Urban environmental clustering to assess the spatial dynamics of Aedes aegypti breeding sites

The identification of Aedes aegypti breeding hotspots in urban areas is crucial for the rational design of control strategies against this disease vector. Remote sensing and geographic information systems offer valuable tools for mapping habitat suitability of a given area. However, predicting species occurrences by means of probability distribution maps based on transversal entomological surveys has limited utility for local authorities. The aim of the present study was to carefully examine the temporal evolution of the number of houses infested with immature stages of Ae. aegypti in each individual neighbourhood and to explore the value of producing environmental clusters generated with information provided by remotely sensed variables to explain the observed differential temporal behaviour. Entomological surveys were conducted between 2011 and 2013 throughout a small town in Argentina registering the number of houses with containers harbouring immature stages of Ae. aegypti. A SPOT 5 satellite image was used to obtain land cover variables, which were subsequently submitted to k-means partitioning for grouping neighbourhoods into four environmental clusters. Finally, a generalized linear model was fitted showing that the number of houses found to be positive for Ae. aegypti was jointly affected by the interaction between environmental clusters and the year of sampling. Moreover, the number of positive houses in one of the clusters was 9.5 times higher (P<0.005, SE=0.37) in 2013 than in 2012, but we did not observe any other statistically significant increases.