Remove or control predators using fencing and/or aerial nets

Supporting evidence from individual studies

1. A controlled, before-and-after study in 1990–1994 in mixed heath and dune habitat in Western Australia, Australia (Risbey et al. 2000) found that a fenced area with cat Felis catus and fox Vulpes vulpes control had similar numbers of geckos and skinks but fewer dragon lizards than unfenced areas with an without fox control. The effects of fencing and predator control cannot be separated. The number of geckos and skinks were similar in the fenced area (with fox and cat control) and unfenced (fox control or no control) area (geckos: 1–2 individuals/trap grid; skinks: 2 individuals/trap grid). Dragon lizard numbers were lower in the fenced area with fox and cat control (fox and cat control: 2 individuals/trap grid) compared to unfenced areas with fox control (fox control only: 5 individuals/trap grid) and unfenced, no control areas (no predator control: 7 individuals/trap grid). In areas with predator control, there was no clear change in reptile numbers from before control began (0–24 individuals/group/year) compared to the three years after control began (0–12 individuals/group/year). In 1991, a mainland peninsula was divided into three areas: one area (12 km²) where cats and foxes were controlled (using electrified fencing, poison baiting, or secondary poisoning by poisoning European rabbits Oryctolagus cuniculus, trapping or shooting); one area (120–200 km²) where foxes were controlled by baiting but cats were not targeted; and one area where no control occurred. Reptiles were monitored with six pitfall-trap and drift fence grids in each area (18 in total). Each grid had eight pitfall traps, 30–50 m apart. Sampling was conducted over three consecutive days in March–April and June–July in 1990–1994 in predator control areas and 1992–1994 in the area without predator control. Reptiles captured included dragon lizards, skinks, geckos, snakes, and a species of monitor lizard and blind snake but only species that could be toe clipped (dragon lizards, skinks and geckos) were included in analysis.

   Study and other actions tested

   Referenced paper

   Risbey D.A., Calver M.C., Short J., Bradley J.S. & Wright I.W. (2000) The impact of cats and foxes on the small vertebrate fauna of Heirisson Prong, Western Australia. II. A field experiment. Wildlife Research, 27, 223-235.

2. A replicated, site comparison study in 1999 in a site of semi-arid shrubland, grasses and sparse woody plants in New South Wales, Australia (Olsson et al. 2005) found that areas within exclosure fences with ongoing fox Canis vulpes control had a higher abundance of small lizard species compared to areas outside the fences. Average captures of all lizard species were higher within the exclosure fences (28 lizards of 11 species; 0.9/trap) compared to outside the fences (13 lizards of 8 species; 0.5/trap) (differences in richness were not tested for statistical significance). The same was true for skinks (inside: 1.2/trap; outside: 0.3/trap). Numbers of geckos was similar inside fences (0.7/trap) and outside (0.7/trap) An area of natural habitat (400 x 200 m) was fenced off in the 1980s as part of a species reintroduction. Poison baiting for foxes was ongoing in the area. Eight pens were established (100 x 100 m in a 4 x 2 design) using chicken wire and an electrified wire. In October 1999, reptile trapping occurred inside and outside exclosures with four lines of pitfall traps (8–14 m long, buckets every 2 m).

   Study and other actions tested

   Referenced paper

   Olsson M., Wapstra E., Swan G., Snaith E., Clarke R. & Madsen T. (2005) Effects of long-term fox baiting on species composition and abundance in an Australian lizard community. Austral Ecology, 30, 899-905.

3. A controlled study in 1998–2005 in a site of dunes and shrubland in South Australia, Australia (Moseby et al. 2009) found that in a fenced area where invasive cats Felis catus, foxes Vulpes vulpes and European rabbits Oryctolagus cuninculus were removed, reptile abundance and species richness were similar for three years, then in the following five years, abundance was lower compared to outside the fenced area and richness was higher in a fenced area where native mammals had not been reintroduced. During the first three years (1998–2000), reptile abundance and species richness were similar inside and outside the fenced area (native mammals reintroduced to fenced area in 1999). In the following five years (2001–2005), the abundance of reptiles was lower inside an expanded fenced area (one area with and one without native mammals) than outside, and richness was higher in one fenced area (no native mammals) than the other fenced area (with mammals) and outside the fence (data reported as statistical model results). A netting fence was constructed in 1997 and all rabbits, cats and foxes were removed. In 1999, locally extinct small mammals were reintroduced to the fenced area. The fenced area was expanded four times in 1999–2005, and one area received no native small mammals. In 1998, twenty-four trapping sites (12 inside the fence, 12 outside) were established. In 1999, six “outside” sites became “inside” sites as the fenced area expanded, and five new “outside” sites were established. Sites were trapped for four nights (6 pitfall traps, and 10 m drift fence) in April 1998–2000 and February 2001–2005.

   Study and other actions tested

   Referenced paper

   Moseby K.E., Hill B.M. & Read J.L. (2009) Arid recovery - a comparison of reptile and small mammal populations inside and outside a large rabbit, cat and fox-proof exclosure in arid south Australia. Austral Ecology, 34, 156-169.

4. A randomized, controlled, before-and-after study in 2004–2006 on a coastal dune site on South Island, New Zealand (Lettink et al. 2010) found that use of predator exclosure fences did not result in increased survival of McCann’s skink Oligosoma maccanni compared to when no exclosure fencing was used. Average change in skink survival before and after installation of exclosure fences did not differ between sites with exclosures (survival changed by 1%) and sites without exclosures (survival changed by -1%). Four sites each were assigned to one of four treatments: exclosure fences (25 x 25 m area, 1 m high chicken wire fence, bird netting on top), exclosure fence and artificial refuges (32 refuges/site); artificial refuges only; and no treatment. Skinks were sampled annually using 4-day pitfall trapping sessions in February and March 2004–2006 with fencing and refuges placed into randomly allocated grids immediately before the second year.

   Study and other actions tested

   Referenced paper

   Lettink M., Norbury G., Cree A., Seddon P.J., Duncan R.P. & Schwarz C.J. (2010) Removal of introduced predators, but not artificial refuge supplementation, increases skink survival in coastal duneland. Biological Conservation, 143, 72-77.

5. A paired sites, controlled, before-and-after study in 1993–1996 and 2007 in chenopod scrubland in South Australia, Australia (Read & Cunningham 2010) found that fencing to exclude predators and herbivores had mixed effects on different reptile species and species groups. One gecko species increased and two geckos decreased in abundance after exclusion fencing was added, compared to before when the same plots were grazed (knob-tailed gecko Nephrurus levis after fencing: 3.3 individuals/plot vs. grazed: 0.3–0.5 individuals/plot; tessellated gecko Diplodactylus tessellatus 0.0 vs. 1.3–1.8; variable fat-tailed gecko Diplodactylus conspicullatus 0.4 vs. 1.5–1.9). See paper for details of other species responses. Five grazed sites and four paired sites of differing grazing pressure were set out in 1993 (low intensity grazing: <12 cattle dung/ha; medium: 12–100; high: >120). Following the initial four years of the study, three of the eight grazing pressure sites were fenced to exclude cattle and predators. Reptiles were sampled for 10 days in summer from 1993–1996 and 2007 using 300 mm long flymesh drift fences with 13 unbaited pitfall traps (500 mm deep x 150 mm wide, 8 m apart). Lizards were marked by toe clips.

   Study and other actions tested

   Referenced paper

   Read J.L. & Cunningham R. (2010) Relative impacts of cattle grazing and feral animals on an Australian arid zone reptile and small mammal assemblage. Austral Ecology, 35, 314-324.

6. A replicated, controlled study in 2009 in open shrubland in Catalonia, Spain (Vilardell et al. 2012) found that fencing nesting sites reduced predation by some species on artificial western Hermann’s tortoise Testudo hermanni hermanni nests and increased the time until predation occurred in one of two trials. In a first trial, artificial Hermann’s tortoise nests in fenced areas survived longer until depredation (15 days) compared to artificial tortoise nests in unfenced areas (<2 days). In a second trial one month later, all fenced and unfenced nests were depredated within three days. Authors report that fencing did not prevent predation by beech martens Martes foina, but other predators in the area (wild boar Sus scrofa, red fox Vulpes vulpes, common genets Genetta genetta and European badgers Meles meles) were successfully excluded (see original paper for details). Predation of artificial tortoise nests (three buried quail Corturnix coturnix eggs) was monitored in a nature reserve in sixteen 100 m² plots which had been cleared to 3% of shrub cover using pruning shears. Half of the plots were enclosed with a mesh fence (200 cm high). In September and again in October 2009, one artificial nest was placed in the centre of each plot. Predation was monitored by trail cameras and visual signs.

   Study and other actions tested

   Referenced paper

   Vilardell A., Capalleras X., Budó J. & Pons P. (2012) Predator identification and effects of habitat management and fencing on depredation rates of simulated nests of an endangered population of Hermann’s tortoises. European Journal of Wildlife Research (formerly Zeitschrift für Jagdwissenschaft 1955-2003), 58, 707-713.

7. A replicated, randomized, controlled study in 2002–2005 in a pine forest in Georgia, USA (Smith et al. 2013) found that constructing fences to exclude predators along with predator removal resulted in higher survival of gopher tortoise Gopherus polyphemus nests and hatchlings compared to areas with no fencing or predator removal. Survival was higher inside fenced areas compared to outside for both nests (fenced: 52 of 78, 66% survived; unfenced: 26 of 73, 35%) and hatchlings (fenced: 74% survived for 1 year; unfenced: 38%). In 2002–2003, four plots (40 ha) were randomly selected and enclosed in 1.1m high mesh fence with electrical wires at the top and bottom. A further four plots were left unfenced.  In 2002–2003, all mammalian predators within the exclosures were live-trapped and removed, and in 2003–2005, further trapping of predators was conducted. Predators that re-entered exclosures were euthanized. In May–June 2003–2005, all tortoise burrows were searched for nests, and active nests were monitored 1–2 times/week for up to 110 days. In 2004, forty hatchlings from 13 different nests were fitted with radio transmitters and monitored for up to a year.

   Study and other actions tested

   Referenced paper

   Smith L.L., Steen D.A., Conner L.M. & Rutledge J.C. (2013) Effects of predator exclusion on nest and hatchling survival in the gopher tortoise. Journal of Wildlife Management, 77, 352-358.

8. A study in 2009–2012 in an area of mixed shrub and grassland in Otago, New Zealand (Norbury et al. 2014) found that survival of captive-bred Otago skinks Oligosoma otagense released into an enclosure was higher for those released when house mice Mus musculus had been eradicated compared to when skinks were released in the presence of mice. Authors reported that post-release survival was higher for skinks released with no mice present (44%) compared to survival of skinks released just prior to reinvasion by mice (15%; see paper for details). Survival of established skinks (2 years after their release) after the mouse reinvasion was higher (91%) than for newly released skinks in the presence of mice (17%). In 2009, a 0.3 ha area was enclosed within a mammal resistant fence (1.9 m high) and over a six month period, all mammals inside the enclosure were eradicated using a range of baited traps. After eradication, 12 captive-bred adult skinks were released in the enclosure following eight weeks in quarantine. In 2011, an additional 16 skinks were quarantined and released. House mice reinvaded during 2012 and were again eradicated using live capture traps and poison bait stations. In 2009–2012, starting 7–10 days after release, skinks were monitored every 15 days by a walking survey of the enclosure.

   Study and other actions tested

   Referenced paper

   Norbury G., Van Den Munckhof M., Neitzel S., Hutcheon A., Reardon J. & Ludwig K. (2014) Impacts of invasive house mice on post-release survival of translocated lizards. New Zealand Journal of Ecology, 38, 322-327.

9. A replicated, randomized, controlled study in 2014–2015 in mixed forest and agricultural land in Georgia, USA (Dziadzio et al. 2016) found that using fencing and overhead netting to control vertebrate predators (as well as nest cage covers) did not reduce predation of gopher tortoise hatchlings Gopherus polyphemus. Gopher tortoise hatchling predation by vertebrate predators in enclosures with overhead netting was the same (3 individuals) as in enclosures with no netting. Enclosures with overhead netting had fewer signs of vertebrate predators (mammals, birds and snakes) compared to those without (signs included raccoon Procyon lotor digging and tracks, no data provided). In May–June 2014, wild-laid gopher tortoise nests were relocated to eight fenced 0.2 ha enclosures (four nests/enclosure, two eggs/nest, 64 total eggs). All nests were covered with cloth cages (30 x 30 x 12 cm). Four of eight enclosures were covered with game farm netting and UV twine to exclude aerial and terrestrial vertebrate predators (mammals, birds and snakes). Four enclosures (two with overhead netting; two without netting) were also treated with insecticide to reduce fire ant numbers. Nests were monitored weekly until two weeks before expected emergence, daily thereafter and excavated after 120 days. Hatchlings were radio tracked (16 individuals each from insecticide-treated and untreated enclosures) from August 2014 to March 2015.

   Study and other actions tested

   Referenced paper

   Dziadzio M.C., Chandler R.B., Smith L.L. & Castleberry S.B. (2016) Impacts of red imported fire ants (Solenopsis invicta) on nestling and hatchling gopher tortoises (gopherus polyphemus) in southwest Georgia, USA. Herpetological Conservation and Biology, 11, 527-538.

10. A replicated, controlled, before-and-after study in 2013–2015 in tropical savanna in the Northern Territory, Australia (Stokeld et al. 2018) found that erecting fencing to exclude feral cats Felis catus (and potentially other carnivores and herbivores) combined with fire suppression increased reptile abundance over time, but effects on reptile species richness were inconclusive. Average reptile abundance doubled over two years in plots with exclusion fencing and fire suppression (2013: 0.3 reptiles/plot; 2015: 0.7 reptiles/plot; results standardised by sampling effort), compared to plots without fencing (2013: 0.6 reptiles/plot; 2015: 0.5 reptiles/plot; results standardised by sampling effort). The effects of fencing and/or fire suppression on species richness was inconclusive (see original paper for details). Cat density in the study area was 0.2 cats/km². Cats were detected at all non-fenced plots during the study. Only one cat was found and removed from a fenced plot (within one week of fence completion). Data were collected from six plots (64 ha) with two each treated with: exclusion fencing and fire suppression; no exclusion fencing but fire suppression; and no exclusion fencing or fire suppression. Exclusion fences (installed December 2013) were 1,800 mm high with a curved floppy section 450 mm at the top of the fence above ground and 550 mm below ground. Fire suppression included 8 m wide firebreaks, early dry season fuel reduction burning around external perimeters, and active fire suppression inside the plots. Reptiles were monitored seasonally (March-April, June-July, October-November) in six transects/plot using drift fences and pitfall traps in 2013–2015. Cats were monitored using camera traps. Abundance of other carnivores and herbivores in/around the study site was not monitored.

    Study and other actions tested

    Referenced paper

    Stokeld D., Fisher A., Gentles T., Hill B.M., Woinarski J.C.Z., Young S. & Gillespie G.R. (2018) Rapid increase of Australian tropical savanna reptile abundance following exclusion of feral cats. Biological Conservation, 225, 213-221.