Skip to main content

Amphibian and reptile road-kills on tertiary roads in relation to landscape structure: using a citizen science approach with open-access land cover data



Amphibians and reptiles are among the most endangered vertebrate species worldwide. However, little is known how they are affected by road-kills on tertiary roads and whether the surrounding landscape structure can explain road-kill patterns. The aim of our study was to examine the applicability of open-access remote sensing data for a large-scale citizen science approach to describe spatial patterns of road-killed amphibians and reptiles on tertiary roads. Using a citizen science app we monitored road-kills of amphibians and reptiles along 97.5 km of tertiary roads covering agricultural, municipal and interurban roads as well as cycling paths in eastern Austria over two seasons. Surrounding landscape was assessed using open access land cover classes for the region (Coordination of Information on the Environment, CORINE). Hotspot analysis was performed using kernel density estimation (KDE+). Relations between land cover classes and amphibian and reptile road-kills were analysed with conditional probabilities and general linear models (GLM). We also estimated the potential cost-efficiency of a large scale citizen science monitoring project.


We recorded 180 amphibian and 72 reptile road-kills comprising eight species mainly occurring on agricultural roads. KDE+ analyses revealed a significant clustering of road-killed amphibians and reptiles, which is an important information for authorities aiming to mitigate road-kills. Overall, hotspots of amphibian and reptile road-kills were next to the land cover classes arable land, suburban areas and vineyards. Conditional probabilities and GLMs identified road-kills especially next to preferred habitats of green toad, common toad and grass snake, the most often found road-killed species. A citizen science approach appeared to be more cost-efficient than monitoring by professional researchers only when more than 400 km of road are monitored.


Our findings showed that freely available remote sensing data in combination with a citizen science approach would be a cost-efficient method aiming to identify and monitor road-kill hotspots of amphibians and reptiles on a larger scale.


Amphibian and reptile species are endangered worldwide, suffering from numerous threats such as habitat modification and fragmentation, diseases, pollution, invasive species or climate change [1,2,3,4]. In Austria, where the current study was conducted, all 20 amphibian species and all 14 reptile species are protected by national conservation laws [5]. Focusing on habitat fragmentation, roads can have various negative effects on many vertebrate species [6,7,8,9]. The most direct negative effect of road traffic on animal populations is through fatal collisions with vehicles, i.e. road-kill [2]. Road-kill does not affect all taxonomic groups in the same way. Amphibians (toads, newts and salamanders) are mostly affected by road-kill when crossing roads during migration between their breeding and hibernation sites; reptiles are even attracted by the favourable microclimate on roads [4, 8, 10,11,12]. Amphibians and some reptile species are even more susceptible to road-kill because they get immobile in response to an approaching vehicle [13, 14]. As mentioned by Rytwinski and Fahrig [15] there are relatively few studies of the effects of roads on amphibian and reptile populations, despite the fact that amphibians and reptiles have significantly more species at risk than mammals or birds. Additionally, available data of road-kills of amphibians and reptiles for Europe is scarce and often not comparable due to different study designs [16] and because of species-specific response patterns [17, 18].

In temperate regions like Central Europe many amphibian and some reptile species require complex landscapes including wetlands for reproduction and woody areas for foraging and hibernation [19]. Hence, the composition of landscape surrounding roads is an important factor influencing the number of road-kills for both amphibian and reptile populations [1, 20, 21]. Some reptile species (e.g. European green lizard, Lacerta viridis, Laurenti, 1768) are more selective to a habitat than other species (e.g. Grass snake, Natrix natrix, Linnaeus, 1758), which migrate long distances from summer to winter habitats. Most adult toads are susceptible to road-kills when migrating to breeding ponds [1, 10], while road-kills of juvenile toads are more dispersed in space and time when moving to hibernation sites in late summer and autumn. In the northern hemisphere most road-kill studies investigate the impact of wide roads with high traffic volumes that are usually fenced off and therefore are a stronger barrier to amphibians and reptiles [6, 22, 23]. However, evidence is increasing that tertiary road networks especially affect small animal species like herpetofauna [24,24,25,27]. Nevertheless, the influences of tertiary roads on amphibian and reptile populations on a landscape level are not often studied. When monitoring herpetofaunal road-kills challenges exist because small road-killed animals disappear quickly [28], the diverse network of tertiary roads is dense and some road-killed species are difficult to identify.

The standard approach to assess the effects of road-kill on animals is to collect data on a regular basis along certain routes [29], but this method is very cost-intensive and time consuming [30]. An alternative is to use a citizen science approach, i.e. involving citizens to report road-kill sightings [29, 31,31,32,34]. However, in most citizen science projects, “presence only” data are collected, which hamper proper statistical analyses of underlying factors for road-kills [35]. The aim of the current study was to test a systematic monitoring approach that would be appropriate to engage citizens in collecting presence and absence data. Results of this pilot study would be a proof-of-concept before engaging the general public, since many challenges (e.g. motivation of citizens) exist in establishing such a citizen science monitoring approach [36]. To test this approach, we used a citizen science software for monitoring road-killed amphibians and reptiles along a fixed bicycle tour on tertiary roads. Additionally, we examine the applicability of open access remote sensing data for a large-scale citizen science approach to describe spatial patterns of road-killed amphibians and reptiles on tertiary roads. The findings will be discussed with respect to their potential for designing a cost-effective large scale road-kill monitoring system based on a citizen science approach.


We monitored road-kills for two activity periods of amphibians and reptiles from March 2014 to October 2015 on a 97.5-km stretch of road in a rural region in eastern Austria. We use the term tertiary road network to summarize (I) agricultural roads (used by farmers and cyclists, speed limit 30 km h−1), (II) cycling paths (used by cyclists only), (III) municipal roads (mainly used by residents, speed limit 50 km h−1) and (IV) interurban roads (mainly used by residents, speed limit 100 km h−1).

Study area

The monitoring was conducted between the Leithagebirge and the Lake Neusiedl in Northern Burgenland, which is located in eastern Austria (Fig. 1). We chose the area for its high biodiversity in amphibian and reptile species and its relatively dense network of tertiary roads [37]. The study route consisted of 61 km of agricultural roads, 26 km of municipal roads, 8 km of interurban roads and 2.5 km of cycling paths, and is partly in the Natura 2000 sites Neusiedler See—Nordöstliches Leithagebirge (north-east of the study route) and Mattersburger Hügelland (south-west of the study route) [38]. The landscape of the study area is topographically heterogeneous with elevations ranging from 115 to 748 m. The climate of the region is considered Continental-Pannonian with 562 mm average annual precipitation and 10.7 °C annual mean temperature (Neusiedl am See; years 1981–2010) [39]. The Leithagebirge is forming the western boundary of the Natura 2000 site and at the same time framing the lake basin. The Leithagebirge exhibits a crystalline basement with accumulated reef lime stones of tertiary origin and is marked by mainly oak and mixed oak forests. In the outlying areas open farmland (vineyards interspersed with trees, bushes and grasslands residues) and dry grasslands dominate. The study area is a biodiversity hotspot in Austria as Pannonian, Alpine and Mediterranean floristic and faunal elements intermingle [38].

Fig. 1

Location of the study route in Eastern Austria (a). Type of road sections of the study route (b)

Data collection

We monitored the selected route from March 2014 to October 2015 by using a smartphone app (Roadkill | SPOTTERON, NINC Media, Vienna, Austria) from the citizen science project Roadkill ( In the citizen science project volunteers collect presence only data of all vertebrate species killed on roads [35, 40]. The app was not adjusted to our approach but rather used as data collection tool in a more systematic way. In the current project, we used the Roadkill app to monitor road-killed amphibians and reptiles along a predefined route on average every 11 days. We monitored road-kills by cycling, since slower traveling speed results in higher detectability especially of smaller species [41]. We stopped at each road-killed animal and filled in the form provided by the SPOTTERON Roadkill application, which included coordinates of the spot, a picture of the animal, species name and number of individuals. All other locations where no road-kill was detected represent absence data. Our monitoring therefore results in a presence/absence dataset of locations along the route at different points in time.

Remote sensing data

We used only open data to follow the idea of open science and to test the applicability of freely available remote sensing data in describing spatial patterns of road-killed amphibians and reptiles. We downloaded the most current CORINE (Coordination of Information on the Environment) land cover data from copernicus land cover monitoring services [42]. We divided the study route in 500 m sections and for each section assessed the surrounding landscape within a 500 m buffer on each side of the road by recording the area of each land cover class. The sum of the area of all land cover classes results in the study area. CORINE land cover (CLC) is a geographic land cover/land use database for a Pan-European region. CLC data provides information on the biophysical characteristics of the Earth’s surface based on images acquired by Earth Observation satellites with a Minimal Mapping Unit (MMU) of 25 ha [43]. The CLC 2012 uses a standardized European level-3 nomenclature consisting of 37 classes [44]. We used only the 14 classes which are present in the study area. The study area consisted of a high number of vineyards, arable land and urban areas (Table 1).

Table 1 CORINE land cover classes in the study area. Land cover classes in descending order of proportional area

Statistical analyses

First, monthly variations of road-kills per year were analysed using Chi squared tests. Second, to find road-kill hotspots on the study route, we used the software KDE+ [45]. KDE+ identifies clusters of road-kills based on kernel density estimation and provides a measure of the significance of a hotspot based on Monte Carlo simulations [46]. Third, we divided the study route in 500 m sections to calculate conditional probabilities to assess the association of a certain land cover class with a road-kill event directly from the sample. Conditional probabilities [P(E|B)] of road-kill events (E) on each land cover class (B) were calculated to analyse which land cover classes are associated with an increased/decreased risk of road-kills, as compared to the overall probability P(E) of sections having a road-kill event. To obtain conditional probabilities, the areal fractions of land cover classes were determined for each section. The probability of a road-kill on a certain land cover class [P(E∩B)] is the total area of B of sections affected by road-kills [A(E∩B)] divided by the total area of all sections (A). [P(E∩B)] was finally divided by the overall availability (i.e., total area fraction) of this land cover class in the study area [P(B)] to obtain its conditional probability. Fourth, to analyse which land cover class is related to the number of road-kills we employed general linear models (GLM) with Poisson distributions. Possible collinearity was handled by means of stepwise model fitting and variance inflation factors (VIF). From all fitted models that do not contain predictors VIF >10, the one with the lowest value in the akaike information criterion (AIC) was chosen.

Note that conditional probabilities and general linear models are performed to 500 m road sections whereas the study route as a whole was used for the kernel density estimation. Chi squared tests and general linear models were performed using the “Rcmdr” package (R Commander Version 2.2–3) [47] in the open source program “R” (R version 3.2.4) [48].

Cost efficiency estimation

A rough estimation of the cost efficiency of our citizen science approach was made to compare costs of our pilot study with a classical monitoring approach. Numbers used for the calculation are based on standard staff costs in Austria, one offer provided by an Austrian engineering office and one offer provided by the software company which developed apps for project Roadkill.


During our monitoring of 20 months, we found 252 road-killed animals (180 amphibians and 72 reptiles) comprising eight species (Table 2). Green toad (Bufo viridis, Laurenti, 1768), common toad (Bufo bufo, Linnaeus, 1758) and grass snake (Natrix natrix) were the dominating species of the investigated amphibian and reptile species, respectively. Most amphibians and reptiles were killed on agricultural roads.

Table 2 Numbers of road-killed amphibians and reptiles found from March 2014–October 2015 on monitored sections of municipal roads (26 km), cycle paths (2.5 km), agricultural roads (61 km) and interurban roads (8 km)

Road-kill reports were not equally distributed across months (amphibians: X2 = 136.44, df = 7, p < 0.001; reptiles: X2 = 34.889, df = 7, p < 0.001). Figure 2 shows that most amphibians were reported in July (n = 64), followed by April (n = 36) and August (n = 30), whereas most reptiles were reported in October (n = 20) and September (n = 19).

Fig. 2

Total numbers of road-killed amphibians (white, n = 180) and reptiles (grey, n = 72) per month from March 2014–October 2015. No road-killed amphibians or reptiles were found between November 2014 and February 2015

Spatial patterns

Applying the KDE+ software to our road-kill monitoring data resulted in several hotspots including sections of 2–37 road-killed amphibians and of 2–8 road-killed reptiles, respectively (Fig. 3; Table 3). The vast majority of amphibian hotspots are next to arable land and suburban areas, whereas most reptile hotspots are located near the reed belt of the lake Neusiedl.

Fig. 3

Amphibian (blue) and reptile (purple) road-kill hotspots calculated with KDE+. Highlighted are the four strongest hotspots of amphibians (AD) and reptiles (EH). Asterisked letters differ the two hotspots in one circle

Table 3 KDE+ results of the strongest four amphibian and reptile hotspots

Green toads and common toads represented 96% of all road-killed amphibians and grass snakes represented 83% of all road-killed reptile species, therefore we focused the following analyses on these three most often found species. Number of road-killed common toads and green toads per 500 m section varied from 0 to 35. The overall probability of a road-killed toad in the study area calculated per section was 0.36. This value was then used as a benchmark for conditional probabilities of road-kill events by land cover class (Table 4). All land cover classes with P(E|B) >0.36 are associated with increased risk of road-killed common toads and green toads, whereas land cover classes with P(E|B) <0.36 constrain the risk. Transitional woodland shrub (CLC 324), Sport and leisure facilities (CLC 142) and Land principally occupied by agriculture, with significant areas of natural vegetation (CLC 243) exhibit the highest conditional probabilities, whereas Industrial or commercial units (CLC 121) and Pastures (CLC 231) were associated with the lowest conditional probabilities. Using the best fitting general linear model focusing on the most often road-killed species common toad and green toad (AIC: 687.67; Table 5), a significant positive relation of Land principally occupied by agriculture, with significant areas of natural vegetation (CLC 243) and Transitional woodland shrub (CLC 324) and amphibian road-kills was seen, indicating that this land cover classes promoted amphibian road-kills. Complex cultivation patterns (CLC 242) and Vineyards (CLC 221) had a significantly negative relationship with the number of road-killed common toads and green toads.

Table 4 Conditional probabilities of road-killed common toads and green toads for all land cover classes [P(E|B)]
Table 5 GLM containing land cover classes as explanatory variables that influence road-kill numbers of green toad and common toad

Number of road-killed grass snakes per 500 m section varied from 0 to 6, with an overall probability of a road-killed grass snake per section of 0.17. Again, conditional probabilities P(E|B) were calculated for all 14 land cover classes (Table 6). Inland marshes (CLC 411) and Sport and leisure facilities (CLC 142) appeared to promote reptile road-kills, whereas Transitional woodland shrub (CLC 324), Industrial or commercial units (CLC 121) and Continuous urban fabric (CLC 111) appeared to constrain them. Using the best fitting general linear model again excluding all but the most often road-killed reptile species grass snake (AIC: 294.79; Table 7), resulted in significantly positive relations of the land cover classes Inland marshes (CLC 411), Sport and leisure facilities (CLC 142) and Complex cultivation patterns (CLC 242) with reptile road-kills. Discontinuous urban fabric (CLC 112) and Land principally occupied by agriculture, with significant areas of natural vegetation (CLC 243) were significantly negative related to the number of road-killed grass snakes.

Table 6 Conditional probabilities of road-killed grass snakes for all land cover classes [P(E|B)]
Table 7 GLM containing land cover classes as explanatory variables that influence road-kill numbers of grass snakes

Cost efficiency estimation

We compared two cases for cost-efficiency estimation (Table 8), monitoring by researchers only and a citizen science approach. Based on our calculations, using an approach involving researchers only would result in 22,000 €. This includes monitoring road-kills by a graduate student on a study route of 100 km over two vegetation periods and recording the surrounding landscape structure and habitat descriptions by an engineering office. Alternatively, the use of citizen science in combination with open access land cover data as we supposed in our pilot study would result in 86,000 €. This includes adjusting the citizen science smartphone application of the project Roadkill to allow for monitoring roads, maintaining the application and professional support of the participants for a 2-year period.

Table 8 Cost-efficiency estimation for the cases researcher and citizen science


To our knowledge, this is among the first studies testing the suitability of a citizen science approach in examining the impact of tertiary roads and the surrounding landscape on amphibian and reptile species by using freely-available remote sensing data.

Road-kills of both amphibians and reptiles were recorded during the whole vegetation period mainly on agricultural roads. The role of tertiary roads for road-kill of endangered species is generally not widely appreciated, however is perhaps more significant in regions where higher ranking roads are already equipped with efficient road-kill mitigation measures. Green toads and common toads represented 96% of all road-killed amphibians with peaks in April, August and September. These two species are also among the most abundant species in the study region [37]. From our results it seems that identifying road-kill hotspots of both species in spring would be most straightforward because of mass migration to the spawning sites during this time [49]. However, when the goal is to assess the effect of road-kill on population dynamics, surveys would need to also include surveys in late summer and autumn when individuals of both species forage in their terrestrial habitat up to 10 km away from the breeding ponds [50]. Grass snakes represented 83% of all road-killed reptile species with road-kill peaks in September and October. Grass snake is the most abundant snake species in the study region [37]. Starting in November, Grass snakes overwinter underground in areas which are not subject to freezing (e.g. compost heaps, burrows of mice) and get active again in March or April [37]. One reason for the peaks of road-killed reptiles in September and October could be that young grass snakes hatch from their nesting sites in August and sprawl into the surrounding landscape crossing our study route. Indeed, when we double-checked the photos of the road-killed reptiles it turned out that about 80% of road-killed grass snakes were juvenile. Additionally, in September and October roads are frequently used by snakes and lizards for basking during the day, but roads during these months are also more frequently used by farmers during wine harvest. Numbers of both amphibian and reptile road-kills per month suggest, that road-kill monitoring should comprise whole activity periods to get a complete overview. Here citizen science would be a suitable approach, since it would be very costly to monitor amphibians and reptiles on tertiary roads covering a wide geographic range in short time periods with a classic approach involving researchers only.

Additionally, KDE+ analyses showed significant road-kill hotspots indicating that road-kills are not randomly distributed in the landscape. This is an important information for nature conservation authorities aiming to mitigate threats for endangered amphibian and reptile species.

Conditional probabilities and general linear models applied in our study showed a positive relationship of the land cover classes Transitional woodland shrub (CLC 324) and Land principally occupied by agriculture, with significant areas of natural vegetation (CLC 243) and common toad and green toad road-kills. It was encouraging to see, that land cover classes based on a rather coarse 500-m grid matched well with the most preferred habitats of the most abundant amphibian species in the study. Green toads and common toads use water bodies only for spawning. Green toads inhabit various kinds of terrestrial sites including gravel pits, field edges, ruderal plots, dry grassland, open forests or suburban areas whereas common toads live mainly in areas with dense vegetation such as forests and scrubland areas, parks or gardens [37]. Grass snake road-kills were positive related to land cover classes Inland marshes (CLC 411), Sport and leisure facilities (CLC 142) and Complex cultivation patterns (CLC 242). Grass snakes, as the most frequently found reptile species in our study inhabit a broad range of open or semi-open habitats, including reed belts, riparian zones, forests, gardens or parks. Grass snakes can be found in the reed belt of the lake Neusiedl in high numbers matched by the CORINE land cover class Inland marshes [37]. We monitored the study route on average every 11 days and might have underestimated the number of road-kills as the persistence time of carcasses could be lower [28, 51]. Notwithstanding this limitation, our current results are in line with previous studies. This is especially important, more frequent monitorings would be difficult over a long time span with a citizen science approach. Generally, the suitability of the CORINE land cover dataset for modelling amphibian and reptile road-kills encourages us to apply our monitoring approach to a broader geographical scale.

Based on our rough cost efficiency estimation, monitoring the influence of land cover on road-kills with a citizen science approach is suitable when monitoring more than 400 km road sections. Below 400 km a conventional monitoring approach with professional researchers only seems to be more efficient. However, this calculation is just a rough estimation and should be treated with care; it is based on Austrian standard staff costs and is calculated for investigating the factor land-cover only. If other factors besides land cover are planned to be investigated, the cost efficiency could be tremendously different.


Overall, our findings confirmed previous results showing that amphibian and reptile species are especially susceptible to road-kill in the vicinity of their preferred habitats [1, 52]. Nevertheless, this is interesting, as we achieved these results using freely available remote sensing data and a survey technique that could easily be adopted on a larger scale using a citizen science approach. We are confident, that the results of this pilot study can be used as basis for other citizen science projects in this field trying to enlarge their study area. A first step to the extension of our monitoring system would be to get an overview of road-killed amphibians and reptiles on a landscape scale by monitoring tertiary road networks potentially using a citizen science approach to cover this wide geographic range [35, 36, 53]. Furthermore, these data could then be used to reduce the impact of road traffic on amphibians and reptiles by installing temporal or permanent mitigation measures.



(akaike information criterion)


(CORINE land cover)


(coordination of information on the environment)


(general linear model)


(kernel density estimation plus)


(variance inflation factors)


(web map tile service)


  1. 1.

    Beebee TJC. Effects of Road Mortality and Mitigation Measures on Amphibian Populations. Conserv Biol. 2013. doi:10.1111/cobi.12063/abstract.

    PubMed  Google Scholar 

  2. 2.

    Rytwinski T, Fahrig L. The impacts of roads and traffic on terrestrial animal populations. Handbook road ecology. 1st ed. West Sussex: Wiley; 2015. p. 237–46.

    Google Scholar 

  3. 3.

    Gibbons JW, Scott DE, Ryan TJ, Buhlmann KA, Tuberville TD, Metts BS, et al. The global decline of reptiles, Déjà Vu Amphibians: reptile species are declining on a global scale. Six significant threats to reptile populations are habitat loss and degradation, introduced invasive species, environmental pollution, disease, unsustainable use, and global climate change. Bioscience. 2000;50:653–66.

    Article  Google Scholar 

  4. 4.

    Andrews KM, Langen TA, Struijk RPJH. Reptiles: overlooked but often at risk from roads. Handbook road ecology. 1st ed. West Sussex: Wiley; 2015. p. 1–9.

    Google Scholar 

  5. 5.

    Gollmann G. Rote Liste der in Österreich gefährdeten Lurche (Amphibia) und Kriechtiere (Reptilia). Rote Liste Gefährdeter Tiere Österr. Checklisten Gefährdungsanalysen Handl. Teil 2 Kriechtiere Lurche Fische Nachtfalter Weichtiere. Wien: Böhlau; 2007. p. 37–60.

  6. 6.

    Fahrig L, Rytwinski T. Effects of roads on animal abundance: an empirical review and synthesis. Ecol Soc. 2009;14(1):21.

  7. 7.

    van der Ree R, Smith DJ, Grilo C. Handbook of road ecology. New York: Wiley; 2015.

    Google Scholar 

  8. 8.

    Forman RTT, Sperling D, Bissonette JA. Road ecology: science and solutions. Washington, DC: Auflage: First Trade Pap; 2003.

    Google Scholar 

  9. 9.

    Andrews KM, Nanjappa P, Riley SPD. Roads and ecological infrastructure: concepts and applications for small animals. Baltimore: Wildlife Management and Conservation; 2015.

    Google Scholar 

  10. 10.

    Glista DJ, DeVault TL, DeWoody JA. Vertebrate road mortality predominantly impacts amphibians. Herpetol Conserv Biol. 2008;3:77–87.

    Google Scholar 

  11. 11.

    D’Amico M, Roman J, de los Reyes L, Revilla E. Vertebrate road-kill patterns in Mediterranean habitats: who, when and where. Biol Conserv. 2015;191:234–42.

    Article  Google Scholar 

  12. 12.

    Teixeira FZ, Coelho IP, Esperandio IB, Rosa Oliveira N, Porto Peter F, Dornelles SS, et al. Are road-kill hotspots coincident among different vertebrate groups? Oecol Aust. 2013;17:36–47.

    Article  Google Scholar 

  13. 13.

    Bouchard J, Ford AT, Eigenbrod FE, Fahrig L. Behavioral responses of northern leopard frogs (Rana pipiens) to roads and traffic: implications for population persistence. Ecol. Soc. 2009;14:23.

    Article  Google Scholar 

  14. 14.

    Mazerolle MJ, Huot M, Gravel M. Behavior of amphibians on the road in response to car traffic. Herpetologica. 2005;61:380–8.

    Article  Google Scholar 

  15. 15.

    Rytwinski T, Fahrig L. The impacts of roads and traffic on terrestrial animal populations. In: Handbook of road ecology, 1st edn. West Sussex, UK: Wiley; 2015. p. 237–46.

  16. 16.

    Elzanowski A, Ciesiolkiewicz J, Kaczor M, Radwanska J, Urban R. Amphibian road mortality in Europe: a meta-analysis with new data from Poland. Eur J Wildl Res. 2009;55:33–43.

    Article  Google Scholar 

  17. 17.

    Kovar R, Brabec M, Vita R, Bocek R. Mortality Rate and activity patterns of an Aesculapian snake (Zamenis longissimus) population divided by a busy road. J. Herpetol. 2014;48:24–33.

    Article  Google Scholar 

  18. 18.

    Robson LE, Blouin-Demers G. Eastern Hognose snakes (Heterodon platirhinos) avoid crossing paved roads, but not unpaved roads. Copeia. 2013;2013:507–11.

    Article  Google Scholar 

  19. 19.

    Semlitsch RD. Critical elements for biologically based recovery plans of aquatic-breeding amphibians. Conserv Biol. 2002;16:619–29.

    Article  Google Scholar 

  20. 20.

    Marsh DM, Trenham PC. Metapopulation dynamics and amphibian conservation. Conserv Biol. 2001;15:40–9.

    Article  Google Scholar 

  21. 21.

    Rytwinski T, Fahrig L. Do species life history traits explain population responses to roads? A meta-analysis. Biol Conserv. 2012;147:87–98.

    Article  Google Scholar 

  22. 22.

    Matos C, Sillero N, Argana E. Spatial analysis of amphibian road mortality levels in northern Portugal country roads. Amphib Reptil. 2012;33:469–83.

    Article  Google Scholar 

  23. 23.

    Sillero N. Amphibian mortality levels on Spanish country roads: descriptive and spatial analysis. Amphib Reptil. 2008;29:337–47.

    Article  Google Scholar 

  24. 24.

    Marsh DM. Edge effects of gated and ungated roads on terrestrial salamanders. J Wildl Manag. 2007;71:389–94.

    Article  Google Scholar 

  25. 25.

    Langen TA, Ogden KM, Schwarting LL. Predicting hot spots of herpetofauna road mortality along highway networks. J Wildl Manag. 2009;73:104–14.

    Article  Google Scholar 

  26. 26.

    Meek R. Where do snakes cross roads? Habitat associated road crossings and mortalities in a fragmented landscape in western France. Herpetol J. 2015;25:15–9.

    Google Scholar 

  27. 27.

    Findlay CS, Houlahan J. Anthropogenic Correlates of species richness in Southeastern Ontario Wetlands. Conserv Biol. 1997;11:1000–9.

    Article  Google Scholar 

  28. 28.

    Santos SM, Carvalho F, Mira A. How long do the dead survive on the road? Carcass persistence probability and implications for road-kill monitoring surveys. PLoS ONE. 2011;6:e25383.

    CAS  Article  PubMed  PubMed Central  Google Scholar 

  29. 29.

    Shilling FM, Perkins SE, Collinson W. Wildlife/roadkill observation and reporting systems. Handbook road ecology. 1st ed. West Sussex: Wiley; 2015. p. 492–501.

    Google Scholar 

  30. 30.

    Costa AS, Ascensão F, Bager A. Mixed sampling protocols improve the cost-effectiveness of roadkill surveys. Biodivers Conserv. 2015;24:2953–65.

    Article  Google Scholar 

  31. 31.

    Vercayie D, Herremans M. Citizen science and smartphones take roadkill monitoring to the next level. Nat Conserv. 2015;11(11):29–40.

    Article  Google Scholar 

  32. 32.

    Shilling FM, Waetjen DP. Wildlife-vehicle collision hotspots at US highway extents: scale and data source effects. Nat. Conserv. 2015;11(11):41–60.

    Article  Google Scholar 

  33. 33.

    Shilling FM. Programs | Global Roadkill Network. Glob. Roadkill Netw. 2015. Accessed 9 Jan 2016.

  34. 34.

    Lee T, Quinn MS, Duke D. Citizen, science, highways, and wildlife: using a web-based GIS to engage citizens in collecting wildlife information. Ecol Soc. 2006;11:11.

    Article  Google Scholar 

  35. 35.

    Heigl F, Stretz RC, Steiner W, Suppan F, Bauer T, Laaha G, et al. comparing road-kill datasets from hunters and citizen scientists in a landscape context. Remote Sens. 2016;8:832.

    Article  Google Scholar 

  36. 36.

    Vercayie D, Herremans M. Citizen science and smartphones take roadkill monitoring to the next level. Nat Conserv. 2015;11:29–40.

    Article  Google Scholar 

  37. 37.

    Schweiger S, Grillitsch H. Die amphibien und reptilien des neusiedler see-gebiets. Naturhistorisches Museum Wien, Nationalpark Neusiedler See—Seewinkel: Wien & Illmitz; 2015.

    Google Scholar 

  38. 38.

    European Environment Agency (EEA). Neusiedlersee - Nordöstliches Leithagebirge (AT1110137). Natura2000 standard data form for special protection areas (SPA), proposed sites for community importance (pSCI), sites of community importance (SCI) and for special areas of conservation (SAC). 2016. Accessed 4 Oct 2016.

  39. 39.

    ZAMG—Zentralanstalt für Meteorologie und Geodynamik. Klimamittel. Klimanormalperiode 1981–2010. 2012. Accessed 22 Sept 2015.

  40. 40.

    Heigl F, Zaller JG. Using a citizen science approach in higher education: a case study reporting roadkills in Austria. Hum Comput. 2014;1. Accessed 21 Jan 2015.

  41. 41.

    Teixeira FZ, Pfeifer Coelho AV, Esperandio IB, Kindel A. Vertebrate road mortality estimates: effects of sampling methods and carcass removal. Biol Conserv. 2013;157:317–23.

    Article  Google Scholar 

  42. 42.

    CLC 2012—copernicus land monitoring services. Accessed 5 July 2016.

  43. 43.

    Copernicus Programme. Copernicus land service—Pan-European component: CORINE land cover. 2015. Accessed 27 May 2016.

  44. 44.

    Bossard M, Feranec J, Otahel J. The revised and supplemented Corine land cover nomenclature. Copenhagen: European environment agency. 2000. Report no.: 38.

  45. 45.

    Bíl M, Andrášik R, Svoboda T, Sedoník J. The KDE+ software: a tool for effective identification and ranking of animal-vehicle collision hotspots along networks. Landsc Ecol. 2015;31:231–7.

    Article  Google Scholar 

  46. 46.

    Bíl M, Andrášik R, Janoška Z. Identification of hazardous road locations of traffic accidents by means of kernel density estimation and cluster significance evaluation. Accid Anal Prev. 2013;55:265–73.

    Article  PubMed  Google Scholar 

  47. 47.

    Fox J, The R. Commander: a basic statistics graphical user interface to R. J Stat Softw. 2005;14:1–42.

    Google Scholar 

  48. 48.

    R Development Core Team. R: A language and environment for statistical computing. Vienna, Austria; 2008. Accessed 21 Sept 2016.

  49. 49.

    Cabela A, Grillitsch H, Tiedemann Franz. Atlas zur Verbreitung und Ökologie der Amphibien und Reptilien in Österreich. Auswertung der Herpetofaunistischen Datenbank der Herpetologischen Sammlung des Naturhistorischen Museums Wien. Wien: Umweltbundesamt. 2001.

  50. 50.

    Trochet A, Moulherat S, Calvez O, Stevens VM, Clobert J, Schmeller DS. A database of life-history traits of European amphibians. Biodivers Data J. 2014. Accessed 1 Apr 2017.

  51. 51.

    Ratton P, Secco H, da Rosa CA. Carcass permanency time and its implications to the roadkill data. Eur J Wildl Res. 2014;60:543–6.

    Google Scholar 

  52. 52.

    Seo C, Thorne JH, Choi T, Kwon H, Park C-H. Disentangling roadkill: the influence of landscape and season on cumulative vertebrate mortality in South Korea. Landsc Ecol Eng. 2015;11:87–99.

    Article  Google Scholar 

  53. 53.

    Cosentino BJ, Marsh DM, Jones KS, Apodaca JJ, Bates C, Beach J, et al. Citizen science reveals widespread negative effects of roads on amphibian distributions. Biol Conserv. 2014;180:31–8.

    Article  Google Scholar 

Download references

Authors’ contributions

FH and JGZ conceived and designed the study; FH, JGZ, KH performed the study; FH, KH and GL analysed the data; FH, JGZ and GL wrote the paper. All authors read and approved the final manuscript.


We want to thank Philipp Hummer and Rainer Holzapfel who developed the website and apps of Project Roadkill ( Comments of two anonymous reviewers helped to improve former versions of this manuscript.

Competing interests

The authors declare that they have no competing interests.

Availability of data and materials

The data set supporting the results of this article is included within the article (and its additional file).


This work was partly funded by the University of Natural Resources and Life Sciences Vienna, Citizen Science Network.

Publisher’s Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Author information



Corresponding author

Correspondence to Florian Heigl.

Rights and permissions

Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (, which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver ( applies to the data made available in this article, unless otherwise stated.

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Heigl, F., Horvath, K., Laaha, G. et al. Amphibian and reptile road-kills on tertiary roads in relation to landscape structure: using a citizen science approach with open-access land cover data. BMC Ecol 17, 24 (2017).

Download citation


  • Anurans
  • Kernel density estimation
  • Landscape ecology
  • Participatory science
  • qgis
  • Road mortality
  • Snakes