1 INTRODUCTION
Climate warming and the increase of extreme climatic events are mostly attributed to the increasing concentrations of greenhouse gases in the atmosphere, especially carbon dioxide (CO2). The coupled biogeochemical cycles of carbon (C) and silicon (Si) in the terrestrial system are deemed a mechanism that impacts the long-term regulation of atmospheric CO2 (Berner, 1992; Parr et al., 2010; Parr & Sullivan, 2011; Song et al., 2012a; Song et al., 2012b). Plants absorb mono-silicic acid from soil solution via their roots (Epstein, 1994; Ma & Yamaji, 2006) and deposit silica (SiO2) in plant tissues as phytoliths (Hodson et al., 2005; Schaller et al., 2013). Approximately 0.1%-6.0% organic C in plants is reported to be incorporated in phytoliths during their formation, and is referred to as phytolith-occluded carbon (PhyOC) (Bartoli & Wilding, 1989; Jones & Milne, 1965; Zuo & Lü, 2011). Phytoliths exist in most plants, and their content in plant tissues varies among plant species, ranging from less than 0.5% in most dicotyledons, 1%-3% in dryland grasses, and up to 10%-15% in the Cyperaceae and wetland Poaceae species (Epstein, 1994). Phytoliths are mainly deposited in the cell wall, cell lumen and intercellular spaces or the extracellular layer in plant issues (Epstein, 2009; Hodson et al., 2005; Ma & Yamaji, 2006; Schaller et al., 2013); after plants’ death, the phytoliths can be incorporated into soil or sediments after the decomposition of plant litter (Blecker et al., 2006). Since phytoliths are resistant to decomposition, PhytOC can be preserved in soil or sediments with the Si-coat protection for several hundred or thousand years, and may account for 82% of the total organic carbon in some old soils (Parr & Sullivan, 2005; Song et al., 2017). The formation and stability of PhytOC in the coupled biogeochemical cycling of C and Si has been increasingly recognized as a promising mechanism of terrestrial ecosystems to sequester atmospheric CO2, which has motivated many researchers to quantify the PhytOC sequestration potential of various systems (Parr et al., 2010; Parr & Sullivan, 2011; Song et al., 2012a).
Grasslands are an important terrestrial ecosystem covering more than one-fifth of the world’s land surface (Scurlock & Hall, 2010). The large distribution area and the high PhytOC concentration in grassland plants, especially Poaceae and Cyperaceae species (Epstein, 1994; Clarkson & Hanson, 1980), make grassland a particularly important long-term C sequestration process (Blecker et al., 2006; Song et al., 2017). Several studies assessed the PhytOC sequestration potential of grasslands (Pan et al., 2017; Qi et al., 2016; Song et al., 2012a). Song et al. (2012a) reported that phytolith and PhytOC production rates in aboveground biomass of grassland were significantly influenced by their aboveground net primary productivity (ANPP); Qi et al. (2016) suggested that the belowground productivity of plants could play a dominant role in PhytOC production in grassland ecosystems; and Ji et al. (2018) reported that Si distribution in the aboveground parts, thus the PhytOC sequestration potential, of grassland plants, varies markedly among plant species and across grassland types. However, the effects of grassland type and species composition on the production rate of phytoliths and PhytOC (including the above- and belowground parts) are still not fully understood. Precipitation is the predominant climatic factor that controls plant species composition and net primary productivity (NPP) of grassland ecosystems in semi-arid steppe region (Bai et al., 2004; Dai et al., 2012; Hou et al., 2014), thus affecting the PhytOC sequestration potential. As such, it is necessary to investigate the PhytOC production rate in different grassland types along climatic gradient, for accurate estimation of PhytOC sequestration potential and prediction of their response to climate changes. In present study, we selected three representative vegetation types in climatically different regions alone a precipitation gradient in Inner Mongolia, that is, the desert steppe, the dry typical steppe, and the wet typical steppe regions, to study their PhytOC production. Specifically, we aimed to investigate the species composition, and measure the phytolith and PhytOC contents of the major plant species, of these grassland types, so as to increase the accuracy in estimating phytoliths and PhytOC production and storage in plant communities, and analyze their relations with climatic factors. We hypothesized that along the gradient of climate aridity increase (that is, precipitation decrease) from the wet dry steppe site, via the dry typical steppe site to the desert steppe site, the grassland NPP would decrease, while the contents of phytoliths and PhytOC in the steppe plants would increase due to the intense plant transpiration but low biomass production.