Spatiotemporal abundance variation of a Hantavirus reservoir and environmental conditions

FRANCISCO POLOP; CECILIA PROVENSAL; VERÓNICA ANDREO; DELIA ENRÍA; XIMENA PORCASI; MARCELO SCAVUZZO; JAIME POLOP

Mendoza

Congreso; 10th International Mammalogical Congress; 2009

Center for Science and Technology - CCT CONICET Mendoza, CONICET, International Federation of Mammalogists (IFM), Sociedad Argentina para el estudio de los Mamíferos (SAREM), entre otras instituciones

The aim of this work was to establish the relationship between spatial and temporal abundance variation of O. longicaudatus and environmental conditions. Rodents were trapped seasonally in two types of habitats: forests and shrublands. 24 forests sites and 15 shrubland sites were selected in each season. A relative density index (RDI) was used to estimate rodent population abundances in autumn since 2004 to 2008. Different environmental variables (NDVI: vegetation index, NDWI: water index and NDSI: soil index) were obtained from a Landsat satellite image. Climatic data was provided by a near by weather station (INTA Esquel). Direct and delayed correlations between the environmental variables and abundance of rodents were applied. We selected autumn 2005 to compare the environmental variables in sites with high and low rodent abundance within and between the two habitats sampled (shrublands and forests) using discriminant and binary logistic regression analyses. O. longicaudatus showed inter-annual fluctuations and differences in abundance between habitats. In Autumn 2005, we registered highest density of the whole sampling period for both habitats. Spring rainfall (previous year) showed a high positive correlation (r=0.94) with rodent relative density in brush habitats. Discriminant analysis showed significant differences (p=0.0047) between pairs of habitat with different rodent abundance (high and low). Environmental variables were different between forests and shrublands and they were also capable of discriminating high from low abundance forests but not within shrublands. Classification matrix showed that 63.6 % of the sites were correctly reclassified in the original habitats. The logistic regression model only showed association whit vegetation related variables (varianceNDVI and MeanNDVI). Spring rainfall and vegetation (NDVI) showed a significant relationship with the reservoir abundance. These associations suggest evidence for indirect effects of rainfall (in rodent) through vegetation. These results allow us to arrive only to partial conclusions about the rodent dynamics.