Contribution au monitoring de populations d'hippopotame commun (Hippopotamus amphibius L.) par l'utilisation de la technologie drone (Parc National de la Garamba, République Démocratique du Congo)

In an era of indisputable decline in populations of large wildlife species, setting up regular and efficient monitoring methods is essential. Because of the difficulties associated with conventional pedestrian and aerial survey procedures, the development of UAV technology turns out to be a potentially interesting alternative. This work focuses on monitoring common hippopotamus (Hippopotamus amphibius L.) with the use of this innovative tool. The search for flight parameters optimization to maximize detection and visibility of animals showed a very small decrease in detection rate and certainty of counts with an increasing flying height and a high variability linked to operators realizing observations. It is recommended to fly at a height of 140 meters above ground level, between 6:30 and 8:45 am at the end of the dry season, with cloud cover and wind as low as possible. Three counts correction methods are proposed to estimate the total population from the only emerged observed animals. The average overall correction factor is 1.252. An attempt to describe the demography of a group by individuals length measurements was also implemented. It allowed to obtain a mean distribution of hippos in three age classes, even if results are not extremely accurate. Finally, an algorithm for the semi-automatic count of hippos on thermal infrared imagery was developed, and brought to a mean error of +3.9%. Taking into account the recommendations proposed in this study for the development of inventories protocols would help dealing with the practical constraints of conventional monitoring methods.