Wetlands Ecol Manage (2015) 23:227–239

DOI 10.1007/s11273-014-9372-4

ORIGINAL PAPER

Spatiotemporal variation in human-wildlife conflicts

along highway BR-262 in the Brazilian Pantanal
Julio Cesar de Souza • Vanessa Pereira da Cunha •

Scott H. Markwith

Received: 14 August 2013 / Accepted: 24 July 2014 / Published online: 2 August 2014
Ó Springer Science+Business Media Dordrecht 2014

Abstract The Pantanal of Brazil, the world’s largest one of the highest in Brazil. IUCN Red List species
freshwater wetland, supports a large diversity of were observed, including lowland tapir, giant anteater,
species and is under continually increasing pressure and marsh deer. At the lowest elevations, during
from human development, including vehicle collisions seasonal wetland inundation the causeway embank-
with wild animals. We examined decadal, annual, and ment is sought for refuge, increasing collision proba-
spatial variation in wildlife-vehicle collisions along bility. Mitigation measures tied to seasonal flood pulse
highway BR-262 in the southern Pantanal, and specif- variation along the causeway may reduce the high rate
ically: (1) what is the rate of roadkill along BR-262, of wild animal mortality and loss of biodiversity.
and did it increased over the previous decade, (2) which
species are frequently victims of collisions and does Keywords Wildlife-vehicle collisions  Roadkill 
this vary annually along the highway’s length, and (3) Mitigation  Threatened species
what anthropogenic or environmental factors may
influence this variation and how can the context of
collisions inform mitigation? We sampled the highway
monthly, between April 2011 and June 2012, stopping Introduction
when roadkill was observed to identify species and
collect GPS coordinates and a photographic record. The Pantanal of west-central Brazil is considered the
Data was analyzed temporally and in relation to traffic world’s largest freshwater wetland, encompassing an
volume and precipitation, and imported into a GIS and area of *140,000 km2 (Potts and Potts 2004). The
analyzed spatially and in relation to elevation. We physical and biological processes in the Pantanal are
recorded 518 carcasses from 40 species over the 15 influenced by the hydrological regime (e.g.Junk and
sampling periods. The rate of roadkill incidents was 1 da Silva 2000), which is characterized by seasonally
every 6.2 km, a tenfold increase since 2002, and is now alternating drought and flood. The sometimes subtle
topographic variation combined with the flooding
regimes create a diversity of habitats that support
J. C. de Souza  V. P. da Cunha
Biological Science, Universidade Federal de Mato Grosso approximately 263 species of fish, 41 amphibians, 113
do Sul, Aquidauana, Mato Grosso do Sul, Brazil reptiles, 582 birds and 170 mammals (MMA (Min-
istério do Meio Ambiente) 2007; Alho et al. 2011;
S. H. Markwith (&)
Nunes 2011; WWF—Brasil 2012; Souza and Lima,
Department of Geosciences, Florida Atlantic University,
Boca Raton, FL 33444, USA 2012). Although it supports less diversity than the
e-mail: [email protected] Amazon and Atlantic Forest, the region has the highest

123
228 Wetlands Ecol Manage (2015) 23:227–239

Fig. 1 Map of the study

area along the stretch of
highway BR-262 between
the city of Anastácio, at the
beginning of segment 0, and
the Paraguay River Bridge,
at the end of segment 210.
Shading is based on
elevation, with elevations
\100 m amsl shaded in
white, and increasing in
50 m increments. Inset of
South America indicates the
area of interest (AOI) for
this research

density of fauna in the Americas. However, anthro- including improvement of a stretch with 4 m tall
pogenic threats to Pantanal fauna are numerous, embankments and numerous bridges, the longest being
including hunting; habitat loss due to deforestation, 1850 m on the Paraguay River. Highway BR-262 is the
conversion to agriculture or ranching, and timber only highway that crosses the Pantanal east and west,
plantations; competition and disease transmission and links the state capital, Campo Grande, and other
from domesticated animals; the pet trade; and mortal- Brazilian states internationally with Bolivia and other
ity from collisions with motor vehicles (Courtenay and MERCOSUL countries (Fischer 1999).
Maffei 2008; Naveda et al. 2008; Miranda and Medri Road networks are essential to supply chains, since
2010; Miranda 2011; Keuroghlian et al. 2013). Several they offer the possibility of rapid collection and delivery
species of conservation interest such as pampas deer of goods to distant regions (CNT Confederação Nac-
(Ozotoceros bezoarticus), marsh deer (Blastocerus ional do Transporte 2006), while also promoting
dichotomus), giant river otter (Pteronura brasiliensis), integration of regions, states, and countries. Although
lowland tapir (Tapirus terrestris), giant anteater (Myr- highways are fundamental for personal ease of move-
mecophaga tridactyla), white-lipped peccary (Tay- ment, commerce, and the growth of cities (Martinelli
assu pecari), and jaguar (Phanthera onca) are found and Volpi 2011); they are one of the most visible and
within its borders. The Pantanal is also the most bird- constant anthropogenic impacts on local environments
rich wetland worldwide, and 117 of its bird species are and fauna (Prado et al. 2006; Rocha 2005). The impacts
included on at least one endangered species list at the include deforestation, fragmentation, migration and
state, national, or international levels. Among them, gene flow barriers, increased edge effects, and mortality
the best known species is the hyacinth macaw (Maehr et al. 1991; Trombulak and Frissell 2000;
(Anodorhynchus hyacinthinus) (Harris et al. 2005). Fischer et al. 2004; Bandeira and Floriano 2004; Rocha
Prior to the mid-1980s the means of transport from 2005; CNT (Confederação Nacional do Transporte)
Miranda to Corumbá, through the Pantanal, was limited 2006; Freitas 2009; Sampaio and Brito 2009).
to navigating the Paraguay River and its tributaries, and Vehicle collisions increase the death rate of many
the Estrada Ferro Noroeste do Brasil railroad. This was species and can cause financial loss, injury, or death for
largely because the existing road between the two cities, drivers as well (Conover et al. 1995; Trombulak and
what eventually became BR-262 (Fig. 1), was unpaved Frissell 2000; Prado et al. 2006). Lalo (1987) estimated
and submerged when the Pantanal lowlands flooded that as many as 1 million vertebrates were killed every
during the yearly wet season. The Mato Grosso do Sul day in the United States, potentially an underestimate
state government decided to pave the road in 1984, on today’s higher volume roadways. Threatened and

123
Wetlands Ecol Manage (2015) 23:227–239 229

endangered species are especially vulnerable where crossroads at the edge of the city of Anastácio (S 20°
roadkill rates exceed the reproduction rates of small 290 57.200 , W 55° 480 50.200 ) and the Paraguay River
populations (Evink 1996; Forman 1995; Forman et al. bridge (S 19° 310 14.400 , W 57° 250 13.100 ) (Fig. 1). The
1997; Forman and Alexander 1998; Glista et al. 2008). highway cuts east and west across the Miranda and
Daily mortality of large numbers of individuals of Aquidauana regions of the southern Pantanal in the
various species is a major cause of biodiversity loss in state of Mato Grosso do Sul. Economically, the
wild fauna (Rosa and Mauhs 2004). Miranda and Aquidauana regions are dominated by
Concomitant with the development of infrastructure cattle ranching, and urban development is only found
in the Pantanal and burgeoning economic success of in the transition zones between the wetland and
Brazil, wildlife mortality along its roads has increased. surrounding uplands, i.e. the cities of Miranda and
Fischer et al. (2004) reported an eightfold increase in Anastácio/Aquidauana. The highway exits the Cerra-
roadkill incidents between 1992 and 2002 on federal do vegetation dominated upland plateau to the east and
roads in the Pantanal. Even with these dramatic south of the Pantanal, crosses the alluvial deposits of
increases and well publicized deaths of iconic species the transition zone (also historically dominated by
such as jaguar along BR-262, the Brazilian government Cerrado vegetation), into the Pantanal floodplain
plans further road development within the ecosystem wetlands dominated by floodplain forests, savannahs,
and is yet to take action to mitigate these incidents even and emergent herbaceous vegetation. There are few
though monthly traffic along BR-262 more than roads in the country that cross regions as rich in flora
doubled between 2002 and 2012 (SEOP (Secretaria and fauna (MMA (Ministério do Meio Ambiente)
de Estado de Obras Públicas e de Transportes) 2012). 2007; Rodrigues et al. 2002).
The objectives of the research reported herein were to The city of Anastácio (192 m above mean sea level
examine decadal, annual, and spatial variation in (amsl)), at the eastern terminus of the sampled stretch,
wildlife-vehicle collisions along highway BR-262 is situated *10 km from the base of the escarpment
between the city of Anastácio and the Paraguay River separating the Cerrado region from the Pantanal basin.
Bridge in Mato Grosso do Sul state in the southern The highest elevation along the sampled highway
Pantanal. The specific research questions included: (1) stretch (260 m amsl) is found 35 km west of Anastá-
What is the rate of roadkill along BR-262, and did it cio, halfway to the city of Miranda. The highway
increase over estimates made by Fischer et al. (2004) a descends quickly to the Miranda River valley, and
decade previous? (2) Which species are frequently then gradually descends through the transition zone,
victims of collisions and does this vary annually along *150 km west of Anastácio, until entering the
the highway’s length? (3) What anthropogenic or Pantanal floodplain wetlands (*85–100 m amsl).
environmental factors may influence this variation in This western section of highway passing through the
mortality and how can the context of collisions inform floodplain wetland is on an elevated solid embanked
mitigation? This research has applications in wetland causeway 4 m above the surrounding wetland, with
regions with seasonally varying hydrologic and faunal occasional short bridges with *5 m wide openings in
interactions and that are undergoing continual or the embankment, at *1 km intervals. The vast
increasing development pressures. In such locations, majority of the length of the highway is a paved two
mitigation measures that protect citizens, their property, lane road, i.e. one lane in each direction, with no
and the vulnerable wildlife living in proximity to median and narrow maintained shoulder.
current and future highways may benefit from design
based on the ecological and anthropogenic contexts. Data collection and analysis

Wildlife roadkill incidents were collected once

Methods monthly between April 2011 and June 2012, a total
of 15 sampling events. The sampled stretch of
Study area highway was driven round trip, beginning at 5:00 am
and typically ending at 5:00 pm, to identify carcasses
This research was conducted along a 215 km long before they were destroyed by the intense vehicular
stretch of highway BR-262 between the highway traffic or removed by scavengers or humans. Where a

123
230 Wetlands Ecol Manage (2015) 23:227–239

carcass was observed, the car was pulled to the final segment was only 5 km in length, so the number
roadside to note the location in latitude and longitude of roadkill incidents was converted to incidents/km for
with GPS (Global Position System), the species was each segment to standardize comparison.
identified where possible, and a photographic record A digital elevation model (DEM) at 90 m resolution
created. was downloaded from the USGS HydroSHEDS web-
We travelled at a speed of 50 km/h to facilitate site (http://hydrosheds.cr.usgs.gov/index.php, acces-
visual observation, with 1 driver and a minimum of sed 7 December, 2012). GPS points were joined with
two dedicated observers. Only carcasses observed in the DEM raster coverage and elevations extracted for
the roadway or the maintained shoulder were counted. each recorded roadkill observation. Mean elevation per
This method likely resulted in an undercount of all 10 km highway segment was calculated, and Pearson’s
mortalities along the roadway. Reasons for a suspected correlation coefficient was used to examine the rela-
undercount include carcass removal by scavengers or tionship between mean elevation and the number of
humans, small animals run over repeatedly, animals roadkill incidents/km within each 10 km segment.
thrown wide of the maintained shoulder, or wildlife Repeated measures Analysis of Variance (ANOVA)
still mobile after collision that later perished away was used to test for significant differences within
from the roadway. The latter two reasons are partic- subjects based on sampling month, between subjects
ularly applicable in areas where forest patches grow grouped by highway zones (i.e. upland ([125 m amsl),
directly up to the maintained shoulder and where the transition zone (100–125 m amsl), wetland proper
highway is elevated and the slopes of the causeway (85–100 m amsl), and construction zone), and inter-
could not be easily viewed from a passing automobile. actions with sampling month and highway zone. This
Descriptive statistics were calculated in order to test was conducted on the number of roadkill incidents/
assess the impact of BR-262 on fauna. The total km for each highway segment each month. Due to a
number of roadkill incidents, estimated number of significant Mauchly’s test for sphericity, we used the
incidents/year (estimated incidents/year = total # of Greenhouse-Geisser correction to the F-ratio to reduce
roadkills/# of sampling events 9 365), total number the potential for Type I error. Pairwise comparisons
of roadkill incidents/sampling event, and kilometers among all months and all highway zones for the
between incidents were calculated. Pearson’s correla- estimated marginal means of the dependent variable
tion coefficient was used to examine the relationships were conducted with the Bonferroni adjustment for
between monthly traffic volume passing through the multiple tests. Repeated measures ANOVA was per-
toll plaza on the Paraguay River Bridge (SEOP formed using SPSS release 19 (IBM Corp 2010).
(Secretaria de Estado de Obras Públicas e de Trans-
portes) 2012), monthly rainfall (mm) measured at
Fazenda Guaicurus (halfway between Anastácio/ Results
Aquidauana and the Paraguay River bridge, S 20° 60
5600 , W 56° 470 46.0800 ), and the total number of There were 518 roadkill incidents total recorded
roadkills recorded at each monthly sampling event. during the 15 months of sampling. We measured a
We also calculated the total and monthly mortality rate of 34.5 roadkills per sampling event, and one
events for individual species. carcass occurred every 6.2 km on average. The
The GPS coordinates for each roadkill incident estimated number of roadkill incidents per year along
were imported into ArcMap 10 (ESRI 2011) and this 215 km stretch of highway BR-262 was 12,604.
visualized on a roads basemap that included BR-262. A total of 40 species were observed as roadkill, of
Due to occasional poor satellite coverage at various these were 21 mammal, 12 reptile, and 7 bird species
times during sampling, 10 % of the collected coordi- (Tables 1, 2). Mammals represented 61 % of all
nates were[100 m from the digitized road, and these roadkill incidents (n = 312), birds 19 % (n = 96),
points were excluded from spatial analysis (reducing and reptiles 13 % (n = 67), and the remaining indi-
the sample size from 518 to 456 roadkills for the viduals could not be positively identified due to being
spatial analysis only). The sampled stretch of highway repeatedly run over (n = 43). The species most
was divided into 10 km long segments, numbered commonly involved in wildlife-vehicle collisions
from the eastern terminus, Anastácio, to west. The were the capybara (Hydrochaeris hydrochaeris) with

123
Wetlands Ecol Manage (2015) 23:227–239 231

Table 1 Scientific names, Scientific name Family Frequency IUCN status

families, roadkill frequency
of the entire study period, Mammals
and species status on the
Blastocerus dichotomus Cervidae 6 VU
IUCN red list of threatened
species for mammals Cebus apella Cebidae 4 LC
Cerdocyon thous Canidae 64 LC
Dasyprocta azarae Dasyproctidae 1 Data deficient
Dasypus novemcinctus Dasypodidae 17 LC
Didelphis albiventris Didelphidae 2 LC
Eira barbara Mustelidae 1 LC
Euphractus sexcintus Dasypodidae 18 Not assessed
Galictis cuja Mustelidae 1 LC
Hydrochaeris hydrochaeris Hydrochaeridae 89 LC
Leopardus pardalis Felidae 4 LC
Lontra longicaudis Mustelidae 1 Data deficient
Mazama spp. Cervidae 5 VU/LC/data deficient
Myrmecophaga tridactyla Myrmecophagidae 19 VU
Nasua nasua Procyonidae 8 LC
Procyon cancrivorus Procyonidae 9 LC
Sylvilagus brasiliensis Leporidae 2 LC
Tamandua tetradactyla Myrmecophagidae 33 LC
Tapirus terrestris Tapiridae 4 VU
Tayassu pecari Tayassuidae 20 VU
IUCN status LC least Tayassu tajacu Tayassuidae 4 Not assessed
concern, VU vulnerable

25 more roadkill incidents than the next most frequent, the interaction effects of month*highway zone was also
the common fox (Cerdocyon thous), followed closely significant (F-ratio = 1.921, p value = 0.024). Pair-
by the caiman (Caiman yacare) (Fig. 2). Two scav- wise comparisons among months of the estimated
engers, southern crested caracara (Caracara plancus) marginal means of incidents/km, applying the Bonfer-
and black vulture (Coragyps atratus), were also roni correction, indicated that the months of February
common victims of collisions, apparently when feed- and March 2012 were significantly different than most
ing on roadkill. Of the species of greatest conservation other sampling months (Table 3), while April 2011 was
interest in the region, the white-lipped peccary and only significantly different than October 2011. The
giant anteater were the only listed species with [10 number of incidents/month showed a significant and
recorded incidents. However, two other species listed strong positive correlation to monthly rainfall measured
as vulnerable on the IUCN Red List (IUCN 2013), the at the approximate mid-point of the sampled stretch of
marsh deer and lowland tapir, were also recorded with highway (r = 0.72, p value = 0.001, Fig. 3).
6 and 4 mortality events, respectively. The temporal pattern observed in Fig. 3 is also
The temporal pattern of mortality was variable, but consistent for all of the most common roadkill species.
demonstrated a potentially cyclical pattern coincident All of the species in Fig. 2 had peak mortality in April
with wet and dry seasons (Fig. 3). Roadkill numbers of 2011, February or March of 2012, or some combi-
peaked substantially in April 2011 and March 2012, nation of these sampling events. In general, the
with a relatively low constant background number of background mortality rate per sampling event was
incidents from June through January. Repeated mea- between 1 and 4 individuals per common species, while
sures ANOVA within subjects, applying the Green- species peaks in incidents ranged as high as 12 caiman,
house-Geisser correction, indicated that month of 14 6-banded armadillos (Euphractus sexcintus), and 16
sampling was a significant effect on the number of capybara during the April 2011 sampling event. The
incidents/km (F-ratio = 8.735, p value \ 0.001), and number of incidents/month showed only a weak to

123
232 Wetlands Ecol Manage (2015) 23:227–239

Table 2 Scientific names, families, roadkill frequency of the trend was centered 170–190 km from Anastácio, where
entire study period, and species status on the IUCN red list of highway shoulder improvements were ongoing during
threatened species for birds, reptiles, and unknown roadkill
the sampling period and speeds were reduced through
Scientific name Family Frequency IUCN the construction zone. Mean elevation showed a
status relatively rapid decrease between 50 and 80 km from
Birds Anastácio, then decreased at a relatively low rate
Amazona aestiva Psittacidae 1 LC through the rest of the study stretch to the Paraguay
Aratinga aurea Psittacidae 1 LC River Bridge. The relationship between roadkill inci-
Asio clamator Strigidae 1 Not
dents/km and mean elevation per highway segment was
assessed moderately negative and significant (r = -0.39,
Buteo Accipitridae 1 LC p value = 0.04). The repeated measures ANOVA
magnirostris between subjects test was significant for the highway
Caracara plancus Falconidae 47 LC zone effect (F-ratio = 11.553, p \ 0.001). Pairwise
Caraiama cristata Cariamidae 3 Not comparisons among highway zones of the estimated
assessed marginal means of incidents/km, applying the Bonfer-
Cathartes aura Cathartidae 11 LC roni correction, indicated that the upland and construc-
Ceryle torquata Alcedinidae 1 Not tion zones were not significantly different from each
assessed
other, nor were the transition zone and wetland proper
Columba picazuro Columbidae 2 Not significantly different (Table 4). However, both the
assessed
upland/construction had significantly fewer kills (not
Coragyps atratus Cathartidae 26 LC
just different) than transition/wetland proper.
Nyctidromus Caprimulgidae 1 Not
albicolis assessed On an individual species level, the distribution of
Tyrannus Tyrannidae 1 LC
incidents along the highway varied with distance from
melancholicus Anastácio. Some species showed a relatively even
Reptiles distribution of incidents along the sample stretch of
Caiman yacare Alligatoridae 60 LC BR-262, i.e. common fox (in 19 of 22 segments), black
Crotalus durissus Viperidae 1 LC vulture (Coragyps atratus), turkey vulture (Cathartes
Drymarchon Colubridae 1 LC aura), southern crested caracara, giant anteater,
corais 6-banded armadillo, and capybara. None of the species
Hydrodynastes Colubridae 1 Not were exclusively found east of Miranda and at the
gigas assessed highest elevations along the highway, i.e. segments
Iguanae iguanae Iguanidae 2 Not 0–60. However, the distributions of roadkill of both
assessed 9-banded armadillo (Dasypus novemcinctus) and
Philodryas Colubridae 1 Not lesser anteater (Tamandua tetradactyla) were strongly
olfersii assessed
skewed towards highway segments above 100 m amsl
Tupinambis Teiidae 1 Not
teguixim assessed
elevation, i.e. segments 0–140. The distributions of
All unknowns 43
some species were skewed towards the lower eleva-
tions along the highway, i.e. west of segment 70.
IUCN status LC least concern Caiman, marsh deer, white-lipped peccary, and col-
lared peccary (Tayassu tajacu) were more commonly
moderately negative non-significant relationship involved in collisions in the transition zone and
(r = -0.34, p value = 0.10, Fig. 3) to traffic volume, wetland proper.
which stayed nearly constant over the entire sampling
period.
The spatial pattern of roadkill was a general increase Discussion
in roadkill rate from the eastern end of the sampled
stretch at Anastácio (i.e. segment 0) to the western end Wild animal mortality along BR-262 continues to
of the stretch at the Paraguay River Bridge (i.e. segment increase at a substantial rate with potentially dramatic
210) (Fig. 4). The major departure from that general effects on common and rare species alike, as well as

123
Wetlands Ecol Manage (2015) 23:227–239 233

Fig. 2 The total number of recorded roadkilled individuals per species over the entire sampling period for the species most commonly
involved in collisions (species with [20 recorded individuals)

Fig. 3 The total number of

recorded roadkilled
individuals per sampling
event (solid line with solid
diamonds), traffic volume
(in thousands of cars) per
month at the Paraguay River
Bridge toll booth (dashed
line and solid boxes), and
monthly rainfall (mm) at
Fazenda Guaicurus (dotted
line and open triangles)
between April 2011 and
June 2012

life threatening consequences for residents and tourists results indicate an estimated single animal collision
frequently traveling the highway. These human con- every 61 km. The current research found another
sequences could have economic impacts on tax payers tenfold increase in animal mortality between 2002 and
and insurers. Fischer et al. (2004) reported an eightfold 2012, i.e. one roadkill every 6.2 km, with an associ-
increase in mortality from 1992 to 2002 along ated 146 % increase in traffic volume between the first
1,350 km of federal roads in the Pantanal region that 6 months of 2002 and the first 6 months of 2012
included BR-262. Calculations based on their reported (SEOP (Secretaria de Estado de Obras Públicas e de

123
 234

123
Table 3 Pairwise mean differences of the estimated marginal mean of the number of incidents/km by month of sampling
2011 2012
Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May

2011 May 0.159

Jun 0.261 0.102
Jul 0.225 0.066 -0.036
Aug 0.238 0.078 -0.024 0.013
Sep 0.226 0.067 -0.035 0.001 -0.012
Oct 0.258* 0.098 -0.004 0.032 0.020 0.032
Nov 0.192 0.033 -0.069 -0.033 -0.046 -0.034 -0.066
Dec 0.170 0.011 -0.091 -0.055 -0.067 -0.056 -0.087 -0.022
2012 Jan 0.224 0.065 -0.037 -0.001 -0.014 -0.002 -0.033 0.032 0.054
Feb 0.054 -0.106 -0.208 -0.172 -0.184 -0.172* -0.204* -0.138 -0.117 -0.170*
Mar -0.025 -0.184 -0.286* -0.250* -0.262* -0.251* -0.282* -0.217 -0.195* -0.249* -0.078
Apr 0.170 0.011 -0.091 -0.055 -0.068 -0.056 -0.088 -0.022 0.000 -0.054 0.116 0.194
May 0.232 0.072 -0.030 0.006 -0.006 0.006 -0.026 0.040 0.061 0.008 0.178* 0.256* 0.062
Jun 0.248 0.089 -0.013 0.023 0.010 0.022 -0.010 0.056 0.078 0.024 0.194 0.272* 0.078 0.016
Bonferroni adjustment for multiple tests used
Significant differences a \ 0.05 among months indicated with an *
Wetlands Ecol Manage (2015) 23:227–239
Wetlands Ecol Manage (2015) 23:227–239 235

Fig. 4 The number of roadkills/km for each 10 km segment of highway BR-262 (columns) and the mean elevation within each
segment (dashed line). Highway segment 0 begins in Anastácio and increases from east to west until the Paraguay River bridge

Table 4 Pairwise mean differences of the estimated marginal highways in the State of São Paulo, recording 81
mean of the number of incidents/km by highway zone species and one mortality event every 20.8 km.
Upland Transition Wetland Although the Pantanal is considered one of the last
bastions of suitable habitat for a number of threatened
Transition -0.069* and endangered species in Brazil and South America
Wetland -0.117* -0.048 as a whole, better quantification of automobile colli-
Construction zone 0.043 0.112* 0.160* sions’ impacts on population viability of protected
Bonferroni adjustment for multiple tests used species is warranted. The absolute numbers of pro-
Significant differences a \ 0.05 among zones indicated with tected species roadkills found herein are relatively
an * small in some cases, but extrapolated to an entire year
the numbers become more alarming, i.e. 4 tapir
Transportes) 2012). Curiously, the rate of increase in carcasses = 97/year, 6 marsh deer = 146/year, 19
wildlife mortality appears to have outpaced the rate of giant anteater = 462/year, and 20 white-lipped pec-
increase in traffic volume, and the probability of being cary = 486/year. The white-lipped peccary is distrib-
involved in a collision with wildlife increased [ uted widely in the region (Biondo et al. 2011), with
fourfold. densities 7.5–9.6 individuals/km2 within parts of the
The mortality rate on BR-262 in the Pantanal is southern Pantanal (Desbiez et al. 2004). The IUCN
greater than at other locations in Brazil where roadkill Red List vulnerable species of greatest concern may be
rates have been calculated. Only Tonin et al. (2009) the giant anteater. Roadkills are explicitly mentioned
found a roadkill rate similar to the results reported as a major factor in the population decline of giant
herein. Along state and federal highways in Rio anteater in the IUCN species description (Miranda and
Grande do Sul state, the rate was one carcass every Medri 2010). Coutinho et al. (1997) estimated that
6.8 km. Other reported roadkill rates in Brazil include there may be as few as 5,000 giant anteaters in the
Prado et al.’s (2006) research along a highway running Pantanal, thus automobile collisions along BR-262
through Parque Estadual Altamiro de Moura Pacheco may account for mortality of close to 1/10 of the
in Goias state, where one roadkill every 11.4 km was population every year. The marsh deer population in
found, and (Prada 2004) sampled 12,440 km of the Pantanal is relatively robust at an estimated

123
236 Wetlands Ecol Manage (2015) 23:227–239

*44,000 individuals by Mourão et al. (2000). In the areas are unavailable for foraging and groups subdi-
vicinity of BR-262, that study estimated population vide and, in undeveloped areas, are largely confined to
density was greatest at the lower elevations between higher elevated forest patches (Alho 2005; Alho and
Miranda and the Paraguay River bridge, and, as Silva 2012). Thus, capybara’s natural wet season
expected, road mortality of marsh deer was greater behavior may attract them to the transition zone and to
along that stretch of highway. The Pantanal region is man-made elevated environments where they are
also considered a stronghold for the lowland tapir, available.
with over 120,000 km2 of potential tapir habitat, A variety of mitigation approaches are used to
representing a priority ‘Tapir Conservation Unit’ for reduce the effects of roads and road mortality on
the survival of the species over the next 100 years wildlife. Motorist’s behavior often requires modifica-
(Trolle et al. 2008). However, population viability tion, involving education, speed limits, warning lights,
analysis in another region in Brazil, the Atlantic Forest and signage. Animal behavior modification is also a
of the Pontal do Paranapanema region, has shown that common approach, including installation of wildlife
high roadkill rates have a major impact on the lowland overpasses and underpasses, fences, verge mainte-
tapir population and may lead to population extinction nance, and measures that warn or scare animals from
in that region (Medici and Desbiez 2012). approaching the roadway (Romin and Bissonette
There are potentially a number of interacting 1996; Forman et al. 2003; Glista et al. 2009).
factors influencing the spatial and temporal patterns Maintenance and vegetation removal along the road
of roadkill rates along BR-262. These factors include verge can ensure that the animals pause before
decreasing elevation along the highway into the entering the roadway, improve driver visibility of
wetland, seasonal wetland inundation patterns, animal approaching animals, and reduce accidents (Bagatini
behavior, potential use of the elevated roadway as 2006; Catella et al. 2010). As with humans, signals
refuge, and minimal development at the lowest that alert animals to traffic may also be a means to
elevations in the wetland proper. In the wet season, decrease roadkill rates. Schafer and Penland (1985)
wetland inundation leaves the causeway and transition found that 52 deer were killed at night along test
zone, where the highway parallels the Miranda River, sections where headlight reflectors along the verge
above the waterline. These dry locations may function were covered, but only six deer were killed when
as refugia with available resources for the abundant reflectors were uncovered and operative. Within the
wildlife typically concentrated in the less developed extensive wetland systems of Florida, USA, fencing
wetlands, and concomitantly increase the probability and animal underpasses installed in coordination to
of vehicle collisions (Fischer 1997; Catella et al. funnel animals under roadways has proven effective
2010). for many years with both herpetofauna and large
Two of the species most commonly involved in mammals, e.g. Florida panthers (Puma concolor
roadway collisions show seasonal migration behaviors coryi) (Aresco 2003; Foster and Humphrey 1995).
that may be representative of behavior across multiple Catella et al. (2010) and Hobday and Minstrell
wetland taxa. The caiman exhibits seasonal variations (2008) argue that an effective measure to reduce
in diet, foraging behavior, habitat selection, and wildlife collisions is increasing the awareness of
fecundity with variation in hydrology. Caiman con- drivers about the behavior of animals, making drivers
gregate in high densities around sparse watering holes better informed and prepared to deal with roadway
in the dry season, but begin to disperse through the encounters. General education likely requires pairing
floodplain when water levels exceed 15 cm (Coutinho with direct roadway information, however, Dique
and Campos 1996; Campos and Magnusson 1995; et al. (2003) evaluated the effect of speed signaling
Alho and Silva 2012). Widespread wet season migra- and warnings to minimize accidents involving koalas
tion may increase the likelihood of contact with in southeast Queensland, but they found no reduction
roadways and subsequent use of the highway for rest, in the rates collisions. Hobday and Minstrell (2008)
body temperature regulation, or attempted passage believe speed reduction can effectively decrease the
over the migration barrier. Capybaras have larger number of incidents. Jones’ (2000) research in Aus-
home ranges and find abundant food in grasslands in tralia found that where speed bumps were installed,
the middle of the dry season, but during floods these the numbers of roadkill were reduced. Research along

123
Wetlands Ecol Manage (2015) 23:227–239 237

BR-262 indicates that forced speed reduction may be Funding Sources Funding for the research described in the
effective. Highway construction between segments submitted manuscript was provided to Julio de Souza by the
following sources: Fundação CAPES, Ministério da Educação
170–190 during sampling required motorists to drive do Brasil, Brasilia/DF Universidade Federal de Mato Grosso do
at a reduced speed, which may have directly influ- Sul.
enced the reduced accident frequency compared to
surrounding segments. However, noise from construc-
tion and the constant presence of humans in the work References
zone may have also kept animals away from the road.
The ecological and anthropogenic context of con- Alho CJR (2005) The Pantanal. In: Fraser LH, Keddy PA (eds)
flict between humans and wildlife along BR-262 in the The world’s largest wetlands—ecology and conservation.
Pantanal is likely similar to that found in many Cambridge University Press, New York, pp 203–271
Alho CJR, Silva JSV (2012) Effects of severe floods and
globally significant flood-pulsed wetlands where droughts on wildlife of the Pantanal wetland (Brazil)—a
development is encroaching as transportation net- review. Animals 2:591–610
works are improving. Efficient designs for mitigation Alho CJR, Camargo G, Fisher E (2011) Terrestrial and aquatic
measures in these locations will take the contexts of mammals of the Pantanal. Braz J Biol 71:297–310
Aresco MJ (2003) Highway mortality of turtles and other her-
interacting factors into account during planning. petofauna at Lake Jackson, Florida, USA and the efficacy
While the search for zero mortality may suggest of a temporary fence/culvert system to reduce road kills. In:
solutions involving complete fencing with frequent Irwin CL, Garrett P, McDermott KP (eds) 2003 Proceed-
wildlife underpasses in wetlands and overpasses in the ings of the international conference on ecology and trans-
portation. Center for Transportation and the Environment,
surrounding higher elevations, financial realities dic- North Carolina State University, Raleigh, pp 433–449
tate choosing the most cost effective solutions, and Bagatini T (2006) Evolução dos ı́ndices de atropelamento de
context driven designs can inform those decisions. vertebrados silvestres nas rodovias do entorno da Estação
The data from the Pantanal suggests that simply ecológica de Águas Emendadas, DF, Brasil, e eficácia de
medidads mitigatórias. Master’s thesis, Universidade de
reducing the *3 month wet season mortality peak Brasilia
to background rates would decrease by [1/3 the Bandeira C, Floriano EP (2004) Avaliação de impacto ambi-
number of wildlife collisions. Thus, cost effective ental de rodovias. Caderno Didático, Santa Rosa, p 69
design may rely on seasonally adjusted mitigation in Biondo C, Keuroghlian A, Gongora J, Miyaki CY (2011) Pop-
ulation genetic structure and dispersal in white-lipped
flood-pulsed wetlands. Mitigation measures, such as peccaries (Tayassu pecari) from the Brazilian Pantanal.
speed reduction, can be tied to inundation patterns J Mammology 92:267–274
with the use of automated depth gauges. Temporally Campos Z, Magnusson W (1995) Relationship between rainfall,
controlled speed zoning and enforcement may nesting habitat and fecundity of Caiman crocodilus yacare
in the Pantanal, Brazil. J Trop Ecol 11:351–358
reduce costs and intrusion on local residents and Catella AC, Thomas WM, Mourão GM (2010) BR-262 no
commerce. Relatively low-cost solutions that reduce Pantanal: cenário de encontros entre homens e animais
the temporal mortality peaks in locations with silvestres. Embrapa Pantanal, Corumbá
greater collision rates may afford the necessary CNT (Confederação Nacional do Transporte) (2006) Atlas do
Transporte, 1st edn. http://www.sistemacnt.org.br/informa
time and resources for governments to determine coes/pesquisas/atlas/2006/index.htm. Accessed 24 May 2012
whether more extensive and constant measures are Conover MR, Pitt WC, Kessler KK, DuBow TJ, Sanborn WA
required, i.e. fencing, and for scientists to study the (1995) Review of human injuries, illnesses, and economic
viability of threatened species under various man- losses caused by wildlife in the United States. Wildl Soc
Bull 23:407–414
agement and mitigation scenarios. Courtenay O, Maffei L (2008) Cerdocyon thous. In: IUCN 2013.
IUCN red list of threatened species. Version 2013.2. http://
Acknowledgments Dr. de Souza would like to thank the www.iucnredlist.org. Accessed 27 Apr 2014
Fundação CAPES, Ministério da Educação do Brasil, Brasilia/ Coutinho M, Campos Z (1996) Effect of habitat and seasonality
DF, and Universidade Federal de Mato Grosso do Sul for on the densities of caiman in Southern Pantanal, Brazil.
support of this research, as well as Andrea G. Dos Santos, J Trop Ecol 12:741–747
Carolina F. de Souza, Geovane G. Ramires, Israel Cardoso, Coutinho ME, Campos ZMS, de Mourão G, Mauro RA (1997)
Jenifer Azambuja, Marcos P. G. Rezende, Sandriane S. Batista, Aspectos ecológicos dos vertebrados terrestres e semi aq-
Vanessa T. Gonçalves for assistance with field sampling. The uáticos no Pantanal. Plano de Conservação da Bacia do
authors would also like to thank two anonymous reviewers for Alto Paraguai (Pantanal)—PCBAP: diagnóstico dos meios
their insightful recommendations. fı́sicos e bióticos: meio biótico. Ministério do Meio

123
238 Wetlands Ecol Manage (2015) 23:227–239

Ambiente, dos Recursos Hı́dricos e da Amazônia Legal, Harris MB, Arcangelo C, Pinto ECT, Camargo G, Ramos-Neto
Brası́lia, pp 183–322 MB, Silva SM (2005) Estimativas de perda da área natural
Desbiez A, Donatti CI, Marques RM, Keuroghlian A, Tomas da Bacia do Alto Paraguai e Pantanal Brasileiro. Relatório
WM, Galetti M, Santos SA, Bodmer RE (2004) Uso de técnico não publicado. Conservação Internacional, Campo
Habitat e Densidades Populacionais de Queixadas, Catetos Grande
e Porcos-Monteiros em duas Áreas do Pantanal Brasileiro. Hobday AJ, Minstrell ML (2008) Distribution and abundance of
VI Congresso Internacional sobre Manejo de Fauna Sil- roadkill on Tasmanian highways: human management
vestre en la Amazonia y Latinoamerica, Iquitos options. Wildl Res 35:712–726
Dique DS, Thompson J, Preece HJ, Penfold GC, Villiers DL, Leslie IBM Corp (2010) IBM SPSS statistics for windows. Version
RS (2003) Koala mortality on roads in south-east Queensland: 19.0. IBM Corp, Armonk
the koala speed-zone trial. Wildl Res 30:419–426 IUCN (2013) The IUCN red list of threatened species. Version
ESRI (2011) ArcGIS desktop: release 10. Environmental Sys- 2013.2. http://www.iucnredlist.org. Accessed 27 Apr 2014
tems Research Institute, Redlands Jones ME (2000) Road upgrade, road mortality and remedial
Evink GL, Garret P, Zeigler D, Berry J (1996) Trends in measures: impacts on the population of eastern quolls and
addressing transportation related wildlife mortality. Pub- Tasmanian devils. Wildl Res 27:289–296
lication no. FL-ER-58-96. Florida Department of Trans- Junk WJ, da Silva CJ (2000) O conceito de inundação e suas
portation, Tallahassee implicações para o Pantanal de Mato Grosso. In: Anais do II
Fischer WA (1997) Efeitos da rodovia BR-262 na mortalidade Simpósio sobre Recursos Naturais e Sócio-econômicos do
de vertebrados silvestres: Sı́ntese naturalı́stica para a con- Pantanal. Manejo e Conservação. EMBRAPA, Corumbá
servação da região do Pantanal. Master’s thesis, Univer- Keuroghlian A, Desbiez A, Reyna-Hurtado R, Altrichter M,
sidade Federal de Mato Grosso do Sul Beck H, Taber A, Fragoso JMV (2013) Tayassu pecari. In:
Fischer WA (1999) Programa Estrada Viva—Volumes 1 e 2: IUCN 2013. IUCN red list of threatened species. Version
Impactos da BR-262 sobre a Vida Selvagem e Proposta de 2013.2. http://www.iucnredlist.org. Accessed 27 Apr 2014
Intervenção. Technical report, GEIPOT Conv. Min.Tran- Lalo J (1987) The problem of roadkill. Am For 50:50–52
portes/UFMS, Brası́lia Maehr DS, Land ED, Roelke ME (1991) Mortality patterns of
Fischer WA, Ramos-Neto MB, Silveira L, Jácomo ATA (2004) Panthers in southwest Florida. Proceedings of the annual
Human transportation network as ecological barrier for conference of the southeastern association of fish and
wildlife on Brazilian Pantanal-Cerrado corridors. In: Irwin wildlife agencies, vol 45., pp 201–207
CL, Garrett P, McDermott KP (eds) Proceedings of the Martinelli MM, Volpi TA (2011) Mamı́feros atropelados na
2003 international conference on ecology and transporta- Rodovia Armando Martinelli (ES-080), Espı́rito Santo,
tion. Center for Transportation and the Environment, North Brasil. Espı́rito Santo. V. Natureza on line 9:113–116
Carolina State University, Raleigh, pp 182–194 Medici EP, Desbiez ALJ (2012) Population viability analysis:
Forman RTT (1995) Land mosaics: the ecology of landscapes using a modeling tool to assess the viability of tapir pop-
and regions. Cambridge University Press, Cambridge ulations in fragmented landscapes. Integr Zool 7:356–372
Forman RTT, Alexander LE (1998) Roads and their major Miranda F, Meritt DA Jr. (2011) Tamandua tetradactyla. In:
ecological effects. Annu Rev Ecol Syst 29:207–231 IUCN 2013. IUCN red list of threatened species. Version
Forman RTT, Friedman DS, Fitzhenry D, Martin JD, Chen AS, 2013.2. http://www.iucnredlist.org. Accessed 27 Apr 2014
Alexander LE (1997) Ecological effects of roads: toward Miranda F, Medri I (2010) Myrmecophaga tridactyla. In: IUCN
three summary indices and an overview for North America. 2012 IUCN red list of threatened species. Version 2012.2.
In: Canters K (ed) Habitat fragmentation and infrastruc- www.iucnredlist.org. Accessed 5 June 2013
ture—proceedings. Ministry of Transport, Public Works & MMA (Ministério do Meio Ambiente) (2007) Cerrado and
Water Management, Delft, pp 40–54 Pantanal: priority areas and actions for biodiversity con-
Forman RTT, Sperling D, Bissonette JA, Clevenger AP, Cut- servation. Biodiversity Series 17, Ministério do Meio
shall CD, Dale VH, Fahrig L, France R, Goldman CR, Ambiente, Brası́lia
Heanue K, Jones JA, Swanson FJ, Turrentine T, Winter TC Mourão GM, Coutinho M, Mauro R, Campos Z, Tomas WM,
(2003) Road ecology: science and solutions. Island Press, Magnusson WE (2000) Aerial surveys of caiman, marsh
Washington deer and pampas deer in the Pantanal wetland of Brazil.
Foster ML, Humphrey SR (1995) Use of highway underpasses Biol Conserv 92:175–183
by Florida panthers and other wildlife. Wildl Soc Bull Naveda A, de Thoisy B, Richard-Hansen C, Torres DA, Salas L,
23:95–100 Wallance R, Chalukian S, de Bustos S (2008) Tapirus
Freitas CH (2009) Atropelamento de vertebrados nas rodovias terrestris. In: IUCN 2013. IUCN red list of threatened
MG- 428 e SP-334 com análises de fatores condicionantes species. Version 2013.2. http://www.iucnredlist.org.
e valorização econômica da fauna. Ph.D. Dissertation, In- Accessed 27 Apr 2014
stituto de Biociências, Universidade Estadual Paulista Nunes AP (2011) Quantas espécies de aves ocorrem no Pantanal
‘‘Júlio Mesquita Filho’’ brasileiro? Atualidades Ornitológicas On-line 160:45–54
Glista DJ, DeVault TL, DeWoody JA (2008) Vertebrate road Potts A, Potts V (2004) Features and conservation of the Bra-
mortality predominantly impacts Amphibians. Herpetol zilian Pantanal wetland. Wetl Ecol Manage 12:547–552
Conserv Biol 3:77–87 Prada CS (2004) Atropelamento de vertebrados silvestres em
Glista DJ, DeVault TL, DeWoody JA (2009) A review of mit- uma região fragmentada do nordeste do Estado de São
igation measures for reducing wildlife mortality on road- Paulo: quantificação do impacto e análise de fatores en-
way. Landsc Urban Plan 91:1–7 volvidos. Ph.D. Dissertaiton, Centro de Ciências

123
Wetlands Ecol Manage (2015) 23:227–239 239

Biológicas e da Saúde, Universidade Federal de São Car- SEOP (Secretaria de Estado de Obras Públicas e de Transportes)
los, São Carlos (2012) Mapa do tráfego total mensal, praça de pedágio,
Prado TR, Ferreira AA, Sobrinha Guimarães ZF (2006) Efeito Porto Morrinho
da implantação de rodovias no cerrado sobre a fauna de Souza JC, Lima JAF (2012) O Pantanal de Forma Sustentável.
vertebrados. Acta Sci Biol Sci 18:237–241 In: Julio César de Souza (ed) Pantanal Produzindo com
Rocha EHD (2005) Impactos dos transportes rodoviários na Sustentabilidade. Universidade Federal de Mato Grosso do
fauna. Master’s thesis, Universidade Federal do Rio de Sul, Campo Grande, pp 7–18
Janeiro, Rio de Janeiro Tonin AM, Amaral BTC, Paier C, de Oliveira CH. (2009) Im-
Rodrigues FHG, Medri IM, Thomas VM, Mourão GM (2002) pactos a fauna silvestre por atropelamento nas rodovias da
Revisão do conhecimento sobre ocorrência e distribuição região norte do Rio Grande do Sul, Brasil. In: Congresso de
de mamı́feros do Pantanal. Embrapa Pantanal, Corumbá ecologia do Brasil, 9. São Lourenço
Romin LA, Bissonette JA (1996) Deer-vehicle collisions: status Trolle M, Noss AJ, Cordeiro JLP, Oliveira LFB (2008) Brazilian
of statemonitoring activities and mitigation efforts. Wildl tapir density in the Pantanal: a comparison of systematic
Soc Bull 24:276–283 camera-trapping and line-transect surveys. Biotropica
Rosa AO, Mauhs J (2004) Atropelamentos de animais silvestres 40:211–217
na Rodovia RS-040. Caderno de Pesquisa Série Biologia Trombulak SC, Frissell CA (2000) Review of ecological effects
16:35–42 of roads on terrestrial and aquatic communities. Conserv
Sampaio RS, Brito PCR (2009) Impactos ambientais causados Biol 14:18–30
pela construção de rodovias. Master’s thesis, Universidade WWF—Brasil (2012) Sustentabilidade: a teoria a prática. http://
Católica do Salvador, Salvador www.wwf.org.br/. Accessed 6 Jan 2012
Schafer JA, Penland ST (1985) Effectiveness of Swareflex
reflectors in reducing deer–vehicle accidents. J Wildl
Manage 49:774–776

123