Ticks have medical and economic importance due to their ability to transmit pathogens to humans and animals. In tropical and sub-tropical countries, tick-borne diseases (TBD) are among the most important diseases affecting livestock and humans. The fast spread of ticks and TBD requires a quick development and application of efficient prevention and/or control programs. Therefore, prior investigations on TBD and related vectors epidemiology, for instance, through accurate epidemiological models, are mandatory. This study aims to develop models to forecast suitable habitat for Rhipicephalus microplus distribution in West Africa. Tick occurrences were assembled from 10 different studies carried out in six West African countries in the past decade. Six statistical models (maximum entropy in a single model and generalised linear model, generalised additive model, random forest, boosted regression tree and support vector machine model in an ensemble model) were applied and compared to predict the habitat suitability of R. microplus distribution in West Africa. Each model was evaluated with the area under the receiver operating characteristic curve (AUC), the true skill statistic (TSS) and the Boyce index (BI). The selected models had good performance according to their AUC (above .8), TSS (above .7) and BI (above .8). Temperature played a key role in MaxEnt model, whereas normalised difference vegetation index (NDVI) was the most important variable in the ensemble model. The model predictions showed coastal countries of West Africa as more suitable for R. microplus. However, some Sahelian areas seems also favourable. We stress the importance of vector surveillance and control in countries that have not yet detected R. microplus but are in the areas predicted to host suitable habitat. Indeed, awareness-raising and training of different stakeholders must be reinforced for better prevention and control of this tick in these different countries according to their status.
Keywords: Rhipicephalus microplus; West Africa; boosted regression tree; ensemble modelling; generalised additive model; generalised linear model; maximum entropy; random forest; support vector machine.
© 2022 Wiley-VCH GmbH.