Introduction 

Patterns of wildlife distribution and abundance have suffered alterations worldwide due to human occupation and related activities ([55]). The agriculture expansion has replaced the original vegetation and fragmented primary environments, thus affecting wildlife patterns of habitat use ([22], [15]). However, studies in agricultural landscapes have shown the use of man-made habitats by the wildlife, such as coffee plantations in Mexico ([38]), banana and coconut plantations in Costa Rica ([24], [23]), cocoa and Eucalyptus plantations in Brazil ([16], [32]) and subsistence agriculture in Nepal ([2]).

A great spatial heterogeneity due to different strata and structural complexity exists in native forests, which results in a significant β-diversity ([37], [44]). In an agricultural landscape, a considerable loss of biodiversity occurs where monocultures are spatially more homogenous. However, the environment of Eucalyptus plantations varies dramatically along their commercial cycle. At the first stage, it assumes a bushy structure that later evolves into a forest, which is totally clear-cut at the end of 6-7 years ([56], [61]). Therefore, these artificial forests present great temporal heterogeneity which may affect patterns of abundance and distribution of resident wildlife, not only for small mammals ([34]), but also for their predators ([60]). In fact, recent studies show these forests are capable of maintaining a resident wildlife ([46], [32], [13], [34]). This study aims at assessing the temporal variation of the specific composition and relative abundance of middle to large-sized mammal species found in Eucalyptus plantations in the State of São Paulo, Southeastern Brazil.

Eucalyptus plantations currently occupy over one million hectares in São Paulo, the richest and most developed state in Brazil. Their main commodities are the pulp and paper ([1]). Together with livestock production, the ethanol agroindustry primarily based on sugarcane and the pulp and paper silvicultural industry primarily based on Eucalyptus are the main agroindustrial sectors of São Paulo state. However, unlike the other two agroindustrial sectors, pulp and paper industries follow Brazilian environmental laws relatively well, due to market pressure posed by certification organizations such as the Forest Stewardship Council (FSC - [59]). Accordingly, the assessment of the distribution of middle to large-sized mammals in Eucalyptus plantations may give significant advances to both the environmental certification process of this sector and the elaboration of public policies which can result in the inclusion of such landscapes in the context of wildlife conservation.

  Materials and Methods 

Study Area

This study was conducted at three Eucalyptus commercial plantations located between Barra Bonita and Jurumirim reservoirs, Paranapanema River Basin, State of São Paulo, Southeastern Brazil (Fig. 1). The native vegetation cover in the region was composed of “Cerradão” (forested savannah) and semidecidual Atlantic forest ([57]). The climate is tropical (Köppen) with rainy summers and dry winters with average temperatures between 19.7 and 21.5 °C ([7]).

Fig. 1 - Location of study sites in Brazil and São Paulo State and the distribution of sampling sites on Eucalyptus stands in the plantation farms assessed in this study. (A): zero to one year old stands; (B): two to three year old stands; (C): four to five year old stands; (D): six to seven year old eucalypt stands.

  Enlarge/Shrink   Download   Full Width  Open in Viewer

The historical use of these lands dates back to the 18^th century with the deforestation for pasture establishment. In the late 19^th century, coffee plantations dominated the region, being replaced by sugarcane in the early 20^th century. At the end of the 20^th century, sugarcane began to be gradually replaced by Eucalyptus plantations, which currently covers approximately 40% of the Alto Paranapanema river basin, with only about 10% of remaining native vegetation left ([28]).

Sampling methodology

The sampling period of this study was from August 2007 to February 2008, including seven two-day campaigns carried out every month. Sand plot for track counts ([62], [42], [48]) and camera-traps ([26], [27]) were used to detect middle to large-sized mammals. Sand plots were cleaned and camera-traps activated on the first day of the campaign, and data collection took place on the second day. The study area was divided into four age ranges of Eucalyptus stands: (i) 0-1 year; (ii) 2-3 years; (iii) 4-5 years; and (iv) 6-7 years (Fig. 1).

The surveys within sand plots took place using trails which were naturally formed by the space between the Eucalyptus tree rows, starting at the gravel roads that divided the stands. Sand plots were placed every 25 meters along each trail, summing up ten sand plots per trail. Four trails were established for each Eucalyptus stand age range. Each trail was considered as a sample unit, totaling 16 sample units. The trails where the sand plots were placed received one camera-trap each, placed 75 meters from the trail entrance. The minimal distance between sampling units was one kilometer. As we considered frequency of occurrence as the dependent variable, independently of the actual species abundances, we could consider the sampling units as independent.

Analytical methodology

Mammals registered in the field surveys were grouped in the following trophic categories (adapted from [49], [50], [20], [11], [12]): (a) carnivores; (b) herbivores; (c) omnivores and (d) insectivores. A list of detected species and their frequency of occurrence have been determined (taxonomic classification followed [63]). Sampling sufficiency was assessed by species incidence curves, with 100 randomizations in a rarefaction procedure (Sobs) by the Mao Tao method ([8]). Specific richness was estimated by the Bootstrap method ([53]). Species incidence curves were adjusted according to the Pearl-Reed trend model ([18]).

Data was tested for normality and homoscedasticity by Kolmogorov-Smirnoff test ([30]) and Levene’s test ([29]), respectively. Eucalyptus stand age ranges were mutually compared regarding species richness and frequency of occurrence for all species, and by trophic categories by Analysis of Variance (ANOVA - [25]) and Analysis of Means (ANOM - [40]). Venn diagrams ([39]) were used for the detection of overlaps among distinct stand ages regarding the species and trophic categories.

  Results 

A total of 15 species of middle to large-sized mammals of 10 families was detected, including the following exotic species: cattle (Bos taurus), domestic dog (Canis lupus (= familiaris)) and the European hare (Lepus spp. - Tab. 1). This represents 86% of the total estimated species richness by Bootstrap, varying from 75.9% in 6-7 years-old stands to 90.6% in 2-3 years-old stands (Tab. 2). The following trophic categories have been detected: herbivores, carnivores, omnivores and insectivores. Their detected and estimated species richness varied from 71.5% (insectivores in 0-1 and 6-7 years-old and carnivores in 0-1 year-old stands) to 95.2% (omnivores in 2-3 years-old stands - Tab. 2).

Data set was homoscedastic and normally distributed (p < 0.05). There was a significant variation in total species richness (F=6.81, df=3, p=0.006) and frequency occurrences (F=7.07, df=3, p=0.005), in relation to plantation age. However, such variation has only been detected in omnivores (F=9.33, df=3, p=0.002; F=9.95, df=3, p=0.001), respectively, for species richness and frequency of occurrence (Fig. 2).

Fig. 2 - Analysis of Means (α<0.05), according to Ott ([40]) for species richness and frequency of occurrence for the whole species group and for omnivorous species by distinct Eucalyptus stand ages.

  Enlarge/Shrink   Download   Full Width  Open in Viewer

Seven species occurred exclusively at a single stand age: Leopardus tigrinus and Pecari tajacu at 0-1 year-old; Sylvilagus brasiliensis at 2-3 year-old; and Euphractus sexcinctus, Nasua nasua, Eira barbara and Puma concolor at 4-5 year-old stands. Two species occurred in all stand ages: Myrmecophaga tridactyla and Dasypus novemcinctus. All trophic categories, with the exception of carnivores (absent in 2-3 and 6-7 year-old stands) occurred at all Eucalyptus stand ages (Fig. 3).

Fig. 3 - Venn’s Diagrams showing overlaps of species (A) and trophic categories (B) occurrence in Eucalyptus plantations of distinct ages. (Dn): Dasypus novemcinctus; (Es): Eupractus sexcinctus; (Mt): Myrmecophaga tridactyla; (Lsp): Lepus sp.; (Sb): Sylvilagus brasiliensis; (Pt): Pecari tajacu; (Msp): Mazama sp.; (Bt): Bos taurus; (Lp); Leopardus tigrinus; (Pc): Puma concolor; (Clf): Canis lupus familiaris; (Ct): Cerdocyon thous; (Cb): Chrysocyon brachyurus; (Eb): Eira barbara; (Nn): Nasua nasua; (Omn): Omnivores; (Carn): Carnivores; (Herb): Herbivores; (Inset): Insectivores.

  Enlarge/Shrink   Download   Full Width  Open in Viewer

  Discussion 

The total species composition found in this study is similar to other studies carried out in the same region ([52], [6], [3]). Although relatively small, the present sampling effort was sufficient to detect 86% of the total estimated number of species of middle to large-sized mammals that use local Eucalyptus plantations in central-south state of São Paulo, in south-eastern Brazil. However, methodological bias for terrestrial species possibly prevented the detection of primates ([51]), as well as the capybara (Hydrochoerus hydrochaeris - [58]) and the Neotropical otter (Lontra longicaudis - [41], [43]), respectively arboreal, semi-aquatic and aquatic species.

The manned-wolf (Chrysocyon brachyurus) and the crab-eating fox (Cerdocyon thous) were the most abundant species at the present study. Both species are omnivorous with generalist diet including plant matter, invertebrates and small vertebrates. Such a diet provides them with a wide feeding niche, allowing them to feed on different resources according to availability ([14]). Armadillos (Dasypus spp.) and giant ant-eaters (Myrmecophaga tridactyla) were also abundant, possibly for eating insects (in particular ants and termites) usually abundant in Eucalyptus plantations ([5], [35]). As a matter of fact, ants are considered the most important “plague” of Eucalyptus plantations in southeastern Brazil, demanding a rigorous control by the pulp and paper industry ([47]). In addition to ants and termites, other opportunistic invertebrates and small vertebrate species can be found in Eucalyptus plantations possibly due to its relatively fast growth (i.e., biomass production - [4], [45], [36]). These species are possibly food resources for middle to large-sized mammals ([9], [54]).

Herbivores, in especial the European hare (Lepus spp.) are relatively abundant in the early stages of Eucalyptus plantations (0-1 year), possibly favored by the abundance of remaining grass still present as the former landscape matrix was formed by exotic pasture land (predominantly Brachyarya = Urochloa spp.). However, later (2-5 years), omnivores and insectivores become more numerous possibly because of the presence of some understory vegetation used by arthropods and small vertebrates. However, in the final stage of commercial Eucalyptus stands (6-7 years) all trophic categories drastically decrease in number of species and frequency of occurrence. At this stage mechanical trimmings become more frequent, resulting in the removal of the understory vegetation in order to facilitate harvest ([10], [61]). Such intense weed control is usually overdone in order to avoid legal problems with local environmental agencies as even an economically viable understory vegetation can be considered as native revegetation that requires especial license to be cut (see Lei No. 11.428, from 22^nd December 2006; Resolução CONAMA No. 338, from 23 February 2007; Resolução CONAMA No. 392, from 25 June 2007; Portaria DEPRN No. 08, from 20 November 1989 and Decreto Estadual 53.027-2008). Such management practice seems rather ineffective for both the pulp and paper industry and the environmental quality as the maintenance of some understory vegetation is undoubtedly better for the wildlife than no vegetation at all. In addition, a less intense weed control could result in cost reduction for the pulp and paper industry and its consumers.

  Conclusions 

The detected temporal variation in the use of Eucalyptus stands by middle to large-sized mammals is possibly related to the relatively high temporal heterogeneity of this environment as well as to the silvicultural practices along its commercial cycle ([19]). According to the “continuum model” ([33], [17], [19], [31]) animal species are distributed as a continuum in the landscape, spatially and temporally, according to the distribution of resources in a pattern that resembles a dégradé more than on a well-defined patchwork pattern. In such model, species will move, use, and therefore, occur on a determined environment at a certain moment depending on its resources availability and the species’ ability to access them ([33]). The present species are relatively common in silvicultural landscapes. However, their use of resources associated with the landscape matrix (i.e., Eucalyptus plantations themselves) are directly associated to the silvicultural practices along commercial plantation cycles. Such principle should be considered by the certification process of the pulp and paper industry ([21]) in order to promote and maintain a sustainable conservation value for silvicultural landscapes in south-eastern Brazil.

  Acknowlegdements 

We would like to thank Júlio César Costa, Gustavo Sigrist Betini, Sabrina Koester-Gobbo and other colleagues for the assistance in the field. This study received financial support from FAPESP (Proc. No. 2006/60954-4), CNPq (Proc. No. 142732/ 2007-9) and Eucatex S/A. LMV has a productive research fellowship from CNPq.

  References