Discussion
In the present study, long-term N loading affected the abundance of nematode trophic groups in shallow soil of a subtropical Chinese fir forest. The N3 treatment, for example, significantly increased total, bacteria- and fungal-feeding nematode abundances but significantly decreased the abundance of plant-feeding group (Fig. 1), which is consistent with the first hypothesis. Because the population dynamics of nematode trophic groups are closely related with their functional roles (Mueller et al. 2016; Shaw et al. 2019), N addition would also affect nematode ecological indices, i.e., MI25, EI, SI, BI, and CI (Fig. 3). For example, N loading significantly decreased the MI25 value at 0-20 cm soil depth but not at deeper soil layer, which supported that N loading only affected bacterial-feeding nematodes in the shallow soil layer. The values of MI25 are determined by c-p2 to c-p4 values (Zhao et al. 2014), however, the values of c-p3 and c-p4 guilds did not respond to N loading (Fig. S1 and S2), so the increase of bacterial- and fungal-feeding nematode abundances to N loading was caused by the increase in the abundance of c-p2 values in our study. Therefore, we considered that the long-lived functional guilds with higher c-p values, e.g., the OP group, are more vulnerable to environmental changes and require longer time to recover from disturbances than fungal- and bacterial-feeding nematodes (Ferriset al. 2001; Liang et al. 2009). Consistent with that view, our results showed that N addition tended to reduce the abundance of OP group. Furthermore, the SI value, which is mainly determined by the abundance of omnivorous and predatory nematodes (Ferris et al. 2001), also declined at 0-20 cm soil depth in response to N loading (Fig. 2d). The SI value is generally found to be high in undisturbed natural systems (Ferris et al. 2001; Zhao et al. 2013), which suggests that N addition may bring some negative effects on soil ecosystem of the subtropical forests.
Another interesting finding is that N loading decreased the EI value, but increased the CI value (Fig. 3). EI reflects resource availability and decomposing capacity of primary decomposers of soil food web, while CI is an indication of fungal-feeding nematodes as the dominant energy channel in soil food web (Ferris et al. 2001; Shaw et al.2019). Therefore, the consideration of both indices provides useful information on soil nutrient status. A previous study suggested that a higher EI value and a lower CI value indicated a resource-rich soil environment dominated by bacterial-feeding nematodes (Liang et al. 2009). Our finding that N loading decreased the EI value and increased the CI value indicates that N loading diminished the relative importance of a bacterial-dominated energy channel. Moreover, the, A significant increase in the BI value indicated a positive effect of N loading on both Fu2 and Ba2 guilds (Fig. S2), since BI value is mainly determined by Fu2 and Ba2 guilds (Zhaoet al. 2014). N loading also increased the abundance of fungal-feeding nematodes, suggesting that the relative importance of fungal energy channel increased with N loading. However, N loading significantly reduced the abundance of plant-feeding nematodes (Fig. 1). The results differed from those of Shao et al. (2017), who found that N inputs did not significantly affect the abundance of plant-feeding nematodes. The reason would be that we used a gradient of N-loading rates, which probably provides more information than a single rate in that Shao et al. (2017). Because the quantity of nitrogen applied regulated the responses of soil nematode communities (Lokupitiya et al. 2000; Wei et al. 2012; Liet al. 2013).
It is worth noting that N addition had no effects on the diversity of soil nematodes (Fig. 2). Both the Shannon-Wiener diversity and Margalef richness were similar with or without N loading and significantly decreased with soil depth. Our results are similar with previous results from a secondary tropical forest (Zhao et al.2014) but not with results from a temperate forest (Sun et al.2013) or a grassland (Wei et al. 2012), where N addition reduced soil nematode diversity. The Simpson dominance index in N2 treatment was significantly higher at 40-60 cm soil depth than at the other two soil depths, indicating that a certain concentration of N addition might increase the dominant position of some nematode species in deep soil. Our results indicated that nematode diversity only showed some dynamics along different soil depth but not N treatments. The explanation would be that the changes of soil nematode community structure did not always reflect the diversity, since the complexity of soil micro-food web in natural ecosystems (Rooney and McCann, 2012; Li et al. 2013). Our recent report also showed that changes in soil microbial community composition would mainly contribute to the litter decomposition process compared with soil microbial diversity (Wu et al. 2019), which supports the phenomenon of soil nematode in this study.
Structural equation modeling was performed to test the second hypothesis how N addition affects the trophic groups of soil nematodes with direct or indirect pathways. We found that N loading reduced the abundance of plant-feeding nematodes either directly or indirectly via an alteration in soil pH. The negative effect of N loading on plant-feeding nematodes may result from the negative effects on plants, for that N addition may reduce plant diversity, aboveground biomass, and fine root biomass of understory vegetation (Lu et al. 2010; Lu et al. 2011; Wuet al. 2013). A decrease in the aboveground biomass and fine root biomass of understory vegetation may lead to a reduction in food resources for plant-feeding nematodes (Wu et al. 2013; Fanet al. 2014). Our previous reports also indicated that N loading decreased soil pH in the studied region (Wu et al. 2013; Shenet al. 2019) and other report also proved it in Northeast China (Liang et al. 2009). Soil acidification may indirectly affect plant-feeding nematodes by decreasing soil base cations and altering the abiotic environment (Chen et al. 2015).
N loading increased the abundance of fungal- and bacterial-feeding nematodes apparently through an increase in the concentration of soil available nutrients, which presumably supported increases in food resources for them. On one hand, the decrease of soil pH in our study did not exert a negative impact on the fungal- and bacterial-feeding nematodes, which was supported by the SEM analysis. Indeed, an increase in the individuals of bacterial-feeding nematodes was measured. Commonly, soil acidification induced by N addition can inhibit soil microbial growth and bacterial biomass accumulation (Rousk et al. 2010; Chen et al. 2015; Li et al. 2018b), which is supposed to lead a reduction of food resources for soil fungal- and bacterial-feeding nematodes, and eventually result in a decline in their abundance. On the other hand, N loading may change the community structure of bacterial-feeding nematodes by changing their environmental adaptations. Generally, soil condition with plentiful food resources are beneficial to those nematodes with short generation cycles (Bongers, 1990; Bongers & Ferris, 1999; Liang et al. 2009), i.e., Ba1 and Ba2 guilds. The elevated N stress made bacterial-feeding nematodes change their survival strategy and transform community structure to abundant r -strategy colonizer (Ba1 and Ba2 guilds) under N addition (Zhao et al. 2015a). The co-occurrence network of nematodes taxa analysis showed that nodes and edges varied in different N treatments (Fig. 4), which supported our inference that N addition can change the community structure and stability of soil nematode communities (Zhao et al. 2015b). All these results indicate that nematodes changed by N addition, whether they can finally reach a new equilibrium in response to gradually elevated N loading requires more studies.