It is extremely difficult to understand the habitat needs and social dynamics of an elephant population, and how to reduce or solve its conflict with people, without knowing how many elephants there are. In open savanna, it isn't too difficult to simply count them, but the only way to get an accurate count of forest-dwelling elephants is through individual identification - in other words getting to know each elephant individually.

Getting to know elephants individually was pioneered by Iain Douglas-Hamilton in his landmark study of the Manyara elephants in the late 1960s. The Amboseli Elephant Research Project perfected the technique and today each elephant over the age of 10 is represented by a 6x4 inch index card on which a couple of mug shots have been pasted. All of the female elephants are organized into a box by family, known as the "Cow Box," and all of the males are organized in a separate box by size class, and known as the "Bull Box." This system works very well if the elephant population is relatively small, or if you already know the elephants pretty well.

In Amboseli, the elephant population now stands at 1,500 individuals and it is becoming very cumbersome (if not almost impossible) to discriminate between literally hundreds of photographs for those who are just getting to know the elephants, or for those who are a little rusty... There have been several attempts to come up with alternative techniques such as using face recognition software, though these have met with mixed results.

Back in 2003 Petter and Joyce worked with programmers in Nairobi to develop an alternative system using a searchable database. The system requires entering the key physical characteristics of each elephants into the database and uploading photographs - which takes time. Once this has been acomplished, however, just entering in a couple of these characteristics (e.g. female, V-notch right ear, tuskless) narrows down the search to just a few animals.

We have now developed such a system for Asian elephants - who have different key characteristics from African elephants - and we believe that this technique can provide projects with a very useful means of monitoring elephant populations.

You can to the right see a screenshot from a pilot ID database populated with 50 elephants from the Minneriya-Kaudulla population.

This article was published 18 June 2009.