Introduction

The control and prevention of rodent-borne diseases largely depends upon understanding the ecology of their hosts, including its spatial and temporal distribution (Mills et al. 1999). Hantavirus pulmonary syndrome (HPS) is a severe endemic disease of the Americas, characterized by a rapid onset of pulmonary edema, followed by respiratory failure and cardiogenic shock (Jenison et al. 1995; Peters and Khan 2002; Martinez et al. 2010). Transmission between humans and rodents occurs via inhalation of virus material in aerosolized excreta of infected rodents (Peters and Khan 2002). In Argentina, seven native sigmodontine species have been identified as hantavirus hosts and five of them have been associated with hantavirus genotypes that cause HPS in humans (Levis et al. 2004; Padula et al. 2007; Palma et al. 2012).

Several studies on habitat use and selection have highlighted the importance of vegetation on abundance, spatial and temporal distribution, and composition of small rodent communities in HPS-endemic areas of central-east Argentina (Busch et al. 1997; Suárez and Bonaventura 2001; Hodara and Busch 2010). However, most of these studies have been conducted in agro-ecosystems and little is known about the environmental and vegetation factors that influence abundance and distribution of small rodent species in natural environments, such as protected areas.

This research was carried out in two national parks located in the floodplain of the Paraná River, where islands are an important geomorphological element (Drago 1981). As part of the floodplain of the Paraná River, these islands are subject to the influence of hydro-sedimentological pulses of the river (Neiff 1990, 1999), triggering substantial changes in vegetation (Franceschi et al. 2010). Depending on the degree and duration of the flood, small mammals may suffer direct mortality, emigrate to non-flooded areas, or remain in temporary refugia (Williams et al. 2001; Jacob 2003; Wijnhoven et al. 2005). Five sigmodontine species have been previously described in the study area (Vadell et al. 2011), two of which are known hantavirus reservoir species (Levis et al. 2004; Padula et al. 2007; Vadell et al. 2011). These species are O. flavescens, an omnivorous species that inhabits grasslands, forests, and irrigated crop-field borders, and A. azarae, an omnivorous species that inhabits principally grasslands and other habitats with high herbaceous cover (Ellis et al. 1997; Gómez Villafañe et al. 2005, 2012; Andreo et al. 2009).

Human activities in and around the islands of the Paraná delta include sustainable and recreational fishing, cattle farming, crop farming, and eco-tourism. Fishermen and farmers often live in temporary, precarious settlements, which sometimes consist of just a plastic roof with no walls, placed near the shore of rivers and streams. These activities and life style involve a very close contact with the natural environment which increases the risk of contact with rodents. Knowledge of the environmental factors that affect rodent abundance can help predict the population dynamics of hantavirus reservoir species, and be used to guide protective measures to prevent human disease (Mills and Childs 1998).

The aim of this study was to evaluate the influence of environmental variables on the macro- and micro-habitat use of small wild rodents in Pre Delta National Park and Islas de Santa Fe National Park, central-east Argentina. In doing so, we address the following questions:

  • Are all small rodent species influenced by the same environmental variables?

  • Which are the main environmental factors determining patterns of rodent distribution and abundance?

  • Are hantavirus reservoir species restricted to a specific habitat type?

Methods

Study Area

The study was conducted in Pre Delta (PDNP; 32°08′S; 60°38′W) and Islas de Santa Fe (ISFNP; 32°16′S; 60°43′W) national parks located in the floodplain of the Paraná River, Argentina, at the upper section of its delta. PDNP occupies an area of 24.58 km2 consisting of both continental and insular areas, while ISFNP occupies 49.60 km2 of islands located mainly in the main channel of the Paraná River (Fig. 1). The climate is temperate humid, with mean monthly temperatures ranging from 1°C during the coldest month to 38°C during the warmest month. Mean annual precipitation is 900 mm, and it is mostly concentrated between October and April. Almost all of the study area is subjected to periodic flooding pulses resulting from water level changes in the Paraná River and its tributaries. These pulses do not occur regularly throughout seasons and years, and they vary in intensity and duration (Neiff and Malvárez 2004; Neiff 2004). The intensity of each flooding also varies according to the location and altitude of each individual site within the study area (Aceñolaza et al. 2003; Neiff 2004).

Figure 1
figure 1

Pre Delta and Islas de Santa Fe national parks.

Full size image

Small mammals were studied at six sites within the study area (two within ISFNP and four within PDNP):

  1. 1.

    Island 1 (ISFNP) is dominated by mixed forests of Salix humboldtiana and Tessaria integrifolia on the main levees, prairies of grasses on slopes, and broadleaf herbs and grasses around permanent or almost permanent inner ponds in the central part of the island. This island has been heavily grazed by domestic cattle for decades.

  2. 2.

    Island 2 (ISFNP) is similar to Island 1 but with a small farm (a house and a barn with some domestic and barnyard animals).

  3. 3.

    The riparian forest (PDNP) is a forest dominated by S. humboldtiana on the main levees, and Sapium haematospermum, Albizia inundata, and Erythrina crista-galli in the low land. Grasses and forbs form a dense understory.

  4. 4.

    The inland forest (PDNP) is a low-diversity forest dominated by S. humboldtiana and T. integrifolia. Grasses and forbs form a dense understory.

  5. 5.

    The marsh (PDNP) is a low-diversity wetland dominated by Cortaderia selloana and Typha latifolia.

  6. 6.

    The pond margins (PDNP) consist of open forests and shrublands surrounding a permanent inland pond. The understory is dominated by grasses of the genus Panicum, Cortaderia, and Eryngium.

Rodent Survey

At each of the six sites, small rodents were live-trapped once each season (Spring: October–December; Summer: January–March; Autumn: April–June; Winter: July–September) from July 2008 to September 2011 (12 trapping sessions in PDNP and 11 trapping sessions in ISFNP). Trapping was conducted at all sites during each trapping session except for the pond margins (PDNP) which was not included in the trapping design until winter 2009 (only eight trapping sessions), and the marsh which was not sampled during spring 2009 because it was inaccessible due to a flooding event. Between 25 and 50 Sherman live traps (15 × 16 × 31 cm) baited with a mixture of peanut butter, fat, and rolled oats were placed at 10-m intervals along transects in each site for three consecutive nights a season. The area covered by traps at each site varied between 2500 and 5000 m2.

For each animal captured, we recorded species according to external or skull morphology (when necessary). Individuals were tagged with a uniquely numbered ear tag and were released at the point of capture.

In October 2008, all O. flavescens captured on the islands of ISFNP were killed in order to reduce the risk of HPS for fishermen and park workers given the occurrence of a fatal case that same year. In the following trapping sessions, only O. flavescens individuals captured in or around the house and the barn on island 2 (where park workers lived) were removed.

For each trapping session at a site, we calculated a capture rate (CR) as an index of population abundance. CR was estimated as

$$ {\text{CR}} = \left[ {{\text{number of individual captured}}/\left( {{\text{number of traps}} \times {\text{number of active nights}}} \right)} \right] $$

A mark-recapture estimate was not used given the low number of recaptured individuals.

Vegetation, Weather, and Hydrological Characteristics

Vegetation

In order to characterize the habitat used by small rodents, we recorded the characteristics of vegetation and substrate using a 1 × 1 m quadrant placed around each trap station. We recorded the percentage of the quadrant covered by green (GGrass) and dry grasses (DGrass), and green (GBroad) and dry broadleaves (DBroad) up to 1 m tall, and the percentage of bareground (Bare). We also measured maximum vegetation height (Height) at each trap station. The percentage of each quadrant covered by each vegetation type (or bare ground) was assigned to one of five cover-classes (0–12.5%, 12.5–25%, 25–50%, 50–75%, and 75–100%). Maximum vegetation height was assigned to one of four classes (<1 m, 1–2 m, 2–3 m, and >3 m). Vegetation was surveyed once during each trapping session from June 2009 to September 2011.

Weather

Weather and hydrological variables were also used to characterize the habitat used by small rodents. Based on the literature (Suárez and Bonaventura 2001; Williams et al. 2001; Bonaventura et al. 2003; Jacob 2003; Hodara and Busch 2010; Gómez Villafañe et al. 2012), we recorded 16 environmental variables that could potentially influence the abundance of small rodents: mean, maximum, and minimum water level during the month previous to each trapping session (MaxWat, MinWat, and MeanWat); number of months since last flooding event (TimeFlood); mean, maximum, and minimum temperature during the month previous to each trapping session (T, MaxT, MinT); and monthly accumulated rainfall during the month previous to each trapping session (Rain). This analysis was also performed applying time lags of one (T30, MaxT30, MinT30, Rain30) and two months (T60, MaxT60, MinT60, Rain60) for temperature and rainfall variables.

Hydrology

Water level data for the Paraná River were provided by Prefectura Naval Argentina (Argentinean National Coast Guard; http://www.prefecturanaval.gov.ar) at Diamante port, 4 km away from Pre Delta National Park. Temperature and rainfall data were obtained from a weather station located in Diamante city. Weather and hydrological variables were recorded throughout the trapping period (July 2008–September 2011).

Statistical Analyses

In order to study habitat use by small rodents, two scales of analysis were defined: at the site scale (macro-habitat) and at the trap station scale (micro-habitat). At macro-habitat scale, the average of each vegetation variable recorded at each trapping station within a site was used, except for the maximum height in which the median value was used. At micro-habitat scale, the data used corresponded to each trapping station. Models at the micro-habitat scale were run separately for each study site. At the macro-habitat scale, site was included as an explanatory variable.

Logistic regressions (McCullagh and Nelder 1989; Nicholls 1991; Crawley 1993) were used to examine macro- and micro-habitat use by small rodents by means of the Generalized Linear Models procedure of R (R-Core-Team 2013) based on a priori models with a combination of explanatory variables (Tables 1, 2). Multiple correlation tests (Zar 1996) were performed among all the explanatory variables in order to avoid redundancy in the models.

Table 1 Macro-habitat Models Showing QAICc Values and Akaike Weights (w i) for All Small Rodents and for Each Separated Species (Aa, Cc, Or, and Of) from Pre Delta and Islas de Santa Fe National Parks.
Full size table
Table 2 Micro-habitat Models for Each Site Showing AIC Values and Akaike Weights (w i) for All Small Rodents and for Each Separated Species (Or, Of, Cc, and Aa) Caught in Pre Delta and Islas de Santa Fe National Parks Between June 2009 and September 2011.
Full size table

At a macro-habitat scale, we explored the relationship between CR (total and per species) and explanatory variables by means of two groups of candidate models: weather–hydrological models (eight models) and vegetation models (seven models; Table 1). Site and season were also included as explanatory variables in both groups of candidate models. Models of both groups were built using the binomial family distribution and the logit link function (Zuur 2009). The number of parameters was limited by the number of observations (Burnham and Anderson 2002).

Models were based on a Quasi-Akaike’s information criterion corrected for small sample size and over-dispersion data (QAICc) (Burnham and Anderson 2002). Models with smaller QAICc values than the null model, and with variables that have a parameter different from zero, were considered selected models. The final model was based on the average of the selected models (Symonds and Moussalli 2011).

For the micro-habitat analysis, the relationship between the presence of rodents (total and by species) at each trap station, and the vegetation variables and season was explored for each site using the binomial family distribution and the clog-log link function (Zuur 2009). Between three and seven candidate models per site derived from combinations of up to three variables were used (Table 2). Model selection was based on Akaike’s information criterion (AIC) (Burnham and Anderson 2002). Models with smaller AIC values than the null model, and with variables that have a parameter different from zero, were considered selected models. The final model was based on the average of the selected models (Symonds and Moussalli 2011).

We consider trapping a valid method to study habitat selection by small rodents because all traps are equally baited and hence their attraction effect on rodents should be equal. Thus, the capture event should be determined by the characteristics of the micro-habitat.

Results

Small Rodent Richness and Composition

A total of 453 individuals of five species were captured 512 times with a trapping effort of 9471 trap-nights. The species were Oxymycterus rufus (167 individuals), Oligoryzomys flavescens (160), Calomys callidus (68), Akodon azarae (42), and the genus Holochilus (15, including 3 H. chacarius determined to the species level).

Maximum rodent species richness was found at the marsh and the pond margins where all small rodent species were captured, followed by the inland and riparian forests and the two islands belonging to ISFNP (Fig. 2).

Figure 2
figure 2

Capture rate (CR) of each species captured from July 2008 to September 2011 in Islas de Santa Fe (island 1 and island 2) and Pre Delta national parks (riparian forest, inland forest, marsh, and pond margins). Trapping on the islands started in October 2008 and in the pond margins in September 2009.

Full size image

O. flavescens was captured at every site in both parks, and was numerically dominant on the islands. In PDNP, O. rufus was the most abundant species and was numerically dominant at the marsh, the pond margins and the inland forest. C. callidus was the most abundant species in the riparian forest. A. azarae was captured in all the sites (Fig. 2).

After the removal of captured O. flavescens during October 2008 on the two islands of ISFNP, these populations crashed abruptly (Fig. 3). By June 2009, both populations showed a recovery after which they declined reaching zero abundance in September 2009 (island 1) and December 2009 (island 2), and did not begin recovering until March 2011 (Fig. 3).

Figure 3
figure 3

Capture rate (CR) of O. flavescens, C. callidus, Holochilus sp, O. rufus, and A. azarae in Islas de Santa Fe National Park : a island 1, b island 2, and Pre Delta National Park: c riparian forest, d inland forest, e marsh, and f pond margins from July 2008 to September 2011. Arrows show removal of all captured O. flavescens individuals. Note different scaling of y-axis for the islands.

Full size image

Habitat Use

The selected weather–hydrological models described the CR of small rodents at PDNP and ISFNP as a function of season and the time elapsed since last flooding (Table 1). Model averaging also showed a positive association (stronger in spring and autumn than in winter) between CR of small rodents and the time since last flooding, contrary to summer where the association was negative (Table 3). The near absence of captures in May 2010 is probably related to the large flooding event which began at the end of 2009 and reached the maximum water level by February 2010 (Fig. 4). The peaks of CR occurred in spring, late winter, and early autumn, and they were related mainly to an increase in the abundance of O. flavescens and O. rufus, which were the numerically dominant species.

Table 3 Averaged Estimates of Relevant Variables for the Macro-habitat Model for Small Rodents from Pre Delta and Islas de Santa Fe National Parks.
Full size table
Figure 4
figure 4

Water level of the Paraná River (gray bars) and total capture rate of rodents (solid line) at Pre Delta and Islas de Santa Fe national parks. Water levels are shown since March 2007, when the last flood before the start of rodent trapping occurred. The dashed line indicates the value over which it was considered a flood event (530 mm).

Full size image

At the micro-habitat scale, small rodents selected patches with low cover of green grasses on island 2, selected habitat patches with high cover of green broadleaves in the riparian forest, and preferred patches with high dry grass cover and high vegetation in the marsh (Tables 2, 4).

Table 4 Average Estimates of Final Micro-habitat Models for Small Rodents from Pre Delta and Islas de Santa Fe National Parks Between June 2009 and September 2011.
Full size table

CR of O. flavescens at the macro-habitat scale was positively affected by the time elapsed since the last flood (estimator: 0.07; SE: 0.01; CI 95%: 0.05; 0.09; Table 1) and by green broadleaf cover (estimator: −0.04; SE: 0.01; CI 95%: −0.05; −0.03; Table 1).

O. flavescens selected patches with low green grass cover on island 2 (estimator: −0.02; SE: 0.01; CI 95%: −0.05; 0.00; Table 2), and with high dry grass cover in the marsh (estimator: 0.03; SE: 0.01; CI 95%: 0.00; 0.05; Table 2). None of the candidate models explained the presence of O. flavescens on island 1 (Table 2). This analysis was not carried out in the pond margins due to the low number of O. flavescens captured in this site.

CR of A. azarae at the macro-habitat scale was negatively affected by vegetation height (estimator: −0.63; SE: 0.21; CI 95%: −0.07; −0.24; Table 1). No weather–hydrological model was selected for A. azarae (Table 1), suggesting that the temperature and the time elapsed since the last flooding did not have an effect on the CR of this species, and that the season and sites do not explain the observed differences within PDNP.

At the micro-habitat scale, A. azarae selected patches with low vegetation at the marsh (estimator: −0.84; SE: 0.40; CI 95%: −1.63; −0.05; Table 2). At the other sites, this species did not select patches based on the variables considered in this study.

CRs of O. rufus differed among sites (Tables 1, 5), with the greatest values in the pond margins followed by the continental forest and the marsh (Fig. 2 and Table 5). This species was least captured in the riparian forest and was not captured on the islands (Fig. 1 and Table 2). Maximum temperature, time since the last flooding, season, and vegetation at a macro-habitat scale did not seem to affect its abundance (Tables 1, 5).

Table 5 Averaged Estimates of Relevant Variables of the Macro-habitat Model for O. rufus from Pre Delta and Islas de Santa Fe National Parks Between July 2008 and September 2011.
Full size table

At a micro-habitat scale, O. rufus selected patches with high cover of green broadleaves at the riparian (estimator: 0.02; SE: 0.01; CI 95%: 0.0; 0.05) and continental forests (estimator: 0.01; SE: 0.01; CI 95%: 0.00; 0.02; Table 2). At the marsh, this species selected patches with high vegetation (estimator: 0.36; SE: 0.18; CI 95%: 0.01; 0.72; Table 2).

CR of C. callidus was associated with the time elapsed since the last flooding event (estimator: 0.08; SE: 0.02; CI 95%: 0.05; 0.12; Table 1). This species was not captured on the islands (ISFNP) and no significant difference was observed in its CR among the sites where it was captured (Fig. 2 and Table 1). Vegetation analyses suggest that the abundance of this species was not associated with the vegetation variables at the macro-habitat scale (Table 1).

At a micro-habitat scale, C. callidus selected patches with high green broadleaf cover at the riparian forest (estimator: 0.04; SE: 0.01; CI 95%: 0.00; 0.08; Table 2), and patches with tall vegetation (estimator: 1.11; SE: 0.32; CI 95%: 0.48; 1.74; Table 2) and high cover of dry grasses (estimator: 0.02; SE: 0.01 CI 95%: 0.00; 0.05; Table 2) at the marsh.

Holochilus sp. was captured on the islands, in the marsh, and in the pond margins (Fig. 2). Individuals were captured during every season except winter (Fig. 3).

Discussion

Two hantavirus reservoir species, O. flavescens and A. azarae, were found in the study area. O. flavescens was captured at every studied habitat and was the most abundant species in Islas de Santa Fe National Park. This species is the reservoir of Lechiguanas genotype, one of the two known genotypes that cause HPS in central-east Argentina (Padula et al. 2007). The presence and high abundance of O. flavescens is of epidemiological importance in the area because the presence and circulation of Lechiguanas pathogenic genotype has been demonstrated in the populations of this species in Pre Delta and Islas de Santa Fe national parks (Vadell et al. 2011). The human activities that take place in this portion of the Paraná delta, such as fishing, cattle, crop farming, and eco-tourism (Gómez Villafañe, pers. com.) encourage the contact between people and sylvan rodents, increasing the risk of contracting HPS. The death of a fisherman who contracted HPS while living in Islas de Santa Fe National Park (island 1) highlights the importance of studying the hantavirus–rodent system, and the need for implementation of prevention measures to reduce virus exposure.

O. flavescens and C. callidus were negatively influenced by flooding. O. flavescens was found in highest numbers on the islands, areas strongly influenced by floods, and was the only species trapped there except for few individuals of the genus Holochilus. As part of the upper delta of the Paraná River, these islands are very recently formed and have a lower altitude than the continental areas. Hence, they are frequently flooded by intermediate and high pulses that surpass the main levees, leaving the whole ground area covered by water. O. flavescens is known as a good colonizer and has been described as a species adapted to arboreal habits (Massoia and Fornes 1965; Udrizar Sauthier et al. 2010; Maroli et al. 2015) and wetlands (Suárez and Bonaventura 2001). This species has been seen nesting in trees during flooding events in Pre Delta National Park (Pre Delta National Park forest rangers, pers. com.) and on other plants at heights between 1 and 2 m above ground level at the border of small channels (Udrizar Sauthier et al. 2010). The great increases in water level of the Paraná River, as occurred in this study, generate floods that sometimes surpass 100 cm above ground level. These floods restrict the access of small rodents to the vegetation at ground level during several days, weeks, or even months. Hence, great floods could affect small rodent populations in a negative way by destruction of their habitat, direct mortality, and by migration of survivors to non-flooded areas (Jacob 2003).

A. azarae, host of Pergamino hantavirus genotype in the area (Vadell et al. 2011; Palma et al. 2012), was present in every habitat of Pre Delta National Park and behaved as a generalist species at a micro-habitat scale in every habitat except for the marsh where it selected patches with low vegetation height. This lack of micro-habitat selection is in agreement with results obtained for this species by other authors in Otamendi Natural Reserve (Buenos Aires province) (Gómez Villafañe et al. 2012; Maroli et al. 2015). In addition, Hodara and Busch (2010) found selection by A. azarae at a micro-habitat scale in maize fields but not in field edges, and suggested an increased selectivity in the “poor habitat” (maize fields), in which individuals can perceive suitable and unsuitable patches. Under this hypothesis, the studied habitats at Pre Delta National Park (inland forest, pond margins, and riparian forest) might all be suitable for A. azarae. While this species has not been associated with a pathogenic genotype of hantavirus, studies should be continued because a potential risk to humans cannot be discarded. High mutation rates and the existence of multiple genotypes give RNA viruses such as hantaviruses the ability to become adapted to changing environments and to overcome barriers to spread (Cleaveland et al. 2007). In this scenario, the fact that A. azarae behaves as a generalist species in Pre Delta National Park is of epidemiological importance and should be taken into account if this species becomes a host of a pathogenic hantavirus in the future.

O. rufus was the most abundant species in every habitat except for the islands. O. rufus used the habitats differently at Pre Delta National Park, with a higher abundance in the pond margins, followed by the continental forest and the marsh, and with a minimum abundance in the riparian forest. This differential use of habitat does not seem due to differences in the general physiognomy of these habitats because they are all mainly moist and close to water bodies, nor to their vegetation because it was not included in the models as a relevant variable to explain O. rufus abundance.

The low number of Holochilus sp. captured in both parks could be due to a sub-estimation produced by the trapping method. This species is considered a strict herbivore and hence might not be attracted to the bait used during this study. This hypothesis is supported by the slight increase in the number of captures seen in autumn 2011, the period in which a high amount of both dead and alive individuals were observed across the middle delta of the Paraná River, which may have resulted from an outbreak of this species which trapping failed to detect. Moreover, two individuals from the genus Holochilus were seen swimming and moving around and over aquatic plants of the family Pontederiaceae in a stream that crosses Islas de Santa Fe National Park, and several others were also observed by park rangers in previous days (M.V. Vadell, personal observation). The construction of nests on plants of the family Pontederiaceae has been reported for Holochilus brasiliensis in the region (Udrizar Sauthier et al. 2010). This genus is considered closely associated with mesic microenvironments, such as grassy marshes, wetlands, and riparian forests (Pardiñas et al. 2010, 2013), as can be found in Pre Delta and Islas de Santa Fe national parks. Two species of this genus, H. brasiliensis and H. chacarius, have overlapping distributions in the study area (Pardiñas et al. 2013), but in this study we could only confirm the presence of H. chacarius in Islas de Santa Fe national park, based on the morphology of the jaws of three specimens. Pre Delta National Park contained a high diversity and abundance of small rodents, while Islas de Santa Fe National Park had a very low diversity. The low diversity in Islas de Santa Fe National Park could be related to its low diversity of habitats or to the intensity and frequency of the floods which could prevent the establishment of some species less adapted to flooding. In Pre Delta National Park, all the species with known distribution in this area were captured with the exception of Oligoryzomys nigripes and Scapteromys aquaticus (IUCN 2014). Massa et al. (2013) found that these species were absent (or in very low numbers) in barn owl’s pellets from different localities close to Pre Delta and Islas de Santa Fe national parks.

Given the susceptibility of hantaviruses to ultraviolet rays (Mills et al. 1995), we suggest creating open camping areas with mowed lawns (approximately 5 cm long) and trimmed shrubs in Islas de Santa Fe National Park. Tents should be placed at least 10 m away from the perimeter of the camping area to reduce contact with rodents living in the surrounding vegetation. We also suggest placing signs warning people about HPS and the risk of camping in densely vegetated areas. Campers should also be warned against keeping food inside tents and touching live or dead rodents, as suggested by Mills et al. (2002). Because Islas de Santa Fe National Park has been recently created and has not been officially open to the public yet, we suggest designing trails and walks that avoid very shadowy sectors in order to minimize the risk of contracting HPS. We also suggest paying special attention to the rodent proofing of buildings, and the use of rodent-proof containers for food and garbage, especially during flooding events when rodents are more prone to seek refuge (Zhang et al. 2007). In addition, special prevention measures should be taken during and after floods in order to minimize the risk of leptospirosis (Bharti et al. 2003; Ahern et al. 2005). Leptospirosis is a very common bacterial zoonosis caused by the spirochetes of the genus Leptospira that can be present in a wide range of animals including rodents, livestock, and domestic pets (Haake and Levett 2010). In spite of the fact that rodent numbers are expected to decrease after floods, the risk of leptospirosis is known to increase with flooding because of the bacteria’s ability to survive in wet soil and water (Bharti et al. 2003; Vanasco et al. 2008). Educational brochures and leaflets should be designed and given to tourist agents and local people in order to create awareness of disease prevention in natural environments.