Seasonal dynamics of phytoplankton in the Gulf of Aqaba, Red Sea
- First Online:
- Cite this article as:
- Al-Najjar, T., Badran, M.I., Richter, C. et al. Hydrobiologia (2007) 579: 69. doi:10.1007/s10750-006-0365-z
- 240 Views
Seawater samples were collected biweekly from the northern Gulf of Aqaba, Red Sea, for Phytoplankton analysis during the period May 1998 to October 1999. Microscopic counts and HPLC methods were employed. Procaryotic and eucaryotic ultraplankton dominated throughout most of the year, with larger nano- and microplankton making up only 5% of the photosynthetic biomass. Moderate seasonal variations in the 0–125 m integrated Chl a contrasted with a pronounced seasonal succession of the major taxonomic groups, reflecting the changes in the density stratification of the water column: Prochlorococcus dominated during the stratified summer period and were almost absent in winter. Chlorophyceae and Cryptophyceae were dominant during winter mixing but scarce or absent during summer. Diatoms and Synechococcus showed sharp and moderate biomass peaks in late winter and spring respectively, but remained at only low Chl a levels for the rest of the year. Chrysophyceae, Prymnesiophyceae and the scarce Dinophyceae showed no clear seasonal distribution pattern. The implications of alternating procaryotic and eucaryote dominated algal communities for the Red Sea pelagic food web are discussed.
KeywordsChlorophyll aPigmentsNutrientsHPLCCHEMTAX GulfAqabaRed Sea
The Red Sea owes its name to the occurrence of intense phytoplankton blooms, especially of the colony-forming filamentous cyanobacteria Trichodesmium (Belogorskaya, 1970), discolouring its surface waters. However, plankton blooms seem to be the exception rather than the rule in this oligotrophic body of water (Badran & Foster, 1998). The Gulf of Aqaba, forming the northeastern extension of the Red Sea, is much like a 1:10 replica of the Red Sea proper, featuring the same general shape along an axial trough and restricted exchange across a shallow sill (Reiss & Hottinger, 1984). Due to its unique depth-to-width ratio (max. depth >1800 m, max. width <25 km: Morcos, 1970) and oligotrophic nature, the steep-sided gulf is also a model ocean, where similar processes governing pelagic production in remote oceanic subtropical gyres may be studied at rowing distance from shore (Reiss & Hottinger, 1984; Hulings, 1989; Hempel & Richter, 2002).
Phytoplankton biomass is generally low, rarely exceeding 1 mg Chl a m−3 (Genin et al., 1995; Badran & Foster, 1998; Badran, 2001; Rasheed et al., 2002; Al-Najjar et al., 2003) with Trichodesmium and other net- (>65 μm) or micro- (>20 μm) phytoplankton comprising only a minor fraction of total photosynthetic biomass (Kimor, 1971; Levanon-Spanier et al., 1979; Lindell & Post, 1995). The bulk of the biomass is comprised of ultraplankton (<8 μm in cell diameter) belonging to three major taxonomic groups: cyanobacteria, prochlorophytes and eucaryotic algae (Lindell & Post, 1995). Although pico- and small nanoeucaryotes have been shown to be abundant during the winter mixing season (Lindell & Post, 1995), their taxonomic composition is not well known yet. There have been separate descriptions of the seasonal dynamics of the micro-, nano- and picophytoplankton communities in the Red Sea (Kimor, 1971; Levanon-Spanier et al., 1979; Lindell & Post, 1995), but a comprehensive simultaneous assessment encompassing the entire phytoplankton size range is still lacking.
Advanced HPLC techniques now allow improved separation and quantification of the 40–50 pigments from all algal classes (Wright et al., 1991) including the very small and fragile cell types (Li et al., 1983a; Platt et al., 1983; Iturriaga & Mitchell, 1986) whose identification and counting by microscopy or flow cytometry are difficult. As many individual algal pigments or pigment combinations and ratios are specific to higher taxa, HPLC analysis can be used diagnostically to identify the presence of individual phytoplankton groups in natural samples. As opposed to the tedious and sophisticated enumeration by epifluoresence microscopy or electron microscopy, HPLC analysis requires no specific preparation techniques for the various size groups, enabling bulk analysis of the samples at the class level (Hooks et al., 1988). This advantage as well as the speed, sensitivity and accuracy of the method made HPLC a powerful tool for the quantitative assessment of phytoplankton communities (e.g. Gieskes & Kraay, 1983, 1986; Klein & Sournia, 1987; Bidigare et al., 1990; Althuis et al., 1994).
Quantitative tests of the agreement between pigment analysis and microscopic analysis in natural waters are still scarce (Wilhelm et al., 1991; Tester et al., 1995), due to difficulties in precisely converting microscopic estimates of phytoplankton cell number, volume and pigment concentrations into carbon. Moreover, there are still some concerns in the general use of diagnostic pigments as tracers of specific algal groups. For example, prasinoxanthin occurs only in prasinophytes, but not all prasinophytes contain prasinoxanthin (Egeland et al., 1995). Some prymnesiophytes do not contain the diagnostic pigment 19 hexanoyloxyfucoxanthin (Jeffrey & Wright, 1994), and some dinoflagellates contain 19 - hexanoyloxyfucoxanthin instead of the marker pigment peridinin (Tangen & Björnland, 1981).
In this study, we combined microscopic and new high performance liquid chromatography (HPLC) techniques to study the seasonal phytoplankton dynamics in the Gulf of Aqaba, Red Sea, in order to understand with finer taxonomic and temporal resolution the roles of the various groups of primary producers in this and, by comparison, other subtropical oligotrophic systems.
Material and methods
Seawater for phytoplankton analysis was collected from the Jordanian section of the Gulf of Aqaba, about 3 km offshore of the Marine Science Station (2927.362 N, 3457.238 E). Daytime biweekly samples were taken from August 1998 through to October 1999. Once a month, night samples were also collected from the same station. Water was sampled with 5-liter plastic Niskin bottles at discrete depth 0, 25, 50, 75, 100 and 125 m, transferred to acid-cleaned 5 l plastic containers, and kept cool and shaded during the transport to the laboratory. Water temperature, depth and salinity were recorded concurrently using a self-recording Ocean Sensor OS 200 CTD. Water density was calculated according to the UNESCO (1981). We calculated the vertical density gradient (Dst) between 25 m and 100 m (g m−3 m−1, or g m−4) as a measure for the stability of the water column.
Microscopic determination of the phytoplankton
Subsamples of two liters were fixed by buffered formalin (final concentration 3 %) and concentrated in cylindrical beakers by sedimentation. Serial sedimentation steps were taken for one month, discarding half of the volume every 7 days. After the final step the volume was corrected to 200 ml and the 10-fold concentrated samples were kept for identification. Samples of 0, 25 and 50 m were pooled to represent the upper euphotic zone and, samples of 75, 100 and 125 m were pooled to represent the lower euphotic zone. Cells were enumerated with an inverted microscope following the Utermöhl (1958) method. Where numbers were sufficient, 100 cells were counted per taxon, yielding 95% confidence levels of about ±20% (Lund et al., 1958). However, for the rare taxa counting of the entire chamber resulted in much lower numbers.
Chlorophyll a fluorescence and nutrient analysis
Chl a determination was carried out fluorometrically as described by Arar and Collins (1997). One liter of seawater was filtered through a 0.45 μm cellulose acetate membrane filter. The filter was then placed in a glass tube wrapped with aluminum foil. 10 ml of 90% acetone was added into the test tube. The membrane filter was ground using a tissue grinder and then kept in a refrigerator at 4°C over night. The mixture was then centrifuged for 5 min at 5000 rpm, and Chl a measured in a13 mm glass test tube using the direct concentration calibration method. 90% acetone was used as a blank.
HPLC Pigment analysis
Another subsample of 2 l was filtered on GF/F-filters (Whatman, 25 mm diameter) with a vacuum of max 0.2 bar. Filters were folded, put into Eppendorff-caps and kept in an ultrafeezer at −80°C.
Filters were ground in 2.5 ml acetone (90 %), together with a mixture of glass beads of 2 and 4 mm diameter in an 11 ml-plastic-tube shaken by a cell homogenizer. The holder for the tube was cooled in a freezer before use. The extract was centrifuged for ten minutes at 5000 rpm, cooled at −10°C. Then the supernatants were taken by a medical syringe and put back into an Eppendorff-cap through a pre filter of 0.2 μm pore size.
Aliquots (110 μl) of the extract were loaded into a thermo separations auto sampler (capable of cooling pigment extracts to 2°C) and mixed with (1:1 v/v) 1 M ammonium acetate (1:1 v/v). Up to 170 μl of the mixture were injected into an HPLC system. The column was a 3 μm Shando Hypersil MOS2 (endcapped), C-8 (6.2–6.8 % carbons), 120 Ao pore size, and 100 × 4.6 mm and maintained at 30°C.
Pigments were separated at a flow rate of 1 ml min−1 by a linear gradient programmed as follows (minutes; % solvent A; % solvent B): (0;75;25), (1;50;50), (20;30;70), (25;0;100), (32;0;100). The column was then reconditioned to original conditions over a further 7 min. Solvent A consisted of 70:30 (v: v) methanol: 1 M ammonium acetate and solvent B was 100% Methanol.
Pigments were detected by absorbance at 440 nm using a Shimadzu SPD-6AV spectrophotometric detector. Chromatographic data were processed using the Pye-Unicam 4880 software and pigments identified by retention time and on-line visible spectroscopy using a Waters 990 diode array detector. Chl a and b standards were obtained from Sigma Chemical Co., and standards for other pigments were purchased from the Water Quality Institute (VKI), Hørsholm, Denmark. The detection limit for all analysis was 1.0 ng l−1.
Data interpretation by CHEMTAX
Chemical Taxonomy (CHEMTAX) is a matrix-factorization program for calculating algal class abundance from concentrations of algal chemosystematic marker photopigments (chlorophyll’s and carotenoids) (Mackey et al., 1996, 1997; Wright et al., 1996). Input for the program consists of a raw-data matrix of photopigments concentrations obtained by HPLC analyses and initial pigment-ratio file. The data matrix is subjected to a factor-minimization algorithm that calculates a best fit pigment ratio matrix and a final phytoplankton class-composition matrix. The class-composition matrix can be expressed as relative or absolute values for specified photopigments. The absolute chlorophyll a contribution of each class is particularly useful because it partitions the total chlorophyll a into major phytoplankton groups. Full discussion, validation, and sensitivity analyses of CEMTAX are provided in Mackey et al. (1996, 1997) and Wright et al. (1996).
Density and nutrients
Phytoplankton biomass as determined by in vitro fluorescence analysis overestimated Chl a by up to more than 100%, particularly during the stratified period (Fig. 2a). Fluorescence data showed better correspondence if both, mono- and di-vinyl derivatives of Chl a (Chl a1 + Chl a2) were considered, but the overall correlation between fluorescence and HPLC values was still weak (R2= 0.33).
Procaryotes dominated throughout the summer, comprising up to more than two thirds of the photosynthetic biomass: Prochlorophytes (Prochlorococcus) alone made up roughly half of the total algal stock in summer, but only less than 5% of all pigments in winter. Rapid build-up and decline of prochlorophytes occurred during the spring and fall transition periods, respectively.
Cyanobacteria were most important in spring and early summer (10–20% of total biomass, Fig. 2c; or up to 7 mg Chl a m−2, Fig. 3), progressively declining to <5% of total Chl a at the peak of mixing period in March.
Eucaryotic algae abounded in winter, accounting for 85–95% of the phytoplankton biomass (Fig. 2c): Already in fall, Cryptophyceae and Chlorophyceae started to build up sizeable populations, each reaching ~25% of the phytoplankton share in winter, and declining thereafter. Cryptophyceae biomass peaked around the end of December at 8 mg Chl a m−2 (Fig. 3), tailing off in the course of winter and spring. Chlorophyceae bloomed twice, following the Cryptophyceae in early January (11 mg Chl a m−2, Fig. 3), and again with the onset of spring in April (10 mg Chl a m−2, Fig. 3). While Cryptophyceae had virtually disappeared by June, Chlorophyceae persisted over summer at a fairly constant ~12% of total crop. Prymnesiophyceae and Chrysophyceae showed little seasonal variability, comprising 10–14% and 13–20% of the phytoplankton, respectively, throughout most of the year (Fig. 2c). The highest concentrations for the former were recorded in December and March, the maximum for the latter in June (Fig. 3).
Dinoflagellates and diatoms together made up less than 5% of the phytoplankton biomass, except for one occasion in March, where a diatom bloom (20 mg Chl a m−2, Fig. 3) contributed to nearly 50% of the total (Fig. 2c). The bloom was preceded by an enrichment of NO3 in the upper 0–75 m of the water column (Fig. 2b).
Nano- and Microphytoplankton
Utermöhl analysis showed that for most of the year the nano- and microphytoplankton community was dominated by small (<20 μm) dinoflagellates, outnumbering the other phytoplankton groups by the factor 3. Much of the variability among this group was due to the 10–20 μm cells displaying strong 2- to 4-month periodicity in abundance (see Supplementary material).
Among the diatoms, Thalassiosira was prevalent throughout the year, with maximum concentrations in winter (100 × 103 l−1 in January, see Supplementary material). The larger Rhizosolenia and Coscinodiscus occurred at lower densities, with maxima in the summer-fall transition period (September through November). Nitzschia distribution was variable throughout, with abundances in the range of 5 × 103–50 × 103 l−1 following a saw-tooth pattern. Chain-forming Chaetoceros were patchily distributed, where brief blooms up to 400 × 103 cells l−1 (March, see Supplementary material) alternated with 2–3 month periods below detection.
Green algae showed also an erratic distribution with concentrations up to 70 × 103 cells l−1 (see Supplementary material). The filamentous cyanobacteria Trichodesmium spp. showed up during the summer/fall (October) and spring/summer (June) transitions, respectively, with up to 40 × 103 cells l−1, being absent for the rest of the study (see Supplementary material).
Succession in the Gulf of Aqaba, Red Sea
The most striking feature of our study is the unambiguous and almost sinusoidal pattern of seasonality in phytoplankton succession, where eucaryote and prokaryote dominated communities were observed to alternate with the wax and wane of vertical mixing and stratification of the water column. The observed symmetry in the fractional contributions of the various taxonomic groups to total phytoplankton biomass is all the more surprising. Short term irregularities in the overall integrated Chl a and nutrients in several sampling events outweighed the long term seasonal changes.
Earlier investigations highlighting the importance of water column stratification and mixing on phytoplankton in the Gulf of Aqaba were based on fluorometric (Genin et al., 1995; Rasheed et al., 2002) and spectrophotometric determinations of bulk Chl a (Badran, 2001; Wahbah & Zughul, 2001) or examinations of only a small part (nano- and microplankton) of the phytoplankton community (Kimor, 1971, 1976; Winter et al., 1976; Kimor & Golandsky, 1977; Kimor et al., 1992). In terms of taxonomic resolution the only study of ultraplankton succession (Lindell & Post, 1995) did not differentiate between the various groups of eukaryotic algae.
Seasonal changes in cellular Chl a levels due to photo acclimation have been shown to affect phytoplankton biomass estimates, leading to an overestimate in shade-adapted algae in well mixed waters (Campbell et al., 1997; Dusenberry et al., 2000). On the other hand, earlier studies from the Gulf of Aqaba suggest that the 125 m integrated Chl a values may in fact underestimate phytoplankton biomass in winter, where a considerable fraction of the algal populations may be mixed below the euphotic zone (Kimor & Golandsky, 1977; Lindell & Post, 1995). Both factors appear to even out to some extent, and Lindell & Post (1995) found good agreement between cell numbers and Chl a values.
The combination of HPLC and microscopic analysis in our study allows a refined classification of phytoplankton distribution patterns with regard to the mixing regime, as follows: (i) taxa exhibiting consistent positive responses to winter mixing (Cryptophyceae, Chlorophyceae); (ii) taxa with short term positive responses to deep mixing (diatoms); (iii) taxa showing negative responses to winter mixing/positive responses to summer stratification (Prochlorophyceae); and (iv) taxa without or with unclear seasonality (Dinophyceae, Prymnesiophyceae and Chrysophyceae) (Li et al., 1983b; Zohary & Robarts, 1998).
The importance of the density structure on phytoplankton succession is reflected in the good association between group-specific Chl a concentrations and the vertical density gradient. Four out of the eight groups showed positive response to mixing, one (Prochlorococcus) increased in biomass with increasing summer stratification and the other three showed weak response. While the general sequence of succession is in general agreement with previous results on ultraplankton (Lindell & Post, 1995) and microplankton (Kimor & Golandsky, 1977) succession in the Gulf of Aqaba, there are differences in the timing, shape and magnitude of the excursions in the stocks of the major phytoplankton groups, which may help elucidate in more detail the factors governing ultraplankton community structure in the Red Sea.
In 1992–93, winter mixing in the Gulf of Aqaba was deep about 600 m, triggering a succession of eukaryotes, Synechococcus and Prochlorococcus in well-mixed, transitional and stratified waters, respectively (Lindell & Post, 1995). Probably as a result of the early and deep mixing in a cold 1992 winter and normal 1993 summer, the growing season for Prochlorococcus appears protracted and skewed compared to our study. Standing stocks were very low already in September 1992, and undetectable between mid-December and early spring, with re-establishment of the population as late as the end of July through to September 1993. Whereas peak concentrations of Prochlorococcus are comparable between 1992/93 and 1998/99, around 12 mg DV-Chl a m−2 (note that the legend in Supplementary material of Lindell & Post (1995) should read mg instead of μg), they were attained much earlier in our study (mid-June) and persisted well into November (Fig. 3). An important difference in our study is also the observation that Prochlorococcus, although absent during the winter and early spring of 1992/93, constituted a viable yet small fraction of the phytoplankton during the 1998/99 winter period.
Synechococcus played a minor role in 1998/99, compared to 1992/93 where it dominated phytoplankton abundance throughout most of the year (October through January and mid-April through late June) including more or less pronounced blooms during the transition periods in fall and spring (Lindell & Post, 1995; Genin et al., 1995). While the massive build-up of Synechococcus during spring 1993 was associated with an abrupt break-down of the eucaryote populations which had gradually accumulated over winter, the moderate and only brief spring increase of Synechococcus in 1999 did not preempt the development of other phytoplankton groups, as shown by the parallel population increases in Prochlorococcus and chlorophyte eukaryotes. No fall increase in Synechococcus biomass was detected in our study. There was also no significant correlation of Synechococcus pigments with any of the chemical parameters, which suggests that its appearance was more likely controlled by biological factors. However, recent experiments suggest avoidance of Red Sea Synechococcus by actively grazing protozoa (Sommer et al., 2002), raising speculations that the population may be kept in check by viral lysis.
With regard to the eucaryotic algae of the Gulf of Aqaba, earlier reports showed enhanced densities during the winter mixing period (Kimor & Golandski, 1977; Lindell & Post, 1995; Yahel et al., 1998) but, as the higher taxonomic resolution in our study reveals, this does not apply to all groups. While some of the taxa (Cryptophyceae and Chlorophyceae) were shown to abound in winter, to the extent of being virtually absent during the stratified season, other (Prymnesiophyceae, Chrysophyceae) occured year-round in sizeable stocks. Overall, eukaryotes appear to be more important than previously thought, generally contributing more than 40%, at times up to 95% of total phytoplankton stocks.
The disparity of the distribution patterns again demonstrates that although the aforementioned groups are all composed of small (<5 μm) microscopically similar flagellated cells, they are subjected to rather different environmental and biotic controls.
This seems to hold also true for the nano- and microplanktonic dinoflagellates and diatoms, whose year-round presence and low concentrations on the one hand, and bloom-forming capacity on the other, suggest differential control by nutrients/light and grazing. Recent experimental work shows that indeed both, bottom-up and top-down control combine to explain the general scarcity of nano- and microphytoplankton in the Red Sea (Sommer, 2000).
In this regard, the occurrence of microplankton blooms suggest that both, temporarilynutrient-enhanced growth and reduced grazing mortality by mesozooplankton (Sommer et al., 2002) allow chain-forming diatoms (Chaetoceros, this study; Thalassiosira, Kimor & Golandski, 1977) to temporarily escape the double-lock of nutrient and grazing limitation during periods of strongest mixing, albeit only for a short period of time. The ephemeral nature and precipitous decline of diatom blooms in the Gulf of Aqaba remains enigmatic, considering that it takes place in spite of elevated levels of NO3 and SiO2 in the water column, and in spite of low overall mesozooplankton grazing impact (A. Cornils, unpubl. data). The maximum depth of the 1975 diatom bloom was at 300 m (Kimor & Golandsky, 1977), i.e. below the critical depth for diatoms, suggesting that the bloom was formed elsewhere. Blooms could thus have formed over the shallow shelf, and reached this depth by lateral advection and synchronous sinking (Niemann et al., 2004) or plume convection (Woelk & Quadfasel, 1996). The general scarcity of diatoms in spite of year-round high levels of SiO2 (Klinker et al., 1978; Badran & Foster, 1998; Badran, 2001; Rasheed et al., 2002) and high concentrations of nutrients in coastal and coral reef waters (Badran & Foster, 1998; Rasheed et al., 2002) could thus be attributed to the extremely restricted aerial extent of shelves in the deep and narrow Gulf of Aqaba.
Grazing and competition for the scarce resources may explain also the high diversity of Dinophyceae, reportedly the group with the largest number of species in the Gulf of Aqaba (Kimor, 1976) displaying peculiar adaptations, feeding modes and associations with other organisms.
Another factor which may have contributed to the observed inter-annual differences in phytoplankton succession in the Gulf of Aqaba is a possible increase in nutrients between the years, favoring eucaryote algae. Nitrate values in our study appear to be almost twice as high (0.3 μmol l−1) as in 1992/93 (0.1–0.2 μmol l−1, Lindell & Post, 1995). This may be due to changes in the depth and volume flow of Red Sea intermediate water into the Gulf of Aqaba (Manasrah et al., 2004). These waters can be a main source of new nutrients and their inflow has been shown to vary between the years (Plähn et al., 2002). Another possible cause currently under debate is the possible eutrophication of the northern gulf due to rapid urban development and extension of fish-farming activities (Zimmer, 2001). Long-term data on the deep water nutrient concentrations of the northern gulf seem to support the trend of nutrient accumulation in deep waters over the last years (Erez et al., 2000).
It thus seems that as mixing depth increases, the nutrient environment changes from chemostat to batch-culture like conditions (Landry & Kirchman, 2002), triggering a temporal sequence of adjustments in the major phytoplankton stocks, where the amplitude of the seasonal excursions seems to vary linearly with maximum mixed layer depth (Fig. 4). Due to the large year-to-year variations in the winter mixing depth in the Gulf of Aqaba (~300 to >800 m, Genin et al., 1995; Farstey et al., 2002), the inter-annual variations in phytoplankton succession may be very large, in fact much larger than the reported inter-ocean differences between the equatorial Pacific and subtropical North Atlantic (Fig. 4). Thus, primary production may be shifted quite drastically from strong bottom-up control in one year (e.g. 1992/93) to mixed bottom-up/top-down control in another (1998/99), in response to external forcing (winter climate).
Top down vs bottom up control
The small spatial and temporal variability in picoplankton communities in oligotrophic oceanic ecosystems has been attributed to the double lock of protistan grazing and viral lysis (Banse, 1992; Fuhrmann, 1999). While the importance of viruses in Red Sea waters remains to be ascertained, the limited data on zooplankton grazing in the Gulf of Aqaba and wider Red Sea suggests a strong grazing control of heterotrophic nanoflagellates on bacteria and of microzooplankton on medium-sized (>5 μm) phytoplankton, but a weak control of meso- and macro-zooplankton (copepods, appendicularians, salps) on the other phytoplankton groups (Al-Najjar, 2000; Sommer, 2000; Post et al., 2002; Sommer et al., 2002). Synechococcus did not appear susceptible to grazing in the Red Sea (Sommer et al., 2002), in contrast to reports of high grazing elsewhere (Burkill et al., 1993, Sakka et al., 2000). The regional differences in grazing may be attributed in part to differences in the protistan community. For example, Synechococcus was shown to sustain ciliate growth (Christaki et al., 1999), but appeared to solicit no growth response in heterotrophic flagellates (Guillou et al., 2001), the latter depending largely on Prochlorococcus and bacteria. Information on the microheterotrophic community, grazing on Prochlorococcus and the role of viruses in the Red Sea are still lacking.
As opposed to the direct effects of grazing and viral lysis on phytoplankton communities, vertical mixing affects phytoplankton community structure indirectly by controlling other environmental factors, such as nutrient concentrations and light levels (Lewis, 1978; Smayda, 1980). Mixing may also affect competition between groups (Kemp & Mitch, 1979) and affect phytoplankton characteristics such as cellular metabolism, sinking rate, and cell integrity of fragile groups (Margalef, 1978; Smayda, 1980).
Near total exhaustion of PO4 over the entire water column indicates that nutrient limitation is imminent in the Gulf of Aqaba (Fig. 1a), as indicated also by the succession of small (<5 μm) to very small (<1 μm) cells (Lindell & Post, 1995; Yahel et al., 1998, 2003), with increasing surface area to volume and higher specific uptake rates (Eppley et al., 1969; Raven, 1986). Concerning competition between osmotrophs for PO4, Moutin et al. (2002) and Tanaka et al. (2003) have showed that cyanobacteria (Synechococcus) have the highest specific affinity for PO4 uptake and the highest Vmax in the Mediterranean, and the appearance of cyanobacteria with N2-fixation capability, supplying as much as 6% of the total N-budget of the Gulf (Lindell & Post, 1995).
In our study we found only a poor statistical correlation between the standing stocks of some phytoplankton groups and ambient nutrient concentrations. This mismatch can be explained in part by the observed temporal lag, between nutrient enhancement following mixing and the phytoplankton response to the nutrient concentrations made available. At the initial stages of mixing when nutrient concentrations are not yet sufficiently high, the nutrient supply across the eroding nutricline is balanced by the demand of the phytoplankton community. Water temperature and irradiance are still relatively high which leads to the rapid increase of eucaryote biomass, particularly cryptophytes and chlorophytes, during the late fall early winter (Fig. 2c and 3). In the later stages of succession the poor match of nutrient concentrations with the biomass of the various phytoplankton groups is very likely explained by intense recycling of material in the euphotic zone (Badran, 2001). Rapid recycling and bacterial remineralization within the euphotic zone also explain the lack of nutrient-exhaustion (Fig. 1a) and the overall high NH4:TIN-ratio (Fig. 1b) (Grossart & Simon, 2002). High-amplitude internal waves recently observed in the Gulf of Aqaba (Plähn et al., 2002), limits of uncertainty in nutrient analysis especially that concentrations for most of the year were close to the detection limits and discreetness of the sampling program as opposed to the continuity of the physical, chemical and biological interactions in the water column may all complicate the picture. Real additional nutrient supply may occur through turbulent mixing of nutrient-rich waters from below that, although can be rapidly manifested biologically into phytoplankton production, its results are not instantaneous.
The only nutrient that showed some association with the phytoplankton community was NO2. This is a short half life time intermediate species, easy to analyze with very small analytical errors, insignificant even in comparison with the short term small concentration changes. It correlated positively with Cryptophyceae and negatively with Prochlorophyceae. The covariance of Cryptophyceae and NO2 is in line with recent field and laboratory results suggesting that the winter accumulation of NO2 in the euphotic zone is due to nutrient-stimulated phytoplankton growth (Al-Qutob et al., 2002). The negative correlation of Prochlorococcus with NO2, by contrast, indicates that the vernal increase of Prochlorococcus may rely in part on the assimilation of NO2 produced by the winter-spring plankton community. Several lines of evidence support the latter hypothesis: the vicinity of the vernal Prochlorococcus maximum at 80–90 m (Lindell & Post, 1995) to the NO2 maximum at 100 m in the stratified Gulf of Aqaba (Badran, 2001; Al-Qutob, et al., 2002); the occurrence of low-light ecotypes of Prochlorococcus at these depths (Moore et al., 1998) capable of utilizing NO2 (but not NO3) as N source (Moore et al., 2002); and the successful culture of Red Sea Prochlorococcus consortia on NO2 as the sole N source for several months (Moore et al., 2002). It is conceivable that NO2 exhaustion at the end of the summer period (Badran, 2001) may be associated with the decline of the Prochlorococcus population, even before the onset of the mixing period. Differences in the nutrient history of the water column may thus help to explain the pronounced inter-annual differences in Prochlorococcus stocks observed in late summer 1992/93 and 1998/99, stratification and light conditions being the same.
We gratefully acknowledge the efforts of the Aqaba Marine Science Station, the Institute für Meereskunde, Kiel and the Center for Marine Tropical Ecology, Bremen in supporting this work. Thanks are due to Gotthilf Hempel for valuable comments, support and advice and to Thomas Hansen and Kerstin Nachtigall for technical support. Our work forms part of the Red Sea Program (RSP) funded by the German Ministry of Education and Science (grant nos. 03F0151A, 03F0245A).