Annual dynamics of the mesozooplankton communities in a highly variable ecosystem (North Adriatic Sea, Italy)

Problem

The northern Adriatic Sea is a very productive region at several trophic levels from phytoplankton to fish. Within this region, the delta of the Po River is particularly rich, with high but variable phytoplankton biomass and production, whose dynamics is mainly related to the spreading of its plume (Franco 1973; Gilmartin & Revelante 1981). Therefore, a marked west-to-east decreasing gradient of phytoplankton biomass and production is often observed (Smodlaka & Revelante 1983), which can be reverted in case of severe and prolonged drought (Specchi & Fonda Umani 1983).

The Gulf of Trieste is an area in which zooplankton is mostly studied, with monthly sampling of a coastal station since 1970 (Cataletto et al. 1995; Kamburska & Fonda Umani 2006). The whole basin has been sampled at different occasions during oceanographic surveys in different seasons on an irregular basis (e.g.Hure & Scotto di Carlo 1969; Comaschi et al. 1998; Hure & Kršinić 1998). Only recently spatial surveys have covered the whole area on a monthly basis in the framework of the MAT (Giani et al. 2005) and INTERREG cooperative projects, allowing the collection of a reference dataset for the interpretation of zooplankton distribution and dynamics (Camatti et al. 2002; Fonda Umani et al. 2005a).

In the northernmost part of the northern Adriatic Sea (Gulf of Trieste and northward of the river Po delta) mesozooplankton is dominated by strictly neritic species with low diversity, and prevalence of Penilia avirostris in summer and Acartia clausi during the rest of the year at the surface. In the southern area, under the influence of the Po River, the zooplankton communities are strongly influenced by the hydrological characteristics of the different water masses. In summer a pycnocline separates surface warmer waters, enriched by nutrients from the Po River, from deeper colder and saltier waters. In the surface water masses, above the pycnocline, the zooplankton communities were characterized by a mixture of coastal species such as Paracalanus parvus and A. clausi; in the deeper water below the pycnocline, two cold relic species, Pseudocalanus elongatus and Temora longicornis, are dominant (Hure et al. 1980; Guglielmo et al. 2002; Kršinićet al. 2007). During winter, when vertical mixing of water masses occurs, the lowest zooplankton abundance and the most homogeneous spatial distribution were detected, with the prevalence of the genera Oithona, Clausocalanus, Calanus and Oncaea.

Fonda Umani et al. (1992) described two distinct areas in the northern Adriatic Sea in terms of biological properties:

The main goal of this paper is to describe the annual dynamics of mesozooplankton in the Northern Adriatic Sea, with the aim of identifying similarities and/or differences in their community structure at the species or higher taxa level as related to geographical and hydrological features.

To our knowledge, this is the first attempt to compare the composition of zooplankton species over the northern Adriatic Sea to assess whether specific composition significantly differs from area to area.

In the present paper we present data collected in monthly sampling of the area in 2001, from the Gulf of Trieste to the Po River area within the multidisciplinary project INTERREG III. The year 2001 was selected for completeness of the dataset in terms of geographical distribution of the sampled stations and because no mucilage events were recorded (Precali et al. 2005); these events are known to strongly impact zooplankton communities (Acri et al. 1999; Fonda Umani et al. 2005a), particularly during summertime. Our aim was also to draw conclusions on the general trends of zooplankton communities in the area, in comparison with previous studies.

Study area

The northern Adriatic Sea (Fig. 1) is a shallow basin located between the northeastern coast of the Italian Peninsula and the western coast of the Balkans.

The general circulation of this sub-basin is influenced by its physiography and by the meteorological patterns (Franco 1973, 1983; Buljan & Zore-Armanda 1976; Franco et al. 1982). The most important factors controlling water density are the wide fluctuation of the surface heat fluxes and the large volume of fresh water inputs. During winter the cold waters diluted by the western riverine inflows remain confined in a coastal belt, separated from the offshore waters by a frontal system. The offshore waters have high salinity, being advected from the southern part of the basins, and are actively mixed by wind-driven surface cooling and mechanical stirring. A significant part of the nutrients from river sources is retained inside the coastal belt, where their concentration becomes very high. Instead, the offshore waters show quite low nutrient concentrations, mainly derived from recycling and advection from the southern waters. Both in the coastal and in the offshore zone the vertical mixing causes a homogeneous distribution along the water column of dissolved and suspended matter. In early spring the inversion of the total heat budget leads to the generation of a thermocline. The dilution due to the Po River inputs increases the buoyancy of the surface layer, which expands to a large part of the northern Adriatic. Successive heating and dilution processes, wind mixing, and northward advection in the summer generate a highly stratified water column. Diffusive transport of the terrigenous and riverine inputs is the prevailing mechanism of nutrient transfer, part of which is retained in the coastal waters. A larger part of nutrients discharged from the Po River is distributed over the whole basin, driven by river plume dynamics (Fonda Umani et al. 1992).

The Gulf of Trieste is the northernmost and shallowest part of the Adriatic Sea. Its main freshwater inputs come from the Isonzo River from the northwestern coast, with highest discharges generally observed in late spring and autumn. In addition to the year-to-year variations, the system is characterized by a typical cold temperate annual cycle with an intense early spring diatom bloom, a nutrient-depleted summer and a second, short-lasting fall bloom (Mozetićet al. 1998; Harding et al. 1999).

In 2001 temperatures ranged from an absolute value of 6.6 °C in winter to a maximum value of 28.1 °C in August; salinity showed a maximum of 38.5 PSU, falling to 8.1 PSU in the area in front of the Po River. In July, the persistence of a confined diluted water mass was observed in the open sea station (Fig. 2a, b) and a clear trophic gradient was observed, encompassing the permanent mesotrophic coastal area, the highly dynamic frontal zone and, at its eastern side, the more offshore oligotrophic area (Bernardi Aubry et al. 2006). Phytoplankton dynamics showed a marked spatial heterogeneity and seasonal fluctuations, but in general, phytoplankton showed a decreasing gradient of abundance and biomass from coastal to offshore waters.

In the Gulf of Trieste surface temperature and salinity variations were tighter, temperature ranging from 8.0 °C to 25.8 °C, and salinity from 34 to 37.9 PSU. From January to June 2001, temperatures remained always higher than 10 °C, even in February, when typically the lowest temperature of 6 °C is registered. Low salinity due to continuous fresh water supplies was also observed. From July to December, lower temperatures and higher salinities than in previous years were registered, which allowed significant intrusions in the gulf of the Modified Levantine Intermediate Water from the southeast (Fonda Umani 2001).

Material and Methods

Six stations were sampled monthly for zooplankton determinations of abundances and taxonomic composition, from January to December 2001 along two coast-to offshore transects (C, E), one station (2E04) southward of the Po River delta, and one station (C1) in the Gulf of Trieste (Fig. 1). In November only station C1 was sampled. Station 2E04 was sampled every month except in May, July, October, November and December. Samples were collected by vertical hauls, from the bottom to the surface, using a WP2 net (0.57 m diameter, 200 μm mesh, filtered volume 3.25–3.75 m³) in the Gulf of Trieste and an Apstein net (0.40 m diameter, 200 μm mesh size, filtered volume 0.50–5.28 m³) at all other stations, and preserved with borax-buffered formaldehyde. The use of WP2 net is highly recommended because of its mesh-to-mouth-area ratio of 6:1 (Sameoto et al. 2000). However, in the present work we have used the smaller Apstein net (5:1 ratio), due to the shallow depth of the coastal station E01 (∼5 m). The diameter of the mouth of this net is likely to be too narrow to collect large zooplankton species, which can escape the net during the haul, and therefore these (e.g. the copepod C. helgolandicus) may have been under-sampled in our survey.

Taxonomic and quantitative zooplankton determinations were performed using a Zeiss stereomicroscope at the lowest possible taxonomic level (species for copepods and cladocerans) on the total sample or on a representative sub-sample. Each sample was poured into a beaker to allow a thorough mixing for random distribution of the organisms. Aliquots of the samples were analysed until at least 1000 individuals were counted.

Diversity was estimated, at the species level, from each station and sampling period, calculating the Shannon–Weaver diversity index (H′; Shannon & Weaver 1949), using the PRIMER 5 software (Clarke & Warwick 2001).

Because the Correspondence analysis (CA) is the most suitable statistical technique to analyse enumerative data (Davis 1986), we used it to highlight possible relationships and groupings between stations and zooplankton species (Benzècri 1980; Ter Braak 1986; Legendre & Legendre 1998). The dataset was organized in a species/station matrix, eliminating the rarest species (< 2% of the abundance). In our analysis, we did not include station 2E04 due to a lack of samplings in some periods. Moreover, we preferred to analyse only the spatial affinity between stations and species (pooling the data of all sampling periods for each station), and not the temporal dynamics, which was evaluated only in a descriptive manner. The taxa included in the matrix contributed 98%of the total zooplankton species composition summed for each station for all the 11 months of sampling. The results were plotted on a Cartesian plane (the F1 and F2 axes of which represent the two main transformations). As it is based on the chi-squared metric, this algorithm automatically weights both low and high frequencies. Statistical analyses were performed using R software (R Development Core Team 2007).

Results

Over the entire sub-basin the main groups of the mesozooplankton showed very wide seasonal fluctuations, with copepods dominating throughout the year, except in summer when cladocerans were dominant. Mean annual abundances were highest at station E01 (20,332 ± 8220 ind·m⁻³), station C1 (9796 ± 1807 ind·m⁻³) and at the offshore station E12 (11,679 ± 7953 ind·m⁻³). Percent contributions of the most abundance species to total zooplankton individuals are reported in Table 1. At stations E01 and E12 the highest average abundance of copepods and cladocerans was observed, with 17,595 and 982 ind.·m⁻³, and 1504 and 9336 ind.·m⁻³, respectively (Fig. 3).

Total zooplankton abundances increased in summer at all stations except E01 where, instead, winter maxima were observed in January due to A. clausi (79,680 ind·m⁻³) and in December by A. clausi, P. parvus and Oncaea spp. (47,360 total ind·m⁻³) (Fig. 4). The rest of the zooplankton community, mainly represented by the meroplankton (i.e. larvae of bivalves, echinopluteii), copepod nauplii and appendicularians, was more homogeneous throughout the year (Table 1, Fig. 4).

A few species dominated the zooplankton community over the whole sub-basin (Table 1), such as the cladoceran P. avirostris, which ranked first as percentage contribution at all stations except at E01 and 2E04 (probably because we missed the July sampling), and the calanoid copepods P. parvus and A. clausi: the former ranking first at 2E04 and second at all stations except C1 and E12. Acartia clausi ranked first at E01 and second at C1, but was less well represented at all other stations. Oncaea spp. ranked third at all stations except at E12. These four taxa comprised more than 50% of the total zooplankton abundance whereas the mean annual percentage contribution of the remaining taxa accounted for much less (Table 1). Penilia avirostris prevailed during the warmer period from June to September, but it was present over the entire basin until December, the highest abundances being observed at E12 and C1 (Fig. 4). Acartia clausi was more abundant at the coastal stations (C06, C10, E01, 2E04, C1), showing spring maxima, except at E01, where it exhibited a more irregular annual cycle with peaks in January, February and December (Fig. 4). Paracalanus parvus registered the highest abundances in winter/autumn with peaks in October and December at C06 and E01, respectively, and lower abundances at C1. Oncaea spp. were more homogeneously distributed, showing an annual pattern similar to P. parvus. The cyclopoids Oithona similis and Oithona nana were more abundant at E01 and C1, respectively; O. similis peaked in late winter and spring, whereas O. nana showed more constant concentrations throughout the year (Fig. 4). Echinopluteii reached the highest abundances during the warm season, with the highest peak in September at C06, whereas the abundance decreased at E01, being replaced by bivalve larvae. The highest abundance of appendicularians was found at C1, whereas siphonophores prevailed in the coastal waters and, particularly, at the stations near the Po River (C1 and 2E04, Fig. 4).

In general, the species diversity was higher in winter and autumn (Table 2), probably due to northward advection of water masses. The lowest diversity was observed at E01 in winter due to the absolute dominance of A. clausi (84%) and at E12 in summer because of the exceptionally high dominance of P. avirostris (82%).

CA allowed separation of the stations into one main group and three isolated stations located (i) in front of the Po River delta (E01), (ii) in the Gulf of Trieste (C1), and (iii) offshore of the E transect (E12). The main group includes all stations situated along the C transect (C06, C10, C15) and station E06 (Fig. 5); these stations have in common species composition and annual dynamics, because their species percentage distribution was similar, with higher specific diversity with respect to all other stations. Their mesozooplankton communities were probably less influenced by land-derived inputs, and can therefore be considered representative of the open waters of the sampled area. In fact, the same species percentage distribution differed considerably in the coastal stations, E01 and C1, and in the farthermost offshore station E12. The latter was characterized by the high abundance of P. avirostris, whereas station E01 was characterized by a dominance of A. clausi, cirriped nauplii and the cladoceran Podon polyphemoides, a strictly neritic species. Station C1 showed higher percentages of siphonophores and appendicularians.

Discussion

The main result of our analyses was the identification of a cluster of stations (C15, C10, C06 and E06) that can be considered representative of the annual dynamics of the zooplankton communities in the open waters of the northern Adriatic Sea. The most abundant species at these stations was P. avirostris, although its dominance was limited to the summer period. This species started to swarm in May and remained present until December. Compared to previous studies, it appears that P. avirostris entered the North Adriatic not earlier than 1914 (Leder 1917), becoming dominant in summer, thanks to its parthenogenetic reproduction, but the period of its presence is expanding in the entire northern sub-basin. Specchi & Fonda (1974) report its first occurrence in the Gulf of Trieste from the last week of June in 1970, at a temperature of 20.6 °C, to a decrease in abundance below detection by the end of November, at a temperature of 13.8 °C. Penilia avirostris diet is very flexible, ranging from small diatoms and bacterivorous microflagellates (Turner et al. 1988) to prymnesiophyceans (Paffenhöfer & Orcutt 1986) and bacteria (Lipej et al. 1997). Katechakis & Stibor (2004) reported significant grazing on elongate diatoms (Nitzschia elongata and Rhizosolenia spp.) or long-chain diatoms (Skeletonema costatum and Thalassiosira spp.) up to 135 μm long. Recent studies in the Gulf of Trieste show that in summer 2000 P. avirostris consumed preferentially diatoms, and particularly the elongate cells of Proboscia (Rhizosolenia) alata (diameter of 2.5–13 μm), which might be ingested by P. avirostris if orientated longitudinally in their filtering current. In summer 2002, instead, this species fed upon autotrophic nanoplankton (Fonda Umani et al. 2005a). In both experiments P. avirostris grazed also on microzooplankton, mostly small ciliates and heterotrophic dinoflagellates.

The apparent extension of its period of appearance might be attributed to the general warming of the system (Russo et al. 2002) as well as to its feeding flexibility, spanning from bacteria to large diatoms, which enables P. avirostris to use all available food resources regardless of their dimensions. Until now, the factors controlling the discontinuous distribution of marine cladocerans during the year, with peaks of abundance followed by rapid decline and absence from the plankton, remain unclear. Some authors suggest that temperature may play an important role in the population dynamics of P. avirostris (Onbé & Ikeda 1995). However, other factors such as food availability, chemical composition of seston, and photoperiod might be also relevant (Egloff et al. 1997). Environmental conditions such as photoperiod, food availability, turbulence, crowding, and predation seem to play key roles in the decline of populations (Stross & Hill 1968; Frey 1982; Fofonoff 1994); however, temperature has been proposed as the main physical factor that controls P. avirostris populations (Gieskes 1971; Onbé & Ikeda 1995). In the eastern coast of the Adriatic Sea, Vidjak et al. (2006) report the unusually high abundance of P. avirostris in December 2003 due to the unusually high temperatures of the surface layer (∼14 °C).

The second most abundant species in the studied area was P. parvus, a small calanoid, which has already been identified as the dominant copepod species in the central and southern Adriatic (Hure et al. 1980; Fonda Umani et al. 1994), although high abundances have never been recorded in the northern Adriatic Sea. This species has not only recently expanded its geographical distribution northward, but it has also replaced other species, becoming dominant. Recent reports from the long-time sampling series in the Gulf of Trieste (36 years) highlight the increasing trend in the abundance of this species in the last 20 years (S. Fonda Umani, unpublished data). These findings can be interpreted as an indication of the general warming of the Adriatic Sea and also point to the oligotrophication of the northern basin in the recent past. This latter feature is probably due to a general improvement in sewage management and the ban on phosphate-containing detergents which were previously massively discharged in the basin by the river outputs. The northern offshore area is presently more similar to the most southern Adriatic Sea, and, in general to the oligotrophic Mediterranean Sea.

The third most represented zooplankton taxa were species belonging to Oncaea, a small poecilostomatoid copepod genus, which was possibly underestimated because of the use of 200-μm mesh size (Kršinićet al. 2007), and which has been reported in the past in the northern Adriatic Sea, but in lower concentrations in the central and southern parts of the basin (Fonda Umani et al. 1994). A significant increase in the abundance of this genus has already been reported in the Gulf of Trieste (Kamburska & Fonda Umani 2006) and has been correlated to the observed temperature increase.

In our study, A. clausi ranked fourth, fifth or seventh (except at E01 and C1), being replaced by P. parvus all over the basin, as opposed to previous studies, where it was reported as dominating the whole northern basin (Fonda Umani et al. 1994). A. clausi and P. parvus have the same average size and probably also share the same diet (Fonda Umani et al. 2005a). Unfortunately, there are no specific grazing assessments for P. parvus in the studied area. Recently, Fonda Umani et al. (2005a) reported that A. clausi has an omnivorous habit, but feeds mainly on diatoms, in particular S. costatum, and also on microzooplankton when diatoms were scarce. Given its preference for diatoms, a general decrease in A. clausi is probably due to less intense diatom blooms, compared to the past, registered in the offshore northern waters (Fornasaro et al. 2006). To confirm this, the exceptionally high abundance of A. clausi in winter and autumn 2001 at E01, a station under the direct influence of the Po River discharge, can be related to the intense diatom blooms recorded that year (Boldrin et al. 2005; Bernardi Aubry et al. 2006). The extremely high tolerance of A. clausi (and other congeneric species) for low salinities, which determines its relevant abundance in many estuaries all over the world (e.g.Alcaraz 1983; Pastorinho et al. 2003), can also justify its huge numbers at the most riverine stations during our surveys, confirming previous data (Specchi & Fonda Umani 1987).

A. clausi was still the dominant copepod in the Gulf of Trieste, and this was the main difference between this station and the cluster of northern Adriatic offshore stations, as defined before. In addition, in the semi-enclosed Gulf of Trieste, an increase of Oncaea spp. and P. parvus abundance compared with the past has been reported (Specchi & Fonda 1981; S. Fonda Umani, unpublished data).

The general low abundances of P. elongatus and T. longicornis recorded during our surveys may be related to a year-to-year fluctuation of their populations, as already observed in the past (Fonda Umani et al. 1994; Acri et al. 1999). However, in the Gulf of Trieste the cold species P. elongatus is still decreasing, mostly due to the increase in the autumn temperatures that delays its appearance and results in a shorter period of occurrence in the water column (S. Fonda Umani, unpublished data).

The offshore station E12 was distinguished from the other stations because of the extremely high P. avirostris abundance in July, when a less saline body of water was observed (Fig. 2b). A similar feature was observed by Tassi Pelati et al. (1983) in November 1980 when they found the absolute dominance of P. avirostris at an offshore station which still presented summer characteristics (strong stratification of less saline waters at surface). Probably, as in July 2001, a surface water body derived from the Po River remained isolated and moved eastward, conserving its biological properties, including zooplankton.

Except at station E01, located within the plume of the Po River, the highest zooplankton abundances were observed in summer, mostly due to P. avirostris, whose mean dry weight (1.2 μg per individual) is relatively low when compared, for instance, with A. clausi (4.6 μg per individual, Fonda Umani 1985). As a consequence, the peak in abundance did not convert into a higher zooplankton biomass. Indeed, in the Gulf of Trieste the peak in zooplankton biomass was registered in April 2001 (Kamburska & Fonda Umani in press) following the March diatom bloom, mainly due to Pseudonitzschia cf. pungens (5 × 10⁵ cells·l⁻¹) and Asterionellopsis glacialis (4–5 × 10⁵ cells·l⁻¹) (Fonda Umani 2001) and to A. clausi, copepod nauplii, appendicularians and O. nana. At the other stations (except E01), the highest spring mesozooplankton abundances were observed in March, 1 month after the most intense diatom bloom (Skeletonema marinoi– 3.5 × 10⁷ cells·l⁻¹) and were due to P. parvus, Oncaea spp. and O. similis (Bernardi Aubry et al. 2006). The specific pattern of station E01 in the initial part of 2001 was probably due to high nutrient loads from an exceptional flood that occurred in the autumn of the previous year (Boldrin et al. 2005), which substantially enriched the coastal area. The Po River and its tributaries showed persistent high discharges later on (Boldrin et al. 2005), and the dissolved nutrient concentrations stayed relatively high and may well have sustained an elevated trophism of the area with intense phytoplankton blooms (Bernardi Aubry et al. 2006), probably reflected in the high concentrations of A. clausi, P. parvus and oncaeids at station E01 later during the year.

Conclusions

Offshore stations with similar biological features could be grouped together and can be considered representative of undisturbed zooplankton dynamics in the open water of the northern Adriatic Sea. Comparing the present results with previous studies we observed significant changes in the zooplankton community composition, mostly due to the general decrease of A. clausi, which was replaced by P. parvus. The only exceptions were the Gulf of Trieste and the Po River plume, where A. clausi was still dominant. In summer, P. avirostris prevailed in the whole basin, with a prolonged period of swarming compared with the past. We also observed an increase in Oncaea spp. when compared to previous data. These observations, together with the indications from the long-term sampling series of zooplankton from the Gulf of Trieste (S. Fonda Umani, unpublished data) point to changes in the dynamics of plankton communities possibly related to the observed general increase of average sea water temperatures of the basin.

In fact, the northern Adriatic Sea in recent years has been undergoing profound changes, mostly due to a temperature increase and a reduction of phosphate loads that strongly impact the zooplankton communities (Kamburska & Fonda Umani 2006). Long-term studies at basin scale are still needed to monitor continuously the evolution of this delicate ecosystem, which is particularly sensitive and exposed to the global climatic changes.