Search

Marine underwater habitats dominated by seagrass Posidonia oceanica play an essential role in fish community assembly, affecting taxonomic and functional diversity, abundance and fish behavior. The value of seagrasses as habitat depends on the spatial arrangement of the seascape elements and the availability of alternative habitats. Little is known about the effect of the seascape context of P. oceanica meadows on fish assemblages in the Mediterranean Sea. To identify P. oceanica meadows' relative importance as a habitat for fishes, fish communities in the Croatian Adriatic Sea were investigated, using SCUBA lure-assisted visual census. The results show a significant effect of different arrangements of P. oceanica meadows' seascape elements and surrounding habitats on fish community structure. Fragmented mosaic meadows with P. oceanica growing directly on and between rocky-algal reefs/boulders had significantly higher fish abundances compared to both types of continuous meadows (bordering rock and bordering sand). Continuous meadows bordering sand harbored the highest number of unique species. Evidence that alternative structured habitats within proximity to seagrass beds may affect the community structure of associated fish assemblages is provided, highlighting the need to consider P. oceanica meadows' seascape context in conservation management and experimental design for fish community structure., Naselja morske cvjetnice Posidonia oceanica igraju važnu ulogu u strukturiranju zajednica riba te utječu na taksonomsku i funkcionalnu raznolikost, brojnost i ponašanje riba. Relativna vrijednost naselja morskih cvjetnica ovisi o prostornoj organizaciji elementa podmorskog okoliša te dostupnosti alternativnih staništa. Međutim, nedovoljno je poznato koji sve čimbenici utječu na strukturu zajednice riba povezanih s ovom morskom cvjetnicom. S ciljem utvrđivanja utjecaja različite prostorne organizacije elementa podmorskog okoliša na zajednice riba, proveden je vizualni cenzus uz pomoć mamca u naseljima cvjetnice P. oceanica u hrvatskom dijelu Jadranskog mora. Utvrđen je značajan utjecaj različite prostorne organizacije elementa podmorskog okoliša na strukturu zajednice riba – u rascjepkanim mozaičnim naseljima u kojima se P. oceanica isprepliće s kamenom podlogom obraslom makro-algama zabilježena je značajno veća brojnost riba u odnosu na kontinuirana cjelovita naselja (uz sediment ili uz kamenitu obalu). U radu su predstavljeni dokazi da dostupnost alternativnih strukturiranih staništa u neposrednoj blizini naselja morske cvjetnice P. oceanica utječe na strukturu zajednice riba koje u njoj žive. Neophodno je uzeti u obzir kontekst prostornog rasporeda elemenata podmorskog okoliša kojima dominira morska cvjetnica P. oceanica prilikom planiranja znanstvenih istraživanja, ali i kod upravljanja priobalnim područjima.

Fish communities within two Croatian MPAs were surveyed using both diver-operated underwater video (DOV) via SCUBA and lure-assisted visual census (Lure-UVC) via snorkling. Both methods were applied at similar depths within replicate MPA locations, Kornati NP (4) and Brijuni NP (2). We predicted that the methods detect similar fish communities, habitat mosaics, and fish habitat preferences. To avoid inter-observer variability- errors, fish and habitat identification in all surveys was done by the CK. Habitat types were defined by a common protocol. Higher species richness (35 > 30) was detected by lure-UVC in Kornati. In Brijuni, DOV detected more species (31 > 25). The MPA-specific communities observed by the two methods were 72% similar within Kornati NP and 92% similar within Brijuni NP (Sørensen Index). Both methods recognized the same six habitats in Brijuni NP and five common habitats in Kornati NP, with one unique observed by DOV. In Kornati NP, Lure-UVC detected significant habitat preferences for 10 (28%) species of which nine preferred transitional habitats, the edges between R and U (RU) and R and P (RP) ; DOV detected significant habitat preferences for 15 (50%) species of which 12 preferred either R or RP transitions. In Brijuni NP lure-VC detected four (13 %) and DOV seven (28 %) species with significant habitat preference, and all eleven species were significantly attracted to transitions from rocky reefs to sedimentary bottoms (RU) and all significantly avoided bare sand (U). We conclude that the two methods are comparable in their ability to sample fish communities, habitats and habitat preferences. Decisions about which method to use should focus on availability of resources and target depths. Lure-UVC is more cost and time efficient because it does not need SCUBA and video-processing but requires the presence of a fish/habitat expert in the field, preferably the same individual. DOV requires costly SCUBA operation and lengthy post- video analysis in the lab but allows for multiple experts relying on permanent records. Sample sizes per unit field-time is high for DOV and depends on the number of divers operating DOV. DOV can be applied at wider depth range than surface-based UVC.

Fish stocks should be monitored with fisheries-independent and non-destructive methods. In Croatia, we are using baited, remote, underwater, stereo video (BRUV) and diver-operated videovisual census (DOV), potentially operated inexpensively by personnel untrained in fish biology, but requiring lengthy and tedious office labor by experts to classify, count, and measure fish. Automated methods will potentially greatly reduce this labor cost and allow greater video dataprocessing per unit time, and increased statistical power for detecting spatio-temporal variation in fish populations and fish communities. We are currently developing computer-vision tools for automated processing of high resolution underwater videos from BRUV and DOV videos taken in the shallow Croatian Adriatic under varied fish assemblage, water, and habitat conditions. The method is based on deep machine learning and consists of three major steps: detection, classification, and movement tracking of fish. The current prototype achieves accuracies for fishspecies classification of 69%, 94% and 98% on the Croatian dataset (794 images), the fish4knowledge 2012 dataset (27370 images) and the seaclef 2015 dataset (22443 images) respectively. The positive correlation between dataset volume and classification accuracy indicates that experts should annotate over 25000 images for the algorithm to be trained to achieve accuracies above 90%. In the near future we plan to embed our algorithms into a lifelong machinelearning framework that continuously improves through incremental learning from newly arriving annotated data. Human-machine collaboration is facilitated by the software tool L3P which allows for annotation within a web-based image-processing engine. This tool enables fish experts to improve analysis algorithms in the cloud without specialized computer-vision knowledge. Our core system achieved best results in a competition to estimate fish abundances within the seaclef 2016 dataset and is foreseen to accomplish several more specialized tasks. We present an innovative machine learning approach that will utilize human-machine collaboration and is suitable for the overall challenge of monitoring marine species and habitats. This work was partially supported by the Croatian Science Foundation, under the project COREBIO (3107).

Lure-assisted visual census, a fisheries-independent and non-destructive method for quantifying fish abundance and benthic habitat, was used to assess fish communities inside Croatian (eastern Adriatic Sea) marine protected areas, and outside these areas at similar locations. Marine protected areas included National Parks Brijuni, Kornati, Mljet, and Nature Park Lastovo. A total of 19678 visual transects were followed at 149 locations throughout the Croatian nearshore at depths from two to seven meters. Relative species abundances were compared within a generalized linear mixed model with MPA protection as a fixed effect and location as a random blocking effect, and species accumulation curves were constructed using Chao, jacknife, and bootstrapping methods. We found the total species richness was slightly higher inside all four MPAs than outside, but communities were highly significantly different inside versus outside all MPAs. The results indicate that organization of fish communities is highly sensitive to 1) regulation of fishing activities, and 2) degradation of bottom habitat, such that commercially valuable predatory species (including Diplodus sargus, D. puntazzo, Mullus surmuletus), and species dependent on intact and undisturbed macro-vegetation (Oblada melanura, Sarpa salpa) and rocky reef habitat (Epinephelus marginatus) are more abundant inside MPAs where habitat is less disturbed. Lure-assisted visual census is a powerful non-destructive method for quantifying the effects of fishing regulations on populations of common and commercially important fish species. This work was partially supported by the Croatian Science Foundation, under the project COREBIO (3107).

Baited remote underwater video is a non-destructive method for censusing fish species abundance that is currently under consideration for use in the Mediterranean Sea for monitoring fish stocks and the effectiveness of fisheries regulations. We tested the method in the eastern Adriatic for effectiveness in detecting impacts of habitat degradation and fish harvest on fish abundance, size, and community assembly. A total of 215 BRUVs were deployed at 10 locations during all months from April to September of 2015 and 2016 at and near Kornati National Park (KNP) in Croatia, at benthic habitats including unconsolidated sand/gravel to rocky reefs, with and without macro vegetation consisting of seagrasses Posidonia oceanica, Cymodocea nodosa, or macroalgae Cystoseira species. We analyzed species abundance using generalized linear models with protection or habitat status as fixed predictor factors, location and date of sampling as a random blocking factor, and the maximum number of individuals visible within any single video frame as the proxy of abundance within a species. To estimate asymptotic total species richness we used the Chao estimator, or standard methods of jacknifing or bootstrapping the dataset. We found very highly significant differences in species richness across 10 sampling locations, with the maximum estimated richness within KNP ranging from 53 to 65 (SE 2 to 10), and outside KNP ranging from 24 to 27 (SE 1 to 2). In addition, we find significantly higher species richness associated with structured and edge habitats, such as rocky reefs, and lower richness associated with uniform seagrass or anthropologically degraded habitat. Taxa found more abundant inside the MPA include species heavily targeted by local fisheries, such as Diplodus species, Spaurus aurata, Dentex dentex, Mullus surmuletus, and schooling pelagic genera such as Spicara. We conclude that BRUV has power to detect responses of individual species both to harvest, and to human habitat disturbance, and can assist in testing hypotheses about fish community assembly in the Adriatic Sea. This work was supported by the Croatian Science Foundation under the project COREBIO (3107).

Trammel-net fish surveys in the Croatian Adriatic between 1993 and 2009 identified increases in abundance of four meso-predator species, hypothesized to be indicators of positive effects from fisheries regulations implemented within those years. An alternative explanation is that ongoing harvest of large predatory species continues to release these populations of indicators. We tested these two hypotheses by comparing fish population densities inside and outside MPAs and no-take areas. If fishing restrictions are responsible for increase in indicator species then 1. densities of indicators should be uniformly higher in more fisheries- restricted zones. If fishing restrictions increase top predators, which then reduce the populations of meso-predators, then 2. densities of predators should increase, and their potential prey decrease, in response to fisheries restrictions. We performed trammel-net surveys at four sites, three within and one outside the Kornati National Park, Croatia in 2009, 2013, 2014 and 2015. We matched the trammel specifications, including habitat and depth, of the 1993-2009 survey. The four indicator species reached highest abundances at different locations. Symphodus tinca peaked within the no-take zone, Pagellus erythrinus at a less protected site within the MPA, and Mullus surmuletus and Scorpaena porcus peaked outside the MPA with no significant differences across MPA sites. This contradicts hypothesis 1. Hypothesis 2 can be addressed by comparing Scorpaena species, which are trophically connected. Two Scorpeana species coinhabited the site with Scorpaena porcus. The larger S. scrofa significantly increased in abundance from 2009 to 2015 within the MPA with significantly higher abundances in the no-take zone for all year/site combinations. The smaller S. notata showed no significant changes across MPA sites and displayed negative trends within the no-take zone. In less protected MPA sites, S. porcus and S. notata displayed opposite abundance trends: positive for the larger S. porcus and negative for the smaller S. notata. These results support hypothesis 2. We conclude that increases in meso-predators is not evidence for the general success of fishing regulations, but of relatively higher vulnerability of large predators to fishing, and subsequent predator release. This work was supported by the Croatian National Science Foundation under the project COREBIO (3107).

Predator - prey relationships are complex and chemical communication plays an important role. The role of chemical communication between Scorpaena porcus and Pomatoschistus spp. was assessed including the added effects of a herbicide. Tank trials featured living free - roaming and confined S. porcus as well as a maquets. This allowed for manipulations of cues perceived by the prey. Trials were performed in chemically polluted and unpolluted seawater. Pomatoschistus spp. assessed predation threats using multiple sources of information. Kairomones of S. porcus were detected by the prey, yet simultaneous visual and olfactory cues resulted in stronger anti - predator responses. Glyphosate disrupted predator - prey interactions.This work was supported by the Croatian Science Foundation under the project COREBIO (3107).

The measurement of degree of health of an organic system is a fundamental goal in both medicine and conservation science. In medicine, a large population of individuals is treated via a three-tiered process: 1) an inexpensive screening test is used to screen out individuals who then are assumed healthy ; 2) those remaining with positive test result are then given a series of more expensive and accurate tests to further screen out healthy individuals ; 3) those remaining with positive tests in both tiers are then given expensive treatment. Conservation science has a similar goal: to distinguish between healthy and impacted ecological populations, processes, and habitats, so that the high cost of intervention is not wasted but applied exactly where it is needed. In the marine environment, assessment of ecological health is especially difficult and costly, and screening tests that are fast and inexpensive can potentially realize large savings in the cost of identifying and conserving marine values and resources. Here we develop a decision tree with associated economic costs in a three-tiered conservation context. We find that, in contrast to medical application with rare disease, in marine environments with high probability of human impact, false positive results have a low economic cost, and the major cost risk is the loss of ecological services due to false negative results, which greatly increase in a two-tiered testing framework, and can become catastrophic even if the second-tier tests have low false-negative error. Ecological screening tests should be avoided in sites likely to be impacted, and instead, accurate assessments using tests with very low false-negative rates (i.e. high statistical power) should be the first line of attack and greatly reduce overall costs. Our analysis shows a fundamental contrast between medical and conservation economics, caused by the high marginal cost of loss of ecological services. The common use of screening tests in marine conservation (e.g. in seagrass monitoring) should be reevaluated and in many cases abandoned to avoid loss of ecological services due to false negative results. There is an ongoing urgent need for the development of fast and inexpensive marine screening tests with low false-negative error rates. This work was partly supported by the Croatian Science Foundation, under project COREBIO (3107).

Interference competition can exclude subordinate species from habitats and alter community structure. An example is the predatory release of mesopredators resulting in changes in their prey and competitor community. Baited, remote, underwater videography is a powerful census tool that can be biased if dominance excludes species from view. At interspecific encounters during BRUV deployments in the Adriatic Sea, Croatia, no such general bias occurred. BRUVS detected few negative co - occurrences for species pairs and they rarely interfered at opportunities near the bait. Most species pairs positively co -occurred and were unique for each habitat type, which indicates that habitat heterogeneity plays an important role in driving fish biodiversity. This research was supported by the Croatian Science Foundation under the project COREBIO (3107).

The Center for Interdisciplinary Marine and MAritime Reseach (CIMMAR) strives to become a Croatian platform to support goals of integrated coastal zone management, sustainability, basic and applied research, and biotechnological applications to develop and apply interdisciplinary tools for the adaptation of Croatian society to the foreseeable environmental changes of the 21st century. At the centre of the sustainability question within ICZM internationally is the sustainability status of aquaculture and its potential to have positive synergies with other stakeholder activities, including tourism. While aquaculture is predicted to grow in stature, economic share, and importance it is also considered to have questionable sustainability status and a negative impact on touristic activities due to various reasons. This can be ameliorated by ecosystem based approaches, including integrated, multi-species, multi-trophic aquaculture (IMTA) and the use of habitat improvement to foster complex natural communities within or near aquaculture facilities. The later increases the local capacity for recreational ecotourism and the enjoyment as well as the local exploitation of wild species.

Demersal fish and benthic macroinvertebrates were censused in spring 2009 in six embayments and one open Adriatic shallow water location. Relative faunal abundance and species richness, diversity, and evenness were calculated and compared across bays and across natural and developed coastline within bays. We found that the faunal communities of the most enclosed bays, Novigrad Sea and Karin Sea, (1) exhibit significantly higher fish richness and faunal diversities than any of the other bays, (2) were more dissimilar from the open Adriatic community, (3) experienced significant reductions in biodiversity at developed compared to natural coastlines and (4) harbor two threatened fish and highest densities of juvenile fish. Their diversity, uniqueness, und sensitivity to development make Novigrad and Karin seas candidates for legal protection.

The European eel, Anguilla anguilla, is critically endangered and according to European council regulations member states need to establish stock recovery measures. Such plans must estimate eel escapement to the sea, thus life stage and habitat-based assessment of coastal eels is crucial, and considering their red-listed status, non-invasive assessment methods are desirable. Within the scope of a pilot eel monitoring, the spatial and temporal distribution of eel have been investigated in the shallow benthos of the Central Croatian Adriatic. Using geo-referenced and video supported SCUBA visual census, transects of up to 1000+ m were followed during the day and then with DGPS sub-meter accuracy re-navigated at various night times. Two bays were investigated at various study sites per bay and over complete moon cycles in the spring and fall of 2009/10 and in the spring of 2011. Eel presence and habitat occupation at the instant of observation were recorded along with their exact geographical location. This permitted the quantification of eel densities and habitat preferences at various scales. Significant positive associations with dense seagrass patches were discovered for the eel, and such habitat preference explained most of the eels’ large scale distribution. Furthermore eel abundance and activity showed significant temporal variation, as they peaked several hours after sunset, at quarter moons, and in the fall. Due to the direct encounter by the observer, additional data, including eel size and behavior are gathered. A future addition of stereo imaging of the censused eels is expected to allow for the identification of life stages (yellow vs. silver eel). The applied method is non-invasive, time and cost effective, combines eel census with simultaneous monitoring of the eels’ benthic environment, can collect data from large areas at various scales, and allows for frequent revisits of monitored locations with a high level of spatial precision. As a future EU member Croatia soon needs to report coastal eel stock assessments and a spatial expansion of the ongoing pilot monitoring scheme with added non-sacrificial life stage recognition may allow for the required estimates of eel escapement to the sea.

A fundamental paradigm of community ecology of the shallow infralittoral is that the most important factor limiting animal populations is the shelter from predators provided by three dimensional structure. However, that structure can also provide ambushing cover for resident predators. Thus the value and role of structured habitat could vary depending on the density of predators and their mode of predation. Our objectives were to assess the direct negative effect of predation by mesopredators on the regulation of prey fish populations, the direct negative effect of top predators on mesopredators, and the indirect positive effect of top predators on prey populations. We used methods of experimental fishing (with hook, bait, and long lines) and fish traps to estimate the relative population abundance of prey, meso- and top predators inside and outside of marine protected areas on the Croatian coast, at several benthic habitats including Posidonia oceanica, unconsolidated sediment, rock, and macroalgae. We found that populations of large predators and small prey species were larger within marine protected areas than outside them, and prey species occurred inside and outside P. oceanica meadows. Outside marine protected areas, top predator populations were very small, and mesopredator populations were larger, showing “mesopredator release”. Outside marine protected areas, small prey species were caught mostly outside of seagrass meadows or at their margins, perhaps due to avoidance of the resident mesopredators using seagrass as ambushing cover. We conclude that top-down effects are important in structuring fish communities in the shallow infralittoral, and the importance of three dimensional structure may vary with the abundance of predator populations and the local fishing pressure.

Meadows of the Mediterranean seagrass Posidonia oceanica have been shown to provide habitat for a high diversity and abundance of adult and juvenile fish species. Its primary habitat value for mobile macrofauna is hypothesized to result from its three-dimensional structure, which provides cover from visual chase predators. However, rocky-algal habitat is also common in the Mediterranean, and also provides abundant three-dimensional structure. The influence of the spatial interaction between different kinds of three-dimensional structure on fish communities in the Mediterranean is unknown but could be substantial. Our main objectives are 1) to evaluate the relative influence of P. oceanica and neighboring rocky-algal habitat on fish community parameters within both habitats and within intervening bare sediment, and 2) to evaluate the effect of spatial scale on the interaction. The method used for community sampling is lure-assisted visual census along belt SCUBA transects. This method has been shown to be more effective in censusing small, cryptic, hidden, ambushing, and fast-swimming species than traditional visual census, and is completely non-destructive. Transects are oriented perpendicular to shore and are designed to census both shallow and deep margins of P. oceanica meadows, with varying extents of nearby bare sediment and rocky-algal habitat. The transects are followed both with traditional visual census and with lure-assistance. We found that fish were more abundant and diverse along the edges between P. oceanica and bare sediment than in the interior of the bed, that total abundance and diversity was greatest in rocky-algal habitat, and that the three habitats contained significantly different fish communities. We conclude that no single habitat is most valuable to fish species at our study sites, and the value of P.oceanica is relative and depends on the proximity of other structured and non-structured habitats.

Benthic communities are dynamic systems that respond constantly due to natural and anthropogenic impacts. Permanent monitoring programs are necessary for effective management of protected and threatened marine areas. The aim of this study was to use the original method of underwater geo-referenced videography to record the baseline condition of the benthic habitats in the shallow infralittoral of Kornati National Park, with special focus on Posidonia oceanica seagrass meadows. The substrate type, depth, density and taxa of aquatic vegetation and presence of ecological elements was recorded for every second of the video. We recorded 52 geo-referenced video-transects, with the total length of 28.4 km, of which Posidonia covered 14.8 km. The median depth of the highest density Posidonia was 12 m, density decreased with increasing depth, and we saw no Posidonia deeper than 32 m. The density of algal cover also decreased with depth, while the settlements of sparse Cymodocea and loose algae occurred at lower depths than where algae or Posidonia dominated. The median depth of the highest density algae and sparse Cymodocea settlements was about 9 m. The median depth of pen shell (Pinna nobilis) was 11 m. Our results confirmed that underwater geo-referenced videography can provide high-quality data (spatial and visual), which can be quickly collected and quantitatively analyzed, and these data can be collected entirely non-destructively. Since the data are geo-referenced it is possible to return to the same location at any time and re-record the same transect, and quantify any change in habitat with known statistical precision. Its low cost, rapid and non-destructive execution and statistical rigor makes this method highly suitable for monitoring benthic habitats.

Functional diversity, the range and variety of functional traits possessed by species in a community, may be positively associated with numerous ecosystem processes ; however, debate continues over the relative information value of taxonomic versus functional trait identity in predicting and explaining these processes, especially in highly diverse communities such as seagrass meadows, the most biodiverse shallow marine ecosystem in Europe. We monitored abundance of juvenile and adult benthic and demersal macrofauna over a 12-month period, day and night, across 40 km of transects over 10 sampling stations within a shallow mosaic of seagrass, bare sediment, rock, and macroalgal habitas in the Novigrad Sea, Croatia, using DGPS/videographic SCUBA visual census. We quantified over 90 functional traits, each defined as a separate node in a binary tree, in observed macrofauna. We found that the most powerful predictor of overall spatiotemporal abundance variation across all habitats, stations, seasons, and diel was the taxonomic genus, which explained 66% of all absolute or relative variation. In contrast, each functional node explained about 3% of the overall abundance variance ; thus, a minimum of 20 nodes was necessary to provide the same predictive power as taxonomic genus. Genus explained 64%, 85%, 76%, and 68% of relative seasonal, diel, habitat, and station abundance of fish respectively (significantly for diel and habitat), while the most informative functional dimension, reproductive mode, explained 16%, 21%, 27%, and 12% of these abundance components (significantly for habitat only). We conclude that taxonomic identity contains unique ecological information not contained in popular categories of functional identity.

Seagrass meadows are often considered the most important animal habitat of the shallow marine benthos. While Croatia has the highest linear extent of shallow benthic habitats per hectare of marine territory of any Mediterranean country, with unusually extensive beds of the seagrass Posidonia oceanica, there is yet no published study of the macrofaunal communities of Posidonia and neighboring habitats in the Croatian Adriatic. Our objective was to document the habitat value of all shallow benthic habitats in Croatia to adult and juvenile fish species. We used lure-assisted visual census in 3650 10-m belt transects at 74 locations spanning the entire Croatian coast from spring to fall 2009. More than 47000 fish of 72 species were observed at six basic habitats – rock, algae, sand, Posidonia oceanica, Cymodocea nodosa, and Zostera noltii – which also occurred in 19 distinct combinations. Highest adult fish abundance and diversity were observed on rock and algae and lowest within Cymodocea, with intermediate values within Posidonia. Mixtures of >2 habitats at the scale of meters showed a higher abundance and diversity of adult fish than

A lure assisted visual census method has been developed to identify habitat related differences in fish abundance, diversity and assemblage structure across twelve study sites on Ugljan Island located in the central Croatian Adriatic. Approximately 150 short visual census transects (each 3 m long) were performed per site in fall 2009 and spring 2010. Almost all observed fish were attracted by and interacted with the lure, which substantially increased their chances of being censused. This novel method is especially advantageous for species that are difficult to detect and identify with traditional visual census methods, including hidden, cryptic, or fast-swimming species. Fish were most abundant and diverse in rocky habitats and least abundant in Cymodocea nodosa meadows. Rocky habitat patches with a minor component of unconsolidated sediment and/or vegetation harbored a more diverse and abundant fish assemblage than patches of rock only, and the same applied to habitat patches of unconsolidated sediments as the dominant component. A cluster analysis of the various habitat types and their combinations based on species’ relative abundances resulted in four main clusters: rock dominated habitats, combinations of bare sediment and Cymodocea nodosa, algae dominated habitats, and the seagrass Posidonia oceanica. The lure assisted visual census method is applicable in shallow (snorkel or SCUBA) and deep waters (SCUBA) and allows for the recording of demographic fish descriptors, like size (age) and sex, as well as of fish-lure interaction behaviors (e.g. active vs. passive lure encounter and frequency of lure ingestion) that can be used to identify and quantify the frequencies of various predation modes as well as to compare the predation risks posed by each mode across habitats and study sites.

The macrofauna of the shallow benthos (0-6m) in Novigrad Sea, Central Croatian Adriatic, was censused in three seasons and over five habitat types at ten sites in 2007/8. The relative abundances of all animal taxa identified within the resulting fifty habitat x site combinations were subjected to a cluster analysis. There was a strong tendency for habitats to cluster together indicating that unique associations of fauna with their preferred habitat exist and that they are spatially stable. Sub-clustering within four of the main habitat types coincided with site variability in faunal richness and diversity. Regardless of clustering, habitat diversity explained 29% of faunal diversity across rock sites and 39% across sparse seagrass sites, while faunal abundance explained 46% of faunal diversity in rock habitats (positive correlation) and 32% at unconsolidated sediments (negative correlation)

Within the scope of a recently started monitoring, the spatial and temporal distribution and activity patterns of the European eel, Anguilla anguilla have been investigated in the shallow water benthos (

Shallow and enclosed bays are prone to experience sedimentation resulting in decreases in slope and overall depth. Seagrass beds are known to foster sedimentation which then either (i) benefits seagrass growth, as the bed expands onto new raised and sufficiently illuminated ground or (ii) hampers it, as fast sedimentation may burry newly growing shoots or interfere with seed germination. The shallow benthos of the westernmost Novigrad Sea, an enclosed bay in the Central Croatian Adriatic, has been monitored by means of videographed, DGPS and depth tracked transects positioned perpendicular to the shore and spanning depths from 0 to maximal 6 m, which is larger then 5.2 m, the maximum depth limit for seagrass in the area. Transects were repeatedly followed and videographed with sub-meter accuracy and subsequent video analyses identified each second (approximately 0.3 transect meters) as either seagrass, rock, algae, or unconsolidated sediment. From June 2007 to June 2008 almost all transects decreased in benthic slope and simultaneously experienced an overall gain in seagrass coverage averaging 40 linear transect meters and equaling an average yearly gain of 22% per transect. We identified most of the seagrass gain as a transition from unconsolidated sediment at various depths but mostly at the deep ends of transects where new shallow ground allowed for additional seagrass growth. Future research will include the continuation of the video monitoring as well as methods to estimate sedimentation rates, and experiments to identify interactions between seagrass growth and sedimentation events.

Fish and invertebrate assemblages associated with seagrass meadows (Zostera marina, Nanoostera noltii, Cymodocea nodosa), rocky-algal communities (Cystoseira), and sedimentary bottoms of various particle sizes (mud, sand, gravel) were investigated in two locations within the Novigrad Sea, a relatively shallow enclosure of the central Croatian Adriatic. Whereas one location is impacted by various anthropogenic disturbances associated with residential housing and tourism, heavy small boat traffic, and various marina and beach constructions, the other location is largely unaffected by anthropogenic disturbances. Utilized was a novel visual census method, which minimizes the well-known problems associated with strip transects and maximizes statistical power per unit effort. Long (ca. 200 m) transects of variable-direction, with video assisted habitat-tracking, DGPS tracking, and depth sounding, allowed for the recording of fish and invertebrate observations simultaneous with various habitat characteristics, as well as geographical location and water depth. Significant positive and negative associations of fish and invertebrate species with specific habitat types were found. Rocky-algal communities were found to have higher fish diversity than seagrasses and unvegetated habitats and all habitats exhibited diurnal variability. Furthermore there were significant differences in macrophyte coverage and faunal diversity comparing the two study locations. The presented study uses a novel, non-destructive, effort and cost efficient method that is highly suitable to elucidate the impact of anthropogenic activities on local abundance and diversity of fish and invertebrates. Therefore it provides a good basis for managing the benthos of targeted populations of fish and invertebrates.

Evidence for competition between two co-occurring seagrass species, Zostera noltii and Z. marina, was found along an isobath transect in Novigrad Sea, Central Croatian Adriatic. The transect included developed (marina) and undeveloped regions of shoreline. Each species’ coverage and presence was evaluated using DGPS-tracked underwater videography. Seagrass shoot density was estimated from SCUBA samples. The two Zostera species exhibited opposite significant changes correlated with the environmental gradient: Z. marina cover and shoot density decreased while Z. noltii cover and shoot density increased with distance from marina. Significantly more of the spatial variation was explained by species interaction than by environmental variables. We conclude that competition is a major process structuring the distribution of Zostera species in the Novigrad Sea.

A lure assisted visual census method has been developed to identify habitat related differences in fish abundance, diversity and assemblage structure across twelve study sites on Ugljan Island located in the central Croatian Adriatic. Approximately 150 short visual census transects (each 3 m long) were performed per site in fall 2009 and spring 2010. Almost all observed fish were attracted by and interacted with the lure, which substantially increased their chances of being censused. This novel method is especially advantageous for species that are difficult to detect and identify with traditional visual census methods, including hidden, cryptic, or fast-swimming species. Fish were most abundant and diverse in rocky habitats and least abundant in Cymodocea nodosa meadows. Rocky habitat patches with a minor component of unconsolidated sediment and/or vegetation harbored a more diverse and abundant fish assemblage than patches of rock only, and the same applied to habitat patches of unconsolidated sediments as the dominant component. A cluster analysis of the various habitat types and their combinations based on species’ relative abundances resulted in four main clusters: rock dominated habitats, combinations of bare sediment and Cymodocea nodosa, algae dominated habitats, and the seagrass Posidonia oceanica. The lure assisted visual census method is applicable in shallow (snorkel or SCUBA) and deep waters (SCUBA) and allows for the recording of demographic fish descriptors, like size (age) and sex, as well as of fish-lure interaction behaviors (e.g. active vs. passive lure encounter and frequency of lure ingestion) that can be used to identify and quantify the frequencies of various predation modes as well as to compare the predation risks posed by each mode across habitats and study sites.