Response of grassland ecosystem function to plant functional traits under different vegetation restoration models in areas of karst desertification

Abstract

Plant functional traits serve as a bridge between plants, the environment, and ecosystem function, playing an important role in predicting the changes in ecosystem function that occur during ecological restoration. However, the response of grassland ecosystem function to plant functional traits in the context of ecological restoration in areas of karst desertification remains unclear. Therefore, in this study, we selected five plant functional traits [namely, plant height (H), specific leaf area (SLA), leaf dry matter content (LDMC), root length (RL), and root dry matter content (RDMC)], measured these along with community-weighted mean (CWM) and functional trait diversity, and combined these measures with 10 indexes related to ecosystem function in order to investigate the differences in plant functional traits and ecosystem function, as well as the relationship between plant functional traits and ecosystem functions, under four ecological restoration models [Dactylis glomerata (DG), Lolium perenne (LP), Lolium perenne + Trifolium repens (LT), and natural grassland (NG)]. We found that: 1) the Margalef index and Shannon-Wiener index were significantly lower for plant species in DG and LP than for those in NG (P<0.05), while the Simpson index was significantly higher in the former than in NG (P<0.05); 2) CWM_H, CWM_LDMC, and CWM_RDMC were significantly higher in DG, LP, and LT than in NG, while CWM_SLA was significantly lower in the former than in NG (P<0.05). The functional richness index (FRic) was significantly higher in DG and LP than in NG and LT, but the functional dispersion index (FDis) and Rao's quadratic entropy index (RaoQ) were significantly lower in DG and LP than in NG and LT (P<0.05), and there was no significant difference between DG and LP, or between NG and LT (P>0.05); 3) ecosystem function, including ecosystem productivity, carbon storage, water conservation and soil conservation, was highest in LT and lowest in NG; and 4) CWM_LDMC (F=56.7, P=0.024), CWM_RL (F=28.7, P=0.024), and CWM_H (F=4.5, P=0.048) were the main factors affecting ecosystem function. The results showed that the mixed pasture of perennial ryegrass and white clover was most conductive to restoration of ecosystem function. This discovery has important implications for the establishment of vegetation, optimal utilization of resources, and the sustainable development of degraded karst ecosystems.

Keywords: ecological restoration; ecosystem function; functional trait; karst desertification; plant diversity.

Figure 1

CWM of grassland ecosystems under different restoration models. Different letters in the same panel indicate significant differences between treatments based on Tukey pairwise comparisons (p<0.05). CWM_H, CWM_LDMC, CWM_SLA, CWM_RL, and CWM_RDMC represent the CWM for plant height, leaf dry matter content, specific leaf area, root length, and root dry matter content, respectively.DG, Dactylis glomerata grassland ecosystem; LP, Lolium perenne grassland ecosystem; LT, Lolium perenne + Trifolium repens grassland ecosystem; NG, natural grassland ecosystem.

Figure 2

Functional trait diversity of grassland ecosystems under different restoration models. Different letters in the same panel indicate significant differences between treatments based on Tukey pairwise comparisons (p<0.05). FRic, FDis, and RaoQ represent the functional richness index, the functional dispersion index, and Rao’s quadratic entropy index, respectively. DG, Dactylis glomerata grassland ecosystem; LP, Lolium perenne grassland ecosystem; LT, Lolium perenne + Trifolium repens grassland ecosystem; NG, natural grassland ecosystem.

Figure 3

Correlation analyses of species diversity, CWM, and functional trait diversity in grassland ecosystems under different restoration models. * indicates significance at the 0.05 probability level. The labels Margalef, Shannon, Pielou, and Simpson represent the Margalef index, the Shannon–Wiener index, the Pielou index, and the Simpson index, respectively. CWM.H, CWM.LDMC, CWM.SLA, CWM.RL, and CWM.RDMC represent the CWM of plant height, leaf dry matter content, specific leaf area, root length, and root dry matter content, respectively. FRic, FDis, and RaoQ represent the functional richness index, the functional dispersion index, and Rao’s quadratic entropy index, respectively.

Figure 4

Principal component analysis of each indicator for evaluation of grassland ecosystem function under different restoration models.

Figure 5

Redundancy analysis of plant functional traits and grassland ecosystem function. The labels Margalef, Shannon, Pielou, and Simpson represent the Margalef index, the Shannon–Wiener index, the Pielou index, and the Simpson index, respectively. CWM.H, CWM.LDMC, CWM.SLA, CWM.RL, and CWM.RDMC represent the CWM of plant height, leaf dry matter content, specific leaf area, root length, and root dry matter content, respectively. FRic, FDis, and RaoQ represent the functional richness index, the functional dispersion index, and Rao’s quadratic entropy index, respectively. EP, CS, WC, and SFC represent ecosystem productivity, carbon storage, water conservation, and soil fertility conservation, respectively.