1 INTRODUCTION
Anthropogenic-driven nutrient inputs, namely nitrogen (N) and phosphorus (P) enrichment of terrestrial ecosystems, have been increasing intensively since the Industrial Revolution (Harpole et al., 2011; Liu et al. 2021; Phoenix et al., 2012). Global cycles of N and P have been respectively amplified by c.100% and c. 400%, respectively, due to intensified human activities (Elser et al., 2007). Being two crucial nutrient elements for growing plant, N and P enrichment can profoundly influence plant growth, survival, and reproduction, with subsequent impacts on community structure and ecosystem functioning (Zhao et al., 2018a). Reproduction is an essential function in the life cycles of plants that determine their fitness (Willson, 1983). Seed production is an important index of reproduction that strongly influences the relative ability of species to disperse and establish as seedlings (Liu et al., 2012; Pierce et al., 2014). In addition, seed production can affect the size and extent of soil seed banks, and contribute to the maintenance of plant diversity and species composition (Luzuriaga et al., 2005). Therefore, understanding the effects of N and P enrichment upon seed production is critical for predicting plant community structure, and consequently, ecosystem functioning.
Nitrogen is a limiting nutrient for plant reproduction in terrestrial ecosystems. Numerous studies have demonstrated that N enrichment tends to augment seed production in plants (Bogdziewicz et al., 2017; Li et al., 2016a; Ma & Herath, 2016; Shi et al., 2017). For example, in a temperate steppe, Leymus chinensis produced more seeds via N enrichment through enhanced spikelet and flower differentiation (Wang et al., 2010). P plays a key role in regulating plant reproductive processes because it can significantly affect the partitioning of assimilation products, flowering phenology, root growth, and seed maturation (Petraglia et al., 2014; Wang et al., 2017). Previous studies have reported that plants under P-deficient conditions allocate little to reproductive growth, which manifests as in a shortened flowering period, reduced seed yield, and weakened dispersal ability (Fujita et al., 2014; Groom & Lamont, 2009). Futhermore, the positive, negative or neutral effects of P enrichment on seed production can all occur in terrestrial ecosystems (Sims et al., 2012b; Singh et al., 2018; Wang et al., 2017). For example, although P enrichment negligibly affected the seed production of Stipa kryloii , it did increase that ofArtemisia frigida in a temperate steppe (Li et al., 2017).
The availability of N and P availability, however, may jointly affect seed production since plant growth is predicted to be co-limited by multiple resources (Graciano et al., 2006; Harpole et al., 2016;Harpole & Suding, 2011; Long et al., 2016; Peñuelas et al., 2013). Simultaneously adding N and P enhances ecosystem primary productivity much more than adding either of them alone (Elser et al., 2007; Harpole et al., 2011; Solis et al., 2013). By contrast, such an interactive effect between N and P addition was not found for plant reproduction in a temperate steppe (Li et al., 2017). Further, the growth and reproduction of plants may respond differentially to the levels of nutrient enrichment (Bowman et al., 2006; Tang et al., 2017). A meta-analysis found that plant productivity in meadow steppe is positively related to N addition under low N addition level, but it decreases with increasing N addition under high N addition level (Tang et al., 2017). Nonetheless, few attempts have been made to empirically investigate how N and P addition rates and their interaction could affect plant reproduction allocation and seed production in terrestrial ecosystems, because most nutrient addition studies only include two levels of nutrient treatments (i.e., control vs. nutrient enrichment).
Grassland is one of the major terrestrial ecosystems and covers 40% of the world’s land area (Adams et al., 1990). In this respect, the temperate steppe in northern China is representative of the typical vegetation of the Eurasian grassland biome (Bai et al., 2010; Su et al., 2018). A comprehensive project that used three levels of N and P addition was begun in April 2012, aiming to examine the effects of nutrient enrichment on community structure and ecosystem functioning in a typical temperate steppe of Inner Mongolia, northern China. As part of this long-term project, the present study was done to examine the interactive effects of N and P addition rates upon seed production of the dominant species Stipa krylovii , which is the most common perennial grass in typical temperate steppe ecosystems in China. We sought to address the following specific questions: (1) Do changes in N and P availability and in N:P ratio alter the seed production of the dominant species in temperate steppe ecosystem? (2) Which factors determine seed production of the dominant species under the different nutrient addition treatments?