Alternative cropping systems are promoted to reduce nitrogen (N) losses in the environment and mitigate greenhouse gas (GHG) emissions. However, these supposed benefits are not fully known, rarely studied together and on the long-term. Here, we studied the N inputs, N exports, soil organic N (SON) storage, N leaching, gaseous N emissions and GHG balance in a 19-yr field experiment comparing four arable cropping systems without manure fertilization, under conventional (CON), low-input (LI), conservation agriculture (CA) and organic (ORG) managements. The N surplus, i.e. the difference between total N inputs and exports, was lowest in LI (43 kg ha(-1) yr(-1)), intermediary for CON and ORG with 63 kg ha(-1) yr(-1) and highest in CA (163 kg ha(-1) yr(-1)). CA and ORG received high amounts of N derived from biological fixation from alfalfa. The annual SON storage rates markedly differed between CA (55 kg ha(-1) yr(-1)) and both CON and LI (13 and 6 kg ha(-1) yr(-1)), with intermediary value in ORG (30 kg ha(-1) yr(-1)). N leaching, calculated using soil mineral N measurements, reached an average of 21 kg ha(-1) yr(-1) and did not significantly differ between treatments, The gaseous N emissions (volatilization + denitrification), calculated as the difference between N surplus, SON storage and N leaching, ranged from 12 kg ha(-1) yr(-1) in ORG to 83 kg ha(-1) yr(-1) in CA. N2O emissions were continuously monitored with automatic chambers during 40 months. They varied from 1.20 kg ha(-1) yr(-1) in LI to 4.09 kg ha(-1) yr(-1) in CA system and were highly correlated with calculated gaseous N emissions. The GHG balance, calculated using SOC and N2O measurements, varied widely between systems: it was highest in CON and LI, with 2198 and 1763 kg CO2eq ha(-1) yr(-1) respectively. In CA, the GHG balance was much more favourable (306 kg CO2eq ha(-1) yr(-1)), despite important N2O losses which partly offset the benefit of SOC storage. ORG was the system with the smallest GHG balance (-65 kg CO2eq ha(-1) yr(-1)), acting as a CO2 sink in the long-term. Similar trends were observed when GHG was expressed per unit of N input or N exported. The N surplus alone was not a good indicator of the N fate in the four agricultural systems. Complementary predictors of N losses and GHG balance are required to obtain a true overview of the C and N environmental impacts of cropping systems. On an operational point of view, these results should lead to investigate the variability of the GHG emissions within each cropping system.