Predicting spatial and seasonal patterns of wildlife–vehicle collisions in high-risk areas†
Hanh K. D. Nguyen
A B , Matthew W. Fielding B C , Jessie C. Buettel B C and Barry W. Brook B C *
+ Author Affiliations
- Author Affiliations
A School of Technology, Environments and Design, University of Tasmania, Sandy Bay, Tas. 7005, Australia.
B School of Natural Sciences, University of Tasmania, Sandy Bay, Tas. 7005, Australia.
C ARC Centre of Excellence for Australian Biodiversity and Heritage, Hobart, Tas. 7005, Australia.
Wildlife Research 49(5) 428-437 https://doi.org/10.1071/WR21018
Submitted: 18 January 2021 Accepted: 28 October 2021 Published: 25 March 2022
© 2022 The Author(s) (or their employer(s)). Published by CSIRO Publishing.
Abstract
Context: Vehicle collisions with wildlife can injure or kill animals, threaten human safety, and threaten the viability of rare species. This has led to a focus in road-ecology research on identifying the key predictors of ‘road-kill’ risk, with the goal of guiding management to mitigate its impact. However, because of the complex and context-dependent nature of the causes of risk exposure, modelling road-kill data in ways that yield consistent recommendations has proven challenging.
Aim: Here we used a multi-model machine-learning approach to identify the spatio-temporal predictors, such as traffic volume, road shape, surrounding vegetation and distance to human settlements, associated with road-kill risk.
Methods: We collected data on the location, identity and wildlife body size of each road mortality across four seasons along eight roads in southern Tasmania, a ‘road-kill hotspot’ of management concern. We focused on three large-bodied and frequently affected crepuscular Australian marsupial herbivore species, the rufous-bellied pademelon (Thylogale billardierii), Bennett’s wallaby (Macropus rufogriseus) and the bare-nosed wombat (Vombatus ursinus). We fit the point-location data using ‘lasso-regularisation’ of a logistic generalised linear model (LL-GLM) and out-of-bag optimisation of a decision-tree-based ‘random forests’ (RF) algorithm for optimised predictions.
Results: The RF model, with high-level feature interactions, yielded superior out-of-sample prediction results to the linear additive model, with a RF classification accuracy of 84.8% for the 871 road-kill observations and a true skill statistic of 0.708, compared with 61.2% and 0.205 for the LL-GLM. The lasso rejected road visibility and human density, ranking roadside vegetation type and presence of barrier fencing as the most influential predictors of road-kill locality.
Conclusions: Forested areas with no roadside barrier fence along curved sections of road posed the highest risk to animals. Seasonally, the frequency of wildlife–vehicle collisions increased notably for females during oestrus, when they were more dispersive and so had a higher encounter rate with roads.
Implications: These findings illustrate the value of using a combination of attributive and predictive modelling using machine learning to rank and interpret a complexity of possible predictors of road-kill risk, as well as offering a guide to practical management interventions that can mitigate road-related hazards.
Keywords: collision risk, habitat features, lasso regression, predictive modelling, random forest, road-kill trends, seasonal influence, sex bias, wildlife road mortality.
References
Allouche, O, Tsoar, A, and Kadmon, R (2006). Assessing the accuracy of species distribution models: prevalence, kappa and the true skill statistic (TSS). Journal of Applied Ecology 43, 1223–1232.| Assessing the accuracy of species distribution models: prevalence, kappa and the true skill statistic (TSS).Crossref | GoogleScholarGoogle Scholar |
Augustin, T, Kneib, T, Boulesteix, A-L, Strobl, C, and Zeileis, A (2008). Conditional variable importance for random forests. BMC Bioinformatics 9, 307.
| Conditional variable importance for random forests.Crossref | GoogleScholarGoogle Scholar | 18620558PubMed |
Barthelmess, EL (2014). Spatial distribution of road-kills and factors influencing road mortality for mammals in Northern New York State. Biodiversity and Conservation 23, 2491–2514.
| Spatial distribution of road-kills and factors influencing road mortality for mammals in Northern New York State.Crossref | GoogleScholarGoogle Scholar |
Bell, C (2012). Colliding human–animal trajectories (road kill!) on a Tasmanian journey. Critical Arts: A South-North Journal of Cultural & Media Studies 26, 272–289.
| Colliding human–animal trajectories (road kill!) on a Tasmanian journey.Crossref | GoogleScholarGoogle Scholar |
Bond, ARF, and Jones, DN (2014). Roads and macropods: interactions and implications. Australian Mammalogy 36, 1–14.
| Roads and macropods: interactions and implications.Crossref | GoogleScholarGoogle Scholar |
Bourel, M, and Segura, AM (2018). Multiclass classification methods in ecology. Ecological Indicators 85, 1012–1021.
| Multiclass classification methods in ecology.Crossref | GoogleScholarGoogle Scholar |
Breiman, L (2001). Random forests. Machine Learning 45, 5–32.
| Random forests.Crossref | GoogleScholarGoogle Scholar |
Campobasso, CP, Di Vella, G, and Introna, F (2001). Factors affecting decomposition and Diptera colonization. Forensic Science International 120, 18–27.
| Factors affecting decomposition and Diptera colonization.Crossref | GoogleScholarGoogle Scholar | 11457604PubMed |
Curlewis, JD (1989). The breeding-season of Bennett’s wallaby (Macropus rufogriseus rufogriseus) in Tasmania. Journal of Zoology 218, 337–339.
| The breeding-season of Bennett’s wallaby (Macropus rufogriseus rufogriseus) in Tasmania.Crossref | GoogleScholarGoogle Scholar |
Cutler, DR, Edwards Jr, TC, Beard, KH, Cutler, A, Hess, KT, Gibson, J, and Lawler, JJ (2007). Random forests for classification in ecology. Ecology 88, 2783–2792.
| Random forests for classification in ecology.Crossref | GoogleScholarGoogle Scholar | 18051647PubMed |
Diaz-Uriarte, R, and de Andres, SA (2006). Gene selection and classification of microarray data using random forest. BMC Bioinformatics 7, .
| Gene selection and classification of microarray data using random forest.Crossref | GoogleScholarGoogle Scholar | 16398926PubMed |
Driessen MM, Hocking GJ (1992) ‘Review and analysis of Spotlight surveys in Tasmania, 1975–1990.’ (Department of Parks, Wildlife and Heritage Australia)
Efron, B (2020). Prediction, estimation, and attribution. Journal of the American Statistical Association 115, 636–655.
| Prediction, estimation, and attribution.Crossref | GoogleScholarGoogle Scholar |
Ellis, WA, FitzGibbon, SI, Barth, BJ, Niehaus, AC, David, GK, Taylor, BD, Matsushige, H, Melzer, A, Bercovitch, FB, Carrick, F, Jones, DN, Dexter, C, Gillett, A, Predavec, M, Lunney, D, and Wilson, RS (2016). Daylight saving time can decrease the frequency of wildlife-vehicle collisions. Biology Letters 12, 5.
| Daylight saving time can decrease the frequency of wildlife-vehicle collisions.Crossref | GoogleScholarGoogle Scholar |
Elzanowski, A, Ciesiolkiewicz, J, Kaczor, M, Radwanska, J, and Urban, R (2009). Amphibian road mortality in Europe: a meta-analysis with new data from Poland. European Journal of Wildlife Research 55, 33–43.
| Amphibian road mortality in Europe: a meta-analysis with new data from Poland.Crossref | GoogleScholarGoogle Scholar |
Evans, MC, Macgregor, C, and Jarman, PJ (2006). Diet and feeding selectivity of common wombats. Wildlife Research 33, 321–330.
| Diet and feeding selectivity of common wombats.Crossref | GoogleScholarGoogle Scholar |
Fielding, MW, Buettel, JC, Nguyen, H, and Brook, BW (2020). Ravens exploit wildlife roadkill and agricultural landscapes but do not affect songbird assemblages. Emu - Austral Ornithology 120, 11–21.
| Ravens exploit wildlife roadkill and agricultural landscapes but do not affect songbird assemblages.Crossref | GoogleScholarGoogle Scholar |
Fisher, DO, and Lara, MC (1999). Effects of body size and home range on access to mates and paternity in male bridled nailtail wallabies. Animal Behaviour 58, 121–130.
| Effects of body size and home range on access to mates and paternity in male bridled nailtail wallabies.Crossref | GoogleScholarGoogle Scholar | 10413548PubMed |
Forman, RTT, and Alexander, LE (1998). Roads and their major ecological effects. Annual Review of Ecology and Systematics 29, 207–231.
| Roads and their major ecological effects.Crossref | GoogleScholarGoogle Scholar |
Friedman, J, Hastie, T, and Tibshirani, R (2010). Regularization paths for generalized linear models via coordinate descent. Journal of Statistical Software 33, 1–22.
| 20808728PubMed |
Gibbs, JP, and Shriver, WG (2002). Estimating the effects of road mortality on turtle populations. Conservation Biology 16, 1647–1652.
| Estimating the effects of road mortality on turtle populations.Crossref | GoogleScholarGoogle Scholar |
Hastie T, Friedman J, Tibshirani R (2009) ‘The elements of statistical learning: data mining, inference and prediction’, 2nd edn. (Springer: New York, NY, USA)
Hobday, AJ (2010). Nighttime driver detection distances for Tasmanian fauna: informing speed limits to reduce roadkill. Wildlife Research 37, 265–272.
| Nighttime driver detection distances for Tasmanian fauna: informing speed limits to reduce roadkill.Crossref | GoogleScholarGoogle Scholar |
Hothorn, T, Hornik, K, and Zeileis, A (2006). Unbiased recursive partitioning: a conditional inference framework. Journal of Computational and Graphical Statistics 15, 651–674.
| Unbiased recursive partitioning: a conditional inference framework.Crossref | GoogleScholarGoogle Scholar |
James G, Witten D, Hastie T, Tibshirani R (2014) ‘An introduction to statistical learning: with applications in R.’ (Springer Publishing Company)
Johnson, CN (1987). Macropod studies at wallaby creek. IV. Home range and movements of the red-necked wallaby. Australian Wildlife Research 14, 125–132.
| Macropod studies at wallaby creek. IV. Home range and movements of the red-necked wallaby.Crossref | GoogleScholarGoogle Scholar |
Laurance, WF, Clements, GR, Sloan, S, O’Connell, CS, Mueller, ND, Goosem, M, Venter, O, Edwards, DP, Phalan, B, Balmford, A, Van Der Ree, R, and Arrea, IB (2014). A global strategy for road building. Nature 513, 229–232.
| A global strategy for road building.Crossref | GoogleScholarGoogle Scholar | 25162528PubMed |
le Mar, K, and McArthur, C (2005). Comparison of habitat selection by two sympatric macropods, Thylogale billardierii and Macropus rufogriseus rufogriseus, in a patchy eucalypt-forestry environment. Austral Ecology 30, 674–683.
| Comparison of habitat selection by two sympatric macropods, Thylogale billardierii and Macropus rufogriseus rufogriseus, in a patchy eucalypt-forestry environment.Crossref | GoogleScholarGoogle Scholar |
Lee, E, Klocker, U, Croft, DB, and Ramp, D (2004). Kangaroo-vehicle collisions in Australia’s sheep rangelands, during and following drought periods. Australian Mammalogy 26, 215–226.
| Kangaroo-vehicle collisions in Australia’s sheep rangelands, during and following drought periods.Crossref | GoogleScholarGoogle Scholar |
Loss, SR, Will, T, and Marra, PP (2014). Estimation of bird-vehicle collision mortality on US roads. Journal of Wildlife Management 78, 763–771.
| Estimation of bird-vehicle collision mortality on US roads.Crossref | GoogleScholarGoogle Scholar |
Lunney D, Munn AJ, Meikle, W (2008) ‘Too close for comfort: contentious issues in human-wildlife encounters.’ (Royal Zoological Society of New South Wales: Australia)
Mallick, SA, Hocking, GJ, and Driessen, MM (1998). Road-kills of the eastern barred bandicoot (Perameles gunnii) in Tasmania: an index of abundance. Wildlife Research 25, 139–145.
| Road-kills of the eastern barred bandicoot (Perameles gunnii) in Tasmania: an index of abundance.Crossref | GoogleScholarGoogle Scholar |
Nguyen, HKD, Fielding, MW, Buettel, JC, and Brook, BW (2019). Habitat suitability, live abundance and their link to road mortality of Tasmanian wildlife. Wildlife Research 46, 236–246.
| Habitat suitability, live abundance and their link to road mortality of Tasmanian wildlife.Crossref | GoogleScholarGoogle Scholar |
Pagany, R (2020). Wildlife–vehicle collisions: influencing factors, data collection and research methods. Biological Conservation 251, 108758.
| Wildlife–vehicle collisions: influencing factors, data collection and research methods.Crossref | GoogleScholarGoogle Scholar |
Polak, T, Rhodes, JR, Jones, D, and Possingham, HP (2014). Optimal planning for mitigating the impacts of roads on wildlife. Journal of Applied Ecology 51, 726–734.
| Optimal planning for mitigating the impacts of roads on wildlife.Crossref | GoogleScholarGoogle Scholar |
Ramp, D, Caldwell, J, Edwards, KA, Warton, D, and Croft, DB (2005). Modelling of wildlife fatality hotspots along the snowy mountain highway in New South Wales, Australia. Biological Conservation 126, 474–490.
| Modelling of wildlife fatality hotspots along the snowy mountain highway in New South Wales, Australia.Crossref | GoogleScholarGoogle Scholar |
Ramp, D, Wilson, VK, and Croft, DB (2006). Assessing the impacts of roads in peri-urban reserves: road-based fatalities and road usage by wildlife in the Royal National Park, New South Wales, Australia. Biological Conservation 129, 348–359.
| Assessing the impacts of roads in peri-urban reserves: road-based fatalities and road usage by wildlife in the Royal National Park, New South Wales, Australia.Crossref | GoogleScholarGoogle Scholar |
Roger, E, Laffan, SW, and Ramp, D (2007). Habitat selection by the common wombat (Vombatus ursinus) in disturbed environments: implications for the conservation of a ‘common’ species. Biological Conservation 137, 437–449.
| Habitat selection by the common wombat (Vombatus ursinus) in disturbed environments: implications for the conservation of a ‘common’ species.Crossref | GoogleScholarGoogle Scholar |
Roger, E, Laffan, SW, and Ramp, D (2011). Road impacts a tipping point for wildlife populations in threatened landscapes. Population Ecology 53, 215–227.
| Road impacts a tipping point for wildlife populations in threatened landscapes.Crossref | GoogleScholarGoogle Scholar |
Rorden, C, Karnath, HO, and Bonilha, L (2007). Improving lesion-symptom mapping. Journal of Cognitive Neuroscience 19, 1081–1088.
| Improving lesion-symptom mapping.Crossref | GoogleScholarGoogle Scholar | 17583985PubMed |
Rubenstein D, Wrangham R (1986) ‘Ecological aspects of social evolution: birds and mammals.’ (Princeton University Press: USA)
Russell, TC, Herbert, CA, and Kohen, JL (2009). High possum mortality on urban roads: implications for the population viability of the common brushtail and the common ringtail possum. Australian Journal of Zoology 57, 391–397.
| High possum mortality on urban roads: implications for the population viability of the common brushtail and the common ringtail possum.Crossref | GoogleScholarGoogle Scholar |
Rytwinski, T, and Fahrig, L (2012). Do species life history traits explain population responses to roads? A meta-analysis. Biological Conservation 147, 87–98.
| Do species life history traits explain population responses to roads? A meta-analysis.Crossref | GoogleScholarGoogle Scholar |
Simpson, K, Johnson, CN, and Carver, S (2016). Sarcoptes scabiei: the mange mite with mighty effects on the common wombat (Vombatus ursinus). PLoS One 11, e0149749.
| Sarcoptes scabiei: the mange mite with mighty effects on the common wombat (Vombatus ursinus).Crossref | GoogleScholarGoogle Scholar | 26943790PubMed |
Skorka, P (2016). The detectability and persistence of road-killed butterflies: an experimental study. Biological Conservation 200, 36–43.
| The detectability and persistence of road-killed butterflies: an experimental study.Crossref | GoogleScholarGoogle Scholar |
Strobl, C, Boulesteix, A-L, Zeileis, A, and Hothorn, T (2007). Bias in random forest variable importance measures: illustrations, sources and a solution. BMC Bioinformatics 8, 25.
| Bias in random forest variable importance measures: illustrations, sources and a solution.Crossref | GoogleScholarGoogle Scholar | 17254353PubMed |
Suissa, S, and Shuster, JJ (1985). Exact unconditional sample sizes for the 2 × 2 binomial trial. Journal of the Royal Statistical Society. Series A (General) 148, 317–327.
| Exact unconditional sample sizes for the 2 × 2 binomial trial.Crossref | GoogleScholarGoogle Scholar |
Svetnik, V, Liaw, A, Tong, C, Culberson, JC, Sheridan, RP, and Feuston, BP (2003). Random forest: a classification and regression tool for compound classification and QSAR modeling. Journal of Chemical Information and Computer Sciences 43, 1947–1958.
| Random forest: a classification and regression tool for compound classification and QSAR modeling.Crossref | GoogleScholarGoogle Scholar | 14632445PubMed |
Tejera, G, Rodriguez, B, Armas, C, and Rodriguez, A (2018). Wildlife-vehicle collisions in Lanzarote Biosphere Reserve, Canary Islands. PLoS One 13, e0192731.
| Wildlife-vehicle collisions in Lanzarote Biosphere Reserve, Canary Islands.Crossref | GoogleScholarGoogle Scholar | 29561864PubMed |
Thuiller, W, Araújo, MB, and Lavorel, S (2009). Generalized models vs. classification tree analysis: predicting spatial distributions of plant species at different scales. Journal of Vegetation Science 14, 669–680.
| Generalized models vs. classification tree analysis: predicting spatial distributions of plant species at different scales.Crossref | GoogleScholarGoogle Scholar |
Torsten, H, Kurt, H, and Achim, Z (2006). Unbiased recursive partitioning: a conditional inference framework. Journal of Computational and Graphical Statistics , 651.
| Unbiased recursive partitioning: a conditional inference framework.Crossref | GoogleScholarGoogle Scholar |
Tredennick, AT, Hooker, G, Ellner, SP, and Adler, PB (2021). A practical guide to selecting models for exploration, inference, and prediction in ecology. Ecology 102, .
| A practical guide to selecting models for exploration, inference, and prediction in ecology.Crossref | GoogleScholarGoogle Scholar | 33710619PubMed |
Trombulak, SC, and Frissell, CA (2000). Review of ecological effects of roads on terrestrial and aquatic communities. Conservation Biology 14, 18–30.
Valero, E, Picos, J, and Alvarez, X (2015). Road and traffic factors correlated to wildlife–vehicle collisions in Galicia (Spain). Wildlife Research 42, 25–34.
| Road and traffic factors correlated to wildlife–vehicle collisions in Galicia (Spain).Crossref | GoogleScholarGoogle Scholar |
Visintin, C, van der Ree, R, and McCarthy, MA (2016). A simple framework for a complex problem? Predicting wildlife-vehicle collisions. Ecology and Evolution 6, 6409–6421.
| A simple framework for a complex problem? Predicting wildlife-vehicle collisions.Crossref | GoogleScholarGoogle Scholar | 27648252PubMed |
Wiggins, NL, and Bowman, D (2011). Macropod habitat use and response to management interventions in an agricultural–forest mosaic in north-eastern Tasmania as inferred by scat surveys. Wildlife Research 38, 103–113.
| Macropod habitat use and response to management interventions in an agricultural–forest mosaic in north-eastern Tasmania as inferred by scat surveys.Crossref | GoogleScholarGoogle Scholar |
Wright, MN, Ziegler, A, and Konig, IR (2016). Do little interactions get lost in dark random forests? BMC Bioinformatics 17, 10.
| Do little interactions get lost in dark random forests?Crossref | GoogleScholarGoogle Scholar |
