Space-time variability of phytoplankton structure and diversity in the north-western part of the Arabian Gulf ( Kuwait ’ s waters )

Studies of the phytoplankton community were conducted in the north-western Arabian Gulf in 2005 and 2006. Seven stations throughout Kuwait’s waters were sampled. Th e infl uence of nutrient-rich freshwaters from the Shatt al-Arab resulted in high phytoplankton productivity characterized by high species diversity with a strong dominance of diatoms, especially in northern Kuwait. Phytoplankton species richness gradually increased from north to south. Spatial distribution of both total abundance and biomass of phytoplankton indicated signifi cant diff erences in species structure and size spectrum of the microalgae. Th e analysis of the temporal and spatial phytoplankton variability (distribution of total abundance and biomass, similarity of species compositions and local community structure) indicated that Kuwait’s northern waters diff ered from areas further south in terms of phytoplankton structure and temporal and spatial variability. Environmental heterogeneity is mainly attributed to the infl uence of the Shatt al-Arab system, which aff ects the temporal and spatial variability of the phytoplankton community.


Space-time variability of phytoplankton structure
and diversity in the north-western part of the Arabian Gulf (Kuwait's waters) Introduction Th e ecology of phytoplankton in the Arabian Gulf has been studied during the last few decades and is relatively well known (e.g.Al-Kaisi 1976, Jacob et al. 1979, Subba Rao et al. 1999, Al-Yamani et al. 2004).Long-term studies of temporal and spatial distributions and the eff ects of physical eff ects on the phytoplankton community in Kuwait's waters have not been reported.Th e main freshwater infl ow into the northern Arabian Gulf is from the Shatt al-Arab River.Seasonal freshwater supply from the Shatt al-Arab has local eff ects on the Gulf 's marine environment, especially on Kuwait's waters.Th e phytoplankton community in the Arabian Gulf is heterogeneous, with species compositions diff ering among localities (Al-Yamani et al. 2004).Th e main objective of this study was to describe the spatial and temporal variability of phytoplankton diversity, species composition and abundance in Kuwaiti territorial waters.

Methods
Daytime phytoplankton surveys in Kuwaiti waters were conducted twice a month from October 2005 through September 2006 at seven stations (Fig. 1).
One-liter samples from the surface layer (1 m depth) were collected by 5-liter Niskin bottles and preserved with acidifi ed Lugol solution.After full sedimentation during at least four weeks, the top water volume was carefully siphoned off without disturbing the precipitated algae (using rubber a hose with curved end).Th e Utermöhl sedimentation method was used for quantitative analysis of the Niskin bottle samples (Utermöhl 1958).Th e concentrated sample was well shaken and an aliquot of 25 ml from each sample was placed in the standard Utermöhl settling counting chamber.After sedimentation during a 24 h period in a well-covered dark desiccator, the area of the settling chamber was examined with a Leica DMIL inverted microscope at ×200 to ×400 magnifi cations.For phytoplankton enumeration, the appropriate area of the chamber was scanned, depending on the abundance of each species.Randomly-selected viewing fi elds were examined for very abundant phytoplankton species, whereas the complete chamber area was scanned for less abundant species.Th e abundance for each phytoplankton taxon was calculated as the number of cells per liter.In total, 76 Niskin bottle samples were examined.
Th e SeaBird SBE-19 CTD profi ler equipped with a Seapoint turbidity meter was deployed at each station to obtain in-situ data for salinity (psu), temperature ( °C) and turbidity (NTU) distribution.Water samples for measuring inorganic nutrients concentrations were collected by a 5-liter Niskin bottle from one meter depth and fi ltered using Whatman GF/C fi lters.Th e automated determination for nitrate and silicate was based on Strickland and Parsons (1972), using a Skalar SUN Flow Analyser.For ammonia concentrations, we employed the phenol-hypochlorite method and added the required reagents immediately after obtaining the water sample.Ammonia concentra-tions were measured in the laboratory using a Beckman DU-650 spectrophotometer after 24 hours of incubation in the dark (Grasshoff et al. 1983).
In order to estimate phytoplankton biomass, the individual volumes of cells (μm3) and biomass as wet weight (mg/L) for each species were calculated according to approximate geometrical fi gures (Hillebrandt et al. 1999).To describe phytoplankton diversity, the Margalef 's richness index, Shannon's heterogeneity index and Pielou's evenness index were used.Similarity between species compositions was calculated by Jaccard and Czekanowski-Sørenesen indices of association.Cluster analysis was applied to generate dendrograms (group average method), based on the Jaccard and Bray-Curtis distance matrixes among samples.Pearson correlation coeffi cients were calculated for estimations of the relationships between environmental variables and the phytoplankton community.Calculation of indices and cluster analysis were performed using Primer 6.1.9software (Primer-E Ltd.).
Diatoms and dinofl agellates were the most diverse groups.Centric and pennate diatoms accounted for the highest diversity with 84 and 50 taxa, respectively.Among the centric diatoms, the most diverse genera were Chaetoceros (22 taxa), Rhizosolenia (12 taxa) and Coscinodiscus (nine taxa).For pennate diatoms, the Nitzschia group was represented by 17 taxa (14 species of the genus Nitzschia and three species of the morphologically close genera Pseudo-nitzschia and Cylindrotheca).Th e genus Pleurosigma was represented by seven species.Of the 56 species of dinofl agellates, over one-half were represented by three genera: Protoperidinium (16 taxa), Ceratium (eight taxa) and Prorocentrum (fi ve taxa).
As a whole, a pronounced prevalence of diatoms was typical for the phytoplankton community in Kuwaiti waters throughout the year.On the average, diatoms contributed 70% to the total species diversity.Th eir prevalence was at a maximum (80% to phytoplankton groups presented here follow the classifi cation scheme of Th rondsen (1997), which was partially modifi ed by Christensen (1962Christensen ( , 1966)).
100%) during the autumn-winter period, especially in November and December, and reduced during the spring and summer (April to July), especially at stations 5, 6 and 7.
Dinofl agellates contributed only 22% to the total species diversity, with a maximum of 40% to 70% during the spring-summer period, and <10% (often <1%) at stations 1 and 2 throughout the study period, probably a result of reduced salinities in the Shatt al-Arab discharge.

Variability of phytoplankton concentrations
Microalgae abundance ranged from 3.06×10 3 to 1.24×10 7 cells/L (1.88 ± 2.59×10 5 cells/L on average) and biomass from 0.03 to 161 mg/L (9.96 ± 24.10 mg/L on average).Diatoms dominated phytoplankton abundance numerically as well as in biomass, accounting for 99% of the latter depending on season.Phytoplankton concentrations, which were obtained in this study, are within the range of those reported previously (Al-Yamani at al. 2004, 2006).

Space-time variability of the phytoplankton structure
Assessments of the spatial and temporal variability of the structure of the phytoplankton community studied are presented in Table 2. High levels of average paired similarity between both the species compositions within stations (0.703) and within year (0.710) as well as small dispersion of these parameters testify a rather high taxonomic homogeneity of the phytoplankton community in Kuwaiti waters and similar trends in seasonal development of phytoplankton at diff erent locations.
For the community as a whole, the high β-diversity value (21.5) indicates heterogeneity in species compositions among the replicates.Average similarity of species structure within samples (average paired samples similarity using the Czekanowski-Sørensen Index) was 0.390 ± 0.141.

Temporal variability of phytoplankton
Analysis of seasonal variability within the phytoplankton community was performed using the hierarchical clustering using the Jaccard Index of similarity.For the samples collected throughout the year, we identifi ed four diff erent periods based on sample similarities (Fig. 2A).Th e community structure of samples within each period showed a higher degree of similarity than that of samples between periods.Cluster analysis found seasonal diff erences as follows: Cluster-1, late winter-spring (January, February and March); Cluster-2, spring (April and May); Cluster-3, summer-early autumn (July and September) and Cluster-4, late autumn-winter (October, November, and December).Each cluster was identifi ed by distinct phytoplankton associations (Fig. 2B, Table 3).

B
Composition of the top-dominant species and their percentage contributions to the total phytoplankton biomass from each isolated period of the year are shown in Fig. 2B.Th e beginning of the year (Cluster-1) was characterized by dominance of large-sized diatoms Guinardia fl accida (53% of the total phytoplankton biomass) and Rhizosolenia cochlea (39%).Late winter phytoplankton development was characterized by minimal values of species diversity and community evenness and maximum concentrations of phytoplankton.Diversity among dominant species during the spring (Cluster-2) was higher due to the appearance of large-and medium-sized diatoms: Palmeria hardmaniana, Rhizosolenia robusta, Eucampia zodiacus, Proboscia indica and Lauderia borealis.In the spring season, the decline of G. fl accida resulted in R. cochlea becoming the dominant species.Th is period was characterized by minimum phytoplankton concentrations; but species diversity and community evenness increased to their highest levels.In summer-early autumn (Cluster-3), the phytoplankton community consisted mainly of R. cochlea complemented by signifi cant concentrations of L. borealis.Th e rather high levels of phytoplankton biomass during the late autumnwinter period (Cluster-4) were supported mainly by R. cochlea and R. robusta populations.Th e maximum values of phytoplankton species richness were found in October to December.Generally, winter months of 2005/2006 were characterized by the main bloom of phytoplankton biomass in the surface waters (up to 60-80 mg/L in some locations), which started in the northern part of Kuwait in December, moving southward through Kuwait Bay during January and February.

Spatial variability of phytoplankton
To estimate the spatial variability within the phytoplankton community, we applied the Bray-Curtis Similarity Index among stations.Cluster analysis identifi ed three different phytoplankton associations in Kuwaiti waters (Fig. 3).
Th e fi rst area outlined (Cluster-1, station 1) was located in Kuwait's extreme northern waters closed to the Shatt al-Arab.Th e second area outlined (Cluster-2) consisted mainly of stations along Kuwait's coast (stations 4, 5 and 6) and included station 2. Th e remaining area (Cluster-3) was restricted to open waters (stations 3 and 7; Fig. 3A).Within each area outlined, distinct phytoplankton associations were found with regard to composition, concentration, species richness and diversity as well as community evenness.Dissimilarity of community structure between phytoplankton from diff erent associations is illustrated in Fig. 3B and Table 4. Th e fi rst association (Cluster-1) greatly diff ered from other locations by the highest phytoplankton concentrations, the minimum values of species richness, diversity and community evenness.Th e high levels of phytoplankton biomass were supported

Coscinodiscus wailesii Guinardia flaccida Lauderia borealis Odontella sinensis Proboscia indica Rhizosolenia cochlea Rhizosolenia robustra Other
by almost total prevalence of the large-sized diatom R. cochlea.Coastal waters were characterized by the dominance of R. cochlea and G. fl accida (the second association, Cluster-2).Phytoplankton composition in open waters (the third association, Cluster-3) diff ered clearly from those of northern Kuwait and the coastal waters due to low densities, despite maximum species richness, diversity and community evenness.In decreasing order of abundance for the off shore stations, the most important species included: G. fl accida, R. cochlea, Proboscia indica, R. robusta and Coscinodiscus wailesii.

Variability of phytoplankton structure along latitudinal gradient
In order to assess macro-scale spatial variability of the phytoplankton community within Kuwaiti waters, we analyzed distributions of species richness and diversity of large taxonomic groups as well as phytoplankton composition along a latitudinal gradient.Figure 4 shows the phytoplankton species richness plotted against latitude.Margalef 's index gradually increased from north to south.Th is trend conformed to a linear regression model, which described 91% of the spatial variability for mean species richness (r 2 =0.91).Phytoplankton composition from northern waters near the Shatt al-Arab estuary was less diverse than that of southern waters.Th e observed increasing trend was supported mainly by an increase of dinofl agellate diversity.
Phytoplankton composition within the northern waters was characterized by high prevalence of diatoms (from 76% to 96% of total species richness; Fig. 5A).If not totally absent, dinofl agellates contributed only 6% to 18% to the total species richness.Th e portion of dinofl agellates in the phytoplankton increased exponentially along the north-south gradient (Fig. 5B).Th e diatom/dinofl agellate ratio was equal to 11 within northern waters (station 1); whereas it was reduced to 2.6 in coastal waters, and even further to 2.3 in open waters (115 diatom species versus 50 dinofl agellates).

Relationship between phytoplankton community and environmental variables
In order to detect the diff erences between various areas within Kuwait's waters, we analyzed the composition of the main environmental factors (salinity, turbidity and nutri-Table 4. Phytoplankton concentrations and community structure within diff erent areas of Kuwait's waters, which were isolated by cluster analysis.Th e cluster numbers correspond to those in Fig. 3A.ent concentrations).Salinity values increased from the north to the south (Fig. 6A), whereas the opposite trend was observed for turbidity and nutrient concentrations along the latitudinal gradient (Figs 6B-D).

Cluster
To estimate the relationships between the phytoplankton community and the main environmental variables, we calculated the Pearson correlation coeffi cients.Correlation analysis was applied to the matrix of annual average values for each variable analyzed (Table 5).
Th e relationships between phytoplankton and environmental variables were not signifi cant in terms of microalgae concentrations.Correlation analysis, however, revealed strong statistically signifi cant correlations among phytoplankton structure and environmental variables.Species richness of phytoplankton community and selected taxonomical groups as well as percentage contribution of dinofl agellates to the community were strongly correlated with salinity (positive correlations with values of 0.82-0.97,p < 0.001), turbidity (negative correlations, r from -0.88 to -0.97) and with  5).For diatoms, we found signifi cant positive correlations with turbidity as well as with nutrient concentrations.Additionally, the high prevalence of diatoms in phytoplankton composition was associated with low salinity (r = -0.98).

Discussion
Th e relatively small geographic area of Kuwait's waters covers a very important transitional zone at the extreme north-western corner of the Arabian Gulf.From north to south, coastal waters of Kuwait extend for 170 km.Th ere is a range of interaction between the Shatt al-Arab River discharge and the Arabian Gulf marine environment.Th e shallow waters of Kuwait are characterized by high biological productivity (Al-Yamani et al. 2004), which are supported mainly by very abundant and diverse phytoplankton communities.
Th e high species diversity of the phytoplankton community (200 identifi ed taxa) is mainly due to diatom algae.Th e assessment of phytoplankton species diversity presented here is close to the maximum number of phytoplankton taxa recorded in the Arabian Gulf area, which is 223 taxa, including 134 diatoms and 86 dinofl agellates (Jacob and Al-Muzaini 1990).Th e latest estimation of diversity in Kuwaiti waters was 220 taxa, including 162 diatoms and 53 dinofl agellates (Al-Yamani et al. 2004).
Th e phytoplankton community in Kuwaiti waters diff ers from the rest of the Arabian Gulf by high prevalence of diatoms and low dinofl agellate species diversity due to the abundance of silicate nutrients in these waters.Th e occurrence of a signifi cant abundance of pennate diatoms, especially benthic taxa (large-and small-sized) diatom algae from periphyton, epipelon and epipsammon associations (genera Pleurosigma, Diploneis, Surirella, Trachyneis, Nitzschia, Entomoneis, Plagiotropis) is also noteworthy.For some of them (such as Pleurosigma spp., Surirella fastuosa, Trachyneis antillarum, and Nitzschia spp.) signifi cant abundances were observed in some inshore locations.
Both the spatial and temporal components contributed to the variability of the phytoplankton community in Kuwaiti waters.During the winter months, the northern area was characterized by the highest concentrations of phytoplankton, whereas the lowest phytoplankton concentrations were observed in open waters in summer.Th e minimum diversity level was associated with the spring months, whereas the maximum was in autumn (October).Phytoplankton species richness gradually increased Th e analysis of space-time phytoplankton variability allowed the clustering of similar samples and hence the identifi cation of the diff erent phytoplankton associations in Kuwaiti waters.Th e northern zone is unique and diff ers from the rest of the study area, which is clearly expressed in the distinctive features of phytoplankton structure and space-time variability.Pronounced diff erences in the northern area are explained by the strong infl uence of lower-salinity waters that are discharged from Shatt al-Arab River and the Shatt al-Basrah channel.
Th e Shatt al-Arab system, which collects the waters of the Tigris, Euphrates and Karun rivers, is the principal fl uvial input to the Arabian Gulf, especially to the northern areas including Kuwaiti waters (Al-Yamani et al. 2004).Seasonal freshwater supply from the Shatt al-Arab appears to have a local eff ect on the marine environment of the examined area.Th e infl uence of the Shatt al-Arab River discharge on the northern Arabian Gulf results in a gradient of environmental conditions, which change according to river fl ow volume.As a result of this interaction, diff erent locations and distinct periods may be identifi ed in Kuwaiti waters.Th ere is a northern zone, which is constantly more dynamic, turbid and rich in nutrients and at the same time less saline.For the other areas of Kuwait's waters, two diff erent time periods were identifi ed: the beginning of the year to May was characterized by higher nutrient concentrations and a decrease in salinity, which corresponds to higher river discharge, while the remainder of the year is characterized by lower nutrient concentrations and higher salinities.
Th ere is a signifi cant correlation among phytoplankton structure and physicochemical variables of Kuwaiti waters.Th e results suggest that salinity, turbidity and inorganic nutrient concentrations (inorganic nitrogen and silicate) were the main factors controlling changes in the phytoplankton community within the area examined.

Figure 1 .
Figure 1.Area of investigation.A Arabian Gulf, with inset showing the greater region in which the sampling area was located B Map of Kuwait showing locations of the stations sampled for phytoplankton in 2005 and 2006 (black dots).

Figure 2 .
Figure 2. Temporal variability of the phytoplankton community.A the dendrogram of cluster analysis (group average method), based on Jaccard's distance matrix among samples (10 months × 200 species; monthly average biomass values) B compositions of dominant species for diff erent phytoplankton associations within each period isolated by cluster analysis.

Figure 3 .
Figure 3. Spatial variability of phytoplankton community.A Dendrogram of the cluster analysis (group average method), based on the Bray-Curtis distance matrix among samples (seven stations × 200 species; annual average biomass values) B Compositions of dominant species in diff erent phytoplankton associations within each area outlined by the cluster analysis.

Figure 4 .
Figure 4. Change of species richness in the phytoplankton community along a north-south latitudinal gradient.Th e dots represent the annual average values of Margalef 's Index ± SD; the solid line represents the linear regression; the dashed lines represent the 95% confi dence interval.

Figure 5 .
Figure 5. Percentage contribution of phytoplankton groups to the total species richness plotted against latitude.A Diatoms B dinofl agellates; annual average values of species richness (number of species) ± SD.Contribution to species richness, % Latitude Latitude

Figure 6 .
Figure 6.Distribution of environmental variables (annual average values ± SD) along the north-south latitudinal gradient.A Salinity B turbidity C nitrate D silicate.

Table 1 .
Diversity of the main phytoplankton groups recorded from Kuwaiti waters in 2005 and 2006;

Table 2 .
Space-time variability of the phytoplankton community structure.

Table 3 .
Phytoplankton concentrations and community structure within diff erent periods of the year, which were isolated by cluster analysis.Cluster numbers correspond to those in Fig.2A.

Table 5 .
Pearson's Correlation Coeffi cients (r) among phytoplankton community and environmental variables measured in Kuwaiti waters in 2005/2006.Th e values in bold represent signifi cant correlations (p < 0.001).