Identifying potential drivers of distribution patterns of invasive Corbicula fluminea relative to native freshwater mussels (Unionidae) across spatial scales

Abstract

This study aimed to identify the importance of ecological factors to distribution patterns of the invasive Clam (Corbicula fluminea) relative to native mussels (family: Unionidae) across seven rivers within the Mobile and Tennessee basins, Southeast United States. We quantitatively surveyed dense, diverse native mussel aggregations across 20 river reaches and estimated mussel density, biomass, and species richness along with density of invasive C. fluminea (hereafter Corbicula). We measured substrate particle size, velocity, and depth in quadrats where animals were collected. Additionally, we characterized reach scale environmental parameters including seston quantity and quality (% Carbon, % Nitrogen, % Phosphorous), water chemistry (ammonium [ ${\mathrm{NH}}_4^{+}$ ], soluble reactive phosphorous [SRP]), and watershed area and land cover. Using model selection, logistic regression, and multivariate analysis, we characterized habitat features and their association to invasive Corbicula within mussel beds. We found that Corbicula were more likely to occur and more abundant in quadrats with greater mussel biomass, larger substrate size, faster water velocity, and shallower water depth. At the reach scale, Corbicula densities increased where particle sizes were larger. Mussel richness, density, and biomass increased with watershed area. Water column ${\mathrm{NH}}_4^{+}$ increased at reaches with more urban land cover. No land cover variables influenced Corbicula populations or mussel communities. The strong overlapping distribution of Corbicula and mussels support the hypothesis that Corbicula are not necessarily limited by habitat factors and may be passengers of change in rivers where mussels have declined due to habitat degradation. Whether Corbicula is facilitated by mussels or negatively interacts with mussels in these systems remains to be seen. Focused experiments that manipulate patch scale variables would improve our understanding of the role of species interactions (e.g., competition, predation, facilitation) or physical habitat factors in influencing spatial overlap between Corbicula and native mussels.

Keywords: filter‐feeders; invasive species; niche overlap; niche partitioning; positive associations; rivers; species interactions.

FIGURE 1

Map of the study area with defined Level III ecoregions. Sample sites are the black points. Focal watersheds are highlighted within major drainage basins

FIGURE 2

Scatter plot of Corbicula density and mussel biomass with square‐root transformed axes (a). Box plots showing variation in substrate particle size (d), velocity (e), Depth (f), Corbicula density (b), and mussel biomass (c) measured in quadrats across seven rivers in the southeastern USA. Boxes cover the first through third quartile of the data; horizontal black line in each box is the median. One data point representing the greatest Corbicula density (2088 individuals m⁻²) is excluded from panels a and b

FIGURE 3

Corbicula probability of occurrence increases in quadrats with more mussel biomass (a), faster velocity (b), and larger particle sizes (c), but decreases in deeper quadrats. Points represent quadrats and are vertically jittered to avoid complete overlap. X axis in (a) is square‐root transformed

FIGURE 4

Ordination of principle components analysis of biotic and abiotic variables measured at the quadrat scale. River names are written in italics and correspond to 95% confidence ellipses for each river and vector arrows indicate latent biotic and abiotic gradients among rivers. “Corbicula” is Corbicula density (individuals m⁻²) and “Mussels” is mussel biomass (g m⁻²), richness, and density (individuals m⁻²). A single vector is shown for mussels because density and richness were strongly collinear (a). Plot of correlation matrix of variables included in quadrat scale generalized linear mixed models. Green indicates significant positive correlations, while pink indicated significant negative correlations. Boxes with an “X” are not statistically significant at p = .05 (b)

FIGURE 5

Correlation matrix of variables measured at mussel bed reaches. Green indicates significant positive correlations, while pink indicates significant negative correlations. Boxes with an “X” are not statistically significant at p = .05