Characterizing permeability at a regional scale where fracture distributions are heterogeneous can be aided by defining hydrostructural domains. A hydrostructural domain approach is applied to a fracture data set for Mayne Island, one of the Gulf Islands in British Columbia, Canada. Fracture domains were defined using changes in fracture intensity, and are represented and modeled using a stochastic, discrete fracture-network approach. Models that statistically honor field data were constructed for representative stations for three hydraulically distinct, hydrostructural domains: “highly” fractured, interbedded mudstone and sandstone (IBMS-SS) (<10-cm spacing), “less” fractured sandstone (LFSS) (>1.0-m spacing), and fault and fracture zones (FZ). The highly fractured IBMS-SS and FZ domains have a greater potential porosity compared to the LFSS domain due largely to greater fracture intensities. The two highly fractured domains (IBMS-SS and FZ) have an average permeability, on the order of 10−13 m2, due to enhanced fracture-network connectivity. In contrast, the LFSS domain has an average permeability of 10−14 m2. The possibility of increased infiltration rates within FZ domains, coupled with a high-storage potential relative to the other domains suggests that fault zones with similar characteristics are likely zones of recharge. As a result, these recharge zones have an increased capacity to store and transmit infiltrated water throughout the interconnected fracture network. This study demonstrates that hydrostructural domain modeling provides a good foundation upon which to simulate flow and transport in regional groundwater resource studies.