Predicting roadkill: a regional-scale model for the Northern Tamandua integrating environmental and road characteristics
Wildlife-vehicle collisions may be the most visible impact of road networks on ecosystems. It has been shown that roadkill does not randomly occur across space and is distributed depending on environmental and ecological factors, such as habitat suitability and landscape connectivity of the interest...
Saved in:
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Pensoft Publishers
2025-02-01
|
| Series: | Neotropical Biology and Conservation |
| Online Access: | https://neotropical.pensoft.net/article/124522/download/pdf/ |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1823856453832671232 |
|---|---|
| author | Silvio Boyat Esmeralda Arévalo-Huezo |
| author_facet | Silvio Boyat Esmeralda Arévalo-Huezo |
| author_sort | Silvio Boyat |
| collection | DOAJ |
| description | Wildlife-vehicle collisions may be the most visible impact of road networks on ecosystems. It has been shown that roadkill does not randomly occur across space and is distributed depending on environmental and ecological factors, such as habitat suitability and landscape connectivity of the interest species. In Costa Rica, the Northern Tamandua (Tamandua mexicana) is one of the mammal species most affected by road mortality. This study aimed to predict roadkill risk for the species in northwestern Costa Rica, considering habitat suitability and landscape connectivity. Roadkill data was retrieved from a local citizen science project and collected through a field survey. Habitat suitability and landscape connectivity models were calibrated on publicly available presence data. The models were then used as input, including additional road characteristics to calibrate a maximum entropy roadkill risk model. The final model had an excellent predictive performance (AUC = 0.989) and identified 108.7 km of road sections throughout the region as high risk, mainly found on primary roads. The most significant variables in the model were road width, traffic speed and habitat suitability. Landscape connectivity did not contribute significantly to the model. This study shows that road characteristics are the main contributors to roadkill regionally and highlights the relevance of the species’ ecological features, such as habitat suitability, in its prediction. Thus, such variables should be considered in the design of mitigation measures for the impacts of roads on wildlife. |
| format | Article |
| id | doaj-art-c420d175d4934abcace0b6266b570c42 |
| institution | Kabale University |
| issn | 2236-3777 |
| language | English |
| publishDate | 2025-02-01 |
| publisher | Pensoft Publishers |
| record_format | Article |
| series | Neotropical Biology and Conservation |
| spelling | doaj-art-c420d175d4934abcace0b6266b570c422025-02-12T08:31:49ZengPensoft PublishersNeotropical Biology and Conservation2236-37772025-02-01201213610.3897/neotropical.20.e124522124522Predicting roadkill: a regional-scale model for the Northern Tamandua integrating environmental and road characteristicsSilvio Boyat0Esmeralda Arévalo-Huezo1Universidad Latina de Costa RicaUniversidad Latina de Costa RicaWildlife-vehicle collisions may be the most visible impact of road networks on ecosystems. It has been shown that roadkill does not randomly occur across space and is distributed depending on environmental and ecological factors, such as habitat suitability and landscape connectivity of the interest species. In Costa Rica, the Northern Tamandua (Tamandua mexicana) is one of the mammal species most affected by road mortality. This study aimed to predict roadkill risk for the species in northwestern Costa Rica, considering habitat suitability and landscape connectivity. Roadkill data was retrieved from a local citizen science project and collected through a field survey. Habitat suitability and landscape connectivity models were calibrated on publicly available presence data. The models were then used as input, including additional road characteristics to calibrate a maximum entropy roadkill risk model. The final model had an excellent predictive performance (AUC = 0.989) and identified 108.7 km of road sections throughout the region as high risk, mainly found on primary roads. The most significant variables in the model were road width, traffic speed and habitat suitability. Landscape connectivity did not contribute significantly to the model. This study shows that road characteristics are the main contributors to roadkill regionally and highlights the relevance of the species’ ecological features, such as habitat suitability, in its prediction. Thus, such variables should be considered in the design of mitigation measures for the impacts of roads on wildlife.https://neotropical.pensoft.net/article/124522/download/pdf/ |
| spellingShingle | Silvio Boyat Esmeralda Arévalo-Huezo Predicting roadkill: a regional-scale model for the Northern Tamandua integrating environmental and road characteristics Neotropical Biology and Conservation |
| title | Predicting roadkill: a regional-scale model for the Northern Tamandua integrating environmental and road characteristics |
| title_full | Predicting roadkill: a regional-scale model for the Northern Tamandua integrating environmental and road characteristics |
| title_fullStr | Predicting roadkill: a regional-scale model for the Northern Tamandua integrating environmental and road characteristics |
| title_full_unstemmed | Predicting roadkill: a regional-scale model for the Northern Tamandua integrating environmental and road characteristics |
| title_short | Predicting roadkill: a regional-scale model for the Northern Tamandua integrating environmental and road characteristics |
| title_sort | predicting roadkill a regional scale model for the northern tamandua integrating environmental and road characteristics |
| url | https://neotropical.pensoft.net/article/124522/download/pdf/ |
| work_keys_str_mv | AT silvioboyat predictingroadkillaregionalscalemodelforthenortherntamanduaintegratingenvironmentalandroadcharacteristics AT esmeraldaarevalohuezo predictingroadkillaregionalscalemodelforthenortherntamanduaintegratingenvironmentalandroadcharacteristics |