Microhabitat partitioning between sympatric intertidal fish species highlights the importance of sediment composition in gravel beach conservation

Abstract

Gravel beaches in the Mediterranean ecoregion represent an economically important and unique habitat type. Yet, burgeoning tourism, intensive coastal development and artificial nourishment of beaches may jeopardize their ecological communities. To date, species that reside on gravel beaches and the consequences of beach alterations are poorly understood, which hampers the development of a sustainable coastal tourism industry along the region's shorelines. Using a simple collection method based on dredging buckets through the intertidal section of beaches, we quantified the microhabitat association of two sympatric clingfish species in the genus Gouania at seven natural and an artificial gravel beach based on sediment characteristics. We hypothesized that slender (G. pigra) and stout (G. adriatica) morphotypes would partition interstitial niche space based on sediment size, which may affect the vulnerability of the species to changes in gravel beach composition due to coastal development. We detected substantial differences in gravel composition within and among the sampled beaches which suggests scope for microhabitat partitioning in Gouania. Indeed, we found significant relationships between species identity and the presence/absence and abundance of individuals in hauls based on their positioning on PC1. Our results suggest that modifications of gravel beaches through coastal development, including beach nourishment, intensifying coastal erosion, or artificial beach creation, may have detrimental consequences for the two species if sediment types or sizes are altered. We posit that, given the simplicity and efficacy of our sampling method and the sensitivity of Gouania species to prevailing gravel composition, the genus could serve as an important indicator for gravel beach management in the Mediterranean ecoregion.

Keywords: Adriatic Sea; Croatia; Mediterranean Sea; cryptobenthic fish; divergent ecological selection; ecological diversification; tourism.

FIGURE 1

Map of the Adriatic Sea, including the study area (red rectangle), the studied Gouania species, their distribution ranges (light and dark blue shading). Additionally, the potential distribution range of G. hofrichteri is shown in light grey (with a northernmost record from Pelješac peninsula).

FIGURE 2

Geological fabric and sedimentologic composition of the eight investigated sites from the northern Adriatic Sea (Istria and Kvarner area). (a) Representative images and locations of the study sites. (b) Relative weight distributions of five investigated size fractions (I: >46 mm, II: 25–46 mm, III: 13–25 mm, IV: 5.5–13 mm, V: 1.5–5.5 mm) from 20 random hauls per site.

FIGURE 3

Principal components analysis (PCA) based on the relative weights of five size fractions (I: >46 mm, II: 25–46 mm, III: 13–25 mm, IV: 5.5–13 mm, V: 1.5–5.5 mm) separately shown for each site. The shape of data points represents the haul category (Gouania adriatica, G. pigra, empty haul or both species in the same haul). Percentages on the x‐ and y‐axis labels indicate explained variation. Loadings for the first two PC axis and their relative contributions are shown in the bottom right. For a better readability of the graph the single locations are plotted in different panels, even though they come from the same PCA.

FIGURE 4

Microhabitat associations of the sympatric Gouania species, G. adriatica and G. pigra. (a) Comparison of relative weight distributions among the five investigated size fractions in hauls that contained either G. adriatica (dark blue) or G. pigra (light blue). p‐values were obtained via Kruskal‐Wallis tests and are in bold font if statistically significant. (b) Differences in the PC1 score of hauls that contained either G. adriatica (dark blue) or G. pigra (light blue), which is mainly associated with changes in the largest size fraction. (c) Logistic regression models showing the probability of occurrence in hauls based on their PC1‐score for the two species. Lines and confidence bands show the model fit, while jittered dots represent the raw data. The seperate density plot for the absence (0) and presence (1) values for both species can be found in Appendix S10. (d) Abundance of G. adriatica and G. pigra in hauls based on their PC1‐score. Lines and confidence bands show the model fit from a zero‐inflated Poisson model, while dots represent the raw data.

FIGURE 5

Temporal and ontogenetic determinants of microhabitat occupation for Gouania adriatica (top) and Gouania pigra (bottom) in Zelenika. The first column indicates changes in total length (TL, in mm) between 2020 and 2021, while the scatterplots display changes in body size across PC1 and PC2 across the 2 years. Each dot in the scatter plots represents a single haul, with shapes indicating the year of collection.