Assessing the spatial distribution of Ixodes ricinus in Scotland using a Bayesian approach

In Northern and Western Europe Ixodes ricinus is the most abundant and widespread tick species and also the vector of several pathogens. Predictive maps with the spatial and seasonal distribution of vectors are important tools in assessing the risk of vector-borne diseases, enabling better targeting of surveillance and control, as well increasing awareness. The objectives of this study are to identify the environmental, climatic, habitat and host related predictors of I. ricinus abundance and to predict I. ricinus abundance in mainland Scotland.

The dataset used comprised questing nymphs counted by scientists during tick abundance surveys performed between 2006 and 2017. To account for the complex structure of the dataset and the existence of spatial heterogeneity, a Bayesian approach was used. Georeferenced variables were selected as possible predictors, based on their biological meaning, statistical significance and lack of correlation with other variables. The model was fitted with two random effects and a zero-inflated Poisson distribution. Candidate models were evaluated using two Bayesian information criteria and the cross-validation method leave-one-out.

The selected model showed positive relationships between tick abundance and temperature in July, roe deer presence and proportion of coniferous and deciduous forest, and a negative association with frost in September. A predictive map was developed based on this model. Although the model choice seemed adequate to capture data complexity, the variability present in the dataset and the lack of data from some areas highlights the need for further data collection to improve the accuracy of predictive maps.