Original scientific article	UDC 574.1/.2(262.3)
Received: 2010-04-02
RELATIONS OF MACROFAUNAL DIVERSITY WITH HABITAT DIVERSITY IN A CENTRAL CROATIAN ADRIATIC LAGOON
Claudia KRUSCHEL & Stewart T. SCHULTZ University of Zadar, Department of Maritime Sciences, HR-23000 Zadar, M. Pavlinovica bb, Croatia E-mail: claudia@claudiakruschel.com
ABSTRACT
The macrofauna of the shallow benthos (0-6 m) in the Novigrad Sea, the 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 habitats 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).
Key words: habitat diversity, faunal diversity, visual census, GPS, videography, Adriatic Sea
RELAZIONE FRA DIVERSITA MACROFAUNISTICA E DIVERSITA Dl HABITAT IN UNA LAGUNA CROATA NELL'ADRIATICO CENTRALE
SINTESI
La macrofauna bentonica di acque superficiali (0-6 m) nel mare antistante Novigrad, in Adriatico centrale, e stata campionata durante tre stagioni in cinque tipi di habitat e 10 siti di campionamento, fra il 2007 ed il 2008. Le ab-bondanze relative di tutti i taxa animali identificati all'interno delle 50 combinazioni habitat x sito, sono state sotto-poste alla cluster analysis. I risultati evidenziano una forte tendenza degli habitat a raggrupparsi, il che indica una marcata relazione fra la fauna e l'habitat preferenziale, nonche la stabilita su scala spaziale. I sotto-ragruppamenti all'interno di quattro habitat principali coincidono con la variabilita dei siti in termini di ricchezza faunistica e diversita. A prescindere dalla cluster analysis, la diversita di habitat spiega il 29% della diversita faunistica in siti di fondo duro, e il 39% in siti con fanerogame marine sparse. L'abbondanza faunistica invece spiega il 46% della diversita faunistica in habitat rocciosi (correlazione positiva) ed il 32% su sedimenti non-consolidati (correlazione negativa).
Parole chiave: diversita di habitat, diversita faunistica, censimenti visivi, GPS, videografia, mare Adriatico
INTRODUCTION
Biodiversity is rapidly being lost in a world increasingly transformed by human activities and there is a need to understand the factors mediating species coexistence and local biodiversity. Shallow enclosed lagoons simultaneously provide nursery habitat for fish and invertebrates, but are especially vulnerable because they have limited exchange of water and fauna with the open sea and already experience natural stresses such as large annual salinity variations, which shape the unique fauna and habitats. Understanding the drivers of diversity permits predictions regarding the influence of anthropogenic activities. Environments with steep gradients in habitat characteristics allow for spatial niche segregation among species and species sorting by habitat (Mouillot, 2007). In such a situation faunal diversity depends primarily on the availability of niches, those directly related to habitat as well as resource niches indirectly related to habitat, such as trophic niches (Wilson, 1999). The total number of niches then is determined by habitat diversity, whose components are habitat heterogeneity and habitat complexity (Ziv, 1998; Guidetti & Bussotti, 2002). Habitat heterogeneity refers to the relative abundance of the various structural components within a habitat, and their variability, while complexity deals with the absolute abundance of the various structural components.
For example, the presence of seagrass at a location often has a positive effect on faunal diversity due both to its structural complexity (Hori ef a/., 2009) and its heterogeneity, which includes spatial variability in seagrass species, shoot density, and shoot morphology (Bell & Westoby, 1986a, b). Likewise, rock habitats are known to have high faunal diversity of cryptobenthic species such as Blennidae and Gobiidae, which increases with increasing complexity and heterogeneity (Macpherson & Zika, 1999; Gratwicke & Spreight, 2005; La Mesa ef a/., 2006; Lingo & Szedlmayer, 2006; Orlando-Bonaca & Lipej, 2007). Even sediments, lacking three dimensional complexity on a scale of macrofaunal body size, are nevertheless reported to support a more diverse macro-fauna when heterogeneous in particle size (Hilbig & Blake, 2000; Jayaraj ef a/., 2008). This is because particle size influences physical variables, including water and organic matter content (Kruschel ef a/., 2009), oxygen concentration (Rosenberg ef a/., 2002), and sediment resuspension rates (Rutzler ef a/., 2000), which all influence the habitat choice of infauna and surface organisms (Alongi & Christofferson, 1992). Generally, increased habitat diversity (complexity and heterogeneity) is expected to increase the number of food resource niches, e.g. epifauna and epiphytes on seagrass (Sirota & Hovel, 2006) and rock (Consoli ef a/v 2008) or infauna (Sciberras ef a/., 2009) in heterogeneous sediments, thereby increasing faunal diversity through reduced
competition when diet overlap is low (Moreno-Rueda ef a/., 2009).
The shallow benthos of the Novigrad Sea is spatially heterogeneous. Fundamentally different habitats with three-dimensional structure (macroalgae, rock, seagrass) and unconsolidated sediments occur in a patchy mosaic at the scale of a few meters. If species sorting and niche availability were in fact the main drivers of faunal assembly and diversity in the Novigrad Sea lagoon, we would predict to observe (1) habitat specific faunal assemblages that are spatially stable (across sites), and that (2) site variability in faunal diversity for each of the basic habitat associated assemblages can be explained by the site variability in habitat diversity.
MATERIAL AND METHODS Geographical location and sample locations
The Novigrad Sea, Croatia (44"12'N, 1530'E) is a protected lagoon of 29 km2 (approximately 8.5 km long) in the eastern Adriatic Sea. It is connected to the open Adriatic by the Velebit Channel via a narrow strait, the Maslenica Channel (Fig. 1). The study area is influenced by freshwater inflow from the Zrmanja River, a few small seasonal creeks, and a canal in the far western corner (Sinovcic ef a/., 2004; Matic-Skoko ef a/., 2007).
Fig. 1: Approximate location of the ten study sites (A-J) in the Novigrad Sea.
Sl. 1: Približne lokacije desetih obravnavanih območij (A-J) v Novigrajskem morju.
The total water volume is approximately 0.5 km3 and reaches maximum depth of 35 m near the channel mouth. Tidal differences are 30 cm or less. Wave heights up to about 1 m occur under strong winds. The shallow benthic habitats (less than about 6 meters depth) include dense macroalgae/rock, unconsolidated sediments, and sparse to dense seagrass on mud, sand, or gravel. The seagrass is most expansive in the low gradient western portion near Posedarje (Schultz et al., 2009). It was this area chosen for the study due to the presence of equal representation of each of the major benthic habitats of shallow protected waters of the Adriatic (seagrass, rock from pebble to bedrock, macroalgae, unconsolidated sand and gravel), all intermixed and neighboring each other on a small spatial scale along a shoreline several kilometers in length (Bakran-Petricioli et al., 2006, Ja-nekovic et al., 2006).
Ground cover/habitat distribution
Along visual census SCUBA transects a total of 5930 m of macroalgae (A), 5500 m of bare rock (R), 9300 m of dense seagrass (SD), 9400 m of sparse seagrass (SS),
and 8400 m of bare, unconsolidated mud, sand or gravel (U) were encountered, with each category scored at a resolution of about 0.3 m. All seagrass vegetation occupied mud, sand or gravel. Algae (Cystoseira and a few unidentified brown and green algae), were attached to rocks of all sizes. For a detailed description of the structural heterogeneity within the five basic habitat categories see Table 1. While all five ground cover types were observed at all sampled depths, algae and rock dominated the shallowest sampling zone (0-2 m), bare unconsolidated ground the intermediate zone (1.5-3.5 m), and seagrass the deepest zone (2.5-5 m) (Fig. 2). Three parallel to the shore transects, one in each sampling zone, where investigated at ten study sites (A-J) that varied in the observed proportions of the five basic ground covers averaged across all three zones (Fig. 3).
Visual census transects
The abundance of macrofauna was censused along 2-m wide SCUBA belt transects (Nagelkerken et al., 2000; Horinouchi et al., 2005). Visual census allows quantification of habitat occupation by each individual
seagrass (sparse, SS and dense, SD with matching			round coverage proportions: 0 - 1)
N Zostera noltii Z Z. marina C Cymodocea nodosa combinations, sequence = dominance ZZ = either NZ or ZN (see Table 2)			1	= SS 0 - 0.25 2	= SS 0.25 - 0.5 3	= SD 0.5 - <1 4	= SD 1
unconsolidated sediments			algae
U D H	1	= mud (< 50 'm) 2	= sand (50 'm - 0.2 mm) 3	= sandgravel (0.2 - 2mm) 4	= gravel (2 mm - 2cm) 5	= sand within rocks 6	= gravel within rocks		A1 Cysoseira spp Aoth other algae
	sediment covered with dead seagrass human wastes		
bare rocks			
R bare rock RT rock covered in algal turf MR rock covered in Mythilus		1	= pebble (2cm - fist) 2	= cobble (fist - head) 3	= rock (head - body) 4	= boulder (> body) 5	= bedrock	
Tab. 1: Key to the identification of microhabitats within the basic habitats. Seagrass is identified by species and visual density (ground coverage proportion), unconsolidated sediments by grain size, algae by species and rock by size and cover. The sediment grain size and rock size categorization is a modified version of Larsonneur (1977) (in UNEP, 1998).
Tab. 1: Ključ za identifikacijo mikrohabitatov znotraj osnovnih habitatov. Habitat morske trave je določen z vrstami in vizualno ocenjeno gostoto (delež pokrovnosti tal), neutrjeni sedimenti z velikostjo delcev, alge z vrstami in skale z velikostjo in poraščenostjo. Kategorizacija velikosti delcev pri sedimentih in velikosti skal po prirejeni Larsonneur verziji (1977) (v UNEP, 1998).
Fig. 2: Relative abundance of the five basic habitats (U: unconsolidated sediments, SS: sparse seagrass, SD: dense seagrass, R: bare rocks, A: algae) across depth related zones (A/R: algae and rocks, U: unconsolidated sediments, SG: seagrass) in which macrofauna was visually censused along parallel transects (one per zone). Sl. 2: Relativna številčnost petih osnovnih habitatov (U: neutrjeni sedimenti, SS: redka morska trava, SD: gosta morska trava, R: neporaščene skale, A: alge) po globinskih območjih (A/R: alge in skale, U: neutrjeni sedimenti, SG: morska trava), v katerih je bila makrofavna vizualno popisana vzdolž paralelnih transektov (eden na območje).
animal, at all substrates and on a small scale (body size; Lipej & Orlando-Bonaca, 2006). This is not possible for other popular methods such as trawl or seine net sampling (Gray et a/., 1998; Hindell & Jenkins, 2005; Hori-nouchi et a/., 2005). Macrofauna observed comprised all fish species and all invertebrates large enough to be detectable by naked eye and without physically disturbing the habitat. All individuals of all taxa were recorded but for the presented analyses, observational units called "groups" were used, defined as spatial clusters of 1 or more individuals of the same species observed at the same moment. For statistical tests, the appropriate unit of replication is probably the group, since individuals within a group are not independent observations, and ninety-eight percent of all groups observed comprised 10 or less individuals. Transects were followed at a speed of approximately 0.3 m/sec during the day and 0.15 m/sec during the night. Transect length ranged from 160 m to 360 m.
Fig. 3: Relative abundance of the five basic habitats (U: unconsolidated sediments, SS: sparse seagrass, SD: dense seagrass, R: bare rocks, A: algae) along long transects within sites (A-J).
Sl. 3: Relativna številčnost petih osnovnih habitatov (U: neutrjeni sedimenti, SS: redka morska trava, SD: gosta morska trava, R: neporaščene skale, A: alge) vzdolž dolgih transektov znotraj območij (A-J).
DG PS/videography
During the daytime sampling, the diver carried a video sensor (Sony, 480 color TVL) that continuously recorded the sea bottom. Simultaneously overlaid on the video image was satellite time recorded every two seconds. Depth was recorded by a 200 KHz, 11.208, single-beam transducer. Horizontal DGPS coordinates (2006 Trimble Pro XRS) were taken with real-time submeter accuracy from radio beacon transmissions to a GPS antenna held by kayak operator directly above the video sensor visible from the surface (Norris et a/., 1997; Dauwalter et a/., 2006; Schultz, 2008). The night dives followed the same transects as the day dives usually within 24 hours, by submeter real-time navigation over the DGPS tracks of that day's dive transects. This had the advantage of not using underwater survey markers or tape measures which have been observed to attract fish. The proportion of each microhabitat, defined as subhabitats within the five basic habitats: rock (R), algae (A), dense seagrass (SD), sparse seagrass (SS), and uncon-solidated sediments (U), occupying the transect, was
measured by analysis of the video (for microhabitat identification key see Table 1). Each second of the video was identified as representing one of these microhabitats, and then the linear extent of transect occupied by each of the microhabitats was calculated from the associated DGPS positions (Schultz, 2008). One second of video represents approximately 0.3 m of transects.
Calculation of fauna and habitat related variables
Total abundance of a faunal species was defined as the number of observational units or groups encountered per meter of visual census transect. Relative abundance of a species within a site/habitat combination was equal to the absolute abundance in that combination divided by the total abundance of all species within that combination. Faunal diversity was calculated as the Simpson's reciprocal index, equal to the multiplicative inverse of the probability that two random groups of fauna are the same species. This we refer to as the "effective" number of species or the number of species in the site/habitat if all species were equally abundant, given the observed probability that two random groups are the same species. Habitat diversity - again as Simpson's reciprocal index, in this case the reciprocal of the probability that two random seconds of transect video showed the same bottom microhabitat. This is referred to in the text as the "expected" number of microhabitat types for a site/ habitat combination. Total microhabitat abundance was equal to the number of video seconds showing a particular microhabitat type within a transect. Relative microhabitat abundance within a given site/habitat was equal to its absolute abundance divided by the total number of seconds in the transect video.
Data analysis
All confirmatory statistical tests were conducted within general linear models, including unequal variance t-tests, single-factor ANOVA and linear regression, after checking for normality and homoscedasticity.
RESULTS
As reported in a previous publication (Schultz et al., 2009), over a total dive transect length of 39 km accomplished in 60 hours of diving, 61,713 animals have been observed: 54,043 fish belonging to 39 taxa and 7670 invertebrates belonging to 46 taxa. Individuals have been recorded in 10,359 observational fish groups and 4,309 observational groups of invertebrates. The total faunal density was 1.6 animals per transect meter (1.4 fish and 0.20 invertebrates), or 0.38 observational units per transect meter (0.27 fish and 0.11 invertebrate).
Habitat specific faunal assemblages
The cluster analyses of the site specific relative abundances of faunal taxa observed in each of the site/habitats revealed seven main significant clusters (Fig. 4), with each cluster representing one of the five basic habitats, including: sparse seagrass (SS, cluster 5), dense seagrass (SD, clusters 1 and 6), rock (R, clusters 2 and 3), unconsolidated sediments (U, cluster 4), and algae (A, cluster 7). For each of the seven main clusters Table 2 gives information on total faunal and microhabitat richness and taxon dominance and indicates in which of the five basic habitats (A, R, SS, SD, U) taxa reach their highest relative abundance. The number of dominant taxa (comprising 90% of the total faunal abundance) within each main cluster positively correlated with sub-cluster habitat richness (Tab. 2). Faunal richness and habitat richness was highest at bare sediment (cluster 4), in sparse seagrass (cluster 5), and in dense seagrass (cluster 6). Taxa comprising 90% of the total faunal abundance in the sparse seagrass (cluster 5) and in the rock habitat (cluster 3) were most even in their relative abundances, whereas the faunal assemblage at bare sediments was dominated by Gobius niger and gobiid juveniles, comprising more than 50% of the total faunal abundance (Tab. 2).
Rock habitat (clusters 2 and 3)
Gobiids (Gobius cobitis (Goco), G. paganellus (Gopa), G. cruentatus (Gocr), and G. bucchichii (Gobu) reached their highest relative abundance in the rock habitat and several blennids (Parablennius sanguino-lentus, P. incognitos, P. gattorugine (Paga), Lipophrys dalmatinus, and L. pavo), and Serranus scriba was found there exclusively. Invertebrates found solely on rocks were the squad lobsters of the genus Galathea, the crab Eriphia verrucosa, the urchin Paracentrotus lividus, and oysters (Ostreidae). The shrimp Palaemon elegans (Pael) had its highest relative abundance here, while labrids (Symphodus ocellatus (Syoc) and S. roissali (Syroi)) were very common, yet had an even higher relative abundance in algal vegetation.
Algae habitat (cluster 7)
The algal vegetation assemblage is characterized by two species of labrids S. ocellatus (Syoc) and S. roissali (Syroi), and the shrimp P. elegans (Pael) all of which also associate with rocks. Algal assemblages of different locations vary primarily in the relative abundances of these labrids and shrimp: as labrids increase the shrimp decreases and vice versa. Like the two labrids, Atherina spp. (Ath) reaches its highest abundance in the algal habitat where at night individuals or small groups hover in canopy gaps within stands of large Cystoseira individuals.
Tab. 2: Comparison of the main site/habitat clusters (Fig. 4) by relative abundance of common taxa (comprising 90% of the total abundance) and total faunal richness (above) and by relative abundance of microhabitats and microhabitat richness. For microhabitat identification refers to Table 1. Highlighted abbreviations for common animal taxa (see text for full species names) indicate in which habitat taxa reached their peak relative abundance. Tab. 2: Primerjava glavnih skupin območje/habitat (Sl. 4) z relativno številčnostjo glavnih vrst (90% celotne številčnosti) in celotnega bogastva favne (zgoraj) ter z relativno številčnostjo mikrohabitatov in bogastva mikrohabita-tov. Za identifikacijo mikrohabitatov glej Tabelo 1. Poudarjene okrajšave za glavne živalske vrste (glej besedilo za polno ime vrste) označujejo habitat, v katerem je vrsta dosegla največjo relativno številčnost.
SD (dense	seagrass)				R (rocks)			U (sediments)		SS (sp	arse	A (algae)	
										seagrass)			
Cluster 1	prop.	Cluster 6	prop.	Cluster 2	prop.	Cluster 3	prop.	Cluster 4	prop.	Cluster 5	prop.	Cluster 7	prop.
relative faunal abundances													
Psmi	0.529	Ath	0.240	Pael	0.341	Goni	0.120	Goni	0.337	Goni	0.190	Ath	0.360
Zoop	0.090	Psmi	0.171	Goni	0.138	Gobu	0.119	Gojuv	0.166	Zoop	0.114	Syoc	0.198
Ath	0.072	Zoop	0.131	Gojuv	0.114	Coco	0.113	Ath	0.140	Ath	0.111	Syroi	0.162
Pael	0.050	Pael	0.085	Ath	0.071	Ath	0.112	Pagu rus	0.063	Brachyura	0.106	Pael	0.147
Ansu	0.045	Ansu	0.062	Brachyura	0.066	Pael	0.111	Pael	0.045	Pagurus	0.100	Brachyura	0.036
Hetr	0.038	Brachyura	0.059	Zoop	0.066	Syroi	0.090	Brachyura	0.044	Gojuv	0.068		
Brachyura	0.034	Goni	0.053	Syoc	0.048	Gocr	0.089	juv2	0.022	Pael	0.063		
Pagurus	0.032	Pagurus	0.035	Paga	0.048	Syoc	0.073	Gobu	0.021	Hetr	0.044		
		Anan	0.028			Brachyura	0.052	Syroi	0.017	Ansu	0.024		
		Syoc	0.028			Gopa	0.033	Syoc Goco Hetr Gocr juvl	0.015 0.013 0.010 0.008 0.008	Psmi Syoc Gobu Syroi juv2	0.024 0.016 0.011 0.010 0.009		
+ 19 taxa	0.099	+ 43 taxa	0.108	+ 4 taxa	0.110	+ 28 taxa	0.087	+ 45 taxa	0.091	+ 43 taxa	0.108	+ 28 taxa	0.097
relative habitat abundances													
ZZ3	0.584	ZZ3	0.386	R1	0.440	R2	0.428	U2	0.244	N2	0.353	A1	0.993
N3	0.196	N3	0.245	R2	0.425	RT3	0.221	U3	0.208	ZZ2	0.175	Aoth	0.007
C3	0.168	C2	0.099	R3	0.108	R3	0.165	D	0.111	N1	0.137		
ZZ4	0.052	ZZ4	0.070	RT2	0.022	R1	0.076	U5	0.107	R2	0.104		
		N2	0.048	RT3	0.006	RT4	0.054	U4	0.093	C2	0.078		
		N4	0.042			RT2	0.045	R2	0.063	R1	0.067		
		C3	0.034			R4	0.005	R1	0.060	Z2	0.022		
		ZZ2	0.023			MR4	0.003	U6	0.050	ZZ1	0.018		
		Z2	0.022			MR3	0.001	R3	0.020	Z1	0.011		
		N1	0.016			R5	0.001	RT3	0.020	RT2	0.009		
		Z1	0.008					RT2	0.006	RT3	0.006		
		C1	0.006					RT5	0.005	RT1	0.006		
		ZZ1	0.001										
								+ 6	0.013	+ 3			
								habitats		habitats	0.010		
Unconsolidated sediment habitat (cluster 4)
Common inhabitants on bare sand/gravel near the rocks in shallow water are the goby Gobius niger (Goni) and hermit crabs of the genus Pagurus. Common inhabitants on bare sand/gravel near the lower limits of the Cymodocea nodosa beds in deeper water include the bivalve Pecten jacobeus, the gastropod Hexaplex trun-culus (Hetr), and the fish Symphodus cinereus. Taxa common in more expansive areas of bare sand/gravel include small gobiids, mainly Pomatoschistus spp., ju-
veniles of G. niger (Gojuv), and bentho-pelagic juvenile fish with schooling behavior (juvl and juv2).
Sparse seagrass (cluster 5)
Species frequently observed in sparse seagrass are associated either with shallower Zostera beds (Zosteris-essor ophiocephalus (Zoop) and Anemonia viridis (Anvi)) or deeper Cymodocea beds and the adjacent deep unvegetated sediments (H. trunculus (Hetr), P. ja-cobaeus, S. cinereus, Spicara smaris).
Fig. 4: Dendrogram of site/habitat combinations based on taxa abundance per linear transect meter relative to the total within each of the five substrates, all sampling times (diel, seasonal) and all sites (A-f) pooled. Significant clusters (multiscale bootstrap resampling probability >0.95) are enclosed in rectangles. Main subclusters representing one of the five basic habitats (see Fig. 3) are numbered. Primary subclusters within main habitat clusters are indicated with the letters "a" and "b" in addition to the cluster number. Total faunal abundance (1st value) and diversity indices (probability, 2nd value, and effective number of taxa, 3rd value) were displayed along with the name of each of the fifty site/habitat combinations.
Sl. 4: Dendrogram kombinacij območje/habitat na podlagi številčnosti taksonov na meter linearnega transekta glede na celoto znotraj vsakega od petih substratov, vseh časov vzorčenja (dnevnih, sezonskih) in vseh območij vzorčenja (A-f). Statistično značilne skupine (multiscale bootstrap resampling verjetnost >0,95) so označene z okviri. Glavne podskupine, ki predstavljajo enega od petih glavnih habitatov (glej Sl. 3), so oštevilčene. Glavne podskupine znotraj glavnih habitatnih skupin so označene s črkami "a" in "b" poleg številke skupine. Celotna številčnost (1. vrednost) in diverzitetni indeks (verjetnost, 2. vrednost, in dejansko število taksononv, 3. vrednost) so pripisani imenu vsake od petdesetih kombinacij območje/habitat.
Dense seagrass (clusters 1 and 6)
The dense seagrass assemblage is characterized by the highest relative abundances of two invertebrates, the snakelock anemone A. viridis (Anvi) and the green urchin Psammechinus microtuberculatus (Psmi), although the abundance of the urchin is seasonal (highest in winter) and patchy (see Fig. 3, cluster 1 is separated from all others on the basis of high relative "Psmi" abundance). Likewise A. viridis (Anvi) have been observed to increase temporarily in spring when they cover seagrass blades in dense patches. Fish taxa reaching their peak abundance in dense seagrass are two seagrass resident predators, the European eel, Anguilla anguilla (Anan), and the grass goby, Z. ophiocephalus (Zoop). Although the labrid S. cinereus prefers deeper sparse Cymodocea and adjacent sediments for foraging, nests of this species
(identified by the presence of a territorial male) are associated with dense seagrass. While Atherina spp. (Ath) frequently and quickly passes in schools over all habitats during the day with little preference, it noticeably seeks out dense seagrass at night, hovering in small groups or individually at a distance above the canopy.
Site variability in habitat specific faunal richness, diversity, and abundance
Rock habitat (clusters 2 and 3)
The two main rock assemblages (clusters 2 and 3) were markedly different in taxon richness (12 vs. 38 taxa) (Tab. 3), yet the likewise large difference in mean diversity (e.g., 5.4 vs. 9 effective taxa) was not quite significant (Tab. 3), while mean faunal abundance was sig-
Tab. 3: Summary of the comparisons of mean site diversity and mean total site abundance within sub-clusters of basic habitats, with t test results given in columns 4-7. In addition sub-cluster richness (column 3), sub-cluster relative proportion of main habitats (column 8), and microhabitat heterogeneity indicators and their relative abundance for each sub-cluster are listed (column 9).
Tab. 3: Povzetek primerjave povprečne diverzitete območja in povprečne številčnosti celotnega območja znotraj podskupin osnovnih habitatov, pri čemer so rezultati t testa zapisani v stolpcih 4-7. Dodane so vrednosti za bogastvo podskupin (3. stolpec), za relativni delež glavnih habitatov po podskupinah (8. stolpec) in indikatorji mikro-habitatne heterogenosti ter njihova relativna številčnost za vsako podskupino (9. stolpec).
habitat	Cluster (richness)	site specific variables	mean	df	t value	P	site habitat proportions	heterogeneity	prop
		Index	3.233333	10	-9.8011	1.910e-06		Z no/t/7	0.613
	1 (27)	probability	0.687	4.146	10.4844	0.0003877	0.05 - 0.15	Z mar/na/C. nodosa	0.386
dense		abundance	1.193333	2.938	-0.0193	0.9858	mean 0.09	very dense seagrass	0.015
seagrass		Index	6.411111					Z no/t/7	0.856
	6 (53)	probability abundance	0.840 1.201111				0.7 - 0.36 mean 0.22	Z mar/na/C. nodosa very dense seagrass	0.14 0.144
		Index	5.350	1.323	-3.257	0.1389		pebbles/cobbles	0.886
	2 (12)	probability	0.8050	1.065	-2.32	0.2467	0.14 - 0.22	turf covered	0.027
rocky		abundance	0.0815	3.012	-3.1079	0.05269	mean 0.18	boulders	none
groundcover	3 (38)	Index probability abundance	9.025 0.8875 2.0750				0.18 - 0.35 mean 0.23	pebbles/cobbles turf covered boulders	0.550 0.320 0.062
		Index	3.633333	8.933	-6.404	0.0001289		gravel	0.255
	4a (44)	probability	0.72000	4.116	-6.3876	0.002796	0.24 - 0.39	within rocks	0.258
unconsolidated		abundance	2.60000	2.003	4.3565	0.04874	mean 0.31	dead seagrass	0.006
sediments		Index	6.387500					gravel	0.235
	4b (47)	probability abundance	0.83875 0.83875				0.08 - 0.27 mean 0.19	within rocks dead seagrass	0.159 0.201
		Index	9.975000	5.608	3.516	0.01406		Z no/t/7	0.636
	5a (41)	probability	0.9000000	7.911	3.7551	0.005697	0.19 - 0.28	Z mar/na	0.220
sparse		abundance	1.717500	3.523	0.7395	0.5058	mean 0.23	C. nodosa	0.144
seagrass		Index	6.933333					Z no/t/7	0.898
	5b (50)	probability abundance	0.8533333 1.211667				0.18 - 0.33 mean 0.25	Z mar/na C. nodosa	0.042 0.060
		Index	4.283333	6.669	1.2528	0.2524			
	7a (29)	probability	0.7483333	6.79	2.1152	0.07345	0.10 - 0.31	Cystose/ra spp.	0.988
algae		abundance	1.121667	3.849	0.531	0.6245	mean 0.18		
		Index	4.633333						
	7b (30)	probability abundance	0.8066667 1.280000				0.16 - 0.20 mean 0.18	Cystose/ra spp.	1.000
nificantly higher for sites in cluster 3. Habitat composition differed in the two groups of rock sites: cluster 2 sites were dominated by pebbles and cobble sized rocks (89%) and lacked boulders, while sites in cluster 3 were to 45% occupied by rocks larger than pebbles and cobbles, of which 6% were boulders (Tab. 3). Also a higher proportion of rocks in cluster 3 sites were covered in turf algae and 0yf//us than in cluster 2-sites (turf: 32% vs. 3%, 0yf//us: 0.4% vs. none) Considering all rock sites, regardless of their cluster position, 29% of the variation among sites in faunal diversity was explained by rock diversity (Fig. 5, Tab. 4). Rock faunal abundance was positively correlated with faunal diversity and explained 46% of its variance (Tab. 4).
Unconsolidated sediment habitat (sub-clusters 4a and b)
The two main unconsolidated sediment assemblages differed significantly in faunal diversity (Tab. 3), yet they had equal taxon richness (44 and 47). Sites in cluster 4a had a significantly higher mean faunal abundance, primarily due to the high abundance of Cob/us n/ger and juvenile gobiids (1.25 and 0.4 per transect meter), comprising 64% of the total faunal abundance. Sediments in 4a-sites did not differ as much in grain size from sediments in 4b-sites (26% vs. 24% coarse sand and gravel) as in spatial relation to neighboring habitats (rock vs. seagrass) (Tab. 3). More of the unconsolidated sediment in 4a-sites than in 4b-sites comprised small patches
Tab. 4: Site faunal diversity as a function of a site's habitat diversity in the rock and sparse seagrass habitat. For each habitat 12 microhabitat types were defined (for details see Table 1) and quantified in DGPS assisted videos of the benthos. Significance of correlation is indicated by p values which are given along with r2 values. Tab. 4: Favnistična diverziteta območja kot funkcija habitatne diverzitete območja za habitat skal in habitat redke morske trave. Za vsak habitat smo določili 12 mikrohabitatnih tipov (za podrobnosti glej Tabelo 1) in jih količinsko opredelili z DGPS video meritvijo dna. Značilnost korelacije je podana v vrednosti p in r2.
	df	t value	p value	F	r2 value
Response: faunal diversity rocky habitat (R)					
habitat diversity	8	1.395	0.0612	4.736	0.2934
faunal abundance	8	2.924	0.0192	8.549	0.4562
Response: faunal diversity sparse seagrass habitat (SS)					
habitat diversity	8 2.620		0.0307	6.862	0.3944
Response: faunal diversity unconsolidated sediment habitat (U)					
faunal abundance	8	-2.277	0.05232	5.184	0.3174
Rock
Sparse seagrass
p = 0.06, r = 0.29
p = 0.03, r = 0.39
Habitat diversity
Habitat diversity
Fig. 5: Site/habitat taxon diversities (effective number of taxa) for the rock (right) and sparse seagrass (left) habitat as a function of a site's habitat diversity (effective number of microhabitats). Each of the 10 sites is represented by a letter (A-f). For site location see Fig. 1, for microhabitat categorization within basic habitats refers to Table 1. Microhabitats were quantified for each second (0.3 m) in DGPS assisted videos of the benthos. Significance of correlation is indicated by p values given along with r2 values on top of each figure (for details of the ANOVA results see Table 4).
Sl. 5: Vrstna diverziteta za območje/habitat (dejansko število vrst) za habitat skal (desno) in redke morske trave (levo) kot funkcija habitatne diverzitete za posamezno območje (dejansko število mikrohabitatov). Vsako od 10 območij je označeno s črko (A-f). Za lokacijo območja glej Sl. 1, za kategorizacijo mikrohabitata znotraj osnovnih habitatov glej Tabelo 1. Mikrohabitati so bili količinsko opredeljeni za vsako sekundo (0,3 m) v DGPS video meritvi dna. Značilnost korelacije je podana v vrednosti p in r2 na vrhu vsakega grafa (za podrobnosti ANOVA rezultatov glej Tabelo 4).
among rocks (26% vs. 16%) while less sediment in 4a-sites than in 4b-sites was covered in dead seagrass (0.6% vs. 20%) (Tab. 3). Considering all sediment sites, regardless of their cluster location, faunal abundance is negatively correlated with faunal diversity and explains 32% of its variation (Tab. 4).
Sparse seagrass (sub-cluster 5a and b)
The two main sparse seagrass assemblages (sub-clusters 5a and b) significantly differed in faunal diversity (Tab. 3) which coincided with marked differences in the sparse seagrass habitat composition. Sparse seagrass in 5b-sites was to 90% comprised of Zostera noltii, the seagrass occupying the shallower benthos (0-2.5 m), while sites clustered in 5a had significantly higher proportions of other seagrass species than 5b-sites: deep (4.5-6m) sparse Cymodocea (2.5 times higher) and sparse Z. marina (4 times higher) (Tab. 3). Considering all sparse seagrass sites, regardless of their association with cluster, site variability in sparse seagrass habitat diversity was significantly correlated with and explained 29% of the site variability in faunal diversity (Fig. 5, Tab. 4).
Dense seagrass (clusters 1 and 6)
The two main dense seagrass assemblages (clusters 1 and 6) significantly differed in faunal diversity (Tab. 3) and in faunal richness (27 vs. 53), which coincided with marked differences in the dense seagrass site coverage (9% vs. 22%), density, and relative abundance of seagrass species. Sites in cluster 6 were dominated by Z. noltii (86%) and had a higher proportion of very dense seagrass (per transect meter coverage =1) than cluster 1-sites (14% vs. 1.5%) (Tab. 3).
Algae (clusters 7a and b)
The sites of the two main algal clusters did not significantly differ in faunal diversity, faunal abundance, or algal habitat diversity (Tab. 3).
DISCUSSION
The study area indeed comprised habitat specific faunal assemblages (Fig. 4). Each of the five basic habitats, algae (A), rock (R), sparse (SS) and dense (SD) seagrass and unconsolidated sediment (U) were clearly dominated by a different set of taxa (e.g., labrids in the algae, four associated gobies in the rocky habitat, Psammechinus tubercularis (Psmi) and Zosterisessor ophiocephalus (Zoop) in seagrass, and juvenile fish (go-biids and various unidentified) at bare sediments (Tab. 2). Also, all habitats were home for a number of exclusive taxa, e.g., Anguilla anguilla (Anan) in seagrass, various blennids and invertebrate species in the rock habi-
tat, and mollusks such as Pecten jacobeaus and Hexaplex trunculus (Hetr) in sparse Cymodocea nodosa and adjacent sediments. These findings are in agreement with habitat preferences calculated by Schultz et al. (2009) for the Novigrad Sea shallow benthos taxa. The observed habitat specific faunal assemblages were spatially stable; taxa did not exhibit different habitat associations at different sites. However, the diversity of the basic assemblages did vary among sites (Tab. 3), which was correlated with microhabitat diversity which to some extent was correlated with site variability in faunal abundance (positive in the rock habitat, negative at bare sediments). Within the Novigrad Sea rock habitat, rock size variability and the coverage with ratio of bare rocks versus rocks covered with turf and mussels may be the most important microhabitat descriptors in explaining faunal diversity variability. In sparse seagrass the most important factor may be seagrass species diversity, which in the Novigrad Sea is linked to depth. In dense seagrass patches site area coverage and density may be most important, while at bare sediments the influence of neighboring habitats (rock vs. seagrass) rather than grain size coincided with differences in faunal diversity. Faunal diversity has been reported to positively respond to a variety of rock characteristics, including: rugosity (Luck-hurst & Luckhurst, 1978), turf coverage (Feireirra et al., 2001), coverage with habitat engineering sessile animals, e.g. mussles (Feireirra et al., 2001; Buschbaum et al., 2008), vertical relief (Gratwicke & Spreight, 2005), and interstitial space variety (Schmude et al., 1998). For the Mediterranean Sea positive relationships between rock habitat diversity and faunal diversity have been found for blennioids by Macpherson & Zika (1999) and for blennioids and other cryptobenthic fish by Patzner (1999). Blennioid diversity has been investigated in greater detail in shallow Adriatic habitats by Orlando-Bonaca & Lipej (2007) who discovered that among other microhabitat descriptors the presence/absence of turf algae and sessile invertebrates, rock size, and the precise position at boulders, all resulted in blenny assemblage variability. A study including a variety of families of rock resident crypto-benthic fish concluded that habitat choice was almost entirely based on whether rocks were covered in vegetation or bare (La Mesa et al., 2006). Letourneur et al. (2003) found for the north-western Mediterranean that rock cover per se and within the rock habitat variability in rock size were most powerful in explaining fish distribution. Garcia-Charton & Ruzafa (1998, 2001) reported from the south-western Mediterranean a positive relation between habitat complexity and fish richness and abundance, with the size variability of boulders, especially the number of large boulders having the largest positive effect on richness and abundance, and Consoli et al. (2008) compared three rocky shores exhibiting significant differences in rock habitat complexity in the central Mediterranean where the most
complex habitat harbored the most abundant fish fauna most likely as a result of increased spatial and trophic niches.
The presence of three seagrass species as well as their high shoot densities in the Novigrad Sea is unique compared to nearby embayments less separated from the open Adriatic and less influenced by fresh water (Stiefel, 2009). Seagrass adapted search-and-attack and ambush predators, such as A. anguilla and Z. ophio-cephalus benefit from this situation (Schultz et al., 2009), as well as the green urchin which comprises >50% of the faunal abundance in sites with high Zostera marina proportions (Tab. 2). In the Novigrad Sea such elevated proportions of Z. marina are a sign of disturbance (Kruschel et al., 2009), which may come along with taxon dominance and reduced faunal diversity (Neira et al., 2007). In the Novigrad Sea the distribution of seagrass species is depth dependent and shallow Zostera seagrass beds support a different fauna from deep Cymodocea beds (Schultz et al., 2009).
Two possible reasons for the observed differences in faunal diversity are that (1) deeper seagrass seems to attract a variety of invertebrates absent from shallower waters, and (2) Zostera beds are avoided by fish taxa such as labrids and gobiids in favor of algal and sediment habitats because of the higher risk of predation posed by the Zostera resident predators A. anguilla and Z. ophiocephalus (Kruschel & Schultz, 2010; Schultz & Kruschel, 2010). Algal-associated labrids such as Sym-phodus ocellatus (Syoc) and S. cinereus as well as the sediment-associated Gobius niger (Goni) are known to prefer seagrass in other locations (Wiederholm, 1987; Guidetti, 2000, 2002; Stiefel, 2009). Diversity differences in the unconsolidated sediment fauna are primarily driven by evenness rather than species richness, as the relative abundance of the most common species G. niger (adults and juveniles) tends to peak in abundance on sediments near rock rather than near seagrass. This distribution may be a response to the risk of predation by and/or to spatial competition with the larger and more successful predator Z. ophiocephalus as has been demonstrated in field and tank experiments by Schultz & Kruschel (2010) and Kruschel & Schultz (2010). Similar cases of competition- and predation-driven niche partitioning among gobiid species has been reported else-
where (Wiederholm, 1987; Schofield, 2003; Malavasi et al., 2005).
CONCLUSIONS
In agreement with our prediction, the Novigrad Sea contains habitat-specific faunal assemblages, and the microhabitat diversity within four of the five basic habitats coincides with site differences in diversity of the associated faunal assemblages. Yet, only in the rock habitat may species diversity be directly linked to microhabitat niche availability. In the seagrass habitat, negative interactions between seagrass resident predators and their potential prey, and depth-related differences in seagrass microhabitat composition may be responsible for the species sorting. Similarly, taxon diversity differences among sediment sites may be a result of species competition (Gobidae) and/or avoidance of predation in the potentially available neighboring seagrass habitat (Gobiidae and juvenile fish). Overall faunal diversity in the Novigrad Sea seems to be driven by habitat diversity, either through increased availability of spatial and resource niches or through facilitation of spatial avoidance of competitors or predators.
Protection of faunal diversity in the Novigrad Sea should therefore focus on maintenance of the mosaic-like distribution of the five basic habitats and their microhabitats. The loss of the rocky-algal habitat would likely result in a greater reduction in biodiversity than the loss of other habitats, as through shoreline development, especially the conversion of the natural sublittoral habitat to concrete seawall. The loss of deep Cymodocea nodosa would also result in the loss of unique biodiversity, as it comprises a unique invertebrate fauna and does not support resident predators such as Zosterisessor ophio-cephalus or Anguilla anguilla. C. nodosa in the Novigrad Sea is threatened by competition with Zostera species, which are stimulated by anthropogenic disturbance (Kruschel et al., 2009), and is most sensitive to increased turbidity due to its restriction to deeper areas. The loss of shallow water sediment patches, the preferred habitat of a variety of juvenile fish, e.g., through Zostera expansion, would likely cause a reduction in the nursery function of the Novigrad Sea, leading eventually to a reduction of fish diversity and abundance.
RAZMERJA MED PESTROSTJO VRST MAKROFAVNE IN PESTROSTJO HABITATOV V LAGUNI V OSREDNJEM HRVAŠKEM JADRANU
Claudia KRUSCHEL & Stewart T. SCHULTZ University of Zadar, Department of Maritime Sciences, HR-23000 Zadar, M. Pavlinovica bb, Croatia E-mail: claudia@claudiakruschel.com
POVZETEK
Popis bentoške makrofavne plitvih predelov (0-6 m) Novigrajskega morja v osrednjem hrvaškem Jadranu je bil izveden v treh letnih časih 2007/08 in v okviru petih habitatnih tipov na desetih območjih. Relativno številčnost vseh identificiranih živalskih vrst v okviru petdesetih habitat x lokacija kombinacij smo obravnavali s klastersko analizo. Pokazala se je močna tendenca grupiranja habitatov, kar kaže na izredno povezanost favne z njenim prednostnim habitatom ter njuno prostorsko stabilnost. Pod-razvrščanje znotraj štirih glavnih habitatnih tipov se ujema z variabilnostjo območja v številu vrst in diverziteti. Ne glede na klastersko analizo pa lahko s habitatno diverziteto pojasnimo 29% favnistične diverzitete na skalnatih lokacijah in 39% na lokacijah, redko poraslih z morsko travo, medtem ko številčnost favne pojasni 46% favnistične diverzitete v skalnih habitatih (pozitivna korelacija) in 32% na neutrjenih sedimentih (negativna korelacija).
Ključne besede: pestrost habitatov, pestrost vrst, GPS, videografija, Jadransko morje
REFERENCES
Alongi, D. M. & P. Christofferson (1992): Benthic infauna and organisms-sediment relations in a shallow, tropical coastal area - influence of outwelled mangrove detritus and physical disturbance. Mar. Ecol. Prog. Ser., 81, 229-245.
Bakran-Petricioli, T., O. Antonic, D. Bukovec, D. Pet-ricioli, I. Janekovic, J. Krizan, V. Kusan & S. Dujmovic
(2006) Modelling spatial distribution of the Croatian marine benthic habitats. Ecol. Model., 191(1), 96-105. Bell, J. D. & M. Westoby (1986a): Abundance of mac-rofauna in dense seagrass is due to habitat preference, not predation. Oecologia, 68920, 205-209. Bell, J. D. & M. Westoby (1986b): Variation in Seagrass Height and Density over a Wide Spatial Scale - Effects on Common Fish and Decapods. J. Exp. Mar. Biol. Ecol., 104(1-3), 275-295.
Buschbaum, C., S. Dittman, J. Hong, I. Hwang, M. Strasser, M. Thiel, N. Valdivia & S. Yoon (2008): Mytilid mussels: global habitat engineers in coastal sediments. Helgol. Mar. Res., 63, 47-58.
Consoli, P., T. Romeo, U. Giongrandi & F. Andaloro (2008): Differences among fish assemblages associated with a nearshore vermetid reef and two other rocky habitats along the shores of Cape Milazzo (northern Sicily, central Mediterranean Sea). J. Mar. Biol. Assoc. U.K., 88(2), 401-410.
Dauwalter, D. C., W. L. Fisher & K. C. Belt (2006):
Mapping stream habitats with a gobal positioning system: Accuracy, precision, and comparison with traditional methods. Environ. Manage., 37(2), 271-280. Feireirra, C. E. L., J. E. A. Goncalvez & R. Coutinho
(2001): Community structure of fishes and habitat complexity on a tropical rocky shore. Environ. Biol. Fish., 61, 353-369.
Garcia-Charton, J. A. G. & A. P. Ruzafa (1998): Correlation between habitat structure and a rocky reef fish assemblage in the southwest Mediterranean. P.S.Z.N.I Mar. Ecol., 19, 111-128.
Garcia-Charton, J. A. G. & A. P. Ruzafa (2001): Spatial pattern and the habitat structure of a Mediterranean rocky reef fish local assemblage. Mar. Biol., 138, 917934.
Gratwicke, B. & M. R. Spreight (2005): The relationship between fish species richness, abundance and habitat complexity in a range of shallow tropical marine habitats. J. Fish Biol., 66, 650-667.
Gray, C. A., R. C. Chick & D. J. McElligot (1998): Diel Changes in Assemblages of Fishes Associated with Shallow Seagrass and Bare Sand. Est. Coast. Shelf Sci., 46, 849-859.
Guidetti, P. & S. Buscotti (2002): Effects of seagrass canopy removal on fish in shallow Mediterranean sea-grass (Cymodocea nodosa and Zostera noltii) meadows: a local-scale approach. Mar. Ecol., 140, 445-453.
Guidetti, P. (2000): Differences Among Fish Assemblages Associated with Nearshore Pos/Gon/a ocean/ca Seagrass Beds, Rocky-algal Reefs and Unvegetated Sand Habitats in the Adriatic Sea. Est. Coast. Shelf Sci., 50, 515-529.
Hilbig, B. & J. A. Blake (2000): Long-term analysis of polychaete-dominated benthic infaunal communities in Massachusetts Bay, USA. Bull. Mar. Sci., 67(1), 147164.
Hindell, G. S. & G. P. Jenkins (2005): Assessing patterns of fish zonation in temperate mangroves, with emphasis on evaluating sampling artifacts. Mar. Ecol. Prog. Ser., 290,193-205.
Hori, M., T. Suzuki, Y. Monthum, T. Srisombat, Y. Ta-naka, M. Nakaoka & H. Mukai (2009): High seagrass diversity and canopy-height increase associated fish diversity and abundance. Mar. Biol., 156, 1447-1458. Horinouchi, M., Z. Nakamura & S. Mitsuhiko (2005): Comparative analysis of visual censuses using different width strip-transects for a fish assemblage in a seagrass bed. Est. Coast. Shelf Sci., 65, 53-60. Janekovic, I., O. Antonic, J. Krizan, D. Bukovec & T. Bakran-Petricioli (2006): Modelling basic physical parameters in the Adriatic Sea as the basis for marine benthic habitats mapping. Ecol. Model., 194, 62-69. Jayaraj, K. A., P. Sheeba, J. Jakob, C. Revichandran, P. K. Arun, K. S. Praseeda, P. A. Nisha & K. A. Rasheed (2008): Response of infaunal macrobenthos to the sediment granulometry in a tropical continental margine-southwest coast of India. Est. Coast. Shelf Sci., 77, 743754.
Kruschel, C. & S. T. Schultz (2010): Lure-assisted visual census: A new method for quantifying fish abundance, behavior, and predation risk in shallow coastal habitats. Mar. Freshw. Res. (/n rev/'ew)
Kruschel, C., S. T. Schultz & S. Dahlke (2009): Evidence for seagrass competition in a Central Adriatic Croatian lagoon. Annales, Ser. Hist. Nat., 19(1), 1-10. La Mesa, G., S. Di Muccio & M. Vacchi (2006): Structure of a Mediterranean cryptobenthic fish community and its relationships with habitat characteristics. Mar. Biol., 149, 149-167.
Letourneur, Y., S. Ruitton & S. Sartoretto (2003): Environmental and benthic habitat factors structuring the spatial distribution of a summer infralittoral fish assemblage in the north-western Mediterranean Sea. J. Mar. Biol. Ass. U.K., 83, 193-204.
Lingo, M. A. & S. T. Szedlmayer (2006): The influence of habitat complexity on reef fish communities in the northeastern Gulf of Mexico. Environ. Biol. Fish., 76, 71-80.
Lipej, L., & M. Orlando-Bonaca (2006): Assessing blen-nioid fish populations in the shallow Gulf of Trieste: a comparison of four in situ methods. Period. Biol., 108, 151-157.
Luckhurst, B. E. & K. Luckhurst (1978): Analysis of the influence of substrate variables on coral reef fish communities. Mar. Biol., 49, 317-323. Macpherson, E. & U. Zika (1999): Temporal and spatial variability of settlement success and recruitment level in three blennoid fishes in the northwestern Mediterranean. Mar. Ecol. Prog. Ser., 182, 269-282. Malavasi, S., A. Franco, R. Fiorin, P. Franzo, P. Torri-celli & D. Mainardi (2005): The shallow water gobiid assemblage of the Venice Lagoon: abundance, seasonal variation and habitat partitioning. J. Fish Biol., 67(Suppl. B), 146-165.
Matic-Skoko, S., M. Peharda, A. Pallaoro, M. Cukrov & B. Baždaric (2007): Infralittoral fish assemblages in the Zrmanja estuary, Adriatic Sea. Acta Adriat., 48(1), 4555.
Moreno-Rueda, G., M. Pizarro, D. Ontiveros & J. M. Pleguezuelos (2009): The coexistence of the eagles Aquila chrysaetos and Hieraaetus fasciatus increases with low human population density, intermediate temperature, and high prey diversity. Ann. Zool. Fenn., 46, 283-290.
Mouillote, D. (2007): Niche-assembly vs. dispersal-assembly rules in coastal fish metacommunities: meta-communities: implications for management of biodiversity in brackish lagoons. J. Appl. Ecolv 44, 760-767. Nagelkerken, I., G. van der Velde, M. W. Gorissen, G. J. Meijer, T. van't Hof & C. den Hartog (2000): Importance of mangroves, seagrass beds and the shallow coral reef as a nursery for important coral reef fishes, using a visual census technique. Est. Coast. Shelf Sci., 51(1), 3144.
Neira, C., L. A. Levin, E. D. Grosholz & G. Mendoza (2007): Influence of invasive Spartina growth stages on associated macrofaunal communities. Biol. Invasions, 9, 975-993.
Norris, J. G., S. Wyllie-Echevierra, T. Mumford, A. Bailey & T. Turner (1997): Estimating basal area coverage of subtidal seagrass beds using underwater videography. Aquat. Bot., 58, 269-287.
Orlando-Bonaca, M. & L. Lipej (2007): Microhabitat preferences and depth distribution of combtooth blen-nies (Blenniidae) in the Gulf of Trieste (North Adriatic Sea). Mar. Ecol., 28, 418-428.
Patzner, R. A. (1999): Habitat utilization and depth distribution of small cryptobenthic fishes (Blenniidae, Go-biesocidae, Gobiidae, Tripterygiidae) in Ibiza (western Mediterranean Sea). Environ. Biol. Fish., 55, 207-214. Rosenberg, R., S. Agrenius, B. Hellman, H. C. Nilsson & K. Norling (2002): Recovery of marine benthic habitats and fauna in a Swedish fjord following improved oxygen conditions. Mar. Ecol. Prog. Ser., 234, 43-53. Rutzler, K., M. C. Diaz, R. W. M. van Soest, S. Zea, K. P. Smith, B. Alvarez & J. Wulff (2000): Diversity of sponge fauna in mangrove ponds, Pelicane bay, Belize. Atoll Res. Bull., 476, 231-248.
Schmude, K., M. J. Jennings, K. Otis & R. R. Piette
(1998): Effects of habitat complexity on macroinverte-brate colonization of artificial substrates in north temperate lakes. J. N. Am. Benthol. Soc., 17(1), 73-80. Schofield, P. (2003): Predation vulnerability of two gobies (Microgobius gulosus; Gobiosoma robustum) is not related to presence of seagrass. Fla. Sci., 68(1), 25-34. Schultz, S. T. & C. Kruschel (2010a): Frequency and success of ambush and chase predation in fish assemblages associated with seagrass and bare sediment in an Adriatic lagoon. Hydrobiologia, 649, 25-37. Schultz, S. T. (2008): Seagrass monitoring by underwater videography: Disturbane regimes, sampling design, and statistical power. Aquat. Bot., 33, 228-238. Schultz, S. T., C. Kruschel & T. Bakran-Petricioli (2009): Influence of seagrass meadows on predator-prey habitat segregation in an Adriatic lagoon. Mar. Ecol. Prog. Ser., 374, 85-99.
Sciberras, M., M. Rizzo, J. R. Mifsud, K. Camilleri, A. Borg, E. Lanfranco & P. J. Schembri (2009): Habitat structure and biological characteristics of a maerl bed off the northeastern coast of the Maltese Islands (central Mediterranean). Mar. Biodiversity, 39, 251-264. Sinovcic, G., M. Franicevic & V. Cikes Kec (2004): Unusual occurance and some aspects of biology of juvenile gilt sardine (Sardinella aurita Velenciennes, 1847) in the Zrmanja River estuary (eastern Adriatic). Appl. Ichthyol., 20, 53-57.
Sirota, L. & K. A. Hovel (2006): Simulated eelgrass Zostera marina structural complexity: effects of shoot length shoot density, and surface area on the epifaunal community of San Diego Bay, California, USA. Mar. Ecol. Prog. Ser., 326, 115-131.
Stiefel, J. (2009): Faunal communities of shallow ben-thic habitats in the Croatian Adriatic. B.Sc. Thesis. Ernst-Moritz-Arndt-University Greifswald, Germany. UNEP (1998): Annex I and II. Draft classification of marine habitat types for the Mediterranean region. WG. 149/3. Mediterranean action plan. Meeting of experts on marine habitat types in the Mediterranean region, SPA/RAC, Tunis.
Wiederholm, A. (1987): Habitat selection and interactions between three marine fish species (Gobiidae). Oikos, 48, 28-32.
Wilson, J. B. (1999): Guilds, functional types and ecological groups. Oikos, 86, 507-522. Ziv, Y. (1998): The effect of habitat heterogeneity on species diversity patterns: a community-level approach using an object-oriented landscape simulation model (SHALOM). Ecol. Model., 1 11, 135-170.