The prediction of methane emissions from high-latitude wetlands is important given concerns about their sensitivity to a warming climate. As a basis for the prediction of wetland methane emissions at regional scales, we coupled the variable infiltration capacity macroscale hydrological model (VIC) with the biosphere–energy-transfer–hydrology terrestrial ecosystem model (BETHY) and a wetland methane emissions model to make large-scale estimates of methane emissions as a function of soil temperature, water table depth, and net primary productivity (NPP), with a parameterization of the sub-grid heterogeneity of the water table depth based on TOPMODEL. We simulated the methane emissions from a 100 km × 100 km region of western Siberia surrounding the Bakchar Bog, for a retrospective baseline period of 1980–1999 and have evaluated their sensitivity to increases in temperature of 0–5 °C and increases in precipitation of 0–15%. The interactions of temperature and precipitation, through their effects on the water table depth, played an important role in determining methane emissions from these wetlands. The balance between these effects varied spatially, and their net effect depended in part on sub-grid topographic heterogeneity. Higher temperatures alone increased methane production in saturated areas, but caused those saturated areas to shrink in extent, resulting in a net reduction in methane emissions. Higher precipitation alone raised water tables and expanded the saturated area, resulting in a net increase in methane emissions. Combining a temperature increase of 3 °C and an increase of 10% in precipitation to represent climate conditions that may pertain in western Siberia at the end of this century resulted in roughly a doubling in annual emissions.