Autor(es):
Carvalho, Filipe 
; Mira, António
Data: 2011
Identificador Persistente: http://hdl.handle.net/10174/3409
Origem: Repositório Científico da Universidade de Évora
Assunto(s): Mediterranean farmland; Redundancy analysis; Variance partitioning; Vertebrates; Road kills
Descrição
We surveyed road kills occurring along a 26-km stretch of a major national road (Portugal) in two different years: 1996 and 2005. For analysis purposes, we divided
the data into seven vertebrate groups: amphibians, reptiles,carnivores, prey mammals (shrews, moles, rodents, rabbits
and hares), hedgehogs, owls and passerines. Main factors influencing vertebrate road casualties were evaluated using redundancy analysis and variance partitioning techniques, focusing on three sets of variables: land cover, landscape metrics and spatial location. We also took into account meteorological conditions and changes in traffic intensity specific to each of the surveyed years. The percentage of
variance explained by the explanatory variables was greater in 1996 (67.5%) than in 2005 (48.1%). Many variables
influencing road kill incidence were common to both years.
The most significantly associated factor was the distance to the Natural Park of Serra de São Mamede (NPSSM): road kills decreased steadily as our survey moved south, away
from the NPSSM border. Moreover, an increased incidence of road losses occurred in forested areas, such as montado and traditional olive groves. As 2005 was a climatically drier year, additional variance factors became prominent, including the distance to water reservoirs, suggesting a greater influence of water availability. Traffic flow increased by almost 150% from 1996 to 2005, which may
explain the overall increase in road kills, with the notable exception of the amphibian group, whose road fatalities
incidences decreased approximately sixfold. We expect that our survey will provide a comprehensive understanding of the most critical factors currently influencing vertebrate road fatalities and aid in improving the effectiveness of mitigation measures to reduce them.
