Distribution of the hantavirus reservoir Oligoryzomys nigripes in Argentina: choosing spatial models for the actual and potential distribution of the black-footed colilargo

Abstract

A first step in understanding the ecology of rodents as reservoirs and their relation with the disease they transmit is to determine their geographical distribution. This distribution can be modeled as a function of environmental variables. We georeferenced an extensive database of hantavirus reservoir Oligoryzomys nigripes (Muridae: Sigmodontinae) records in Argentina and used generalized linear models (GLMs) and genetic algorithm for rule-set prediction (GARP) to model the presence probability of this rodent as a function of multiple environmental variables. The GLMs correctly classified 86% of the sites and gave a good prediction area. The GLMs with a spatial term resulted in a probable presence area that matched the rodent occurrences too tightly, and thus was not useful to speculate on potential distribution. The GARP model resulted in a broader probable presence area for the black-footed colilargo than regression models. We suggest that the GLMs without spatial term reflect the actual distribution and should be considered for hantavirus most urgent control plans, whereas the GARP model could be regarded as the most widespread potential distribution and thus considered in long-term plans.