Modeling longterm viability in African elephants

The longer-term viability of African elephant populations in semiarid regions was modeled by Peter Armbruster and Russell Lande. Their stochastic model of individual births and deaths used twelve 5-year age classes and discrete 5-year time steps. The main feature of their model is the range of environmental fluctuations incorporated in the form of 10-year (mild), 50-year (moderate), a) b)

Figure 7.8

Population viability analysis in Asian elephants. Probability of survival for 100 years is plotted as a function of initial population size for populations subject to demographic stochasticity, moderate environmental stochasticity, and small chances of catastrophes: (a) populations with varying potential intrinsic growth rates r and adult male : female ratio of 1 : 4; (b) populations with varying adult male-to-female ratios and potential intrinsic growth rate r = 0.02. (From Sukumar 1992, 1995b.)

15 20 25 30 35 40 45 50 55 60 65 70 15 20 25 30 35 40 45 50 55 60 65 70

Initial population size Initial population size

Figure 7.8

Population viability analysis in Asian elephants. Probability of survival for 100 years is plotted as a function of initial population size for populations subject to demographic stochasticity, moderate environmental stochasticity, and small chances of catastrophes: (a) populations with varying potential intrinsic growth rates r and adult male : female ratio of 1 : 4; (b) populations with varying adult male-to-female ratios and potential intrinsic growth rate r = 0.02. (From Sukumar 1992, 1995b.)

and 250-year (severe) droughts. The demographic data for their model came from the East African studies, particularly those in Tsavo National Park, Kenya, by Richard Laws and later by Timothy Corfield.

Births were modeled by relating age at first reproduction and intercalving interval to density in accordance with the functions derived for the Fowler-Smith model. In addition to the normal background rates of mortality, these were also varied under the three different drought regimes, which in turn were based on John Phillipson's (1975) analysis of rainfall patterns in Tsavo. While Phillipson described a mild 10-year drought superimposed on a more severe 50-year drought, Armbruster and Lande added a very severe 250-year drought based on the probability of three consecutive drought years. Carrying capacity constraints were placed by varying habitat or reserve area available to elephants from about 50 to 2,500 km2.

A baseline simulation was run for 625 years of an initial population of 11 males and 11 females in an area of 10,000 km (roughly the size of Tsavo East National Park). Using normal mortality rates without environmental variance, the population grew to an equilibrium density of 1.2 individuals/km following the classical logistic growth curve. At the lower population densities, the growth rate averaged about 3% per year. The population was again simulated by incorporating mortalities during droughts, but now starting with the equilibrium population size of over 12,000 individuals. When this is simulated over 5,000 years, the volatile dynamics seen from sharp decreases and increases suggest that drought-driven deaths could have a considerable impact on populations (fig. 7.9).

Subsequent simulations were also initiated with a density of 1.2 elephants/ km2 and the population at stable age distribution. Populations in one of the six habitat areas (50 km2, 125 km2, 250 km2, 500 km2, 1,250 km2, 2,500 km2) were simulated over 1,000 years. The simulations show that an area of 1,250 km2 with an initial population of 1,500 elephants is needed to ensure a 99% probability of persistence for 1,000 years (see chapter 9).

Armbruster and Lande also explored how various options of culling would affect the survival of populations. They found that the probabilities of survival did not decrease appreciably until culling reached about 50% of the carrying capacity. This was especially true for habitat areas above 1,250 km2.

The Armbruster-Lande model is useful in looking not only at future survival of elephant populations, but also at how they may have behaved in the past under natural cycles of drought in semiarid African regions. A crucial assumption in their model is that practically no young elephant would survive a very severe drought. Their demographic data also came entirely from the Tsavo population. Data from a broader spectrum of habitats in the African savannas would enable the model parameters to be adjusted for more generalized results to emerge on population viability. In particular, the extraordinarily detailed demographic information from Amboseli would provide interesting comparisons with the Tsavo situation. While the model results may be applicable to the extremely fluctuating semiarid zones of the elephant's range, it would

12,000

Figure 7.9

Simulation of total population size over 5,000 years in an African elephant population subject to 10-year, 50-year, and 250-year droughts of varying severity. (From Armbruster and Lande 1993. Reproduced with the permission of the Society for Conservation Biology, U.S.A.)

Time (years)

Figure 7.9

Simulation of total population size over 5,000 years in an African elephant population subject to 10-year, 50-year, and 250-year droughts of varying severity. (From Armbruster and Lande 1993. Reproduced with the permission of the Society for Conservation Biology, U.S.A.)

still be necessary to modify it for the more benign habitats, such as moist forests in Africa and in Asia.

0 0

Post a comment