Energetic benefits and costs (Griffiths 1975, Creel & Creel 2002) Mechanisms of selection Search images Prey vulnerability Habitat characteristics relating to hunting or escape (Kruuk 1972, Husseman et al. 2003)
Potential to disturb evolutionarily stable predator-prey relationships Water provisioning in KNP: lions and rare antelope (Harrington et al. 1999) Extensive habitat alteration due to agricultural practises (Sweitzer et al. 1997) Reduction of suitable habitat into smaller patches (mesopredator release - Crooks & Soulé 1999)
Predator-proof fencing limits movements of predators and prey Spate of reintroductions (Marnewick et al. 2007, Funston 2009, Davies-Mostert et al. 2009) High perimeter-to-area ratios increase likelihood of contact with fences Suggested that carnivores modify their hunting behaviour (van Dyk & Slotow 2003; Rhodes & Rhodes 2004) different species, size classes, sex and condition, overall hunting success, etc Implications for the sustainability of reserves to support carnivores
Large carnivore ~25kg Endangered (Globally, Nationally) Social, pack-living Cooperative rearing of young Hunting Coursing predator Kills likely when prey become exhausted, mean chase 0.57 km, 0.05 m-4.60 km (Creel & Creel 1995) Likely to exert higher selection for animals in poor condition than ambush predators, southern Zimbabwe (survival of the fittest - Pole et al. 2004) High energy requirements (Gorman et al. 1998) 15MJ/day for 3.5 hour day Working border collie – 8.2MJ in 6 hours Pole et al. 2004
1. Hluhluwe-iMfolozi Park 2. Madikwe Game Reserve 3. Pilanesberg National Park 4. Venetia Limpopo Nature Reserve 5. Marakele National Park 6. Tswalu Kalahari Reserve 7. Mkhuze Game Reserve 8. Balule Game Reserve 9. Thanda Private Reserve Davies-Mostert, 2010 Note: - Most sites not much bigger than AWD home range - Isolated
van Dyk & Slotow, 2003 Rhodes & Rhodes, 2004 Direct observations supplemented with patrol data Unsystematic
Do patterns of prey selection support earlier findings from Save Valley Conservancy? What influence do perimeter fences have on prey selection? Do fence-impeded kills comprise a greater proportion of larger prey species and prime age animals than those that are not fence- impeded? Is the physical condition of prey better among fence-impeded kills? Do wild dogs spend more time actively hunting in close proximity to fences so at to maximise the advantage they confer?
Direct observations of kills Date and time Location Species Age (tooth wear) Sex (adults and sub adults only) Femur marrow condition Visual score, % marrow fat Compare with venison hunts Estimated edible biomass Allocated size category Small <25 kg Medium 25-90 kg Large >90 kg Distance from fence line, fence- impeded or not
Straight-line distance of kill from packs previous resting location Includes hunting AND traversing Catch per unit effort measured in kg/km
Perimeter fence buffered Number of locations compared to proportional area Activity vs. resting Exclude denning season locations
392 kills in 3 years Sub-sample of 316 for this fence analysis 0.41 kills/activity period (n=723) Kill rates were correlated to pack size Per capita consumption was 2.09 kg/dog/day Impala accounted for 78.1% of dietary items Only juveniles of larger species were taken (eland, gemsbok, wildebeest) and only rarely 41% of kills were fence-impeded
Large kills: >90kg, solid circles Medium kills: 25-90 kg, empty circles Small kills: <25 kg, crosses Den sites: squares n=316
Juvenile impala killed less frequently on the fence Adult female kudu killed more frequently on the fence Adult and sub adult male kudu killed more often on the fence but sample sizes are too small SexAge Not fence- impeded Fence- impeded RatioTotal Proportion on fence No.Prop.No.Prop. ImpalaFemaleAdult320.17180.140.8500.36 Sub adult80.0440.030.7120.33 > Juvenile20.010.020.00 Unknown30.0210.010.540.25 MaleAdult220.12130.100.9350.37 Sub adult50.0330.020.980.38 Unknown10.010.010.00 UnknownAdult20.02-21.00 Unknown50.034 1.290.44 AllJuvenile300.1660.050.3360.17 Kudu, GreaterFemaleAdult110.06290.233.9400.73 Sub adult70.0470.051.5140.50 Unknown10.01-11.00 MaleAdult10.0150.047.360.83 Sub adult10.0140.035.950.80 UnknownAdult10.010.010.00 Unknown10.01-11.00 AllJuvenile90.0580.061.3170.47 Bushbuck50.0310.010.360.17 Cow10.01-11.00 Duiker, Common80.0440.030.7120.33 Eland10.010.010.00 Gemsbok40.0210.010.450.20 Hartebeest, Red20.010.020.00 Jackal10.010.010.00 Steenbok80.045 0.9130.38 Warthog110.0650.040.7160.31 Waterbuck80.0440.030.7120.33 Wildebeest, Blue20.011 0.730.33 Total1881.001281.001.03160.41
Greater proportion of large prey caught on the fence ( 2 =33.05, p<<0.001) Median mass of fence-impeded kills > than unimpeded kills, 32.9 vs. 25.0 kg (W=25667.0, p<<0.001) Fence-impeded kills comprised 54.1% of total edible biomass captured Edible biomass was the most significant variable in predicting whether a kill was fence- impeded or not (positive)
Proportion of males and females killed on fence did not differ for impala or kudu Adult impala males killed on the fence had higher % marrow fat than those killed away from the fence (Mann-Whitney, W=111.0, p=0.012, n=28) Kudu males killed on fence had highest % marrow fat of all, but small sample sizes
Un- impeded CriterionFence- impeded 2.9Mean distance from resting to kill (km) 2.5 12.2Median catch per unit effort (kg per km) 27.3 Important consequences: Biomass consumed per activity period influences inter-kill interval Hunting is risky!
Zone (distance from fence in km) Area (km 2 ) Proportion of total reserve area Number of active locations Proportion of active locations Preference index (PI) Normalised preference index (Log 10[ PI+1]) 0-18,1080.26970.662.550.55 1-26,5180.21230.160.750.24 2-3.56,5440.21140.090.460.16 3.5-54,9050.1690.060.390.14 >55,5070.1750.030.190.08 Total31,5821.001481.00 The pack was also more likely to be found closer to the perimeter fence when active than when resting (χ 2 = 7.94, p=0.047, 3 d.f.)
12 confirmed breaches in 723 follow periods 4 kudu 3 impala 1 juvenile zebra 1 steenbok 3 unidentified More suspected!
Fence-impeded kills SpeciesDifferent species composition (e.g. kudu females) SizeGreater biomass per kill AgeMore adults (impala, kudu) SexNo difference for main prey species Body conditionBetter condition (e.g. adult male impala) Shorter chase distances Greater catch per unit effort Fewer hunting forays Benefits for AWDs Fitness benefits? - Survival? - Reproduction?
Davies-Mostert, 2010 * Includes data from all reserves, even those where fence-hunting is uncommon
Wild dog predation is at least partly compensatory Supports SVC survival of the fittest study Wild dogs actively make use of fences as a hunting tool Supports earlier studies in SA Fences cause shifts in patterns of prey selection with potential shifts in predator-prey dynamics Potentially reduces the compensatory nature of predation Consequences for the ability of small reserves to support predator populations in the long-term
Magriet van der Walt Melanie Boshoff Herta Martin Lynda Hedges Richard Selamolela and Venetia scouts Kath Potgieter Pat Fletcher Graham Main Warwick Davies-Mostert Lesley Sutton Many others!