Summer autumn distribution and abundance of the hantavirus host, Oligoryzomys longicaudatus, in northwestern Chubut, Argentina

Transcription

1 Journal of Mammalogy, 93(6): , 2012 Summer autumn distribution and abundance of the hantavirus host, Oligoryzomys longicaudatus, in northwestern Chubut, Argentina VERÓNICA ANDREO,* CECILIA PROVENSAL, SILVANA LEVIS, NOEMÍ PINI, DELIA ENRÍA, AND JAIME POLOP Consejo Nacional de Investigaciones Científicas y Técnicas (CONICET), Avenida Rivadavia 1917, CP C1033AAJ, Buenos Aires, Argentina (VA) Departamento de Ciencias Naturales, Universidad Nacional de Río Cuarto, Ruta 36 Km 601, Agencia Postal N8 3, CP 5800, Río Cuarto, Córdoba, Argentina (VA, CP, JP) Instituto Nacional de Enfermedades Virales Humanas Dr. Julio I. Maiztegui (INEVH), Monteagudo 2510, CP 2700, Pergamino, Buenos Aires, Argentina (SL, NP, DE) * Correspondent: veroandreo@gmail.com We examined population density of Oligoryzomys longicaudatus (colilargo) and prevalence of Andes virus (ANDV) antibody at regional and landscape spatial scales in northwestern Chubut Province (Argentina) and contrasted it with climatic variables recorded by meteorologic stations near the study area. Mice were trapped in late summer early fall (March April) for 3 years ( ). The composition of the rodent assemblage and species representation in the community varied among years, regions (forest, ecotone, and steppe), and landscape units (Nothofagus and Austrocedrus forests, sweet briar shrublands, and without sweet briar shrublands). Colilargos occurred in all regions and landscape units within the study area, from dense forest to open habitats such as steppe. The species dominated the rodent assemblages of ecotone and forest at a regional scale and the assemblages in sweet briar shrublands and Austrocedrus forests at a landscape scale. Abundance of colilargos also varied among periods, regions, and landscape units. Antibodies to ANDV were found in all regions but not in every landscape unit. Thus there is a potential for human hantavirus pulmonary syndrome (HPS) cases to occur not only in forests and shrublands, but also in steppe. At a landscape scale, Nothofagus forests appeared to pose a higher risk than Austrocedrus in wet years, because colilargo abundance and ANDV antibody prevalence were significantly greater. Within ecotone, sweet briar shrublands posed greater risk than habitats without sweet briar. Sweet briar shrublands were the landscape unit with the highest colilargo abundances during the driest periods. Sweet briar shrublands may play an important role in HPS dynamics, and should be considered when designing prevention policies. Key words: Andes virus, Argentina, hantavirus, Oligoryzomys longicaudatus, Patagonia Ó 2012 American Society of Mammalogists DOI: /11-MAMM-A Hantaviruses are zoonotic, host-specific RNA viruses that persistently infect rodents of the families Muridae (subfamily Murinae) and Cricetidae (subfamilies Arvicolinae, Neotominae, and Sigmodontinae [Schmaljohn and Hjelle 1997]). In North and South America, hantaviruses cause hantavirus pulmonary syndrome (HPS), an acute respiratory illness fatal in 30 50% of cases (Butler and Peters 1994). Since the recognition of HPS in the United States in 1993, outbreaks of the disease have been confirmed in many countries of the Americas, where at least 23 hantavirus genotypes have been described; about half of them are known to cause HPS. In general, each virus is associated with a single primary rodent host species, in which it establishes a chronic, asymptomatic infection that involves the shedding of infectious virus into the environment in rodent urine, feces, and saliva. These characteristics are key to the transmission of the virus to humans, which occurs mainly by inhalation of infectious aerosols of rodent excreta and secretions (Glass et al. 1993), and among rodents, which may be by a variety of mechanisms, including bites and aggressive encounters among adult animals (Childs et al. 1988, 1994; Glass et al. 1988; Mills et al. 1997). In Argentina, 4 distinct HPS-endemic areas have been recognized (Padula et al. 2007; Sosa-Estani et al. 2001): Northwest (Salta, Jujuy, and Formosa provinces), Northeast

2 1560 JOURNAL OF MAMMALOGY Vol. 93, No. 6 (Misiones Province), Central (Buenos Aires, Santa Fe, and Entre Ríos provinces), and Southern (Neuquén, Río Negro, Chubut, and Santa Cruz provinces). Both hantaviruses and hosts differ among endemic areas and HPS cases are unevenly distributed in space and time in the 4 regions. The colilargo (Oligoryzomys longicaudatus) is the major reservoir of Andes virus (ANDV), associated with HPS in southern Argentina (Levis et al. 1997, 1998; López et al. 1996). The colilargo is a widespread species with high prevalence of ANDV antibody (Cantoni et al. 2001; Larrieu et al. 2003; Padula et al. 2004; Piudo et al. 2005; Polop et al. 2010). The colilargo shows a marked flexibility in habitat use, although it is primarily found in woods and shrublands in Chile and southwestern Argentina (Larrieu et al. 2003; Murúa and González 1982; Pearson 1983; Piudo et al. 2005). The colilargo also has been captured in disturbed habitats such as borders of cultivated fields and peridomestic settings, grasslands, and pastures (Monjeau et al. 1998; Pardiñas et al. 2003; Pearson 1983; Piudo et al. 2005; Polop et al. 2010). Host population and pathogen prevalence may vary on regional, landscape, and local scales (Mills 1999). Many rodent species exhibit distinct habitat preferences and different habitat use patterns, so the risk of human disease may be more precisely defined by describing those differences in host distribution, density, and prevalence of infection among the distinct habitats represented in an area (Mills 1999). In addition to increasing our knowledge of basic wildlife habitat relationships, an understanding of the population dynamics and habitat associations of these rodent reservoirs is critical for an understanding of disease cycles in nature and is the 1st step toward any reservoir control or risk-reduction program. Although there are studies of abundance and habitat distribution of O. longicaudatus in other Patagonian provinces and in northwestern Chubut Province particularly, here we propose a different approach to understand habitat use by colilargos. We considered 2 different scales, regional and landscape, thus encompassing the complete rainfall gradient from forest in the west to steppe in the east. A west east gradient in the number of HPS cases in northwestern Chubut has already been observed (Dirección Área Programática Esquel, Chubut Province, Argentina). We assess how host and virus populations are distributed along the same gradient. Our main objective was to examine distribution and abundance of colilargos and distribution of ANDV among habitats, in northwestern Chubut Province, Argentina. Our specific goals were to determine the rodent assemblage composition, abundance of O. longicaudatus, and ANDV antibody prevalence in colilargos at regional and landscape spatial scales in late summer early fall (end of reproductive period), and to analyze the climatic variability of the study area and contrast it with rodent abundances. MATERIALS AND METHODS Study area. The study area in northwestern Chubut Province, Argentina, encompassed approximately 2,000 km 2, between S and S and W and W (Fig. 1). The topography is primarily mountainous with most features formed by glacial processes (Cabrera and Willink 1973). The climate is cold temperate with most rainfall occurring between April and September (2,000 mm in the west to,250 mm in the east). The study area is composed of 2 main phytogeographic regions, Subantarctic forest and Patagonian steppe. The Subantarctic forest contains the deciduous forest district (Cabrera and Willink 1973) characterized by 2 Nothofagus species, N. antarctica (ñire) and N. pumilio (lenga), and by the conifer Austrocedrus chilensis (ciprés de cordillera) in lower numbers. Patagonian steppe in the study area was characterized by highly covered grass steppes dominated by Festuca pallescens (coirón dulce) and shrub grass steppes with a total cover of 50% dominated by Stipa speciosa (coirón amargo), Stipa humilis (coirón llama), Adesmia campestres (mamuel choique), Berberis heterophylla (calafate), and Poa lanuginosa (pasto hilo). Between forest and steppe there is a transition zone or ecotone characterized by Nothofagus or Austrocedrus in open forest patches, shrublands, and grasslands. This natural transition zone has widened both toward forest and steppe due to human activities (wood extraction, logging, fire, exotic species introduction, cattle farming, and tourism). Rosa rubiginosa (sweet briar), an exotic invasive shrub species, is particularly abundant in this transition area. Selection of sampling units. We identified 3 geographic regions in the study area following the west east gradient in precipitation: forest, ecotone, and steppe. At a landscape scale or mesoscale, within forest we selected 2 vegetation units to perform rodent trapping: forests dominated by N. antarctica (Nothofagus forests) and forests dominated by A. chilensis (Austrocedrus forests) and within the ecotone, we selected 2 types of shrublands, those dominated by native species (without sweet briar shrublands) and those dominated by R. rubiginosa (sweet briar shrublands); fruits of this species constitute one of the most important food items for colilargos in the study area (Sbriller and Sepúlveda 2007). We did not identify different landscape units within steppe. Criteria for the selection of these vegetative units included accessibility, representation in the area, homogeneity, and ease of identification. Rodent trapping and serology. Rodent trapping was conducted once a year for 3 years ( ), in late summer early fall (March April). Rodents were trapped using 72 traplines per year: 24 in forest (12 in Nothofagus and 12 in Austrocedrus forest), 24 in ecotone (12 in sweet briar shrubland and 12 in shrubland without sweet briar), and 24 in steppe. Each trapline consisted of 20 traps (10 live-capture traps and 10 snap traps alternating) placed at 5-m intervals and baited with a mixture of peanut butter and cow fat. All traplines were checked for 3 consecutive days. The minimum distance between traplines was 200 m. For all rodents captured, species, sex, body mass, body and tail length, reproductive status, and presence of scars were recorded. Rodents captured using live traps were anesthetized with methoxyflurane gas, and blood

3 December 2012 ANDREO ET AL. O. LONGICAUDATUS IN SOUTHERN ARGENTINA 1561 samples for antibody detection were obtained from the retroorbital sinus. The anesthetized animals were euthanized by overdose of anesthesia before necropsy and liver sample extraction. Rodents captured with snap traps were processed in the same way. Blood and liver samples were sent to the National Institute of Human Viral Diseases (Pergamino, Buenos Aires Province, Argentina) where, using an enzymelinked immunosorbent assay, the infection status of individuals was assessed. Research on live animals was performed in a humane manner and followed guidelines for the care and use of animals approved by the American Society of Mammalogists (Sikes et al. 2011). Handling of rodents followed standardized safety guidelines recommended by the United States Centers for Disease Control and Prevention (Mills et al. 1995). Climatic data. Temperature (mean, maximum, and minimum) and precipitation from 2006 to 2009 for 14 meteorologic stations 200 km around the study area were provided by Instituto Nacional de Tecnología Agropecuaria (Esquel), Servicio Meteorológico Nacional (Buenos Aires), and Hidroeléctrica Futaleufú (Esquel). We also obtained monthly values of the Southern Oscillation index ( These data were collected in order to analyze the climatic variability of the study area and its surroundings during the study period, and then relate and contrast it to rodent abundances observed in each sampling period. Data analysis. To estimate rodent abundance we used the relative density index (RDI), which weights the relationship between the number of animals captured and the number of trap-nights used: captures RDI ¼ 3 100; ðtraps 3 nightsþ a where a is the number of traps that were closed but empty. We assumed that this index is correlated to the population abundance, and if so, it is useful to compare different sites and moments. The rodent assemblage composition for the study area in general and for each region and landscape unit was determined FIG. 1. Study area in northwestern Chubut Province in Argentina. by the proportion of each species relative to the total number of animals captured. Prevalence of infection was calculated as the proportion of antibody-positive animals divided by the total number tested. Prevalence was estimated for each species, sample site or trapline, landscape unit, and region. Because data failed to meet the assumptions of normality and variance homogeneity, we applied nonparametric Kruskal Wallis and Mann Whitney tests to compare general rodent assemblage and abundance and prevalence of O. longicaudatus among regions and landscape units, respectively. Colilargo region and landscape associations were evaluated, as suggested by Mills et al. (1991), by deriving an expected number of captures for the species in each unit by multiplying the total number of individuals captured by the proportion of trap nights in that unit. This expected value was compared to the observed number of captures in a particular unit and chisquare values were derived. Deviations of observed from expected values were standardized as percentages of expected values. Significant deviations from expected associations were tested using 95% Bonferroni confidence intervals (Byers et al. 1984). We compared temperature and precipitation among years for each meteorologic station alone and for all stations together. Comparisons were performed using analysis of variance and a posteriori Tukey tests when data met the normality and variance homogeneity assumptions and Kruskal Wallis and Mann Whitney tests when it did not. All statistical analyses were performed using R (R Development Core Team 2010). RESULTS Rodent assemblage composition. We captured 760 individuals during the 3 sampling periods, from March 2007 to March 2009 with a total of 12,960 trap-nights (4,320 trapnights/year). The rodent assemblage in the study area was formed by Abrothrix longipilis, O. longicaudatus, Loxodontomys micropus, Abrothrix olivaceus, Eligmodontia morgani, Reithrodon auritus, Phyllotis xanthopygus, Geoxus valdivianus, Chelemys macronyx, and Ctenomys sp. A.

4 1562 JOURNAL OF MAMMALOGY Vol. 93, No. 6 longipilis and colilargos comprised more than 50% of the rodent community (Fig. 2a). The general assemblage composition (number of different rodent species) varied among years as well as did species representation (percentage of each species relative to the total number of animals captured) in the community (Fig. 2b). Likewise, assemblage composition and species representation varied among regions (Fig. 3) and landscape units. The steppe had the highest total richness in the 3 periods studied (Fig. 3). In the ecotone and forest, the rodent assemblage was dominated by the colilargo, even though its proportion relative to the total number of animals captured was lower in 2008 and 2009 (Fig. 2b). In 2007, colilargos dominated the rodent assemblage in all landscape units. However, in the next periods, colilargos only dominated the assemblage of sweet briar shrublands and Austrocedrus forests. General rodent assemblage and colilargo abundance. The total number of rodents captured varied among years. Steppe had the highest total relative density (in terms of relative density index) in the 3 periods studied (2007: K 2 ¼ 5.93, P ¼ 0.05, n ¼ 72; 2008: K 2 ¼ 29.53, P ¼ 0.000, n ¼ 72; 2009: K 2 ¼ 47.15, P ¼ 0.000, n ¼ 72). There were no significant differences between forest and ecotone abundance values in any period considered (Fig. 4). At a landscape scale, we only detected significant differences in total abundance between forest types in 2007 (U ¼ 16.5, P ¼ 0.001, n ¼ 24), with Nothofagus forest showing higher abundance than Austrocedrus forest. Colilargo relative abundance varied both at regional and landscape scales among sampling periods. In 2007, the forest had the highest abundance value and steppe the lowest (K 2 ¼16.05, P ¼ 0.000, n ¼ 72), with no statistically significant (P. 0.05) differences between forest and ecotone (Fig. 4). In 2008, colilargos were more abundant in ecotone and in 2009, in steppe but differences among regions were not statistically significant (2008: K 2 ¼ 3.12, P ¼ 0.20, n ¼ 72; 2009: K 2 ¼ 4.23, P ¼ 0.120, n ¼ 72). At a landscape scale (Fig. 4), sweet briar shrublands had significantly (P, 0.05) higher abundances than shrublands without sweet briar both in 2008 (U ¼ 102, P ¼ 0.016, n ¼ 24) and 2009 (U ¼ 96, P ¼ 0.036, n ¼ 24). Within forests, we only found significant differences in colilargo abundances during 2007 (U ¼ 11, P ¼ 0.000, n ¼ 24), when Nothofagus forest had the highest abundance of colilargos in the 3 periods considered. Colilargo region and landscape association. Colilargos showed a significant degree of heterogeneity in region and landscape associations. Captures in forests during the 1st sampling period, for example, were 70% greater than expected, whereas in steppe captures of this species were almost 80% less than expected (v 2 2 ¼ 57.09, P ¼ 0.000). In 2008, the negative association with steppe persisted, but colilargo captures in the ecotone were twice that expected (v 2 2 ¼ 10.33, P ¼ 0.005). Finally, in 2009, colilargos were significantly overrepresented in steppe and underrepresented in forests (v 2 2 ¼13.76, P ¼ 0.001). At a landscape scale, colilargos were positively associated with Nothofagus forests only in 2007, where captures were 68% greater than expected (v 2 2 ¼ 40.01, P ¼ 0.000). Although FIG. 2. General rodent assemblage composition a) for the 3 sampling periods altogether and b) for each sampling period in northwestern Chubut Province, Argentina. A.l.: Abrothrix longipilis; A.o.: Abrothrix olivaceus; Ch.m.: Chelemys macronyx; C.sp.: Ctenomys sp.; E.m.: Eligmodontia morgani; G.v.: Geoxus valdivianus; L.m.: Loxodontomys micropus; O.l.: Oligoryzomys longicaudatus; Ph.x.: Phyllotis xanthopygus; R.a.: Reithrodon auritus. statistical tests were not significant, colilargos were captured 20% and 33% more frequently than expected in Austrocedrus forests in 2008 and Within ecotone, colilargos were significantly overrepresented in sweet briar shrublands (100%) in 2008 (v 2 2 ¼ 12, P ¼ 0.000) and 2009 (v 2 2 ¼ 5, P ¼ 0.025).

5 December 2012 ANDREO ET AL. O. LONGICAUDATUS IN SOUTHERN ARGENTINA 1563 FIG. 3. Rodent assemblage composition for each region and sampling period in northwestern Chubut Province, Argentina. A.l.: Abrothrix longipilis; A.o.: Abrothrix olivaceus; Ch.m.: Chelemys macronyx; C.sp.: Ctenomys sp.; E.m.: Eligmodontia morgani; G.v.: Geoxus valdivianus; L.m.: Loxodontomys micropus; O.l.: Oligoryzomys longicaudatus; Ph.x.: Phyllotis xanthopygus; R.a.: Reithrodon auritus. Andes virus infection. In 2007, 5 (3.3%) of 152 colilargos captured in the study area were hantavirus antibody positive. Of those 5 infected colilargos, 4 were trapped in the forest region (n ¼ 87) and 1 in the ecotone (n¼54). At a landscape scale, the 4 infected colilargos from forest were captured in Nothofagus forests (n ¼ 73), and the only infected mouse from ecotone was trapped in a sweet briar shrubland (n ¼ 26). All infected colilargos were males and all but 1 were adults. In 2007, we also captured a positive specimen of L. micropus in the steppe. Two colilargos were trapped in the same steppe sampling site but neither had antibody. In 2008, we found only 1 positive A. longipilis in sweet briar shrubland where 3 antibody-negative colilargos were captured. Finally, in 2009, infection was only detected in steppe, where 2 colilargos (n ¼ 17) were found to be hantavirus antibody positive. The low number of infected colilargos observed during the 3 sampling periods prevented the application of statistical comparisons of prevalence among regions and landscape units. Climate. There were significant differences in climatic records from meteorologic stations near the study area and in Southern Oscillation index values in the periods before samplings were carried out. The period before the 2007 sampling was characterized by negative Southern Oscillation index values and the periods before 2008 and 2009 samplings by positive Southern Oscillation index values (Fig. 5). Spring summer rainfall before 2007 sampling was higher than rainfall recorded in spring summer before 2008 and 2009 sampling periods (comparison among periods for all meteorologic stations together: F 2,242 ¼ 7.20, P ¼ 0.000). Fig. 6 shows the total spring summer rainfall recorded in each meteorologic station during Winter precipitation, on the other hand, was significantly lower in 2007 (before 2008 samplings, for all meteorologic stations, X ¼102 mm, SD ¼ 92.8 mm) compared to winters of 2006 (for all meteorologic stations, X ¼196 mm, SD ¼ mm) and 2008 (for all meteorologic stations, X ¼199 mm, SD ¼ mm; comparison among periods for all meteorologic stations together: F 2,163 ¼ 7.85, P ¼ 0.000). There also were differences in temperature among the periods before sampling. Summer mean temperature was lower before 2007 (for all meteorologic stations, X ¼13.28C, SD ¼ 1.588C) sampling than before the other 2 sampling periods, that is, summer was cooler than the following periods (2008: for all meteorologic stations, X ¼ C, SD ¼ 2.138C, 2009: for all meteorologic stations, X ¼ C, SD ¼ 1.778C;

6 1564 JOURNAL OF MAMMALOGY Vol. 93, No. 6 FIG. 4. Abundance values expressed as relative density index (RDI) a, c, e) for the general rodent assemblage and b, d, f) for Oligoryzomys longicaudatus in each region and landscape unit in northwestern Chubut Province, Argentina ( ). comparison among periods for all meteorologic stations together: F 2,134 ¼ 20.24, P ¼ 0.000). Moreover, spring summer maximum temperature was higher before the 2008 sampling (for all meteorologic stations, X ¼ C, SD ¼ 4.088C) compared to the period before the other sampling periods (2007: for all meteorologic stations, X ¼ C, SD ¼ 3.428C, 2009: for all meteorologic stations, X ¼ C, SD ¼ 3.228C; comparison among periods for all meteorologic stations together: F 2,165 ¼ 17.94, P ¼ 0.000). DISCUSSION We found that the ANDV host occurred in all regions and landscape units studied, from dense forests to more open steppe. Relative abundances, however, varied among regions and landscape units, both within and among sampling periods. We found ANDV in all regions but not in every landscape unit and this also varied among periods. The occurrence of colilargos in all regions and landscape units is consistent with several studies suggesting a marked flexibility in habitat use by this species, which is characterized by opportunistic behavior, high mobility, and large home ranges (Murúa and González 1986; Murúa et al. 1986). Pearson and Pearson (1982) cited O. longicaudatus as a wide-range species, occurring in southern Argentina from the zone of approximately 3,500 mm of rainfall in the west (dense N. pumilio forests) to less than 500 mm annually in the east (steppe). In Chile, distribution of O. longicaudatus encompasses contrasting landscapes ranging from semidesert thornscrub in the north (Mediterranean region) to southern Magellanic rain forests (Mann 1978). Colilargos also have been found in grasslands and peridomestic settings in Patagonia (Piudo et al. 2011; Polop et al. 2010). Although occurring in every type of habitat studied, colilargos were most commonly found in forest and ecotone (even those periods when abundances were lower, colilargos still dominated rodent

7 December 2012 ANDREO ET AL. O. LONGICAUDATUS IN SOUTHERN ARGENTINA 1565 FIG. 5. Southern Oscillation index (SOI) monthly values from 2006 to assemblages of forest and ecotone), which is consistent with studies suggesting that, despite its generalist habits, O. longicaudatus prefers humid habitats with high cover (Kelt et al. 1994; González et al. 2000; Lozada et al. 2000; Monjeau et al. 1998; Pearson 1983; Pearson and Pearson 1982; Piudo et al. 2005; Polop et al. 2010). This also is in agreement with association analyses performed in this study, both at regional and landscape scales. We could assume then, that environmental differences among regions and landscapes are responsible for the different abundances and habitat associations observed. In previous analyses we have observed that colilargos consume a high proportion of the sweet briar fruit both in forests and shrublands year-round (Sbriller and Sepúlveda 2007). Therefore, it seems likely that this shrub is providing colilargo populations with both refuge and food resources. Nevertheless, Murúa and González (1982) have suggested that colilargos select microsites with high foliage density and thick understory in southern Chilean forests because they offer protection from predators. Abundance differences observed among sampling periods, in forest and ecotone in particular, are probably related to some environmental or climatic condition that differed among years and somehow diminished suitability of these regions the last 2 summers (2008 and 2009). A high sensitivity to climatic factors and fluctuations in resource levels was suggested for O. longicaudatus in Chile (Murúa et al. 1987). Consistent with this concept, we observed that the period before 2007 samplings was characterized by an El Niño event, already known to positively affect population dynamics of O. longicaudatus through an increase in precipitation and subsequent increase in primary productivity (Murúa et al. 2003). The periods before the 2008 and 2009 sampling, in contrast, were characterized by positive values of the Southern Oscillation index (La Niña), which are related to decreased precipitation. Values of this index were consistent with precipitation recorded in meteorologic stations around the study area. Spring summer rainfall before the 2007 sampling was higher than rainfall recorded in spring summer before the 2008 and 2009 sampling. We analyzed particularly spring summer total rainfall because we suspect it might have a stronger effect on rodent populations because it is coincident with the reproductive season (Guthmann et al. 1997; Murúa et al. 1986; Pearson 1983). We also recorded differences in temperature among the periods before samplings. We suspect that these climatic differences among years may be related to rodent abundance differences, likely acting indirectly through primary productivity, the onset of vegetative growth, and the production of seeds and fruits, affecting food and refuge resources. When matching climatic conditions with region and landscape associations and abundances recorded, colilargos appeared to be abundant in Nothofagus forests only during wet summers, but were always found in sweet briar shrublands and forests closer to transition zone, such as Austrocedrus forests. This is in agreement with previous studies on habitat FIG. 6. Total spring summer rainfall previous to each sampling period for several meteorologic stations nearby the study area Instituto Nacional de Tecnología Agropecuaria (Esquel), Servicio Meteorológico Nacional (Buenos Aires), and Hidroeléctrica Futaleufú (Esquel). in northwestern Chubut Province, Argentina.

8 1566 JOURNAL OF MAMMALOGY Vol. 93, No. 6 preferences of the species in Argentina and Chile, as described above. Moreover, some authors have observed that, in southern forests of Chile, colilargos are absent from forest during summer, probably because of migration to other areas or different altitudes within the forest (Murúa and González 1979). However, a different pattern is observed in northern Chile, where colilargos occur in aguadas (mesic vegetation with standing water) and scrublands in wet years (El Niño events), but disappear from these environments in dry years (e.g., La Niña events). The increase in population abundance during and after El Niño events may reflect an increase in the availability of food or a reduction in water stress, allowing populations to move out of the aguadas (Meserve et al. 2003). The characteristic that has most strongly established this species as a focus of epidemiologic research in the last decade is that it is the reservoir of ANDV (Levis et al. 1998). ANDV antibodies were confirmed in all regions but not in every landscape unit; in Austrocedrus forests and shrublands without sweet briar we did not detect infection. We did capture colilargos in those units, however. Therefore, we cannot infer that ANDV is absent from those landscapes, but might assume that the prevalence of infection is lower. Numbers of infected mice, although low, were concordant with records from other areas of Patagonia (Piudo et al. 2005, 2011). The highest number of infected mice was observed when abundances were high and in the landscape unit where colilargo abundances were highest. Therefore, the probability of humans contracting HPS may be greater in forest habitats, particularly in Nothofagus forest, which is where most cases have occurred (Carbajo and Pardiñas 2007). Examination of our data indicates, however, that this might be especially true for wet years, because we did not detect infection in forests in summers 2008 and Previous studies have not detected infection in sylvan areas of steppe (Cantoni et al. 2001; Larrieu et al. 2003; Piudo et al. 2005, 2011). On the contrary, we found 2 infected colilargos and an individual of L. micropus with antibodies for ANDV. This constitutes the 1st record of ANDV in a steppe sylvan area. Therefore, we should not disregard the importance of this type of habitat in the dynamics of the host, virus, and disease. We have characterized distribution and abundance of O. longicaudatus populations and ANDV in northwestern Chubut along a precipitation gradient from forest in the west to steppe in the east for 3 consecutive summers. The presence of colilargos in every region and landscape unit studied and the presence of the virus in almost every landscape unit should be considered when designing prevention programs against HPS. The distribution of the reservoir species may indicate the maximum potential extent of the disease (Mills and Childs 1998). Therefore, according to our findings there is the potential for human HPS to occur not only in forests and shrublands, but also in steppe. At a landscape scale, forests dominated by N. antarctica appeared to represent higher risk than those dominated by Austrocedrus in wet years, because colilargo abundances and ANDV antibody prevalence were significantly greater. Within ecotones, sweet briar shrublands represented higher risk habitats than shrublands without sweet briar where we captured colilargos only in the 1st sampling period, and no mice were infected. Sweet briar shrublands were the landscape unit with the highest colilargo abundances during the driest periods. These shrublands have an extra risk factor: from March to May, local inhabitants collect sweet briar fruits for jam making and cosmetic industries. Previous studies already have found association between colilargo captures and the abundance or cover of sweet briar (Lozada et al. 2000; Pearson 1983). Moreover, we have observed that fruits of sweet briar are one of the most common food items of colilargos year-round (Sbriller and Sepúlveda 2007). So, this particular type of shrubland may play an important role in HPS disease dynamics and should be especially considered when designing preventive polices. Further studies at different scales are needed to elucidate climatic and environmental variables associated to occurrence and abundance of O. longicaudatus in order to gain a deeper understanding of the system dynamics and allow prediction of those times and places of higher risk of hantavirus transmission to humans. RESUMEN Nuestro objetivo fue examinar la densidad poblacional de Oligoryzomys longicaudatus (colilargo) y la prevalencia de virus Andes (ANDV) a escala regional y de paisaje en el noroeste de la provincia de Chubut, Argentina. Los muestreos se llevaron a cabo durante el verano tardío otoño temprano (Marzo Abril) por un período de 3 años (2007 a 2009). La composición de la comunidad de roedores y la representación de cada especie en la misma varió entre años, regiones (bosque, ecotono y estepa) y unidades de paisaje (bosques de Nothofagus ydeaustrocedrus, matorrales de rosa mosqueta y matorrales sin rosa mosqueta). O. longicaudatus fue encontrado en todas las regiones y unidades de paisaje del área de estudio. Esta especie dominó los ensambles de roedores de bosques y ecotono a escala regional y los ensambles de los matorrales de rosa mosqueta y bosques de Austrocedrus a escala de paisaje. La abundancia de colilargos varió entre períodos, regiones y unidades de paisaje. Se detectaron anticuerpos contra ANDV en todas las regiones, pero no en todas las unidades de paisaje. Por lo tanto, el riesgo de enfermedad en humanos existe no sólo en bosque y ecotono, sino también en estepa. Además, a escala de paisaje, los bosques de Nothofagus parecieron implicar un mayor riesgo que los de Austrocedrus en años húmedos, ya que las abundancias de colilargos y la prevalencia de ANDV fueron significativamente mayores en los primeros. Dentro del ecotono, los matorrales de rosa mosqueta significaron un mayor riesgo que los matorrales sin rosa mosqueta. Los primeros constituyeron la unidad de paisaje con las mayores abundancias de colilargos durante los períodos secos. Estos matorrales podrían tener un papel importante en la dinámica del Síndrome Pulmonar por Hantavirus y debieran ser considerados a la hora de diseñar medidas de prevención.

9 December 2012 ANDREO ET AL. O. LONGICAUDATUS IN SOUTHERN ARGENTINA 1567 ACKNOWLEDGMENTS This research was made possible by grants from the Fondo para la Investigación Científica y Tecnológica (FONCYT), Consejo Nacional de Investigaciones Científicas y Técnicas (CONICET), and Fundación Mundo Sano. We thank F. Argel, B. San Martin, and F. Polop for collaboration in the fieldwork and J. N. Mills for reviewing an earlier version of the manuscript. We also thank Instituto Nacional de Tecnología Agropecuaria (Esquel), Servicio Meteorológico Nacional (Buenos Aires), and Hidroeléctrica Futaleufú (Esquel) for providing us with climatic information. We are grateful to two anonymous reviewers who provided valuable comments and suggestions on early versions of the manuscript. LITERATURE CITED BUTLER, J., AND C. PETERS Hantaviruses and hantavirus pulmonary syndrome. Clinical Infectious Diseases 9: BYERS, C. R., R. K. STEINHORST, AND P. R. KRAUSMAN Clarification of a technique for analysis of utilization availability data. Journal of Wildlife Management 48: CABRERA, A. L., AND A. WILLINK Biogeografía de América Latina. Serie de Biología. Programa Regional de Desarrollo Científico y Tecnológico, Departamento de Asuntos Científicos, Secretaria General de la Organización de los Estados Americanos, Monografía 13: CANTONI, G., ET AL Seasonal variation in prevalence of antibody to hantaviruses in rodents from southern Argentina. Tropical Medicine and International Health 6: CARBAJO, A. E., AND U. F. J. PARDIÑAS Spatial distribution model of a hantavirus reservoir, the long-tailed colilargo (Oligoryzomys longicaudatus), in Argentina. Journal of Mammalogy 88: CHILDS, J. E., G. E. GLASS, G.W.KORCH, AND J. W. LEDUC The ecology and epizootiology of hantaviral infections in small mammal communities of Baltimore: a review and synthesis. Bulletin of the Society for Vector Ecology 13: CHILDS, J. E., ET AL Serologic and genetic identification of Peromyscus maniculatus as the primary rodent reservoir for a new hantavirus in the southwestern United States. Journal of Infectious Diseases 169: GLASS, G. E., J. E. CHILDS, G.W.KORCH, AND J. W. LEDUC Association of intraspecific wounding with hantaviral infection in wild rats (Rattus norvegicus). Epidemiology and Infection 101: GLASS, G. E., A. WATSON, J.LEDUC, G.KELEN, T.QUINN, AND J. CHILDS Infection with a rat-borne hantavirus in US residents is consistently associated with hypertensive renal disease. Journal of Infectious Diseases 167: GONZÁLEZ, L. A., R. MURÚA, AND C. JOFRÉ Habitat utilization of two muroid species in relation to population outbreaks in southern temperate forests of Chile. Revista Chilena de Historia Natural 73: GUTHMANN, N., M. LOZADA, A.MONJEAU, AND K. HEINEMANN Population dynamics of five sigmodontine rodents of northwestern Patagonia. Acta Theriologica 42: KELT, D. A., L. MESERVE, AND B. K. LANG Quantitative habitat associations of small mammals in a temperate rainforest in southern Chile: empirical patterns and the importance of ecological scale. Journal of Mammalogy 75: LARRIEU, E., ET AL Seroprevalencia de hantavirus en roedores y casos humanos en el sur de Argentina. Revista Brasilera de Epidemiología 6: LEVIS, S., ET AL Genetic diversity and epidemiology of hantaviruses in Argentina. Journal of Infectious Diseases 177: LEVIS, S., J. E. ROWE, S.MORZUNOV, D.A.ENRÍA, AND S. ST. JEOR New hantaviruses causing hantavirus pulmonary syndrome in central Argentina. Lancet 349: LÓPEZ, N., P. PADULA, C.ROSSI, M.E.LAZARO, AND M. T. FRANZE- FERNANDEZ Genetic identification of a new hantavirus causing severe pulmonary syndrome in Argentina. Virology 219:1 4. LOZADA, M., N. GUTHMANN, AND N. BACCALA Microhabitat selection of five sigmodontine rodents in a forest steppe transition zone in northwestern Patagonia. Studies on Neotropical Fauna and Environment 35: MANN, G Los pequeños mamíferos de Chile. Gayana, Zoología 40: MESERVE, P. L., D. A. KELT, B.MILSTEAD, AND J. R. GUTIÉRREZ Thirteen years of shifting top-down and bottom-up control. BioScience 53: MILLS, J The role of rodents in emerging human disease: examples from the hantaviruses and arenaviruses. Pp in Ecologically-based rodent management (G. R. Singleton, L. A. Hinds, H. Leirs, and Z. Zhang, eds.). Arrawang Communication Group, Canberra, Australia. MILLS, J., AND J. E. CHILDS Ecologic studies of rodent reservoirs: their relevance for human health. Emerging Infectious Diseases 4: MILLS, J., J. E. CHILDS, T.G.KSIAZEK, C.J.PETERS, AND W. M. VELLECA Methods for trapping and sampling small mammals for virologic testing. United States Department of Health and Human Services, Atlanta, Georgia. MILLS, J. N., B. A. ELLIS, K.T.MCKEE, J.I.MAIZTEGUI, AND J. E. CHILDS Habitat associations and relative densities of rodent populations in cultivated areas of central Argentina. Journal of Mammalogy 72: MILLS, J., ET AL Patterns of association with host and habitat: antibody reactive with Sin Nombre virus in small mammals in the major biotic communities of the southwestern United States. American Journal of Tropical Medicine and Hygiene 56: MONJEAU, J. A., E. C. BIRNEY,L.GHERMANDI,R.S.SIKES,L.MARGUTTI, AND C. J. PHILLIPS Plants, small mammals, and the hierarchical landscape classifications of Patagonia. Landscape Ecology 13: MURÚA, R., AND L. A. GONZÁLEZ Distribución de roedores silvestres con relación a las características del hábitat. Anales del Museo de Historia Natural Valparaíso 12: MURÚA, R., AND L. A. GONZÁLEZ Microhabitat selection in two Chilean cricetid rodents. Oecologia 52: MURÚA, R., AND L. A. GONZÁLEZ Regulation of numbers in two Neotropical rodent species in southern Chile. Revista Chilena de Historia Natural 59: MURÚA, R., L. A. GONZÁLEZ, AND M. LIMA Population dynamics of rice rats (a hantavirus reservoir) in southern Chile: feedback structure and non-linear effects of climatic oscillations. Oikos 102: MURÚA, R., L. A. GONZÁLEZ, AND P. L. MESERVE Population ecology of Oryzomys longicaudatus philippii (Rodentia: Cricetidae) in southern Chile. Journal of Animal Ecology 55: MURÚA, R., P. L. MESERVE, L.A.GONZÁLEZ, AND C. JOFRÉ The small mammal community of a Chilean temperate rain forest: lack

10 1568 JOURNAL OF MAMMALOGY Vol. 93, No. 6 of evidence of competition between dominant species. Journal of Mammalogy 68: PADULA, P., ET AL Transmission study of Andes hantavirus infection in wild sigmodontine rodents. Journal of Virology 78: PADULA, P., ET AL Pathogenic hantaviruses, northeastern Argentina and eastern Paraguay. Emerging Infectious Diseases 13: PARDIÑAS, U. F. J., P. TETA, S.CIRIGNOLI, AND D. H. PODESTÁ Micromamíferos (Didelphimorphia y Rodentia) de norpatagonia extra andina, Argentina: taxonomía alfa y biogeografía. Mastozoología Neotropical 10: PEARSON, O. P Characteristics of a mammalian fauna from forests in Patagonia, southern Argentina. Journal of Mammalogy 64: PEARSON, O. P., AND A. K. PEARSON Ecology and biogeography of the southern rainforests of Argentina. Special Publication Series, Pymatuning Laboratory of Ecology, University of Pittsburgh 6: PIUDO, L., M. J. MONTEVERDE, S.GONZÁLEZ CAPRIA, P.PADULA, AND P. CARMANCHAHI Distribution and abundance of sigmodontine rodents in relation to hantavirus in Neuquén, Argentina. Journal of Vector Ecology 30: PIUDO, L., M. J. MONTEVERDE, R.S.WALKER, AND R. J. DOUGLASS Rodent community structure and Andes virus infection in sylvan and peridomestic habitats in northwestern Patagonia, Argentina. Vector-borne and Zoonotic Diseases 11: POLOP, F. J., ET AL Temporal and spatial host abundance and prevalence of Andes hantavirus in southern Argentina. EcoHealth 7: RDEVELOPMENT CORE TEAM R: a language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. Accessed 23 March SBRILLER, A., AND L. SEPÚLVEDA La rosa mosqueta, el colilargo patagónico y el hantavirus. Desde la Patagonia 5: SCHMALJOHN, C., AND B. HJELLE Hantaviruses: a global disease problem. Emerging Infectious Diseases 3: SIKES, R. S., W. L. GANNON, AND THE ANIMAL CARE AND USE COMMITTEE OF THE AMERICAN SOCIETY OF MAMMALOGISTS Guidelines of the American Society of Mammalogists for the use of wild mammals in research. Journal of Mammalogy 92: SOSA-ESTANI, S., O. D. SALOMÓN,A.O.GÓMEZ,M.L.ESQUIVEL, AND E. L. SEGURA Diferencias regionales y síndrome pulmonar por hantavirus (enfermedad emergente y tropical en Argentina). Cadernos de Saude Publica, Rio de Janeiro 17: Submitted 26 May Accepted 19 June Associate Editor was Victor Sánchez-Cordero.