Engineering the wettability and microstructure of gas diffusion layers offers a powerful strategy to optimize water management in polymer electrolyte fuel cells. To this goal, we recently developed a radiation grafting technique to synthesize GDLs with patterned wettability. Despite the promise of this approach, current designs are empirically-driven which hampers the development of advanced wettability patterns. To design materials with improved transport characteristics over a range of water saturations, physically representative models can be employed for the bottom-up design of gas diffusion layers with local variations in hydrophilicity. In this paper, pore network models using topology and size information extracted from high resolution tomographic images of three common gas diffusion layer materials are presented with patterned wettability. We study the influence of the substrate microstructure, the hydrophobic coating load, and the hydrophilic pattern width. It is shown that tuning the wettability leads to enhanced phase separation and increased diffusive transport which is attributed to decreased gas phase tortuosity. The network model elaborates on previous experimental studies, shedding light on the effectiveness of the radiation pattern transference and the importance of matching the masking pattern with the substrate microstructure. The work opens the door for exploration of advanced patterns, coupled with flow from gas flow field designs.