Role of environmental variables and seasonal mixing in dynamics of the phytoplankton community in a Tropical Highland Lake Ardibo, Ethiopia

Abstract This study aimed to assess the spatial and seasonal variations of the phytoplankton community in response to environmental variables in Lake Ardibo. The study was done from October 2020 to September 2021 and physicochemical parameters such as water temperature, pH, dissolved oxygen, electrical conductivity, turbidity, alkalinity, Secchi-depth, nitrate, ammonium, silicon dioxide, soluble reactive phosphorus, and total phosphorus were measured using standard methods. ANOVA results indicated that all physicochemical parameters except turbidity had significant variation (p < 0.05) in all seasons. A total of 66 phytoplankton species grouped into seven phyla, including Bacillariophyta (23 taxa), Chlorophyta (21 taxa), Cyanophyta (14 taxa), Euglenophyta (3 taxa), Charophyta (2 taxa), Dinophyta (2 taxa) and Cryptophyta (1 taxon) were identified. Phytoplankton biomass showed significant spatial and seasonal variation with maximum cell density during the pre-rainy season, and low density in the main rainy season. Redundancy analysis (RDA) revealed that nutrients had a positive affinity with the phytoplankton community, whereas electrical conductivity, water temperature, and pH showed a negative relationship with phytoplankton abundance. Temporal rather that abiotic factors far influenced phytoplankton dynamics in this deep tropical lake. The analysis postulates that partial atelomixis during the pre-rainy months favors higher phytoplankton biomass with functional groups such as desmids and heavy diatoms contributing most during this mixing period, while wet season mixing depressed algal biomass due to increased turbidity with nutrients not being limiting during these times.


Introduction
Phytoplankton composition, community structure, species diversity, and abundance in the aquatic environment are controlled by various environmental variables, mainly salinity, silicate, phosphate, nitrate, nitrite, ammonia, temperature, pH and the effect of grazing by zooplankton (Oseji et al. 2018;Vajravelu et al. 2018;Belokda et al. 2019).Environmental changes in hydro-chemical and physical parameters shapes the distribution and abundance of many phytoplankton species (Vajravelu et al. 2018;Graco-Roza et al. 2021), and therefore, phytoplankton are known for their potential as relevant bioindicators of environmental changes in all aquatic systems (Belokda et al. 2019).
Phytoplankton is one of the biological indicators of environmental changes in the aquatic ecosystem due to its sensitivity to changes in the environment (Reynolds 2006;Wojciechowski et al. 2017a;Wojciechowski et al. 2017b) and therefore reflect, existing water quality status and a good insight before algal bloom formation (Shekhar et al. 2008;McQuatters-Gollop et al. 2009;Nin cevi c-Gladan et al. 2015;Bu zan ci c et al. 2016;Inyang and Wang 2020).Understanding the structure, dynamics, and diversity of phytoplankton is vital to measure the suitability of water quality for a healthy ecosystem and the trophic condition of the water body (Oseji et al. 2018;Mishra et al. 2019;Liu et al. 2021).
Changes in the structure of algal can alter other trophic levels, stimulating or limiting other interdependent aquatic organisms.Previous studies on freshwater bodies indicated that changes in phytoplankton composition reflect variation in water quality and are helpful information for changes in physicochemical variables and biotic interactions (Das et al. 2018;Ula nczyk et al. 2021).
Spatio-temporal changes in the physicochemical parameters can impact the succession of phytoplankton communities and provide opportunities to understand the level of water quality parameters in a particular area (Inyang and Wang 2020).Fluctuating these parameters is closely associated with short-term changes in the water retention time, flux regime, outflow, and level (Mishra et al. 2019).Frequent reorganization of the phytoplankton composition and cell abundance would depend on physical processes with the availability of light, nutrients, and oxygen (Baliarsingh et al. 2015).Some records of heavy phytoplankton blooms in deep, tropical South American lakes were ascribed to the phenomenon 'atelomoxis'.Atelomixis is one robust water movement occurring once a day (during night) and extending either to whole water column (dry periodfull atelomixis) or is restricted to epilimnion (rainy periodpartial atelomixis).This complex stratification pattern is a driving force upon phytoplankton community structure and the determinant role of atelomixis has been explored in some other tropical and subtropical systems in Brazil (Bouvy et al. 2003;Lopes et al. 2005;Becker et al. 2009).
Lake Ardibo is one of the crater lakes situated near the edge of the northwestern rift escarpment.This lake has been used for irrigation and plays a significant role in local fishing activity and watering of animals (Ayenew and Demellie 2004;Metekia 2019).Earlier studies on Lake Ardibo fishery resource, hydrology and water balance were published (Ayenew and Demellie 2004;Yesuf et al. 2012;Metekia 2019).However, information about physichochemical variables and phytoplankton communities of Lake Ardibo are is still scarce and unknown.In this study, we wanted to see whether the pattern of seasonal succession of the phytoplankton community in Lake Aardbo also verified atelomixis in this tropical, deep, highland lake.The present study was carried out to assess the response of the phytoplankton community structure to the environmental variables and seasonal mixing in Lake Ardibo.

Study area
Lake Ardibo is situated in the northeastern part of Ethiopia, bounded by 39.75 to 39.78 E longitudes and 11.20 to 11.27 N latitudes within the Awash River Basin (Figure 1).
The lake have closed drainage system with no surface outflow (Yesuf et al. 2012).The lake area and its average depths are about 15.8 km 2 and 25.45 m (Yesuf et al. 2012).The catchment is highly cultivated and characterized by scattered bushes and natural grazing lands.The area's climate condition is sub-humid and characterized by four seasons such as, main-rainy season, post-rainy season, and pre-rainy season.The lake support the local community through fishing, grazing for cattle, irrigation and other services.It is also provides habitat to different aquatic life.There are exotic fish species commercially important found in this Lake introduced for fisheries and as weed control, including Nile tilapia (Oreochromis niloticus) and Common carp (Cyprinus carpio), respectively (Asnake and Mingist 2018).Three sampling sites, MEN.(Menafesha), represent littoral areas (settlement), Goro (GOR) found around the agricultural areas, and open water (OP) denoting the pelagic area of the Lake.Sampling sites were selected based on human impact, livestock, and agricultural activities.

Environmental variables
Environmental variables in the water were measured and collected from the three sites of Lake Ardibo (Figure 1) from October 2020 to September 2021 during the post-rainy season (POR), dry season (DR), pre-rainy season (PRA), and the main-rainy season (MRA).Samples were collected two times of each season.Water temperature, pH, and electrical conductivity (EC) values were recorded in-situ using a portable pH/EC/TDS/temperature combined HANNA, HI 991301 model instrument.An oxygen meter of the model Hanna H198186 meter was used to measure the amount of dissolved oxygen (DO) by dipping the electrode into the water, and the reading was noted for a stable value.Transparency of the water was also measured by lowering a 20 cm diameter circular disc (Secchi disc) down to the water column.The turbidity of the water was also determined with the use of a portable turbidity meter.
Water samples collected from each sampling site were used to analyze the concentration of nitrate, ammonium, silicon dioxide, soluble reactive phosphorus, and total phosphorus.The analysis of the nutrients was done seasonally using a spectrophotometric method in the Limnology laboratory of Addis Ababa University.Nitrate was measured with sodium salicylate method, ammonium with indophenol blue method (APHA 1995), silica with the molybdosilicate method, soluble reactive phosphate, and total phosphate with ascorbic method (APHA 1999).

Phytoplankton collection and analysis
Both 15 lm and 30 lm mesh size nets were used to collect phytoplankton samples from the open water and two littoral areas of Lake Ardibo.About 30 liter of water sample was filtered at each sampling location and 1 L of composite water sample was placed in glass bottles, and 10% of neutral Lugol iodine solution was added to the bottles containing phytoplankton for fixation.The collected samples were then transported to the Limnological laboratory of Addis Ababa University.The water samples in the laboratory were transferred to a measuring cylinder of 50 mL capacity and stored in darkness for sedimentation.After 48 h, the supernatant of about 90% of the total sample volume was slowly siphoned off without disturbing the sedimented algae. 1 mL of concentrated subsample was transferred to a Sedgewick-Rafter counting chamber and allowed to settle before counting (H€ otzel and Croome 1999).Taxa were then identified according to their morphology and the units (cells, colonies, and filaments) were counted under Nikon inverted light microscope (NIKON TS100, Germany).The species photography was taken using a digital camera mounted on the inverted microscope at the magnification power of 100Â and 400Â.The taxa identification was made based on the various taxonomic literature on phytoplankton identification keys (Whitford and Schumacher 1973;Komarek and Anagnostidis 2005).The cell number (cells ml À1 ) of the lake water was quantified based on H€ otzel and Croome (1999) and Wetzel and Likens (2000) mathematical formula.

C cells ML
where N ¼ number of cells or units counted, A ¼ area of field (1mm 2 ), D ¼ depth of a field (Sedgwick-Rafter chamber depth-1mm) and F ¼ number of field counted.

Biomass estimation
Lake water samples were filtered through GF/F (Whatman) filters under a low vacuum and frozen at À20 C. The pigments were then extracted using 90% acetone overnight at 4 C, and concentration chl-a was measured at a wavelength of 665 nm and 750 nm following the method described in Talling and Driver (1963).

Statistical data analysis
Statistical analysis and data plotting were performed using excel for Windows 10 and SPSS (version 26).Variations of physicochemical parameters and phytoplankton biomass across sites and within seasons were performed with one-way ANOVA followed by Tukey-HSD test.The association between environmental variables and phytoplankton species was examined through canonical multivariate analysis using CANOCO for windows 4.5 (Ter et al. 2002).Leps and Smilauer (1999) explained that when the length of the longest gradient is <3, the species data indicate linear response to physico-chemical variables.Thus, redundancy analysis (RDA) was carried out because the detrended correspondence analysis (DCA) test result of response data showed that the length of the longest gradient was about 1.1.

Physicochemical parameters
Most of the physical variables showed significant variations (p < 0.05; Table 1) except temperature and pH, both in space and time.The average water temperature values varied between 21.07 C (dry) to 22.68 C (main-rain), and the mean pH measurement of the surface water ranged from 9.02 (dry) to 9.70 (main-rain) (Table 1).The highest dissolved oxygen concentration (7.71 mg L À1 ) was observed during the main rainy season, with super-saturation of about 116.6%.All the inorganic nutrient contents showed significant differences in the studied seasons (ANOVA: p < 0.05).The highest peak value of 78.20lgL À1 nitrates (NO 3 -N) occurred during the main rainy season, and the lowest value of 14.54lgL À1 was measured in the post rainy period.There was marked temporal fluctuation in the mean value of ammonia (NH 3 -N), with the highest value of 221.26 lgL À1 detected in the dry season and the lowest value of 8.67 lgL À1 during the pre-rainy season.The highest concentrations of soluble reactive phosphorus (SRP, mg L À1 ) with 57.10 were found in the main rainy period.However, its levels of concentration were under the limits of detection of the method of analysis used during the dry season, as presented in (Table 2).The concentration of total phosphorus (TP) ranged from 52.88 lgL À1 to 81.07 lgL À1 .The mean concentrations of silicon dioxide (SiO 2 ) were variable during the study time, with the highest value of 14.30 mgL À1 during the post rainy season (Table 2).The difference in water temperature per meter depth was slight, and differences from top to bottom were less than 1 C in all sampling periods.Based on depth profiles, the concentration of oxygen relatively varied down the water column and maximum in the surface water, while it declined with increasing depth and markedly dropped at 15 m depth during the pre-rainy and main rainy seasons.The seasonal maximum (7.78 mgL À1 ) and minimum(0.2mgL À1 ) concentration of DO was recorded at 2 m and 15 m depth in the rainy (August) and the minor rainy season (May), respectively.

Phytoplankton species composition and abundance
The identified phytoplankton contained 66 phytoplankton taxa belonging to 7 taxonomic groups (Table 3).The most significant component was dominated by Bacillariophyta, containing 23 species (35%), followed by Chlorophyta with 21 species (32%) Cyanophyta was the third most common group with 14 species and accounted for 21% of the total phytoplankton cell identified.Euglenophyta (3 species, representing 5%), Charophyta, and Dinophyta phyla had two species, each of them constituted about 3% of the total species identified, while Cryptophyta was the least dominant group (1 species, representing 1%).

Relative abundance of phytoplankton taxa
In this study, Cyanophyta was the predominant group throughout the annual cycle, and exhibited 61% of the total phytoplankton abundance in the samples.The second most abundant algal group was the Chlorophyta, which varied from 1.54 to 7.81%, with a mean value of 19%.The contribution of Charophyta to the total phytoplankton abundance in the Lake was 7%, dominated by Cosmarium contractum.The contribution of Bacillariophyta was 7%, which was dominated by Fragilaria, Aulacoseira, and Pinnularia viridis.The last group of phytoplankton Dinophyta (6%) was the least abundant, with Peridinium spp. the major contributor.

Seasonal and spatial variation in phytoplankton abundance and chlorophyll-a concentration
A numerically significant number of species was recorded during the post rainy season (October), while the least species were observed in the main rainy season (August).Bacillariophyta, Chlorophyta, and Cyanophyta were observed in all sampling periods, while others were identified in some sampling periods.Considering sites, the percentage contribution of Cyanophyta is the highest due to Anabaena ambigua species at both MEN (accounting for 66.93%) and GOR (accounting for 50.60%), while phylum Chlorophyta contributes 41.65% of the total phytoplankton abundance at the open water (Figure 2(a)).It is noted that the phytoplankton abundance in Lake Ardibo showed distinct seasonal variations.The highest abundance of phytoplankton occurred during the pre-rainy season due to Cyanophyta, about 83.03%, the most significant proportion of its abundance reached by Aphanothece minutissima.On the contrary, the percentage contribution of the Bacillariophyta group to the total phytoplankton abundance was small (1.16%) in the pre-rainy season (Figure 2 Figure 4 illustrates that the distribution of phytoplankton groups at different water column depths is varied.The dinoflagellates, diatoms, and green algae are widely distributed at the top of the Lake Surface, while the Charophyta group is highly distributed at 5 m depth of the Lake.However, the distribution of the cyanobacteria groups increased down the water column and was dominant at 10 m depth.
The Chlorophyll-a boxplot (Figure 5) exhibited temporal and spatial variations during the investigated period.The mean values of surface water chlorophyll-a concentration was 2.41, 2.48, and 4.11, mg L À1 in littoral (MEN), vegetated areas (GOR), and the open water (OP), respectively.The observed maximum concentration of Chl-a of the surface water at GOR resulted from high nitrate levels compared to the other two studied sites of the Lake (see Table 2).Phytoplankton biomass production was not significant at the studied sites although relatively higher at GOR (Figure 5   occurred during the pre-rainy season.It is probably related to water mixing events during the post-rainy season associated with atelomixis, because the composition of the phytoplankton assemblage was 68% dominated by Charophyta, Bacillariophyta and Dinophyta during this period.

Phytoplankton community patterns with environmental variables
A redundancy analysis (RDA) tri-plot was employed to investigate the relationship between environmental factors and main phytoplankton groups at the study sites of Lake Ardibo.
The results indicated that the first and second axes accounted for 100% of the cumulative percentage variance of species-environment relations (Table 4).The first axis accounts for 95.4%, and the second axis explains only 4.6% of species-environment relations (Table 4).Phytoplankton density is closely associated with water environmental factors and most significantly positively and negatively affected by TP and EC, respectively.There was a positive relationship between the abundances of Kirchnerilla, chlorella, Synedra ulna, Chroococcus minor, Staurastrum bullardii, Cosmarium contractum, Scenedesmus bijugatus, Peridinium spp., Cymbella sp. and Crucigenia with NO 3, SiO 2 , TP, Alka and reactive phosphate (SRP) at OP (Figure 6).Species such as Navicula antonii.Nitzschia dissipata and Anabaena ambigua are highly influenced by turbidity and DO.However, other phytoplankton species include Fragilaria sp., Aulacoseira, Epithemia sorex, Pinnularia viridis, Merismopedia, Cyanodictyon endophyticum, Tetraedron minimum, Aphanothece minutissima, Oocystis lacustris and Gomphonema acuminatum were negatively correlated with EC, NH 3, water temperature, and pH at MEN and GOR.
We found a wide seasonal variation in environmental parameters that greatly influence the phytoplankton assemblages (Table 5).Many predominant species include Pinnularia,

Physico-chemical parameters
The high water temperature measured in Lake Ardibo is similar to the previous record for the same lake (21.7-23.7 C; Asnake and Mingist 2018) and also the nearby lake, Lake Hayq (21 C-26 C; Fetahi et al. 2014) and (20 C-24 C; Tessema et al. 2020), although they are higher than those reported by Degefu et al. (2014) and Degefu and Schagerl (2015) in the highland lakes of Dendi and Wonchi.Asnake and Mingist (2018) recorded higher oxygen concentration of Lake Ardibo (12.06-17.53ppm) than the present study but the results of earlier studies in Lake Hayq (Fetahi et al. 2014;Tessema et al. 2020) were closer to this report.The lowest mean values of dissolved oxygen was observed in the dry season which was also recorded by Fetahi et al. (2014).A contrasting seasonal trend was observed for nutrients in the present study.The maximum nitrate concentrations was measured during the rainy season which agreed with the findings of Tibebe et al. (2018), Wagaw et al.(2021), Symader andBierl (1998), andKalaiarasi et al. (2012) and which could be due to external nutrients inputs and overland flow during this period.In contrast, the lower nitrate level in the post-rainy period could be due to conducive atelomictic mixing event which increased the consumption of nitrate by photosynthetic organisms (Gogoi et al. 2019).
The mean concentration of TP and SRP in Lake Ardibo are in the intermediate range as compared to other Ethiopian highland Lake Tana (0.862 mg L À1 , 0.326 mg L À1 Tibebe  et al. 2019) in the rainy season.However, TP did not show any seasonal pattern.In the contrary, the concentration of soluble reactive phosphate fluctuated temporally with a seasonal peak during the rainy periods, probably due to the diffuse sources and allochthonous input through surface runoff (Becker et al. 2009) and internal loading through water column mixing.
The possible reason for the high accumulation of SiO 2 in the post rainy season is attributed to surface mixing and relative low abundance of diatoms (Rao et al. 2018) or because of sediment and dissolution processes of volcanics, dominantly basalt of the area as described by Demlie et al. (2007).The present study found that the Secchi-depth measurements varied between sampling periods.The low transparency measurements in the pot-rainy season might be due to settling of the silt that entered the lake through runoff in the rainy period (Tibebe et al. 2019).

Phytoplankton composition, abundance
Among the phytoplankton taxa, diatoms with 23 species were the most species-rich, which conforms with a previous observation made by Belokda et al. (2019), Kumar et al. (2020) and Wagaw et al. (2021).The lowest species diversity was recorded in Class Cryptophyceae with one genus (Cryptomonas).The diatoms flora contributed the highest number of species during this and in previous studies.Kumar et al. (2020) and Wagaw et al. (2021) reported that Chlorophyta had highest contribution followed by Bacillariophyta to the total number of phytoplankton species.A similar observation was made by Bode et al. (2017) in their investigation that the diatom flora was the dominant taxa, implying that high nutrient supply and turbulence conditions prevailed.Throughout the study period, blue-green algae had a high relative abundance contribution to the total phytoplankton, which was dominated by species of Aphanothece minutissima and Cyanodictyon endophyticum.The abundance of phytoplankton taxa in the aquatic ecosystem gives some insight about its water quality status (Celekli and € Ozt€ urk 2014).The dominance in the composition of diatom flora could be associated with moderate nutrient enrichment of the water (Sabir and Fakhruddin 2007;Basavaraja et al. 2012; Inyang and Wang 2020), and the higher abundance of blue-green algae can be used as bio indicators of changing environmental conditions of the lake (Okoth et al. 2009;Basavaraja et al. 2012;Kumar et al. 2020).

Seasonal changes of phytoplankton communities and biomass
Variations in the composition and abundance of phytoplankton community at different seasons are associated with environmental drivers, including major algal nutrients, light availability, temperature, grazing pressure, water mixing regimes, and hydrological patterns (Reynolds et al. 2001;Schagerl and Oduor 2008).Our results showed clear seasonal variations in species composition and abundance of different taxonomic groups of phytoplankton.Boyer et al. (2000) asserted that assessment of chlorophyll-a concentration is an important indicator and scientifically meaningful to monitor water body changes with various environmental stressors.During the post rainy season, the algal taxa composition and biomass production were high because of good light conditions (Wondie et al. 2007), possible atelomicitic mixing in the surface water column and related to stable hydrographical conditions (Babu et al. 2013).On the contrary, algal diversity and chlorophyll-a production was limited in the main rainy season, possibly attributable to the reduction of light penetration from the surface down to the water column due to the presence of high suspended matters that comes from the lake catchment through surface runoff and effect of heavy rainfall (Wondie et al. 2007;Kumar et al. 2020).The phytoplankton biomass recorded in the present study is lower than in earlier reports (Fetahi et al. 2014;Degefu and Schagerl 2015;Tibebe et al. 2019;Tessema et al. 2020).The low phytoplankton biomass values of Lake Ardibo are probably related to the taxonomic composition of predominant algal groups, and their average cell size, which accounts for the variation in the Chl-a biomass (Schagerl et al. 2022) and the overall ecological deterioration that the lake has undergone since the previous studies.
The present study observed a seasonal pattern in phytoplankton community succession.Seasonal shifts between the picophytoplankton and microphytoplankton groups were noticed by which the picophytoplankton mostly Cyanophyta and Chlorophyta were prevailing during the dry and pre rainy season, while the contribution of microphytoplankton (Bacillariophyta) increases to main rainy period and decreases in the dry season (Figure 2(b)).The seasonal cycle of phytoplankton species could be driven by changes in environmental variables such as, nutrient concentrations, water temperature, light intensity and water column stability (Oseji et al. 2018).The increase in nitrate concentrations, especially during the long rainy season could promote the growth of microphytoplankton mainly the diatoms.It is plausible to suggest that, the high abundance of copepods (Adem et al. submitted for publication) and their grazing activity during the dry season may stress the growth of large size group of phytoplankton as reported by (Oseji et al. 2018).
The highest peak abundance of phytoplankton appeared in the pre rainy season following mixing periods represented by small coccoid colonial algae (Aphanothece minutissima and Aphanocapsa), Cyanodictyon endophyticum, and also the heterocystous species from the filamentous genera of Anabaena ambigua.Temperature and nutrient concentration are significant factors facilitating the succession, distribution, and composition of phytoplankton communities.The higher temperature (pre rainy months) and high rainfall (main rainy months) most likely favored the growth and dominance of filamentous cyanobacteria due to increasing temperature and nutrient inputs from land runoff (Jindal et al. 2014;Striebel et al. 2016;Liu et al. 2021).The dominance of blue-green algae coincides with the high levels of nutrients, mainly nitrate and phosphate, during and after mixing periods group is in good agreement with Cardoso et al. (2017), who showed that the species has efficient pigments that favor them owing to their structural and physiological adaptation for buoyancy regulation during mixing events.As discussed above, the peak abundance of cyanobacteria after the lake mixing confirmed the poor water quality condition and contribute to the disappearance of Peridinium from the lake during the dry season (Zohary et al. 2012).

Response of phytoplankton community structure to water quality parameters
The RDA analysis exhibited that the phytoplankton species' abundance in Lake Ardibo is influenced by physicochemical variables at different sampling sites.The dominant group of Bacillariophyceae (Fragilaria sp., Aulacoseira, Epithemia sorex, Pinnularia viridis, and Gomphonema acuminatum) mostly negatively correlated with EC, NH 3, water temperature, and pH.Similar observations were noted in Hongmen Reservoir by Liu et al. (2021), who found a negative association of diatom density with water temperature.
Species of Synedra ulna and Cymbella sp.responded explicitly well to the NO 3, SiO 2 , TP, Alka, and reactive phosphate (SRP) at OP (Figure 6) and others (Navicula antonii.and Nitzschia dissipata) influenced by turbidity and DO.At the same time, Synedra and Fragilaria had a close relation to increased SiO 2 .A similar pattern of positive correlation between diatoms, alkalinity, nitrate, molybdate silicon, and total phosphorus was reported by Fetahi et al. (2014), El-Serehy et al. (2018) and Wagaw et al. (2021).Recent studies noted that the structure of diatom community in Lake Shalla and inflowing hotsprings was controlled by pH, EC and DO (Wagaw et al. 2022).The present results corroborated with the previous studies as small-sized cyanobacteria and green algae species rapidly outcompete at higher temperatures while the proportion of diatoms responds positively with decreasing temperature (Schabh€ uttl et al. 2013;Kozak et al. 2015;Rao et al. 2018;Duong et al. 2019).The available NO 3, SiO 2 , TP, Alka, and SRP in the open water are closely associated with the abundance of Charophyta and Dinophyceae.These heavy phytoplankton taxa are known to dominate during periods of atelomictic mixing in deep tropical lakes (Barbosa et al. 2011).
The phytoplankton biomass at littoral areas was positively and moderately correlated only with turbidity and oxygen.The observed greater association of chl-a concentration with turbidity at the study station is consistent with the observations made in Mohammadabad Reservoirs, Gorgan, Iran (Kordi et al. 2012).During the post rainy period of 2020 the lake showed marked seasonal influence of the diatom flora with high biomass production.Rao et al. (2018) stated that pH had a direct positive impact on Chl a concentration in the aquatic environment.The present results suggested that the low phytoplankton biomass production in most of the sampling seasons could be attributed to negative impact of pH measurements.The higher pH value in all sampling period (Table 1) could attributed to limit available CO 2 in the system required for photosynthesis production by phytoplankton (Okello and Kurmayer 2011).During the pre-rainy season, the predominance of Cyanobacterial groups was strongly correlated with partial water mixing driving factors, especially the increased temperature and total phosphorus concentrations (Figure 7).Cyanobacteria filamentous and colonial groups correlated to high temperature and nutrients (Jindal et al. 2014;Striebel et al. 2016;Cardoso et al. 2017;Liu et al. 2021) and also due to grazing resistance (Haney 1987).However, in the dry and rainy months, the abundance of phytoplankton species was limited and negatively influenced by NH 3 concentrations.Thus, during rainy periods when turbidity was high, and water column light penetration was low, diatom growth was lower, leading to decreased relative abundance within the phytoplankton community (Duong et al. 2019).

Atelomictic mixing and phytoplankton functional groups
Lake Ardibo is one of the tropical crater lakes found on the northern mountain chains of Ethiopia and is a reasonably deep Lake with (Z max ¼63.9m) enough to become seasonally stratified.Barbosa et al. (2011) and Cardoso et al. (2017) discussed the effect of the diel and seasonal thermal regime on the vertical distribution and composition of phytoplankton groups in some tropical and subtropical lakes in South America.Lake Ardibo showed seasonal mixing during the sampling time of the dry season (February 2021).Similar sudden mixing events happened in January 2008 at Lake Hayq (Fetahi et al. 2011) and complete mixing in the dry season (Fetahi et al. 2014), and the same phenomenon was observed at a tropical mountain crater lake Wonchi (Degefu et al. 2014).Microscopic inspection of the dominant taxa during the dry season confirmed that, Cosmarium contractum was observed most frequently; this phenomenon could be associated with partial atelomixis of the lake in congruence with previous observations in deep tropical lakes.For instance the success of Cosmarium laeve in Lake Kivu was related with repeated partial mixing of the water column (Rugema et al. 2019) that favors the taxa to have the capacity for buoyancy regulation due to their structural and physiological adaptation mechanisms and grazing resistance from zooplankton (Stoyneva et al. 2012;Rugema et al. 2019).In addition massive number of Peridinium spp.were observed from October until December (post rainy season) and with small number from July to September (main rainy months).The occurrence of Peridinium spp.during the post rainy period and high rain fall season most likely nutrient reach organic soils from the lake catchment of agricultural areas (Zohary et al. 2012), especially associated with high Ammonium levels during such period (Fetahi et al. 2014).The lower occurrence of the species in the dry and pre rainy seasons could be associated with the effect of allelochemicals that are released from cyanobacterial groups of algae (Vardi et al. 2002), beside the lack of complete mixing and nutrient limitations during these periods.
Interestingly, atelomixis-dependent desmids, mainly the diatoms and Charophyta (Cosmarium contractum), mostly occurred in the epilimnion of Lake Ardibo (Figure 4) throughout the sampling periods.These study findings agreed with previous research works (Gunkel and Casallas 2002;Tavera and Mart ınez-Almeida 2005;Souza et al. 2008;Fetahi et al. 2014;Degefu and Schagerl 2015) which pointed out the role of atelomixis as driving force which enable planktonic desmids species to remain within the epilimnion zone.According to Tavera and Mart ınez-Almeida (2005), partial atelomixis has ecological contributions to sustaining desmids in the deep layers of temperate lakes under limited light and nutrient conditions.

Conclusion
In summary, ANOVA results confirmed that most environmental variables showed significant variations among the sampling seasons.The maximum number of phytoplankton communities was counted in the post rainy period.The peak abundance of phytoplankton was observed during the pre-rainy season due to complete atelomixic mixing but lowest in the main rainy months because of turbidity, low stability and runoff effects.Incomplete atelomixis might be the driving force for the presence of heavy phytoplankton groups in this relatively deep tropical lake ecosystem.As evidenced from the RDA triplot, the growth and abundance of phytoplankton were mainly governed by physico-chemical dynamics which was seasonally influenced by atelomictic mixing in this deep tropical lake.Further, this study gives baseline information for the future sustainable management plan of the lake resources.

Notes on contributors
Adem Mohammed completed his first Degree in Biology at Haramaya University in 2009, Master of Science in Limnology, Fisheries and Aquatic Ecotoxicology from Hawassa University in 2012.He is a Ph.D. candidate in Water Management specializing in Aquatic Ecosystems Management at Addis Ababa University, Ethiopia.He has 5 peer-reviewed published articles.Adem is currently working as lecturer and researcher at Wollo University.
Professor Seyoum Mengistou holds a Ph.D. in Biology (Limnology) from Waterloo University, Ontario, Canada.He has been Head of Department of Biology, Addis Ababa University, Director of Zoological Natural Museum, Editor-in-Chief of the Ethiopian Journal of Biological Sciences, Founding member of the Ethiopian Fisheries and Aquatic Sciences Association (EFASA), Member of the Biological Society of Ethiopia, and Member of Society International of Limnology.He has mentored and supervised over 20 Ph.D. students and has published over 95 articles in reputable international and national science journals.
Dr. Tadesse Fetahi has a Ph.D. in Aquatic Ecology from Vienna University in 2010, Austria and he has over 17 years of professional experience.He has been the President and Founding member of the Ethiopian Fisheries and Aquatic Sciences Association (EFASA).Dr. Tadesse has been engaged in the design and effective implementation of several national and international collaborative research projects, much of which is used to support graduate students.Currently, he is serving as Director of Research Office of Addis Ababa University.

Figure 1 .
Figure 1.Location of Lake Ardibo and sample sites.
(b)).Generally, there were changes in the contributions of the different phytoplankton over time, with the order of Cyanophyta > Chlorophyta > Bacillariophyta in the pre rainy season, Bacillariophyta > Cyanophyta > Dinophyta > Chlorophyta in the main rainy season, Charophyta > Bacillariophyta > Chlorophyta > Dinophyta > Cyanophyta in the post rainy season, Cyanophyta > Chlorophyta > Charophyta > Bacillariophyta in the dry season (Figure 2(b)).Spatially, the mean abundance of phytoplankton was varied.The peak algal abundance were found in the open water, followed by GOR and MEN (Figure 3(a)).The highest abundance of phytoplankton recorded in the short rainy period, and the lowest mean abundance occurred during the main rainy season (Figure 3(b)).

Figure 2 .
Figure 2. Mean percentage contribution of algal groups to total phytoplankton abundance at the different sites (a) seasons (b) of Lake Ardibo during (October 2020-September 2021).
(a)).On the other hand, seasonal variations of phytoplankton biomass was significant (p < 0.05) with Chl-a concentrations ranging from 0.58 mg L À1 in August (2021) main rainy season to 8.44 mg L À1 in October 2020 (Figure 5(b)) post-rainy period.The peak of phytoplankton biomass observed in the postrainy season did-not coincide with peaks of total phytoplankton cell abundance that

Figure 4 .
Figure 4. Vertical distribution of phytoplankton groups at different depths of Lake Ardibo (October 2020-September 2021).

Table 1 .
Seasonal variation of physical parameters of Lake Ardibo, Ethiopia.

Table 2 .
Seasonal variations of inorganic nutrients from Lake Ardibo, Ethiopia.

Table 3 .
List of identified phytoplankton species collected from Lake Ardibo, Ethiopia.

Table 4 .
Redundancy analysis result of phytoplankton species versus sites and physico-chemical parameters.

Table 5 .
Redundancy analysis results from phytoplankton species versus seasons and physico-chemical parameters.