Abstract The ability of pipelines to store gas by increasing their operating pressure, or linepacking, is a common operational practice used to mitigate future operational uncertainty. The optimal operation of a gas pipeline network considering linepacking is determined by weighing the trade-off between storing linepack and compressor power consumption. Existing compressor performance models do not accurately capture the rigorous nonlinear operating relationships, and the more accurate widely-used models are computationally complex. This paper develops a novel integer-linear data-driven compressor performance model which is shown to be both more accurate than the best existing model, and less computationally complex. An integer-linear gas transportation model that captures future operational uncertainty using a two-stage multi-period stochastic framework is introduced and solved in a case study on a subnetwork of the Norwegian natural gas network. The case study demonstrates the novel model is highly accurate and can be optimized quickly enough for real-time decision support.