1. Introduction
The deep sea, the part the ocean deeper than 200 m, represents the vastest ecosystem on Earth and provides ecosystem goods and services that are crucial to human wellbeing [
1,
2,
3,
4]. Among the biodiversity hotspots hosted by the deep sea ecosystem, cold-water coral (CWC) communities represent complex three-dimensional habitats that can provide reproductive areas and refuge to a large variety of valuable fish and invertebrates of commercial interest, both in the adult and juvenile stages [
5,
6,
7,
8]. CWC communities may act as a feeding area, a refuge from predators and fishing activities, and a spawning and nursery area for many fish species, and these areas generally show higher diversity and abundance than in adjacent soft bottom areas [
6,
8,
9]. Although several studies have reported higher diversity and densities of ichthyofauna associated with CWC, it is still difficult to demonstrate whether the CWC habitat or its complexity are the attractive factor for fish species [
9,
10,
11], Most studies, in fact, reported species that are distributed at comparable depth and common also in different types of habitats such as soft and rocky bottom areas. Although the fish fauna associated with CWC habitats may be not exclusive of this habitat, this species seems to benefit from the shelter provided by the structures built by the CWC and from the enhanced trophic conditions. A higher density of zooplankton, in fact, represents a potential trophic resource for planktivorous fish and can lead to a higher density of small invertebrates which are prey for benthic feeders and scavengers [
6,
8,
12]. These ecosystems support a high biodiversity and high biomass, and they are impacted by commercial fishing activities causing extensive damage to CWC, resulting in productive ecosystems being transformed into coral rubble [
13,
14,
15].
The Apulian margin (central Mediterranean) is characterized by the presence of an almost continuous belt of vulnerable marine ecosystems (VMEs) characterized by the presence of CWC communities whose real extension is still poorly understood. The Food and Agriculture Organization (FAO) indicates the following criteria for defining what constitutes a VME: (1) uniqueness or rarity; (2) functional significance of the habitat; (3) fragility; (4) life history traits of component species that make recovery difficult; and (5) structural complexity [
16]. The exploration of the Apulian margin from the southern Adriatic to the northern Ionian Sea (central Mediterranean) has led to the discovery of several CWC communities distributed between 300 and 1100 m depths that can represent a network of VMEs mostly built by the colonial scleractinians
Madrepora oculata and
Desmophyllum pertusum [
17,
18,
19]. CWC habitats have been described on the Gondola slide off Manfredonia, inside Bari Canyon (BC), off Monopoli, Otranto, Tricase, and up to the Santa Maria di Leuca (SML) CWC province, which is the largest living occurrence described in the Mediterranean Sea; lastly, off the Porto Cesareo marine protected area, living colonies of
Dendrophyllia cornigera have also been observed down to a 217 m depth [
17,
18].
The exploration of fragile, heterogenous, and complex habitats like those built by CWC species requires the use of low-impact sampling techniques, such as experimental longlines or video systems, such as baited remote underwater video surveys (BRUVSs). A baited lander represents an effective method to explore sensitive habitats with complex geomorphology, such as seamount, canyon, and CWC communities, as well as their associated benthopelagic biodiversity [
20,
21,
22,
23,
24,
25,
26]. In particular, the baited lander is a low-impact non-extractive tool for collecting data on megafauna diversity, abundance, and behaviour without damage to habitat former organisms and associated vulnerable species [
10,
21,
22,
23,
24,
27,
28]. An experimental longline with a small number of hooks is a tool that can allow the capture of fish fauna with low impact in a heterogeneous and complex habitat, providing complementary information to that provided by BRUVSs [
21,
29]. Therefore, the aim of this study is to provide a further contribution to the knowledge of fishes and benthopelagic fauna distributed in the CWC-VMEs along the Apulian margin using low-impact sampling tools.
4. Discussion
The utilization of two low-impact sampling techniques allowed for the collection of new information on the distribution and abundance, size, and maturity of fishes associated with deep CWC-VMEs along the Apulian margin. The use of a baited lander allowed for the detection of species which can be barely sampled with traditional sampling tools and provided observations on the behaviour of some cartilaginous fishes.
The faunal assemblages of the VMEs are representative of the deep Mediterranean Sea and are also related to the presence of scavenger species attracted by the odour plume produced by the baits. Although the presence of several species in common in all the VMEs, differences among the VMEs could be explained by the different densities of some species. The highly significant difference between BC and SML and between Mn and SML detected by the MEMO lander can be explained with a higher abundance of sharks recorded both in BC and Mn and by the absence of G. melastomus in the video recorded with the MEMO lander in SML.
Although a wide period of sampling (2010–2019) could be subjected to environmental changes, the samplings for this study were conducted in deep sea habitats characterized by stable environmental conditions and small fluctuations in the sea bottom temperature [
42]. The VMEs explored are also characterized by occasional fishing pressure due to the presence of conservation measures such as the FRAs instituted in BC and SML and the presence of an irregular morphology of the seabed. Moreover, information on the species assemblages of the trawlable muddy bottom of the northwestern Ionian Sea was collected by Maiorano et al. [
43], who found no significant variations in the distribution and abundance of the fish species collected inside VMEs during the present study. Lastly, the species observed with the MEMO lander and collected using the experimental longline are characterized by high longevity that could minimizes the effects of a long period of sampling.
The most abundant teleost fish in all the VMEs explored with the MEMO lander was
P. bogaraveo, whereas in the same VMEs, using the experimental longline, the most abundant species was
H. dactylopterus. This could be due to the different depths and period of the day explored with the two different sampling tools and the vertical migratory behaviour of
P. bogaraveo, which shows a shallower distribution during the daytime and a deeper distribution due to migration during dusk and early nighttime [
44,
45]. Capezzuto et al. [
46] detected a relevant device effect for
H. dactylopterus, showing higher abundances for a longline than for a baited lander. This species is a typical sit-and-wait ambush predator, feeding mainly on benthic crustaceans and fishes, as well as on planktonic organisms [
47].
H. dactylopterus is frequently associated with submarine canyons and CWC habitats [
6,
26,
48,
49]. This species, in fact, is the most abundant species collected close to corals in the eastern Ionian Sea [
50], in the northwestern Ionian Sea CWC province [
8], in the Quirra Canyon (Tyrrhenian Sea) [
51], and in French Mediterranean submarine canyons [
52]. This habitat preference can be explained by the enhanced availability of zooplankton and small crustaceans, which are suitable prey for
H. dactylopterus [
47,
49].
Although
M. merluccius is one of the most abundant species collected using the longline in all the VMEs, its recorded abundances with the MEMO lander in the same VMEs were low. This could be explained with the daily vertical migration carried out by this species and the different period of the day explored with the two different sampling tools.
M. merluccius, in fact, feeds in mid-water or near the surface during the night and spends extended periods of time near the seabed in the daytime [
53,
54]. Using both sampling techniques,
C. conger was recorded in all the VMEs explored in this study. This species is a large opportunistic feeder, feeding mainly on bethopelagic and benthic prey, with fishes as main prey group [
55]. This species shows a preferential distribution in complex habitats such as rocky bottom areas and those built by CWC [
7,
8,
55].
Phycis blennoides was also observed with lower densities than those detected using the longline in the same VMEs. This highlights the importance of using different low-impact tools to have more complete information on the biodiversity of the megafauna in heterogenous and complex deep sea sensitive habitats. The two sampling tools, in fact, have a different efficacy and selectivity. The dimension of the hook used in the longline can be selective in terms of species composition and dimension of the individuals, whereas the presence of the lights on the baited lander could attract or lead to light avoidance in some species. Given these differences, these sampling tools should be used simultaneously to obtain reliable information on ichthyofauna biodiversity. However, the comparison of the two tools is not an objective of this study.
P. bogaraveo showed a higher abundance of mature females in BC, whereas a higher abundance of immature and maturing specimens was collected in Mn and SML. This species seems to prefer complex habitats such as CWC, canyons, and seamounts, where usually larger individuals are more abundant [
8,
44,
56].
P. bogaraveo shows an ontogenetic habitat shift with juveniles up to 180 mm in TL mainly distributed in shallower waters and muddy bottoms, and larger individuals distributed in deeper water preferably characterized by a three-dimensional habitat [
56,
57]. The residency of sub-adults and adults of
P. bogaraveo at the Condor seamount (Azores, mid-north Atlantic) was confirmed through acoustic telemetry [
44]. The strict association of this species with CWC habitats was also confirmed by D’Onghia et al. [
8], comparing megafauna distribution in coral versus non coral habitats; this species, in fact, was collected exclusively in coral habitats.
P. bogaraveo, moreover, was exclusively collected inside BC using the longline [
58]. These findings suggest that this species is more abundant in habitats less accessible to trawling such as CWC habitats and canyons.
The fact that a male specimen of
C. granulosus was recorded after one year in the same VME off Monopoli is of crucial importance, since this can give an indication of some form of site fidelity that this shark shows to the VME, in which it can find trophic resources or protection from fishing activities carried out on surrounding muddy bottoms.
C. granulosus is a large deep-water shark that lives in the outer continental shelf and upper slope of the Mediterranean Sea from 100 to 1200 m in depth and is classified as Critically Endangered in this basin. It is a very active feeder and usually preys on teleost and squids [
59,
60].
D. licha, a shark classified as Vulnerable in the Mediterranean, was the species that was observed returning most often in the same deployment, especially in Mn. This could be due to the very slow swimming speed measured for this species [
61], leading to very short and limited movement from the MEMO lander and indicating probable fidelity to the area. This may also be true for other species, such as
H. griseus and
D. oxyrinchus, but only a greater number of observations could confirm the fidelity to the area.
The presence of commercial species and cartilaginous fishes and the presence of large and sexually mature individuals of
G. melastomus, H. dactylopterus, and
P. bogaraveo in all the VMEs confirm that the network of CWC-VMEs along the Apulian margin can act as a network of refuge areas for some species exploited during fishing activities in the surrounding muddy bottoms [
5,
7,
8,
18,
58,
62,
63]. CWC communities and submarine canyons indeed represent suitable areas in which these species can spend crucial phases of the life cycle such as reproduction and spawning, thus providing an essential fish habitat (EFH) for species threatened by anthropogenic impacts such as fishing activities carried out on the seabeds surrounding the VMEs. Bottom trawling is one of the most important anthropogenic threats to CWC ecosystems [
6,
13,
64,
65,
66] but the indirect impact of habitat destruction on the demersal resources is still poorly explored.
The main impact of trawling on CWC communities is mechanical damage and the destruction of the three-dimensional structures of the colonies. The impact of bottom trawling on CWC habitats has been widely documented in the Atlantic and in the Mediterranean Sea [
13,
14,
15,
35,
67,
68,
69,
70,
71]. In SML, Savini et al. [
35] recorded trawling traces using an ROV and D’Onghia et al. [
13] observed the presence of longlines entangled in corals and trawl scars using towed cameras. In northwestern Sicily (southern Tyrrhenian Sea), the most important anthropogenic impact on CWC habitats, observed using an ROV, was represented mainly by longlines and ropes entangled with or hanging between rocks and organisms, causing heavy impacts on the community [
65]. In addition to the direct destruction caused by fishing activities, bottom trawling alters the sedimentary condition through the resuspension of large amount of sediment due to the mechanical effects of the fishing gear [
69,
70].
The governance of VMEs requires information regarding biodiversity and the presence of endangered species and the identification and protection of habitats that can act as spawning and nursery areas for these species [
6,
72]; this necessitates the utilization of low-impact sampling methods in order to propose new and more effective conservation measures. A baited lander is a low impact non-extractive sampling method, and its non-destructive nature allows for its deployment in structurally complex habitats, such as CWC communities and submarine canyons. A BRUVS is a passive sampling tool that enables the investigation of the abundance and behaviour of rare and threatened species and the exploration of sensitive and vulnerable habitats that could be damaged by traditional sampling tools [
21,
24,
73]. BRUVSs allow for the direct observation of species behaviour and enable the analysis of macro- and megafauna interaction and feeding behaviour [
23,
27,
74]. Lastly, the videos recorded by the benthic lander can be examined by different observers, allowing for impartial and repeatable data collection [
24,
27,
73]. However, some difficulties should be considered when using a baited lander. The precise positioning of a benthic lander can be difficult, especially in complex habitats in which a free-falling lander could miss the target. Moreover, a baited lander attracts principally scavenging species due to the presence of the bait, which make a BRUVS a selective sampling tool [
24]. Lastly, although the videos are permanent and can be analyzed several times, the correct identification of ichthyofauna to the species level can be difficult.
Currently, the conservation measures adopted for the CWC ecosystems along Apulian coasts are the fishery-restricted areas (FRAs) established by the General Fishery Commission for the Mediterranean (GFCM). In particular, the SML CWC province FRA was established in 2007, whereas the BC CWC province FRA was established more recently in 2021. Although CWC ecosystems are included in the list of VMEs and in Annex I of the Habitat Directive, none of the CWC areas along the Apulian margin have been designated as a Natura 2000 site, which is an important European conservation tool based on the Habitat (92/43/EEC) and Birds (2009/147/EEC) Directives [
75,
76]. The institution of a network of high-sea marine protected areas and offshore Natura 2000 sites could represent a more effective conservation measure that can ensure the protection of vulnerable ecosystems and species in combination with the management of deep-water fishery resources.