The influence of weather conditions (temperature and wind) on cyanobacterial bloom development in the Gulf of Finland (Baltic Sea)

Cyanobacterial blooms are common in the Baltic Sea. They are dominated by Aphanizomenon flos-aquae and Nodularia spumigena and take place in July–August. Investigations of bloom development using different approaches have been carried out in the Gulf of Finland during recent years. The ship-of-opportunity technique allows to observe the upper layer dynamics from meso- to basin-wide scale with high temporal and spatial frequency at low cost. Unattended measurements on board a commercial ferry along a transect between Tallinn and Helsinki have been conducted for 3 years (1997–1999). The influence of weather conditions—temperature and wind—on the cyanobacterial bloom development was investigated. The formation of cyanobacterial blooms was favoured by warm and calm weather, while in cold and windy conditions other species formed mass occurrences. Water temperature has been found to be the main factor controlling the initiation of the bloom, in general, while vertical stratification appeared to be the critical factor determining the intensity of the bloom at species level. The spatial distribution of the cyanobacterial bloom was determined rather by the wind-forced advection than by the possible vertical transport of nutrients in the areas of the observed upwelling events.     

Distribution Patterns of Isomorphic Cold-Water Dinoflagellates (Scrippsiella/Woloszynskia Complex) Causing ‘red tides’ in the Baltic Sea

During the latest years medium-sized (15–30 μm), single-celled dinoflagellates have been reported to form blooms in the northern Baltic Proper and the Gulf of Finland in winter and spring. Recent studies (Kremp et al., 2003. Proceedings of the 7th International conference of Modern and Fossil Dinoflagellates, September 21–25, Nagasaki, Japan, 66 pp.) indicate that those blooms are caused by two isomorphic species – Scrippsiella hangoei (Schiller) Larsen, and a new species, tentatively belonging to the genus Woloszynskia. Until now there has been no report on how widely distributed these phytoplankton species are in the Baltic Sea. In this study, the occurrence of Scrippsiella/Woloszynskia complex in the entire Baltic Sea was investigated, by using monitoring data from 1997 to 2003. The species occurred in a salinity range from 2 to 8 PSU. Highest concentrations were observed at salinity 4.5–6.5 PSU. Maximum cell densities of Scrippsiella/Woloszynskia complex in the water column were mainly obtained in April or in the beginning of May by the water temperature <3 °C prior to stratification was formed. In the central Gulf of Finland, the second maximum was found in 1999 and 2002 by the temperature >6 °C. Bloom formations in the Baltic Proper and in the Gulf of Finland may not only be explained by optimum temperature and salinity, but also with other factors e.g. high nutrient concentrations and good seeding conditions from the sediments.     

A Fuzzy Logic Model to Describe the Cyanobacteria Nodularia spumigena Blooms in the Gulf of Finland, Baltic Sea

A fuzzy logic model to describe the seasonal evolution of Nodularia spumigena blooms in the Gulf of Finland was built and calibrated on the basis of monitoring data. The model includes three phosphate sources: excess phosphate after the annual spring bloom and parameterised phosphate transport to the upper mixed layer by turbulent mixing and upwelling events. Surface layer temperature and wind mixing form the physical conditions controlling the growth of N. spumigena. Model simulations revealed that phosphate input caused by turbulent mixing and upwelling have to be taken into account to achieve the best fit with observed data. Testing the fuzzy model for early prediction of maximum N. spumigena biomass about a month before the usual occurrence of blooms, gave good results. The potential use of the model for prediction of bloom risk at a certain location along the Estonian or Finnish coast was tested. The bloom transport velocities used in the fuzzy model were pre-calculated by a 3D numerical circulation model for different wind regimes.     

Growth rates and elemental composition of Alexandrium monilatum, a red-tide dinoflagellate

The combined effects of temperature and salinity on growth of Alexandrium monilatum were studied in laboratory cultures. This toxic, red-tide dinoflagellate grew faster with higher temperatures, up to a maximum of approximately 1 division per day at 31 °C. Salinities above 15 psu had a lesser effect on growth rate, as might be expected for an estuarine species. Growth rates of cultures exposed to natural light and temperature fluctuations were comparable to laboratory cultures. The minimum N cell quota suggested that high N flux would be required to support bloom development. A literature survey of documented A. monilatum blooms indicated that within US waters, blooms occur in July–September in nearshore or estuarine regions of the Gulf of Mexico and the Florida Atlantic coast. Temperature and salinity measured during blooms correspond to the optimal growth conditions of the laboratory cultures. Nevertheless, the occurrence of A. monilatum blooms is sporadic compared to the occurrence of seemingly optimal growth conditions. Laboratory growth experiments predict when blooms of this species are unlikely due to low growth rates, but so far cannot predict individual blooms.     

Parasitism as a biological control agent of dinoflagellate blooms in the California Current System

Amoebophrya is a marine parasite recently found to infect and kill bloom-forming dinoflagellates in the California Current System (CCS). However, it is unknown whether parasitism by Amoebophrya can control dinoflagellate blooms in major eastern boundary upwelling systems, such as the CCS. We quantified the abundance of a common bloom-forming species Akashiwo sanguinea and prevalence of its parasite (i.e., % infected cells) in surface water samples collected weekly from August 2005 to December 2008 at the Santa Cruz Wharf (SCW), Monterey Bay, CA. Additionally, we measured physical and chemical properties at the SCW and examined regional patterns of wind forcing and sea surface temperature. Relative abundance of the net phytoplankton species was also analyzed to discern whether or not parasitism influences net phytoplankton community composition. Epidemic infection outbreaks (>20% parasite prevalence in the host species) may have contributed to the end or prevented the occurrence of A. sanguinea blooms, whereas low parasite prevalence was associated with short-term (≤2 weeks) A. sanguinea blooms. The complete absence of parasitism in 2007 was associated with an extreme A. sanguinea bloom. Anomalously strong upwelling conditions were detected in 2007, suggesting that A. sanguinea was able to outgrow Amoebophrya and ‘escape’ parasitism. We conclude that parasitism can strongly influence dinoflagellate bloom dynamics in upwelling systems. Moreover, Amoebophrya may indirectly influence net phytoplankton species composition, as species that dominated the net phytoplankton and developed algal blooms never appeared to be infected.     

Dynamics of potentially harmful microalgae in a confined Mediterranean Gulf—Assessing the risk of bloom formation

The population dynamics of potentially harmful microalgae was investigated in the semi-enclosed shallow Gulf of Kalloni, Greece (Aegean Sea, Eastern Mediterranean), during a 2-year period from August 2004 to March 2006. A total of 21 potentially harmful microalgae (bloom-forming and/or toxic) were identified including 3 diatoms and 18 dinoflagellates. The densities of each species were analyzed in time and space and in relation to environmental parameters. Some species such as Alexandrium insuetum, Heterocapsa circularisquama, Karlodinium veneficum, Scrippsiella trochoidea, and Ceratium spp. developed high cell concentrations, particularly during a Pseudo-nitzschia calliantha winter bloom. Other species such as Dinophysis caudata, Ostreopsis ovata, Prorocentrum minimum, and Protoperidinium crassipes were rare or appeared in small numbers. Densities of the most abundant species were closely associated with freshwater nutrient-rich inputs during winter, being negatively correlated with temperature and salinity and positively correlated with nitrogen. The spatial distribution of the abundant species exhibited a marked increase towards the inner part of the gulf, close to the main freshwater inputs, whereas some species were mainly concentrated in the dilute surface layer (1 m depth). Examination of the abundance–occupancy relationship revealed that the species more prone to bloom are those with wide spatial distribution and frequent presence throughout the year such as the diatom P. calliantha. Although blooms of cyst-forming species are rarer, an increased risk can be foreseen under favorable resource supply and environmental conditions during winter.     

Cyanobacterial blooms and water quality in Greek waterbodies

The cyanobacterial species composition of nine Greek waterbodies of different type and trophic status was examined during the warm period of the year (May–October). Cyanobacterial water blooms were observed in all waterbodies. Forty-six cyanobacterial taxa were identified, 11 of which are known to be toxic. Eighteen species are reported for the first time in these waterbodies, 8 of which are known to produce toxins. Toxin producing species were found in all of the waterbodies and were primarily dominant in bloom formations (e.g., Microcystis aeruginosa, Anabaena flos-aquae, Aphanizomenon flos-aquae and Cylindrospermopsis raciborskii). Cosmopolitan species (e.g., M. aeruginosa), pantropic (e.g., Anabaenopsis tanganyikae) and holarctic species (e.g., Anabaena flos-aquae) were encountered. Shallow, eutrophic waterbodies had blooms dominated by Microcystis species and were characterized by phytoplankton association M. Anabaena and Aphanizomenon species of association H were dominant in waterbodies with low dissolved inorganic nitrogen and thermal stratification in the summer. Total cyanobacterial biovolumes (CBV) ranged from 7 to 9,507 cm³ m⁻³ and were higher than Alert Level 2 and Guidance Level 2 (10 cm³ m⁻³; World Health Organization; WHO) in seven of the waterbodies. Chlorophyll a concentrations ranged from 6 to 90,000 mg m⁻³ and were higher than Alert Level 2 and Guidance Level 2 (50 mg m⁻³; WHO) in eight of the waterbodies. There is also an elevated risk of acute toxicosis (Guidance Level 3; WHO) in five waterbodies. Water of an undesirable quality, hazardous to humans and animals occurs in several Greek waterbodies.     

Occurrence of cyanobacterial blooms in the baltic sea in relation to environmental conditions

Cyanobacterial blooms occur regularly in summer in central parts of the Baltic Sea. They are mainly composed of Aphanizomenon sp. and Nodularia spumigena. Both species have almost similar ecological requirements and can roughly be considered a uniform functional group. In order to identify factors that might favour bloom development, water quality data from monitoring programmes were compared with bloom distribution. A salinity from 3.8 to 11.5 PSU proved important for the spatial distribution of the bloom development. The bloom's onset was triggered by temperatures approximating 16°C provided that global radiation was > 120 W/m² (daily mean) and wind speed was < 6 m/s. Nutrient concentrations decreased immediately before the bloom. The bloom's development ceased with poor weather conditions characterized by low irradiation or high wind speed.     

First observation of microcystin-LR in pelagic cyanobacterial blooms in the northern Baltic Sea

Cyanobacterial bloom samples from the Gulf of Finland (northern Baltic Sea) were collected in July 2003 and analyzed for microcystins and nodularins, cyanobacterial peptide hepatotoxins, by ELISA, HPLC-UV and LC-MS. The blooms consisted mainly of the genera Nodularia, Anabaena and Aphanizomenon. The main hepatotoxin in the samples was nodularin-R (Nod-R), all the samples also contained demethylnodularin-R. The presence of microcystin-LR was confirmed in three locations out of nine by multiple reactant monitoring on the triple quadrupole mass spectrometer. This is the first reported finding of microcystins in the Baltic Sea from the open sea area. Anabaena was the likely producer of microcystin-LR in the samples.     

Phytoplankton Spring Bloom Intensity Index for the Baltic Sea Estimated for the years 1992 to 2004

We introduce an index for estimating the annual phytoplankton spring bloom intensity in the Baltic Sea. It is based on chlorophyll a estimates calculated from automatically sampled fluorescence and chlorophyll a measurements on board cargo ships from 1992 to 2004. The intensity is described by an index including information on the chlorophyll a concentration and duration of the spring bloom period. In all of the years studied, the spring bloom was most intense in the Gulf of Finland. In the Gulf of Finland and the Northern Baltic Proper there was a slight tendency for the bloom to start earlier in the spring.     

Toxic cyanobacteria (blue-green algae) in Finnish fresh and coastal waters

A survey of the occurrence of toxic blooms of cyanobacteria in Finnish fresh and coastal waters was made during 1985 and 1986. Toxicity of the freeze-dried water bloom samples was tested by mouse-bioassay (i.p.). Forty-four per cent (83/188) of the bloom samples were found to be lethally toxic. Hepatotoxic blooms (54) were almost twice as common as neurotoxic ones (29). Anabaena was the most frequently found genus in toxic and non-toxic blooms and it was present in all neurotoxic samples. Statistical associations were found between hepatotoxicity and incidence of Microcystis aeruginosa, M. viridis, M. wesenbergii, Anabaena flos-aquae and Anabaena spiroides. Neurotoxicity was statistically associated with Anabaena lemmermannii, Anabaena flos-aquae and Gomphosphaeria naegeliana. Isolation of strains of cyanobacteria confirmed the occurrence of hepatotoxic and neurotoxic strains of Anabaena, as well as hepatotoxic strains of Microcystis and Oscillatoria species.Toxic blooms caused cattle poisonings at three different lakes during the study period. Toxic blooms also occurred in drinking water sources. Our study shows that toxic cyanobacteria are more common in Finnish lakes than would be expected on the basis of animal poisonings. The results of this study show the existence of toxic cyanobacteria in Finnish water supplies and the need for their continued study as agents of water based disease.     

Phytoplankton vertical distributions and composition in Baltic Sea cyanobacterial blooms

We studied the vertical structure of the phytoplankton community in two toxic cyanobacterial blooms in the offshore Baltic Sea. In 1994, vertically separated potentially toxic, diazotrophic and mixotrophic species (belonging to Cyanophyceae, Dinophyceae and Prymnesiophyceae) dominated. In 1997, picocyanobacteria, mainly in colonies, made up 40–50% of the total phytoplankton carbon biomass in the top 20 m both day and night. Colony-forming species of picocyanobacteria seem to be occasionally important and hitherto underestimated in the Baltic Sea.We found species-specific depth distribution patterns. Nodularia spumigena and Anabaena spp. were observed mainly above 10 m depth, while Aphanizomenon sp. was mostly found deeper, especially at night. Dinophysis norvegica was only abundant near the seasonal pycnocline and showed very limited diurnal migration. Other flagellates, including small Cryptophyceae and 10 identified Chrysochromulina species, occurred down to 40 m depth. Their vertical migration may help to retrieve nutrients from below the summer pycnocline.We conclude that considerable differences in dominating functional groups may occur between years/bloom stages, and that the vertical distribution pattern of many species is recurring at similar environmental conditions, suggesting species-specific niche-separation.     

Monitoring Karenia brevis blooms in the Gulf of Mexico using satellite ocean color imagery and other data

Harmful algal blooms (HABs) of Karenia brevis are a recurrent problem in the Gulf of Mexico, with nearly annual occurrences on the Florida southwest coast, and fewer occurrences on the northwest Florida and Texas coasts. Beginning in 1999, the National Oceanic and Atmospheric Administration has issued the Gulf of Mexico HAB Bulletins to support state monitoring and management efforts. These bulletins involve analysis of satellite imagery with field and meteorological station data. The effort involves several components or models: (a) monitoring the movement of an algal bloom that has previously been identified as a HAB (type 1 forecast); (b) detecting new blooms as HAB or non-HAB (type 2); (c) predicting the movement of an identified HAB (type 3); (d) predicting conditions favorable for a HAB to occur where blooms have not yet been observed (type 4). The types 1 and 2 involve methods of bloom detection requiring routine remote sensing, especially satellite ocean color imagery and in situ data. Prediction (types 3 and 4) builds on the monitoring capability by using interpretative and numerical modeling. Successful forecasts cover more than 1000 km of coast and require routine input of remotely sensed and in situ data.The data sources used in this effort include ocean color imagery from the Sea-Viewing Wide Field-of-View Sensor/OrbView-2 satellite and processed using coastal-specific algorithms, wind data from coastal and offshore buoys, field observations of bloom location and intensity provided by state agencies, and forecasts from the National Weather Service. The HAB Bulletins began in coordination with the state of Florida in autumn of 1999 and included K. brevis bloom monitoring (type 1), with limited advisories on transport (type 3) and the detection of blooms in new areas (type 2). In autumn 2000, we improved both the transport forecasts and detection capabilities and began prediction of conditions favorable for bloom development (type 4). The HAB Bulletins have had several successes. The state of Florida was advised of the potential for a bloom to occur at the end of September 2000 (type 4), and the state was alerted to the position of blooms in January 2000 and October 2001 in areas that had not been previously sampled (type 3). These successful communications of HAB activity allowed Florida agencies responsible for shellfish management and public health to respond to a rapidly developing event in a timely, efficient manner.     

A survey of toxicity of cyanobacterial blooms in Lake Ladoga and adjacent water bodies

Twentyfive cyanobacterial blooms in Lake Ladoga and adjacent water bodies were studied in the summer of 1990–1992. Toxicity of the water bloom material for mice was detected in 9 cases. The maximal tolerable doses (MTD) of the material extracted from biomass varied within 3–30 mg kg^–1 mouse body weight; 50% lethal doses (LD₅₀) were within 45–125 mg kg^–1. Toxic water blooms were registered in Karelian lakes and in the Neva Bay, Gulf of Finland. Cyanobacterial samples collected on the eastern coast of Lake Ladoga proved to be non-toxic. The species identified in toxic bloom material included Anabaena circinalis, A. flos-aquae, A. lemmermannii, Anabaena sp., Aphanizomenonflos-aquae, Gloeotrichia echinulata, G. pisum, Microcystis aeruginosa and Oscillatoria sp. These data suggest that toxic forms of cyanobacteria are widespread in Karelian lakes belonging to the drainage basin of Lake Ladoga.     

Community development and modes of phosphorus utilization in a late summer ecosystem in the central Gulf of Finland, the Baltic Sea

The development of a filamentous, nitrogen-fixing cyanobacterial bloom was followed during July–August 1990 in a stratified basin in the central Gulf of Finland, Baltic Sea. Hydrography, dissolved inorganic, particulate and total nutrients, chlorophyll a, alkaline phosphatase activity, ³²PO₄-uptake and phytoplankton species were measured. The study period was characterized by wind-induced mixing events, followed by marked nutrient pulses and plankton community responses. Phosphate uptake was highest throughout the study period in the size fraction dominated by bacteria and picocyanobacteria (< 2 µm) and the proportion of uptake in the size fraction 2–10 µm remained low (2–6%). Higher phosphate turnover times were observed in a community showing signs of enhanced heterotrophic activity. The bloom of filamentous, nitrogen-fixing cyanobacteria Aphanizomenon flos-aquae was promoted by a nutrient pulse with an inorganic nutrient ratio (DIN:DIP) of 15. The results show that the quality, frequency and magnitude of the physically forced nutrient pulses have an important role in determining the relative share of the different modes of phosphorus utilization and hence in determining the cyanobacterial bloom intensity and species composition in the Baltic Sea.     

Identification of the coastal zone of the central and eastern Gulf of Finland by numerical modeling,measurements, and remote sensing of chlorophyll <Emphasis Type="Italic">a</Emphasis>

A combination of numerical modeling results with measurement and satellite imagery data was used during the biologically active period for the determination of the coastal zone extent in the central and eastern Gulf of Finland. Adopting the approach that the coastal zone can be identified by the spatial distribution of biotic parameters, spatial variations and gradients of chlorophyll a (chl-a) concentrations were analyzed. The results showed that chl-a concentrations vary in a wide range over the biologically active period. During heavy blooms, the coastal zone may appear occasionally and depend on the spatial distribution of the bloom. On average, clear limits of the coastal zone could be defined for the central and eastern Gulf of Finland. In the central Gulf of Finland, water and material exchange are rather intensive, and the coastal zone is narrower than in the eastern Gulf. In the easternmost part of the Gulf of Finland, chl-a concentrations were permanently high in an area of about 100 km width due to the discharge of the Neva River. The study has shown that gradients of chl-a spatial distribution can be applied for determining limits of the coastal zone extent. The standardized gradient of zero is shown to be a threshold separating the coastal zone (standardized gradients > 0) from the open sea (standardized gradients < 0). Guest editors: J. H. Andersen & D. J. Conley Eutrophication in Coastal Ecosystems: Selected papers from the Second International Symposium on Research and Management of Eutrophication in Coastal Ecosystems, 20–23 June 2006, Nyborg, Denmark     

Toxic and noxious phytoplankton in Big Glory Bay,Stewart Island,New Zealand

Diurnal vertical profile sampling of the water column, during a fish killing bloom of the raphidophycean alga Heterosigma akashiwo, revealed a phytoplankton population otherwise composed almost entirely of a variety of dinoflagellates. Of these Glenodinium danicum, Dinophysis acuta, Polykrikos schwartzii, Ceratium furca and Gyrodinium spirale were predominant. The distribution of the major species within the phytoplankton were documented and evidence of synchronous vertical migration of H. akashiwo, G. danicum and P. schwartzii was observed. Extracts of shellfish obtained during the bloom and tested by mouse bioassay showed no PSP toxicity but a marginal degree of DSP toxicity. During a subsequent one year phytoplankton monitoring programme another potentially noxious species (Chaetoceros convolutus) appeared and the seasonal reoccurrence of species present during the bloom (e.g. H. akashiwo) was observed. Important year to year differences in the summer phytoplankton (diatom versus flagellate dominated populations) were apparent and analysis of climate data showed that these differences related to different weather conditions prevailing during the two summer periods sampled. The data suggest the fish killing bloom was giving a chance to develop by a prolonged period of warm, calm weather (during which several heavy rainfall events occurred) leading to stable hydrographic conditions (i.e. stratification) and an increase in the retention time of water within the bay.     

‘Ferry-Boxes’ and data stations for improved monitoring and resolution of eutrophication-related processes: application in Southampton Water UK,a temperate latitude hypernutrified estuary

To provide detailed observations of algal bloom development in Southampton Water which is a hypernutrified, macro-tidal estuary (mean tidal range 3.2 m, low suspended load <100 g m^–3), a ferry running between Southampton and Cowes on the Isle of Wight, was fitted with an instrument package (Ferry-Box). Measurements were made of temperature, conductivity, turbidity, and chlorophyll-fluorescence at a data rate of 1Hz. For comparison a data station which measured the same variables was operated at a fixed site in the estuary. In 1999 the Ferry-Box achieved reliable operation with a data return over 95%, for the fixed data station the return was 92%. From this data spatial and temporal variations in chlorophyll a concentrations have been mapped. The maps show the development of blooms in different areas of the estuary, through the spring and summer, in relation to tidal and weather conditions. In 1999 conditions were such that the spring bloom increased in intensity through a spring tide (maximum chlorophyll a 55 mg m^–3), which coincided with calm weather with high light levels (irradiance). This was followed by a sequence of seven blooms, the development of which can be related to changes in the tidal energy, irradiance and nutrient supply.     

Development and application of LSU rRNA probes for Karenia brevis in the Gulf of Mexico, USA

The brevetoxin producing dinoflagellate, Karenia brevis, is the target of several monitoring and research programs in the Gulf of Mexico, where it forms extensive and frequently long-lived annual blooms that can cause human intoxication and fish kills, as well as severe economic losses to coastal communities. Rapid, reliable methods for the detection and enumeration of K. brevis cells, as well as their discrimination from morphologically similar species, are valuable tools for managers and scientists alike. Our aim was to produce a species-specific molecular probe that would serve as a tool to facilitate the efficient and reliable detection of K. brevis in the Gulf of Mexico. We sequenced a fragment of the large-subunit ribosomal RNA gene (LSU rDNA) from five K. brevis cultures isolated from the Texas Gulf coast, the Florida Gulf coast, and the Atlantic coast of Florida, and detected no differences among these isolates. A consensus sequence was thus compiled and compared to a previously published sequence from Karenia mikimotoi, the closest known phylogenetic relative to K. brevis, for the purpose of identifying unique K. brevis signature sequences. Fluorescently-labeled (FITC) oligonucleotide probes targeting these regions of the K. brevis LSU rRNA were designed to include at least two base pair differences, as compared to K. mikimotoi. Among seven probes designed, one uniquely identified all K. brevis isolates to the exclusion of all other species tested (Kbprobe-7), including a Gulf of Mexico K. mikimotoi isolate (Sarasota, FL) and several additional Gymnodinium species, as well as other dinoflagellate, diatom, and raphidophyte taxa. Importantly, K. brevis cells in samples taken during a 2001 bloom, fixed with a mixture of modified saline ethanol and 10% formalin, and stored at 4 °C for 7 months were successfully labeled with Kbprobe-7. In addition, preliminary analysis of labeled cells by flow cytometry revealed that K. brevis could be distinguished from K. mikimotoi in solution, suggesting other potential applications of this probe.     

Disturbance events affecting phytoplankton biomass,composition and species diversity in a shallow,eutrophic, temperate lake

The seasonal changes in phytoplankton biomass and species diversity in a shallow, eutrophic Danish lake are described and related to different disturbance events acting on the phytoplankton community.Both the spring diatom maximum and the summer bloom of the filamentous blue-green alga, Aphanizomenon flos-aquae (L.) Ralfs, coincided with low values of phytoplankton species diversity and equitability. Diatom collapse was mainly due to internal modifications as nutrient depletion (Si, P) caused by rapid growth of phytoplankton, and increased grazing activity from zooplankton. A large population of Daphnia longispina O.F. Müller in June effectively removed smaller algal competitors, thus favouring the development of a huge summer bloom (140 mm³ l^–1) of Aphanizomenon flos-aquae. Heavy rainfall and storms in late July increased the loss of Apahnizomenon by out-flow and disturbed the stratification of the lake. These events caused a marked decline in phytoplankton biomass but had no effect on species diversity. A second storm period in late August circulated the lake completely and was followed by a rapid increase in phytoplankton diversity, and a change in the phytoplankton community structure from dominance of large, slow-growing K-selected species (Aphanizomenon) to small, fast-growing r-selected species (cryptomonads).