Skip to main content
Dryad

Data for: Invasion by an exotic grass species homogenises native freshwater plant communities

Cite this dataset

Figueiredo, Bruno (2022). Data for: Invasion by an exotic grass species homogenises native freshwater plant communities [Dataset]. Dryad. https://doi.org/10.5061/dryad.hqbzkh1kp

Abstract

A growing body of evidence has shown that biological invasions cause shifts in species composition of communities in space and time. Although biological invasions are considered a major driver of biotic homogenisation worldwide, most previous studies are conducted at small spatial scales and over short time periods, which may have underestimated the impacts of exotic species on native communities.

Using a unique dataset of aquatic plants sampled in 235 sites over 12 years (2007–2010 and 2015–2019) in a large reservoir (Itaipu Reservoir; 1,350 km²), we analyzed how the invasion of a non-native grass (Urochloa arrecta) affects the species richness, ecological uniqueness (i.e., local contribution to beta diversity – LCBD) and temporal β–diversity of native plant communities.

From 3,934 surveyed plant communities, U. arrecta was recorded in 2,888 samples and it was absent from 1,046 samples. Overall, species richness and ecological uniqueness of native plant communities were markedly lower in sites invaded than non-invaded by U. arrecta. From 2007 to 2019, the ecological uniqueness of native plants was 60% lower in the invaded than non-invaded sites.      Whereas in invaded sites the species loss was the dominant mechanism driving native communities over time, in non–invaded sites the gain of new native species was the primary mechanism underlying community trajectories. Moreover, comparing native plant communities before and after the invasion of U. arrecta, species richness, ecological uniqueness and species gains of native plant communities decreased, whereas species losses increased after the invasion of U. arrecta. Finally, the positive relationship between native biodiversity and precipitation was stronger in sites non-invaded than invaded by U. arrecta.

Synthesis: Our findings provide comprehensive evidence that an invasive plant is decreasing the spatial and temporal β–diversity of native plant communities through declining species richness, rather than simply correlating with them. This suggests that U. arrecta is driving native plants to become less diverse and homogeneous after the invasion, both spatially and temporally. Our findings illustrate that at broad scales, aquatic plant communities may become increasingly homogeneous with the increasing number of biological invasion events taking place worldwide. 

Methods

We analysed a 12-years data set (2007–2010 and 2015–2019) with presence-absence data for aquatic plants recorded in the Brazilian shoreline of the Itaipu Reservoir, a large water body filled in 1982 and located in the frontier between Brazil and Paraguay (24º15’–25º33’ S and 54º00’–54º37’ W; Figure 1). This reservoir covers an area of 1,350 km², has a mean depth of 22 m and its arms (the sampled sites) vary from oligotrophic to eutrophic (Thomaz et al., 2009). The local mean annual temperature is 22 ºC and the mean annual rainfall is 1,400 mm. More detailed information on environmental data can be found in Thomaz et al. (2003) and Pulzatto et al. (2019). It is not certain when U. arrecta first colonized this reservoir, but this species has been recorded since the first sampling carried out in Itaipu in 1996 (S. M. Thomaz, unpublished).

            We used aquatic plant data collected between 2007 and 2010 as well as between 2015 and 2019 in 235 sites located in the Reservoir arms, totalling 12 years of sampling (2007–2019). Reservoir sites were surveyed during the Aquatic Macrophytes Monitoring Program of the Itaipu Reservoir, which was designed to include sites in a wide range of environmental conditions. To provide a comprehensive assessment of aquatic plants, samplings were performed during the dry and wet seasons in each of the 235 geo-referenced sites (except in 2008, when only one sampling was conducted). In particular, the 235 sites were distributed along eight arms (flooded tributaries) of the Itaipu Reservoir, with the number of sampling sites within each arm (26–30) being proportional to its total area (Fig. 1). In total (between 2007 and 2019), there were 3,934 samples. The survey followed a standardized sampling protocol with the same sampling effort across all sites. Specifically, plant species occurrence in each site was assessed by three observers on a boat moving at a slow and constant velocity (Thomaz et al., 2009). These observers recorded plant species present along a ca. 100–m stretch from the shoreline of each site. To do so, the emergent vegetation was inspected visually and used a rake with a 4–m pole to collect submerged plants. Species not immediately recognized in the field were collected for identification in the laboratory (see Thomaz et al. 2009 for further details).

Invaded and non-invaded sites

Across the 235 sites, we considered all sites where U. arrecta was present as invaded sites, while other sites where U. arrecta was not present were considered non-invaded. Considering that each of the 235 sites was monitored over 12 years, this allowed us to analyse how the invasion of U. arrecta changed at these sites over time. For instance, whether a given site was not invaded by U. arrecta at a time (t1), this site was then considered not invaded at this time. However, if the same site was invaded by U. arrecta at another time (t2), then this site was considered invaded at this time point. Therefore, the same site may have become invaded or not invaded by U. arrecta over time. The opposite was also true; that is, if at a location at a time (t1) U. arrecta was present, then this site was considered invaded, but if at this same location at time (t2) U. arrecta was absent, this site was considered non-invaded.

Climate predictors

            Daily total precipitation (i.e., the sum of rainfall volume (Kg/m2/s)) and daily mean temperature (ºC) values were obtained for each sampling site and extracted from the RNCEP package in R (Kemp et al., 2012). The RNCEP database has a spatial resolution of 2.5 x 2.5o and a temporal resolution of 6 h (Kemp et al. 2012).

Funding

Itaipu Binacional