In October 2009, a diesel pipe connecting two storage tanks at Carlini station (Antarctica) cracked, leaking fuel and contaminating the surrounding soil which, at that time, was covered by a thick snow layer. Three months later, the snow began to melt and the diesel moved with runoff water from the fuel tanks through the residential and laboratory area, towards the Potter Cove. Five months after the spill, soil samples were taken at 10 sites close to, as well as further away, from the spill site. Marine sediment samples were also taken from Potter Cove basin, with the aim to monitor the bacterial communities in relation to the presence of hydrocarbons. One year later, in January 2011, sampling was repeated at approximately the same sites and additional sites that could not be sampled before due to the snow cover. At this time, neither smell nor visual appearance of diesel residues were noticeable on the surface or at 30 cm-depth soil, and measured total petroleum hydrocarbons decreased in the most affected sites, except in front of the pool containing the tanks where the spill originated. At this site, an underground leaking pipe had caused a more recent small event of acute contamination that is being used as a proxy to better understand the effect of the main spill on bacterial communities at Carlini station area. The impact of the spill on soils and sediments was monitored by total carbon and nitrogen content, C/N ratio, presence of hydrocarbons and differences in bacterial community structure. DGGE based on the 16S rRNA gene was used to identify the main changes in community composition over the 14-months period. Most samples showed a quite high diversity few months after the spill, as well as one year later, and only marginal changes in the taxonomic structure were observed during the study period. However, the bacterial community composition in the samples taken near the cracked underground pipe, which experienced a larger input of diesel and less time between impact and sampling, showed clear changes. The DGGE gels exhibited a reduction in diversity and an enrichment of specific bacterial groups. Clone libraries revealed enrichment in Actinobacteria and one Betaproteobacterium closely related to Polaromonas napthalenivorans, indicating an actively growing hydrocarbon-degrading community. After one more year (January 2012), there was no visual presence or smell of hydrocarbons at this site, the same that had been observed after the main spill. Our results show that in Potter peninsula the bacterial communities undergo a fast shift to lower diversity and higher dominance when a sudden spill occurs on a soil with certain history of exposure to hydrocarbons. The unusually fast disappearance of the spilled diesel in such a cold habitat is favoured by the kind of diesel used (containing mostly light aliphatic hydrocarbons), the sandy texture of soil, snowmelt runoff, summer rains and a well adapted microbiota, with a promising oil-bioremediation potential