INTRODUCTION
Direct adverse impacts of global warming on tropical forests, including decelerated plant growth due to high respiration costs (Feeley et al. 2007; Vlam et al. 2014) and increased tree mortality (Aleixoet al. 2019) attributed to high risk of hydraulic failure (Adamset al. 2017), have well been documented. However, an increasing number of studies have recently shown that global warming and its concomitant drought can also alter the interaction between pathogens and their host trees, for example, by increasing the load of infectious diseases (Anderson et al. 2004) or changing the severity of plant diseases (Swinfield et al. 2012; García-Guzmán et al.2016; Bachelot et al. 2020; Milici et al. 2020). Any change in environmental conditions due to climate change can potentially affect disease severity as predicted by the framework of the classic plant disease triangle (Chakraborty 2005; Liu & He 2019; Miliciet al. 2020; Romero et al. 2021). Therefore, predicting the net response of forest tree species to global warming requires quantification of not only the direct responses of tree species (e.g., ecophysiological stress) (Adams et al. 2017), but also indirect responses, such as those associated with changing natural enemies (e.g., soil-borne pathogens) (Thompson et al. 2010; van der Puttenet al. 2016; Bachelot et al. 2020). This is particularly true for soil-borne pathogenic microbes, which form the most important biotic interactions with plant roots (Wardle et al. 2004) and play a major role in maintaining plant diversity in tropical forests via a negative plant–soil feedback (PSF) (Bell et al. 2006; Manganet al. 2010; Bagchi et al. 2014; Eck et al. 2019).
Primarily driven by soil-borne microbes (Mariotte et al. 2018), PSF occurs due to changes in soil microbes induced by the presence of plants, which in turn influences plant performance (Bever et al.1997). PSF can be either positive or negative (Bever 2003; Kandlikaret al. 2019), with the direction and magnitude depending on the relative influence of beneficial (e.g., ectomycorrhizal fungi, EcM fungi) (Corrales et al. 2018) and antagonistic microbes (e.g., plant-pathogenic fungi) on plant species as well as the effect of climate (Liu & He 2019; Pugnaire et al. 2019). The pathogen-induced Janzen–Connell (JC) effect (Janzen 1970; Connell 1971) is a type of net negative PSF, important to maintaining biodiversity in the tropics (Bellet al. 2006; Mangan et al. 2010; Bagchi et al.2014; Eck et al. 2019; Schroeder et al. 2020), and its strength is typically estimated based on the mortality of seedlings (or relative growth rate of seedlings) that are infested by host-specific soil-borne fungal pathogens. Increased temperatures may not be optimal for the reproduction and growth of pathogens in the tropics because tropical temperatures are already on the upper limit for many microorganisms (Thompson et al. 2010; van der Putten et al. 2016; Liu & He 2019; Bachelot et al. 2020), and thus the role of JC effect in maintaining biodiversity can potentially be weakened by the further increase of temperatures. However, the response of PSF to global warming is more than the response of the pathogen-based JC effect as differential responses of antagonistic versusbeneficial microbes to warming could change the net PSF (Liu & He 2019).
In this study, we conducted a three-year warming experiment using open-top chambers (OTCs) to test how seedlings of two tree species,Ormosia semicastrata (Fabaceae) and Cyclobalanopsis patelliormis (Fagaceae), with contrasting nutrient-acquisition strategies via soil-borne microbes, responded to elevated temperature in a tropical montane rain forest in Hainan Island, China. Our aim was to examine warming-induced changes in the relative abundances of plant-pathogenic fungi and EcM fungi in bulk soils and quantified effects of such changes on seedling survival of the two tree species. We expected plant disease severity for O. semicastrata to vary if warming alters soil pathogenic fungi. The severity could be heightened or weakened, depending on the response of plant-pathogenic and/or EcM fungi to warming. In contrast, we expected that warming would not affect pathogen-induced seedling mortality for C. patelliormis due to Hartig net of EcM fungi that protects the host roots from pathogenic attack (Marx 1972; Albornoz et al. 2017; Tedersoo et al.2020).