Estimating Abundance of African Wildlife

Ground count techniques can be divided into total counts and sample counts, carried out on foot or by vehicle. Sample counts by vehicle can be divided into terrain counts, following the same principles as sample counts on foot, and road counts, whereby the sample count is restricted to the existing road-system. The same principles apply to total ground counts on foot as to those conducted by vehicle, with the only difference that dense vegetation may not be accessible by vehicle.

The line-transect technique is relatively easy to apply in the field, but the underlying theory and the mathematics involved are complicated, and consequently a computer and appropriate software are essential tools. The underlying theory has been described in a series of papers. The two most important papers were a monograph by Burnham et al. (1980), followed by the volume by Buckland et al. (1993), which documents the most recent developments in the science of line-transect methodology.

The technique and the field procedures relating to aerial total counts were first summarised by Norton-Griffiths (1975). Although the concept is simple, the design of an aerial census requires careful consideration to minimise error and bias. The main objective of an aerial census is to describe accurately the total number of a particular target species, and its spatial distribution over the study area. The census requires at least two observers, each counting on a different side of the aircraft, to scan the entire study area, as the aircraft flies along parallel flight lines that are between 500 m and 2 km apart. In the case of a hippo count the observers scan rivers and pools, while for puku, reedbuck and oribi they scan dambos.

Aerial sample counting is the most frequently used method to assess the abundance of large ungulate species on the African continent. Since its first use in the 1950s, the technique has changed in many aspects. The accuracy has improved beyond recognition, mainly as a result of the introduction of the GPS as an aid in navigation, and a better understanding of the potential sources of bias. The principles of an aerial sample count are the same as those that apply to other methods of sample counting discussed in previous chapters. As with other sample techniques, aerial sample counts estimate the number of animals by counting in a small part of the survey area. This is known as the sample. The density estimate for the sample is then extrapolated to the whole survey area. The calculations are similar to those for sample counts on the ground.

Certain indirect techniques, such as the mark/recapture method, or estimating abundance through radio tracking or by using DNA analysis, are special cases. Although these methods are used when only part of a population is visible to the observers, they do not come under the category of index techniques, because they lead to estimates of population size.

Counting techniques discussed in previous chapters produce absolute estimates of density, whereas index counts yield relative densities or density indices. Relative densities can be defined as the density of one population relative to another, or the density of a population at time t relative to its density at time t + 1. Thus, relative densities are useful only in temporal and spatial comparisons, and only if the conditions and methodology used are exactly the same for the areas and time periods compared. When conditions and methodology are consistent, index counts are more cost-efficient than direct counts, and can be used for trend analysis or as feedback for law-enforcement operations. A density index should be a measurable correlative of density, preferably with a linear trend on absolute density. A linear relationship can be expressed as Absolute Density = a + b Density Index, where a is the density when the index is zero and b is the increase of density per unit increase of index. In wildlife counts, however, the trend of density indices on absolute density is often logarithmic (Figure 8.1), but can easily be transformed to make density linear on the index.

Animal spoor or footprints per unit area can be used as an index of abundance (Van Dijke et al., 1986; Koster and Hart, 1988), while footprint measurements can be used to approximate the age-structure of an elephant population (Western et al., 1983), or to determine absolute density and distribution of elephant and black rhino (Kelly and Beer, 1994; Jachmann, 1984a), or large cats, such as lion, leopard and cheetah (Smallwood and Fitzhugh, 1993). Both elephants and rhinos are sufficiently heavy to render footprints visible for extended periods in a variety of soil types and habitats. When the population is small (< 40 individuals) and isolated (no migration), footprint measurements provide a means to identify individual animals. This technique may be combined with dropping measurements (circumference measurements of individual boli) to estimate abundance. Although the technique is relatively simple, leading to accurate estimates of abundance when used for small populations of solitary black rhino, its applicability for a gregarious species such as elephants is more complicated. Due to its limitations, the technique does not have a wide application in the field, but nevertheless may be useful under certain conditions. With only few small isolated pockets of black rhinos remaining in the wild, these conditions are found more frequently.