Several methods can be used to survey elephant populations. The appropriate technique for a particular area will depend on the question being asked, the habitat, and resources available to conduct the survey. Each survey method has advantages and disadvantages that researchers must take into account before choosing a method.
Animals can be surveyed using direct methods (counting the animals themselves) or indirect methods (counting signs or signals, such as dung, tracks, or sounds). After a method is chosen, the researchers map out the study area, which is the total area they are surveying. If a study area includes several different habitats, such as grassland and forest, or mountains and plains, the study area is divided into strata, or sections, for each habitat. The methods are then adapted to fit the conditions within each stratum.
Within each study area or stratum, researchers can choose to perform a total count or a sample count. A total count is when the entire study area is surveyed, while a sample count is when parts of the study area are surveyed and the results from those parts are applied to the entire study area to estimate the total population size.
When a sample count is performed, sampling areas can be in the shape of strips, blocks (or polygons), and points. Sampling areas are selected throughout the study area or stratum in either a random or regular fashion. Random sampling is usually preferred because it avoids bias, but in some cases, placing the sampling areas at regular (or equal) intervals is preferred.
When sampling in strips, researchers first draw transect lines and the strips are the areas to the right and left of the transect line. Transect lines are typically drawn at regular intervals. Blocks are polygons selected around random points and the entire area within the block is surveyed. A point survey entails stationing an observer, an audio recorder, or a camera at random points in the stratum for a period of time. If researchers are measuring sounds the distance at which sound can be detected around the point determines the area that is being sampled.