Providing evidence to improve practice

Action: Install barrier fencing along roads Amphibian Conservation

Key messages

  • Seven of eight studies (including one replicated and two controlled studies) in Germany, Canada and the USA found that barrier fencing with culverts decreased amphibian road deaths, or decreased deaths provided that the fence length and material were effective. One found that low numbers of amphibians were diverted by barriers during breeding migrations.
  • One replicated study in the USA found that barriers at least 0.6 m high were required to prevent green frogs and leopard frogs climbing over. Two studies in the Netherlands and USA found that treefrogs and 10% of common toads climbed over barrier fencing during breeding migrations.

Supporting evidence from individual studies


A study in 1984–1985 of a barrier fence and wildlife tunnel in Lower Saxony, Germany (Buck-Dobrick & Dobrick 1989) found that many common toads Bufo bufo and common frogs Rana temporaria went around the end of the barrier fence and were killed on the road during breeding migrations. In 1985, deaths were reduced by lengthening the fence. Initially, fences 350 m long were installed on both sides of the road. A concrete tunnel was located in the centre of the fences. Common toads and common frogs were monitored in March–April. Toads were tagged.




A replicated study in 1986 of 114 sites including at least 60 amphibian barrier fences, 11 road closure sites and 23 hand-collected human assisted crossings in Nordrhein-Westphalia, Germany (Feldmann & Geiger 1989) found that a total of 131,061 amphibians were protected from death on roads. Between one and 116,515 individuals of 14 species were recorded at each barrier fence, road crossings or hand-collected crossing. The majority of the 60 barrier fences to protect amphibians were constructed from polythene and averaged 600 m in length (range: 30–3,000 m). Animals were collected by hand alone at 23 sites and at 11 sites roads were closed for migrations. Nine sites had a combination of two of the interventions and for 20 sites it was unknown which of the interventions were used.



A before-and-after study in 1987–1988 of a barrier fence and two amphibian tunnels in the Mittelgebirge region of West Germany (Meinig 1989) found that once an effective fence was installed, numbers of migrating amphibians killed on the road during the breeding migration decreased. Prior to the new fence numbers killed were 109/night, compared to just 20 in 1987 and 30 in 1988. Overall, 85% of amphibians recorded at the fence passed through the tunnels. The total number of individuals captured at the fence and surroundings during the spring migration were 2,432 in 1987 and 2,050 in 1988. Of 211 toads marked at the fence in 1987, 68% were recaptured at tunnel exits within five days. Two drain channels with metal grid roofs were installed in the road in 1981. A more effective fence of plastic fabric similar to wire mesh (1 m high) was installed at tunnel entrances and parallel to the road in 1987. Pitfall traps were set at each end of the fence and at tunnel exits.



A study of barrier fencing between five amphibian tunnels in Overveen in the Netherlands (Zuiderwijk 1989) found that 10% of the population of 2,000–3,000 common toads Bufo bufo climbed over the fencing during breeding migrations. The remaining toads walked along the fence, but only 4% used the tunnels. The others were captured in pitfall traps and carried across the road. The cast-iron tunnels had been installed nine years before the study. The road had permanent barrier fencing.



A controlled, before-and-after study in 2001–2002 of a barrier wall linking culverts along a highway in Florida, USA (Dodd, Barichivich & Smith 2004) found that the wall significantly decreased amphibian road deaths, apart from treefrogs (Hylidae), which could climb over. A total of 19 amphibian road-kills were found on the 3 km section with barrier, compared to 326 kills on the 500 m section with no barrier. Treefrogs were excluded from these figures. Treefrog mortality increased after construction of the barrier and culverts (from 149 to 194 over three survey sections). In 2001, a 1 m high concrete wall with a 15 cm overhang was erected along the highway, parallel to a wetland prairie. The wall extended 3 km on each side of the road. Concrete culverts under the highway were increased from four to eight. The highway and grass verge were monitored from 200 m before the start of the barrier until 200 m past the end. Monitoring was undertaken on three consecutive days from dawn each week from March 2001 to March 2002.



A study of drainage culverts modified with diversion fencing in Texas, USA (Jochimsen et al. 2004) found that fencing reduced road-kills in its vicinity, but aggregations of dead toads were recorded at the barrier endpoints. No Houston toads Bufo houstonensis used the culverts, which became impassable when flooded. Short sections of steel diversion fencing were added to existing drainage culverts to guide toads from known migration routes into the culverts.


A before-and-after study in 2000–2003 of temporary fencing along a highway to a culvert by Lake Jackson, Florida, USA (Aresco 2005) found that 70% of amphibians and reptiles (not including turtles) were diverted from the highway towards the culvert. Twelve amphibian species were recorded along the barrier. Fences diverted 74% of the 1,088 upland and semi-aquatic amphibians and reptiles from the highway (at fence: 74%; dead on road: 26%). Twenty-two percent of the 299 aquatic animals were also diverted (alive at fence: 22%; dead at fence: 2%; dead on road: 76%). In particular, the fence diverted small frogs and toads. Some species were significantly underestimated. The temporary fence was installed along the highway to divert animals to a culvert in April 2000 (700 m; 0.4 m high) and September 2000 (600 m). Monitoring was undertaken 1–4 times/day by walking the fence and checking the road and culvert until November 2003.


A study in 2004–2008 of a barrier wall leading to a culvert under a new highway through upland forest in New Hampshire, USA (Merrow 2007) found that the wall only diverted small numbers of amphibians towards the culvert. Small numbers of spotted salamanders Ambystoma maculatum and wood frogs Rana sylvatica were found moving along the wall. However, small numbers were also found crossing the road in areas without a wall or culvert. There was no evidence that amphibians used the tunnel during the first three years. The diversion wall was at least 0.3 m high and extended from the culvert to a stone-lined stream channel on one side and a larger pedestrian culvert on the other. Spring amphibian migrations were monitored for three years after construction.



A replicated study in 2005–2006 of different height barrier fencing in a Wildlife Management Area, New York, USA (Woltz, Gibbs & Ducey 2008) found that fences of at least 0.6 m excluded most green frogs Rana clamitans and leopard frogs Rana pipiens. Fences 0.6 m high were more effective at excluding frogs (97–100%) than 0.3 m fences (77–80%). Only one leopard frog climbed over the 0.9 m high fence. Opaque, corrugated plastic fences were used to construct three nested, circular enclosures of heights 0.3, 0.6 and 0.9 m. Local green frogs (n = 135) and leopard frogs (n = 187) were placed in the centre of each arena and left for 15 min to attempt to scale the fences.



A controlled study in 2009 of wildlife culverts with barrier fencing along a new highway through wetlands near Whistler, Canada (Malt 2011) found that drift-fencing or barriers directing amphibians towards culverts significantly reduced road-kills. Road-kill rates were reduced by over 50% along road sections with ≥ 50 m of drift-fencing or barriers compared to those with no barriers (2–8 vs 15–17 killed/50 m section). Additional fencing was therefore installed. Eight wildlife culvert underpasses were constructed along the section through the wetland. Drift-fences were installed to funnel animals towards culverts. Barrier walls were also installed to prevent migration along some sections. Amphibians were monitored using road-kill surveys, remote cameras at culvert entrances and a mark‐recapture study of red‐legged frogs Rana aurora.


Referenced papers

Please cite as:

Smith, R.K., Meredith, H. & Sutherland, W.J. (2017) Amphibian Conservation. Pages 9-65 in: W.J. Sutherland, L.V. Dicks, N. Ockendon & R.K. Smith (eds) What Works in Conservation 2017. Open Book Publishers, Cambridge, UK.