Drained agriculturally used peatlands are hotspots for greenhouse gas (GHG) emissions, especially carbon dioxide (CO2) and nitrous oxide (N2O). To reduce GHG emissions and simultaneously maintain intensive grassland use, raising water levels by subsurface irrigation (SI) is controversially discussed. Both, intensive grassland use and installations of SI may require grassland renewal. We investigated an experimental intervention site (INT) (SI target water levels: -0.30 m) and a deeply drained reference site (REF), both intensive grassland on deep bog peat. After installation of the SI system, a mechanical grassland renewal took place at INT. At both sites, CO2 (eddy covariance), N2O and methane (manual closed chamber technique) were measured. Additionally, soil water was analyzed for nitrogen species. Here, we report on the initial year of GHG measurements including grassland renewal and rising water levels. Overall, GHG emissions were strongly influenced by grassland renewal at INT. Despite progressively rising water levels, soil moisture in the upper centimeters was low and thus grass growth was slow, resulting in an almost complete loss of harvest. This resulted in a net ecosystem carbon balance (NECB) of 4.64 ± 1.03 t C ha−1 containing only 0.57 ± 0.09 t C ha−1 harvest at INT, while NECB at REF was 6.08 ± 1.74 t C ha−1 including harvest from five grass cuts. Methane fluxes were negligible at both sites. Nitrous oxide emissions dominated the GHG balance at INT. With 144.5 ± 45.5 kg N2O–N ha–1 a–1, they were much higher than at REF (3.9 ± 3.1 kg N2O–N ha–1 a–1) and any other values published so far. Peak fluxes occurred when nitrate concentrations in soil water were extremely high, soil moisture was increased, and vegetation development was struggling. This study highlights the risk of grassland renewals on peat soils regarding yield losses as well as high GHG emissions.