Effect of Physico-Chemical attributes on the abundance and spatial distribution of fish species in Lake Tanganyika, Burundian coast

The water of Lake Tanganyika is subject to changes in physical and chemical characteristics and resulting in the deterioration of water quality to a great pace. The current study was carried out to assess the physical and chemical characteristics of water at 4sampling stations of Lake Tanganyika and intended, firstly to make an inventory and a taxonomic characterization of all fish species found in the study sites, secondly to determine the pollution status of the selected sites and the impact of physico-chemical parameters on the abundance and spatial distribution of fish species in the Lake. The results obtained regarding the taxonomy and abundance of fish species showed that a total of 75 fish

The water of Lake Tanganyika is subject to changes in physical and chemical characteristics and resulting in the deterioration of water quality to a great pace. The current study was carried out to assess the physical and chemical characteristics of water at 4sampling stations of Lake Tanganyika and intended, firstly to make an inventory and a taxonomic characterization of all fish species found in the study sites, secondly to determine the pollution status of the selected sites and the impact of physico-chemical parameters on the abundance and spatial distribution of fish species in the Lake. The results obtained regarding the taxonomy and abundance of fish species showed that a total of 75 fish species belonging to 12 families and 7orders existed in the 4 selected sampling stations. Besides, it has been proved that Kajaga and Nyamugari stations were heavily polluted while Rumonge and Mvugo Stations were moderately polluted. Karl Pearson's correlation coefficient calculated between physico-chemical parameters values and the number of fish species showed strong positive and negative correlation which revealed that physico-chemical parameters have respectively a high influence on the increase and on the decrease of fish species amount in the study environment and at the same time, one way ANOVA and Tukey's Honestly Significant Difference test (Tukey's HSD) showed that the influence of the sampling stations on the abundance of fish species is highly significant (P= 0.007). Int. J. Adv. Res. 7 (12), 410-424 411 undesirable effects such as a decrease in biodiversity and an increase in tolerant species to pollution is observed (Roxane and Reinhard, 2015).
Nevertheless, freshwater fish are one of the most threatened taxonomic groups (Darwall and Vie, 2005) because of their high sensitivity to the quantitative and qualitative alteration of aquatic habits (Laffaille et al., 2005;Kang et al., 2009;Sarkar et al., 2008). Due to the habitat destruction and defragmentation (Cuizhang et al.,2003), water abstraction, industries and private use (Szollosi-Nagy,2004; Ricciardi and Rasmussen,1999;Gibbs, 2000;Dawson et al.,2003), increasing surface water pollution (Lima-Junior et al.,2006), global climate change impacts (Leveque et al.,2005;Mas-Marti et al.,2010) and overfishing, it has been realized that many fish species have become highly endangered, fish richness and diversity are rapidly declining and many species have become extinct (Pompeu and Alves,2003;Pompeu and Alves,2005; Shukla and Singh,2013; Mohite and Samant,2013;Joshi,2014). As a consequence, the Freshwater fish are often used as bioindicators for the assessment of the water quality, river network connectivity or flow regime (Chovance et al., 2003). The water quality assessment is therefore essential to identify magnitude and source of any pollution load and provides significant information about the available resources for supporting life in a considered ecosystem. Any change in the water quality has direct influence on biotic communities where different species of flora and fauna exhibit great variations in their responses to the environment (Watson and John, 2003).
The present study was conducted on Lake Tanganyika which is one of the largest lakes of Africa and second biggest lake considering the area after Lake Victoria .At the world level, Lake Tanganyika is the longest freshwater lake and holds second position in terms of volume and depth after Lake Baïkal (Wetzel, 1983). It is surrounded by four countries (Burundi, Democratic Republic of Congo, Tanzania and Zambia) sharing its entire perimeter unequally. However, the surface waters of Lake Tanganyika are highly polluted by different harmful contaminants from human activities in large cities and are subject to frequent fluctuations in their physico-chemical characteristics and to desiccation (Wetzel, 2001) due to sudden changes in weather Conditions. That is why the current study was undertaken to assess the present status of fish community structure and the impact of physico-chemical characteristics of water on the abundance, diversity and spatial distribution of fish species in four sampling stations.

Materials and Methods:-Study area:-
As the lake has a long perimeter (1838km) shared between four countries (Burundi, Tanzania, Democratic Republic of Congo and Zambia), the data collection on fish species caught in the lake, and water sample for laboratory analyses was carried out at 4 study sites (Kajaga, Nyamugari, Rumonge and Mvugo) belonging to the Burundian Littoral and the distance separating the selected study sites was at least 20km. The figure1 and table1 below show the geographical location of the study areas:  Sampling and field data collection:-Fish species sampling and taxonomic identification:-The sampling of fish species was carried out twice per month during three months (January, February and March both for 2017 and 2018) .Fish samples were collected from various sampling sites with the help of local fishermen using different types of nets namely gill nets, cast nets and drag nets and much other valuable information were obtained by physical verification and interview with the people residing nearby the selected stations ( Figure 2).  The comparative study of the spatial variations of the diversity of fish population for the studied stations was carried out using two commonly used indices: Jaccard (1908) and Sorensen (1948) coefficients which show the similarity or dissimilarity between fish species recorded in the sampling stations on the basis of the presence-absence of species. Jaccard's Index: S j This index can be modified to a coefficient of dissimilarity by taking its inverse: Jaccard's dissimilarity coefficient = 1-S j Sorensen's Index: S s 413 This measure is very similar to Jaccard's measure and can also be modified to a coefficient of dissimilarity by taking its inverse: Sorensen's dissimilarity coefficient = 1-S s , Where: S j = Jaccard's similarity coefficient S S = Sorensen's similarity coefficient C = Number of species common or shared between two sampling station A = Number of species present only in the first sampling station B = Number of species present only in the second sampling station Water samples for physico-chemical analyses:-During the present investigation, field data collection has lasted 6months, at 3 months per year (January, February and March both for 2017 and 2018) and the various outings were always conducted in the morning time. The water samples for Physical and chemical analyses were collected from different Study sites with plastic containers in the morning time. The containers were thoroughly washed and sterilized to avoid extraneous contamination. All samples were adequately labeled and transported immediately to the laboratory for analyzing of different parameters. Some physical and chemical parameters such as Temperature, Electrical Conductivity, pH and Dissolved Oxygen have been measured in-situ (at the time and on the sampling sites of sample collection) using Electrometric method (conductivity meter and pH-meter) while the remaining parameters were determined in the Laboratory using the standard methods (APHA, 2005; Trivedy and Goel, 1986). The methods adopted for water quality analysis and the used instruments are listed in the table2 below: Table 2:-Analytical methods adopted to determine quality of lake water.

Data analyses:-
All the Statistical analyses were carried out using the software SPSS.20.0 and those analyses include: A descriptive analysis to describe the minimum, maximum, average and standard deviations corresponding to the Physicochemical parameters values; Pearson's correlation analysis to assess pair wise associations between variables (Physico-chemical parameters values and Fish species amount) and the strength of their relations; One-way analysis of variance (ANOVA-1) to show the effect of physico-chemical parameters on the variation of fish species number in sampling stations; Tukey's Honestly Significant Difference test (Tukey's HSD) which is the one way ANOVA post hoc non parametric test to make the comparison among the means of fish species number from different samplings stations.

Physico-chemical characteristics:-
Monitoring physico-chemical parameters is very important for studying the influence of these parameters on the distribution of various components of biodiversity in headwater stream (Sharma et al., 2007). Water quality is influenced by geological, hydrological, climatic and anthropogenic factors (Boon et al., 1992; Bartram and Balance, 1996). In the present study, the physical and chemical parameters evaluated were Turbidity (Tur), Temperature (   Regarding the fish species distribution per orders (figure 3B), it has been realized that the order Perciformes is the most dominant with 51species (68%), followed by Siluriformes order with 13 species (17%), then Clupeiformes order with 4species (6%). Characiformes order with 3species (4%) and Cyprinodontiformes order with 2species (3%) occupy respectively the fourth and the fifth positions while Synbranchiformes and Cypriniformes order hold last position with one specie (1%) each.    Effect of change in physico-chemical attributes of water on the abundance of fish species:-For checking the link established between the water quality and the abundance of fish species, Pearson's correlation analysis was performed. The results (Table 5) showed that the amount of fish species is negatively correlated to thirteen parameters and positively correlated to six parameters; with strong and weak relation. From the Table 5, it has been found that the increasing of fish species amount in the sampling stations is:

Parameters
Significantly Significantly and strongly related to the increase in value of Iron (r=0.908, p<0.05), strongly related to the increase in value of Temperature(r=0.823), weakly related to the increase in value of Transparency(r=0.45), Chlorides (r=0.185) and Calcium hardness (r=0.101).
Lastly, a very weak positive relationship is established between the fish species amount and Total Alkalinity (r=0.011), which shows that Total Alkalinity has almost no influence on the abundance of fish species in the sampling stations. Fish diversity in relation to pollution:-Pollution status of the water at sampling stations:-Water pollution occurs when untreated waste is thrown into water bodies. Polluted water can lead to destruction of plants and organisms living in the aquatic ecosystem and can also be harmful to peoples, plants and animals that use it. The assessment of the pollution status of the sampling stations water was focused on the analysis of the major conventional pollutant (Biochemical Oxygen Demand and Chemical Oxygen Demand) which are directly related to organic pollution. BOD is similar in function to chemical oxygen demand (COD, they both measure the amount of organic compounds in water. However, COD is less specific because it measures everything chemically oxidizable, rather than the levels of biodegradable organic matter. COD is helpful in terms of water quality by measuring the effect of an effluent on the receiving body, much like (BOD). COD range in unpolluted surface water is less than or equal to 20 mg.L -1 (Chapman, 1997). In the present study, the COD value ranged from 20-67.5mg.L -1 and the general mean was 34.25±20.77mg.L -1 (table  3). Kajaga station appeared to be polluted by both sewage and industrial wastes as it shows high COD value with average of 67.5mg.L -1 . Nyamugari, Rumonge and Mvugo stations show respective mean values of 28mg.L -1 , 21.5mg.L -1 and 20mg.L -1 (table 3). Since COD in unpolluted surface water is ≤ 20 mg.L -1 (Chapman, 1997), all stations appeared to be polluted and the pollution stage is reflected by the BOD value. The BOD content of water ranged from 6.25 to 14 mg.L -1 with a general mean of 9.51±3.18 mg.L -1 (table 3). Kajaga and Nyamugari stations appeared to be polluted by sewage and industrial wastes as they have high BOD Concentration with respective averages of 14 and 10.3mg.L -1 . Rumonge and Mvugo stations show low mean value of 7.5 and 6.25mg.L -1 , respectively (table 3). According to Adakole (2000), Connor and Richard (2016), The present study revealed that waters of Mvugo and Rumonge stations were moderately polluted ,while waters at Kajaga and Nyamugari sites were heavily polluted during the investigation period.
Effect of pollutants on diversity and distribution of fish species and identification of pollution indicator fish:-75 fish species have been identified in all sampling stations (see Table 4) and three categories of fish species have been distinguished (Table 7), depending on their adaptation level to pollution (Table 6):

Sensitive Species to pollution or Polluosensitive species:
Species living exclusively at Mvugo and Rumonge station which are moderately polluted. In this category, 21 (or 28%) species have been recorded and the presence of these species can be used as indicators of slightly polluted environment.

Resistant Species to pollution or Polluoresistant species:
Species exclusively inhabiting at Kajaga and Nyamugari stations which are heavily polluted. In this category, 15 (or 20%) species have been identified and the presence of these species can serve as indicators of highly polluted environment.

Tolerant Species to pollution or Polluotolerant species:
Species adapted for living in all sampling stations, both heavily and moderately polluted. In this category, 39 (or 52%) species have been identified.
The table 6 summarizes the pollution status of the sampling stations and the acclimation level to pollution of the fish species inhabiting the respective stations while the table 7 shows the identification and distribution of fish species based on their acclimation level to pollution.

Similarity between fish species richness of sampling stations:-
The similarity between fish species recorded in the sampling stations was determined using similarity indices. The most used indices are similarity coefficients of Jaccard (1908) and Sorensen (1948). These indices are used to compare objects on the basis of the presence-absence of species and are so very simple measures of beta biodiversity, ranging from 0 (when there are no common species between two communities) to 1 when the same species exist in both communities). A smaller index indicates less similarity in species composition between different habitats (Condit et al., 2002;Nshimba, 2008). The table 8 shows Similarity Index between the fish species composition of sampling stations, which is calculated using Jaccard and Sorensen's Method.  Effect of sampling sites on the abundance of fish species:-Tukey's Honestly Significant Difference test (Tukey's HSD) and One-way Analysis of Variance (ANOVA-1) both at the 5% level were performed respectively to make the averages comparison and to assess the effect of the sampling sites on the abundance of fish species. The results of one-way Analysis of variance (ANOVA-I) ( Table 9) indicated that the influence of the study stations on the abundance of fish species is highly significant (p= 0.007). It means that the variation of fish species in number depends on the environmental conditions. The differences among pairwise averages number of fish species from the sampling stations are shown by Tukey's HSD multiple comparison test in the table 10 and it has been reflected that the mean difference of fish species amount between stations is significant (p<0.05) for Kajaga and Rumonge sites (p=0.036), Nyamugari and Rumonge sites (p=0.006), Nyamugari and Mvugo sites (p=0.016). The comparison of the average number of fish species using Tukey's HSD at the 5% level classifies the 4sampling stations into 3homogeneous subsets of averages A, B and C (Table 11). Indeed, the averages belonging to the same homogeneous subset are not significantly different (e.g: Nyamugari and Kajaga or Kajaga and Mvugo or Rumonge and Mvugo stations) whereas the averages belonging to different homogeneous subsets are significantly different because the subsets A, B and C are different.

Conflict Of Interest:-
The authors declare that they have no conflict of interest to report with respect to the research, authorship and publication of this article. field data collection.
Lastly, the local people residing nearby Lake Tanganyika, especially the fishermen are also thankfully acknowledged for their co-operation and help during the time of field data collection.

Conclusion:-
The specific objective of the present investigation is an attempt to assess the current pollution aspect of the sampling sites and the spatial variations of physico-chemical parameters (pollutants) and their influences on the abundance and spatial distribution of fish species.
The results obtained regarding the taxonomy and abundance of fish species showed a total of 75 fish species belonging to 12 families and 7 orders from all selected sampling sites and the relative diversity index of families has indicated that Rumonge site holds first position with an average of 46 species distributed into 9 families, Mvugo site in the second position with 42 species distributed into 11 families, Kajaga site in third position with an average of 35 species distributed into 11 families and Nyamugari site appeared as the poorest with an average of 28 species distributed into 6 families. Besides, the similarity index between sampling stations proved that Rumonge and Mvugo pair have a high similarity coefficient (Sorensen index=0.67) which indicated that many fish species are common or shared between Mvugo and Rumonge stations and the environmental conditions prevailing in these two stations are almost the same.
Regarding the pollution status of the sampling locations and the impact of water quality on the abundance and fish diversity, it has been proved that Kajaga and Nyamugari stations were heavily polluted while Rumonge and Mvugo Stations were moderately polluted. In this regards, three categories of fish species have been distinguished, based on their adaptation level to pollution like (i) polluosensitive species, (ii) polluoresistant species and (iii) polluotolerant species. On the other hand, Karl Pearson's correlation coefficient calculated between physico-chemical parameters values and the number of fish species showed strong positive correlation (for Temperature and Iron) and negative correlation (for Turbidity, Potential of Hydrogen , Electrical Conductivity, Total Dissolved Solids, Total carbon, Total Phosphorus, % of Oxygen saturation, Dissolved Oxygen and Biochemical Oxygen Demand) which revealed that physico-chemical parameters have a high influence on the increase and the decrease of fish species amount in the study environment and at the same time, one-way Analysis of variance (ANOVA-I) and Tukey's Honestly Significant Difference test (Tukey's HSD) have showed that the influence of the study stations on the abundance of fish species is highly significant (P= 0.007).