Using multiple scales of observation in studying the fractures of the bedrock increases the reliability and representativeness of the respective studies. This is because the discontinuities, i.e. the fractures, in the bedrock lack any characteristic length and instead occur within a large range of scales of approximately 10 orders of magnitude. Consequently, fracture models need to be constructed based on representative multi-scale datasets.In this paper, we combine a detailed bedrock fracture study from an extensive bedrock outcrop area with lineament interpretation using light detection and ranging (lidar) and geophysical data. Our study offers lineament data in an intermediary length range (100-500 m) missing from discrete fracture network modelling conducted at Olkiluoto, a nuclear spent-fuel facility in Finland. Our analysis provides insights into multi-scale length distributions of lineaments and fractures and into the effect of glaciations on lineament and fracture data. A common power-law model was fit to the lineament and fracture lengths with an exponent of 1.13. However, the fractures and lineaments might follow distinct power laws or other statistical distributions rather than a common one. When categorising data by orientation, we can highlight differences in length distributions possibly related to glaciations. Our analysis further includes the topological, scale-independent fracture network characteristics. For example, we noticed a trend of decreasing apparent connectivity of fracture networks as the scale of observation increases.