Spatial variation in the direct and indirect effects of plant diversity on soil respiration in an arid region

Abstract

Soil respiration (Rs) is a key ecosystem function that controls the terrestrial energy balance and element cycling. Quantifying how plant diversity affects Rs is critical for predicting the impact of global plant diversity loss on ecosystem function. However, it is unclear how soil factors and plant diversity are spatially interrelated under natural environmental gradients, and how plant diversity affects the spatial variation of Rs on the basis of soil factors. Our objectives were to quantify the direction and magnitude of direct and indirect effects of plant diversity on Rs. We assessed spatial variability in the relationships between soil factors (soil chemistry and microclimate), multiple dimensions of plant diversity (taxonomic, functional, and phylogenetic diversity), and Rs using a new combination of geographically weighted regression and structural equation modeling based on survey data from three plots (river bank, transitional zone, and desert margin plots) along an environmental gradient in northwest China. Reduced plant diversity had a negative impact on Rs. However, when the effects of soil factors on Rs and spatial non-stationarity were considered simultaneously, the relationship between plant diversity and Rs changed in magnitude and direction. From the river bank to the desert margin, the positive effect of taxonomic and functional diversity on Rs gradually weakened, while the effect of phylogenetic diversity on Rs changed from a negative to a positive effect. Plant diversity and soil factors together explained 55%–75% of the spatial variation in Rs. Among them, the contribution of plant diversity to the spatial variation of Rs gradually decreased from the river bank to the desert margin, while the contribution of soil factors gradually increased. Functional diversity (17.39%–18.93%) and phylogenetic diversity (10.75%–19.53%) better explained the spatial variability of Rs compared with taxonomic diversity (2.67%–6.01%). The results provide strong evidence that plant diversity, especially functional and phylogenetic diversity, is a key driver of Rs, expanding our understanding of the relationship between plant diversity and Rs. These findings more accurately reveal the changing characteristics of carbon dynamics in desert ecosystems, and provide a methodological framework for studying the spatial variability of the relationship between plant diversity and ecosystem function.