Uncollected polythene fragments size-dependently alter soil multifunctionality via mediating bacterial drivers in dryland

Le résumé fourni par la source

Plastic pollution has emerged as a crucial driver of environmental change [1]. Despite human efforts towards recycling and clean-up, current practice makes it difficult to collect fragments that are <100 cm² in size and this poses a significant challenge for environmental sustainability [2]. Uncollectible fragments might modify soil ecosystem functions via mediating soil microbial communities, particularly soil multifunctionality (SMF) [3–5]. Currently, plastic fragments steadily accumulate in the environment and are believed to have negative impacts on soil function [2,6,7]. However, it remains unclear how the uncollected plastic pollution will affect SMF with increasing accumulation (over the next 30 years) and varying sizes. To address this widespread concern, a continuous 3-year field investigation was conducted in irrigated dryland. The results showed that various soil functions exhibited differentiate responses to plastic fragments (Fig. 1a–d). Overall, relative to the control group without plastic fragment addition (CK), the nitrogen storage function of the soil was significantly reduced in the area treated with fragment addition. Under low concentrations, nitrogen stocks initially decreased and then increased with increasing fragment sizes (P < 0.05). At high concentrations, however, there was no significant trend (Fig. 1a and Fig. S2). On the other hand, the enzymatic activity and DNA concentration functions were markedly enhanced, yet this was highly dependent on the fragment sizes and concentrations (Fig. 1c and e, and Figs S3 and S4). Moreover, medium-sized fragments significantly reduced crop production function across the concentrations. In contrast, under high concentrations, the addition of small fragments (Small-H) evidently promoted crop production function whereas that of large fragments had no pronounced effect (Fig. 1d and Fig. S5). The interactive effects of fragment sizes and concentrations were generally evident on nitrogen stocks and crop production. Finally, the overall carbon storage function was basically unaffected by plastic fragment addition (Fig. 1b and Fig. S6). Polythene fragment pollution conditionally alters soil multifunctionality via the mediation of bacterial drivers. (a–f) Effects of plastic fragment pollution on sole soil functions (Z-score) and multifunctionality. (g) Piecewise structural equation model (PiecewiseSEM) accounting for the hypothesized direct and indirect relationships between soil functional groups, bacterial community characteristics (including co-occurrence patterns) and soil multifunctionality. R2 denotes the proportion of variance explained. (h) Co-occurrence patterns in the soil bacterial network as affected by plastic fragments pollution. The sizes of the nodes (ASVs) are proportional to the number of connections. Only nodes that were significantly correlated with each other (Spearman's > 0.7, P < 0.05) were connected (edges). (i) Correlation between dissimilarity in bacterial community characteristics and dissimilarity in soil multifunctionality. (j) The standardized effects (λ) of the components in the model on the SMF derived from PiecewiseSEM. a + b + c + d, total effect. b + c, mediating effects of bacterial community characteristics. c + d, mediating effects of bacterial co-occurrence patterns. For each function, data followed by different lower-case letters indicate the significant differences at P < 0.05. ***P < 0.001; **P < 0.01; *P < 0.05. In total, the SMF index was significantly lowered by 33.2% and 35.3% in response to small and medium-sized fragments, respectively, relative to CK (P < 0.05). However, there were no significant changes in SMF for large fragments (Fig. 1f). Specifically, compared with large fragments, small and medium-sized fragments exhibited disruptive effects on soil aggregate structure and hydrothermal status (Fig. S7). In particular, medium-sized fragments displayed an absolutely negative impact on plant growth (Fig. S5). The above phenomenon generally aggravated the micro-environmental instability of the soil [8]. Interestingly, a high concentration of fragments generally resulted in higher SMF regardless of size. However, the interactions between sizes and concentrations were not significant (Fig. 1f). A general trend was that the contrasting sizes and concentrations of plastic fragments significantly altered multiple soil ecosystem functions and bacterial community characteristics (Figs S2–S6 and S8–S11). Microbial diversity and soil functionality were positively correlated in both natural and agricultural ecosystems [9,10]. In the plastic-contaminated agricultural soils, the linear regression model showed significant positive correlations between SMF dissimilarity and bacterial community characteristic dissimilarity (P < 0.001) (Fig. 1i and Figs. S11 and S12). On the other hand, increasing plastic residues significantly altered soil bacterial communities, as evidenced by notable changes in co-occurrence patterns (Fig. 1h, Fig. S13 and Table S1). Relative to CK, the average clustering coefficient tended to increase in the groups with plastic fragment addition. However, the number of nodes showed a decreasing trend (Fig. S13). In addition, there was an increasing tendency in either the positive edges or the average degree in the small-fragment and medium-fragment groups, while no significant differences were observed in the large-fragment group relative to CK (Fig. S13). Notably, the small and medium-sized fragments altered the node connectivity by reducing the number of nodes with increased positive edges, resulting in a denser network. This densification may have simplified the co-occurrence patterns, potentially reducing the network complexity and stability [11]. Certainly, the competition among microorganisms enhanced network robustness [12]. To a large extent, the alterations in the network structure represent the ecological adaptive responses of soil microorganisms to plastic fragments [13]. In the present study, the impacts of fragment concentrations on network structures were statistically insignificant but became significant when interacting with different sizes (Table S1). Surprisingly, there was only a slight change in the stability of the network in the large-fragment groups (Fig. S13). Yet, small and medium-sized fragments brought about more pronounced effects on bacterial co-occurrence relationships than did large ones, due to their larger contact area with the soil matrix. Given that SMF could be affected by soil bacterial community characteristics, the associations of SMF with bacterial co-occurrence patterns were also analysed (Figs. S13–S15). There were strong correlations between the topological properties of the network and SMF (Fig. S15). Here, the network centrality node, network connector, positive edge, node number and average degree were identified as the major factors affecting SMF (Fig. S15). It can be argued that the microbial network structure played a crucial role in shaping SMF (Fig. 1j and Fig. S16). On the other hand, random forest modeling demonstrated that bacterial community characteristics (including co-occurrence patterns) exhibited a greater contribution to SMF relative to nitrogen or carbon storage when considering multiple soil drivers (Fig. S16). Use of the Piecewise structural equation model explained 73% of the variance in SMF. Specifically, the majority of key functional groups displayed positive effects on SMF, including microbial activity (λ = 0.244), soil properties (λ = 0.278), bacterial community characteristics (λ=0.286), bacterial co-occurrence patterns (λ = 0.167) and crop production (λ = 0.125), respectively (Fig. 1g). In total, the characteristics of the bacterial community were as influential as or more influential than the other functional groups, which was consistent with the outcome from the random forest analyses (Fig. 1g and j, and Fig. S16). Mechanistically, the impacts of uncollected plasti

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.