1. Invasive species pose significant challenges to successful restoration efforts worldwide. A strategy to reduce invasions is to establish communities consisting of species with varied ecological strategies. These strategies typically align along the conservative and plant size axes, and more recently, along a belowground collaboration axis. However, we lack understanding of how the diverse ecological strategies of Cerrado grass species, their combinations, and their interactions with soil conditions can mitigate invasions.
2. Here, we investigated how native grass communities composed of species with different ecological strategies affect the invasion success in two soil types of abandoned pastures in the Cerrado. Specifically, we tested the hypothesis that greater above- and belowground functional diversity reduces exotic species invasion. We also evaluated whether the isolated effects of native species on invasion were positive or negative.
3. We installed an experiment with species richness ranging from zero to eight native grass species. In November 2019, we sowed species combinations to create communities composed by species with different ecological strategies. We quantified the aboveground biomass of exotic species as a measure of invasion. To characterize the species' ecological strategies, we measured five functional traits.
4. Functional diversity of maximum height and specific root length (SRL) had the highest predictive power, however, the most parsimonious model included only SRL diversity, which represents the collaboration axis. Native aboveground biomass was also negatively related to exotic species biomass. Furthermore, invasion was greater in less stressful soil conditions but did not interact with diversity. The effect of native species varied from facilitation to competition, with the annual fast-growing native species favouring invasion.
5. Synthesis and applications. Our results show that greater functional diversity of combined above- and belowground traits reduces invasion success, shedding light on an underexplored role of SRL diversity. The competitive and facilitative effects of different native species highlight the need for careful selection of the species to be used in restoration programs. Furthermore, the absence of interaction between diversity and soil types highlights the need for an integrated management of the functional composition and edaphic factors to increase resistance to invasion in these Neotropical grass communities.

In November 2019, a biodiversity experiment was established, sowing 302 plots with different species richness levels and dominance, utilizing eight native Cerrado grass species. Following two growth seasons, in October 2021, aboveground biomass of exotic grasses was harvested from 4m² plots, while that of native species was collected from two 12.5 x 25 cm subplots within these larger plots. The biomass samples were then oven-dried at approximately 60°C to a consistent weight and subsequently weighed. Soil types for each plot were determined by visual assessment, categorized as rocky or clayey based on the presence of laterites. In February 2022, five functional traits of the native species were measured: leaf mass per area (LMA), maximum vegetative height (H), root tissue density (RTD), specific root length (SRL), and total root length (TRL). These measurements were taken from six isolated individuals per species near the experiment. Leaf samples for LMA were hydrated for 24 hours, scanned for area estimation using ImageJ software, and then dried and weighed. Maximum vegetative height was measured from five adult individuals in native areas. The root system was assessed by washing soil monoliths and analyzing root segments stained with methylene blue. Root length and volume were quantified using RhizoVision Explorer, and total root length was derived by multiplying specific root length by total root mass. Rao’s quadratic entropy (FDQ) was calculated for each plot as an index of functional diversity, producing 31 distinct FDQ values representing various trait combinations.