Strong Environmental Filtering Based on Hydraulic Traits Occurring in the Lower Water Availability of Temperate Forest Communities

Abstract

The trait-based approaches have made progress in understanding the community assembly process. Here, we explore the key traits that may shape community assembly patterns of the same community type but within different water availabilities. Natural Quercus wutaishanica forests were chosen as a suitable study system to test the difference between economic and hydraulic traits across water availability on the Loess Plateau (LP, drought region) and Qinling Mountains (QL, humid region) of China. A total of 75 plots were established separately in two sites, and 12 functional traits (seven hydraulic traits and five economic traits) of 167 species were studied. Community-weighted mean trait values and functional diversity indices were compared between the two sites. Canonical component analysis was performed to infer whether the changes of community traits and their relationships are driven by intraspecific variation or species turnover. Evidence for likely community assembly processes was tested using the null model to determine whether functional structure among seven hydraulic traits and five economic traits was dominated by different ecological processes between two sites. We found that forests in the Loess Plateau and Qinling Mountains showed different hydraulic and economic traits. Hydraulic and economic traits coupled at the community level were driven by species turnover. Hydraulic traits showed more significant convergent patterns on LP than that in QL. Our results suggest a strong environmental filtering process occurred in hydraulic-based community assembly in the temperate forest with low water availability. Reveal the relationship of hydraulic and economic traits at the community level. Emphasize the critical role of multi-dimensional traits selecting like hydraulic traits in community ecology.

Keywords: Loess Plateau; community assembly; environmental filtering; hydraulic traits; null model.

FIGURE 1

Box-plot comparing community-weighted mean value of (A) maximum vessel diameter, (B) Huber value, (C) stomatal length, (D) stomatal density, (E) the leaf turgor loss point, (F) branch wood density, (G) vein density between LP and QL. Different letters indicate significant differences (P < 0.05).

FIGURE 2

Box-plot comparing community-weighted mean value of (A) leaf dry matter content, (B) leaf dry mass per area, (C) leaf thickness, (D) leaf tissue density, (E) maximum net photosynthetic rate between LP and QL. Different letters indicate significant differences (P < 0.05).

FIGURE 3

Community-weighted trait bi-plot for canonical correlation analysis, showing the effects of functional traits on the community species composition and plot distribution between the Loess Plateau and Qinling Mountains. (A) The effect of functional traits on the plot distribution. (B) The effect of functional traits on the community species composition. The ordination is based on community-weighted mean traits, with traits weighted for each species by its relative abundance in the community. Trait abbreviations are given in Table 1.

FIGURE 4

Difference of functional diversity components in relation to hydraulic traits and economic traits between LP and QL. (A) Functional richness (FRic); (B) functional evenness (FEve); (C) functional divergence (FDiv); (D) Rao’s quadratic entropy (RaoQ); (E) one-dimensional functional richness (FDis); different letters above bars indicate statistically significant differences at P < 0.05.

FIGURE 5

The difference of community functional structures based on single traits between the Loess Plateau (LP) and the Qinling Mountain (QL). The functional structures were measured as standardized effect sizes (SES). (A) Test on single hydraulic traits in tree stratum. (B) Test on single economic traits in tree stratum. (C) Test on single hydraulic traits in the whole site. (D) Test on single economic traits in the entire site. Specifically, SES > 0 represents functional divergence, whereas SES < 0 indicates functional convergence. Each point shows the mean SES values among all plots within the LP and the QL. Error bars represent standard errors. Asterisks indicate significant differences between SES with 0. Different letters above bars (ab) indicate statistically significant differences at P ≤ 0.05 between the LP and the QL in the same trait.

FIGURE 6

The difference of community functional structure [standardized effect size (SES)] based on multiple traits between the Loess Plateau (LP) and the Qinling Mountain (QL). (A) Test on multiple hydraulic and economic traits in tree stratum. (B) Test on multiple hydraulic and economic traits in the whole site. Standardized effect size > 0 represents functional divergence, whereas SES < 0 indicates functional convergence. Red point and blue point show the mean SES values within the LP and the QL. Error bars represent standard errors. Asterisks indicate significant differences between SES with 0. Different letters above bars (ab) indicate statistically significant differences at P < 0.05 between the LP and the QL in the same trait.