Using High Definition Satellite Images to Monitor Saigas

Censuses are an important component of monitoring the status of wildlife populations. Different techniques are used for different species, on account of their specific biological characteristics. The European population of the saiga antelope (Saiga tatarica) inhabiting northwest Caspian Lowlands (eastern part of the Republic of Kalmykia and south-western part of the Astrakhan Region) is in a critical condition. According to expert estimates, its size is no more than 5000 individuals. Past saiga censuses have typically relied on line transect surveys, either aerial (using mainly low flying helicopters) or ground based (using several vehicles). These methods have a negative impact either on the animals themselves or on the plant and soil cover of the areas they inhabit. Besides, line transect surveys of species that tend to move while being surveyed (as is the case with the saiga) are prone to substantial random error.

In recent years, attempts have been made to conduct saiga censuses by means of aerial photography using drones. However, this method, albeit far less invasive than conventional line transect surveys, is still not adequately developed, and it also has certain inherent limitations. Satellite remote sensing may offer a worthwhile alternative to the above methods. State of the art satellite imaging systems are capable of providing highly detailed photographs that allow identification of relatively small objects, including certain species of mammals. Pilot censuses using high resolution satellite images have already been undertaken for a number of Arctic marine mammals, including the polar bear, with promising results.

The aim of this study was to determine whether individual saiga can be reliably identified in high resolution satellite images, with a view to developing a new noninvasive method for an accurate assessment of saiga population size.

To read the report please visit the ResearchGate website. You’ll need a login to read the full report.

%d bloggers like this: