Antoine Marchal

Conserving wildlife amongst the cotton fields. A third of a century of experience at the Nazinga Game Ranch, Burkina Faso.

Established in the early 1970 as a participatory wildlife production area, the Nazinga Game Ranch turned into an island of conservation surrounded by cultivation. We asked ourselves how long-term ungulate trends are affected in a context of continuous human pressure. To find out, we compiled and analysed the data of yearly line-transect counts of mammals carried out since 1985. Results showed that large species such as the elephant (Loxodonta africana) and large antelopes increased or showed stable populations. In contrast, medium and small ungulates showed continuously decreasing trends. During the same period, rainfall, water availability from artificial water points and the crop encroaching outside Nazinga Game Ranch increased. After an initial significant reduction, illegal human signs increased. However, we showed that human signs were positively correlated with the abundance of large ungulates but negatively correlated with the abundance of medium and small ones. In conclusion, this study showed that some isolated mammal populations could be restored and maintained in the long term, in spite of being surrounded by highly cultivated areas.

Identification of the Anteroposterior and Mediolateral Position of Lion Paws and Tracks Using 3D Geometric Morphometrics

Estimating the distribution and status of animal populations is crucial in various fields of biology. Monitoring species via their tracks is controversial due to unreliable recording techniques, manipulator bias and substrate variation. Furthermore, subjective identification of the foot that produces each track can lead to significant errors, for example, when assigning tracks made by different feet from the same individual to different individuals. The aim of this research was to develop an accurate, consistent and objective algorithm to identify the anteroposterior (hind/front) and mediolateral (right/left) position from digital threedimensional (3D) models of African lion (Panthera leo) paws and tracks using geometric morphometrics. We manually positioned 12 fixed landmarks on 132 paws and 182 tracks recorded in 3D using digital close-range photogrammetry. We used geometric morphometrics to evaluate and visualize the shape variation between paws and between tracks along the anteroposterior and mediolateral axes, and between paws and tracks. The identification algorithm using linear discriminant analysis with jack-knifed predictions reached a maximum accuracy of 95.45% and 91.21% for paws and tracks, respectively. We recommend the use of this objective position identification algorithm in future studies where tracks are compared between individual African lions.