We aimed to characterize an environmental niche driving the distribution of Crimean-Congo haemorrhagic fever (CCHF) in Turkey, using a georeferenced collection of cases reported between 2003 and 2008 and a set of climate and vegetation features. We used mean monthly air temperatures and Normalized Derived Vegetation Index (NDVI) values, at a resolution of 0.10 degrees, as well as climate features at and below the surface. We computed significant differences in monthly variables between positive and negative sites, within the range of distribution of the tick vector. Seasonal climate (growth season and summer length, accumulated temperatures in winter) and vegetation components (anomalies in NDVI data) were analysed. Fragmentation of habitat was obtained from NDVI monthly data at a resolution of I km. Neither single climate or vegetation variables, nor any individual seasonal component, accounted in both space and time for the delineation of areas of disease although accumulated temperatures in winter consistently showed lower values in areas where the disease was reported. Coherent and significant differences between disease-containing and disease-free sites were found when habitat fragmentation and connectivity were examined. High fragmentation and connectivity were unambiguously associated with sites where disease is reported and accounted for the spatial spread of cases in 2003-2008. CCHF cases were always associated with areas of highly fragmented and well-connected patches within the range of the tick vector, while there were no reports from areas with low fragmentation. There was a linear relationship between degree of fragmentation and case incidence. The implications of these findings are discussed with reference to the concept of disease spread through networks of connected spots with high densities of infected vectors and social factors driving different human activities in sites of high fragmentation.