Habitat preference of fish species as revealed by lure-assisted visual census

To pursue the COREBIO goal of understanding the ecological processes driving organization of fish communities, we estimated the effect of habitat heterogeneity on fish community structure. We compared fish abundance, richness, and diversity in homogeneous habitat patches and at heterogeneous habitat transitions along visual census transects randomly sampled. Within 135 discrete shallow water sites that spanned the N-S and E-W range of the Croatian Coast, 158000 individual fish were observed. We predicted that species seek structural heterogeneity due to their predation mode or due to their ability to exploit the more varied resources at edge habitats, and that edge assemblages are unique from those in source habitat. We found a significant heterogeneity effect: edges associated with rock, sand, and Posidonia seagrass habitats contained significantly higher fish abundance, species richness, and diversity than homogenous habitats. No edge effect was found after excluding species more abundant at edges, indicating that the edge effect is not based on the random movements of fish originating from multiple adjacent source habitats. We identified eight species with a significant overall preference for edge habitats, “edge-seeking” species. More than half of these are known as wait-chase predators and all but one wait-chase predator species had absolute highest abundances in heterogeneous habitats. No other guild displayed such a significant association with edges. Species observed to be significantly more abundant in heterogeneous habitats only, are known to have diets with significantly higher taxon diversity. This indicates that edges may be attractive to any fish with general feeding habits. Our results indicate that heterogeneous habitats attract a unique fish assemblage and are preferred by several edge-adapted species, including some with economic importance. Current management of Mediterranean fish biodiversity should recognize the conservation value of edge habitats.

We performed over 10000 lure-assisted, visual-census, field transects of fish communities at outer coast study sites in Croatia, for an average of 102 transects per site (minimum 12, maximum 377). Sampling was performed between May and October, with a mean of over 3000 transects per year from 2014 to 2016. Sites were inside and outside five MPAs with different levels of harvest regulation and enforcement. Sites differed in protection status, shoreline development, the presence and extent of urchin barrens, mean transect depth (1.4 m to 8.1 m), and benthic habitat. Benthic habitats included rocky reefs and unconsolidated and vegetated (algae, Posidonia oceanica, Cymodocea nodosa) sediments.

In order to match observed fish with the small-scale habitat they occupied at the time of observation, proportional cover of all present basic habitat types, including: rock (R), bare sediment (U), algae on sediment (A), Posidonia oceanica (P), and Cymodocea nodosa (C) have been estimated for each transect. Based on these estimates, dominant habitat types were identified as combinations of basic habitat types that had a relative coverage of ≥ 25% within individual transects. Details of the method can be found in Kruschel and Schultz (2012). Dominant habitats with only one contributing habitat (R, U, A, P, C) were categorized as homogeneous, while dominant habitats that included distinct patches of more than one contributing habitat were categorized as heterogeneous (all possible combinations). Heterogeneous habitats are therefore functionally identical to habitat transitions or habitat edges.

To uncover relationships between the functional identity of fish and their habitat use (homogeneous vs. heterogeneous), the observed 82 species and 16 taxa identified to a higher taxonomic level were categorized into seven groups using the following functional traits: carnivores (including omnivores) were grouped by predation mode and mobility as either substrate dwellers, or ambush predators, or cruisers, or wait-chase predators as defined in Kruschel and Schultz (2011). Planktivores were identified by diet and food item information given by the database Fishbase (Froese and Pauly, 2000), and in Bell and Harmelin-Vivien (1983) and in Guidetti and Sala (2007). Small epibenthic/cryptobenthic fish defined by Kovačić et al. (2012) and with a common length < 10 cm as defined in Fishbase by Froese and Pauly (2012) were grouped together regardless of their mode of predation, locomotion, or trophic position and primarily based on their common small size and their close association with the benthos. This functional group also included epibenthic juveniles (<5 cm) that showed sustained physical contact with the substrate, including all juvenile sized gobies. Benthopelagic juveniles < 5 cm in length (JUV) were grouped based on their common small size and their lesser association with the bottom substrate (no sustained physical contact with the substrate).

Of the 53 taxa that had at least one significant association with one of the five homogeneous or their derived edge habitats, 14 had higher abundances in at least one homogeneous and at least one heterogeneous habitat, 20 were significantly more abundant only in homogenous habitats, and 19 taxa were significantly more abundant only in heterogenous habitats (see the following table). Of the 53 taxa that had at least one significant association with one of the five homogeneous or their derived edge habits, 18 had their absolute highest abundance in one of the homogeneous habitats and 28 at one of the edges.

Thirty-two taxa had overall significantly higher abundances in homogeneous habitats. Nine taxa were overall significantly more abundant in heterogeneous habitats. Eight taxa, identified to species, significantly preferred edge habitats over homogeneous source habitats -- the “edge seeking species”, including: Gobius cruentatus, Gobius geniporus, Diplodus vulgaris, Diplodus annularis, Serranus cabrilla, Serranus hepatus, Mullus surmuletus, and Coris julis. When excluding all species with higher abundance at edges, this edge effect disappeared. Instead, significant positive effects of habitat homogeneity on fish richness and diversity were found.

Of the total number of 122 taxa that fitted only one of the seven functional groups, 43 were more abundant in heterogeneous and 79 were more abundant in homogeneous habitats (Fig. 3). Significantly higher numbers of ambush predator species (14 > 1) and EBCB taxa (24 > 4) were more abundant in homogeneous habitats than in heterogeneous habitats. Significantly higher numbers of wait-chase predator species (10 > 1) were more abundant in heterogeneous habitats (edges) than homogeneous habitats.

Overall, we observed a significant positive edge effect across all habitats and sites. Thirty-six percent of all taxa with significant habitat associations preferred exclusively heterogeneous habitats, 26% favoured some homo- and also some heterogeneous habitats, and 38% showed exclusive preference for some homogeneous habitats. We demonstrated that the edge-preferring species are known to exploit a more varied diet than the non edge-preferring species. The observed edge effect was clearly based on edge-preferring species and not on random movements of fish associated with the source habitats. Fish with higher frequency at edges are solely responsible for the observed overall significant edge effect on abundance and richness and diversity at rock, unconsolidated sediments, and Posidonia beds. The majority of these edge-seeking species are wait-chase predators that are adapted to edges by their unique predation mode, while others seem to be generalist feeders capable of exploiting the more varied food offered at edges. Habitat edges are an abundant component of the natural benthic environment, create unique conditions and support a unique fish assemblage. Efforts to understand and conserve the marine environment would benefit from a program of explicit recognition and quantification of habitat edges and their impacts on the faunal community.