Physicochemical Characterization of Dembi Reservoir Water for Suitability of Fish Production, Southwest Ethiopia

Reservoir water physicochemical characteristics provide important information about water suitability for fish production. Accordingly, the study aimed to characterize the physicochemical characteristics of Dembi reservoir water for sustainable fish production. The study was conducted in Dembi reservoir during the dry season. Water samples were collected in triplicate from selected 10 sampling sites of the reservoir water using manually prepared water sampler made from polyvinyl chloride (PVC) tube. The depth integrated sampling technique was employed to take water samples for all physicochemical characteristics analysis. From the selected 14 physicochemical characteristics, four (temperature, electrical conductivity, pH, and dissolved oxygen) were tested onsite using a multisystem HQ4d electronic meter (probe), whereas the rest 10 water quality characteristics were tested in the laboratory. The result showed that the current average depth of the dam was 5.6 ± 1.61 m. The overall mean values of the water quality characteristics at different sites of the reservoir were as follows: turbidity (26.4 ± 0.44 FTU), total hardness (22.2 ± 0.51 mgL−1), NO3 (5.4 ± 0.48 mgL−1), NO2 (0.3 ± 0.11 mgL−1), NH4 (2.1 ± 0.06 mgL−1), PO4−3 (1.7 ± 0.27 mgL−1), total alkalinity (52.5 ± 0.91 mgL−1), and BOD5 (2.7 ± 0.24 mgL−1). There was a significant difference (p < 0.05) in all physicochemical characteristics among 10 sampling sites of the reservoir water. The recorded values of all physicochemical characteristics, except NO2, NH4, and PO4−3, were found within the recommended standard limit for fish production. The change in reservoir water depth and increase in nutrients shows the presence of sediment siltation and nutrient enrichment. Therefore, proper watershed management practices and waste management should be carried out for sustainable water quality maintenance and fish production.


Introduction
Fish is one of the major nutritional and commercially important foods in the world. It is a source of nutrition for 4.3 billion people worldwide with 15-20% of their intake of animal protein and in some countries over 50%, and the demand for fsh and fshery products is predicted to remain increase because of population growth, economic development, and changes in food habits [1]. However, sustainable fsh production is highly dependent on water quality and availability [2].
Water is an essential life-supporting medium for fsh and other aquatic organisms. It essentially provides all fsh needs, such as food, oxygen, and other helpful environments for fry fsh of Nile tilapia (Oreochromis niloticus Linnaeus, 1758) and Redbelly tilapia (Coptodon Zilli (Gervais, 1848)) has been introduced in the dam since 1995, aimed to supply fsh for the surrounding community [5].
Dembi reservoir is estimated to have an area of 72 km 2 and a fshery potential of about 383 tons/year [6]. However, the current production of fsh from the reservoir is much lower, and even the fsh are very small in size. Despite, the southwestern highlands of Ethiopia are the source of water for major rivers and reservoirs, yet, human activities such as agriculture expansion, population increment, and urbanization have led to considerable environmental alteration [7,8]. For example, the Bench Maji zone natural forest including the upper part of the Dembi reservoir forest reduced from 20% in 1986 to 12% by 2001 [9]. Following agriculture and settlement expansion, surface water decreases in stream base fow and dissolved oxygen (DO) concentration [10,11]; and increases in nutrient and sediment concentration, periphyton production, and water temperature [12][13][14]. Furthermore, several wet cofee processing stations have been installed and operational on the upper part of the reservoir. In wet cofee processing stations, cofee by-products (cofee pulp and efuents) are disposed of on the river, thereby, causing water pollution such as high TSS, total ammonia, total nitrogen, COD, BOD 5 , and pH [15]. Similarly, pollutants from urban nonpoint sources add a signifcant amount of phosphorus and nitrogen to surface water [16].
Knowledge of seasonal reservoir water quality is important for sustainable fsh production and management decisions [17]. According to Makinde et al. [18] and Tsegay and Zebib [13], seasonal diference plays a signifcant role in the reservoir water quality characteristics. In this regard, Gebremichael and Fentahun [19] evaluated selected reservoir water quality properties (temperature, pH, turbidity, nitrate, calcium hardness, conductivity, and total hardness) of the Dembi reservoir during the wet season of 2017. However, no study has been undertaken on reservoir water quality during the dry season. In addition, variability in the spatial distribution of nutrients and sediment determines the reservoir water quality characteristics [20].
Terefore, this study is important to increase our insight on dry season reservoir water quality and the spatial extent of reservoir water quality characteristics. In addition, it provides baseline information for further monitoring and tracking changes in the water quality and maintaining the quality of the reservoir water for sustainable fsh production. Tus, the study aimed to determine the physicochemical characteristics of Dembi reservoir water for suitability of fsh production during the dry season, with the following specifc objectives: (i) To determine the level of selected physicochemical characteristics of Dembi reservoir water (ii) To determine the spatial water quality variation inside the reservoir (iii) To compare the current water quality of the reservoir with the recommended water quality characteristics for productive fsheries

Description of the Study Area.
Te study was conducted in Dembi reservoir, the upper reaches of the White Nile basin. Te reservoir is located at about 14 km from Mizan Teferi, the capital of Bench Sheko Zone, and 599 km from Addis Ababa, the capital of Ethiopia. Dembi reservoir is located at 6°56'-7°0'N latitude and 35°30'-35°36'E longitude. Te reservoir is agroecologically located at the lowland elevation, 1440 m a.s.l ( Figure 1). Te average yearly rainfall in Aman near Mizan Teferi, the main town of Bench Maji Zone, is 1603 (±404) mm y −1 [21]. Te average air temperature ranges from 13 to 27°C.

Water Sampling Techniques.
Te water samples were taken during the main dry season (February) in 2019. A preliminary feld visit was made using a topographic map and Garmin GPS (model 76CSX ) to fully understand the land features around the reservoir, water fow pattern and shape of reservoir for locating the study area's representative water sampling points. Te depth of the reservoir water was measured by a water pipe having a 15 m height with a gauging number marked outside. A total of sixty (60) sampling points were measured to identify the depth contour of the reservoir. A total of ten stations were selected for water quality assessment at the Dembi reservoir ( Figure 1). Te depth integrated sampling technique was employed to take water samples for all physicochemical analyses using 600 ml plastic bottles. Te bottles were thoroughly washed with distilled water and rinsed repeatedly with water to be sampled. Triplicate water samples from each sampling site were taken in triangle form. Te water samples were collected by a manually prepared water sampler from a polyvinylchloride tube (PVC), having 10 m in height and 0.178 m in diameter. Height gauging marks were written on the PVC, and the water sample was collected using gauged PVC tube from the sampling points at 1 m intervals. Each 1 m PVC tube height has a separated internal tube compartment and an inlet hole at the top for the water entrance.
All bottled water samples were capped immediately, stored in an icebox, and transported immediately to Sebeta National Fisheries and Other Aquatic Life Research Center and Haramaya University' Chemistry Laboratory. To avoid decomposition, water samples were immediately fltered in the laboratory using a water jet vacuum pump at low pressure before nutrient analysis.

Physicochemical Analysis of Water
2.3.1. In Situ Analysis. Reservoir water' physicochemical characteristics such as pH, temperature, conductivity, and dissolved oxygen (DO) were tested onsite by using a multisystem HQ4d electronic meter (probe).

Ex Situ (Laboratory) Analysis.
Te laboratory analysis of physicochemical characteristics of water samples was performed at Sebeta National Fisheries and Other Aquatic Life Research Center and Haramaya University, following the standard methods for the examination of water and wastewater [22,23]. Te reservoir water turbidity was determined by using a digital turbidity meter (formazin turbidimeter), total hardness by the titration method, NO 3 -N by the sodium-salicylate method, NO 2 -N by the colorimetric method, NH 4 -N by the indophenol blue method, total phosphorus (P) by frst digesting the unfltered sample using potassium-peroxodisulphate oxidation, PO −3 4 − P by the ammonium molybdate method, total alkalinity by the titration method, calcium (Ca 2+ ) by fame atomic absorption spectrometry (AAS), and biological oxygen demand (BOD) by Azide modifcation of Winkler's titrimetric method by determining dissolved oxygen contents of the samples before (D1) and after fve days (D2) of incubation at 20°C.

Data Analysis.
Te data obtained from the study were managed in Microsoft Excel. Te sources of samples were categorized as inlet, open, and outlet sites to enable analysis. Te diference in water physicochemical characteristics between the reservoir sites was tested by one-way ANOVA using SPSS (software version 22), and the mean diference was compared by the least signifcant diference (LSD) at a 5% level of signifcance. Moreover, an overall mean value of the physicochemical characteristics of reservoir water was compared with standard values for productive fsheries.

Physical Properties of Dembi Reservoir Water
3.1.1. Depth. Te high depth of the reservoir was recorded at P6, P5, P8, P7, and P4, with an overall mean depth of 5.6 ± 1.61 m (Table 1). Tere was a signifcant diference in the reservoir water depth between the 10 sampling stations (p < 0.01). It was 17 m during the reservoir construction time since (1991) [24] and has shown a signifcant reduction from 17 m to 5.6 m in between 28 years. Tis may be due to the reservoir sediment siltation because of soil erosion from the upper catchment agriculture, grazing, and settlement lands. Similarly, Vacher and Quinn [25] reported that human disturbance on the upper catchment can cause a dramatic change in water depth level.

Turbidity.
High reservior water turbidity was recorded at P4, P1, and P3, with an overall mean value of 26.38 ± 0.44 FTU (Table 1). Tere was a signifcant diference (p < 0.05) in water turbidity level between the diferent sampling stations of the reservoir water. Since those sites (P4, P1, and P3) were located on an inlet site of the reservoir, the presence of higher turbidity may be due to the presence of river-transported and accumulated suspended sediments and low dilution efect. Similarly, Abate et al. [26] reported higher turbidity in a site near and around the inlet of Lake Hawassa.
3.1.3. Temperature. High reservoir water temperature was recorded at P4, P1, P2, P3, and P10, with an overall mean value of 26.74 ± 0.36°C (Table 1). Tere was a signifcant diference (p < 0.05) in the level of water temperature among diferent sampling stations. Te presence of high temprature on inlet and outlet sites (P4, P1, P2, P3, and P10) may be associated with relatively higher turbidity and low water depth on those sites. High water turbidity increases water temperature due to the trapping of heat by turbid water and suspended sediments. Similarly, Tilahun and Ayale [30] observed higher water temperatures in the shallow depth of water at the Selameko reservoir.

Electrical Conductivity.
High reservoir water conductivity was recorded at P1, P2, P10, and P3, with an overall mean value of 51.03 ± 1.14 μScm −1 (Table 1). Tere was a signifcant diference (p < 0.05) in water conductivity value among the reservoir water sampling stations. Te higher conductivity value recorded at P1, P2, P3, and P10 may be due to an increase in the concentration of salts and ions. Similarly, Woldeab et al. [29] reported a higher level of EC in a site near and around the inlet of the Gilgel Gibe reservoir.

Chemical Properties of Dembi Reservoir Water
3.2.1. pH. Te reservoir water high pH value was recorded at P8, P5, P7, P4, and P6, with an overall mean value of 7.7 ± 0.07 (Table 2). Tere was a signifcant diference (p < 0.05) in water pH value among the diferent sampling sites. Te presence of higher pH in the middle sites (P8, P5, P7, P4, and P6) may be related to the presence of low nitrate and high water depth on those sites. Higher water depth and volume improve the dilution and bufering capacity, which in turn increased the pH value on the middle sites.

Dissolved Oxygen (DO).
High dissolved oxygen in the reservoir was recorded at P5, P9, and P8, with an overall mean value of 4.6 ± 0.12 mgL −1 ( Table 2). Tere was a signifcant diference (p < 0.05) in dissolved oxygen content among the diferent sampling sites. Te higher DO content  Te values are given as mean ± SD. Biochemistry Research International was recorded at P5, P9, and P8, this may be due to lower turbidity values and temperature found on those sites, associated with relatively low biological degradation activities. Furthermore, it may be due to east-west moving cool windwave action, which might contribute to the increase in oxygen on middle sites. Similarly, Adeniji [35] and Ibe [36] reported that water temperature, water depth, wind action, and amount of biological degradation activities determine the level of dissolved oxygen content. Te overall mean value of dissolved oxygen was found within the recommended limit for productive fsheries (4-5 mgL −1 ) by Rao et al. [37] and (3-5 mgL −1 ) Bhatnagar and Devi [4]. However, it is lower than the DO level in Lake Hawassa (17.9 mgL −1 ) by Abate et al. [26]. In general, the presence of low DO in Dembi reservoir water may be related to the microbial decomposition of organic wastes from agriculture, municipal solid wastes, cofee by-products (cofee pulp and efuent), and dead aquatic vegetation. Similarly, Srivastava et al. [38] reported that decomposing of organic matter, dissolved gases, mineral waste, and agricultural runof play a great role in decreasing DO content of the reservoir water.

Total Hardness (TH).
High reservoir water total hardness was recorded at P1, P10, P2, and P3, with an overall mean value of 22.16 ± 0.51 mgL −1 ( Table 2). Tere was a signifcant diference (p < 0.05) in the value of water hardness among diferent sampling stations. Higher values of water hardness recorded at P1, P10, P2, and P3 may be associated with the presence of higher calcium and magnesium content at those sites, which emanated from decomposed organic wastes from agriculture, municipal solid waste, and cofee byproducts. Similarly, APHA [23] reported that higher values for calcium are related to organic wastes (sewage) and weathering of Ca-rich rocks from the upper catchment.

Nitrate-Nitrogen.
Te reservoir water high nitrate content was recorded at P1, P3, P4, and P2, with an overall mean value of 5.44 ± 0.48 mgL −1 ( Table 2). Tere was a signifcant diference (p < 0.05) in the nitrate content among the diferent sampling sites. Te higher nitrate content recorded at the inlet sites (P1, P3, P4, and P2) may be related to the presence and decomposition of inorganic and organic material transported from agricultural land, municipal waste, and cofee processing station. Similarly, Zelalem et al. [46] reported higher levels of nitrate in the inlet site of the Selameko reservoir.

Nitrite-Nitrogen.
High reservoir water nitrite content was recorded at P1, P4, and P3, with an overall mean value of 0.26 ± 0.11 mgL −1 ( Table 2). Tere was a signifcant diference (p < 0.05) in the nitrite content of the reservoir water among the diferent sampling sites. Te higher nitrite content recorded on those sites (P1, P4, and P3) may be due to the decomposition of organic materials transported from agricultural lands, settlements, and cofee processing stations. Similarly, Abate et al. [26] reported a higher level of nitrite in the inlet side of Lake Hawassa.

Ammonia (NH 4 -N).
Te reservoir water high ammonia content was recorded at P1, P4, P3, and P2, with an overall mean value of 2.13 ± 0.06 mgL −1 ( Table 2). Tere was a signifcant diference (p < 0.05) in the ammonia content of water among diferent sampling sites. Since those sites (P1, P4, P3, and P2) were found on an inlet site of the reservoir, the presence of higher content may be associated with the presence of decomposed organic material which is transported from cultivated, grazing, and municipal solid waste dumping sites. Similarly, Osman and Kloas [42] reported a higher level of ammonia on the inlet side of the reservoir water.
Te overall mean value of ammonia was found to be higher than the recommended limit for productive fsheries by (<0.2 mgL −1 ) Santhosh and Singh [34], Bhatnagar and Singh [43] and Bhatnagar and Devi [4]. Furthermore, it was found higher than the ammonia level in Geray reservoir (0.06 mgL −1 ) by Mohammed et al. [18].

Total Phosphorus (P).
High reservoir water phosphorus content was recorded at P1, P4, P2, P3, P10, and P7, with an overall mean value of 2.97 ± 0.16 mgL −1 ( Table 2). Tere was a signifcant diference (p < 0.05) in total phosphorus content of reservoir water among the diferent sampling sites. Higher disposal of phosphate from domestic sewages and surface runof from phosphate-containing fertilizers can lead to higher values of orthophosphate and total phosphorous in reservoir water [44]. Hence, the presence of higher total phosphorous content at P1, P4, P2, P3, P10, and P7 may be due to inorganic and organic nutrient accumulation and decomposition in the inlet sites mainly delivered from upper catchment agriculture, settlement, and wet cofee processing station.
Te overall mean value of total phosphorus was found within the recommended limit for productive fsheries (0.01-3 mgL −1 ) by Bhatnagar and Devi [4]. Furthermore, it was found higher than the total phosphorus content in Lake Hawassa and Chamo (0.03 mgL −1 and 0.18 mgL −1 , respectively) by Girma and Ahlgren [45].

Soluble Reactive Phosphorus.
Te reservoir water high phosphate content was recorded at P1, P3, P4, and P2, with an overall mean value of 1.67 ± 0.27 mgL −1 ( Table 2). Te phosphate content of the reservoir water showed a signifcant diference (p < 0.05) among the diferent sampling sites. Te presence of higher phosphate content on those sites (P1, P3, P4, and P2) may be due to higher disposal of phosphate source materials from domestic sewages and runof from agriculture land, municipal solid waste, cofee processing stations, and settlement land. Similarly, Abate et al. [26] and Mohammed et al. [18] reported a higher level of phosphate at the inlet site of reservoir water.
Te overall mean value of phosphate at the reservoir was found beyond the recommended limit for productive fsheries by Stone and Tomforde [32] 0.06 mgL −1 . Tis implies that the reservoir water was highly polluted by phosphate. Furthermore, it was found higher than the phosphate content in Lake Hawassa (1.12 mgL −1 ) by Abate et al. [26] and lower than in Gomit reservoir (1.96 ± 2.54 mgL −1 ) by Zelalem et al. [41].
3.2.9. Total Alkalinity. High reservoir water total alkalinity content was recorded at P1, P2, and P3, with an overall mean value of 52.54 ± 0.91 mgL −1 ( Table 2). Tere was a signifcant diference (p < 0.05) in the alkalinity content of reservoir water among diferent sampling sites. Te higher total alkalinity value recorded at P1, P2, and P3 may be associated with the presence of weathered rocks, waste discharge, and microbial decomposition of organic matter on those sites. Similarly, Mohammed et al. [18] reported higher content of alkalinity on an inlet side of the Geray reservoir.
Te overall mean value of total alkalinity at the reservoir was found within the recommended limit for productive fsheries (25-100 mgL −1 ) by Bhatnagar and Devi [4]. Furthermore, it was found lower than the total alkalinity value in Gomit reservoir (91.7 ± 42.2 mgL −1 ) and higher than in the Selameko reservoir (44.5 ± 4.17 mgL −1 ) by Zelalem et al. [41].

Calcium.
High reservoir water calcium content was recorded at P4, P1, P2, and P10, with an overall mean value of 4.24 ± 0.11 mgL −1 ( Table 2). Tere was a signifcant difference (p < 0.05) in calcium content at the reservoir water among the diferent sampling sites. Te higher calcium content recorded at P4, P1, P2, and P3 may be due to the presence of sewage from the settlement, municipal solid waste, efuents from the cofee processing station, and weathering of calcium-rich materials from the upper catchment.

Biological Oxygen Demand (BOD).
Te reservoir water high BOD content was recorded at P1, P4, P3, and P2, with an overall mean value of 2.68 ± 0.24 mgL −1 ( Table 2). Tere was a signifcant diference (p < 0.05) in the level of BOD among diferent sampling sites. Te presence of higher BOD content at P1, P4, P3, and P2 may be associated with the presence of higher nitrate and phosphate content on those sites. Similarly, Abate et al. [26] reported a higher level of BOD on the inlet site of Lake Hawassa.
Te overall mean value of BOD at the reservoir was found within the recommended limit for productive fsheries (<10 mgL −1 ) by Santhosh and Singh [34] and (3-6 mgL −1 ) Bhatnagar and Devi [4]. Furthermore, it was found in line with the BOD content in Gilgel Gibe reservoir (2.56 mgL −1 ) by Bizuneh et al. [29] and lower than the BOD content in Lake Hawassa (117 mgL −1 ) by Abate et al. [26].

. Conclusion and Recommendation
Water quality is the most important factor afecting fsh health, growth, and production. Te reservoir water physicochemical characteristics are spatially variable in Dembi reservoir. Most importantly, Dembi reservoir water quality was suitable for fsh production, as confrmed by most of the physicochemical characteristics of the reservoir water. However, the concentration of NO 2 -N, NH 4 -N, and PO 4 -P may cause stress and unsuitable condition for fsh production. A signifcant reduction in the reservoir water depth from 17 m to 5.6 m in between 28 years period, and the increase in nutrient level implies siltation problem and nutrient enrichment. Terefore, proper watershed management practices and waste management should be carried out for sustainable water quality maintenance and fsh production in the reservoir.

PVC:
Polyvinyl chloride BOD: Biochemical oxygen demand CO 2 : Carbon dioxide EEPCo: Ethiopian Electric Power Corporation DO: Dissolved oxygen GPS: Global positioning system APHA: American Public Health Association AAS: Atomic absorption spectrophotometry LSD: Least signifcant diference FTU: Formazin Turbidity Units FAO: Food and Agriculture Organization TH: Total hardness

Data Availability
Te data used to support this study are given in Tables 1  and 2.

Conflicts of Interest
Te authors declare that they have no conficts of interest.