Climate variability can affect the epidemiological patterns of human infectious diseases both directly and indirectly. Direct mechanisms are those involving the effects of the physical variables (temperature; humidity etc) upon pathogens, vectors and, less frequently, on the biology of human hosts. However, most of the effects of climate on the health of human populations are mediated through complex ecological and social pathways. Straightforward correlation between a climate variable and variations in the intensity, frequency and distribution of infectious diseases of humans are rarely observed. They occur against a background of social vulnerability and environmental structure in a well defined territorial unit. The landscape level of analysis seems to be the most appropriate for the assessment of the impacts of climate as regards infections. The main advantage is that, at the scale of landscapes, under a relatively uniform climatic pattern, we can account for the differences of the social-environmental variability and its relative role in the production of infections, which should be part of any assessment of climate/health linkages. The landscapes, either natural, cultivated or built are easily recognized entities for operational purposes and the role played by human societies in shaping the physical environment is best seen at the landscape level. Therefore it should be the preferred locale for interventions to reduce the vulnerability and mitigate the impacts of climate. These aspects are illustrated with a comparison of the effects of the drought associated with the ENSO phenomenon in two regions of Brazil: the Amazon and the northeast.