Increasing inputs of invasive N‐fixing Acacia litter decrease litter decomposition and associated microbial activity in streams

Abstract

Nitrogen (N)‐fixing Acacia species are often aggressive invaders outside their native range. When invading native riparian temperate forests, they can decrease tree species diversity, alter the quality of litter inputs to streams and increase water N concentration. Although the effects of riparian tree species diversity and nutrient enrichment on litter decomposition and associated microbial decomposers have been widely studied, their individual and combined effects remain poorly understood, especially in streams flowing through forests invaded by Acacia species. Here, we assessed the effects of litter diversity (species evenness) and water N concentration on the decomposition of native and Acacia litter, and the activity and community structure of associated microbial decomposers. Litter of Castanea sativa (C) and Acacia melanoxylon (A) was enclosed in fine‐mesh bags in a total of five litter evenness treatments (100%C, 75%C + 25%A, 50%C + 50%A, 25%C + 75%A and 100%A), and immersed in a stream flowing through a native forest (native stream) and a stream flowing through a forest invaded by Acacia species (invaded stream). Litter decomposition rates and microbial decomposer activity differed among litter evenness treatments, generally decreasing as the proportion of A. melanoxylon increased. When considered individually, C. sativa litter decomposition and associated microbial activity did not differ among treatments. For A. melanoxylon, decomposition rates did not differ among treatments, whereas microbial activity was generally lower in treatments with higher or even proportions of C. sativa. Litter diversity had (small) antagonistic effects on litter decomposition in streams. However, litter treatments affected by diversity (species evenness) effects differed between streams, suggesting that effects can be modulated by water N concentration. Litter decomposition rates and microbial decomposer activity were higher in the invaded than in the native stream, probably as a consequence of the higher water N concentration in the former stream. However, the magnitude of the effects was small owing to the fact that water N concentration was still in the oligothrophic range in the invaded stream. Overall, our results suggest that the increasing proportion of N‐fixing Acacia species in invaded deciduous riparian forests will affect litter decomposition rates and microbial decomposer activity, and alter aquatic hyphomycete community structure, most probably as a result of decreases in the diversity and quality of litter inputs to streams, and increases in water N concentration. However, the magnitude of the effects resulting from decreases in litter input diversity and quality (due to increases in Acacia contribution) into invaded streams will probably be larger than those resulting from increases in water N concentration, thus overall litter decomposition will decrease. These impacts will possibly alter nutrient cycles in aquatic food webs that depend on riparian detritus, with implications for stream functionin...

Effects of Acacia invasion on water quality, litterfall, aquatic decomposers, and leaf litter decomposition in streams

“…Higher aquatic nutrient concentrations in invaded streams were expected since the same tendency was found in previous studies performed in the same streams (for few weeks in autumn/winter; Pereira & Ferreira, 2022), but also in streams flowing through areas invaded by other N-fixing species (Goldstein et al, 2009;Mineau et al, 2011;Stewart et al, 2019;Wiegner et al, 2013) or colonised by native N-fixing species (Compton et al, 2003;Shaftel et al, 2012). In invaded riparian areas, the continuous accumulation of high amounts of N-rich Acacia litter ensured the continuous release of N (Incerti et al, 2018;Marchante et al, 2019) and P compounds to the soil (Sitters et al, 2013;Souza-Alonso et al, 2015;Yelenik et al, 2004), due to the slow decomposition of phyllodes and other recalcitrant components of Acacia trees (e.g., pods; Madureira & Ferreira, 2022).…”

“…Increases in aquatic nutrient concentrations promoted by Acacia invasion have been shown before to stimulate aquatic hyphomycete activity in leaf litter Pereira & Ferreira, 2022).…”

“…Furthermore, contributions of reproductive components may be higher in Acacia stands than in deciduous forests, considering the massive production of reproductive structures by Acacia species (Milton, 1981) when compared to natives (Abelho & Graça, 1996; Molinero & Pozo, 2004; Pozo et al., 1997). The shedding of high amounts of N‐rich Acacia litter may also change the quality of litter inputs for aquatic decomposers (Madureira & Ferreira, 2022; Pereira et al., 2021; Pereira & Ferreira, 2021, 2022), and further contribute to increased aquatic nutrient concentrations.…”

“…Although the effects of Acacia invasion on terrestrial ecosystems are well known (reviewed by Le Maitre et al., 2011; Lorenzo & Rodríguez‐Echeverría, 2015; Souza‐Alonso et al., 2017), studies addressing their effects on stream communities and processes (e.g., leaf litter decomposition) are recent, few and restricted to the Fynbos biome in South Africa and deciduous forests in central Portugal (Lowe et al., 2008; Pereira et al., 2021; Pereira & Ferreira, 2021, 2022; Railoun, 2018; Railoun et al., 2021). The invasion of the Fynbos biome by Acacia mearnsii altered leaf litterfall dynamics and increased litterfall rates (Railoun et al., 2021), and changed macroinvertebrate communities through alterations in the abundance of functional feeding groups (Lowe et al., 2008), but did not affect macroinvertebrate density associated with decomposing leaf litter nor leaf litter decomposition rates in streams (Railoun, 2018).…”

“…variance[PERMANOVA];Anderson et al, 2008), since data failed to meet the assumptions of ANOVA after transformation. Even though PERMANOVA is mostly used in multivariate analysis, when run using the Euclidean distance, it holds the same F-statistic and equivalent pvalue, obtained by permutations, as univariate ANOVAs (e.g.,Caspi et al, 2019;Orozco-Aceves et al, 2015;Pereira & Ferreira, 2022).…”