Slope position- mediated soil environmental filtering drives plant community assembly processes in hilly shrublands of Guilin, China.

Background and aims: A major goal of community ecology focuses on trying to understand how environmental filter on plant functional traits drive plant community assembly. However, slopes positions- mediated soil environmental factors on community-weighted mean (CWM) plant traits in shrub community has not been extensively explored to analyze and distinguish assembly processes. Methods: Here, we surveyed woody shrub plant communities from three slope positions (foot, middle, and upper) in a low hilly area of Guilin, China to assess differences in functional trait CWMs and environmental factors across these positions. We also measured the CWMs of four plant functional traits including specific leaf area, leaf dry matter content, leaf chlorophyll content, and leaf thickness and nine abiotic environmental factors, including soil water content, soil organic content, soil pH, soil total nitrogen, soil total phosphorus, soil total potassium, soil available nitrogen, soil available phosphorus, and soil available potassium. We used ANOVA and Tukey HSD multiple comparisons to assess differences in functional trait CWMs and environmental factors across the three slope positions. We used redundancy analysis (RDA) to compare the relationships between CWMs trait and environmental factors along three slope positions, and also quantified slope position-mediated soil environmental filtering on these traits with a three-step trait-based null model approach. Results: The CWMs of three leaf functional traits and all soil environmental factors except soil pH showed significant differences across the three slope positions. Soil total nitrogen, available nitrogen, available potassium, and soil organic matter were positively correlated with the CWM specific leaf area and leaf chlorophyll content along the first RDA axis and soil total potassium, total phosphorous, and soil water content were positively correlated with the CWM leaf dry matter content along the second RDA axis. Environmental filtering was detected for the CWM specific leaf area, leaf dry matter content, and leaf chlorophyll content but not leaf thickness at all three slope positions. Conclusions: Ultimately, we found that soil environmental factors vary along slope positions and can cause variability in plant functional traits in shrub communities. Deciduous shrub species with high specific leaf area, low leaf dry matter content, and moderate leaf chlorophyll content dominated at the middle slope position, whereas evergreen species with low specific leaf area and high leaf dry matter content dominated in slope positions with infertile soils, steeper slopes, and more extreme soil water contents. Altogether, our null model approach allowed us to detect patterns of environmental filtering, which differed between traits and can be applied in the future to understand community assembly changes in Chinese hilly forest ecosystems.