Edge-effects of fences on elephant movement patterns: Implications - - PowerPoint PPT Presentation

Edge-effects of fences on elephant movement patterns: Implications for small reserves

Abi Tamim Vanak, Maria Thaker & Rob Slotow University of KwaZulu-Natal

Humans, fences and wildlife

Keep wildlife out

Keep wildlife in

• Positive

– Reduce human-wildlife conflict – Manage wildlife/livestock populations – Highway fencing - reduce collisions – Prevent spread of diseases

Effects of fences

• Negative

– Constrain movement – Prevent dispersal – Cut migration routes – Physical contact – entanglement, electrocution

Effects of fences

Effects of fences

• Population level responses
• Behavioural responses?

– Linear feature -> Habitat edge – Edge-effects on behaviour – Extent of edge-effects

Elephant as ecosystem engineers

• Megaherbivore

– Ecosystem engineers

• Elephant movement

behaviour affects many ecological processes

Elephants and vegetation

• Concern over deleterious

effects of elephants on vegetation

• Large trees

– Foliage utilization – Breaking of large branches – Debarking – Pushing over

• In free-ranging populations

– Cut migratory routes – Restrict seasonal movement

• Revisit areas near the fence

– Over-utilisation close to fences

Fences and elephants

• Most South African

reserves are fenced

• Reintroduced

populations

– soft release

• Aversive conditioning

to electric fence

Fences and elephants

Elephants and fences

• Close proximity to the fence will affect

elephant movement

– 1) Elephant will “bunch-up” near the fence

• Over-utilisation of resources near the fence

– 2) If elephant show aversive behaviour then avoidance of fence

• Or movement rates near fence will be faster
• Greater net-daily displacement

Elephants in Pilanesberg NP

• Six GPS collared females

– Represents six herds

• Data collected 4-6 hourly

– from 2004-2007

• Based on the autocorrelation

function, sampling interval gap

• f 24 h yielded spatially

independent movement paths

(ACF = 0.019, Q = 0.277, P = 0.6)

Elephant movement paths

• Movement rates
• Daily net displacement

– In relation to fence

• 3806 daily paths

– Subsampled to >24hour – 1113 dry season paths – 790 wet season paths

Model fitting

1 2 3 4 5 6 7 8 9 1 2 3 4

• Minimum distance from fence (km)

Daily net displacement (km)

Dry season

5

T

10

Wet season

1 2 3 4 5 6 7 8 9

Minimum distance from fence (km)

1 2 3 4

Daily net displacement (km)

• 10

T

Dry season AIC ΔAIC w Piecewise regression 19280.32 0.971 Sigmoidal function 19287.32 7 0.029 Linear regression 19364.9 84.58 0.000 Wet season Piecewise regression 13849.51 0.986 Sigmoidal function 13858.01 8.48 0.014 Linear regression 13882.76 33.25 0.000

Model fitting

Edge-effects of fences

Piece-wise regression analysis

Vanak, Thaker and Slotow 2010 Biol. Cons. 143: 2631-2637

(β = - 0.01) (β = 0.04) (β = - 0.47) (β = - 0.75)

Fences and Elephant movement

• Daily net-displacement was higher near fence

in both seasons

• Effect dissipates at 2.6 km in dry season &

3.8 km in wet season

• Fence effects over-ride seasonal effects

– Lower net displacement in dry compared wet season

Edge-effects of fences

Piece-wise regression analysis

Vanak, Thaker and Slotow 2010 Biol. Cons. 143: 2631-2637

(β = - 0.01) (β = 0.04) (β = - 0.47) (β = - 0.75)

Dynamic vs. Static analyses

On either side of the spatial breakpoint in movement behaviour:

• No difference in proportion of locations
• No difference in habitat composition

(wet season: χ2 = 9.15, p = 0.10; dry season: χ2 = 7.64, p = 0.18)

• Few differences in habitat selection

– Grassland: selected in the centre, avoided near the fence – Acacia mellifera: avoided in the centre, selected near the fence

By using movement metrics

– Discriminate non-linear behavioural responses – Spatially explicit fence effects

Cascading edge-effect of fences

• “Streaking” behaviour
• Lower tortuosity near fences

– Lowered foraging – Transfer of foraging pressure to center of reserve

• Elephant transmit edge-effect
• f fences to rest of reserve

Cascading edge-effect of fences

• Reserve shape
• High edge-to-area ratio exacerbates edge-effect of fences
• Carrying capacity of reserves is smaller

Solution?

• Drop fences to optimize reserve size and shape

500 m = 10%

• f under-

utilized space

Pongola Game Reserve

Acknowledgements

• Amarula Elephant Research Programme
• University of KwaZulu-Natal Research Office
• PPC Cement South Africa
• National Research Foundation
• Northwest Parks and Tourism Board