The relationship between phytoplankton and fish in nutrient-rich shallow Lake Qarun, Egypt

Lake Qarun is the major inland saline and turbid lake in the northern part of El-Fayoum Depression (central Egypt, ~80 km southwest of Cairo, at the margin of the Nile Valley, 29°30’N, 30°40’E). The length of this lake from east to west is about 40 km and the breadth at the widest point is about 6.7 km. The surface area of the lake is about 243 km², the volume is 924 million m³ and it is located at 43 m below sea level (Anonymous 1995). The deepest point (8.3 m) is located in the northwest. The non-irrigated northern shores of the lake are virtually devoid of vegetation and mark the beginning of the Western Egyptian Desert. The lake has no connection with the sea, being located 320 km south of the Mediterranean coast of Egypt, and is directly supported by the Nile River via the Bahr Yussef canal. Since the early 18^th century, the lake has received mainly agricultural wastewater (Anonymous 1995). An increase in the water level can be attributed to the leakage from adjacent groundwater aquifers and from the Wadi El-Rayan Depression (El-Sayed & Guindy 1999; Mansour & Sidky 2003). The total water draining annually into the lake is about 395 million cubic meters (data provided by the Irrigation Department, El-Fayoum). Previously, the lake supported a moderate level of fisheries and caused a decline in water quality (Meininger & Atta 1994).

Water samples were collected seasonally from the subsurface water layer using a 1.5 dm³ Ruttner sampler from ten sites in Lake Qarun in winter (February), spring (May), summer (August) and in autumn (September) 2012. The lake comprises three main subareas, where water-sampling sites 1, 2 and 3 represent the east, while 4, 5, 6 and 7 represent the middle and 8, 9 and 10 represent the west of the lake (Fig. 1). The fish catch in the lake was carried out at 11 landing sites (Senoris, Abu-Neema, Shakshouk, Abu-Soliman, Abu-Shanab, Kahk, El-Lokanda, El-Rawashdia, El-Saaida, Ayuob and Qarun, from east to west respectively), located in the southern part of the lake. The landing sites 1-3 belong to the east, sites 4-8 to the middle and 9-11 to the west subareas.

Figure 1

During sampling, pH was measured using an Orion Research Ion Analyzer (399A), electrical conductivity – using a conductivity meter (S.C.T.33 YSI), salinity – a portable Hydro Lab equipment, mod. Multi 340I/SET WTW, transparency – a black-white Secchi disk and water temperature – an ordinary thermometer. Water samples were analyzed according to the APHA (1996) procedures. Colorimetric techniques were used for the analysis of nutrients: formation of a reddish purple azo-dye for NO₂, Cd reduction for NO₃, the phenate method for NH₄ and stannous chloride reduction for PO₄. Total nitrogen and total phosphorus (after mineralization) were also analyzed colorimetrically.

A defined volume of water was filtered through a glass microfiber filter (GF/F) and the filter with residue was then foiled and refrigerated for pigment analyses. The chlorophyll a (Chl a) concentration retained on filters was extracted in acetone (90%) overnight and then prepared and measured spectrophotometrically and calculated by applying the trichromatic equation. The Chl a was chosen as the best descriptive parameter of the total phytoplankton biomass.

The phytoplankton samples were preserved with 4% neutral formalin and Lugol’s iodine solution (Margalef 1974) and then transferred into a glass cylinder. Phytoplankton cells settled for 5 days (APHA 1996). The phytoplankton samples were then siphoned and concentrated to a fixed volume and transferred to plastic vials for microscopic examination. The drop method was applied to count and identify the phytoplankton species. Triplicate samples (5 μl) were taken and examined under an inverted microscope at magnifications of 400× and 1000×. The identification of taxa followed e.g. Huber-Pestalozzi (1961, 1983), Komárek & Anagnostidis (1986, 1989, 1999), Krammer & Lange-Bertalot (1986, 1988, 1991a, 1991b), Anagnostidis & Komárek (1988), Popovský & Pfiester (1990).

Thirteen field trips were conducted in the three subareas in 2012, in December and February (referred to as winter) and in July and September (referred to as summer) and 78 fishermen were interviewed. The fishing season in the lake lasted throughout the year except for 3 months (January, May and June) as a fishery management plan. The fish catch and species composition in each subarea were investigated. Fish from different types of fishing gears and methods were collected and identified to the species level according to Whitehead et al. (1986) and FishBase (2012).

Non-parametric methods (due to lack of normally--distributed data) were applied to test the significant changes of environmental variables in Lake Qarun using the Kruskal-Wallis test. The relationships between physicochemical and biological variables in Lake Qarun were confirmed by calculating the Spearman’s rank correlation coefficient with STATISTICA version 10. The percent similarity of phytoplankton taxa was determined based on a cluster analysis (Multi-Variate Statistical Package, Kov. Comp. Serv. 1985-2009). As predicted, the relationships were statistically significant at the significance level of 0.05.

The trophic state of Lake Qarun was determined based on the Trophic Level Index (TLI) (Burns et al. 2005), which includes four partial modules based on the Secchi disk depth (SDD) and the concentrations of chlorophyll a (Chl-a), total phosphorus (TP) and total nitrogen (TN). The final TLI is an average from values obtained separately for each module: TLI_SDD, TLI_Chl, TLI_TP and TLI_TN. Models of the annual physicochemical and biological changes were determined from transformed data on phytoplankton density, chlorophyll a concentration, Secchi disk depth, total and dominant fish catches from all sampling sites.

Lake Qarun is a saline inland lake characterized by a significantly differentiated salinity (K-W test, p=0.001) across the lake, which varied between 11.1 PSU and 37.8 PSU during the study period of 2012, with some increasing tendency from the east to west subareas. The electrical conductivity (EC, 16.5-49.1 mS cm^-1) and the mineral forms of N and P were statistically significantly differentiated between the sampling sites (Fig. 2). Similar findings with a high EC and large spatial variability in nitrite, nitrate and ammonium concentrations were recorded in August 2011 (El-Shabrawy et al. 2015). The lowest values of EC were related to the highest total nitrogen (max 6.45 mg dm^-3), mineral nitrogen (the sum of nitrite, nitrate and ammonium; max 2.34 mg dm^-3) and total phosphorus (max 0.63 mg dm^-3) concentrations and the lowest salinity at sites 1 and 7, i.e. the east and middle subareas, respectively. The separation of nutrient-rich sites 1 and 7 from the others is connected with the agricultural water discharge from the drainage systems of El-Bats and El-Wadi mentioned by El-Shabrawy et al. (2015). Regarding such a division of areas, the Secchi disk depth (0.3-1.5 m) significantly increased (K-W test, p=0.010), whereas the chlorophyll a concentration (2.7-69.3 μg dm^-3) significantly decreased (K-W test, p=0.004) along the sites i.e. from east to west. The water temperature ranged from 12.1 to 29.1°C and pH varied from 7.9 to 8.6. The other parameters were approximately similar. According to the TLI-system (a trophic state monitoring tool) proposed by Parparov et al. (2010), the assessment indicated hypereutrophic waters at almost all sampling sites in Lake Qarun, excluding sites 6 and 8 (eutrophic) (Table 1). Such a high trophic level was caused mainly by high concentrations of TN and TP, which is consistent with findings from other water bodies (Napiórkowska-Krzebietke et al. 2013, NapiórkowskaKrzebietke & Dunalska 2015). Comparing the data on nutrient enrichment in Lake Qarun, summarized by El-Shabrawy & Dumont (2009), Abdel-Satar et al. (2010) and El-Shabrawy et al. (2015), a clear increasing trend occurred in the concentrations of nitrites, nitrates, ammonium, orthophosphates and total phosphorus from 1955 to 2006. The assessment of trophic conditions also indicated a more advanced eutrophic level of this lake in the period of 1995-2006 (Table 1).

Figure 2

In total, 134 species were recorded in Lake Qarun, including 60 species of Bacillariophyceae, 35 species of Chlorophyceae, 13 species of Cyanophyceae, and less than 10 species each of Dinophyceae, Chrysophyceae, Cryptophyceae and Euglenophyceae. Previous studies suggested that the species richness was much lower, e.g. in 2001 only 49 phytoplankton taxa were recorded (Fathi & Flower 2005).

The phytoplankton density in Lake Qarun ranged from 75 to 140 × 10⁴ cells dm^-3 in the spring (Fig. 3A). A similar density was noted in autumn. The most abundant phytoplankton was observed in summer and winter – its density reached 935 × 10⁴ cells dm^-3 and 815 × 10⁴ cells dm^-3, respectively. The annual phytoplankton growth in Lake Qarun indicated the enhanced density in its east subarea and a decreasing tendency toward the west subarea with the minimum at site 8. The spatial variation of the phytoplankton density could refer to local trophic conditions, especially inflows of contaminants (Abdel-Satar et al. 2010). Such findings are reported for large lakes situated within urban or agricultural catchments (e.g. Dembowska et al. 2015). Comparing the annual variation in phytoplankton taxonomic composition at the class level, Bacillariophyceae and Dinophyceae co-dominated in phytoplankton assemblages (Fig. 3B), with a similar decreasing tendency toward the west as it was observed in the total density. The highest contribution of Dinophyte was recorded at the first site, and diatoms dominated in assemblages at the other sites in the Bacilariophyceae:Dinophyceae ratio of up to 3:1 (site 9). Furthermore, such a large difference between sites 1-5 and 6-10 was confirmed by the cluster analysis based on the percentage similarity in the taxonomic structure (Fig. 4).

Figure 3

Figure 4

The seasonal variation of phytoplankton structure indicated that diatoms predominated in spring and winter, whereas dinophytes predominated in summer assemblages (Fig. 5). In autumn, both groups co-dominated. The contribution of other groups was much smaller, with a distinct and seasonally similar presence of Chlorophyceae and Cyanophyceae. Bacillariophyceae were dominated by Cyclotella glomerata Bach. (up to 37.2% of the total density), C. operculata (C.Agardh) Bréb. (up to 22.2%), Lindavia ocellata (Pant.) T. Nakov et al. (up to 20.2%), L. kuetzingiana (Thwait.) T. Nakov et al. (up to 12.0%) and Ulnaria ulna (Nitzsch) P. Comp. (up to 25.0%) (Table 2). These species abundantly occurred mainly in spring and winter. The small-sized Cyclotella have generally low nutrient and temperature requirements (Brettum, Andersen 2005). The species were also described as dominants in other Egyptian lakes (Gharib & Abd El-Halim 2006; Hussian et al. 2015). The second predominant class was Dinophyceae, which was mostly represented by Nusuttodinium aeruginosum (F. Stein) Y. Tak. & T. Hor. and Prorocentrum micans Ehr. with the maximum density of 77.5% and 46.4% (of the total density) in summer and autumn, respectively. Chlorophyceae were dominated by e.g. Chlamydomonas globosa J. W. Snow, Crucigenia tetrapedia (Kirch.) Kuntze and Schroederia nitzschioides (G. S. West) Korsch. Among Cyanophyceae, chroococcales Aphanocapsa elachista var. conferta West & G. S. West and Chroococcus dispersus (Keiss.) Lemm. were most abundant, whereas Mallomonas ploesslii Perty and Mallomonas helvetica Pasch. were most abundant among Chrysophyceae.

Figure 5

The main fishing gear used in the lake were trammel nets, beach seines, handlines and traps. There were four types of trammel nets which differed in their design, characters, dimensions and mesh sizes according to their target species. The first trammel net (Ghazl Bolti) targeted mainly tilapias, particularly Tilapia zillii. This worked effectively in the east and middle subarea of the lake. The second trammel net (Ghazl Bory), which targeted Mugil cephalus worked mainly in the northern part of the lake and in deeper water. The third trammel net (Ghazl Mossa) was set mainly on the bottom to catch sole fish, e.g. Solea solea and S. aegyptiaca. It operated mainly in the middle and western part of the lake. The fourth trammel net (Ghazl Fahhar) targeted mullet fish species, other than M. cephalus, such as Liza ramada, L. aurata and L. saliens. It was mainly active in the middle subarea. Beach seines with two main types, Gorafet Bory and Gorafet Zardina, were also set in the lake. Gorafet Bory targeted mainly M. cephalus and it operated in the northern beach of the lake, while Gorafet Zardina was present mostly in the southern beach and targeted anchovy. Other fishing gears, handlines and traps targeted mostly T. zillii in the eastern and middle subarea.

The total catch of fish and seafood was estimated using the yearbook statistics of GAFRD (2013). A significant increase in the Lake Qarun catch was recorded from 2003 to 2012 (excluding 2006) (Fig. 6). By analogy, an increasing trend of the total national production was also recorded in Egypt (Samy-Kamal 2015). According to El-Serafy et al. (2014), the fish production level in Lake Qarun is rather low compared to other Egyptian lakes. During the 2012 fishing season, the total catch in Lake Qarun was estimated at 4410 tons according to GAFRD (2013). In winter, the catch was 1731 tons (ca. 39.35%), and 2679 tons (ca. 60.7%) in summer. The catch in the lake was composed mainly of tilapias (dominant T. zillii while Oreochromis aureus, O. niloticus and Sorthedon galileous were sporadic), mullets (M. cephalus, L. saliens, L. ramada and L. aurata), soles (Solea solea and S. aegyptiaca), sea-bream (Sparus auratus), sea-bass (Decintrarachus labracus), anchovy (Engraulis encrasicolus) and shrimps (Metapenaeus stebbingi and Penaeus semisulcatus). The species composition in the whole lake was seasonally recorded as follows: in winter, the dominant species were tilapias (mainly T. zillii), soles (S. solea and S. aegyptiaca), mullets (M. cephalus, L. saleins, L. aurata and L. ramada) and shrimps (M. stabbingii and P. semisulcatus), which contributed about 32.1%, 25.0%, 24.6% and 17.2% of the lake catch, respectively. The sea-bass, eels (Anguilla anguilla) and others (E. encrasicolus and Gubius niger) each contributed below 1%. In summer, the catch composition showed a higher species diversity and included mullets (32.0%), tilapias (25.0%), shrimps (22.9%), soles (19.2%), D. labracus (0.3%), eels (0.1%), Siganus luridus (0.1%), S. auratus (0.1%) and other species (0.3%). Similar fish and shrimp taxa were also frequently recorded in the national catch composition in Egypt (Samy-Kamal 2015).

Figure 6

The fish and seafood catches in Lake Qarun according to subareas are presented in Table 3. The production of the east subarea was about 27.2% (1199 tons) of the annual lake catch (11.3% in winter and 15.9% in summer). The middle subarea produced about 35.7% (1573.7 tons), most of which (26.0%) was caught in winter and the rest in summer. The catch of the west subarea accounted for about 37.1% (1637.3 tons) of the lake catch (about 30.0% in summer and about 7.1% in winter).

The fish species structure was differentiated in subsequent subareas. In the east subarea, T. zillii (67.2%), Solea spp. (17.2%) and M. cephalus (13.9%) accounted for over 98% of the total catch in winter, similar to the summer catch (T. zillii – 49.7%, M. cephalus – 37.3% and Solea spp. – 12.4%) (Fig. 7). In the middle subarea, the dominant species were mainly E. encrasicolus (39.1%), M. cephalus (26.1%) and T. zillii (9.0%) in winter, while T. zillii (65.7%), Solea spp. (25.3%) and L. saliens (6.1%) in summer. In the west subarea, M. cephalus (58.0%) and Solea spp. (38.7%) co-dominated in winter, and M. cephalus (60.7%), T. zillii (16.4%) and Solea spp. (16.0%) co-dominated in summer.

Figure 7

The highest phytoplankton density and chlorophyll a concentrations were recorded in the east subarea (Fig. 8A). The highest water transparency was recorded in the middle (site 6) and in the west part (site 8), whereas the lowest transparency was determined in the east subarea (site 1). The maximum nutrient concentrations were recorded at site 1 (mineral nitrogen and orthophosphates) and site 7 (total phosphorus), characterized by the strongest exposure to pollutants (El-Shabrawy et al. 2015). The largest total catch of fish and seafood, with M. cephalus as a dominant fish, came from the west subarea in summer (Fig. 8B), along with the lowest phytoplankton. Similar results were found for some Polish lakes, where conditions with less abundant phytoplankton and high oxygen concentrations were suitable for some coregonids (Napiórkowska-Krzebietke et al. 2016). Furthermore, the peak of phytoplankton density was accompanied by a decline in the fish catch in the east subarea.

Figure 8

The correlations between the physicochemical and biological variables in Lake Qarun were tested with Spearman’s rank correlation coefficient and are presented in Table 4. The large catch of M. cephalus was significantly correlated with the total catch in Lake Qarun (R=0.829). The second dominant fish, i.e. T. zillii was significantly positively correlated with Dinophyceae, concentrations of TP and ammonium, TLI and pH. Statistically significant negative correlations were found between Solea spp. and the total phytoplankton density, Dinophyceae density and the concentrations of chlorophyll a, nitrite and nitrate. Salinity and electrical conductivity were significantly positively correlated with Solea spp. In addition, the relationship between other fish species and orthophosphates was statistically significant, but negative.

In Lake Qarun, four major groups: mullets, soles, tilapias and shrimps (with different diet preferences and direct/indirect pressure on phytoplankton) have been recorded in the total catch for many years. Mullets can be detritivorous (Porter et al. 1996), benthivorous (Cardona & Castelló 1989) and planktivorous (Cardona et al. 1996; Torras et al. 2000; El-Shabrawy & Khalifa 2007). Abdel-Malek (1980) found that the lake mullet (Mugil cephalus, Liza saliens, L. aurata and L. ramada) can feed on benthic animals and plants in a different ratio, i.e. L. saliens and M. cephalus feed mainly on benthic algae and detritus, and L. ramada prefers benthic algae (El-Shabrawy & Khalifa 2007). However, Cardona (2001) found that the spring diet of mullets in a coastal lagoon based on algae was a response to detritus reduction, but in autumn it shifted to detritus and diatoms. Solea aegyptica feeds mainly on the lake macrobenthos (El-Shabrawy & Khalifa 2007). Tilapia zillii is an omnivorous species, feeding mainly on phytoplankton: Bacillariophyceae (26.8%), Rhodophyceae (24.2%) and Cyanophyceae (12.4%), while polychaetes, crustaceans, mollusks contributed only 9.4% (Abdel-Malek 1980). Macrobenthos and zooplankton constitute the main food (27% and 6.6%, respectively), while phytoplankton account for only 18.3% (El-Shabrawy & Khalifa 2007). Regarding shrimps, Prawn kerathurus is predominantly carnivorous, but sometimes also omnivorous (Ishak et al. 1980). The euglenoids (Euglena sp. and Phacus sp.) were the most important food items in the gut of Metapenaeus stebbingi (dominant in shrimp catch) during the summer of 1993 (El-Shabrawy & Khalifa 2007).

Variables Total catch Mugil cephalus Tilapia zillii Solea spp. Engraulis encrasicolus Others Total catch - 0.829 n.s. n.s. n.s. n.s. Mugil cephalus 0.829 - n.s. n.s. n.s. n.s. Tilapia zillii n.s. n.s. - n.s. n.s. n.s. Solea spp. n.s. n.s. n.s. - n.s. n.s. Engraulis encrasicolus n.s. n.s. n.s. n.s. - n.s. Other fish species

Strong or slight physicochemistry-phytoplankton--fish correlations were previously mentioned e.g. by Reid et al. (2000) or James et al. (2003) in seawaters and by Hansen & Carey (2015) in temperate waters, where fishery (through top-down control) can be an important contributor to environmental changes. However, some reckless actions (e.g. overstocking or excessive feeding of fish in aquaculture systems) can lead to enhanced water degradation (Legaspi et al. 2015). In Lake Qarun, the complexity of biological and chemical interactions (e.g. food availability, competition, predation) and the additional input of pollutants (e.g. heavy metals, pesticides) affect the abundance and structure of aquatic organisms (e.g. Abou El-Geit et al. 2013; El-Shabrawy et al. 2015). The earliest response to the variability of this lake ecosystem is shown by phytoplankton, especially its spatial heterogeneity.