Author(s): Valerio, Francesco ; Godinho, Sergio ; Salgueiro, Pedro ; Medinas, Denis ; Manghi, Giovanni ; Mira, António ; Pedroso, Nuno ; Ferreira, Eduardo ; Craveiro, João ; Costa, Pedro ; Santos, Sara
Date: 2023
Persistent ID: http://hdl.handle.net/10174/34959
Origin: Repositório Científico da Universidade de Évora
Context Road impacts on biodiversity are increasing worldwide. Few attempts have been made to integrate multiple taxonomic groups into roadkill mitigation plans, while using remotely sensed habitat suitability and functional connectivity. Objectives We pinpoint high-risk road locations (road planning units) for 19 woodland species from different taxonomic groups (non-flying mammals, birds, and bats) to enhance prioritisation and versatil- ity of roadkill mitigation plans.Methods In Southern Portugal, we collected species occurrence data, roadkill, and high-resolution satel- lite imageries, along 15 years. We identified remotely sensed habitat metrics, in turn weighted together with functional connectivity models and road metrics to estimate roadkill vulnerability, using random forests. The roadkill cumulative risk across species is then estimated, as well the likelihood variation within and between taxonomic groups to verify prediction consistency. Results Remote sensing information thoroughly explained habitat suitability, identifying similar met- rics within each group, and non-uniform environ- mental tolerance across species. Functional connec- tivity and habitat suitability significantly explained mortality, highlighting connected woodlands and neighbouring matrices. The roadkill cumulative risk endorses a conspicuous prioritisation of road plan- ning units for implementing mitigation structures use- ful for multiple species, with high precision and low probability variation within each group. Some dis- crepancies in prediction consistency still emerge after group comparisons regarding bats. Conclusions We provide novel insights for multi- taxa ecological responses and roadkill evaluations, demonstrating a possible spatial prioritisation in mortality patterns from species with different traits. The identified road units support resilience and mul- tifunctionality over long-term, enabling to assist cost-effective mitigation plans. Findings ultimately offer versatility during the mitigation planning phase throughout the identification of road sub-optimal units, and opportunity costs given their potential for different taxa.
