1. INTRODUCTION

Desert ecosystems, the arid areas (approximately 45.6 million km2, covering at least 35% of Earth’s land surface) most environmental sensitive to changes in climate and human disturbances, are suffering from increasing land degradation and biological diversity loss (Bastin et al., 2017; Schimel, 2010; Wang et al., 2019a). In arid and semiarid regions, fragile steppe grassland ecosystems are especially prone to desertification, which results in the destruction of plant communities, soil physical texture, and nutrient losses. Restoring and rehabilitating desertification grassland has been a serious issues in ecology (D’Odorico et al., 2019). Soil microorganisms have a predominant potential to reveal and regulate the changes in soil ecosystem functions and services, so characterizing the microbial successional patterns and their interactions with plants and soils is essential for increasing our understanding of the mechanisms of ecological restoring and rehabilitating, improving our capacity to predict the responses of ecosystems to human disturbance, and optimizing the design of large-scale restoration projects.
Soil multifunctionality has been widely used to quantify and compare the ability of various ecosystems to support overall functionality (Manning et al., 2018). Numerous studies have widely emphasized the significance of multifunctionality in superficial soils (top ~20 cm) due to their predominant ecological importance (Delgado-Baquerizo et al., 2017; Delgado-Baquerizo et al., 2016; Wagg et al., 2014; Wagg et al., 2019; Zheng et al., 2019), where the density, activity, and diversity of microorganisms are higher than those in deep soils (i.e., below 20 cm) (Fierer et al., 2003). However, due to the large bulk of soil profiles, deep soils may accumulate abundant soil nutrients, leading to an enhanced ability to provide multifunctionality during ecosystem development. For example, compared to superficial soils, deep soils in typical forest and grassland ecosystems can store over two-thirds the organic carbon and nearly equal amounts of phosphorus due to the long-term input from plant roots and soil leaching (Balesdent et al., 2018; Upton et al., 2020; Zhou et al., 2018). Indeed, increasing evidence has confirmed that a large percentage of soil multifunctionality is hidden in deep soils (Jiao et al., 2018; Upton et al., 2020). Intriguingly, in desert ecosystems, the desertification process can result in a decrease in the water-holding capacity of superficial soils and a steady increase in the soil leaching capacity (D’Odorico et al., 2013; D’Odorico et al., 2019; Neilson et al., 2017), leading to the enhanced downward elemental (i.e., carbon (C), nitrogen (N), and phosphorus (P)) transfer responsible for the persistent accumulation of nutrients in deep soils. Thus, it is expected that deep soil multifunctionality in desert ecosystems plays increasingly important roles in regulating and buffering overall ecological service functions.
Soil microbial diversity and community composition drive and determine soil multifunctionality in various ecosystems (Delgado-Baquerizo et al., 2016; Wagg et al., 2014; Zheng et al., 2019). Different influences of biodiversity and community composition on soil multifunctionality have been investigated (Delgado-Baquerizo et al., 2017; Wagg et al., 2014). For example, the bacterial and fungal diversity in superficial soils (top ~20 cm) positively influenced multifunctionality (Wagg et al., 2019), especially in arid ecosystems (Delgado-Baquerizo et al., 2016), while particular taxa of bacteria and fungi but not their total richness and abundance determined the resistance of superficial soil multifunctionality in arid ecosystems (Delgado-Baquerizo et al., 2017). Furthermore, although such vast soil microbes in superficial soils may play a central role in driving nutrient turnover with supportive effects on soil multifunctionality, microbial diversity and particular microbial attributes in deep soils share indispensable ecological drivers (Delgado-Baquerizo et al., 2019; Jiao et al., 2018). Indeed, deep soil contains a significant portion (35% ~ 50%) of total microbial diversity and biomass, suggesting its non-negligible ecological functions (Eilers et al., 2012). Increasing evidence, based mostly on results from ecosystems with typical development (reforested ecosystems) and grassland ecosystems with high rainfall, suggests that particular microbial attributes (i.e., individual species, particular communities and their cooperation) rather than alpha-diversity levels (i.e., total abundance, species richness, and the Shannon index) may play more important roles in driving deep soil nutrient (i.e., C, N, and P) cycling with direct feedback effects on soil multifunctionality (Jiao et al., 2018; Upton et al., 2020), as particular microbial attributes compose significant regulators of microbial growth and interactions and predominant functional attributes (e.g., soil C cycling or N mineralization) (Delgado-Baquerizo et al., 2017). In contrast to the above ecosystems with typical development and less drought stress, ecosystems undergoing desertification display enhanced environmental stress gradients (i.e., decreasing plant productivity and soil nutrients and increasing soil leaching), while the dynamics of soil microbiomes and their contributions to regulating deep soil multifunctionality remain largely unexplored, particularly the contributions of particular microbial attributes.
Herein, we assess the importance of vertical soil microbiomes and their contributions to multifunctionality during the desertification process in desert ecosystems. We used three sites in different typical desertification stages (potential, moderate, and severe desertification) that represent the process of desertification. The multifunctionality and diversity of soil bacteria, fungi, and archaea were investigated at soil depths of 0-100 cm. We tested the hypotheses that (1) multifunctionality in deep soils (20-100 cm) would, at least partially, be equally as important as that of superficial soils (0-20 cm) in the desert ecosystem and (2) the combined effects of the particular species richness of soil microbiomes would better predict multifunctionality in desert ecosystems, especially in deeper soil profiles.