Nitrogen enrichment is pervasive in forest ecosystems, but its influence on understory plant communities and their stoichiometric characteristics is poorly understood. We hypothesize that when forest is enriched with nitrogen (N), the stoichiometric characteristics of plant species explains changes in understory plant diversity. A 13 year field experiment was conducted to explore the effects of N addition on foliar carbon (C): N: phosphorus (P) stoichiometry and understory plant species richness in a subtropical Chinese fir forest. Four levels of N addition were applied: 0, 6, 12, and 24 g m^-2 yr^-1. Individual plant species were categorized into resistant plants, intermediate resistant plants, and sensitive plants based on their response to nitrogen addition. Results showed that N addition significantly decreased the number of species, genera and families of herbaceous plants. Foliar N:P ratios were greater in sensitive plants than resistant or intermediate resistance plants, while intrinsic water use efficiency showed an opposite trend. However, no relationship was detected between soil available N and foliar N, and soil N:P and foliar N:P ratios. Our results indicated that long-term N addition decreased the diversity of understory plants in a subtropical forest. Through regulating water use efficiency with N addition, sensitive plants change their N:P stoichiometry and have a higher risk of mortality, while resistant plants maintain a stable N:P stoichiometry, which contributes to their survival. These findings suggest that plant N:P stoichiometry plays an important role in understory plant performance in response to environmental change of N.

Plant investigation and sampling analyses

Understory plant communities were investigated in September 2011, 7 years after the start of the N addition treatment (Wu et al. 2013). One 5 m × 5 m subplot was established in each 20 x 20 m plot and all plants taller than 5 cm were recorded. Plant richness (family, genera and species) and percentage cover were evaluated within each subplot. Understory plant species were divided into two functional groups: woody plants (including tree seedlings, shrubs and woody vines) and herbaceous plants. In September 2016 (13 years after N addition started), the understory plants were re-assessed using the same methods as in 2011 (Wu et al. 2013). We divided the understory species into three resistance types: resistant plants (RP), intermediate resistant plants (IRP) and sensitive plants (SP) based on the mean presence and absence of the plant species recorded in 2011. Briefly, plant species found in plots with any of the three N addition treatments and in the control (N0) plots, were classified as RP. Plant species found in the N0, N1 and N2 treatment plots, but not in the N3 treatment plots, were categorized as IRP. Finally, plant species found only in N0 treatment plots were categorized as SP.

After the second plant community assessment (species presence and abundance in each treatment plot) in 2016, plant leaves were immediately collected in each plot to measure leaf C, N, P and their C:N:P stoichiometric traits. Leaves from 22 plant species were collected from the twelve plots of four treatments for foliar chemical analysis. For each individual plant, 5 mature leaves were sampled and bulked to make a composite sample. The fresh plant leaves were oven-dried at 60℃ then ball milled prior to analysis of ¹³C and ¹⁵N natural abundances, and total C, N and P. Foliar P was analysed by persulfate oxidation followed by colorimetric analysis (Bao 2000), whereas foliar C and N were analyzed by Elemental Analyzer (Flash 200 EA-HT, Thermo Fisher Scientific, Inc., USA). Foliar δ¹³C and δ¹⁵N was analyzed by Isotope Ratio Mass Spectrometer (Deata V Advantage, Thermo Fisher Scientific, Inc., USA). The standards for foliar ¹³C and ¹⁵N were Pee Dee Belemnite and atmospheric N₂. The calculation as the following:

δ¹³C (δ¹⁵N) = (R_sample/R_standard—1) × 1000 (‰),

where R represents the isotope ratio (¹³C/¹²C or ¹⁵N/¹⁴N) from samples or standards. The analytical precision for δ¹³C and δ¹⁵N were better than 0.1‰ and 0.2‰, respectively. Based on the value of δ¹³C, we calculated the intrinsic water use efficiency (iWUE) according to the descriptions from (Lu et al. 2018).