Exploring hybrid consensus models to assess roadkill

Karanasios, Panagiotis; Wunderlich, Rainer Ferdinand; Mukhtar, Hussnain; Chiu, Hao-Wei; Lin, Yu‐Pin

doi:10.1016/j.jenvman.2021.112886

Cited by 3 publications

(4 citation statements)

References 82 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Our ndings contrast with those of Karanasios et al (2021) who found that the accuracy of KDEMs for herpetofauna roadkill was slightly better (outstanding; AUC > 0.9) than that of MaxEnt models (excellent; AUC 0.8-0.89). Our roadkill data was irregularly distributed and sparse over large areas, and roadkill KDEMs may perform better for denser distributions like those in the four road segments studied by Karanasios et al (2021). Among the KDEMs for roadkill of different herpetofauna taxa, the AUC was highest for Gulf Coast Toad, likely because it had a more clumped distribution (Fig.…”

Section: Discussioncontrasting

confidence: 99%

“…Variables associated with human population density were highly ranked across MaxEnt roadkill models for all herpetofauna taxa we studied. Although some recent studies have obtained excellent performance of MaxEnt herpetofauna roadkill models without anthropogenic variables (e.g., Karanasios et al 2021), the results of this study support inclusion of human development variables in correlative herpetofauna roadkill SDMs. In broad strokes, we found that areas with open space to medium development, low to moderate population density, and low to moderate road density are all areas with the highest probabilities for herpetofauna roadkill.…”

Section: Maxent Roadkill Models Variable Importancesupporting

confidence: 44%

“…The expansion of road corridor networks has been associated with increased faunal mortality (roadkill) for many types of animals; however, herpetofauna are among the most vulnerable taxa (Jochimsen et al 2004, Glista et al 2008, Colino-Rabanal and Lizana 2012, Cosentino et al and inter-population gravity models (Beaudry et al 2008). Various correlative species distribution models (SDM) have been used for mapping roadkill probability, such as logistic regression (Visintin et al 2016), MaxEnt (Ha and Shilling 2018, Kantola et al 2019a, Yue et al 2019, Chyn et al 2021), a combination of SDM methods (Fabrizio et al 2019, and kernel density estimate models (KDEMs) (Karanasios et al 2021). Regional herpetofauna roadkill correlative SDMs have been developed using MaxEnt in Australia (Santori et al 2018), Taiwan (Yue et al 2019, Chyn et al 2021), and Slovenia (Sillero et al 2019), and using multi-algorithm ensemble SDMs in Taiwan (Lin et al 2019), but never within North America.…”

Section: Introductionmentioning

confidence: 99%

“…It is important to compare different roadkill modeling approaches for their accuracy and usefulness in facilitating regional conservation planning efforts across various scales. Karanasios et al (2021) reported the only comparison between roadkill MaxEnt models and KDEMs of which we are aware, and they found accuracy was higher for KDEMs compared to MaxEnt models in a localized herpetofauna roadkill study involving four road segments in Taiwan.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Where the toad crosses the road: Multi-method Texas herpetofauna roadkill modeling for conservation planning

Chyn¹,

Tracy

Wright

et al. 2023

Preprint

View full text Add to dashboard Cite

Road mortality is a significant contributor to the global decline in herpetofauna, making regional roadkill models important for conservation planning. We used the iNaturalist Roadkills of Texas community science volunteered geographic information database to develop the first regional herpetofauna roadkill models in North America for: (1) all Texas herpetofauna combined; (2) Texas Horned Lizards (Phrynosoma cornutum), a Texas state threatened species; and (3) Gulf Coast Toads (Incilius nebulifer), a common generalist species. We novelly compare the utility of extrapolative feature-selected MaxEnt models to interpolative kernel density estimate models (KDEMs) of roadkill to inform conservation planning for sensitive Texas herpetofauna. A random subset feature selection algorithm was used to identify subsets of five to seven of 55 variables optimizing Maxent roadkill model performance and for ranking variables. MaxEnt roadkill models had acceptable to excellent accuracy, outperforming KDEMs for Texas herpetofauna combined and the Texas Horned Lizard. Both model types had poor accuracy in projecting Gulf Coast Toads roadkill, however, KDEMs had much lower complexity. Roadkill models of both types for various taxa revealed roadkill risk areas for partially sympatric sensitive species having insufficient data for individual models, including the Texas Tortoise (Gopherus berlandieri) and Houston Toad (Anaxyrus houstonensis). Top ranked variables in feature-selected MaxEnt roadkill models included distance to residential roads, human population density, percent cover developed open space, and spatial density of non-roadkill herpetofauna occurrences. Identified high roadkill risk areas should be further surveyed to refine roadkill projections and inform mitigation planning to support sensitive Texas herpetofauna conservation efforts.

show abstract

Section: Discussioncontrasting

confidence: 99%

Section: Maxent Roadkill Models Variable Importancesupporting

confidence: 44%