Spatiotemporal distributions of potentially toxic elements’ dynamics in Lake Tana and its major tributary rivers, Ethiopia, Upper Blue Nile Open Access

Water is a vital resource for all forms of life, including aquatic organisms. Human lifestyles are closely interconnected to lakes and rivers, and historically, civilizations have developed around these water bodies (Liu et al. 2014; Gibling 2018). Most developmental activities occur near rivers, wetlands, lakes, and other aquatic ecosystems because they offer a range of services that contribute to economic growth (Brummett et al. 2013).

Quality water ecosystems play a significant role in poverty reduction and in achieving the sustainable development goals, particularly SDG 6 (clean water and sanitation), SDG 14 (life below water), SDG 15 (life on land), SDG 2 (sustainable agriculture and food production), and SDG 11 (urban development and resilience (Bhaduri et al. 2016; Yang et al. 2020). However, introducing harmful substances into these ecosystems and encroaching on their habitats disrupts their normal functions and benefits. One pressing concern is the presence of potentially toxic elements (PTEs) in aquatic ecosystems, which significantly impact both the ecosystem and its inhabitants (Saeed et al. 2023).

PTEs refer to harmful elements. Their dynamics are affected by various factors, such as agricultural runoff, industrial activities, urbanization, and mining (Daripa et al. 2023; Das et al. 2023). The major sources of metal contamination include leaching and the release of metals from landfills, livestock and poultry manure, road work, automobiles, and other urban-related activities (Khairunnash et al. 2022). Additionally, agricultural practices, such as the use of pesticides (including herbicides, insecticides, and fungicides), fertilizers, and plastics, contribute significantly to potentially toxic metal pollution (Briffa et al. 2020). For instance, agricultural phosphorus fertilizers have been identified as a source of cadmium (Cd), copper (Cu), and zinc (Zn) (Alengebawy et al. 2021).

Once PTEs enter aquatic ecosystems, they undergo various processes, including adsorption, bioaccumulation, and sedimentation (Muluye et al. 2024). The levels of these processes depend on the concentration and distribution of elements in the ecosystem. Moreover, the dynamics of the elemental composition, in terms of quality and quantity, can be influenced by climate change (Soomro et al. 2023). Research studies have shown that PTEs are prevalent in lake water and sediment, due to anthropogenic factors such as agriculture and flower farming (use of intensive fertilizers and pesticides), and urbanization (discharges of untreated wastes and emissions) (Niu et al. 2021), the spatiotemporal studies also showed that the PTEs have higher trends near urban peripheries and industrial surroundings (Jiang et al. 2022).

The concentrations and distributions of elements in lakes and their tributary rivers can vary spatially and temporally, depending on hydrological factors, such as water flow and ecological factors (Zalewski 2002). The biology of lakes and rivers also regulates the hydrological content of metals (Zalewski 2002). When the concentration of elements exceeds maximum permissible limit (carrying capacity) of the ecosystem, the ecosystem's function becomes disturbed. This is directly related to plankton dynamics. PTEs may significantly affect the plankton community, impacting its abundance, diversity, and overall community structure. This, in turn, directly affects the fish community, from feeding to reproduction. Scientific evidence has shown that high concentrations of PTEs can impair the reproduction, feeding habits, and growth rates of plankton communities (Aránguiz-Acuña et al. 2018). These problems escalate through the food chain, affecting the food web and increasing trophic levels (Saidon et al. 2024). Changes in plankton dynamics have implications for energy flow and the geochemical cycle (Quere et al. 2005). Moreover, PTEs in aquatic ecosystems impair the developmental growth of fish, enhances the development of anomalies, and declines the fish survival rate, potentially leading to fish extinction (Khayatzadeh & Abbasi 2010). The PTE concentration has not yet been determined in the largest lake, Tana, Ethiopia.

The study area, the Lake Tana subbasin, is the largest subbasin within the Abay Basin. It is a national growth corridor, experiencing escalated agricultural farming in its watershed, flower farming, intensive recession local farming, and nearby urbanization (Minale 2020). It also provides intensive irrigation purposes for local communities, such as Megech-Seraba, Tana Asrate, Tana Mekonta, Tana Wenjeta, and Tana Zegie (Mequanent & Mingist 2019). These practices have been reducing the inflow to the lake (Tesfaye 2023). Furthermore, in most practices, the community uses fertilizers, pesticides, and metal-emitting inputs without expert recommendations. As a result, the lake and its tributaries are highly vulnerable to pollution. Consequently, the quality of water is deteriorating and might cause public concerns (Ewnetu et al. 2014). Meanwhile, the lake water serves as a drinking water source for humans and animals, as well as for other domestic purposes for the nearby communities (Banchamlak 2021; Tesfaye 2023).

Previous studies have also shown that the water quality of the lake has deteriorated over time, particularly in urban and agricultural runoff routes (Goshu et al. 2017; Banchamlak 2021). For instance, the lake is highly infested with water hyacinth in the north and eastern parts of the lake (Dersseh et al. 2019), which has an implication of pollution presence, and decreases its potential for fish populations (Tewabe 2015). A recent study indicated the existence of PTEs in the southern gulf of the urban peripheries of the lake (Sishu et al. 2024). The other study done in one of the major tributary rivers, Megech, indicated a significant amount of metals from the watershed, particularly Gonder town periphery (Engdaw et al. 2022). Apart from such proxy studies, Lake representative research has not been conducted on PTEs in the Tana subbasin.

Despite growing concerns over pollution, comprehensive baseline data, regarding the concentration and dynamics of PTEs in Lake Tana and its tributaries, remain lacking. Given Lake Tana's vital role as a source of drinking water, agricultural irrigation, recreational values, and fisheries, this research aimed to (1) examine the spatial and temporal dynamics of PTEs within the lake system and its watershed; (2) identify major anthropogenic sources of PTE contamination, including agricultural, industrial, and urban activities; (3) assess the ecological impacts of PTEs on aquatic organisms, particularly plankton and fish communities, and (4) provide evidence-based recommendations for sustainable lake and watershed management under increasing pollution pressure.

The study was conducted in Lake Tana in Ethiopia, which is located at the latitude of 12° 0′ N and longitude of 37° 15′ E. Lake Tana is the largest freshwater lake in the country, with a surface area of 3,046 km,² a volume of 29.6 km,³ and a shore length of 431 km (Kebedew et al. 2020). It is a crater lake formed two million years ago, due to the volcanic blocking of the Blue Nile River (Gashaye 2016), and has been recorded as the UNESCO first biosphere in the country since 2014. It has seven medium to large perennial and 40 small seasonal tributary rivers that drain into the lake. Among perennial rivers, the Gilgel Abbay, Megech, Dirma, Gumera, and Rib rivers account for 95% of the inflow to Lake Tana (Vijverberg et al. 2009). The lake has a mean depth of 9.7 m and a maximum depth of 14.8 m. The subbasin receives 1280 mm of mean annual rainfall (mm year⁻¹) (Abebe et al. 2017), with 250–330 mm of maximum rainfall per month during July and August, which accounts for nearly two-thirds of the annual precipitation. The average seasonal air temperature reaches 21.1 °C in the dry season and 18.4 °C in the rainy season.

The surrounding wetlands and lakes have rich biodiversity, including various flora, fauna, endemic fishes, such as Labeobarbus, and birds, such as globally threatened waterfowl. However, the lake and surrounding areas are highly pressurized by the communities used for recession farming, grazing, boat making, fishing, and so on (Amsalu et al. 2007). The land use types in the lake basin are farmlands, water bodies, wetlands, forests, shrubs, rangelands, grasslands, and settlements. Among them, 71% were cultivated land, 9% were grazing, 6% were infrastructure, 3% were forest, and 11% were other land types (Stave et al. 2017).

The subbasin is predominately characterized by different soil types based on elevation. In the highlands, nitosols and cambisols are prevalent; these soils are fertile and well-drained. At mid-elevation, luvisols and vertisols are found, which are moderately fertile and used for farming activities, particularly cereal cultivation. In the lowland plains and river mouths, fluvisols and vertisols are dominant, consisting of clay-rich soils (Setegn et al. 2009; Stave et al. 2017).

To determine the PTEs in the lake and tributary rivers, water samples were collected following standard sampling methods (American Public Health Association 1998) from the study sites twice a year during the dry season (March–April 2023) and wet season (August–September 2023).

Water samples were collected via clean, acid-washed polyethylene bottles, starting from a depth of 50 cm to the surface at a uniform rate, allowing time to overflow before being covered with screw caps. Before sampling, the bottles were rinsed three times with water to ensure that the media were similar. The samples were then stored in an ice chest at 4 °C and transported with thermo-insulated containers to be frozen at −20 °C in a deep freezer. The analysis was carried out at HORTICOOP ETHIOPIA PLC, Bishoftu.

Fifty milliliters of each water sample was taken into a digestion vessel, and 4 ml of concentrated nitric acid (HNO₃) and 1 mL of hydrochloric acid were added to the sample. The samples in the digestion vessels were placed on a hot plate, and the mixture was heated at 250 °C for 3 h until the digestion volume was reduced to 25 mL. After digestion, the sample-containing vessels were allowed to cool to room temperature and filtered through a filter paper in a 50 mL Erlenmeyer flask to remove any particulate matter. Finally, the filtered samples were refilled with deionized water in a 50 mL standard flask to ensure a consistent volume for analysis. All elements, such as Ca, Na, Mg, K, Ni, Mn, Fe, Cu, Co, As, Pb, Cd, Zn, Hg, Cr, and B, were analyzed through direct aspiration of the sample mixture into an Inductively Coupled Plasma Optical Emission Spectrometer (ICP‒OES, ARCOS ICP-OES, Germany).

Before starting the operation of the ICP–OES instrument, several parts were checked, including the instrument condition, gas supply, sample introduction system, cooling system, optical system software and calibration, and safety checks. To start analyzing PTEs, the nebulizer, spray chamber, torch, sample introduction tubing, glassware, and autosampler cups were cleaned with 10% v/v HNO₃, thoroughly rinsed with deionized water, and allowed to dry. All the measurement conditions used for the ICP‒OES were as follows: plasma power (1,400 W), pumping speed (30 rpm), coolant flow (13 L/min), auxiliary flow (0.8 L/min), nebulizer flow (0.73 L/min), optical temperature of 15.05 °C (range: 14.0–16.0 °C), OSC exhaust (285.70 Imp/S) (minimum: 170.0 lmp/s), OSC temperature of 51.53 °C (range: 40–70 °C), OSC impedance of 5,365 Ohms, HVPS current (607 mA), HVPS voltage (3,260 V), flow light tube (0.90 L/min), nebulizer pressure (1.86 bar), and main argon pressure (7.75 bar).

The calibration and standardization of the spectral method were performed following the standard protocols of the instrument (Agilent Technologies 2018). Standardization was performed daily, which is a quick procedure that corrects the measurement intensity. Thus, the correct concentrations were obtained via the original calibration curve. The relationship between concentration and intensity had a strong positive correlation coefficient (r) across a range of wavelengths (184.95–324.754 nm) for each PTE. The minimum detection limits of the instrument ICP–OES ranged from 0.00043 mg/L (As) and 0.01257 mg/L (Co). The instrument detection limits, wavelengths, and correlation coefficient (r) values at the time of operation are described in Table 1.

The HPI was calculated and evaluated for nine PTEs: Cd, Pb, As, Cr, Ni, Cu, Zn, Mn, and Fe.

The critical pollution index value or threshold value is set at 100 (Kone et al. 2019). An index value below 100 indicates low levels of heavy metal contamination, posing no health risks. When the index reaches 100, it signifies a threshold risk, suggesting potential adverse health effects. If the HPI value exceeds 100, the water is deemed unsuitable for drinking and domestic use (Tiwari et al. 2015; Kone et al. 2019).

The instrument was calibrated with standard solutions of known concentration for each of the 16 metals by Sigma-Aldrich. The prepared black and control samples were run to ensure the accuracy and precision of the measurements. Triplicate samples, spiked samples, and standard reference materials were used to validate the quality of the analysis. All stock solutions were imported from an inorganic venture in Germany.

IBM SPSS (version 25), R software (3.4), and ArcGIS (10.8.2) were used for statistical analysis and plotting. Spatial and temporal variations in the PTEs were determined. The associations between PTEs in the water across the sampling sites were examined by canonical multivariate analysis (CANOCO) for Windows 4.5 (Lepš & Šmilauer 2003).

In the current study, the concentrations of the above four metals were measured in the dry season, with Ca ranging from 13.128 to 29.5 mg/L, Mg from 3.319 to 20.657 mg/L, K from 0.49 to 1.72 mg/L, and Na from 4.042 to 20.068 mg/L. In the wet season, the concentrations ranged as follows: Ca from 11.715 to 44.737 mg/L, Mg from 2.479 to 16.693 mg/L, K from 0.621 to 1.983 mg/L, and Na 3.709 to 12.985 mg/L (Figure 2). Overall, these findings were generally within the optimum ranges at most sampling sites in both seasons, except for K, which fell below the optimum levels in each season. According to the EPA (1999), the essential elements should be maintained within the following optimum ranges: Ca (20–100 mg/L), Mg (2–20 mg/L), K (2–20 mg/L), and Na (10–100 mg/L) for the aquatic ecosystem to function effectively.

At four sites (S2, S7, S8, and S15), the concentrations of Na, Ca, and Mg elements were lower. Potassium levels were also lowest at six sampling sites (S2, S1, S3, S7, S8, and S16). The highest mean concentrations of magnesium and potassium were recorded at the Megech River sampling point S9, with values of 20.282 mg/L ± 0.521 and 1.683 mg/L ± 0.026, respectively. In contrast, the highest mean concentration of Na was found at Dirma River sampling point S11, 18.992 mg/L ± 1.231, with a range of 17.563–20.068 mg/L during the dry season. Additionally, the highest mean concentration of calcium was also at the Dirma River site (S11), recorded at 44.488 mg/L ± 0.217, with values ranging from 44.338 to 44.737 mg/L during the wet season.

The high concentrations of Ca, Mg, and Na from the analyzed samples in the two main tributary rivers, Megech and Dirma, might be due to their origin in the Armachiho highlands and the tectonic geological characteristics of the area that inhabited the carbonated elements (Nigate et al. 2023) that regulate the lake demand of the essential elements, as illustrated in Figure 2. Similar findings were reported in a study conducted on seven Ethiopian rift valley lakes (Zinabu 2002). The elevated concentrations during the dry season could be attributed to the natural geological sources of nutrients and the lower influx of water from the catchment. However, these findings are under the permissible limits of WHO guidelines (Fortin 2023). Those metals are essential for human health and play substantial roles in various body functions, such as bone health, nervous function, and fluid balance (Zoroddu et al. 2019). Both sodium and potassium are alkali metals that are highly reactive in water and are essential for nerve function and maintaining fluid balance in the body. Zoroddu et al. (2019) described magnesium and calcium, both alkaline earth metals, are important for bone health, muscle function, and enzyme activity.

The mean concentrations and standard deviations of six trace elements iron (Fe), manganese (Mn), zinc (Zn), copper (Cu), cobalt (Co), and nickel (Ni) in water samples collected from study sites during both the dry and wet seasons are presented in Tables 1 and 2, respectively. These trace elements are essential metals for various biological processes in aquatic organisms when present in small quantities (He et al. 2005). However, if these metals exceed permissible limits in aquatic ecosystems as indicated in Table 3, they can become toxic, leading to impaired metabolism, disrupted osmoregulation, stunted growth, reproductive issues, and other health complications (Gomes et al. 2019).

The concentrations of Fe, Mn, Zn, Cu, and Co tended to increase during the wet season, whereas Ni was highest during the dry season shown in Supplementary Material Figure 1. This might be due to anthropogenic activities and the inflow of trace elements from the upper catchment of the lake. The elemental concentration levels are described independently as follows:

The concentration of iron varied range from 0.39 ± 0.0 to 61.03 ± 0.43 mg/L in the dry season while it was in the range of 1.37 ± 0.03 to 34.87 ± 0.63 mg/L in the wet season (Tables 3 and 4). The highest mean concentration of Fe was 41.12 mg/L at sampling site S4, with values ranging from 21.2 to 41.04 mg/L. In contrast, the minimum mean concentration was 1.17 mg/L at the S1 sampling site, with values ranging from 0.97 to 1.37 mg/L during the dry season. There was also significant spatial variation during the wet season (p < 0.05), with the lowest mean concentration of 1.37 mg/L measured at the sampling site S1 and the highest mean concentration of 34.87 mg/L from sampling site R. The concentration in the wet season was significantly lower than that in the dry season (P < 0.05). However, the concentrations of iron in all the sample sites in both seasons were higher than the maximum permissible limits of the WHO standard (0.2 mg/L) and (FAO 2022) (0.3 mg/L). This might also be associated with the overall high runoff of iron-containing inputs from tributary rivers as well as the geological characteristics of the lake itself. Similarly, a study by Engdaw et al. (2022) reported a high concentration of iron (4.58 mg/L) in the Megechi River within the Lake Tana subbasin.

Table 4

Pearson correlation matrix and level of significance of 16 PTEs in water samples from Lake Tana and its tributary rivers

View Large

The tributary rivers were contributing to the lake in the following order: Rib (34.87 mg/L), Gilgel Abay (22.71 mg/L), Gumera (20.21 mg/L) and Gelda (20.18 mg/L), respectively. This is directly related to watershed characteristics, which include intensive farming and Fe reach areas that feed Fe into the lake through runoff. The lowest concentrations of Fe (1.37 and 2.15 mg/L), were recorded on the Bahirdar city side of the lake, which experiences urbanization and few farming activities in its watershed. A concentration of Fe above 14 mg/L has also been reported in Indian surface water in the Nega Thrust of Assam-Arakan Basin rivers and ponds (Kshetrimayum & Hegeu 2016). There were no significant differences (P < 0.05) in the Fe concentration distributions across seasons. However, there was a significant difference (P < 0.05) in the spatial distribution of Fe among the study sampling sites.

The average concentration of Fe (0.97–61.04 mg l⁻¹) in the present study was higher than the concentration found in River Awash, Ethiopia (1.11–4.12 mg/L) (Eliku & Leta 2018). The presence of iron above the limit may inhibit the absorption of essential nutrients such as Ca and Zn by aquatic organisms as well as those who drink excessive amounts of Fe-containing water (Gurzau et al. 2003), although Fe is crucial for the existence of life through oxygen transport (hemoglobin) and electron transfer.

The average concentration of Mn was 0.24 to 0.27 mg/L. The concentration and/or contamination of Mn may occur naturally in aquatic ecosystems. It may also be found in aquatic ecosystems associated with industry or the periphery of urbanization. High concentrations of Mn in aquatic ecosystems may cause toxicity. The study conducted in an Indian surface water oxbow lake reported a higher concentration of 4.242 mg/L Mn than the current study (Kshetrimayum & Hegeu 2016).

The highest mean concentration of cu (0.136 mg/L) was found in the wet season, with values ranging from 0.135 to 0.138 mg/L at the S8 sampling site on the Rib River side of the lake (Table 3). On the other hand, the lowest mean concentration of Cu in the wet season was (0.072 mg/L) at the S10 sampling site on the Megech River side of the lake. During the dry season, there was also significant variation among the sampling sites (P < 0.05), with the highest value of 0.127 mg/L at sampling site S4 and the lowest value (0.025 mg/L) at the OT2 sampling site (Table 2). The concentration of cu showed a significant temporal variation (P < 0.01). However, the specific variation did not significantly change. Similarly, Cu (0.11–0.17 mg/L) was reported in one of the major tributary rivers of Tana Lake upstream of the Megech River (Engdaw et al. 2022). In the present study, the mean concentration of Cu (0.025–0.136 mg/L) was similar to that reported by Mola (2018) in the Awash River. When the concentration of Cu exceeds the permissible limit, it has significant health impacts on humans through accumulation in organs, causing Wilson's disease, instead of being removed from the body, resulting in excreta, acute gastroenteritis, hepatic and kidney disease, and potentially death (Briffa et al. 2020).

The highest mean concentration of nickel (Ni) was measured during the dry season at 0.328 mg/L, with values ranging from 0.325 to 0.33 mg/L at the S16 outlet sampling site of the lake (Table 2). In contrast, the lowest average concentration of Ni was found at the S5 sampling site, where it was 0.028 mg/L. Additionally, there was significant variation in Ni concentration during the wet season (p < 0.05), with the highest value recorded at sampling site S7 at 0.073 mg/L and the lowest at the S16 outlet of the lake, which was 0.038 mg/L (Table 3). The analysis indicated significant seasonal variation (P < 0.05) in the distribution of Ni concentrations. The present study revealed that the average concentration of Ni in Lake Tana (0.038 to 0.33 mg/L) was greater than the concentration of Ni found in the River Awash, Ethiopia (0.02–0.2 mg/L) (Eliku & Leta 2018), and in Lake Asejire (0.2 mg/L) (Utete & Fregene 2020). The concentration of Ni was below the regulatory limits, with an average of 0.12 mg/L in the dry season and 0.05 mg/L in the wet season. The Ni concentration was significantly higher at the outlet of Lake Tana in the Abay River, whereas its concentration is relatively lower at Gumera S5 and outlet two S17 sites compared to other sites. The lowest Ni concentration, 0.039 mg/L, was measured on the Gumera side of the lake and in the center of the lake. The highest concentration of Ni is associated with urbanization and limited agricultural activities in the watershed. This is directly related to watershed characteristics, including intensive farming and nickel reach areas that contribute Ni to the lake through runoff.

The trace elements mentioned above are essential metals for various biological processes in aquatic organisms when present in small quantities (He et al. 2005). However, if these metals exceed permissible limits in aquatic ecosystems as indicated in Table 3, they can become toxic, leading to impaired metabolism, disrupted osmoregulation, stunted growth, reproductive issues, and other health complications (Gomes et al. 2019).

The concentrations of the three detected PTEs (Cd, B, and Cr) generally increased at the sampling points during the dry season compared to the wet season. In contrast, Hg levels rose at the sampling points during the wet season compared to the dry season. Furthermore, Pb and As exhibited inconsistent trends at the sampling points in both seasons (Figure 3).

The highest mean concentrations of lead (Pb) were found at the S4 and S10 sampling sites during the dry season, with values of 0.234 and 0.209 mg/L, respectively. These concentrations ranged from 0.205 to 0.256 mg/L and 0.178 to 0.246 mg/L, respectively (Figure 3). In contrast, the lowest lead concentrations were observed at the S5 and S16 sites, with mean values of 0.03 and 0.047 mg/L, respectively ranges of 0.03 to 0.234 mg/L and 0.025 to 0.064 mg/L. During the rainy season, the highest Pb concentrations were detected at the S7 and S8 sites (Rib River side of the lake), ranging from 0.121 to 0.164 mg/L and from 0.119 to 0.152 mg/L, with average values of 0.14 and 0.136 mg/L, respectively. This increment encompassed the eastern and northeastern parts of the lake, as shown in the spatial analysis in Figure 5. However, all the sampling points had Pb concentrations above the permissible limits set by Codex and WHO for drinking water (Fortin 2023). This contamination may be attributed to the urban peripheries of Gonder Makisegnit and Debirtabore towns, fragile watersheds, and the use of agrochemicals. Like in the current study, Malik & Nadeem (2011) reported Pb concentrations between 0.09 and 0.69 mg/L in the Rawal Lake Reservoir. The other study by Hiruy et al. (2024) also reported Pb concentration in the range of 0.0548–0.066 mg/L in Lake Babogaya, Ethiopia. Moreover, Eliku & Leta (2018) reported Pb concentrations in the Awash River, Ethiopia, ranging from 0.31 to 1.36 mg/L. However, these Pb values did not exceed the international guideline of 5 mg/L for irrigation water (FAO 2022).

The highest mean concentration of arsenic (As) during the dry season was 0.046 mg/L at the GI.T sampling site, with values ranging from 0.042 to 0.048 mg/L (Figure 3). The lowest mean concentration (0.02 mg/L) was detected, at the OT1 sampling site with a ranging from 0.015 to 0.025 mg/L. In the wet season, the highest mean concentration (0.004 mg/L) was detected at the S11 (Dirma River) sampling site, with values ranging from 0.038 to 0.048 mg/L. These concentrations exceeded the international guideline value of 0.01 mg/L for drinking water (Fortin 2023), and the values were nearly equal to the international guideline value of 0.1 mg/L for irrigation water (FAO 2022). The higher values might be associated with the input from the Gelda River, which is affected by flower and Khat farming in the watershed. A previous study conducted by Obinna & Ebere (2019) stated that arsenic and copper levels could result from the practical agricultural use of pesticides.

Boron was detected during the dry season at most sampling points, except at the Gelgel Abay sampling site (Figure 4). The highest concentration was detected at the S6 sampling site, with an average value of 0.48 mg/L, which is on the side of the Gumera River of the lake. However, during the wet season, B was not detected at ten sampling sites, and the highest concentration was detected at the S13 sampling site, with a mean value of 0.196, ranging from 0.186 to 0.201 mg/L. The values of B were below the limits set for drinking water (0.5 mg/L) (WHO 2020; Fortin 2023).

A relatively high concentration of chromium was detected at the S6 sampling site, Gumera River side of the lake, with an average value of 0.502 mg/L, ranging from 0.498 to 0.507 mg/L, during the dry season. This value exceeded the permissible limits set by the FAO/International Water Management Institute (IWMI) at 0.1 mg/L (Drechsel et al. 2023). This might be associated with the agricultural practices in the fogera plain, such as the intensive use of fertilizers and pesticides, and the input from Gumera and Rib rivers into the lake. Furthermore, during the dry season, the reduced dilution effects may increase the concentration of Cr. In the wet season, elevated chromium concentrations were measured at the S8 sampling site, Rib River side of the lake, with an average value of 0.092 mg/L, ranging from 0.085 to 0.1 mg/L comparable to the other sites (Figure 5). This might be due to the nearby recession farming practices and the inputs of tributary rivers of the Rib and Gumera rivers. The lowest concentration of chromium was detected at the S6 and S5 sampling points. They had mean values of 0.008 mg/L for both, ranging from 0.005 to 0.011 mg/L and 0.007 to 0.009 mg/L, respectively. Most samples collected during the wet season were below the permissible limit (0.005 mg/L) for drinking and irrigation water (Fortin 2023). Malik & Nadeem (2011) reported concentrations of chromium ranging from 0.29 to 2.78 mg/L in the Rawal Lake Reservoir. Similarly, Eliku & Leta (2018) reported concentrations of Cr ranging from 0.3 to 0.98 mg/L in the Awash River, Ethiopia.

For correlation analysis, Pearson correlation was performed to determine the relations between the PTEs in the water sample and their possible sources and distributions (Akoto et al. 2021). The Pearson's correlation coefficients and the significance levels (P < 0.01, P < 0.05) for the concentrations of the 16 PTEs in the water samples are shown in Table 4. Pearson correlation analysis is a fundamental statistical method used to assess the associations between variables. The analysis reveals that a correlation coefficient (r > 0.5) indicates a strong association between the metals (p < 0.01). A coefficient ranging from 0.3 to 0.5 indicates a moderate correlation, which was significant at a P-value of 0.05. However, (r < 0.3) indicates a weak or no relationship between the metals.

There were 30 positive correlations and three negative correlations among the 16 metals analyzed in Table 4. Calcium (Ca) exhibited a strong positive association with magnesium (Mg) and sodium (Na) at a P-value of 0.01 and a moderate association with potassium (K). Mg also showed a strong positive association with K and Na. K had a strong association with Na and B and a moderate association with Cd and B. Fe demonstrated a strong positive association with Co, Mn, Hg, Pb, Zn, and Cu in that order. Mn had a strong positive association with Co, Pb, Hg, and Zn but a moderate association with Cu. Zn was strongly positively associated with Co, Cu, and Hg. Cu had a strong positive association with Co and Hg and a moderate association with Pb. In contrast, Cu exhibited a negative association with Ni, with a P-value of 0.05. Ni had a strong positive association with Cr. Co showed a strong positive association with Hg and Pb but a negative association with Cd. Cd also had a negative association with Hg. Hg was positively associated with Pb. Uniquely, As was not correlated with any of the metals in the water samples from the study area.

In this study, the highest correlation coefficients (r > 0.85) were observed as follows: (r = 0.95) between Na and Mg, (r = 0.92) between Cr and Ni, (r = 0. 91) between Hg and Co, and (r = 0.88) between Co and Fe (Table 4). Strong correlations between metals occur in aquatic ecosystems because of the interactions between and within metals (Gomes et al. 2019).

Principal component analysis (PCA) was carried out to reduce the number of detected metals and to simplify and infer potential sources of their contamination in the lake, either from human activities or natural sources (Olando et al. 2020). It also helps in managing complex datasets and obtaining insightful observations for further analysis. The PCA was computed using all the trace metal concentrations found in the water samples during the dry and wet seasons. The output of the four principal components of the PCA explained 85.68% of the total variance extracted at Eigenvalues > 1. Among them, component 1 and component 2 explained 35.7% and 25.61% of the total variation, respectively (Table 5). The line plot of the combined PC indicated the spatial distributions of the metals across the sampling points (Figure 6). This analysis revealed multiple sources of metals from nature on the basis of the geochemical characteristics of the watershed, along with the land use/land cover state, intensive agricultural practices, and other anthropogenic sources of the urban periphery.

Figure 6 shows that the major sources of Na, Mg, K, Ca, Cd, and B are at the Megech and Dirma Riversides. These findings indicate that the sources of these metals are from the watershed of the Armachiho Mountain regime, which serves as the origin of the two rivers. The Gelda River is the major source of As, Cu, Pb, Hg, Co, Fe, Mn, and Zn. The Gelida River has shown major sources of Ni and Cr. The majority of the metal sources are from the nature and periphery of large towns such as Gonder, Debirtabor, and Bahirdar.

In the present study, the loading values (>0.5) were taken as the most significant to the variation. In the first PC axis, PTEs such as Co, Fe, Hg, Mn, As, Cu, Pb, and Zn presented relatively high-loading factors, thus defining PC axis 1. The second PC axis was explained by variables, including Na, Mg, Ca, and K. Eigenvalues, variance, cumulative variance, and loadings for PC1 and PC2 are illustrated in Table 5.

The mean concentrations of the nine PTEs (Fe, Zn, Cu, Cd, Pb, AS, B, and Cr) were used to calculate the HPIs. Table 6 provides the seasonal dynamics of the HPI, along with a detailed analysis that includes the maximum permissible limit, weight of each parameter (W_i), seasonally monitored data, subindices, and the determined HPIs. The critical limit of the HPI is 100 (Milivojević et al. 2016), and if the value of the HPI is below 100, it is a low pollution level that can be used for drinking and other domestic purposes from a chemical composition perspective, whereas if the value is between 100 and 200, its pollution level is medium, whereas above 200, the pollution level becomes high and needs chemical removal or treatment for domestic and drinking purposes.

The mean HPI values of the study area were 371.51 and 434.7 for the wet and dry seasons, respectively (Table 7), which revealed that the study area in both the wet and dry seasons was under high heavy metal pollution levels in the lake, as categorized by Milivojević et al. (2016). This might be due to the presence of PTEs in the water (Edet & Offiong 2002). The pollution level in the dry season is greater than in the wet season. During the dry season, the main contributors to the HPI were the four PTEs; Pb, As, Fe, and Cd in that order, which are among the nine metals included in the HPI analysis. Whereas in the wet season, the three PTEs: Pb, As, and Cd were the major contributors (Table 8). This study's findings align with Ojekunle et al. (2016), who reported a mean HPI value of 518.55, significantly greater than the critical value of 100. This variation is due to the influence of industry and chemical weathering, which contribute substantial amounts of PTEs in the water body. In contrast, Manoj et al. (2012) reported HPI values of the river Subarnarekha in India below the critical pollution index value of 100.

The seasonal dynamics of the metal quality index across the sampling points are illustrated in Table 7. The mean concentrations of the fifteen measured PTEs (Na, Ca, Mg, Fe, Mn, Zn, Cu, Co, Ni, Cd, Pb, As, Hg, B, and Cr) were used to calculate and estimate the MQI of Lake Tana. The threshold for drinking water quality on metal composition is indicated by MQI values less than 1 (Pal et al. 2017).

In this study, the mean MQI values for both the wet and dry seasons were above the limit, with values of 2.36 and 4.412, respectively. During the dry season, a higher MQI value was found at the S4 sampling point on the Gelda River side of the lake, followed by S12, S7, and S8, with values of 7.89, 4.50, 3.50, and 3.04, respectively. This may be due to the fact that the concentrations of most metals were higher than at other sites, as well as the practices and the existence of growth corridors and investments in the Gelda River micro-watershed, along with poor watershed management. As a result, the release of a significant amount of PTEs into the tributary could have been expected. At ten other sampling sites, such as S1, S2, S3, S5, S6, S9, S11, S13, S14 and S17, the MQI values were between 1 to 2, which are nearly at the critical limits.

Similarly, during the wet season, the highest value was found at the S4 sampling site on the Gelda River side of the lake, followed by, S7, S8, S13, S3, and S5, with values of 11.386, 10.08, 7.14, 6.52, 5.689, and 5.486, respectively. This might be also due to the pressure from the watershed and intensive farming in the nearby floodplain area contributing inputs to the lake. However, a study conducted on the Yamuna surface water in India reported a higher Water quality index (WQI) value (23.5) (Pal et al. 2017) than the current study. In contrast, lower MQI values were found at the S2 sampling site, with a value of 0.993, which is under the limit and can be used for drinking purposes (Pal et al. 2017). Furthermore, there were lower MQI values between 1 and 2 at the S15, S1, and S16 sampling sites in that order.

The concentrations of Fe, Pb, Cd, Cr, As, and Hg surpassed the national Ethiopian water quality standards for drinking water (EDWQS 2017) (Table 8). The presence of high concentrations of PTEs in aquatic ecosystems poses significant public health issues and ecological risks (Rahman & Singh 2019). Particularly the metals Pb, As, Cd, Cr, and Hg are a serious threat to aquatic ecosystems (Raj & Maiti 2020).

Additionally, the concentrations of Fe, Pb, Cd, As, Cr, and Hg were higher than the standards set by the British Aquatic Ecosystem Water Quality guidelines (British ECCS 2021). However, regarding the irrigation water standard set by FAO/IWMI, the concentrations of Pb, As, and Cr were below the permissible limits (Drechsel et al. 2023). The presence of higher concentrations in the lake might lead to multiple aquatic and public health issues. For instance, the study conducted by Tchounwou et al. (2012) reported that exposure to high concentrations of Fe in the aquatic ecosystem might lead to discoloration, influence the growth of aquatic plants by altering nutrient availability, and disturb the food web chains. A disturbed food web leads to a decrease in the existence of fish, which reduces food security activities. In another regard, Mushtaq et al. (2020) reported that elevated Fe levels in aquatic environments increase the existence of pathogenic organisms, as most pathogenic organisms need Fe for their growth.

From an aquatic risk perspective, the presence of significant amounts of PTEs might threaten the aquatic life of the lake by entering and accumulating in the organisms' bodies, such as the liver, muscles, or other organs, in the form of bioaccumulation. This can affect the movement, feeding, behavior, and reproduction of the organisms (Sharma et al. 2024)). For instance, Cd slightly exceeded the international standards and might concerned for aquatic health effects. Because it is a highly toxic element, exposure to Cd leads to bioaccumulation in macro-organisms, including fish, and causes physiological biochemical disorders in fish. The study done by Al Mazed et al. (2022) showed exposure to Cd reduced fertility, damage to the liver, and kidney, bone fragility, and cancer risk in the lung and prostate.

From a public health perspective, humans can be exposed to contaminants by drinking, contacting, and eating fish from the lake, which could have public health concerns. For instance, the current study found that Cr is above the permissible limits of most international standards, which is exposure to hexavalent Cr might cause allergic reactions or potent carcinogenic effects, such as lung cancer and respiratory problems (Putra et al. 2024; Reif & Murray 2024). Similarly, exposure to Pb may result in developmental delays, cognitive deficits, and behavioral changes (Gudadhe et al. 2024; Zou et al. 2024). Exposure to Hg is particularly concerning as it can cause neurological cardiovascular, and developmental disorders, especially in children and young people (Pant et al. 2024; Zafar et al. 2024). Arsenic was also found 0.027 mg/L exceeding the drinking water standard of WHO (Fortin 2023). This poses a carcinogenic risk to the skin, lungs, and bladder (Al Mazed et al. 2022). Moreover, in accordance with the current study, Hu et al. (2024) and Sharma et al. (2024) stated that the PTEs cause disruption of the food web cycle and result in public health concerns.

Thus, contaminations of aquatic ecosystems with PTEs lead to serious public and ecological health implications. Regular monitoring and management plans for regulating the concentrations are essential to mitigate the potential impacts and effects of PTEs.

This study examines the spatial and temporal distributions of PTEs in Lake Tana and its major tributary rivers. It has analyzed the presence of 16 elements in water samples, assessed pollution levels using various indices and guidelines from different countries, and identified potential sources through multivariate analysis tools. The results indicated that concentration patterns during both the dry and wet seasons were similar, with the elements ranked from highest to lowest as follows: Fe > Ca > Na > Mg > K > Mn > Cr > Zn > Co > B > Ni > Cu > Pb > As > Cd > Hg. The concentrations of seven PTEs, including Cu, Zn, Hg, Ca, Co, Mn, and Fe, are relatively high during the wet season. In contrast, the other seven PTEs, such as Mg, K, Na, Ni, Cd, Cr, and B are more concentrated in the dry season. Lead (Pb) and arsenic (As) exhibited irregular trends across various sampling sites and seasons. This indicates significant spatial and temporal variation in most PTEs across the study sites. The study found that the Lake Tana ecosystem and its main tributary rivers contain substantial amounts of PTEs within the subbasin, posing risks to public and aquatic health. The multivariate analysis revealed that the primary sources of Na, Mg, K, Ca, Cd, and B are the watersheds of the Megech and Dirma rivers in the Armachiho Mountain regime, including contributions from Gonder town. Whereas, As, Cu, Pb, Hg, Co, Fe, Mn, Ni, Cr, and Zn primarily originate from the watershed of the Gelda, Gumera, and Rib rivers, along with the peripheries of Adis Zemen and Deibretabore towns. The presence of PTEs in the lake may be attributed to intensive agricultural practices, such as the use of fertilizers and pesticides, coupled with poor watershed management, geological characteristics of the lake and its catchment, and other human activities. Concentrations of WHO 2011 Fe, Mn, Pb, As, and Cd exceeded the permissible limits set by the national Ethiopian drinking water standards and the drinking water guidelines. The HPI indicated that Pb, As, and Cd are major concerns in the study area. Similarly, the MQI identified the major elements with index values above 2.0, such as Fe, Pb, Cr, Cd, and As. These findings suggest that the lake water is not chemically potable unless treated. Consequently, these factors pose potential risks to aquatic life and public health in the subbasins that rely on water from the lake and its tributary rivers. Further investigations into the major PTEs in the biometrics and sediments of the lake and Tributary Rivers are recommended. Additionally, the EPA and the Ministry of Water and Energy should regularly monitor the quality of the lake and its main tributary rivers.

The authors express their gratitude to the African Center of Water Management at Addis Ababa University for providing budgetary support for the sampling activities. We also extend our appreciation to the Ministry of Water and Energy and the HORTICOOP ETHIOPIA PLC for their laboratory services and willingness to support any inquiries. We are thankful to our field assistants and hydro geologist expert Mr Dawit for their invaluable help.

The African Center of Excellence for Water Management at Addis Ababa University generously funded this research.

Balew Yibel conceived and designed the study, collected and acquired the data, analyzed and interpreted the data, and drafted the manuscript. Authors Andualem Mekonnon and Ermias Deribe contributed to the design of the study, supervised the assessment, and provided feedback on the manuscript. All the authors read and approved the final manuscript.

All relevant data are available from an online repository or repositories.

The authors declare there is no conflict.