Policelli et al_Invasive ectomycorrhizal fungi can disperse in the absence of their known vectors_Metadata
Description
MANUSCRIPT TITLE: Invasive ectomycorrhizal fungi can disperse in the absence of their known vectors AUTHORS: Policelli, N; Horton, TR; Kitzberger, T; Nuñez, MA. METADATA DESCRIPTION: Metadata includes raw data from the greenhouse experiment (biomass -above and belowground-, EMF colonization, germination, and survival of plants). It also includes field data: vegetation survey of the different sites together with uprooted surface and number of feces. For data analysis details please refer to the Methods section of the manuscript. ABSTRACT: Positive interactions between non-native species can accelerate their invasion rate and exacerbate their impacts. This has been shown for non-native mammals that disperse invasive ectomycorrhizal fungi (EMF), in turn facilitating the invasion of trees. While wind is the main short-distance dispersal mechanism for most EMF species, mammal-mediated dispersion is assumed to be the main mechanism of EMF long distance dispersal, being particularly critical for hypogeous truffle-like EMF species. We asked whether the absence of non-native mammals is an obstacle for pine invasion given the lack of invasive EMF being dispersed. We studied EMF species colonization and pine trees’ growth in soil from mainland sites where non-native mammals are highly abundant, and lake islets in which they have been historically absent. Contrary to what we expected, we found invasive EMF, including hypogeous species, in sites where invasive mammals have been historically absent. Pine trees grew equally well and had the same EMF colonization in soil from mainland and islets. Two non-native invasive EMF species were present in the islets, one of them hypogeous. The absence of invasive mammals does not hinder the invasion of EMF, which are critical for pine invasion. Alternative mechanisms of EMF dispersal, such as saltation, bird dispersal, or human dispersal, can be involved in their arrival to native stands. The presence of invasive EMF makes native sites vulnerable to pine invasion, even in the absence of mammalian dispersers.
Files
Steps to reproduce
To analyze the change in dry biomass of the different plant species according to the type of soil in the greenhouse experiment, we assumed a normal distribution of the response variable “biomass”. We used a linear mixed effects model adjusted for maximum likelihood. Each species was analyzed separately, as well as the above and the belowground tissue. The variable "soil type" was included as a fixed factor with three levels, "sterile control" as a fixed factor with two levels (yes, no) and "Block" as a random factor in the model. The model was validated using the Shapiro-Wilk test for normality, in turn testing for homoscedasticity. Given the lack of homoscedasticity, the structure of variances was modeled using the identity function (varIdent), estimating a different variance for each level of the variable "soil type". To analyze root colonization by EMF fungi in the case of P. menziesii in the different types of soil from the greenhouse experiment we assumed a binomial distribution of the response variable "root colonization”, calculated as the number of colonized root tips / total number of root tips. We used a generalized linear mixed model based on Laplace approximation and the logit link function. In all cases, Tukey tests were used to analyze the differences between treatments. All analyzes were performed with the statistical software R 3.4.1. To explore and visualize the similarities between the three types of sites in the percent cover of each plant species, we performed a Principal Coordinates Analysis (PCoA). In turn, we included the categorization of species into native and non-native. The proportion of endemic, native and non-native species was plotted separately for each type of site. To analyze the differences in the number of feces between the different sites, we assumed a Poisson distribution of the response variable "number of feces in 100 m2". We used a Hurdle Poisson model that considers zero values as a separate category from the rest of the values, for zero-inflated data. We included soil type as a fixed factor and "block" as a random factor in the model. To analyze the differences in the soil physicochemical variables, we carried out an analysis of the variance for each variable separately, incorporating as a fixed factor the type of soil with three levels: mainland, islet, and mainland near plantation.