The invasion of exotic plants in the river-lake ecotone has seriously affected the nutrient cycling processes in wetland soil. The South American species Alternanthera philoxeroides (Mart.) Griseb. is rapidly invading the river-lake ecotone in subtropical China, and has become the dominant species in the river-lake ecotone. However, there have been few studies on the effects of A. philoxeroides invasion on soil phosphorus (P) cycling and bioavailability in this ecotone. Herein, we measured the bioavailable P fractions, physicochemical properties and nutrient content in the surface soils of the native plant (Zizania latifolia (Griseb.) Turcz and Nelumbo nucifera Gaertn.) communities and the adjacent invasive A. philoxeroides communities in three river-lake ecotones with different nutrient substrates in the subtropical Dongting Lake basin over a three-year period to reveal the effects of A. philoxeroides invasion on the morphology and concentrations of soil bioavailable P. The principal coordinate analysis results showed that A. philoxeroides invasion significantly altered the bioavailable P concentrations in the soil of native plant communities in the different river-lake ecotones, and this effect was not disturbed by the heterogeneity of the soil matrix. However, the effects of invasion into different native plant communities on the fractions of soil bioavailable P were different. Compared with native Z. latifolia and N. nucifera communities, A. philoxeroides invasion increased the concentration of inorganic P by 39.5% and 3.7%, respectively, and the concentration of organic P decreased by 32.7% and 31.9%, respectively. Meanwhile, the invasion promoted P cycling and accumulation in the river-lake ecotone, which resulted in average decreases in the soil N:P and C:P ratios of 7.9% and 12.5%, respectively. These results highlight the impact of exotic plant invasions on nutrient cycling in wetland ecosystems in the river-lake ecotone, and this process may be detrimental to the late recovery of native plants.

Prior to data analysis, log10 and square root arcsine transformations were conducted to ensure normality and homogeneity of variance. One-way ANOVA was used to test the effect of site or plant type on the soil bioavailable P fraction, and Tukey’s test was used to compare the differences in bioavailable P contents among sites or plant communities. Two-way repeated-measures ANOVA was used to analyze the effects of the site and invasion time and their interaction on soil bioavailable P fractions. A linear mixed model (LMM) was used to analyze the effect of plant invasion on the native community, in which plant invasion was set as a fixed variable, sites and invasion time were set as random variables, and soil physicochemical properties, nutrient contents and stoichiometric ratios were set as continuous dependent variables. The analysis was performed using the "lme4" package in R, and the "glmm.hp" package was used to calculate R² values for fixed and random effects (Nakagawa & Schielzeth, 2013). The standardized major axis (SMA) was used to analyze the scaling relationships of the P content and N:P and C:P ratios in aboveground and belowground plant organs. The "SMATR" package in R was used for the scaling relationship analysis, and a likelihood ratio test was performed to compare differences in the allometric index between different native plants and their adjacent invasive plant species (Hu et al., 2021; Warton et al., 2012). Principal coordinate analysis (PCoA) was used to test the differences in soil bioavailable P fractions among different plant communities, and the method was multivariate analysis of variance (Adonis test) based on the Bray‒Curtis distance. Linear regression models were constructed to analyze the relationships between soil bioavailable P fractions and soil physicochemical properties in different plant communities. The differences between the slopes of linear regression models were compared using the "diffslope" function in R (Nekola & White, 1999). The "vegan" package was also used for the redundancy analysis of the effect of soil physicochemical properties on soil bioavailable P fractions, and "rdacca.hp" was used to perform a hierarchical analysis to determine the contribution of a single physicochemical factor to the bioavailable P fractions (Lai et al., 2022). Pearson’s correlation analysis was conducted to study the relationship between soil bioavailable P contents and soil physicochemical properties in plant communities. The analyses described above were performed using SPSS 23.0 (SPSS, Chicago, Illinois, USA) and R Version 4.1.1 (R Development Core Team, 2020).

Output units are described in the ReadMe_file.rtf file.