[en] This report describes the methods, analyses, and conclusions of the modeling team in the production of the 2010 revision to the geological discrete fracture network (DFN) model for the Olkiluoto Site in Finland. The geological DFN is a statistical model for stochastically simulating rock fractures and minor faults at a scale ranging from approximately 0.05 m to approximately 565m; deformation zones are expressly excluded from the DFN model. The DFN model is presented as a series of tables summarizing probability distributions for several parameters necessary for fracture modeling: fracture orientation, fracture size, fracture intensity, and associated spatial constraints. The geological DFN is built from data collected during site characterization (SC) activities at Olkiluoto, which is selected to function as a final deep geological repository for spent fuel and nuclear waste from the Finnish nuclear power program. Data used in the DFN analyses include fracture maps from surface outcrops and trenches, geological and structural data from cored drillholes, and fracture information collected during the construction of the main tunnels and shafts at the ONKALO laboratory. Unlike the initial geological DFN, which was focused on the vicinity of the ONKALO tunnel, the 2010 revisions present a model parameterization for the entire island. Fracture domains are based on the tectonic subdivisions at the site (northern, central, and southern tectonic units) presented in the Geological Site Model (GSM), and are further subdivided along the intersection of major brittle-ductile zones. The rock volume at Olkiluoto is dominated by three distinct fracture sets: subhorizontally-dipping fractures striking north-northeast and dipping to the east that is subparallel to the mean bedrock foliation direction, a subvertically-dipping fracture set striking roughly north-south, and a subvertically-dipping fracture set striking approximately east-west. The subhorizontally-dipping fractures account for most (65 % - 75 %, depending on fracture domain) of the total observed fracture intensity at Olkiluoto. Inside a given fracture domain, fracture intensity is adequately represented as a Gamma distribution at the 10 m scale and larger, though near-surface depth-dependence is noted in the foliation-subparallel set in a few fracture domains. There is little difference between open and sealed fractures, though flowing fracture intensity decreases rapidly below approximately -200 masl. Fracture locations are modeled assuming a 3D Poisson point process. Two alternative size models are assessed; the models differ in terms of whether the power-law scaling exponent is derived solely from surface outcrop data (OSM) or from a combination of outcrop-scale, tunnel-scale, and deformation-zone scale data (TCM). The consequences of the uncertainty in the model parameterization is identified and quantified, with the largest effects (on the order of five to seven times) due to uncertainty in the size model. (orig.)