Generalizing and transferring spatial models: A case study to predict Eurasian badger abundance in Atlantic Spain

Even when spatially explicit models are published in accessible international journals, they are rarely reused by the scientific community. This is unfortunate, as these models contain useful information to develop further research and to support decision-making policies. In the absence of a major study on Eurasian badger (Meles meles) abundance in Atlantic Spain, and given the potential role of this species in Mycobacterium bovis epidemiology, we aimed to predict broad-scale badger abundance by generalizing published models for sett suitability within the UK (one calibrated for Northern Ireland and another one for England and Wales). The UK models used fine-resolution environmental predictors that were not available for Spain. Thus, we generalized the models using the outputs of the published models as response variables and calibrated new models using broad-scale environmental predictors. The new model derived from that for Northern Ireland accurately predicted the badger abundance (field data for 48 1 km x 1 km squares in 12 localities) in Atlantic Spain, and indicated a high potential for the species in lowland pastoral areas. The mean density of badgers in the study area was 3.81 adults/km(2) (3.0 +/- 1.3 adults/group), which is higher than in Mediterranean areas in Spain, but lower than in some areas in England where badgers contribute to M. bovis maintenance. We provide the first example of generalization of published spatial models, and confirm that this procedure allows for more efficient use of research funding, by generating new information of relevance, in this case study, for badger management and for understanding M. bovis epidemiology in Atlantic Spain. (C) 2013 Elsevier B.V. All rights reserved.