Aims Hantaviruses are zoonotic, aetiological agents maintained by rodents of the family Muridae. The occurrence of hantavirus in rodent hosts has been correlated to a number of climatic and environmental factors, including landscape structure. To date, most of these correlative studies have been conducted at moderate to fine spatial resolution. Our aim is to determine whether land cover classes defined at a mapping resolution of 1 km 2 are associated with rodents with antibodies to hantavirus in Paraguay. Location The Republic of Paraguay. Methods A total of 362 rodents from 10 species known to host hantaviruses were tested for the presence of hantavirus antibodies, resulting in 27 seropositive individuals. This data base was then combined with a map of six land cover types derived from coarse resolution remote sensing data to create a series of contingency tables, which were used to relate serostatus to land cover type using nonparametric tests of proportions and qualitative comparison of observed and expected values. Results There was a significant difference in habitat association between seropositive and seronegative rodents when species were pooled. Seropositive rodents were found with disproportionately high frequency in areas where human disturbance in the form of intensive and mosaic agricultural landscapes was present. Main conclusions Human-disturbed land cover classes have a detectable relationship to the hantavirus serostatus of host population rodents when observed at coarse spatial resolutions. Although coarse-grained analysis does not lead to any conclusions as to why agricultural land cover is more likely to harbour seropositive rodents, the relationship between them could form the basis for a monitoring system designed to relate land cover change to potential viral outbreaks in rodents and humans.
