CollarData

Eight adult male baboons within our focal troop were cage-trapped before being sedated by a certified veterinary and fitted with a custom-built tracking collar. Collars weighed less than 2.5% of the body mass of the baboons and were approved for use by Swansea University Ethics Committee (Swansea University IP-1314-5). Collars were equipped with sensors recording GPS every 5 minutes and acceleration in 3 axes at 40 Hz. Data were recorded from 7:30 am to 6:00 pm and only data recorded between sunrise and sunset according to the South African Astronomical Observatory were selected resulting in 6,325 fixes. Any pair of successive fixes more than 1 km apart were assumed to be errors, and removed, resulting in 6,274 fixes. Ad-hoc checks of the data where baboons cross known landmarks indicate positional accuracy of <10m for processed GPS data. Latitudes and longitudes are given in WGS84. Acceleration was recorded continuously at 40 Hz for 5 individuals and at 20 Hz for one individual (LM). Acceleration raw data were first decomposed into static and dynamic acceleration with a running mean of two seconds then we calculated the Vector of the Dynamic Body Acceleration (VeDBA) calculated from dynamic acceleration data. We superimposed the GPS data with (1) the map of the risk of deterrence, obtained via interview with baboon rangers (scores ranging from 0 for low risks of being deterred to 22 for high risks); (2) the habitat map were habitats were categorised as one of five broad habitat categories (fynbos, trees, meadows, vineyards and urban areas) based on researcher knowledge of the site and images from Google Earth (accessed 06/06/2014).