Greenhouse gas emissions (GHG) resulting from forage production contribute a major share to ‘livestock’s long shadow’. A 2-year field experiment was conducted at two sites in northern Germany to quantify and evaluate the carbon footprint of arable forage cropping systems (continuous silage maize, maize–wheat–grass rotation, perennial ryegrass ley) as affected by N-fertiliser type and N amount. Total GHG emissions showed a linear increase with N application, with mineral-N supply resulting in a steeper slope. Product carbon footprint (PCF) ranged between –66 and 119 kg CO₂eq/(GJ net energy lactation) and revealed a quadratic or linear response to fertiliser N input, depending on the cropping system and site. Thus, exploitation of yield potential while mitigating PCF was not feasible for all tested cropping systems. When taking credits or debts for carbon sequestration into account, perennial ryegrass was characterised by a lower PCF than continuous maize or the maize-based rotation, at the N input required for achieving maximum energy yield, whereas similar or higher PCF was found when grassland was assumed to have achieved soil carbon equilibrium. The data indicate potential for sustainable intensification when cropping systems and crop management are adapted to increase resource-use efficiency.