ANPP and species richness
Across the nine years of fertilization management, compared with the control, Si fertilization only, N fertilization only, and N+Si fertilization significantly increased the mean ANPP by 34.74% in Si, 31.21% in N70, 36.50% in N140, 27.88% in N210, 50.47% in N70Si, 74.02% in N140Si, and 57.48% in N210Si (Fig. 1a). The combination of N and Si fertilization led to a higher mean ANPP than the equivalent amount of N fertilization only. Also, significant interactive effects of Si fertilization, N fertilization, and year on ANPP were observed (Table 1).
Compared with the control, the treatments of N fertilization only and N+Si fertilization led to statistically significant declines in species richness, by 6.67% in N70, 19.66% in N140, 28.14% in N210, and 13.61% in N210Si, while the N70Si and N140Si treatments showed no pronounced effects. The application of N incorporated with Si (N70Si, N140Si and N210Si) significantly increased the species richness by 4.83%, 10.35% and 15.55%, respectively, compared to that without Si fertilization under the equivalent amount of N fertilization only (N70, N140 and N210) (Fig. 1b). Furthermore, the fertilization of N, fertilization of Si, and year had significant interactive effects on species richness (Table 1).
Insert Figure 1 here
Insert Table 1 here
Dominant species stability and species dominance
The dominant species stability and species dominance under all types of fertilization are shown in Fig. 2. The temporal stability of dominant species increased significantly by 21.01%, 23.34%, 13.23%, 40.71% and 36.84% in the N140, N210, N70Si, N140Si and N210Si treatments, respectively, while the treatments of Si and N70 had no statistically significant effects (Fig. 2a). Compared with the control, the N140, N210, N140Si and N210Si treatments increased the dominance by 41.69%, 89.08%, 7.66% and 31.92%, respectively (P < 0.05), while the Si, N70 and N70Si treatments had no significant impacts on species dominance (Fig. 2b, P > 0.05). The fertilization of N+Si had significant interactive effects on dominant species stability and species dominance (Table 2).
Insert Figure 2 here
Species asynchrony, population stability and community temporal stability
Population stability decreased with increasing N fertilization level. Si treatment significantly increased the population stability (Fig. 3a). The measure of species asynchrony was less than 1 (Fig. 3b), indicating that compensatory effects took place among different plant species. In comparison to the control, the N70, N140, N210 and N210Si treatments resulted in statistically significant decreases in species asynchrony by 7.09%, 7.47%, 19.49% and 8.76%, respectively, while the Si treatment significantly increased species asynchrony by 6.04%, but the N70Si and N140Si treatments produced no statistically significant effects (Fig. 3b).
In comparison to the control, Si fertilization only significantly increased community temporal stability by 9.97%, while all N fertilization and all N+Si fertilization treatments significantly decreased community temporal stability. Moreover, N+Si fertilization resulted in higher community temporal stability than under the equivalent amount of N fertilization only (Fig. 3c). N and Si fertilization had interactive effects on species asynchrony, population stability, and community temporal stability (Table 2).
Insert Figure 3 here
Insert Table 2 here
Major factors contributing to community temporal stability
Correlation analysis results indicated that species richness was significantly positively related to population stability (Fig. 4a, R2 = 0.568), community temporal stability (Fig. 4c, R2 = 0.599), and species asynchrony (Fig. 4f, R2 = 0.715). Species asynchrony was significantly positively related to community temporal stability (Fig. 4b, R2 = 0.481), while species dominance was positively related to the dominant species stability (Fig. 4d, R2 = 0.145) and community temporal stability (Fig. 4e, R2 = 0.549). The temporal stability of dominant species, meanwhile, was negatively related to community temporal stability (Fig. 4g, R2 = 0.274) and population stability (Fig. 4h, R2 = 0.432), and the relationship between population stability and community temporal stability was positively significant (Fig. 4i, R2 = 0.580).
Insert Figure 4 here
SEM demonstrated that the increase in species asynchrony induced by Si can directly increase the community temporal stability. Meanwhile, the increase in dominant species stability induced by N fertilization can directly decrease the community temporal stability; and whilst Si fertilization can increase the species richness, N fertilization can decrease it and further impact the degree of species asynchrony, thereby inducing an increase or decrease in community temporal stability. However, Si fertilization can reduce species dominance but N fertilization can increase it, thereby inducing a decrease or increase in community temporal stability (Fig. 5).
Insert Figure 5 here