two species does not necessarily indicate competition (Putman 1996; Arsenault & Owen-Smith 2002). It is possible that the grazing habits of domestic livestock may result in feeding facilitation for kangaroos by making grass more accessible or stimulating grass regrowth. Similar processes have been observed in tall floodplain grasslands where African elephants Loxodonta africana Blumenbach have exposed grasses for buffalo Syncerus caffer Sparrman and topi Damaliscus lunatus Burchell (Vesey-Fitzgerald 1960). In addition, woodland clearance and pasture improvement in ACT over the past 50 years have probably led to habitat facilitation for kangaroos, resulting in population increases. In Africa, increased elephant numbers led to widespread destruction of woodlands and the subsequent development of new grazing lands, leading to increases in oryx Oryx gazella L. and zebra Equus burchelli Gray (Parker 1983).
There has been some research on competition between kangaroos and sheep but this has been in the arid rangelands where there is considerable dietary divergence between the two species at different times of year, which may limit competition (Caughley, Shepherd & Short 1987; Edwards, Dawson & Croft 1995). To date, there has been no research in the temperate areas of Australia to examine the impacts of M. giganteus on farmland and domestic livestock. If M. giganteus does have a negative impact on farm productivity, there is a case for developing more effective kangaroo population control methods or some form of financial remuneration for farmers in accordance with numbers of M. giganteus utilizing their properties (Kleijn & Sutherland 2003). This could be particularly important for conservation of remnant woodland in landscapes where kangaroos have limited opportunities for dispersal. In these cases animals spend considerable time feeding to maximize food intake (Jaremovic & Croft 1987) with consequent high impacts on pasture biomass and vegetation condition. Conversely, it is also possible that, in temperate areas, M. giganteus, although highly visible to farmers, may have little impact and therefore may require little or no management. Clarification of this issue is important in the debate on land management and protection of remnant vegetation.
This study was approved by the Australian National University Animal Experimentation Ethics Committee (R.DN.59.00), New South Wales National Parks and Wildlife Service (A2888), and Environment ACT (LT2001887). We thank Carmen Huckel and Rosie Smith for assistance with fieldwork and preparation of data. Staff at Environment ACT, Wildlife Research and Monitoring, provided logistical support and help with data analysis, especially Don Fletcher, Mark Dunford and David Shorthouse. The staff at Googong Foreshores Reserve and Tidbinbilla Nature Reserve provided valuable support and assistance. We thank the Lowe and Gorman families for access to their properties. Jeff Wood and Christine Donnelly provided assistance with statistical analyses. David Lindenmayer commented on the manuscript.