Northern lakes and ponds are abundant and emit large amounts of the potent climate forcer methane to the atmosphere at rates prone to change with amplified Arctic warming. In spite of being important, fluxes from surface waters are not well understood. Long-term measurements are lacking and the dominant and irregular transport mode ebullition (bubbling) is rarely quantified, which complicate the inclusion of lakes and ponds in the global methane budget. This thesis focuses on variations in emissions on both local and regional scales. A synthesis of methane fluxes from almost all studied sites constrains uncertainties and demonstrates that northern lakes and ponds are a dominant source at high latitudes. Per unit area variations in flux magnitudes among different types of water bodies are mainly linked to water depth and type of sediment. When extrapolated, total area is key and thus post-glacial lakes dominate emissions over water bodies formed by peat degradation or thermokarst processes. Further, consistent multiyear measurements in three post-glacial lakes in Stordalen, northern Sweden, reveal that seasonal ebullition, primarily driven by fermentation of acetate, can be predicted by easily measured parameters such as temperature and heat energy input over the ice-free season. Assuming that most water bodies respond similarly to warming, this thesis also suggests that northern lakes and ponds will release substantially more methane before the end of the century, primarily as a result of longer ice-free seasons. Improved uncertainty reductions of both current and future estimates rely on increased knowledge of landscape-level processes related to changes in aquatic systems and organic loading with permafrost thaw, as well as more high-quality measurements, seldom seen in contemporary data. Sampling distributed over entire ice-free seasons and across different depth zones is crucial for accurately quantifying methane emissions from northern lakes and ponds.