You are here

Counting elephants the right way: Study describes methods to overcome inaccuracy in estimating the population size of social mammals

Photo: Siddharth Kankaria / Research Matters


Can computer simulations help plan and improve population size estimation of wild animals? Could they, then, help in the conservation of endangered species such as the Asian elephant (Elephas maximus)? Yes, says a recent study by a team of researchers at the Jawaharlal Nehru Center for Advanced Scientific Research (JNCASR), Bangalore. Using various computer simulations and the Asian elephant as an example, the researchers have assessed how accurately population size of a social species that does not live in fixed groups can be estimated. “In general, computer simulations can help us conduct virtual experiments wherein the researcher can control various aspects of the system, and thus derive precise consequences under different assumptions and scenarios”, says Manan Gupta, the first author of the study.

In the study of ecology, much importance is given to population sizes of wild animals. Knowing the population size, age-sex ratio and understanding population dynamics can give many clues about the future of the species in a region. Dwindling population sizes can indicate a risk of extinction, whereas rapidly growing population sizes can stem from changing food resources, predator-prey dynamics, or climatic conditions in the ecosystem. Since population sizes are crucial, various techniques exist to estimate them, with ‘mark- recapture’ being one of the most popular techniques.

In the ‘mark-recapture’ technique, a part of the population is captured and marked with a unique identifier, like a number or a tag. These animals are then released and, after a predetermined amount of time, another set of animals is captured. The number of marked individuals recaptured is then used to estimate the actual population size of the species. Often, depending on the species, animals are not physically captured and marked, but are “captured and marked” via collecting and genotyping dung samples or by photographing and individually identifying animals using natural physical characteristics. While mark-recapture techniques work well for solitary species like tigers, previous literature has suggested that this method may introduce a bias when estimating population sizes of animals with complex social structures.

In the case of Asian elephants, female elephants form social groups. These herds are also observed to break up and associate with other herds and then re-join, exhibiting fission-fusion dynamics. Adult male Asian elephants are largely solitary. “If the two sexes in a species exhibit different kinds of social structure (which is the case in Asian elephants), the sex ratio calculated from the population size estimates of males and females may be an over- or an under-estimate, depending on how social structure affects bias in population estimation.” explains Dr. Vidya, the corresponding author of the study.

The researchers used a computer programming platform called MATLAB to create individual-based simulations, taking into account different social natures of animals. They tested for various traps densities and population densities to understand how these factors also affect the population estimation of social animals through the ‘mark-recapture’ method. Through computer simulations, they have identified which parameters need to be tested to minimize bias. “Our work is not specific to elephants, and the simulation framework can be tweaked to study any other type of social species”, Dr. Vidya adds.

Of the three mark-recapture models tested, the researchers found that an estimation model (POPAN) that accounted for movement of animals in and out of the study area across the study period, performed the best. However, even in this model, the accuracy of estimation was higher when the sampling or trapping was randomly performed rather than uniformly, and when adult density was higher.

“Computer simulations are particularly important for field studies because of the tediousness of field data collection, the impossibility of controlling various factors in the field, and the impossibility of knowing the correct information beforehand. Therefore, if one wanted to test which mark-recapture estimator performed best, it would be very difficult to find out based on only field studies. In such cases, simulations are useful”, explains Dr. Vidya on the importance of this study.

This paper highlights possible biases in estimating population sizes of social animals, by considering Asian elephants as a case-study. It also describes how the biases introduced by social organisation can be overcome if certain sampling strategies are used, thus making the findings applicable to similar studies on other mammals with complex social structures and large home ranges like dolphins and whales. “Using our simulation frameworks, field ecologists can design studies to obtain better population size estimates of elephants and other social species”, signs off Dr. Vidya.