Evaluating factors affecting amphibian mortality on roads : the case of the Common Toad Bufo bufo , near a breeding place

The Common Toad Bufo bufo is the amphibian with the highest rates of road mortality in many European countries. This elevated incidence of road kills has frequently been associated with migration to breeding sites. In this study, we analysed the mortality of the Common Toad in the road network in Catalonia (NE Spain), and investigated the related causative factors on four roads near a breeding site in the Pyrenees. Results suggest that the high mortality rate is due to a combination of factors: toad abundance, traffic density and quality of water bodies for breeding. On the road with the highest incidence of road kills we investigated whether deaths occurred at specific spots or in a random manner. The road was divided into 500 m sections and each section was classified according to biotic (type of vegetation) and abiotic (presence of streams, roadside topography) variables. Multiple correspondence analysis showed that sections with streams crossing under the road had the highest mortality rate, suggesting that such water bodies flowing into the breeding pond are the toads’ main migratory pathways for hibernation and breeding. As toads use the same migratory routes each year, it is critical to identify areas with a high potential mortality so that efficient measures can be designed to increase wildlife permeability, and thereby reduce habitat fragmentation. This methodology could be applied in other areas with high amphibian mortality.

The Common Toad Bufo bufo, present in most parts of Europe, is the amphibian with the widest Palaearctic distribution (Lizana, 2002).It is abundant and unfortunately the amphibian most frequently found dead on roads in many European countries (Langton, 1989a).Bufo bufo mortality has frequently been associated with migration to breeding sites (Feldmann & Geiger, 1989;Orlowski, 2007;Van Gelder, 1973), when hundreds of specimens are killed by traffic during the reproductive period (Cooke, 1995;Frétey et al., 2004).However, reports on biotic (i.e.vegetation) and abiotic (i.e.roadside topography) data related to toad mortality are scarce (Clevenger et al., 2003).This aspect is relevant, since every year toads use the same migratory pathways from hibernation to breeding sites.Areas with migratory routes that present potentially high mortality should be determined and efficient measures should be adopted to reduce road kills.
The aims of this study were: 1) to detect road mortality of the common toad in the study area, 2) to determine the influence of traffic density on toad mortality and, 3) to examine whether road kills occurred randomly or at specific spots related to toad migratory pathways.

Material and methods
The study was conducted in La Pobla de Segur (542 m a.s.l., Lleida province, Spain), a mountainous region located in the Pyrenees (41º 15' N and 0º 58' E, fig.1).Field work was carried out in the valley of the Noguera Pallaresa River, in an area where the river forms the Sant Antoni reservoir (267 hm 3 ).The topography is moderately abrupt with several streams that flow into the reservoir.Potential vegetation is composed of oak (Quercus faginea) although often replaced by pine plantations (e.g.Pinus nigra) and, in areas with low slope, by agricultural fields.All the roads studied in this work were surveyed by car at 20 km/h speed and recently killed common toads and other amphibians were recorded.At this speed it is possible to observe all kind of road-kills, hence avoiding biases due to the size of the amphibian species sampled (see Montori et al., 2001 for methodological details).
To determine whether the common toad mortality was important with respect to other amphibian causalities in the study area a 20-km stretch of a road near La Pobla de Segur was surveyed every two weeks from May 2000 to April 2001, to detect the species with highest mortality rates.Bufo bufo road kills in the study area were compared with respect to results on another 45 roads in the Catalan road network , using the same methodology (Montori et al., 2004).
To find whether traffic density and size of breeding sites were relevant factors to explain toad mortality dead toads were searched in four stretches of road around La Pobla de Segur.These roads differed in traffic density and type of breeding pond (fig.1): a. C-13, between La Pobla de Segur and Tremp (11 km), the main road, running alongside the dam, with high traffic density as it links big cities further south (Barcelona, Lleida) with the tourist area of the Pyrenees; b.N-260, from La Pobla de Segur in the direction of Sort (5 km), with the same traffic density but running close to the Noguera Pallaresa River; c. N-260, from La Pobla de Segur going towards Senterada (5 km), with half the former traffic density and running close to the Flamisell River, a medium-size river which flows into the Noguera Pallaresa River in La Pobla de Segur, and d.A local road from La Pobla de Segur to Aramunt (7 km), with low traffic density, running alongside the dam on the opposite side to that of the C-13.No correction measures to reduce wildlife mortality (e.g.drift fences and tunnels) have been adopted on these roads to date.The common toad is an explosive breeder (Wells, 1977) that selects spring periods with high temperature and rainfall to move from hibernation to breeding sites (Sinsch, 1988;Zuiderwijk, 1989).For this reason, these four roads were surveyed during the peak of migratory activity, which occurred on 5-7 April 2001, during a rainy episode, when hundreds of specimens migrated from hillsides to the dam for breeding.Mortality on these four roads was compared by the number of killed toads per km, IKA (see Carretero & Rosell, 2000).
To assess whether toad road kills were randomly distributed on roads running close to breeding points we studied only the C-13 stretch between La Pobla de Segur and Tremp as it runs alongside the dam (fig. 1) and showed the highest IKA score.The 11 km section between these two localities was divided into 500-meter segments and toads were counted in each of them.Segments were separately characterized according to the type of vegetation, type of roadside, roadside slope, presence of local roads and presence of streams that cross under the road.Association between these variables and the abundance of toads was assessed by a multiple correspondence analysis.

Common toad mortality in the Pyrenees
On the road surveyed from May 2000 to April 2001, we found three species of amphibians in road kills: the midwife toad Alytes obstetricans, the common toad Bufo bufo and the common frog Rana temporaria.Most of the species present in this area (Llorente et al., 1995) were thus affected, although the common toad was that most frequently killed (92% of occurrence).Consistent with these results, preliminary data of standardized transects on 45 roads in Catalonia (NE Spain) showed that the common toad Bufo bufo is the amphibian most frequently killed on roads (29% of observations, n = 1,240, unpublished data from the authors) in this area.Half-mountainous and rainy regions as our study area showed the highest mortality rates of common toads in Catalonia.

Toad mortality related to traffic density and breeding sites
The numbers of Bufo bufo found killed on the four roads surveyed is shown in table 1.The IKA scores revealed differences between roads according to traffic density and potential reproductive sites.Roads with high density traffic reached high IKA scores when they ran close to the dam (C-13 road), whereas they showed lower values running alongside rivers (N-260 from La Pobla de Segur to Sort).Furthermore, the local road from La Pobla de Segur to Aramunt ran close to the reservoir and, despite its low traffic density, the IKA score was higher than in the N-260 road.

Factors affecting toad mortality
In the C-13 road (La Pobla de Segur-Tremp), 99 common toads recently killed in the 11 km section were recorded.Toads did not appear randomly along the road.The number of toads did not increase from the closest to the farthest segment to the head of the dam (Spearman r = 0.04, p = 0.8, n = 21).Although there were toads in all the segments, the distribution differed significantly between them (range from 1 to 23 toads), suggesting that certain spots concentrated many specimens.The multiple correspondence analysis (fig.2) revealed that segments with high mortality were characterized by: 1) the presence of streams crossing under the road, 2) low slope of the roadside, and 3) vegetation characterized by a mosaic of natural vegetation and agricultural lands.The variable best associated with mortality was the presence of streams crossing under the road.Segments with a stream had more kills than those without (median = 5, mean = 8.9 ± 2.8 toads in 7 segments with stream; median = 2, mean = 2.5 ± 0.4 toads in 14 segments without stream; Z = 10, p = 0.004).

Discussion
Differences in the numbers of amphibians killed on different roads may be due to several factors as toad abundance and quality of breeding ponds in the road vicinity.Stretches of N-260 exhibited fewer road kills than C-13 and the local road (table 1).These differences suggest that toad populations are smaller in the vicinity of mountainous rivers like the Noguera Pallaresa and the Flamisell than in the hills around the dam.In Poland, road-kills were closely related to toad abundance and to the size of water bodies for breeding (Orlowski, 2007).Our results agree with these conclusions as the Sant Antoni reservoir is the biggest water body for breeding in the study area as mountainous rivers have only small ponds for breeding due to the waterspeed.This finding supports the fact that the presence of quiet or slow-moving water seems to be Bufo bufo's only requirement for breeding (Augert & Guyetant, 1995;Lizana, 1997).Reservoirs in the Pyrenees like the Sant Antoni dam seem to be suitable water bodies to maintain important breeding populations of the common toad.Differences in toad road kills between the two roads that run alongside the dam (C-13 and the local road, fig. 1  and table 1) support the role of traffic density in toad mortality as has been extensively reported (Fahring et al., 1995).
Intensive surveys of roads near the breeding ponds proved an effective method to detect migratory paths from hibernation sites to reproductive ponds.In the study area, we detected that common toads followed small streams to move from hibernation to breeding sites.Generally, toads exhibit a high fidelity to their natal pond (Sinsch, 1989) and, like other amphibians, non-random migratory routes through specific corridors are followed every year (Langton, 1989b;Dodd & Cade, 1998;De la Torre & Sobrino, 2001).The streams that flow into the Sant Antoni reservoir in La Pobla de Segur seem suitable routes for toad migration because amphibian displacements are sensitive to all barriers, both natural and anthropic (Sinsch, 1988(Sinsch, , 1992;;Ray et al., 2002), whereas natural routes originated by rivers present few obstacles.Measures to minimise toad mortality should focus on spots where streams cross under roads near the breeding sites.The section with the highest number of killed toads in the C-13 road includes a hillside of 40º slope with herbaceous and shrub vegetation.In this section, a stream crosses under the road through a 30 m long, 10 m wide, and 8 m high tunnel.High mortality in this spot was detected in previous visits, most  (Slater, 1989(Slater, , 1994, pers. obs.), pers. obs.).In April 1998April , 2000April and 2001, , this point had been visited and 64, 46 and 82 kills were detected, respectively implying that multiple road kills occurred every year at this point and suggesting that the stream crossing the road in this section could be a key migratory pathway for common toads.Unfortunately, we lack information concerning the proportion of the breeding populations which were killed by traffic.However, the fact that the road is not new suggests that a sufficient number of specimens manage to cross the road annually and road kills are not of particular relevance for the breeding population.
Other results are, however, interesting in other ways.Despite the existence of tunnels for streams, many toads climbed the embankment onto the road and were killed by traffic.This observation suggests that road permeability by big tunnels is not always an effective measure to prevent toad mortality during migration.The use of some sort of low fencing or barrier to prevent road access for migrating amphibians and ideally guide them to a tunnel entrance may contribute to decreasing toad mortality (Galet, 1995;Marshall et al., 1997;Rosell et  1997; SCV, 1997;Rosell & Velasco, 1999).Moreover, new infrastructures like highways and high speed railways use viaducts when they cross threatened ecosystems.This seems to be a suitable solution to increase wildlife permeability and reduce landscape fragmentation.In short, the present study shows that streams that flow into breeding points may be effective migratory routes for common toads to follow between breeding and terrestrial hibernation sites.Once detected, these natural pathways, partially blocked by roads, should be given high attention and taken into account when applying measures to reduce mortality (Schlupp & Podloucky, 1994).Although the common toad is a widespread species in Spain, its high mortality rate and the increasing road network suggest it is necessary to pay attention to breeding areas, particularly in very disturbed regions.

Fig. 2 .
Fig. 2. Plot of the multiple correspondence analyses to detect the variables linked to the presence of killed common toads Bufo bufo in the C-13 road.Variables (categories): Toad presence (yes, no); Presence of streams (yes, no); Presence of local ways (yes, no); Slope of the road side (high, medium, low); Type of vegetation (natural vegetation, fields, nat.veget.+ fields).

Table 1 .
Number of common toads Bufo bufo killed and IKA scores (no./ km) of four roads in La Pobla de Segur (Lleida, Spain): R.