Acta Oecologica 69 (2015) 78e86

Contents lists available at ScienceDirect

Acta Oecologica
journal homepage: www.elsevier.com/locate/actoec

Original article

Amphibian diversity increases in an heterogeneous agricultural

landscape
Cecilia Guerra a, b, *, Ezequiel Ara

oz a, c
a
Consejo Nacional de Investigaciones Cientcas y T
ecnicas (CONICET), Argentina
b
Instituto de Herpetologa, Fundacio
n, PC 4000 Tucuma

n Miguel Lillo, Miguel Lillo 251, San Miguel de Tucuma
n, Argentina
c
Instituto de Ecologa Regional, Facultad de Ciencias Naturales, Universidad Nacional de Tucuma
n, Horco Molle (CC 34), Yerba Buena, PC 4107 Tucuma

n,
Argentina

a r t i c l e i n f o a b s t r a c t

Article history: As a group amphibians are the vertebrates most affected by anthropic activity, particularly by agriculture.
Received 26 May 2015 The rapid advance of the agricultural frontier makes it important to identify the role of agroecosystems
Received in revised form as habitat supply for amphibians. We analyzed the differences in amphibian assemblages and pop-
13 September 2015
ulations between habitats with different plant covers and different degrees of human intervention in
Accepted 14 September 2015
Available online xxx
northwestern Argentina. For three years we conducted 114 high frequency trap samplings to quantify
abundance, specic composition and species richness of amphibian assemblages in three habitat types
(lemon and sugarcane crops and secondary forest) of a piedmont agroecosystem of Tucuma
n province.
Keywords:
Anurans
Crops hosted more species and individuals than secondary forests, but the specic composition of forest
Croplands was different from that of crops suggesting that they could be complementary. Although the assemblage
Disturbances abundance of every observation responded to climate, the strong effect of sampling year was not related
Yungas to climatic factors suggesting that there might be long term uctuations that were not analyzed. We also
Argentina found that responses to agricultural practices were species specic, so no generalizations about these
practices should be done. Our study shows that cultivated areas are not hostile environments for am-
phibians since they can lodge huge amphibian populations and that the occurrence of disturbances
associated to agricultural practices is not necessarily reected in a decrease in the abundance and
richness of amphibians in the short term. However, our results also show that forested lands are
necessary to lodge some specialist species which are very rare in the croplands. This suggests that
environmental heterogeneity generated by the combination of natural and cultivated patches can in-
crease biodiversity at landscape scale because this allows the coexistence of species related to either kind
of environment.
2015 Elsevier Masson SAS. All rights reserved.

1. Introduction causes of this decline (e.g. Collins and Storfer, 2003). For example,
several studies showed that amphibians are strongly affected by
Amphibians are the group of vertebrates most affected by an- environmental conditions (e.g. Prado and Rossa-Feres, 2014) and
thropic activity; nearly 32% of their species being threatened or some of them describe the effects of habitat fragmentation on
extinct (IUCN, 2014; Wake and Vredenburg, 2008). For the last amphibians species richness (e.g. Peltzer et al., 2003) or show that
decades the decline of amphibians has been recognized as a global the advance of the agricultural frontier affects amphibian assem-
phenomenon which was assessed quantitatively (Houlahan et al., blages through different mechanisms (e.g. Hazell et al., 2001;
2000), in all the continents and in all the habitat types (Lavilla, Johansson et al., 2005). Recurrent mechanical disturbances asso-
2001). Different processes have been mentioned as possible ciated to agricultural practices alter the soil surface affecting the
sites for reproduction (wetlands, ponds and pools), shelter and
hibernation of amphibians (Bishop et al., 1999; Knutson et al.,
* Corresponding author. Miguel Lillo 251, San Miguel de Tucuma
n, PC 4000 2004). Pollution with agrochemicals has been regarded as a rele-
Tucuma
n, Argentina. vant factor in the decrease of the species richness and abundance of
E-mail addresses: [email protected] (C. Guerra), [email protected] amphibians (Mann et al., 2009). Taking into account the accelerated

oz).
(E. Ara

http://dx.doi.org/10.1016/j.actao.2015.09.003
1146-609X/ 2015 Elsevier Masson SAS. All rights reserved.

oz / Acta Oecologica 69 (2015) 78e86
C. Guerra, E. Ara 79

advance of the agricultural frontier, it is of great importance to southeast direction.

analyze the role of agroecosystems in biodiversity conservation and The farm is mainly covered with sugarcane and lemon planta-
as habitat supply for amphibians (Peltzer et al., 2006). Such ana- tions but in some areas, which are not apt for mechanized agri-
lyses should take into account the strong association between culture, remnants of natural forest are the main land cover. The
amphibians and environmental variables. resulting spatial pattern is a mosaic of croplands and secondary
It is widely accepted that climatic variables, mainly temperature forests which are connected with the Yungas natural vegetation of
and rainfall are relevant abiotic factors inuencing amphibians at the Sierra de San Javier. We selected six sites of Lules Farm to
individual and population levels (e.g. Parris, 2004; Urbina Cardona sample the amphibian assemblages: two within the piedmont
et al., 2006) due to the effects of climate on their history of life forest, two within the sugarcane crops and two within the lemon
(Prado et al., 2005) and physiology (Pounds et al., 2006). In sub- plantation (Table 1). All the sites had similar topographies and were
tropical areas, the activity of amphibians and the dynamics of their located at approximately 20 m from a water course, which consti-
populations are strongly inuenced by the climatic seasonality; the tutes an attractor for amphibians. To avoid spatial interaction be-
reproduction of most species occurs in the warm season of the year tween sampling sites they were located at least at 450 m from each
(Conte and Machado, 2005). Thus, depending on their life histories other. The distance between sites with the same cover type was at
and ecophysiological characteristics, species differ in their toler- least at 1200 m. Average distance between all sampling sites was
ance to climate conditions (see Wells, 2007) which determines the 1541 m.
dynamics of amphibian populations (Sanchez et al., 2007). For
example, generalist species -that use a broad spectrum of the 2.2. Sampling methods and design
environment- have wide habitat tolerance and are more abundant
in disturbed environments than specialist species, which are We used the pitfall trapping method, modied from Corn
restricted in their distribution (Dash and Mahanta, 1993). More- (2001). The traps consisted of plastic cylindrical 33 cm diameter
over, it has been observed that species may respond differentially and 39 cm height buckets, with side perforations to prevent
not only to climate but to the occurrence of disturbances (e.g. ooding. In every site, we placed 20 pitfall traps arranged into ve
Pineda and Halffter, 2004). Agricultural practices can be considered rows separated by 40 m. In every row we placed four traps, sepa-
as disturbances within croplands (Zhao and Neher, 2013). rated by 10 m. It was not possible to install the fences suggested by
Amphibian populations are dynamic; population dynamics are not Corn (2001), which increase the efcacy of the capture method,
simple and their response to disturbance events may not be im- because they would have interfered with the agricultural tasks. We
mediate, since delays are frequent in ecological processses. So it is conducted 114 samplings from February 2007 to April 2010, which
important to evaluate the mid term effect of these disturbances on totaled 1131 days.
the dynamics of amphibian populations and assemblage. Since Traps were checked once a week during the spring and summer
climate and disturbance may affect each species differentially season and every fortnight in the autumn and winter season when
(Wells, 2007), the dynamics of the properties of the assemblage the activity of amphibians decreases almost completely. Yet the
would also change in response to these variables (Parris, 2004; interval between samplings was not homogeneous, so the interval
Pineda and Halffter, 2004). These factors may affect populations length was taken into account in the computation of the sampling
and assemblage of amphibians to different extents; it is important effort. As the objective of the survey was to perform a long-lasting
to assess the particular contribution of every factor to the dynamics study of the abundance of amphibian we did preliminary surveys to
of populations of amphibians. decide on the optimal interval between samplings taking into ac-
Our study aims at assessing the joint effect of habitat type, count the trade-off between the sampling effort and the survival of
agricultural practices, climate and sampling year on abundance and amphibians. In the survey we found few dead individuals due to
species richness and composition of amphibians. Our three specic dehydration, only in extremely hot days, and it was possible to
aims were: 1) to assess the potential of each habitat type to host identify them to the species level. Only once we found a snake in
amphibian populations and assemblages, and to detect potential the pitfalls which could be a potential predator of amphibians.
complementary effects between them; 2) to assess differential re- Although we acknowledge that there might be some loss of infor-
sponses of amphibian populations and assemblages to habitat type, mation (e.g. escaped or predated specimens), the method provides
agricultural practices and climate and to evaluate whether pop- a good indicator of the abundance of amphibians and it is partic-
ulations and assemblage uctuated through the years; 3) to ularly useful when comparisons between environments are per-
determine which kind of agricultural practice is more harmful to formed. In every sampling we identied the species of the captured
the amphibian populations and assemblages. specimens and then we freed them near the capture area. During
the rst sampling year, 258 captured individuals were marked by
2. Materials and methods means of toe clipping but, due to the noticeably low number of
recaptures (11), we discontinued the procedure in the successive
2.1. Study area sampling years.
The Famailla
Station of the National Institute of Agropecuarian
We carried out samplings in the Lules Farm (26530 2000 S; Technology (INTA, 2011) provided meteorological records of daily
65200 3200 W), located in the Lules Department, province of minimum, maximum and mean temperature. Rainfall data, which
Tucuma
n. The farm lies in the Yungas piedmont area between 450 is spatially variable in montane environments was provided by
and 750 m.a.s.l. (Sesma et al., 2010) and is characterized by a sub- Lules Farm, who also provided information on the different agri-
tropical climate with an average annual temperature of 19  C and a cultural practices performed in the croplands under study: me-
marked seasonality (Brown and Malizia, 2004). Rainfall regime is chanical interventions (weeding and harvesting of lemon crops and
monzonic with dry winters and rainy summers (Grau et al., 2010). harvesting in sugarcane patches) and application of herbicides,
The most important hydrological system in the area is the basin of insecticides, fungicides and fertilizers.
the Lules River. The water of the system has good physical and
chemical characteristics, with high values of dissolved oxygen and 2.3. Data analyses
low conductivity (Sesma et al., 2010). Some streams of this basin are
channeled for irrigation and ow along the farm in a northwest to To evaluate the variation in species composition of amphibian
80
oz / Acta Oecologica 69 (2015) 78e86
C. Guerra, E. Ara

Table 1
Description of the six sampling sites.

Geographical coordinates Cover type Vegetation Plant location Plant harvesting Main agricultural practices
height
Mechanical Agrochemical
interventions application

Sugarcane crop 1 26540 0200 S; 65190 4400 W Gramineous up to 6 m Furrow ridges in rows Annually harvested Plant harvesting Herbicides,
Sugarcane crop 2 26520 5700 S; 65200 1200 W separated by approximately insecticides
80 cm and fertilizers
Lemon plantation 1 26530 4700 S; 65200 3900 W Ever green 3 to 6 m Furrow ridges in rows Not harvested Weeding and Herbicides,
Lemon plantation 2 26530 1300 S; 65200 0900 W fruit tree separated by approximately (perennial plants) fruit harvesting insecticides,
8m fungicides
and fertilizers
Piedmont forest 1 26530 2900 S; 65210 2600 W Yungas forest up to 30 m It is the natural forest
Piedmont forest 2 26520 4900 S; 65200 4000 W

Sources: Gonz

alez Sicilia, 1960; Grau et al., 2010; Humbert, 1974; Lules Farm.

assemblages between habitat types, taking into account the sam- the parameters of two components; a binomial distribution which
pling years and seasons of the year, we conducted a non parametric models the amphibian presence and a Poisson distribution which
multidimensional arrangement (NMDS, Kruskal and Wish, 1978), models individual counts. This approach has the advantage of
with PC ORD (McCune and Mefford, 1999). The aim of this analysis dealing with count data with no need of transformations at the
was to assess variation between different combinations of condi- time that it helps in the identication of possible causes of the
tions, which are considered to have signicant ecological implica- excess of zeros (e.g. detection probability, Wenger and Freeman,
tions for amphibians. In each combination more than one single 2008). The models were parameterized with the zeroin function
observation were clumped, thus we rearranged the raw data of from the pscl package (Jackman, 2011) of R (R Development Core
observation in order to generate composites of abundance for every Team, 2011). We compared 45 models resulting from the combi-
combination. We grouped the abundance of every species consid- nation of three binomial and 15 Poisson components (Appendix 1),
ering the habitat type, the sampling year and the season of the year, for the abundance and species richness of the assemblage. We also
and performed a single NMDS based on their Bray Curtis distance modeled the abundance of the ve most abundant species, which
matrix (Legendre and Legendre, 2012). To statistically evaluate the presented at least 5 records in each habitat type (for example, we
compositional distances between the grouping variables we per- did not model the abundance of Oreobates discoidalis, for which we
formed Kruskal Wallis analyses on the scores for each axis, and had only 1 record in sugarcane and 2 records in lemon crops). For
used Dunn test to determine which treatments differed the binomial component, we evaluated one explanatory variable at
signicantly. a time; none, minimum temperature or rainfall. For the Poisson
To evaluate the association of habitat type, sampling year, component, we used different combinations of four sets of
climate, and agricultural practices with the presence and abun- explanatory variables: habitat type (forest, sugarcane crop and
dance of amphibians we compared the tting of a series of gener- lemon crop), sampling year (rst, second and third), climate (mean
alized linear models (GLM) for the whole assemblage (Gelman and temperature and rainfall) and agricultural practices (mechanical
Hill, 2007). In this analysis we aimed at identifying the key vari- interventions, herbicides, insecticides and fertilizers application).
ables that controlled the abundance of amphibians in every sam- We did not include minimum and maximum temperature because
pling, so we used the raw data of abundance of each sampling. We they were highly correlated with mean temperature. Data on
used a zero inated Poisson distribution to simultaneously estimate agricultural practices were provided by Lules Farm. Weeding and
harvesting of lemon crops and harvesting in sugarcane patches
were clumped within mechanical interventions. We grouped fun-
gicides and insecticides in the same category because they were
applied simultaneously in lemon crops (Appendix 2). For the ana-
lyses of the effect of these disturbances we considered one subse-
quent sampling to the agricultural practice. All the models included
traps nights as an offset to take into account the survey effort. In
previous analyses we evaluated the interactions between explan-
atory variables. This generated a multitude of models that did not
improve the modeling. We decided to keep the simplest version to
assess the individual effect of groups of factors because we were
interested on their additive effects and we decided to avoid the
proliferation of models. We used the Akaike information criterion
(AIC) to jointly evaluate the complexity and the goodness of t of
every model to the data. As we intended to nd a single model to
explain the different dynamics of amphibians we used a hybrid
criterion. We kept the best common model for every response
variable whenever it was not signicantly overcome by an alter-
native model, but we kept the alternative model when it signi-
cantly overcome the common one.
Fig. 1. Non parametric multidimensional arrangement (NMDS) of composites based on In order to further evaluate the short term response of
the abundance of amphibian species. Composites summarize the assemblage infor-
amphibian populations and assemblages to the agricultural prac-
mation of single observations based on their habitat type, sampling year and season of
the year. AeW: autumn and winter; SeS: spring and summer. The numbers (1e3) tices (application of agrochemicals and mechanical interventions),
correspond to three consecutive sampling years. Final stress 10.41, p 0.004. we carried out superposed epoch analyses (SEA, Haurwitz and

oz / Acta Oecologica 69 (2015) 78e86
C. Guerra, E. Ara 81

Table 2
Results of the best generalized linear model to t the abundance and species richness of the assemblages and the abundance of the ve most abundant species: Estimate
(standard error). dAIC: differential Akaike information criterion, df: degree of freedom, min. temp.: minimum temperature, mean temp.: mean temperature, mech.: mechanical
interventions, herb.: herbicides application, insec.: insecticides application, fert.: fertilizers application. Bold numbers indicates no signicative effect of the variable (p > 0.05).

GLM Binomial distribution Poisson distribution

Habitat type Sampling year Climatic vbles. Agricultural practices

dAIC Df Weight Min. temp. Rainfall None Cane Lemon second Third Mean temp. Rainfall Mech. Herb. Insec. Fert.

Total abundance 0 13 0.89 0.69 e 0.82 0.38 1 2.23 0.52 0.02 0 0.39 0.20 1.13
(0.24) (0.04) (0.05) (0.03) (0.04) (0.01) (0.01) (0.04) (0.04) (0.03) (0.06)
Total richness 0 13 0.95 0.29 e e 0.26 0.42 0.90 1.87 0.44 0.04 0.35 0.22 0.33 1.41
(0.1) (0.1) (0.13) (0.09) (0.11) (0.01) (0.01) (0.10) (0.10) (0.08) (0.16)
Rhinella arenarum 0.5 13 0.43 0.2 e e 0.46 0.22 1.07 1.93 0.56 0.01 0.22 0.56 0.24 0.67
(0.04) (0.05) (0.07) (0.06) (0.06) (0.01) (0.01) (0.06) (0.08) (0.05) (0.15)
Pleurodema borellii 2 13 0.21 0 e e 0.17 0.75 0.79 1.8 0.45 0.06 0.01 0.47 0.29 1.26
(0.04) (0.21) (0.18) (0.16) (0.17) (0.02) (0.01) (0.13) (0.21) (0.14) (0.2)
Physalaemus biligonigerus 0 13 1 0.34 e e 2.63 3.26 0.09 1.4 0.19 0.04 0.28 0.63 0.41 0.59
(0.06) (0.33) (0.34) (0.18) (0.2) (0.02) (0.01) (0.14) (0.14) (0.12) (0.2)
Odontophrynus americanus 0 9 0.43 0.2 e e 2.84 1.35 2.31 3.72 0.31 0.09 e e e e
(0.09) (0.55) (0.49) (0.47) (0.47) (0.09) (0.03)
Leptodactylus latinasus 0 13 1 0.18 e e 1.16 1.45 1.11 2.11 0.25 0.01 0.26 0.52 0.38 0.57
(0.03) (0.08) (0.09) (0.06) (0.06) (0.01) (0.01) (0.06) (0.06) (0.05) (0.1)

Brier, 1981). SEA is a re-sampling technique that permits statisti- count model (Poisson distribution) that included all the explaining
cally evaluating the temporal response of a variable to the occur- variables: habitat type, sampling year, mean temperature and
rence of an event. It is specially designed to deal with rainfall, and agricultural practices (Table 2).
autocorrelation effects. We matched the dates of agricultural The presence of amphibians was positively associated with the
practices with the intervals between samplings and considered, as minimum temperature. Total abundance was positively related
focal time, the rst sampling after the agricultural practice was with cultivated patches, and the magnitude of the effect of sugar-
carried out (time lag 0). To avoid the effect of seasonality on the cane on total abundance was signicant and higher than that of
practices (e.g. harvest takes place during winter) we previously lemon crops. Species richness was positively associated with sug-
built binomial GLM to account for the probability of occurrence of arcane and negatively with lemon crops. The second and third year
each type of practice in response to temperature. In the SEAs we showed a decrease in the total abundance and richness per sam-
used this modeled probability to simulate the occurrence of every pling, being the effect of the third year bigger than that of the
practice. For the SEAs we considered a time window ranging from second year. Total abundance and species richness were positively
the sampling previous to the agricultural practice (time lag 1) to associated with mean temperature and rainfall. When we consid-
three subsequent samplings (time lag 3). In contrast with the ered one subsequent sampling from the agricultural practice, we
assessment through GLM, these analyses evaluated the evolution of observed positive responses of total abundance and richness to
abundance in a longer time (up to three samplings after the mechanical interventions and to the application of fertilizers and
application) and within a single habitat type. insecticides. By contrast, we observed a negative response of total
abundance and species richness to the application of herbicides
3. Results (Fig. 2A, Appendix 5 and 6).
Taking into account up to three subsequent samplings to the
During the study, we captured 4928 individuals from 12 species: agricultural practice the patterns changed a little. After the appli-
Leptodactylus chaquensis, Leptodactylus latinasus, Leptodactylus cation of insecticides there was a decrease in total abundance in
mystacinus, Physalaemus biligonigerus, Pleurodema borellii, Pleuro- subsequent samplings which was not detected in the GLM (SEA
dema tucumanum (Leptodactylidae), Odontophrynus americanus time lag 2 p < 0.001, Fig. 3A).
(Odontophrynidae), Oreobates discoidalis (Craugastoridae), Rhinella The GLM that best tted the abundance data of amphibian
arenarum and Rhinella schneideri (Bufonidae), Phyllomedusa sau- populations of the ve most abundant species was the one that
vagii, Scinax fuscovarius (Hylidae). L. chaquensis, L. mystacinus and combined the binomial model that included minimum tempera-
P. tucumanum were only captured in the croplands. Oreobates dis- ture and the count model (Poisson distribution) that included all
coidalis was captured mainly in the forest and only occasionally in the proposed variables: habitat type, sampling year, mean tem-
the crops (Appendix 3). perature and rainfall, and agricultural practices (Table 2). The best
The species composition strongly varied between seasons GLM of O. americanus, did not include agricultural practices in the
(autumnewinter and springesummer) (Fig. 1, axis 1, Kruskal count model. The presence of the ve most abundant species was
Wallis H 23.82, p < 0.001). Although less marked, there were positively associated with the minimum temperature. Cultivated
signicant compositional differences between croplands and for- patches presented higher abundance than forests, and the magni-
ests (Fig. 1, axis 2, H 14.68, p < 0.001). The segregation between tude of the effect of sugarcane on the abundance of R. arenarum and
the crops and the forest was more marked during the spring and O. americanus was higher than that of lemon crops. The effect of
summer period than in the autumn and winter period. The lemon crops was more important in the increase of abundance of
segregation of forest sites between seasons was more marked than P. borellii, P. biligonigerus and L. latinasus. The second and third year
that of the crops. No segregation of sites based on species showed a decrease in the abundance per sampling, being the effect
composition was observed between the sampling years (Fig. 1, of the third year bigger in all the cases. The mean temperature and
Appendix 4). rainfall were positively associated with amphibian abundance.
The GLM that best tted the total abundance and species rich- When we considered one subsequent sampling to the agricultural
ness data of amphibian assemblages was the one which combined practice, the short term response of abundance to herbicides
the binomial model that included minimum temperature and the application was negative and the response to fertilizers was
82
oz / Acta Oecologica 69 (2015) 78e86
C. Guerra, E. Ara

Fig. 2. The explanatory variables that best explained the presence and abundance of amphibians of a series of generalized linear models (GLM). Normalized point estimate (Z value)
of every explanatory variable in the zero inated Poisson model that best explained the presence and abundance of amphibians. Data correspond to 114 observations of three
habitat types through three sampling years. The represented model obtained the lowest AIC value in a contest of 45 zero inated Poisson models which combined three binomial
and 15 count components. The two rst bars correspond to Z value of the explanatory variable of the binomial component (intercept and minimum temperature or rainfall). The rest
of the bars correspond to Z value of explanatory variables of the count component. A: amphibians assemblage, B: Leptodactylus latinasus, C: Odontophrynus americanus, D: Phys-
alaemus biligonigerus, E: Pleurodema borellii, F: Rhinella arenarum.

positive in all the cases. The short term response of R. arenarum agricultural practice we observed different patterns. After the ap-
abundance to mechanical interventions and insecticides applica- plications of insecticides there was a decrease in abundance of
tion was positive, while those of P. borellii, P. biligonigerus and L. latinasus and O. americanus (Fig. 3B, C). After the application of
L. latinasus were negative (Fig. 2B to F, Appendix 5 and 6). herbicides there was a decrease in abundance of O. americanus,
Taking into account until three subsequent samplings to the P. biligonigerus and P. borellii (Fig. 3D to F). In most cases, the

oz / Acta Oecologica 69 (2015) 78e86
C. Guerra, E. Ara 83

Fig. 3. The abundance of amphibians in a time window that included different agricultural practices. The short term responses of abundance were obtained through superposed
epoch analyses (SEA) which average the information of samplings within the observation windows. The observation window ranges from 1 sampling before the agricultural practice
(time lag 1) to three samplings (time lag 1, 2 and 3) after the agricultural practice (time lag 0). Condence intervals are generated through 10000 re-samplings which randomize
the date of the agricultural practice. A to C: Decrease in abundance of amphibians assemblage, Leptodactylus latinasus and Odontophrynus americanus after the insecticides
application. D to F: Decrease in abundance of O. americanus, Physalaemus biligonigerus and Pleurodema borellii after the herbicides application.

response of abundance was observed two samplings after the 4. Discussion

intervention, which was not evaluated through the GLMs (SEA time
lag 2, p < 0.001). Habitats with human intervention surveyed in this study lodged
84
oz / Acta Oecologica 69 (2015) 78e86
C. Guerra, E. Ara

higher species richness and abundance of amphibians than sec- uctuations could be as diverse as variations in soil humidity, in the
ondary forests. This pattern could be explained by the presence of hydroperiod or in the availability of food and shelter places. How-
depressions on the soil in the cultivated patches. In these de- ever, we cannot discard the existence of a decline trend, which
pressions temporary ponds were formed, which were used by could be explained by global factors (Collins and Storfer, 2003) or
L. chaquensis, P. biligonigerus, P. borellii, R. arenarum and local factors such as agricultural practices, whose long term effects
S. fuscovarius for their reproduction. In contrast, in forest patches were not assessed in this study.
there were only ephemeral streams after heavy rains, which could The responses of abundance and richness of amphibians to
not be used for reproduction. Our results are supported by some agricultural practices show the existence of an effect of these dis-
studies that evidenced a positive association between diversity of turbances on amphibians. An other evidence of the effects that
amphibians and agricultural areas (e.g. Attademo et al., 2005; habitat types can exert on amphibian assemblages was the absence
Kolozsvary and Swihart, 1999) because they can reproduce in of O. discoidalis (Craugastoridae) in the croplands. This species
wetlands, ponds and pools of disturbed habitat (Prado and Rossa- presents direct development and inhabits in the leaf litter from
Feres, 2014). On the other hand, higher abundance of amphibians mountain forest (Lavilla et al., 2010). Their terrestrial reproduction
could be associated with their diet since other studies demon- demands high humidity to avoid the desiccation of eggs (Suazo
strated that they are great consumers of arthropods pest in other Ortun~ o et al., 2008). Due to their lifestyle and reproductive mode,
croplands (e.g. Attademo et al., 2007a, 2007b; Peltzer et al., the habitat degradation by agricultural expansion constitutes a
2010). threat to this sensitive species (Lavilla et al., 2010). In other studies,
At assemblage level and landscape scale, the contribution of the the abundance and richness of Craugastoridae family were higher
forest to the overall specic composition may be more signicant in forest than in disturbed habitats (e.g. Urbina Cardona and
than any of the crops. The loss of pristine areas could entail the loss London ~ o, 2003; Zorro Ceron, 2007). Pearman (1997) propose that
of specialist taxa with the consequent homogeneization of the some species from this family could be bioindicators of tropical
assemblage. For example, species that thrive in mature forests such humid forest quality, because they thrive under a specic combi-
as O. discoidalis, are put at risk while opportunistic species such as nation of environmental conditions.
L. chaquensis, L. mystacinus and P. tucumanum, selected both crops. Our study supports the idea that environmental heterogeneity
These opportunistic species inhabit on the soil, usually near pools generated by the combination of natural and cultivated patches,
and ooded areas; they are common in disturbed environments allows the coexistence of species related to either kind of envi-
(Heyer et al., 2004, 2010; Lavilla and di Tada, 2004) and they ronment, increasing biodiversity at landscape level (agro-
typically inhabit the Chaco region. Their invasion in the Yungas ecosystem). This is due to the environmental heterogeneity which
would be associated with the advance of the agricultural frontier in implies a variety of habitats and resources that allow the coexis-
this region. On the other hand, L. latinasus, O. americanus, tence of species with different environmental requirements
P. biligonigerus, P. borellii and R. arenarum were most generalist in (Benton et al., 2003), including amphibians (e.g. Rodrigues da Silva
their habitat selection. Various studies have reported a similar et al., 2012a, 2012b), and can give stability to the amphibian
pattern of spatial distribution in habitat specialists and habitat assemblage (Guerry and Hunter, 2002). On the contrary, numerous
generalists frogs (e.g. Dash and Mahanta, 1993; Peltzer et al., 2006; studies report lower species richness or biological diversity in
Urbina Cardona and London ~ o, 2003). monocultures or in those agroecosystems where environmental
Amphibian assemblages and populations responded in a similar heterogeneity is scarce (e.g. Hole et al., 2005; Stoate et al., 2001). In
way to the climatic variables, showing the tight relationship be- turn, the internal regulation of natural ecosystems occurs through
tween temperature and rainfall with their history of life (Prado complex uxes of energy and nutrient (Altieri and Nicholls, 2000).
et al., 2005) and physiology (Pounds et al., 2006), independently These autoregulation mechanisms could be lost with agricultural
of habitat conditions (Appendix 6). This could be due to the com- intensication and landscape simplication associated with
bination of the ectothermal condition of amphibians and the highly monocultures, turning them into ecologically vulnerable environ-
seasonal climate that dominates the study area. Other studies ments and dependent on high agrochemicals levels (Altieri and
performed in other areas with seasonal climate have found similar Nicholls, 2000). Moreover, our study reinforce the importance of
relationships between amphibians and climatic factors (e.g. Conte small forest fragments to amphibians diversity, because they pro-
and Rossa Feres, 2007; Parris, 2004). In fact, the composition of vide resources that are critical for amphibians, such as habitat
species differed between the autumnewinter and the spring- refuge and dispersal corridors, and preserve hydrologic regimes
esummer periods, and higher richness and abundance of am- and water quality (Rodrigues da Silva et al., 2011).
phibians in the springesummer period than in the autumnewinter
period were recorded in previous analyses which were not reported Acknowledgments
here. The relationship of amphibian abundance to the mean tem-
perature and rainfall could be due to the inuence of these vari- We would like to thank different persons and institutions that
ables in their reproductive activity and in the abundance and facilitated and improved this study. R. Montero, P. Blendinger, R.
availability of their preys (Martori et al., 2005; Prado et al., 2005; Lajmanovich and M. Halloy provided academic advice during the
Sanchez et al., 2007). study. R. Montero, P. Blendinger and the anonymous reviewers
Although no compositional shift was observed between sam- made useful suggestions on previous version of this manuscript;
pling years (Fig. 1), a marked decrease in the overall abundance was Engineer A. Garca was the administrator of the Lules Farm who
reported (Table 2). This trend could be explained by a long term authorized us to carry out the eld survey on behalf of the owners
decline or by natural interannual uctuations which are frequent in (Nougues S.A.) and provided meteorological records and informa-
amphibian populations (Storfer, 2003). These uctuations have tion of the agricultural practices; Engineer A. Valeiro from INTA
been attributed to different factors, mainly to the variation of cli- Famaill
a provided detailed meteorological records and biblio-
matic conditions due to their relationship with the amphibian graphic material; M.J. Salica, F. Zaffanella, J.L. Galvan, J.C. Stazzonelli,
reproductive activity (Prado et al., 2005). Although the mean S. Marinaro, P. Presti, M.E. and L. Guerra, P. Zelaya, C. Ju
arez Vocos,
temperatures and rainfall during the three years of study were M. Sandoval, G. Rodrguez, M. Ayup, R. Seitz, D. Baldo, R. Medina
similar, the occurrence of inter or intraannual variations in the and M.L. Ponssa assisted in the eld data collection; F. Zaffanella
microclimatic conditions should not be discarded. Such improved appendix 2; A. Malizia helped with the NMDS; L. Macchi

oz / Acta Oecologica 69 (2015) 78e86
C. Guerra, E. Ara 85

improved Fig. 1. CG acknowledges CONICET for a doctoral and Instituto Nacional de Tecnologa Agropecuaria, Estacio
n Experimental Agropecuaria
Famailla
, 2011. Database: Sntesis Agrometeorolo
gica. http://anterior.inta.gov.
postdoctoral grant; CIUNT (G315, 430 and 519), Idea Wild Foun-
ar/f/?urlhttp://anterior.inta.gov.ar/famailla/info/sintesis_agromet.htm/
dation and CONICET (PIP 1112008010 2422 and PIP (accessed 13.05.11.).
11220110100875) for nancial support; and director of the Instituto International Union for Conservation of Nature and Natural Resources, 2014.
de Herpetologa FML, E. Lavilla, for providing a nice place to work Announcing the IUCN Red List of Threatened Species. Available. http://www.
iucnredlist.org/initiatives/amphibians/analysis/ (accessed 3.05.14.).
and bibliographic materials, and for allowing access to the database Jackman, S., 2011. Pscl: Classes and Methods for R Developed in the Political Science
of Herpetology Collection. EA acknowledges Fundacio
n Bunge & Computational Laboratory. Stanford University, California. Available. http://pscl.
Born for a postdoctoral grant. stanford.edu.
Johansson, M., Primmer, C.R., Sahlsten, J., Merila, J., 2005. The inuence of landscape
structure on occurrence, abundance and genetic diversity of the common frog,
Appendix A. Supplementary data Rana temporaria. Glob. Chang. Biol. 11, 1664e1679.
Knutson, M.G., Richardson, W.B., Reineke, D.M., Gray, B.R., Parmelee, J.R., Weick, S.E.,
2004. Agricultural ponds support amphibian populations. Ecol. Applic. 14 (3),
Supplementary data related to this article can be found at http:// 669e684.
dx.doi.org/10.1016/j.actao.2015.09.003. Kolozsvary, M.B., Swihart, R.K., 1999. Habitat fragmentation and the distribution of
amphibians: patch and landscape correlates in farmland. Can. J. Zool. 77,
1288e1299.
References Kruskal, J.B., Wish, M., 1978. Multidimensional scaling. In: Uslaner, E.M. (Ed.), Sage
University Papers Series on Quantitative Applications in the Social Sciences.
Altieri, M., Nicholls, C.I., 2000. Teora y pra
ctica para una agricultura sustentable. In: Sage Publications, California, pp. 7e11.
Red de formacio
n ambiental para Ame
rica Latina y el Caribe. Programa de las Lavilla, E.O., 2001. Amenazas, declinaciones poblacionales y extinciones en anbios
Naciones Unidas Para el Medio Ambiente, Me
xico DF, p. 257. argentinos. Cuad. Herpetol. 15 (1), 59e82.
Attademo, A.M., Peltzer, P.M., Lajmanovich, R.C., 2005. Amphibians occurring in Lavilla, E.O., di Tada, I., 2004. Pleurodema tucumanum. Database: the IUCN Red List
soybean and implications for biological control in Argentina agricsulture. Agric. of Threatened Species. http://www.iucnredlist.org/details/57292/0 (accessed
Ecosyst. Environ. 106, 389e394. 3.05.14.).
Attademo, A.M., Cejas, W., Peltzer, P.M., Lajmanovich, R.C., 2007a. Phenology in the Lavilla, E.O., Cortez, C., Reichle, S., De la Riva, I., Ko hler, J., 2010. Oreobates Dis-
diet of Chaunus arenarum (Anura: Bufonidae) in a soybean eld of Co
rdoba coidalis. Database: the IUCN Red List of Threatened Species. http://www.
province, Argentina. Rev. Esp. Herpetol. 21, 41e48. iucnredlist.org/details/56564/0 (accessed 3.05.14.).
Attademo, A.M., Peltzer, P.M., Lajmanovich, R.C., 2007b. Feeding Habits of Phys- Legendre, P., Legendre, L., 2012. Numerical Ecology. Elsevier, Amsterdam, p. 990.
alaemus biligonigerus (Anura, leptodactylidae) from soybean eld of Co
rdoba Mann, R.M., Hyne, R.V., Choung, C.B., Wilson, S.P., 2009. Amphibians and agricul-
province, Argentina. Russ. J. Herpetol. 14 (1), 1e6. tural chemicals: review of the risks in a complex environment. Environ. Pollut.
Benton, T.G., Vickery, J.A., Wilson, J.D., 2003. Farmland biodiversity: is habitat het- 157 (11), 2903e2927.
erogeneity the key? Trends Ecol. Evol. 18, 182e188.
nez, C., 2005. Temporal variation and size
Martori, R., Aun, L., Gallego, F., Rozzi Gime
Bishop, C.A., Mahony, N.A., Struger, J., N.G., P., Pettit, K.E., 1999. Anuran develop- class distribution in a herpetological assemblage from Co
rdoba, Argentina.
ment, density and diversity in relation to agricultural activities in the Holland Cuad. Herpetol. 19 (1), 35e52.
river watershed, Ontario, Canada (1990-1992). Environ. Monit. Assess. 57 (1), McCune, B., Mefford, M.J., 1999. PC-ORD Multivariate Analysis of Ecological Data
21e43. Version 5.0. MjM Software, Oregon.
Brown, A.D., Malizia, L.R., 2004. Las selvas pedemontanas de las Yungas. Cienc. Hoy Parris, K.M., 2004. Environmental and spatial variables inuence the composition of
14 (83), 53e63. frog assemblages in sub-tropical eastern Australia. Ecography 27, 392e400.
Collins, J.P., Storfer, A., 2003. Global amphibian declines: sorting the hypotheses. Pearman, P.B., 1997. Correlates of amphibian diversity in an altered landscape of
Divers Distrib. 9, 89e98. Amazonian Ecuador. Conserv. Biol. 11, 1211e1225.
Conte, C.E., Machado, R.A., 2005. Riqueza de espe
cies e distribuia ~o espacial e Peltzer, P.M., Lajmanovich, R.C., Beltzer, A.H., 2003. The effects of habitat frag-
temporal em comunidade de anuros (Amphibia, Anura) em uma localidade de mentation on amphibian species richness in the foodplain of the middle Parana
Tijucas do Sul, Parana
, Brasil. Rev. Bras. Zool. 22 (4), 940e948. river, Argentina. Herpetol. J. 13, 95e98.
Conte, C.E., Rossa Feres, D.C., 2007. Riqueza e distribuia ~o espao-temporal de Peltzer, P.M., Lajmanovich, R.C., Attademo, M.A., Beltzer, A.H., 2006. Diversity of
anuros em um remanescente de Floresta de Arauca
ria no sudeste do Parana
. anurans across agricultural ponds in Argentina. Biodiv Conserv. 15, 3499e3513.
Rev. Bras. Zool. 24, 1025e1037. Peltzer, P.M., Attademo, M.A., Lajmanovich, R.C., Junges, C.M., Beltzer, A.H.,
Corn, P.S., 2001. Cercas en lnea recta y trampas de pozo. In: Heyer, W.R., Sanchez, L.C., 2010. Trophic dynamics of three sympatric anuran species in a
Donelly, M.A., McDiarmid, R.W., Hayek, L.C., Foster, M.S. (Eds.), Medicio
n y soybean agroecosystem from Santa Fe Province, Argentina. Herpetol. J. 20 (4),
monitoreo de la diversidad biolo
gica: Me
todos estandarizados para anbios. 261e269 (9).
Smithsonian Institute Press, Washington, pp. 104e113. Pineda, E., Halffter, G., 2004. Species diversity and habitat fragmentation: frogs in a
Dash, M.C., Mahanta, J.K., 1993. Quantitative analysis of the community structure of tropical montane landscape in Mexico. Biol. Conserv. 117, 499e508.
tropical amphibian assemblages and its signicance to conservation. J. Biosci. 8 Pounds, L.A., Bustamante, M.R., Coloma, L.A., Consuegra, J.A., Fogden, M.P.L.,
(1), 121e139. Foster, P.N., La Marca, E., Masters, K.L., Merino Viteri, A., Puschendorf, R.,
Gelman, A., Hill, J., 2007. Data Analysis Using Regression and Multilevel/hierarchical Ron, S.R., S
anchez Azofeifa, G.A., Still, C.J., Young, B.E., 2006. Widespread
Models. Cambridge University Press, New York, p. 651. amphibian extinctionsfrom epidemic disease driven by global warming. Nature
Gonza
lez Sicilia, E., 1960. El cultivo de los agrios. Instituto Nacional de Inves- 439, 161e167.
tigaciones Agrono
micas, Valencia, p. 805. Prado, C.P., Uetanabaro, M., Haddad, C.F.B., 2005. Breeding activity patterns,

n, estructura y dina
Grau, H.R., Paolini, L., Malizia, A., Carilla, J., 2010. Distribucio
mica reproductive modes, and habitat use by anurans (Amphibia) in a seasonal
de los bosques de la sierra de San Javier (Tucuma
n, Argentina). In: Grau, H.R. environment in the Pantanal, Brazil. Amphib-reptil 26, 211e221.
(Ed.), Ecologa de una interfase natural-urbana: la sierra de San Javier y el Gran Prado, V.H.M., Rossa-Feres, D.C., 2014. Multiple determinants of Anuran richness
San Miguel de Tucum
an. EDUNT, Tucuma
n, pp. 33e50. and occurrence in an agricultural region in South-eastern Brazil. Environ.
Guerry, A.D., Hunter, M.L., 2002. Amphibian distributions in a landscape of forests Manag. 53, 823e837.
and agriculture: an examination of landscape composition and conguration. R Development Core Team, 2011. R: a Language and Environment for Statistical
Conserv. Biol. 16 (3), 745e754. Computing. R Foundation for Statistical Computing, Vienna. Available. http://
Haurwitz, M.W., Brier, G.W., 1981. A critique of the superposed epoch analysis www.R-project.org.
method: its application to solar-weather relations. Mon. Weather Rev. 109, Rodrigues da Silva, F., Prado, V.H.M., Rossa-Feres, D.C., 2011. Value of small forest
2074e2079. fragments to amphibians. Science 332, 1033.
Hazell, D., Cunningham, R., Lindenmayer, D., Mackey, B., Osborne, W., 2001. Use of Rodrigues da Silva, F., Panin Candeira, C., Rossa-Feres, D.C., 2012a. Dependence of
farm dams as frog habitat in an autralian agricultural landscape: factors anuran diversity on environmental descriptors in farmland ponds. Biodiv
affecting species richness and distribution. Biol. Conserv. 102, 155e169. Conserv. 21 (6), 1411e1424.
Heyer, R., Reichle, S., Silvano, D., Lavilla, E.O., di Tada, I., 2004. Leptodactylus cha- Rodrigues da Silva, F., Oliveira, T.A.L., Gibbs, J.P., Rossa-Feres, D.C., 2012b. An
quensis. Database: the IUCN Red List of Threatened Species. http://www. experimental assessment of landscape conguration effects on frog and toad
iucnredlist.org/details/57118/0 (accessed 3.05.14.). abundance and diversity in tropical agro-savannah landscapes of southeastern
Heyer, R., Silvano, D., Reichle, S., Lavilla, E.O., di Tada, I., 2010. Leptodactylus mys- Brazil. Land Ecol. 27 (1), 87e96.
tacinus. Database: the IUCN Red List of Threatened Species. http://www. Sanchez, L.C., Peltzer, P.M., Manzano, A.S., Lajmanovich, R.C., 2007. Din
amica de un
iucnredlist.org/details/57147/0 (accessed 3.05.14.). ensamble de anuros en un humedal del tramo inferior del ro Parana
, Argentina.
Hole, D.G., Perkins, A.J., Wilson, J.D., Alexander, I.H., Grice, P.V., Evans, A.D., 2005. Interciencia 32, 463e470.
Does organic farming benet biodiversity? Biol. Conserv. 122, 113e130. Sesma, P.J., Guido, E.Y., Puchulu, M.E., 2010. Geologa de la ladera oriental de la
Houlahan, J.E., Findlay, S., Schmidt, B.R., Meyer, A.H., Kuzmin, S.L., 2000. Quantita- sierra de San Javier, soporte fsico para la gestio
n territorial. In: Grau, H.R. (Ed.),
tive evidence of global amphibian population declines. Nature 404, 752e755. Ecologa de una interfase natural-urbana: la sierra de San Javier y el Gran San
Humbert, R.P., 1974. El cultivo de la can ~ a de azcar. Compan ~ a Editorial Continental Miguel de Tucuma
n. EDUNT, Tucum
an, pp. 19e32.
S.A, Me
xico DF, p. 719. Stoate, C., Boatman, N.D., Borralho, R.J., Carvalho, C.R., de Snoo, G.R., Eden, P., 2001.
86
oz / Acta Oecologica 69 (2015) 78e86
C. Guerra, E. Ara

Ecological impacts of arable intensication in Europe. J. Environ. Manage 63, Wake, D.B., Vredenburg, V.T., 2008. Are we in the midst of the sixth mass extinc-
337e365. tion? A view from the world of amphibians. Proc. Natl. Acad. Sci. U. S. A. 105,
Storfer, A., 2003. Amphibian declines: future directions. Divers Distrib. 9, 151e163. 11466e11473.
Suazo Ortun ~ o, I., Alvarado Daz, J., Martnez Ramos, M., 2008. Effects of conversion Wells, K.D., 2007. The Ecology and Behavior of Amphibians. The University of Chi-
of dry tropical forest to agricultural mosaic on herpetofaunal assemblages. cago Press, Chicago and London, p. 1148.
Conserv. Biol. 22 (2), 362e374. Wenger, S.J., Freeman, M.C., 2008. Estimating species occurrence, abundance, and
Urbina Cardona, J.N., London ~ o, M.C., 2003. Distribucio
n de la comunidad de her- detection probability using zero-inated distributions. Ecology 89, 2953e2959.
petofauna asociada a cuatro a
reas con diferente grado de perturbacio
n en la Isla Zhao, J., Neher, D., 2013. Soil nematode genera that predict specic types of
Gorgona, Pacco colombiano. Rev. Acad. Colomb. CEFN 27, 105e113. disturbance. Appl. Soil Ecol. 64, 135e141.
Urbina Cardona, J.N., Olivares Pe
rez, M., Reynoso, V.H., 2006. Herpetofauna diversity Zorro Ceron, J.P., 2007. Anuros de piedemonte llanero: diversidad y preferencia de
and microenvironment correlates across a pasture-edge-interior ecotone in microha
bitat. Tesis de Grado. Facultad de Ciencias, Ponticia Universidad Jav-
tropical rainforest fragments in the Los Tuxtlas biosphere reserve of Veracruz, eriana de Bogota
. Available. http://www.javeriana.edu.co/biblos/tesis/ciencias/
Mexico. Biol. Conserv. 132, 61e75. tesis88.pdf/ (accessed 17.05.14.).