Commercial harvesting of Kangaroos in Australia

by Tony Pople and Gordon Grigg
Department of Zoology, The University of Queensland
for Environment Australia, August 1999
Chapters 10,11,12 and 13 and Appendix 1 provided by staff at Environment Australia



Despite a history of harvesting kangaroos by Europeans that dates to the early days of settlement (see Chapter 5), limits to the distribution of macropods (see Chapter 3) and changes in their numbers can usually be attributed to changes in land use and environmental factors rather than harvesting per se (Cairns and Kingsford 1995; see Table 1). An exception is perhaps the now extinct Toolache Wallaby, although land clearance in the late 1800s and predation by foxes probably precipitated their decline (Caughley and Gunn 1996). Nevertheless, as identified in Chapter 2, harvesting may influence both the dynamics and the composition of kangaroo populations.

A population can be defined simply as a group of individuals of one species in a particular area. What is relevant to discussing the effects of harvesting on a population is the size of that area. Harvesting invariably occurs in a patchy manner across the landscape, potentially resulting in a continuum of harvesting effects on population dynamics and composition. Over time, movement of kangaroos between unharvested and harvested areas will tend to even out these differences, but this will also depend on the size of the harvested area and the size of and the proximity to unharvested areas. The following discussion addresses harvesting effects on a broad scale of > 400 km2. Populations in large, unharvested areas (e.g. nature conservation reserves, Chapter 12) also provide an important contrast with harvested populations, allowing the effects of harvesting to be assessed.

It is worth noting that the dynamics and composition of many unharvested populations do not necessarily represent those of kangaroo populations prior to European settlement. Marked differences would be expected because of a combination of lower densities and predation by dingoes (see Chapter 3) and aboriginal hunters. The composition and dynamics of present day populations experiencing dingo predation may be more representative of pre-European populations. Dynamics of unharvested populations

In order to determine the effects of harvesting upon a population, its unharvested dynamics must be considered first. While rainfall drives the dynamics of arid and semi-arid zone kangaroo populations (see above), these fluctuations may be modified by changes in population composition (age structure and sex ratio). It is well recognised that the observed maximum exponential rate of increase of a population may be much higher than the theoretical maximum (rm) determined from a stable age distribution and balanced sex ratio. Essentially, populations with fewer juveniles and males will have higher rates of increase. Bayliss (1985b) and Pople (1996) calculated maximum rates of increase for red kangaroo populations with altered sex ratios and age structures, estimating values as high as 0.67, considerably higher than the theoretical maximum of 0.25 for a population with a stable age distribution and balanced sex ratio. Numerical response models, based on empirical data, have suggested maximum rates of increase for red kangaroo populations of 0.33-0.58 in western New South Wales (J. Caughley et al. 1984; Bayliss 1985b, 1987) and 0.38-0.92 in South Australia (Cairns and Grigg 1993). As expected, estimates for western grey kangaroos are slightly lower (J. Caughley et al. 1984; Bayliss 1985b, 1987; Cairns and Grigg 1993). Caughley (1987a) used a value of 0.40 in his interactive model of a red kangaroo-pasture system in western New South Wales.

Juvenile survival rate has been identified as the poorest and most variable of all kangaroo population classes. Under deteriorating environmental conditions, populations are likely to show, in order (Hanks 1981):

  1. Increased juvenile mortality rate,
  2. Delayed sexual maturity,
  3. Reduced fecundity and
  4. Increased adult mortality rate.

Such a sequence has been found throughout Australia in populations of red kangaroos (Frith and Sharman 1964; Newsome 1965b; Shepherd 1987; Pople 1996), eastern grey kangaroos (Kirkpatrick and McEvoy 1966; Stuart-Dick 1987), western grey kangaroos (Norbury et al. 1988; Arnold et al. 1991) and common wallaroos (Sadlier 1965; Russell and Richardson 1971; Clancy and Croft 1992).

Male-biased mortality is widespread among ungulates and other herbivorous mammals, particularly during food shortages and in high density populations (Clutton-Brock et al. 1982). A number of causes have been suggested, including greater nutritional requirements and increased predation and disease risk. For kangaroos, mortality during drought appears to be strongly male biased. Pople (1996) reported a fall in the percentage of males in an unharvested population following two years of drought from 46% to only 25% ( Figure 23). Unharvested populations that are female-biased have been reported elsewhere for red kangaroos (Newsome 1977; Johnson and Bayliss 1981; Robertson 1986; Edwards et al. 1994), eastern grey kangaroos (Jarman and Southwell 1986; Quin 1989; Shelly 1997), western grey kangaroos (Norbury et al. 1988; Arnold et al. 1991) and common wallaroos (Clancy and Croft 1992). The higher energetic costs of males, incurred through their having larger body size, higher growth rates and greater ranging, may explain these mortality patterns.

Fluctuating populations of red kangaroos have been considered to exhibit centripetality: a weak influence of population density on food supply pulls the population towards an equilibrium state, but a highly erratic rainfall means that state is never reached. The environmentally induced changes to population age structure and sex ratio described here would act to strengthen centripetality.

Much of this discussion has centred on arid and semi-arid zone populations of kangaroos, where commercial harvesting concentrates. Populations in more mesic areas and at higher latitudes may well fluctuate to lesser degree. Shepherd and Caughley (1987) suggested a threshold of 30% (occurring east of Goondiwindi in eastern Australia) for the coefficient of variation (CV) of annual rainfall, above which the environmental variance reveals more about an animal's ecology than the environmental mean. Shepherd and Caughley (1987) centred their discussion on the concept of carrying capacity and its value in fluctuating environments and this was elaborated on by McLeod (1997). Below their threshold CV, herbivore densities will be relatively steady, with a strong feedback between food supply and herbivore numbers. Above this value, fluctuations in the environment and herbivore numbers are expected. Such an assumption is important in monitoring kangaroo populations. It is feasible to monitor arid and semi-arid zone populations relatively frequently, but monitoring of more mesic populations relies on ground surveys which are labour intensive and time consuming.

Dynamics of harvested populations

The response of a red kangaroo population to harvesting will not depend solely upon the extent of density-dependent compensation. Harvests that are selective by sex or age may further alter the population's potential rate of increase, rp. The extent of compensation may vary also with the timing of the harvest in relation to periods of mortality and reproduction. This is particularly relevant to harvesting in a variable environment. Rate of increase of red kangaroo populations is not a simple function of density; responding to a variable environment over which the kangaroos have only a weak influence (Caughley and Gunn 1993).

The compensatory response of a population to harvesting can be viewed with respect to ecological carrying capacity (K). For a hypothetical population which displays logistic growth in a stable environment, harvesting will keep it below K. The MSY will be taken at economic carrying capacity (Caughley 1976) which, in this case, is K/2. In a fluctuating environment, K varies, as will the position of the harvested population relative to K. When the population is below K, harvest mortality is less likely to be compensated for. When the population is near K, or above it, harvest mortality is more likely to be compensated for (Bartmann et al. 1992).

For a fluctuating population of red kangaroos, whose density is often caught either well above K or far below it, harvesting at a constant rate should strengthen centripetality. Assuming regulation is solely extrinsic, the population will experience greater compensatory mortality and reproduction during droughts than when food is abundant. Conversely, if harvest rate increases during drought, then centripetality may be weakened (Caughley 1987a; see below). Comparing properties in western Queensland in which red kangaroos were and were not harvested, Pople (1996) found juvenile body condition and survival rates were better on harvested properties, although immigration confounded this result. Breeding success was also greater on harvested properties, but only when pasture conditions were poorest, suggesting differences may be manifested only during drought.

Concluding that compensation would be negligible during periods of abundant food, ignores the effect of harvesting on population composition ( Figure 23). Selective harvesting may raise rp by altering the sex ratio, but this effect may be swamped during drought. As a result, the disparity in rp between an unharvested and a selectively harvested population will be greatest during periods of abundant food. In this case, the compensation is not provided by improved survival rates or individual reproductive success, but by an increased reproductive rate in the population as a whole. Whether the extent of a population's fluctuations should be altered by wildlife managers is contentious. While a more stable population may provide the kangaroo industry with a less variable supply of animals, there may be sound ecological reasons for allowing fluctuations in numbers(see Chapter 2).

graph - Currawinya 1991graph - Currawinya 1993graph - Terrick Terrick

Figure 23 Age structures of red kangaroos Age structures of red kangaroo populations on two properties in western Queensland.
23a an unharvested population on Currawinya, at the start of a drought in 1991
23b Currawinya, at the tail end of the drought in 1993
23c a heavily harvested population on Terrick Terrick in 1991
Shaded parts of the 0-1 age class represent pouch young; the remainder represent young-at-foot. Bars represent the proportion of each age class in the population greater than or equal to 1 year old. Determined from samples of greater than 250 kangaroos that were shot at random (after Pople 1996).

In practice, the gains in sustained yield from a selective harvest that were suggested above and by Pople and Cairns (1995) for red kangaroos, may not be that large. In western Queensland, Pople (1996) reported age distributions of heavily harvested populations of red kangaroos of 42% 1 year old with an adult sex ratio of 29% male. Such a population has an rp of 0.34, given realistic rates of age-specific rates of survival and fecundity (Bayliss 1985b; Pople 1996). On the other hand, the age distribution of an unharvested red kangaroo population (14% 1 year old and an adult sex ratio of 44% male), actually had a slightly higher rp of 0.38 (see also Figure 23). A selective harvest can generate a higher rate of increase over an unselective harvest, but only if those gains are not lost through an overharvest of mature females. Obviously, these comparisons do not include the gains in rp achieved from a reduction in density.

While the optimal harvest strategy (at least for the sustained yield from a wild population) is a male-biased harvest that is not selective by age (Caughley 1977), this is not the aim adopted by kangaroo shooters. The potential for this suboptimal selective harvest exists partly because of the open access system of harvesting for kangaroos. Exacerbating this lack of incentive to conserve the resource, many graziers in western Queensland allow several shooters to harvest kangaroos on their properties, with the particular shooters changing from year-to-year.

Harvesting may also change the social organisation through the removal of males >10 years old, who, in unharvested populations, appear to have the greatest breeding success (Croft 1980, 1981, 1989). Fewer old females, whose breeding success may (Stuart-Dick and Higginbottom 1989; Ashworth 1995; Moss 1995) or may not (Pople 1996) be greater than younger females, are found on heavily harvested properties. An alteration to the adult age structure may therefore alter reproductive output by increasing the proportion of inexperienced breeders in the population and, by reducing competition among males, potentially compromising female selection of suitable mates. Notably, Pople (1996) found no reduction in fecundity rates in heavily harvested red kangaroo populations compared with unharvested populations.

Much of the preceding discussion has been of populations in relatively large areas where movement in and out of the population (immigration and emigration) has negligible effect on its dynamics. Changes in numbers are therefore considered the net result of births and deaths. Given the mobility of most kangaroo species, harvesting at the scale of a 150 km2 grazing property will be influenced by such movement. Both Edwards et al. (1994) and Pople (1996) reported red kangaroos rapidly recolonising areas that had been heavily depleted through shooting. In many areas, particularly the semi-arid woodlands and forested tablelands, a large proportion of the population will remain unharvested, because shooters simply cannot access them. McCullough's (1996) patchwork harvesting system may well inadvertently apply to kangaroo harvesting in many areas of Australia.