MARINE ECOLOGY PROGRESS SERIES Vol. 247: 1–16, 2003 Published February 4 Mar Ecol Prog Ser Seasonal variations in the dynamics of microbial plankton communities: first estimates from experiments in the Gulf of Trieste, Northern Adriatic Sea Serena Fonda Umani1, 2,*, Alfred Beran1 1Laboratorio di Biologia Marina, Via Piccard 54, 34010 Trieste S. Croce, Italy 2Dipartimento di Biologia, Università di Trieste, Via Weiss 2, 34127 Trieste, Italy ABSTRACT: Heterotrophic plankton grazing was studied in the Gulf of Trieste (Northern Adriatic Sea) from November 1998 to August 1999 using the dilution method. Four sets of experiments were carried out quarterly in order to assess the impact on communities of both phototrophic and heterotrophic prey. Four different trophic models were observed: during the autumn microzooplankton fed on small dinoflagellates, and phototrophic (PNAN) and heterotrophic nanoflagellates (HNAN), but not on the abundant bulk of diatoms. The entire initial HNAN standing stock was removed; we therefore did not observe any mortality of bacteria, whose biomass was the highest in the whole period. In late winter, the intense diatom bloom of Lauderia annulata remained almost untouched as the microzooplankton fed only on a less abundant small sized diatom (Chaetoceros). Microzooplankton also fed on HNAN, halving the mortality of bacteria induced by HNAN grazing only. In late spring, microzooplankton grazed effectively on a large array of prey (small diatoms, PNAN and HNAN). Reduction of bacterial mortality, exerted by microzooplankton through grazing on HNAN, was less evident, possibly due to direct microzooplankton grazing on bacteria. During the summer, we observed an intense grazing on bacteria by microzooplankton, which shifted from the usual nano-sized prey organisms, due to their extreme paucity, to bacteria. In conclusion, microzooplankton grazing was highly selective and vari- able due to the prey composition and to the predator community structure, which were investigated at the species to genus level. Microzooplankton was unable to control a bloom of large-sized diatoms, but showed a high level of control on most of the PNAN fractions. The result of this selection contributed significantly to the shaping of the phytoplankton community structure. Microzooplankton controlled HNAN biomass even more efficiently with relevant indirect effects on bacterial mortality. KEY WORDS: Microzooplankton herbivory · Microzooplankton bacterivory · HNAN bacterivory · Microbial food web · Seasonal trophic models Resale or republication not permitted without written consent of the publisher INTRODUCTION microzooplankton and HNAN has been studied in the Gulf of Trieste (e.g. Fonda Umani et al. 1995); however, The Northern Adriatic and the Gulf of Trieste have the mortality induced by grazing on phototrophic and been intensely investigated since the 19th century (e.g. heterotrophic plankton components has only been Fonda Umani 1996). However, before this project, inferred by distribution analyses (e.g. Del Negro et al. there were no studies on the grazing impact of either 2001). Direct determination of the relative grazing microzooplankton (20 to 200 µm protozoa of various rates of microzooplankton and HNAN, and quantifica- taxa, and small metazoa larvae and nauplii) or hetero- tion of specific prey mortality rates is critical for under- trophic nanoplankton (HNAN; 2 to 20 µm nanoflagel- standing the patterns of energy and material flows lates and ciliates). The occurrence and distribution of through the pelagic food web. Microzooplankton and *Email: [email protected] © Inter-Research 2003 · www.int-res.com 2 Mar Ecol Prog Ser 247: 1–16, 2003 HNAN are able to select food and display rapid re- toring chl a concentration following dilution, several sponses to changes in food availability. Therefore, they researchers estimated taxon- or pigment-specific mor- play significant roles in structuring plankton com- tality rates using pigment analysis by HPLC (e.g. munities (e.g. Paranjape 1990, Fahnenstiel et al. 1995, Strom & Welschmeyer 1991, McManus & Ederington- Froneman & Perissinotto 1996, Verity et al. 1996, James Cantrell 1992, Verity et al. 1993, Waterhouse & & Hall 1998, Lessard & Murrell 1998). Studies in other Welschmeyer 1995, Latasa et al. 1997, Schlüter 1998) marine environments have shown that through their and flow cytometry (e.g. Landry et al. 1995, Recker- grazing impact, microzooplankton and HNAN can mann & Veldhuis 1997, Kuipers & Witte 1999, Stelfox- control phytoplankton production (e.g. Gifford 1988, Widdicombe et al. 2000). We conducted microscopical Verity et al. 1993, Burkill et al. 1995, Cotano et al. 1998) counts of microzooplankton and phytoplankton (e.g. and dynamics (e.g. Landry et al. 1993, Strom & Strom McManus 1995, Verity et al. 1996, Nejstgaard et al. 1996, Latasa et al. 1997, Ruiz et al. 1998, Stelfox- 1997, Fonda Umani & Zanon 2000), and enumeration Widdicombe et al. 2000) as well as play a significant of phototrophic and heterotrophic nano- and pico- role in nutrient regeneration (Goldman et al. 1987). plankton by epifluorescence microscopy (e.g. Camp- Our knowledge of protozoan grazing in controlling bell & Carpenter 1986, Verity et al. 1993, Ayukai 1996, energy fluxes from producers to higher trophic levels James & Hall 1998). has steadily increased over the last 2 decades, leading The study was carried out under a 3 yr monitoring to a new perception of the marine food webs that project (INTERREG 2, Italy and Slovenia) from July has dramatically changed from the classic grazing 1998 to June 2001 in the Gulf of Trieste (Northern food chain to a more complex multivorous food web Adriatic Sea), conducted on a biweekly to monthly (Pomeroy 1974, Azam et al. 1983, Rassoulzadegan basis. The sampling grid comprised 30 stations; in 3 of 1993, Legendre & Rassoulzadegan 1995), where the these, we analysed most of the biotic components in microbial loop plays a pivotal role. In most of the the water column (e.g. pico-, nano-, microphyto- and oceanic environments, major fluxes of organic matter microzooplankton compositions). The dilution experi- move via dissolved organic matter to bacteria and the ments were performed in the most coastal ‘biological’ microbial loop. A significant part of this energy can station. Based on the monitoring results and the histor- be channelled to higher trophic levels via protozoan ical data set, we selected 4 distinct periods which grazing on bacteria. correspond to 4 different environmental conditions: Many studies have considered the sum of the graz- the moment of the first annual diatom bloom which ing impact of microzooplankton and HNAN (e.g. appeared in February (e.g. Fonda Umani 1992), the Caron et al. 1991, Murrell & Hollibaugh 1998), but only shift from diatoms to nanoflagellates in the phyto- few measured the relative effects of HNAN and mi- plankton composition in May (e.g. Malej et al. 1995), crozooplankton in combined experiments (Kuuppo- the lowest yearly phytoplankton biomass in August Leinikki 1990, Reckermann & Veldhuis 1997) or the (e.g. Mozetic et al. 1998) and the autumn diatom bloom predator communities composition (e.g. Gifford 1988, (e.g. Malej et al. 1995). Paranjape 1990, Verity et al. 1993, Froneman & Peris- In the 4 periods, we performed separate parallel sinotto 1996, Froneman et al. 1996, Strom & Strom dilution experiments with natural assemblages to 1996, James & Hall 1998, Dolan et al. 2000). assess the grazing impact of both microzooplankton The main goals of our study were: (1) to quantify the and HNAN. Therefore, we were able to consider carbon fluxes through the microbial community; (2) to microzooplankton and HNAN separately as potential identify any prey selectivity exerted by heterotrophic grazers, as well as microphytoplankton, and photo- communities; and (3) to analyse the relationships be- trophic (PNAN) and heterotrophic (HNAN) nano- tween the microzooplankton and HNAN grazing im- plankton as potential prey. This also enabled us to pact on picoplankton prey. Therefore, we attempted to distinguish between microzooplankton and HNAN analyse both predator and prey communities at genus grazing induced mortality of bacteria. to species level, and determined both specific growth and grazing rates of the predators represented by microzooplankton and HNAN, as well as specific MATERIALS AND METHODS growth and mortality rates of the prey represented by bacteria, nanoplankton and microphytoplankton. Study area. The Gulf of Trieste is a semi-enclosed A number of approaches have been used to measure shallow system (maximum depth 25 m), in the north- the grazing impact of both micro- and nano-predators ernmost eastern part of the Northern Adriatic Sea on a wide range of prey, but most used the dilution (Fig. 1). It stretches from the Isonzo (Soca) River, the method (Landry & Hasset 1982, reviewed by Dolan et most important source of fresh waters to Punta Salvore al. 2000). In addition to the classical approach of moni- (Istrian peninsula). It is mostly controlled by pulsing Fonda Umani & Beran: Seasonal variations in microbial plankton communities dynamics 3 The surface layer (from 0 to 5 m), usually flows clockwise towards Trieste (Stravisi 1983). The primary production, which is around 50 g C m–2 yr–1 (Faganeli & Malej 1981, Fonda Umani 1991), is supported by >10 µm producers in late winter and by the <10µm size class in the rest of the year (Malej et al. 1995). The composition and annual dynamics of pico-, nano-, microphyto-, microzoo- and net-