BackgroundCulicoides imicola KIEFFER, 1913 (Diptera: Ceratopogonidae) is the principal vector of Bluetongue disease in the Mediterranean basin, Africa and Asia. Previous studies have identified a range of eco-climatic variables associated with the distribution of C. imicola, and these relationships have been used to predict the large-scale distribution of the vector. However, these studies are not temporally-explicit and can not be used to predict the seasonality in C. imicola abundances. Between 2001 and 2006, longitudinal entomological surveillance was carried out throughout Italy, and provided a comprehensive spatio-temporal dataset of C. imicola catches in Onderstepoort-type black-light traps, in particular in Sardinia where the species is considered endemic.MethodsWe built a dynamic model that allows describing the effect of eco-climatic indicators on the monthly abundances of C. imicola in Sardinia. Model precision and accuracy were evaluated according to the influence of process and observation errors.ResultsA first-order autoregressive cofactor, a digital elevation model and MODIS Land Surface Temperature (LST)/or temperatures acquired from weather stations explained ~77% of the variability encountered in the samplings carried out in 9 sites during 6 years. Incorporating Normalized Difference Vegetation Index (NDVI) or rainfall did not increase the model's predictive capacity. On average, dynamics simulations showed good accuracy (predicted vs. observed r corr = 0.9). Although the model did not always reproduce the absolute levels of monthly abundances peaks, it succeeded in reproducing the seasonality in population level and allowed identifying the periods of low abundances and with no apparent activity. On that basis, we mapped C. imicola monthly distribution over the entire Sardinian region.ConclusionsThis study demonstrated prospects for modelling data arising from Culicoides longitudinal entomological surveillance. The framework explicitly incorporates the influence of eco-climatic factors on population growth rates and accounts for observation and process errors. Upon validation, such a model could be used to predict monthly population abundances on the basis of environmental conditions, and hence can potentially reduce the amount of entomological surveillance.