Graphical abstract Abstract On the premise of maximizing nitrogen elimination in wastewater and recharging the effluent directly to groundwater, two groups of subsurface wastewater infiltration systems (SWIS), a turfgrass-covered system (S-G) and a bare soil system (S-B), were examined under a pollution load of 600 kg N ha−1 yr−1 to investigate the contribution of turfgrass in removing sewage nitrogen. The results showed that compared with S-B, the concentrations of total nitrogen (TN) and nitrate nitrogen (NO 3 −-N) in S-G effluent were significantly lower. In 2014 and 2015, the average concentrations of NO 3 −-N in S-B effluent were 38.9 and 36.8 mg L−1. None of the NO 3 −-N concentrations in S-B effluent satisfied the international standard for drinking water (10 mg L−1 NO 3 −-N). While the average concentrations of NO 3 −-N in S-G effluent in 2014 and 2015 were 2.96 and 0.80 mg L−1, the proportions of NO 3 −-N concentrations in S-G effluent, which were below 10 mg L−1 NO 3 −-N, were 87.7% and 97.9%, respectively. The TN removal efficiency and removal quantity of the S-G were significantly higher than those of the S-B. In 2014 and 2015, the TN removal quantities of S-G were 646 and 647 kg ha−1 yr−1, respectively, and the TN removal efficiencies reached 98.1% and 99.1%, for which the contribution efficiencies of turfgrass on nitrogen removal were 52.5% and 61.0%, respectively. Meanwhile, the TN removal quantities of S-B were 269 and 204 kg ha−1 yr−1, and the TN removal efficiencies were 40.7% and 31.3% in 2014 and 2015, respectively. These obtained results indicated that plants play an important role in SWIS under these conditions. [ABSTRACT FROM AUTHOR]