3.1.1. Spatiotemporal variations of nitrogen in water
In water, the spatial variation of nitrogen pollutants (Fig. 2). As shown in Fig. 2a, the consistence of NO3−-N was between 0.31 and 1.46 mg/L. On the whole, the consistence of NO3−-N fall and then raise along flow direction. In the regulation pond, the NO3−-N consistence droped by 0.13 mg/L, which may be caused by the adsorption of sediment in the regulation pond (Tunçsiper 2020). Along the water flow direction, NO3−-N flowed from the regulation pond to the SSF CW and SF CW, respectively. In the SSF CW, the consistence of NO3−-N changed greatly, and other study has reported that due to the presence of a large number of microbial (Zhao et al. 2021), NH4+-N could be easily converted into NO3−-N by nitrification. Meanwhile, the NO3−-N consistence decreased to 0.36 mg/L in the SFW, which may be related to the absorption of plants such as reeds. From the landscape lakes to the water outlet, the NO3−-N consistence increased slightly from 0.31 to 0.42 mg/L. This is attributed to the microorganism produced part of NO3−-N through nitrification (Li et al. 2019).
Similarly, the spatial changes of NH4+-N were shown in Fig. 2b. Overall, the consistence of NH4+-N between 0.15 and 0.6 mg/L and showed first decreased and then increased along flow direction. Compared with other functional areas, the content of NH4+-N in SSF CW was relatively low (0.15 mg/L). Our obtained results were consistent with previous research, who assumed that nitrification by microorganisms and adsorption by substrates will occur in SSF CW (Langergraber et al. 2009; Zhao et al. 2021). The spatial variation of NO2−-N in hybrid CW was shown in Fig. 2c. The maximum consistence of NO2−-N was not more than 0.022 mg/L in hybrid CW, which could be ignored. This could be due to NO2−-N was the intermediate product of nitrification, and easy to be converted (Li et al. 2018). At the water inlet, the NO2−-N consistence was relatively high, which may be caused by the discharge from the sewage treatment plant.
As shown in Fig. 2d, the consistence of TN ranged from 1.72 to 3.64 mg/L, which indicated a downward trend overall and slightly increased at the outlet. It could be inferred from the Fig. 2 that NO3−-N has a large proportion in TN and was the most important form of nitrogen.
Correspondingly, Fig. 3 showed that the temporal variation of nitrogen from August to December in hybrid CW. Overall, the consistence of NO3−-N first decrease and then increased with time (Fig. 3a). As the temperature decreases slowly, plants were still undergoing absorption despite the hindrance of nitrification, causing the content of NO3−-N dropped from 1.02 to 0.28 mg/L before October. Next, its gradually increased to 0.74 mg/L in December, which could be attributed to the large number of plant deaths and could not absorbed by temperature decrease. Meanwhile, when the temperature was lower than 15℃ in December, nitrification and denitrification could not be carried out effectively, leading to the slightly accumulated of NO3−-N.
Analogously, Fig. 3b showed the temporal variation of NH4+-N. Observed that the NH4+-N consistence shows an overall upward trend with time (from 0.10 to 0.63 mg/L), which could be due to the slowing down of nitrification and the large number of plant death as the temperature gradually decreases (Ding et al. 2018). As shown in Fig. 3c, the consistence of NO2−-N changes with time. Interestingly, the fluctuation range of NO2−-N content was smaller in time. Hence, it could be included that the temporal and spatial changes of NO2−-N were consistent.
The temporal variation of TN consistence was shown in Fig. 3d. From August to September, it showed an upward trend from 2.20 mg/L to 2. 66 mg/L. The variation trend of TN consistence from September to December was similar to NO3−-N. The TN consistence showed a downward trend from 2.66 to 1.70 mg/L, and then gradually increased to 2.05 mg/L in December. It can be concluded from Fig. 3 that NO3−-N accounts for the highest proportion of TN in the temporal variation, which was also reported in previous studies (Li et al. 2015).
3.1.2. Spatiotemporal variation of nitrogen in sediments
The spatial variation of nitrogen pollutant consistence in sediments of hybrid CW was shown in Fig. 4. As shown in Fig. 4a, the content of NO3−-N in sediments flowed from the water inlet to the SSF CW and the SF CW, and all of them showed a slight downward from 2.37 to 1.19 and 1.37 mg/kg, respectively. This is attribute to the substrate adsorption of in the SF CW and the plants absorption in the SSF CW (Saeed and Sun 2012). In the landscape lake, the NO3−-N consistence decreased to 0.72 mg/kg, which may be due to the low DO in the landscape lake sediments. At the same time, in the hypoxic environment, the nitrification of microorganisms slows down, and it is conducive to the progress of denitrification (Hong et al. 2019). After that, it showed a slight upward trend at the water outlet (up to 1.32 mg/kg), which may be due to the flow rate of the water outlet was slow and the sediment fully adsorbed NO3−-N in water.
The spatial variation of NH4+-N in sediments was shown in Fig. 4b. At the water inlet, the NH4+-N consistence a was 103.28 mg/kg. The NH4+-N consistence from the regulation pond to the water outlet was 51.52 to 68.48 mg/kg. Except for the water inlet, the NH4+-N consistence in other regions has little difference. However, it was obviously higher in sediments than that in water, which was mainly attributed to two aspects: on the one hand, the adsorption of the matrix leads to the migration of NH4+-N in water to sediments (Wu et al. 2018); On the other hand, the NH4+-N in sediments mainly came from the mineralization process of total organic nitrogen (TON) (Zhu et al. 2019). And the low DO in the sediments slows down nitrification, which lead to an upward trend of NH4+-N accumulation in the sediments. According to Fig. 4c, the consistence of NO2−-N was extremely low due to poor stability (range from 0.009 to 0.032 mg/kg).
The spatial variation of TN in sediments was shown in Fig. 4d. The variation trend of TN consistence in sediments was similar to NH4+-N. The consistence of TN ranged from 252.40 to 414.84 mg/kg. TN at the water inlet was high (414.84 mg/kg), then decreased to 358.76 mg/kg in the regulation pond, and with little change in the SSF CW and SF CW (337.88 and 347.00 mg/kg respectively). It decreased to 252.40 mg/kg in the landscape lake, and final recovered slightly to 253.60 mg/kg.
The temporal variation of NO3−-N in sediments was shown in Fig. 5a. The NO3−-N consistence changed in a small range from August to December (ranging from 0.82 to 1.37 mg/kg), which is related to the consistence of DO. Low consistence of DO in sediments, is key limiting nitrification for removal of NO3−-N.
The NH4+-N consistence fluctuated greatly from August to December. In August, the consistence of NH4+-N was only 24.00 mg/kg. This is related to the migration of pollutants in water. By comparison, it is found that the consistence of this pollutant in the water is also low. After that, the NH4+-N consistence showed an upward trend in September and October (95.12 and 91.76 mg/kg, respectively). This may be due to the increase of NH4+-N content in water, resulting in a corresponding increase in sediments, and the mineralization of organic nitrogen also produced part of NH4+-N (Zhu et al. 2019). However, the NH4+-N consistence in November and December is 56.80 and 60.16 mg/kg, respectively. Since the effect of temperature on microorganism activity and mineralization, NH4+-N consistence decreased slightly (Ding et al. 2018). As shown in Fig. 5c, the NO2−-N consistence remained at a low level (much less than 0.032 mg/kg) and didn’t change significantly over time.
The variation trend of TN in sediments from August to December was similar to that of NH4+-N. It increased from 246.36 to 414.84 mg/kg (Fig. 5d), then decreased to 282.88 mg/kg in November, and then increased slightly in December (345.92 mg/kg). Therefore, it could be known that NH4+-N accounts for the highest proportion of TN, which had been similarly reported in previous studies (Zhu et al. 2019).