The excitation energy levels of two-dimensional (2D) materials and their one-dimensional (1D) nanostructures, such as graphene nanoribbons (GNRs), are strongly affected by the presence of a substrate due to the long-range screening effects. We develop a first-principles approach combining density functional theory (DFT), the GW approximation, and a semiclassical image-charge model to compute the electronic band gaps in planar 1D systems in weak interaction with the surrounding environment. Application of our method to the specific case of GNRs yields good agreement with the range of available experimental data and shows that the band gap of substrate-supported GNRs are reduced by several tenths of an electronvolt compared to their isolated counterparts, with a width and orientation-dependent renormalization. Our results indicate that the band gaps in GNRs can be tuned by controlling screening at the interface by changing the surrounding dielectric materials.