Canada�s peatlands hold more than half of soil organic carbon stored in all Canadian soils and >90% of these peatlands are in the boreal and subarctic regions, which are susceptible to climate warming. Climate models project the rate of warming will continue through the 21st century, with the greatest warming occurring during the non-growing season (NGS). Given that a major process responsible for carbon dioxide (CO2) emissions and microbial respiration, increases with warming even at sub-zero temperatures, the CO2 emissions during the NGS are expected to increase as the climate warms. Here, we present the governing environmental variables of NGS CO2 emissions in a peatland to infer how these emissions may respond to future climate warming. We developed a machine-learning model whose results imply that changes in soil moisture, soil temperature, snow-cover, and photosynthesis are the primary drivers of net CO2 fluxes during the NGS. The model was applied to a 13-year continuous record of eddy covariance flux measurements at Mer Bleue Bog. We determined that only seven variables were needed to reproduce carbon fluxes, which were most sensitive to net radiation above the canopy, soil temperature, wind speed and soil moisture. Next, we used regional climate projections to forecast future changes in peatland net ecosystem exchange of CO2 during the NGS. Under the highest radiative forcing scenario, the NGS Mer Bleue peatland CO2 emission rates could experience a 103% increase by 2100. Our results highlight the potential for a strong positive climate feedback loop from accelerated peatland carbon loss.