In the future, the risks of oil spills in the Arctic marine ecosystem will increase due to growing economic activities linked to the retreating ice cover and climate change. Microorganisms play a crucial role in biodegradation processes. However, little is known about the self-cleaning capacity of the natural microbial communities in cold, mostly sea ice-covered, nutrient-limited Arctic marine systems. Here we hypothesize that physical oceanographic conditions and site-specific biogeochemical properties such as nutrient variations define spatio-temporal oil-biodegradation rates and community patterns. Thus, we deployed, in our study areas in West Greenland, adsorbents coated with thin oil films at three locations along Vaigat Strait at 300 and 500m water depth and one in Godhåbsfjord at 650m. After 1- and 3.5-months exposure time, 16S rRNA gene amplicon sequencing on the harvested biofilms revealed strong spatio-temporal variations in the bacterial community composition. For example, after one month, gammaproteobacterial P13-46 and the obligate alkane degrader Oleispira (both Oceanospirillales) were the most abundant taxa in a rather diverse community. After 3.5 months, the β-diversity decreased and the oil-degrading community shifted towards polycyclic aromatic hydrocarbon degraders showing higher abundances of gammaproteobacterial Marinicella, C1-B045, and Alphaproteobacteria compared to the one-month community. At Vaigat Strait, communities from different sites clustered closely together and were distinct from the Godhåbsfjord community (spatial variation: R2=0.53 one month; R2=0.37 3.5-months). Furthermore, spatial community variations are mainly caused by site-specific biogeochemical properties - β-diversity and nutrient concentrations are correlated, which in turn determines the microbial biodegradation capacity that contributes to a recovering ecosystem.