University of Ghana http://ugspace.ug.edu.gh TREND EVALUATION OF RAW WATER QUALITY FROM KPONG DAM (VOLTA RIVER) USING STATISCAL ANALYSIS AND TIME SERIES PREDICTION MODEL THIS THESIS IS SUBMITTED TO THE UNIVERSITY OF GHANA, LEGON IN PARTIAL FULFILMENT OF THE REQUIREMENT FOR THE AWARD OF MPHIL IN NUCLEAR AND ENVIRONMENTAL PROTECTION BY DAVID BAKOMNAAH (10551750) JULY, 2018 i University of Ghana http://ugspace.ug.edu.gh DECLARATION This thesis is the result of research work undertaken by David Bakomnaah in the Department of Nuclear Sciences and Applications, School of Nuclear and Allied Sciences, University of Ghana, under the supervision of Professor Joseph Richmond Fianko and Mr. Abass Gibrilla. No part of this work has been presented for another degree in this university or elsewhere except for references to works of other researchers which have been duly acknowledged and cited. David Bakomnaah ………………… …………………… (student) signature date Prof. J. R. Fianko ………………… …………………… (Supervisor) signature date Mr. Abass Gibrilla ………………… …………………… (Co-supervisor) signature date ii University of Ghana http://ugspace.ug.edu.gh DEDICATION To my wife, Mama Rose for her invaluable assistance and morale support. iii University of Ghana http://ugspace.ug.edu.gh ACKNOWLEDGEMENTS The successful completion of this work owes a great deal to the contributions of many individuals who cannot go unnoticed. My heartfelt gratitude goes to Professor J. R. Fianko, my supervisor, and Mr Abass Gibrilla (co-supervisor), without whose sincere guidance, advice and constructive criticisms, its completion would not have been possible. I sincerely appreciate the invaluable tuition and direction of all lecturers of the Department of Nuclear and Environmental Protection at the University of Ghana, Legon. Finally, my thanks go to Mr. Chris, the chief production officer at the 37 Branch of the Ghana Water Company Limited for freely given me the data needed for this work. iv University of Ghana http://ugspace.ug.edu.gh TABLE OF CONTENTS DECLARATION ................................................................................................................ ii DEDICATION ................................................................................................................... iii To my wife, Mama Rose for her invaluable assistance and morale support. .................... iii ACKNOWLEDGEMENTS ............................................................................................... iv LIST OF FIGURES ............................................................................................................ x ABSTRACT ...................................................................................................................... xii CHAPTER ONE ................................................................................................................. 1 INTRODUCTION .............................................................................................................. 1 1.0 Overview .......................................................................................................................... 1 1.1 Background of the Study ................................................................................................. 1 1.2 Study area profile ............................................................................................................. 4 1.3 Statement of the problem ................................................................................................. 7 1.4Aim and Objectives of the study ....................................................................................... 7 1.4.1 MAIN OBJECTIVE: ................................................................................................ 7 1.4.2 Specific Objectives: .................................................................................................. 7 1.6 Scope and Limitation ....................................................................................................... 9 1.7 Thesis Structure: .............................................................................................................. 9 LITERATURE REVIEW ................................................................................................. 10 2.0 Introduction .................................................................................................................... 10 2.2 Surface Water Quality in Ghana- An Overview: ........................................................... 13 2.3 The Problem of Pollution ............................................................................................... 15 2.4Input Sources of Pollutants ............................................................................................. 18 v University of Ghana http://ugspace.ug.edu.gh 2.4.1 Non-Point Sources .................................................................................................. 19 2.5.2 Point Sources .......................................................................................................... 19 2.5 Studies on Water Quality from other countries ............................................................. 20 2.7 Water Quality Parameters .............................................................................................. 27 2.7.1 Physico-chemical Parameters ................................................................................. 28 2.7.1.1Temperature .......................................................................................................... 29 2.7.1.2 pH ......................................................................................................................... 30 2.7.1.3 Dissolved Oxygen (DO)....................................................................................... 31 2.7.1..4 Biological Oxygen demand and Chemical Oxygen Demand ............................... 32 2.7.1.5 Total Alkalinity .................................................................................................... 32 2.7.1.6 Ammonia.............................................................................................................. 33 2.7.1.7 Phosphorous ......................................................................................................... 34 2.7.2 Faecal Coliforms ..................................................................................................... 35 2.8 Some Applications of Time Series Analysis .................................................................. 36 METHODOLOGY ........................................................................................................... 38 3.1 Basic conceptsof statistical tools employed ............................................................. 38 3.1.1 Water quality index (WQI) concept ........................................................................ 38 3.1.2 Time Series Analysis .............................................................................................. 43 2.1.3 Mann-Kendal test For Trend (Mann, 1945, andKendall, 1975 )............................. 44 3.1.4 Autoregressive (AR) Models .................................................................................. 45 3.1.5 Moving Average (MA) Models .............................................................................. 45 3.2 Best Model Identification and Selection ........................................................................ 48 vi University of Ghana http://ugspace.ug.edu.gh 3.3 Forecast accuracy measures ........................................................................................... 49 CHAPTER FOUR ............................................................................................................. 51 RESULTS AND DISCUSSIONS ..................................................................................... 51 4.1 Introduction .................................................................................................................... 51 4.2 Physical characteristics .................................................................................................. 54 4.3 Nutrient .......................................................................................................................... 54 4. 4. Chemical characteristics .............................................................................................. 56 4.5.3 Comparing the WQI scores across years from 2011 to 2016. ..................................... 61 4.6 Trend analysis of the Water Quality Index over the years ......................................... 63 4.6.1 Time series plot of the monthly Water Quality Index of the Kpong Dam. ............. 63 4.7. Identifying the appropriate trend model of the series ................................................... 66 4.8 Predicting WQI of the Kpong Dam using ARIMA model ............................................ 68 4.8.1 Testing for White Noise in the series ...................................................................... 70 4.8.2 Identifying the Autoregressive (p) and Moving Average (q) orders of the ARIMA model ............................................................................................................................... 71 4.8.3 Residual diagnostics of fitted ARIMA (3, 1, 1) model ........................................... 74 CHAPTER FIVE .............................................................................................................. 81 CONCLUSIONS & RECOMMENDATIONS ................................................................. 81 5.0Conclusion ...................................................................................................................... 81 5.2 Recommendations .......................................................................................................... 82 REFERENCES ................................................................................................................. 83 APPENDIX ....................................................................................................................... 94 vii University of Ghana http://ugspace.ug.edu.gh LIST OF TABLES Table 3.1: Water Quality criterial for clasifying surface waters……………………………..40 Table 3.2 A sample calculation of WQI for the Kpong dam (Volta River) in Jan. 201151….40 Table 4.1 Summary of physico-chemical Parameters of the raw water from Kpong Dam. All values except Ec. and pH are in mg\l………………………………………………………...60 Table 4.2: Mean Water Quality Index values across years…………………………………..72 Table 4.3: Output of the ANOVA test…………………………………………………..…...73 Table 4.4: Water Quality Index data of Kpong Dam………………………………………...75 Table 4.5: Output of the Mann-Kendall test…………………………………………………77 Table 4.6: Prediction errors of the fitted trend models………………………………………79 Table 4.7: Mann-Kendall trend test of differenced series……………………………………83 Table 4.8: Portmanteau White Noise test……………………………………………………85 Table 4.9: Fitted ARIMA models…………………………………………………………...87 Table 4.10: Fitted ARIMA (3, 1, 1) model…………………………………………………88 viii University of Ghana http://ugspace.ug.edu.gh Table 4.11: Modified Box-Pierce or Ljung Box Test……………………………………….89 Table 4.12: Lagrange Multiplier (LM) test of homoscedasticity…………………………...90 Table 4.13: Forecasted WQI values of the Kpong Dam for the Next 2 Years………………….94 ix University of Ghana http://ugspace.ug.edu.gh LIST OF FIGURES Figure3.1: Map of the Study rea..............................................................................................6 Figure 4.1: Levels of nutients in raw water samples from Kpong dam (2011 – 2016)………65 Figure 4.2: Anionic content of raw water from Kpong dam (2011 – 2016)…………………67 Figure 4.3: Metal content of raw water samples from Kpong Dam (2011 – 2016) ………...69 Figure 4.4: Trace metal content of raw water samples from Kpong Dam (2011 – 2016)……70 Figure 4.5: Monthly variations of WQI in 2011 to 2016, depicting water quality levels……70 Figure 4.6: Yearly water quality variations of the Kpong Dam from 2011 to 2016…………71 Figure 4.7: Time Series Plot of WQI of the Kpong Dam.........................................................75 Figure 4.8: Fitted Quadratic trend model.................................................................................82 x University of Ghana http://ugspace.ug.edu.gh Figure 4.9: Combined Time Series plot of the original data and differenced series................84 Figure 4.10: PACF plot of the WQI series at lag 1..................................................................86 Figure 4.11: ACF plot of the WQI series at lag 1....................................................................87 Figure 4.12: Residual Plots of the fitted ARIMA (3, 1, 1) model...........................................91 Figure4.13 Predicting the WQI for one year..........................................................................92 xi University of Ghana http://ugspace.ug.edu.gh ABSTRACT This research evaluates raw water quality status from Kpong dam of the Volta River, Eastern part of Ghana, to assess its quality levels over a six-year period and predict future quality trend. The dataset identified twenty-six physicochemical constituents in each monthly sample analyzed. Hydrochemical analysis showed variation in contaminant levels for some constituents which ranged from low to considerable contamination (WHO, 2006). PO 2-4 levels were relatively high. Results of WQI indicated general quality degradation with minimal variations across the years with respect to surface water quality rating criteria (WC,2003). The forecasted figures with the fitted ARIMA (3, 1, 1) model also suggest minimal degradation of the water quality in the light of the current prevailing conditions. It is therefore recommended that efforts be made to prevent further pollution of the Volta River. xii University of Ghana http://ugspace.ug.edu.gh LIST OF ABBREVIATIONS: WQI Water Quality Index AR Autoregressive ARMA Autoregressive Moving Average MA Moving Average WHO World Health Organization UN United Nations GWCL Ghana Water Company Limited VRA Volta River Authority WRC Water Resource Commission xiii University of Ghana http://ugspace.ug.edu.gh CHAPTER ONE INTRODUCTION 1.0 Overview The chapter presents the background of the study, the area profile, the statement of the problem and objectives of the study. It also includes the justification of the study, the scope and limitation and the thesis organisation. 1.1 Background of the Study Water, a prime unit of life, and most valuable national asset, forms the chief element of ecosystem and man’s existence (Bartram and Ballance, 1996). The link between water and life can be seen in the fact that about 70% of the mass of the human body and about 50-97% of that of other life forms is made up of water (Buchholz, 1998). Water quality is a key feature of any water body which plays a crucial role in health, quality control studies, water planning and management. One limitation that most people do not fully realize, care about, or even know about is safety of our water bodies. Globally, many factors are threatening the sustainable use of freshwater resources, notably, rivers, lakes, ponds, ground water etc. (Raymond et al., 2008).Though nature herself is sometimes blamed of water quality variability (Miller and Tetlow, 1989), water quality of water sources tends to deteriorate gradually with human actions, such as hydrological perturbations (Booker and Woods, 2014), changes in land use (Seeboonruang, 2012), inputs of hazardous chemical species and nutrients (Gevrey et al., 2010), coupled with changes in other physical, chemical and biological characteristics of water (Vanlandeghem et al., 2012), 1 University of Ghana http://ugspace.ug.edu.gh which results in a sequence of environmental problems, for example, shortage of safe drinking water (Bao et al., 2012), deterioration of aquatic environments (Hu and Cheng, 2013), and sudden uprising of epidemic diseases (Zhao et al., 2012). With environmental pollution becoming an increasingly global canker, the issue of water quality has attracted serious attention from public and governments. The primary concern of the public is the status of water quality, so water quality assessment has been extensively studied globally (Nives, 1999). However, governments worldwide are not only concerned about the status, but also about the future trends of water quality (Crosa et al., 2006) and so measures for the protection of water quality is very crucial. Besides the need of Dams for drinking water, dams play a vital role in various sectors of economy such as agriculture, livestock production, forestry, industrial activities and hydropower generation. In 1982, the Kpong Dam was created on the Volta River to provide potable water to the eastern part of Accra, the national capital. Apreliminary survey through various sub-catchments of the Kpong dam of the Volta River reveals array of agricultural and socioeconomic activities including aquaculture and small-scale surface mining undertakings without proper coordination and regulations. According to Dorner et al (2007), such activities cause deterioration of surface water quality. It is being noted that growth of human population is intricately linked to extent of pollutants loads in water bodies (Akuffo, 1998). This strong bond could be traced from three major input sources of pollutants to water bodies: Urbanisation, Industrialisation and large-scale agricultural development. This is the situation along the sub catchments of the Volta River, which poses a potential threat to its water quality as well as the Kpong Dam. According to Barlwin, 2009, dams affect the river system 2 University of Ghana http://ugspace.ug.edu.gh below dams, altering characteristics such as pH, temperature, flow rate and morphology. Therefore, monitoring of water quality in the Dam is of extreme importance. An important part of any water quality monitoring program is the reporting of results to both managers and the public who may need the water quality information expressed in a concise form for decision making. Sometimes it is difficult to assess water quality from a large number of water quality parameters. Traditional methods to evaluate water quality are based on the comparison of experimentally determined parameter values with existing local guidelines, which does not provide a global summary on the spatial and temporal trends in the overall water quality (Debels et al., 2005; Kannel et al., 2007). This practice was widely used in rivers (Xu et al., 2012), reservoirs, and lakes (Xu, 2005) to produce results describing status of water quality and compliance with official guidelines. The advantage of this approach is that it provides a wealth of data and information. However, since one water test sample can have more parameters determined on it, such numerous variables do not normally convey any meaningful information to the layman, decision maker, policy makers, water managers or the public (Ansa Asare et al., 2013). To overcome this difficulty or water quality decision making problem and ensure efficient management of water bodies, a comprehensive index method, known as water quality index can mathematically be used to reduce the bulk nature of water quality data to produce a single value which can provide a general and readily understood description of the water quality status (Darko et al., 2013). In this way many methods such as the rank correlation method (El-Shaarawi et al., 1983), non-parametric Mann Kendall tests method as well as other Time Series methods can also be successfully used to determine whether the water quality has consistently improved, deteriorated, or remained constant over long-time scales, or the water quality has changed significantly because of natural or anthropogenic causes, so that potential causal mechanisms can be explored for the water-quality changes. Accordingly, this study is 3 University of Ghana http://ugspace.ug.edu.gh set to apply Times Series as well as Mann Kendall Test method to the monthly water quality index data of the Kpong dam to detect water quality changes from 2011 to 2016 and predict future trend. 1.2 Study Area Profile The Kpong Dam (Fig. 1) lies within coordinates of 6° 08' to 6° 12' N and 0° 07' to 30' E. It was created in 1982 through damming River Volta mainly to augment the need for energy for the smelting of aluminium at VALCO in Tema (VRA, 2014) and to satisfy the demand for potable water supply for the capital city, Accra. The Kpong Dam, also known as the Akuse Dam, is located about 24 km downstream of Akosombo Dam, and is about 80 km from the city of Accra (World Bank, 2014). The dam currently provides water to all parts of Accra and its environs, supports irrigation undertakings, as well as fisheries. The Dam is 14 km long, 2.2 km wide and has total surface area of 40 k m2 with mean depth of 5 m (Volta River Authority, 1981). The normal surface elevation is estimated at 14.37 km with maximum of 15.24 km (Volta River Authority, 1996). The Volta River forms a major part of the coastal river systems in Ghana and is the largest of the three river systems across the country (Warm, 1998). It is the main drainage channel that spans from the north to south of Ghana and across other West African countries. Owing to its topographical location it plays a very significant role in the history of the socio-economic development of Ghana. The river has three main tributaries, notably the Black and White Volta Rivers, the Oti River and the Lower Volta, including Lake Volta. 4 University of Ghana http://ugspace.ug.edu.gh The river has a low-lying plateau with undulating topography and isolated ridges. Climatic conditions are wet and dry semi equatorial zones with temperature averaging between 27 to 28oC. Rainfall is moderate with the seasonal average being 780 – 2160mm. and relative humidity 77 – 85% 1500 mm. The catchment of the dam lies in the coastal tropical zone where rainfalls occur twice in a year and peaks in June and September, while dry periods span between December and March. Presently the river serves an estimated 1.5 million in Accra alone (GWCL Records, 1999) and over two million people living within the Basin or close to it. The river is said to provide water for domestic, agricultural and industrial activities and serves as a disposal facility for both domestic and industrial waste. This has brought untold stress on the river to an extent that it is known to be over-stressed environmentally since various signs of water quality deterioration have been observed in the river water. The main economic activities in the catchment are fishing and crop farming. Major crops include maize, cassava, sugarcane and vegetables. Untreated household waste liquids are discharged into the river. 5 University of Ghana http://ugspace.ug.edu.gh Figure 1: Map of the Study area 6 University of Ghana http://ugspace.ug.edu.gh 1.3 Statement of the problem Man’s need for water is not only a function of quantity but also of the quality of the water (Kowalkowski et al., 2007). There is the need to understand the variations of raw water quality as it has a bearing on the risk associated with the final use of the water particularly abstracted for potable water production (WHO, 2006). The Kpong dam is a very important source of water supply for local communities around Kpong and other areas including the Accra-Tema Metropolitan area. This source of water however risks getting polluted by several anthropogenic activities that occur within the Volta River catchment. Unregulated mining activities, poor waste disposal, bad farming practices and extensive human and animal contact with the river are common. To control pollution of the river and avoid incurring high cost in water treatment, it is necessary to adopt water quality protection measures including water quality trend monitoring tools to determine levels of pollution over time scales in order to explore potential causal mechanisms to protect life from effects of pollution of the water. 1.4Aim and Objectives of the study 1.4.1 MAIN OBJECTIVE: To assess raw water quality status of the Kpong dam and establish the WQI trend. 1.4.2 Specific Objectives: 1. To perform trend analysis by plotting Time Series using monthly raw WQI data from 2011 to 2016 2. To identify trend and the best fit times series model to the water quality data by employing statistical models. 7 University of Ghana http://ugspace.ug.edu.gh 3. To forecast raw water quality of the Kpong Dam for the next 2 years on a monthly basis. 1.5 Justification of the Study problem statement Man’s need for water is not only a function of quantity but also of the quality of the water (Kowalkowski et al., 2007). There is the need to understand the variations of raw water quality as it has a bearing on the risk associated with the final use of the water particularly abstracted for potable water production (WHO, 2006). The Kpong Dam which has been in operation over decades is of great significance to the nation as it supplies potable water to the fast-growing population. The effect is that water managers must keep taps flowing without compromising quality for future generations. However, this source of water risks getting polluted by several anthropogenic activities that occur within the Volta River catchment. These include galamsey activities, poor waste disposal, bad farming practices and extensive human and animal contact with the river. To control pollution of the river and avoid incurring high cost in water treatment, it is necessary to adopt water quality protection measures including water quality trend monitoring tools to determine levels of pollution over time scales in order to explore potential causal mechanisms to salvage life from effects of pollution. The outcome of this research would therefore be crucial to the Ghana Water Company Limited (GWCL), Kpong, Akuse and the academia. The GWCL and the Volta River Authority would depend on the model and forecasted values to guide their operations. Also, researchers in the academia would use the study as literature in other related areas. This information would again help provide a system-wide synopsis of water quality of Ghanaian Rivers. 8 University of Ghana http://ugspace.ug.edu.gh 1.6 Scope and Limitation Out of the many dams on the Volta River, a lengthy raw water quality data from 2011 to 2016 were only accessible from the Kpong dam treatment plant. The study seeks to compute monthly WQI to determine raw quality class of the Dam over the period, forecast and predict future trend quality scores of the Dam. Other setbacks include time constraints and the difficulties in obtaining relevant materials. 1.7 Thesis Structure: This study is being put into five chapters. Chapter one (1) is a brief introduction and background of the work, including the study areas, problem statement and objective of the study. It also presents the justification and limitations of the study. Chapter two (2) highlights related literature on the topic with ideas of different authors whose findings have been defined in relation to the topic under study. Chapter three (3) focuses on methodological review in the light of mathematical and statistical tools that are relevant to the analyses of the data. Basically, the study seeks to use WQI concept, Mann Kendal Test and time series model for the analyses. Chapter four (4) deals with the data collection and analysis, and the findings from the application of the various time series models. In the same way, chapter five (5) consists of summary, conclusion and recommendations. 9 University of Ghana http://ugspace.ug.edu.gh CHAPTER TWO LITERATURE REVIEW 2.0 Introduction In this chapter, there is a review of work of many authors concerning concept definitions and various researches done. Research works, empirical work and authors’ opinion are looked at. Below are the focuses of the review. Globally, water quality studies have accumulated enough information on the effects of quality degradation factors predisposing the population to future risk. Knowledge of the effects of water quality degradation is difficult to understand due to the complexity of a variety of pollutant input sources. Therefore, to ensure adequate amount of water and of appropriate quality within a surface water body like rivers, reservoirs and lakes, an Integrated Water resource management (IWRM) approach is needed (Rivers-Moore and Jewitt, 2007). With rapid increase in population especially in developing nations like Ghana, the need for an intensive and integrated water resources management to protect the integrity of the water resources is of extreme importance. Transboundary pollution effects resulting from upstream countries that share the Volta River with Ghana has exacerbated the status of pollutant levels. It is certain in literature that about 44% of the total fresh water in the country is required for domestic purposes, with 54% and 3% for agricultural and industrial sectors respectively (Boateng, 1995). The acute water shortage problems currently affecting many sectors of the economy of the country have telling consequences on the need for effective water quality monitoring. Pollution and land degradation are also undermining the already stressed resource base. A variety of contaminants enter the water bodies in Ghana because most manufacturing industries do not have treatment plants and so release the generated raw waste directly into 10 University of Ghana http://ugspace.ug.edu.gh them especially the surface water bodies like the rivers. Moreover, raw domestic liquid is also discharged directly into these water bodies. The influx of these wastes coupled with the indiscriminate disposal of wastes largely from anthropogenic activities have led to the death of water bodies like the Korle Lagoon in Accra, Sakumono Lagoon in Tema and Fosu Lagoon at Cape Coast. Presently, several reports from the mass media about the extent of generative state of the Densu River and its reservoirs are astonishing. The Volta River basin is probably the next on the series since similar catchment activities also exist in this river too. Currently there is numerous consumer complains about the taste and odour characteristics of water from the Kpong Dam. Apart from this deteriorative nature of our water bodies, it is being speculated that the nation may face an additional problem of insufficient water supply by the turn of the century. The need to stop over reliance on the Volta River is indeed most timely. The Densu River which forms part of the coastal river systems of Ghana and serves some significant communities in the capital city of Ghana, Accra West, Nsawam, and Koforidua is said to be in a bad state due to contaminant loads. This plays a significant role in Ghana’s water monitoring problems due to its topographical location though relative to another river system. The problem of water quality and sustainable development cannot be said to be given full focus without considering the problem of pollution and quality of the Volta River as well as its dams if the aim of potable water for all Ghanaians by the year 2020 is to be achieved (Warm, 1998). This is because a country’s economic development rests on the healthy state of the people, which also relies to a large extent on the quality of water available. This necessitates for studies into the water bodies in Ghana in relation to their water quality and monitoring. An insight into the quality status of water from the Volta River Basin will not 11 University of Ghana http://ugspace.ug.edu.gh only help enhance the numerous methodologies of water quality monitoring in the basin but will also help to improve the most important aesthetic properties like taste and odour of treated water from the Kpong Dam since the quality of any finished product has a direct link to the quality of the raw materials used for its purification process. 2.1 Surface Water Chemistry Water quality studies focuses on the presence of unwanted substances in water and their effects on people and aquatic health. Water of acceptable quality for one purpose may be considered to be of unacceptable/undesirable quality for some other use. Water quality studies therefore lies on the concentrations as well as types of substances that are present in water for the determination of its intended uses. Quality of water is therefore described in terms of its physico-chemical and biological characteristics which make it suitable for a particular application (Khalil et al, 2011). Fresh water bodies like rivers, streams etc. are dynamic aquatic ecosystems and so are able to remove a significant part of incoming loads of pollutants but this capacity is limited and strongly relies on the internal local characteristics of the system (Howarth et al., 1996). That is, they are able reorganise and adapt to changes naturally upon receiving small amounts of wastes due to dilution effects (Spellman, 1996). On the degree of dilution, there may be significant increases in concentrations of various water quality parameters. Increase in levels of organic substances encourage the growth of decomposers such as bacteria and fungi, which breakdown the biodegradable organic substances into their cells and into basic substances like carbon dioxide, nitrates, sulphates and phosphates (Moran et al., 1986). Higher concentrations of organic load above the capacity of a water body makes it 12 University of Ghana http://ugspace.ug.edu.gh unfit for normal aquatic life due to oxygen deficits especially for organisms that are sensitive to changes in oxygen levels (Smith, 1974). Haslam (1995) noted that contaminants entering surface waters either move downstream with the water or be deposited on the bed, accumulating particularly in silt. Within a water body, physico-chemical agents as well as biological agents may neutralize the effect of contaminants. However certain contaminants are resistant to such biogeochemical processes even when in minute levels. Moreover, other substances are decomposed slowly so downstream purification is ineffective if concentrations are high or if further impurities are continuously added downstream, for example from a series of point sources. Grimaldi and Simonds (1989) cautioned that the days when natural methods in streams (e.g., the self- purification process) automatically compensated for increase in anthropogenic pollution are over. It is therefore essential that effluent discharged into a water course is of high quality and the degree of pollution is such that the self-purifying capacity of the river is not overloaded. 2.2 Surface Water Quality in Ghana- An Overview: Ghana has a huge hydrological structure, the Akosombo Dam which covers about 4% of the area of the country; quite enough to satisfy various water needs in the country but climatic conditions that are affecting regular rainfall trends has under scored its intended purpose (Warm, 1998). However, the emerging water problem that is affecting every part of the country economy is conferring a very different picture. It is certain that in the previous years, water crisis was very uncommon, though only urban dwellers had access to safe drinking water, majority mostly the rural dwellers had to go long distances for water especially during dry seasons when water bodies are dried up. Today, the image is different and water crisis is becoming pervasive and persistent. Drought has increased and underscored the weaknesses of 13 University of Ghana http://ugspace.ug.edu.gh the existing water supply problems being experienced in every sector of Ghana (Warm, 1998). Over rise in human population within Ghana and in other West African neighbour countries which share the Volta River with Ghana have increased the demand for water for it diverse purposes. This has resulted in acute water shortages, which is affecting various facets of the Ghanaian economy (Warm, 1998). In Ghana, however, inadequate amount of water is not the only problem but its quality as well. According to Bugri, (1998), the availability of safe water could have eliminated more than 50% of water related diseases affecting Ghanaians. He asserted that eradicating guinea- worm disease was still a task for the department of Public Health because places which have been cleared have been re-infected by people from areas where there is no safe drinking water. The dominant water-borne and water-related diseases including typhoid, dysentery, schistosomiasis and onchocerciasis which have been the root causes of death due to ingestion of polluted water common in rural areas have become a point of interest that need attention (EPA, 1997). From an environmental perspective, the primary principle states that water should be made available in a potable form for the entire population with minimal effort, and that its availability on a sustainable basis should be guaranteed (EPA, 1997). The World Health Organisation guideline for drinking water quality also indicates that, a supply of drinking water should be sufficient in quantity, wholesome and not injurious to health (WHO, 1996) but this is not the picture in Ghana. According to the Ghana Environmental Protection Agency, 1997, only 30% of the Ghanaian population have ready access to potable water. It is also noted that although Ghana is blessed with water resources besides gold and diamond seasonal shortages of water is prevalent. The leading cause of this social canker is primarily 14 University of Ghana http://ugspace.ug.edu.gh linked to lack of proper monitoring systems and inadequate use of available technologies (EPA, 1997). Colerangle (1994) linked the decline in water quality in the country to population growth. Warm, (1998), reported that the elements responsible for the poor water quality in the country are the belief of its sufficiency and so is handled anyhow, weak financial positions of water agencies, catchment degradation, impacts of transboundary water uses, poor, centralized and uncoordinated management policy frameworks, coupled with population growth and lack of sufficient data on the quality of water from our water resources 2.3 The Problem of Pollution The issue of pollution results from the discharge of inadequately treated or raw wastes on the surface of the land, in-shallow pits and in some cases directly into water bodies. The main industrial sector polluters are the breweries, textile mills, food and beverage and petroleum refining companies. Most of these are in the Accra-Tema Metropolis (Warm, 1998). The mines also pollute the waters with their liquid and solid waste from gold, bauxite and diamond ore processing plants, which are discharged, untreated into nearby water bodies. The Pra and Ankobrarivers suffer the most from these mining activities (Warm, 1998). Agricultural practices also contribute a great deal to the pollution problem. These include pollution from use of agro-chemicals such as fertilizers, pesticides, and insecticides. Leaching of soil nutrients also takes place where lands are deforested through bush fires, lumbering, and fuel-wood harvesting. The nutrients find their way into water bodies to impact on their quality. The Densu River is presently said to be the most polluted river from human, industrial and agricultural waste. Because of this cost of treatment to make the Densu water suitable for domestic use in the Accra-Tema Metropolis is one of the highest in the country 15 University of Ghana http://ugspace.ug.edu.gh (Warm, 1998). Indeed, pollution can be said to be the number one threat to water quality management system in the country. The Environmental Protection Agency (EPA) of Ghana, which has the responsibility to control pollution, emphasizes lack of data and information as the factors hindering the proper management of pollution. This is because without having adequate knowledge of the nature of our water resources, appropriate steps cannot be taken to conserve, preserve or improve upon the quality status of our water resources (EPA, 1997). An evaluation carried out in 1995 by the Water Resources Institute (WRI), showed that, there were weaknesses and gaps in the collection of water cycle data. In particular, surface water quality was not monitored on a regular basis and data on groundwater, sediment discharge and soil moisture were hardly collected. According to the report, present evaluation of water quality is based on limited parameters. In particular organic micro-pollutant, pesticides and other toxic substances were not normally evaluated. This is because the specialised equipment to do so are not available (Lamie,1995). Nitrogen and other elements or metals are vital or essential micronutrients for growth and metabolism in plants and microorganisms while others like Cd, Cr and Pb have no known biological function but are proved to be harmful beyond certain limit (Marschner, 1995; Bruins et al., 2000) but are environmental concern because of their persistence (Armitage et al., 2007), toxicity, and ability to be incorporated into food chains (Forstner and Wittman, 1983). Agriculture and urban activities are known in literature to be the major sources of nitrogen, phosphorus and pesticides to aquatic ecosystems. Chemical elements from these sources are difficult to measure and regulate because they derived from multiple activities dispersed over wide areas of land and are variable in time due to effects of weather. In aquatic ecosystems, excess of the nutrients leads to diverse problems such as growth of harmful algal blooms, 16 University of Ghana http://ugspace.ug.edu.gh oxygen depletion followed by fish kills and subsequent loss of biodiversity and other problems. Nutrient excessiveness seriously degrades aquatic habitats and impairs the use of water for industrial, recreation and domestic purpose. It is certain on review of scientific literature that nutrient enrichment is the main cause for the widespread eutrophication problems in surface waters especially rivers. In the United States, excess application of nitrogen containing fertilizers and manure production on farm lands creates surplus nitrogen which is mobile in many soils and often leaches to downstream aquatic ecosystems, and which can also volatilize to the atmosphere, redepositing elsewhere and eventually reaching other aquatic environments (Beaton et al., 1995). In freshwater, blooms of cyanobacteria (blue-green algae) are a prominent symptom of over enrichment of nutrients (Kotak et al., 1993, McComb and Davis 1993, Smith 1998). These blooms continue to a wide range of water-related problems including foul odours, unpalatability of drinking water and formation of trihalomethane during chlorination in treatment plants (Palmostrom et al., 1988, Kotak et al., 1994). Release of water soluble neuro and hepatotoxins from dead cyanobacterial blooms are harmful to livestock when ingested and could cause a serious health related risk to humans (Lawton and Codd 1991, Martin and Cooke 1994) A newly discovered toxic flagellate has been associated with the mortality of finfish on the U.S. Atlantic coast (Burkholder et al., 1992). The highly toxic, volatile substance produced by this alga can cause long term neurological damage to people who can encounter it. Presently, uncontrolled disposal of urban wastes, run-off, atmospheric deposition and domestic and industrial discharge as key/major sources of aquatic pollution (wassawa, 1997) has made water quality monitoring a matter of great concern in streams and river systems. 17 University of Ghana http://ugspace.ug.edu.gh The human activities that influence the Volta river water quality can be traced through small scale industries, farming, domestic waste water discharge and uncontrolled open defecations. Discharge of heavy metals pollutants from industrial activities, are the major source of water quality impairment. The river basin is also subject to receive surplus chemical substances from the farmlands utilizing N-containing fertilizers and pesticides; this can be attributed to the high population of people living around the basin and its tributaries. Empirically, the volume of water released by a river in a certain amount of time has been used extensively as a covariate in the analysis of the physico-chemical and biological characterization and in the establishment of the status of water quality criteria for rivers being assessed for discharge of waste effluents, based on low flow rate conditions. However, levels of contaminants influence water chemistry among other parameters with varying interactions in different rivers (Chapman and Esterby, 1996). Natural factors such as a change in climate, for instance, a dry season could attribute to water quality fluctuations of rivers and streams (Cho et al., 2004). It is certain from literature that dynamic changes in patterns of water utilization due to human civilization has also led to uncontrolled influx of all forms of wastes into aquatic systems affecting the quality of these essential part of our natural environment (Song and Kim, 2006). In 2007, Kan et al., made an empirical observation that industries, factories, outputs from agricultural sectors, intensive and extensive use of fertilizers and pesticides coupled with waste liquid effluent treatment works are the key potential sources of pollution. 2.4Input Sources of Pollutants Entry points of pollutants in inland waters are generally classified as point or non-point sources: 18 University of Ghana http://ugspace.ug.edu.gh 2.4.1 Non-Point Sources Entry points of pollutants into water systems through Non-point source generally emanates from multiple areas like runoffs from agricultural lands, cleared forest areas, construction sites, roads, and precipitation, atmospheric deposition, seepage from underground tanks and hydrological modifications (Miller, 1988) [cited in Spellman, 1996]. Among these multiple pollutant sources, wastes coming from agricultural fields are of most significant interest to environmentalists in recent years. A report by Mason (1990) indicates that farm wastes are estimated to be more than 200 times as hazardous (in terms of BOD) as treated domestic effluent. Accordingly, Chapman (1992) also observed that an underlining distinction between a point source and a nonpoint source is that discharges from a point source could be managed by conventional methods whereas non-point sources may not. 2.5.2 Point Sources Point sources of pollution as defined by Chapman (1992), pollutants result from discrete conveyances/single outlets, such as discharge pipes like effluents from factories, treatment and industrial plants. Major entry points of pollutants to surface waters especially freshwater bodies like rivers, streams, lakes and ponds originate from the collection and discharge of domestic wastewater, industrial effluents or certain agricultural activities, such as animal husbandry. As is known in a literature report by Chapman (1992), that the discharge of sewage, both raw and inadequately treated is probably still the greatest point source of pollution to the world’s waters. Other important sources include mine and industrial effluents. 19 University of Ghana http://ugspace.ug.edu.gh 2.5 Studies on Water Quality from other countries An investigation carried out by Zhang et al., 2010 on the influence of anthropogenic activities on river water in China indicated that the water problems of the river Basin were primarily related to pollution by nutrients, heavy metals as well as sediments and soil in the watershed. This confirmed the findings of Chen et al., (2004), Tang et al., (2008) and Zhang et al., (2009) further observed that heavy metals contamination is considered to be point source which is primarily discharged from smelting and heavy industrial enterprises in addition to other nonpoint sources like domestic waste water and agricultural run- off. Elemental analysis conducted on a water sample from river Dzindi, Madanzhe and Mvudi in Thohoyandou, South Africa by Okonkwo et al., (2004) using a Varian Spectra AA 220 atomic absorption spectrophotometer indicated a seasonal variation in the levels of Cd, Cu, Pb and Zn. The amounts of these metals were generally lower for the dry season than the wet season. This can be attributed to runoff from land into the rivers during the wet season. It was also observed that the Pb concentrations in water samples from all the rivers were significantly higher than the values for the other metals in both wet and dry seasons. These high levels shown by Pb may be attributed to the deposition of Pb particulates on the roads next to the rivers especially during precipitation. The high concentrations of Pb in Rivers may also have been influenced by the effluents from the nearby sewage treatment plant and a waste dumping site respectively. Also, the agricultural activities around the rivers may have contributed to the observed high levels of Pb and Cd levels, since these metals can occur as impurities in fertilizers and in metal-based pesticides and compost manure. Sekabira et al., (2010) have recently investigated the heavy levels and extent of contamination of the Nakivubo channelized stream water in Kampala, Uganda. The results revealed that the 20 University of Ghana http://ugspace.ug.edu.gh concentration of Pb, Cu and Mn have exceeded the WHO (2008) standard for drinking water at most of the sites and maximum permissible limit of discharge of waste water into the environment (100um/L). Water resources management studies and water quality in particular has been one of the areas that draws attention. Iguisi et al, (1999) reported the presence of heavy metals in the Kubanni river with very significant concentration levels. Butu, (2002) in his study on variation in concentration of selected heavy metals in the lower and upper regions of the Galma dam, Zaria, reported the presence of heavy metals in high concentration within the Galma dam. Ahmed and Tanko (1994a or b) observed that concentration of calcium, magnesium, sodium, manganese, nitrate, chromium and lead vary spatially and seasonally within the Challawa drainage basin in Kano. Lead, chromium and manganese were observed to be in high concentration which could be a threat to the health of both animals and plants. The increase in chemical pollution of water in the basin was due to increasing use of chemicals in Agricultural activities and industrial discharge. Akan et al., (2007) found high levels of heavy metals; copper, cobalt, zinc, iron, lead, manganese and chromium within the Challawa drainage basin. All observed concentration was above USEPA and WHO permissible limit for drinking water. Also observed is the level of phosphate exceeding the FEPA tolerance limit for drinking water and water meant for domestic uses. The Kano River basin serves as the main source of water supply to metropolitan Kano. A study by Bichi and Anyata, (1999) within the basin reveals that the Salanta river receives the highest pollution from industrial discharge with COD of 8557.4mg/l, total solids of 21 University of Ghana http://ugspace.ug.edu.gh 16934.6mg/l, hardness of 1349.6mg/l CaCO3, and ammonia- nitrogen of 5150.0mg/l. The Challawa river had COD of 598.7mg/l, total solids of 1609.9mg/l, hardness of 1332.0mg/l CaCO3 and ammonia nitrogen 400mg/l. Both rivers empty into the Kano River where the COD was 1166.9mg/l, total solids 1458.0mg/l, hardness 2506.8mg/l CaCO3 and ammonia- nitrogen 530mg/l. Although these rivers are used extensively for water supply, irrigation and fishing, the quality of water was found to be unsuitable for these purposes. It is based on the foregoing that it’s necessary to carry out a research that will assess the impact on quality of water body overtime by elements that were found to be in high concentrations by previous operations. 2.6 Water Quality Studies on Dams and Rivers in Ghana The Kpong Dam and many others of the Volta River were built to produce hydro-energy to enhance socio-economic development and bring prosperity to the peoples of Ghana. However, one of the major environmental problems has been the invasion, development and spread of aquatic pollutants causing nuisance in the reservoirs created and rivers upstream and downstream of these reservoirs. This has necessitated the need for many research undertakings. A study carried out by Mensah (1976) on possible input sources of contaminants into the kpong dam showed that industrial facilities such as the Akosombo and Juapong Textile factories release unprocessed wastes into the lower Volta River. Akosombo Textiles Limited (ATL) alone was reported to be consistently releasing 118,650,000 litres of unprocessed wastes monthly into the Volta River. The wastes contained salts and were of a higher temperature. He noted that despite the dilution effect by the large volume of river water, the water was deeply coloured. In situ measurements of physical parameters like temperature, 22 University of Ghana http://ugspace.ug.edu.gh pH, BOD and chemical parameters such as ammonia were found to be high, indicating incidence of contamination. Fauna at the point of discharge were greatly affected when compared with those found upstream and downstream at the point of discharge of the wastes. Biney (1977) in pre-impoundment chemical studies of the Lower Volta River found that differences with respect to chemical constituents of water samples from different sites were not very significant with the exception of the sample from ATL which recorded high concentrations of parameters measured. He reported that the alkalinity had been increasing since September 1976. The recovery capacity of the river was generally good since for most constituents, the concentrations below the ATL discharge point were not very different from upstream the point of discharge. He therefore concluded that the level of pollution of the Lower Volta was low since most of the parameters, which indicate the presence of pollution, occurred in low quantities. Amoah (1989) also reported that effluents from ATL had lowered the dissolved oxygen of the river and increased suspended solids, the latter making the water more turbid. He added that effluents from the Akosombo Sewage Treatment Plant encouraged the growth and multiplication of indicator bacteria when discharged into the river. Larmie (1993) undertook an assessment of the water quality characteristics of the Lower Volta River. He observed two important sources of pollution within the Lower Volta Basin as textile and domestic sewage effluent releases at Akosombo. Contrary to Mensah’s (1976) report on the high temperature conditions of the river after the discharge of the ATL effluents, Larmie (1993) reported that the effluents from the ATL industries did not seem to have any impact on the temperature conditions of the Lower Volta River. Larmie (1993) also observed that the river water was generally alkaline in character. He recorded pH values ranging from 6.7 to 7.6 in 1990 and 23 University of Ghana http://ugspace.ug.edu.gh 1991 respectively. He also observed that the effect of the textile effluent on the dissolved oxygen (DO) content of the river was more pronounced in the dry season than in the wet. The range of the DO values was between 5.0 and 7.5 mg/L. He recommended the use of the Lower Volta River as a raw water source for domestic water supply as far as from the Akosombo dam to Agordome. In a comparative study of the nutrient status of the Weija Lake and the Kpong reservoir in Southeast Ghana. Ansa- Asare and Asante (1998) recorded a mean temperature range of 28.9-30.8 °C for the Kpong reservoir. Over the 5-year study period, temperatures were more or less constant. The pH was also about neutral with a range of 6.7-7.0 and alkalinity values ranged from 40.3-52.0 mg/L as CaCO3. Low suspended solids concentrations with a mean value of 4.32 mg/L were also recorded. Ansa-Asare and Asante (1998) also observed that the yearly average orthophosphate levels were mostly moderate in relation to natural levels (0.02 mg/L). The yearly mean values ranged between a minimum of 0 mg/L to a maximum of 0.08 mg/L. Nitrate-nitrogen concentrations ranged from a yearly mean of 0.2 to 0.5 mg/L while ammonia-nitrogen recorded yearly mean values ranging from 0 mg/L to 0.4 mg/L. The distribution of ammonia concentration showed a gradual increase over the 5- year period at both Weija and Kpong. Ansa-Asare and Asante (1998) however observed that on the average, there was complete nitrification of ammonia to nitrate over the 5-year period. The Kpong reservoir recorded moderate dissolved oxygen concentrations. The five-year monthly means ranged from 4.6 mg/L in January to 9.0 mg/L in September. BOD values recorded were comparatively lower than the Weija. The 5-year monthly mean concentrations ranged from 2.2 mg/L in May to 4.22 mg/L in March. 24 University of Ghana http://ugspace.ug.edu.gh A study conducted by Biney (1990) on Ghanaian freshwater and coastal ecosystems concluded that the Korle lagoon located at the heart of the capital, Accra was the most polluted water body in Ghana. This is because the Korle lagoon catchment is the most industrialized and most populated in the Greater Accra region, which has resulted in uncontrolled increases in the quantity and diversity of discharges that reach the lagoon. Before the insurgence of urban and industrial activities in the Korle catchment, the Korle lagoon was clean enough to support aquatic lives, the dominance of which were tilapia, mullets. However, the gross pollution of the catchment over the years has led to a complete reduction of biodiversity especially the major fishery. Biney (1996) indicated that substances which contribute to increases in Biochemical Oxygen Demand (BOD), originating from domestic sewage, industrial effluents and land runoff are the most important sources of pollution in all the catchments. Of the calculated total daily BOD of 140 x 103 kg in 1990, 68 % was estimated to have been received by the Odaw/ Korle/ Chemu 1 catchment, 16 % by the Sakumo II and 5 % by the Kpeshie/ OsuKlorte catchment. Amuzu and Leitmann (1991) undertook an Environmental Profile of Accra and reported that most surface waters in the Greater Accra region were seriously polluted as a result of the indiscriminate disposal of raw household and industrial waste into the water resources. They also brought to light that acute pollution is confined mostly to waters in urban and industrial areas. Nana- Amankwah et al. (1995) in a study of the impact of development and urbanization on the Nima Creek in Accra reported that, the middle and lower segments of the creek were heavily polluted with human faecal waste. This condition was attributed to high population densities and poor standard of living in the area. Industrial wastes as well as domestic wastes are polluting several other water bodies in Ghana. Mensah (1976) in a survey of water quality and pollution of inland and coastal waters of Ghana, observed that 25 University of Ghana http://ugspace.ug.edu.gh most industries in Ghana pollute inland waters and make them unfit for several purposes. For example, some diamond-mining companies at Akwatia, Kade and Edubiase discharged mine wastes containing considerable amounts of suspended solids; mostly silt, about 18,000 mg/L into the Birim River. The resultant turbidity of the river downstream of the mines could be as high as 800 NTU. Dadzie- Mensah (1999) in a similar study reported that the Ankobra River is significantly polluted due to mining activities in the Prestea area. Chemicals such as arsenic, cyanide and mercuric compounds from the extraction of gold are washed directly into the river, polluting it to such an extent that areas downstream no longer benefit from the rich resources of the river. Ameka et al., (2000) conducted a limnology study on the Weija dam and other fresh water bodies in Ghana and observed seasonal variation in biological, physical, and geological properties. Amuzu, (1973) conducted similar studies on the Barekese reservoir. The results indicated that the water problems in barekese and Weija reservoirs are primarily related to water pollution by nutrients. The study further observed that heavy metal and nutrient pollution are point source which primarily discharged from nearby industrial enterprises in addition to other non-point sources like domestic waste water and agricultural runoff. Some comparative studies on nutrients such as nitrogen and phosphorous concentration levels on Weija and Kpong over a 5-year period from 1993-1997 was carried out by Ansa-Asare, 1992. The results showed a seasonal distribution with the nutrients in the weija dam being found to be higher than the Kpong dam. Seasonal variation also showed higher concentrations of nitrates in the wet season at Weija than at Kpong. The monthly trend of nitrates showed a unique pattern in the months of March and May, the main ploughing periods for farming in the Weija catchment area when most fertilizer is applied. In general, over the 5-year period, there were no trends in nutrient levels. However, sulphates showed a gradual decline in 26 University of Ghana http://ugspace.ug.edu.gh spatial distribution from 1993-1997, and phosphate index in January at both Weija and Kpong was high. A similar comparative study on Weija, Brimsu and Inchaban reservoirs in Ghana in 2004 by Bosque-Hamilton again further confirms earlier works. The results indicated that the Weija dam was well mixed and aerated but less transparent with high aquatic plants. Ansa Asare et al., (2013) undertook water quality monitoring and assessment of the South western and the Coastal Rivers Systems of Ghana from 2005 to 2008 for the Water Resources Commission of Ghana (WRC). The Southwestern and the Coastal Rivers Systems cover approximately 30% of the total drainage basins of Ghana. Adapted Water Quality Index (WQI) was used as a tool to classify the overall ambient water quality at 19 different stations. The index classified water quality into one of four categories: good (Class I, >80), fairly good (Class II, 50 - 80), poor (Class III, 25 - 50), and grossly polluted (Class IV, <25). Evaluation of the waters with the WQI indicated that most Ghanaian waters are currently in Class II, the fairly good water quality state, but with variations in this range within the seasons and stations, and from one water body to the other. Potroase in the Densu basin had the best water quality during the study period while Nsawam had the poorest. 2.7 Water Quality Parameters Discharge of dynamic surface water systems like rivers has been widely used as an indicator that is possibly predictive of the outcome in water quality evaluation and in the establishment of water quality guidelines for rivers that are being assessed for disposal of liquid effluents based on low flow rate scenarios. However, varying interactions in different dynamic hydrological systems causes changes among water quality constituents (Chapman, 1996; Esterby, 1996). This could be linked to the fact that numerous environmental stressors such as 27 University of Ghana http://ugspace.ug.edu.gh drought or climate impact greatly on dynamic surface water bodies which can culminate to changes in the quality of the water (Atasoy et al., 2006). The goals and needs of a research on the status of raw water quality in water body informs the selection of its quality evaluation parameters (Bartram and Balance, 1996). The main water quality parameters, namely, temperature, dissolved oxygen and pH are extensively considered in any quality assessment because they influence many chemical reactions in water chemistry with DO, the important ingredient for the sustenance of aquatic ecosystem biodiversity (Chapman, 1996). Water quality of a given water body is assessed in terms of the requirements of its levels of physical, chemical, radiological and biological parameters. 2.7.1 Physico-chemical Parameters Parameters that are dominantly used to describe physical qualities of water include turbidity, temperature, colour, taste, pH and odour. Colour as an important water quality parameter is requirement in potable water for its aesthetic value. Though, it has no known health effect but due to sensory aspect people object strongly to water that offends their sense of sight, taste or smell (WHO, 1984). The colour of a water sample may be due to the presence of dissolved or suspended colloidal particles resulting from mainly decaying leaves or microscopic plants, which tends to give the water a characteristic colour (APHA, 1976). Almost 50% of colour in a waterbody results from influence of the colloidal fraction of organic acid compounds (Pemmanen, 1975). The guideline for colour limits in drinking water is less than 15 true colour units (TCU). This could be achieved by removing colour impacted by organic acids through centrifugation or filtration techniques (Sawyer, and McCarthy, 1967). 28 University of Ghana http://ugspace.ug.edu.gh Turbidity in water is a measure of how turbid or murky the water appears. Turbid water is also unacceptable for aesthetic reasons. It is caused by suspended matter such as clay, silt, tiny fragments of organic matter, and microscopic organisms (WHO, 1984). Turbid water is of grave concern in drinking water, because suspended particles provide hiding places for microbes, thus complicating their detection in drinking water (Brock, 1966) and from the action of disinfectants (WHO, 1984). According to Downing and Crowley (1976), low land surface water systems are comparatively often murky than upland surface waters due to influence of dispersive fine soil particles such as clays which often make them more murky than upland waters which flows over rocks (AWWA, 1990). It is gathered from literature that a Nephelometric readings of turbidity values ranged from 3.8 to 8.4 NTU due to presence of coliform organisms upon disinfection with chlorine (Clarke, 1964). 2.7.1.1Temperature Many aquatic life forms like fish require certain optimum conditions of temperature to live and reproduce. Generally, a change in temperature in water significantly alters the balance and health of an aquatic environment. The basic reason for this is that the solubility of dissolved oxygen decreases as the temperature of the water goes up. Aquatic environments undergo temperature variations along with fluctuations of normal weather conditions. These changes in weather conditions occur periodically and in many water systems, over periods of 24 hours (Chapman, 1996). Several factors influenced the temperature of water but it may exceed recommended limits of standards due to exposures to radiation (Washington State, 1998). Temperatures above set limits has shown that it is an important determinant of biodiversity of aquatic species since it directly influences the rate at which the organisms use oxygen to burn food for energy within a water body. Though temperature is of little direct 29 University of Ghana http://ugspace.ug.edu.gh importance in public water supplies on the basis that many people prefer cold drinking water, it plays critical roles in the treatment of waste water and water pollution control. Effectiveness of a biological liquid waste treatment plants are more efficient at elevated temperatures. In regions of very low temperatures, optimum temperature ranges are maintained by sheltering treatment plants in enclosures. Water temperature is therefore an important constituent for any water resources monitoring system (Rivers-Moore and Jewitt, 2007; Zhang, 2008). 2.7.1.2 pH The pH level of any liquid is an indicator of the degree of its acidity or alkalinity on a scale of 0 to 14 (A. APHA, 1998). On the pH scale range of (0-14), reading of seven is "neutral", the pH of pure water. Readings below seven indicate acidic conditions with the hydrogen ions being more than the hydroxyl ions in solution, while readings above seven indicate the water is alkaline, or basic. The usual pH for fresh waters is six to nine with many water systems around this pH an indication of existence of biological life. However, WHO (2011) stipulates that drinking water should have a pH range of 6.5 to 8.5. Naturally occurring fresh waters have a pH range between six and eight. The pH of water particularly is a crucial factor that influences the solubility and availability of nutrients, and how they can be metabolized by various aquatic species. Water pH is therefore a critical parameter in water quality assessment as it influences many biological and chemical reactions within a water body and all physical, biological, radiological and chemical processes associated with water supply and treatment (Chapman, 1996). 30 University of Ghana http://ugspace.ug.edu.gh The pH levels in surface waters can render the water unusable for all or some activities (Washington State, 1998). The pH of water preferably should not vary by ± 0.5 outside the background range and by ± 1.0 in the natural occurring range (DeCesare and Connors, 2002). Surface water pH can be relatively higher in low discharge since water is rich in solutes characteristic of ground water (Calles et al., 2007). pH can sensitively indicate variations in water quality and is affected by dissolved substances (Yang et al., 2008). Very high levels of acidity or alkalinity in a water sample may indicate the presence of industrial or chemical pollution, though acidity and alkalinity also occur naturally. Carbon dioxide from the atmosphere, or from the respiration of aquatic organisms, causes acidity when dissolved in water by forming carbonic acid. Dissolved carbonates or bicarbonates of sodium, calcium, or magnesium cause natural alkalinity. Contact between water and minerals in the ground the major source of these substances. 2.7.1.3 Dissolved Oxygen (DO) Dissolved oxygen is commonly considered to be one the most important water quality variable in dynamic hydrological systems such as rivers, lakes, dams and streams. This parameter is of interest to most environmental hydrologists because with other parameters, the less there is in the water, the better is the quality. But the situation is the opposite for D.O. the higher the amount of dissolved oxygen in water is, the better the water quality is. Fluctuations in water temperature have a significant inverse effect on dissolved oxygen concentrations because the cooler the water temperature the more dissolved oxygen it can hold. Depending on the temperature and biological activities of aquatic organisms in a water body D.O can vary periodically (Chapman, 1996). 31 University of Ghana http://ugspace.ug.edu.gh Dissolved oxygen has no direct effect on public health, but potable water with very little or no oxygen tastes flat and may be objectionable to people. Dissolved oxygen does play a part in the rate of organic matter degradation by microbes in water bodies (Reckhow, 1994). The amount of dissolved oxygen is a determinant of the survival of organisms (DeCesare and Connors, 2002). The introduction of biodegradable pollutants into surfaces waters such rivers especially from runoff fields during and after a precipitation event alter dissolved oxygen levels (Kannel et al., 2007) 2.7.1.4 Biological Oxygen demand and Chemical Oxygen Demand Biological Oxygen Demand (BOD) is a measure of the need for oxygen by microbes to break down organic pollutants in water under aerobic conditions. Usually, the BOD is used as direct measure of the total amount of biodegradable organic matter as well as the strength of sewage in streams, lakes and river waters. BOD decreases the amount of dissolved oxygen available to other life forms. This measurement is obtained over a period of five days and is expressed in mg/ℓ. Discharge of wastes containing high concentrations of organic materials into water bodies, results in localized areas of Dissolved Oxygen depletion (Chapman, 1996). Sources of these organic pollutants into water bodies are largely from municipal sewage, industrial liquid effluents and multiple source pollutants cause water quality issues to downstream users (Poo et al., 2007). This scenario is characteristic of the catchments of the Volta River where downstream users are at risk of upstream activities. Odour, taste, and colour are physical are physical characteristics of potable water that are important for aesthetic it values. 2.7.1.5 Total Alkalinity Alkalinity of a solution is measure of its neutralizing effect of an acid (Hemond, 1990) and is usually expressed in mg/ℓ. Alkalinity and acidity are correlated to pH when there is no acid- 32 University of Ghana http://ugspace.ug.edu.gh neutralizing effect. However, as most natural waters contain weak acids alkalinity is usually analysed as well as pH in water quality evaluation (Chapman, 1996). The total alkalinity of water is of function of the concentration of mainly bicarbonate (HCO -3 ), carbonate (CO 2- 3 ) and hydroxyl (OH-) ions and is expressed as mg/ℓ of CaCO3 (Dougherty et al., 2007). Changes in total alkalinity in a water body are influenced by changes in its flow characteristics (Brydstenet al., 1990) and its natural trends is attributed to the presence or absence of carbonate rock (Kney and Brandes, 2007). It can also be brought about by biological processes such as denitrification mediation processes in water which increases alkalinity in river water (Kannel et al., 2007). 2.7.1.6 Ammonia In the environment, nitrogen occurs in many forms and takes part in many biochemical reactions. Generally, four forms of nitrogen that are of significance in water quality studies are organic nitrogen, ammonia nitrogen(NH +4 -N)), nitrite nitrogen (NO2-N), and the nitrate nitrogen(NO3-N). The oxidation of organic nitrogen in aqueous solution by biological mediation processes forms ammonium ion (NH +4 ) ion (DWAF, 1996), however, in aqueous solution the un-oxidized ammonia exists in equilibrium with the ammonium ion. Total ammonia is the sum of these two forms (Chapman, 1996). In rivers, ammonia losses tend to be associated with surface runoff and erosion rather than subsurface flow (Heathwaiteet al., 1996). In sewage treatment effluent, the ammonium ion along with urea tends to prevail above other nitrogen compounds (Kannel et al., 2007). Among the species of nitrogen, nitrate levels in water are perhaps the most serious problem. Excessive nitrate levels in rivers, lakes etc. causes eutrophication (encourage the rapid growth of microscopic plants called algae) which degrades water quality (). Excessive nitrate concentrations in drinking water pose an immediate and serious health threat to infants less 33 University of Ghana http://ugspace.ug.edu.gh than three months of age. The nitrate ion reacts with blood haemoglobin, reducing the bloods ability to carry oxygen; this produces a condition called blue baby or methemoglobinemia, in asphyxia (Fan and Steinberg, 1996). Owing to its high toxicity widespread occurrence in water, the allowable concentration limits are set at 10mg/l for drinking water (WHO, 2011). 2.7.1.7 Phosphorous Like nitrate, phosphorous is crucial to the creation of eutrophication of rivers, lakes, although, its presence is known to have little effect on health. In natural waters, phosphorous exists as both dissolved and particulate species. In natural waters and in wastewaters, phosphorus occurs mostly as dissolved orthophosphates (PO 3-4 ) and polyphosphates, and organically bound phosphates (Chapman, 1996). “Orthophosphate, is that phosphorus which is immediately available to aquatic biota which can be transformed into an available form by naturally occurring processes” (DWAF, 1996). Nitrate and Phosphorous are the major nutrients responsible for growth, reproduction and consequently eutrophication in aquatic ecosystems. Anthropogenic activities are attributable for the influx of phosphate and nitrates into waters resulting from sewage or from agricultural runoff containing fertilizers and organic wastes from animal droppings (Chapman 1996; Chen et al., 2007). The problem of Eutrophication is exacerbated when nutrient loads are high with low renewal rate. Unlike BOD, Chemical Oxygen Demand (COD) is a parameter of water quality that measures all organics, including non-biodegradable matter. It is a chemical test using a strong oxidizing agent potassium dichromate, sulfuric acid, and heat to oxidize the materials present in water 34 University of Ghana http://ugspace.ug.edu.gh bodies. The amount of oxygen equivalent of the organic matter that would have been used to break down the organic matter is a measure of COD. COD is widely used as a measure of the susceptibility to oxidation of the organic and inorganic materials present in water bodies and in the wastes from sewage and industrial plants (Chapman, 1996). It is useful for the determination of wastewater quality requirement discharged into receiving waters to limit their impact and can be used as an index for organic pollution (DWAF, 1996; Chen et al., 2007). 2.7.2 Faecal Coliforms One of the most significant characteristics of good quality water is that it be free of disease causing organisms -pathogenic bacteria, viruses, protozoa, or parasitic worms. The most important biological indicator of water quality and pollution used in public health technology is the group of bacteria called faecal coliform bacteria that are always present in the intestinal tract of all warm-blooded animals and billions are eliminated with body wastes. Faecal coliforms bacteria may indicate the presence of disease causing microbes which live in the same environment as the faecal coliform bacteria. The measurement is expressed as the number of organisms per 100 mℓ sample of water (FC/100mℓ). To evaluate possible faecal contamination, faecal coliform count is used as it tells the presence of faecal bacteria in water bodies. The Washington State (1998) reported that the prime cause of deterioration of water quality in streams is faecal coliforms. According to MNM Consultants (2002) high coliform counts are responsible for the impairment of use of surface waters in Swaziland. Faecal pollution can escalate in the summer warm and dry months (Assaf and Saadeh, 2008). This was also confirmed by Sinclair et al. (2009) who 35 University of Ghana http://ugspace.ug.edu.gh observed that during low flows the outlet discharge of streams analysed exceeded the national standards set for faecal coliform counts. 2.8 Some Applications of Time Series Analysis Time series analysis has vast scope in geology, ocean technology, seismology, and it has also been applied to many hydrological and climatological situations. For instance, time series studies have been carried out for analyzing rainfall data (e.g. Mirza et al., 1998; Pugachevaet al., 2003; Astel et al., 2004), streamflow data (Radziejewskiet al., 2000; Fanta et al., 2001; Adeloye&Montaseri, 2002), flood data (Westmacott & Burn, 1997; Robson et al., 1998; Cayan et al., 1999; Douglas et al., 2000; Zhang et al., 2001; Cunderlik & Burn, 2002), infiltration data (Schwankletal., 2000), and surface water quality data (Higashinoet al., 1999), as well as for generating rainfall data (Janos et al., 1988), determining water consumption patterns (Maidment &Parzen, 1984), detecting trends in evapotranspiration and wind speed (Hameed et al., 1997; Raghuwanshi &Wallender, 1997), and for detecting climatic changes (Kite, 1989; Khan, 2001). ARIMA modelling techniques have been applied in many fields of research. For example, Aidoo (2010) applied ARIMA model on the monthly inflationary rates in Ghana. He indicated that Ghana faces a macroeconomic problem of inflation for a long period of time. The problem in somehow slows the economic growth in this country. Using monthly inflation data from July 1991 to December 2009, we find that ARIMA (1,1,1) (0,0,1)12can represent the data behaviour of inflation rate in Ghana well. Based on the selected model, we forecast seven (7) months inflation rates of Ghana outside the sample period (i.e. from January 2010 to July 2010). The observed inflation rate from January to April which was published by 36 University of Ghana http://ugspace.ug.edu.gh Ghana Statistical Service Department fall within the 95% confidence interval obtained from the designed model. The forecasted results show a decreasing pattern and a turning point of Ghana inflation in the month of July. Again, Cui (2011) researched on the topic: “ARIMA Models for Bank Failures: Prediction and Comparison,” Cui said that the number of bank failures has increased dramatically over the last twenty-two years. A common notion in economics is that some banks can become “too big to fail.” Is this still a true statement? What is the relationship, if any, between bank sizes and bank failures? In this thesis, the proposed modeling techniques are applied to real bank failure data from the FDIC. In particular, quarterly data from 1989: Q1 to 2010: Q4 are used in the data analysis, which includes three major parts: 1) pairwise bank failure rate comparisons using the conditional test (Przyborowski &Wilenski, 1940), 2) development of the empirical recurrence rate (Ho, 2008) and the empirical recurrence rates ratio time series; and 3) the Autoregressive Integrated Moving Average (ARIMA) model selection, validation, and forecasting for the bank failures classified by the total assets. 37 University of Ghana http://ugspace.ug.edu.gh CHAPTER THREE METHODOLOGY In calculating WQI, modelling and predicting water quality of the Kpong dam, a time series data was required. Therefore, secondary raw water quality data was obtained from the Ghana Water Company Limited, Accra, from 2011 to 2016 on monthly basis. American Public Health Association (APHA) Standard Methods (APHA, 1998), were employed for the monthly analysis of the various physicochemical parameters. Using long term data was to help ascertain the changes that might have occurred over the years as the water travels on the Volta River course to the dam. Tables, graphs, Moving Average (MA), Autoregressive (AR), ARIMA, WQI and Mann Kendall Test were statistical tools employed. In gathering literature for the study, the following sources were consulted for materials: The University of Ghana Library, GWCL Annual Reports (2009 – 2012), and the internet. The SPSS (version 17.0), and MINITAB (version 14.0), WQI calculator (XSTAT 1.0 on excel spreadsheet) were used in the analysis. 3.1 Basic concepts of statistical tools employed 3.1.1 Water quality index (WQI) concept Water quality, simply put is a single number description of a water-body quality based on aggregation of several water variables determined at a time (Ansa – Asare, et al., 1998). There are several water indices but the CCME (2001) WQI is used for this study. The CCME Water Quality Index (1.0) is based on a formula developed by the British Columbia Ministry of Environment, Lands and Parks and modified by Alberta Environment. The CCME WQI provides a mathematical framework for evaluating ambient water conditions relative to water 38 University of Ghana http://ugspace.ug.edu.gh quality objectives. It combines three factors: scope- the number of variables not meeting water quality objectives; frequency - the number of times these objectives are not met; and amplitude - the amount by which the objectives are not met. The index produces a number between 0 (worst water quality) and 100 (best water quality). These numbers are divided into 5 descriptive categories to simplify presentation (Table 1). Table 3.1: WQI classification by CCME (2001). WQI Quality Description No. 95-100 Excellent water quality is protected with a virtual absence of threat 1 or impairment; conditions very close to natural or pristine levels. 80-94 Good water quality is protected with only a minor degree of 2 threat or impairment; conditions rarely depart from natural or desirable levels. 65-79 Fair water quality is usually protected but occasionally 3 threatened or impaired; conditions sometimes depart from natural or desirable levels. 45-64 Marginal water quality is frequently threatened or impaired; conditions 4 often depart from natural or desirable levels. 0-44 Poor water quality is almost always threatened or impaired; 5 conditions usually depart from natural or desirable levels. The CCME (2001) Water Quality Index (WQI) is calculated from the following equation: √F1 𝟐+F2 𝟐+F3 𝟐 CCMEWQI = 100 - ( ...............eqn. 1 𝟏.𝟕𝟑𝟐 39 University of Ghana http://ugspace.ug.edu.gh Calculation ofF1 andF2 is relatively simple but F3 requires some additional steps F1(Scope) represents the percentage of variables that do not meet their objectives at least once during the time period under consideration (“failed variables”), relative to the total of parameters measured: 𝑵𝒖𝒎𝒃𝒆𝒓 𝒐𝒇 𝒇𝒂𝒊𝒍𝒆𝒅 𝒑𝒂𝒓𝒂𝒎𝒆𝒕𝒆𝒓𝒔 F1 = ( )x 100 ----------(2) 𝑻𝒐𝒕𝒂𝒍 𝒏𝒖𝒏𝒎𝒃𝒆𝒓 𝒐𝒇 𝒑𝒂𝒓𝒂𝒎𝒆𝒕𝒆𝒓𝒔 F2 (Frequency) represents the percentage of individual tests that do not meet objectives (“failed tests”): 𝑵𝒖𝒎𝒃𝒆𝒓 𝒐𝒇 𝒇𝒂𝒊𝒍𝒆𝒅 𝒕𝒆𝒔𝒕𝒔 F2 = ( )x 100-----------------(3) 𝑻𝒐𝒕𝒂𝒍 𝒏𝒖𝒏𝒎𝒃𝒆𝒓 𝒐𝒇 𝒕𝒆𝒔𝒕𝒔 F3 (Amplitude) represents the amount by which failed test values do not meet their objectives. F3 is calculated in three steps. (i) The number of times by which an individual concentration is greater than (or less than, when the objective is a minimum) the objective is termed an “excursion” and is expressed as follows. When the test value must not exceed the objective, which is the 𝑭𝒂𝒊𝒍𝒆𝒅 𝑻𝒆𝒔𝒕 𝒗𝒂𝒍𝒖¡𝒆¡ case for this study: excursion¡ = ( )-1 --------------(4a) 𝑶𝒃𝒋𝒆𝒄𝒕𝒊𝒗𝒆¡ For the cases in which the test value must not fall below the objective: 𝑶𝒃𝒆𝒄𝒕𝒊𝒗𝒆𝐣 Excursion= ( )-1-------------(4b) 𝑭𝒂𝒊𝒍𝒆𝒅 𝑻𝒆𝒔𝒕 𝑽𝒂𝒍𝒖𝒆𝒋 40 University of Ghana http://ugspace.ug.edu.gh (ii) The collective amount by which individual tests are out of compliance is calculated by summing the excursions of individual tests from their objectives and dividing by the total number of tests (both those meeting objectives and those not meeting objectives). This variable, referred to as the normalized sum of excursions, or nse, is calculated as: ∑𝒏𝒊=𝟏 𝒆𝒙𝒄𝒖𝒓𝒔𝒊𝒐𝒏𝒊nse = ------------(5) # 𝐨𝐟 𝐭𝐞𝐬𝐭𝐬 (ii) F3 is then calculated by an asymptotic function that scales the normalized sum of the excursions from objectives (nse) to yield a range between 0 and 100. 𝒏𝒔𝒆 F3 = ( ) ----------------(6) 𝟎.𝟎𝟏(𝒏𝒔𝒆)+ 𝟎.𝟎𝟏 Once the factors have been obtained, the index itself can be calculated by summing the three factors as if they were vectors. The sum of the squares of each factor is therefore equal to the square of the index. Table 3.2: Sample calculation of WQI for the Kpong dam (Volta River) in Jan. 2011. DATE 1-Jan 8-Jan 15-Jan 22-Jan OBJECTIVE pH 6.9 6.9 6.9 6.9 8.5 Temp 27.6 28 27.8 27.2 29.4 EC 72.1 68.5 64.2 77.2 1500 TDS 31 32 31 31 1000 S.S 3 3 2 4 5 ALK 29 31 30 30 200 TH 22 23 24 22 500 Cl- 1.5 1.5 1.5 2 250 SO4-2 4 3 4 2 250 f- 0.08 0.12 0.16 0.2 1.5 No3-N 1.2 1.4 1.23 1.4 10 NO2- 0.009 0.01 0.008 0.001 0.2 PO4-P 0.3 NH4-N 0.02 0 0.02 0.07 1.5 Mn 0.3 0.31 0.3 0.3 0.1 41 University of Ghana http://ugspace.ug.edu.gh Cr 0.04 0.06 0.05 0.02 0.05 Al 0.001 0.001 0 0.01 0.2 Cu 0.06 0.09 0.16 0.24 1 Fe 0.2 0.12 0.09 0.1 0.3 Zn 0.02 0.05 0.01 0.03 0.12 As 0 0 0 0 0.01 Underlined values do not meet the objective The number of parameters higher than safe limits is 1 (i.e. Mn). The total 1 number of parameters is 21. Therefore: f1= ( ) x100 = 4.762 21 The number of tests not meeting safe limits is 3, and the total number of tests is 3 84. Missing data are excluded (i.e. none). Therefore: f2 = ( ) x100 = 3.571 84 Calculation of excursions, their normalized sum, and f3: 0.3 0.31 0.3 Excursion =( ) − 1 + ( ) − 1 + ( ) − 1= 6.1 0.1 0.1 0.1 2+2.1+2 Nse =( ) = 6.1/84 = 0.0726 84 0.0726 F3 = ( ) =0.0726/0.0107 = 6.785 0.01(0.0726) + 0.01 With the three factors now obtained, the index can be calculated: √4.7622+3.57162+6.7852 CCMEWQI = 100 - ( ) = 9.02578/1.732 = 100- 5.211 = 1.732 94.789 Given the category ranges suggested in the table 3.1 above, the water quality at January 2011 would be rated as GOOD based on this year data. Similar calculations were done for all the months in each year from 2011 to 2016. 42 University of Ghana http://ugspace.ug.edu.gh 3.1.2 Time Series Analysis A time series is a sequence of data values, usually consisting of successive measurements or observations on quantifiable variable(s), made over a time interval (Cochrane, 2005). Time series analysis is therefore a process of using statistical tools to model and explain a time-dependent series of data values in order to extract meaningful statistics and other characteristics of the data (Leonard & Wolfe, 2005). In Time series analysis, observations which are obtained at regular time intervals are correlated using models. Such time series models include MA, AR, ARIMA, GARCH, TARCH, EGARCH, FIGARCH, CGARCH and ARIMA among others but the MA, AR, ARMA, and ARIMA models were used. According to Cochrane (2005), time series can be represented as a set of observations XT, each one being recorded at a specific time T; written as: {X1, X2 ,...Xt} or {XT}, where T = 1, 2,...t --------------eqn.(7) If a time series has a regular pattern i.e. trend, then a value of the series should be a function of past values. If X is the target value that is to be modelled and predicted, and Xt is the value of X at time t, then the goal is to create a model of the form: Xt = f(Xt-1, Xt-2, Xt-3, …, Xt-n) + et -------------------eqn.(8) Where Xt-1 is the value of X for the previous observation, Xt-2 is the value two observations ago, etc., and et represents noise that does not follow a predictable pattern (this is called a random shock). Values of variables occurring prior to the current observation are called lag values. If a time series follows a repeating pattern, then the value of Xt is usually highly correlated with Xt-cycle where cycle is the number of observations in the regular cycle (DTREG, 2010). For example, monthly observations with an annual cycle often can be modeled by: Xt = f(Xt-12)----------------------eqn.(9) 43 University of Ghana http://ugspace.ug.edu.gh 2.1.3 Mann-Kendal test For Trend (Mann, 1945, and Kendall, 1975) The Mann (1945) and Kendall (1975) test is a non-parametric test for identifying trends in time series data. Since water quality data usually exhibit characteristics such as non-normality missing values, values below detection limits, and serial dependence, nonparametric methods are robust and can handle these problems easily. The test compares the relative magnitudes of sample data rather than the data values themselves (Gilbert, 1987). Conventionally, Mann Kendall test statistic S is computed by evaluating the data values as an ordered time series. Thus, S is incremented by 1 if a data value from a later time period is higher than a data value from an earlier time period and vice versa. The net results of such decrements and increments yield the final value of S. A high positive value of S is an indicator of increasing trend and a very low negative value indicates a decreasing trend. Variance of S, VAR (S), Normalized test statistic Z and its associated probability level of significance (95% typically) are respectively calculated. Trend is said to be decreasing if Z is negative and the calculated probability is greater than the level of significance. The trend is said to be increasing if the Z is positive and the calculated probability is greater than the level of significance. If calculated probability is less than the level of significance, there is no trend. In this study, XSTAT package in Excel Spreadsheet was used to process the input time-series data, perform the trend analysis, and report the results (table 4.5). 44 University of Ghana http://ugspace.ug.edu.gh 3.1.4 Autoregressive (AR) Models An autoregressive model is the simplest linear regression ARIMA type models in which a value at the previous time is used to predict a value at the present time. The order (p) of AR model indicates how many previous times that is used to predict the present time.AR models can be analyzed with one of various methods, including standard linear least squares techniques. They also have a straightforward interpretation. For instance, the estimation of forecast for WQI value, ?̂?𝒕 , using only previously recorded history of WQI value is termed AR model of order p and is represented by an equation as follows: ?̂?𝒕 = 𝝁𝟎 + 𝜶𝟏?̂?𝒕−𝟏 + 𝜶𝟐?̂?𝒕−𝟐 + … … + 𝜶𝒑?̂?𝒕−𝒑 + 𝒆𝒕--------------------------eqn. (10) 𝒑 ?̂?𝒕 = 𝝁𝟎 + ∑𝒌=𝟏 𝜶𝒌?̂?𝒕−𝒌 + 𝒆𝒕 Where ?̂?𝒕 is the forecast variable and it is the current value at time t. ?̂?𝑡−1 , ?̂?𝒕−𝟐, ……, ?̂?𝑡−𝑝are the previous values of WQI. 𝛼 ( k= 1, 2,….., 𝜌) are the model parameters to be estimated. The right side of the model, explanatory variable that predict WQI consists of three parts 𝜇0 a constant part,𝑒𝑡, a random error part and the AR summation (Maia et al., 2008). 3.1.5 Moving Average (MA) Models Similarly, it is also possible to regress past values of a series using only past random errors (or residuals) as explanatory variables. This generates an equation for the model called moving average model (MA), represented as follows; 𝒆𝒕 = 𝜽𝟎 − 𝜽𝟏𝒆𝒕−𝟏 + 𝜽𝟐𝒆𝒕−𝟐 + … … + 𝜽𝒒𝒆𝒕−𝒒 + 𝒂𝒕……..eqn.(11) 45 University of Ghana http://ugspace.ug.edu.gh Where 𝒆𝒕 is the random error at time, and the previous random error terms are 𝒆𝒕−𝟏, 𝒉𝒆𝒕−𝟐,.,, 𝒆𝒕−𝒒. The𝜽𝟎,𝜽𝟏, ………𝜽𝒒 are known as its parameters and the 𝒂𝒕 is its random walk. A dependency relationship, according to Da Veiga et al., (2014), tends to be established between successive errors (such as between 𝒆𝒕−𝟏and 𝒆𝒕−𝟐 ) and the equation is called a moving average model (MA). An autoregressive moving average (ARMA) model is created from a finite linear combination of past values of the series [(AR(p) model] and a finite combination of past errors [MA(q) model (Maia et al., 2008). The combination of the two model processes into ARMA (p, q) process is represented as follows; ∑𝒑 ∑𝒒?̂?𝒕 = 𝝁𝟎 + 𝒌=𝟏 𝜶𝒌?̂?𝒕−𝒌 + 𝒆𝒕 + 𝒌=𝟏 𝜽𝒌 𝒆𝒕−𝒌 -------------------eqn. (12) ARMA model can only be used on stationary time series. But in practice, many of the time series are non-stationary because the characteristic of the underlying stochastic process often changes over time (Maia et al., 2008; Pereira, 2004 ). Thus, to extend the use of the ARMA model for nonstationary series, it is necessary to difference the time series. In this situation, the model becomes what is called the autoregressive integrated moving average (ARIMA). The “I” in ARIMA implies that the dataset undergoes differentiation. That is, upon completion of the modelling, the results undergo an integration process to produce final predictions and estimates (Tularam, 2010). Therefore, accordingly, ARIMAhas three model parameters AR(p), I(d) and MA(q) all combined to form ARIMA (p, d, q) model, generally known as the Box Jenkins Methodology, where p = order of autocorrelation, d = order of integration (differencing), also termed as order of differencing operator ( ∆𝑑?̂?𝑡 ) and, q = order of moving averages. 46 University of Ghana http://ugspace.ug.edu.gh The practical problem in modelling is to decide on the most appropriate values for the parameters p, d and q which best specify the ARIMA model (Da Veiga et al., 2014). In order to achieve the best fit, one needs to examine the autocorrelation function (ACF) and partial autocorrelation function (PACF) of the time series dataset. By the use of the graphs of ACF and PACF, the degree of homogeneity of the parameter (d); that is, the number of times that the series needs to be differentiated to produce a stationary series can be determine. Augmented Dicky-Fuller test (ADF) can also be used to statistically test for stationarity during ARIMA modelling (Stock & Watson, 2007). After stationarity, has been established, the graph plots of autocorrelation and partial autocorrelation functions are examined to determine possible specifications of p and q parameters. The model selection criteria have been discussed by Bowerman et al., (2005). Significant spikes outside the confidence band of ACF and PACF graphs are also indicative of the likely order parameters for MA(q) and AR(p) respectively. Before the equation of ARIMA model can be written, the lags and operators need to be understood. To express the underlying differenced ARIMA models, the concept of the backshifts (lags) operators, B, and difference operator , is used. As explained by (Out et al., 2014). These have no mathematical meaning order than to facilitate the writing of models that would otherwise be extremely difficult to express. The backshift is defined as 𝐵𝑚𝑌𝑡 = 𝑌𝑡−𝑚. For instance, 𝐵𝑌𝑡 = 𝑌𝑡−1, 𝐵 12𝑌𝑡 = 𝑌𝑡−12 . The difference operator is presented as ∇ 𝑑 = (1 − 𝐵)𝑑, where the differenced ‘d’ are taken to achieve stationarity in the time series data. Using these notations, the following equations can be deduced. These expressions were adopted from the study by Otu et al., (2014); 𝒑 The general AR (p) = ∑𝒌=𝟏 𝜶𝒌𝒀𝒕−𝒌 + 𝒆𝒕from (1) becomes 47 University of Ghana http://ugspace.ug.edu.gh AR (p) = 𝟏 − 𝜶 𝑩 − 𝜶 𝑩𝟐𝟏 𝟐 − ⋯ 𝜶𝒑𝑩 𝒑 ---------eqn. (13) Where 𝛼(𝐵) is the autoregressive operator of order p The general MA (q) model ∑ 𝒒 𝒌=𝟏 𝜽𝒌 𝒆𝒕−𝒌 + 𝒂𝒕(equation 2) can be presented as; MA (q) = 𝒆𝒕 − 𝜽𝟏𝒆𝒕−𝟏 − 𝜽𝟐𝒆𝒕−𝟐 − ⋯ − 𝜽𝟑𝒆𝒕−𝒒 = 𝜽(𝑩)𝒆𝒕 + 𝝁 ---------eqn.(14) Where (B) is the moving average operator of the q. The general stationary series is written as 𝒁𝒕 = 𝛁 𝒅𝒀 𝒅 𝒕 = (𝟏 − 𝑩 𝒀𝒕) The general ARIMA (p, d, p) model is expressed as (Equation 4) and (Equation 5) becomes (𝟏 − 𝑩𝒅)(𝟏 − 𝜶 𝑩 − 𝜶 𝑩𝟐𝟏 𝟐 − ⋯ − 𝜶𝟐𝑩 𝒑)𝒀𝒕 = (𝟏 − 𝜽𝟏𝑩 − 𝜽 𝟐 𝟐𝑩 − ⋯ − 𝜽𝟑𝑩 𝒒)𝒆𝒕(𝟏 − 𝑩) 𝒅 𝜶 (𝑩)𝒀𝒕 = 𝜽(𝑩)𝒆𝒕-------------eqn.(15) After data analysis , the ARIMA( p, d, q) model has been speficied , the coefficients of the parametres of the model is calculated. Validation of the overall model fit is done using the Ljung -Box test statistics of the residuals of the ARIMA model. 3.2 Best Model Identification and Selection The Ljung-Box statistic would be used to identify the best model. The Ljung-Box statistic, also called the modified Box-Pierce statistic, is a function of the accumulated sample autocorrelations, rj, up to any specified time lag m. As a function of m, it is determined as where n = number of usable data points after any differencing operations. The Ljung-Box test can be defined as follows: H0: The data are independently distributed (i.e. the correlations in the population from which the sample is taken are 0, so that any observed correlations in the data result from randomness of the sampling process). 48 University of Ghana http://ugspace.ug.edu.gh Ha: The data are not independently distributed. The choice of a plausible model depends on its p-value for the modified Box-Pierce if is well above 0.05, indicating “non-significance.” In other words, the bigger the p-value, the better the model. 3.3 Forecast accuracy measures In order to ensure adequacy of the model, the residuals of the predicted WQI need to be examined for forecast accuracy measure and for validation. Forecast accuracy measure is important because it help researchers or managers to evaluate the performance of a given method and or help choose the appropriate method among alternatives (Stevenson, 2009). Mean Absolute Deviation (MAD), also called mean absolute error (MAE), measures the accuracy of fitted time series values. It expresses accuracy in the same units as the data, which helps conceptualize the amount of forecast error. The equation of MAD is presented as follows; 𝒏 𝟏 𝑴𝑨𝑫 = ∑ ⎹ 𝒚 − ?̂? ⎹ 𝒏 𝒕 𝒕 𝒕=𝟏 Where 𝑦𝑡 equals the actual wqi at time t, ?̂?𝑡equals the forecast value of fitted model and n equals the number of observations. The 𝑦𝑡 − ?̂?𝑡 is the forecast error for each forecast period ahead. The problem with MAD, however, is that the value depends on the magnitude of the product being forecast (Rasyid& Adhiutama , 2014). Rasyid and Adhiutama (2014) explained further that if the forecast item is measured in say thousands, the MAD becomes very large. In order to avoid this problem, authors recommended the use of mean absolute percentage error (MAPE) instead. Mean Absolute Percentage Error (MAPE) is calculated as 49 University of Ghana http://ugspace.ug.edu.gh the average of the absolute difference between the forecasted and actual values, expressed as a percentage of the actual values (Rasyid, A.; Adhiutama, 2014); 𝒏 𝟏 𝒚𝒕 − ?̂?𝒕 𝑴𝑨𝑷𝑬 = ∑ ⎹ ⎹ × 𝟏𝟎𝟎 𝒏 𝒚𝒕 𝒕=𝟏 Mean Squared Deviation (MSD) or mean squared error (MSE) can also be used to measure the accuracy of forecast. It is always computed using the same denominator, n, regardless of the model (Ravinder, 2013). MSD is a more sensitive measure of an unusually large forecast error than MAD. 𝒏 𝟏 𝑴𝑺𝑫 = ∑(𝒚𝒕 − ?̂? ) 𝟐 𝒏 𝒕 𝒕−𝟏 There are numerous forecast studies which used one or combination of these accuracy measures to compare the performance of using different forecast method. For instance, Rasyid and Adhiutama, (2014) used MAPE and mean absolute error (MAE) to compare Moving Average, Exponential Smoothing, Holt’s Model, ARIMA, and Transfer Function, during forecasting of demand for PT Telkom’s Internet Service. Medellin et al. (2013) and Maia et al. (2008) also used MAPE and Mean squared deviation (MSD) as forecast accuracy measure in selecting best model. The current study likewise used MAPE and MSE as forecast accuracy measures. 50 University of Ghana http://ugspace.ug.edu.gh CHAPTER FOUR RESULTS AND DISCUSSIONS 4.1 Introduction This section introduces: the comparison of the various physicochemical parameters with existing local and international guidelines; the WQI data; various models and discussion of findings. 4.2 Physicochemical Parameters The yearly mean data of physico-chemical parameters of raw water from the Kpong Dam of the Volta River are presented in Table 4.1. The results are compared with World Health Organization (WHO) guidelines for drinking water (WHO, 2004), background levels for tropical surface waters (Stumm and Morgan, 1981) and the international average for fresh water since communities in the catchment of the Volta River use the raw water for domestic purposes. The levels that are above the guideline values are identified and discussed. 51 University of Ghana http://ugspace.ug.edu.gh Table 4.1: Summary of physico-chemical Parameters of the raw water from Kpong Dam. UNITS WHO, 2016 2015 2014 2013 2012 2011 Min Max Mean 2004 PH 6.94 6.94 6.94 6.96 6.98 6.90 6.90 6.98 6.5- 8.5 Color TCU 1.31 1.42 2.33 1.83 4.17 1.75 1.31 4.17 2.1 4 1.5 Turb NTU 1.45 1.40 3.64 1.75 2.56 1.93 1.40 3.64 2.12 5 Temp 0C 29.4 . 27.64 27.75 27.80 27.70 27.38 27.63 27.38 27.80 27.65 µScm-1Ec 118.10 71.13 74.38 66.62 61.98 70.48 61.98 118.10 77.11 1500 T.D.S mg/L 52.75 43.92 34.33 30.58 31.00 31.50 30.58 52.75 37.35 1000 S.S mg/L 3.75 1.92 3.17 3.08 2.75 2.67 1.92 3.75 2.89 5 T.SS mg/L 56.50 45.83 37.50 33.67 33.75 34.17 33.67 56.50 40.24 1005 Sal mg/L 0.09 0.17 0.17 0.00 0.00 0.00 0.00 0.17 0.07 200 Alk mg/L 32.25 29.58 30.17 28.92 29.50 29.50 28.92 32.25 29.99 200 T. mg/CaCO3/L 500 Hard 22.88 21.85 22.25 21.50 21.00 22.17 21.00 22.88 21.94 Cl- mg/L 4.75 3.29 3.46 2.58 2.54 2.25 2.25 4.75 3.15 250 SO 2- mg/L 4 4.13 2.33 3.17 2.67 2.67 2.75 2.33 4.13 2.95 250 F- mg/L 0.08 0.50 1.20 0.15 0.19 0.11 0.08 1.20 0.37 1.5 NO3- mg/L 10 N 2.39 0.99 1.86 0.20 1.20 1.49 0.20 2.39 1.35 NO - mg/L 2 0.02 0.03 0.01 0.01 0.01 0.01 0.01 0.03 0.01 0.2 PO4- mg/.L 0.3 P 2.71 0.35 0.65 0.35 0.21 0.00 0.00 2.71 0.71 NH4- Mg/L 1.5 N 0.05 0.02 0.03 0.08 0.03 0.07 0.02 0.08 0.05 Ca mg/CaCO3/L 12.50 15.33 12.50 13.00 14.17 14.00 12.50 15.33 13.58 50 Mg mg/CaCO3//L 13.38 6.50 9.75 7.67 6.83 8.17 6.50 13.38 8.72 250 Cr mg/L 0.05 0.05 0.07 0.04 0.04 0.04 0.04 0.07 0.05 0.05 Al mg/L 0.01 0.01 0.03 0.02 0.00 0.00 0.00 0.03 0.01 0.2 52 University of Ghana http://ugspace.ug.edu.gh Cu mg/L 0.08 1.28 0.26 0.14 0.04 0.14 0.04 1.28 0.32 1 Fe mg/L 0.12 0.14 0.19 0.27 0.09 0.14 0.09 0.27 0.16 0.3 Zn mg/L 0.45 0.43 0.33 0.45 0.35 0.38 0.33 0.45 0.40 3 As Mg/L 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.01 53 University of Ghana http://ugspace.ug.edu.gh 4.2 Physical characteristics The mean water temperature and pH of the Kpong Dam from 2011 – 2016 were quite normal and typical of fresh waters. Temperatures ranged from 27.38 to 27.80 0C while pH was between 6.90 and6.98. The surface water was generally weakly acidic. The pH values recorded were within the recommended limit of 6.5- 8.5 for drinking water (WHO, 2004). Throughout the monitoring period, the turbidity levels were within the WHO guideline value of 5 NTU (WHO, 2004) and background levels 0–5 NTU (WRC, 2003) (Table 4.1). The results also indicate that raw water from Kpong Dam was soft and fresh since the total dissolved solids (TDS) of the water samples did not exceed 500 mgL-1 (Kattan, 1995). The TDS recorded was between 30.58mgL-1in 2013 and 52.75mgL-1in 2016 while electrical conductivity (Ec) fluctuated between 61.98µScm-1in 2012 and 118.10µScm.-1indicating low variability in salinity.The electrical conductivity registered does not give cause for concern as it makes the raw water from the dam suitable for direct domestic use. 4.3 Nutrient Anthropogenic enrichment of the Volta River with nutrients and other oxygen consuming materials can change nutrient dynamics, deplete oxygen and change abundance and diversity of aquatic plants and animals as well as the production of undesirable effects and cultural eutrophication. Nutrients in water bodies mainly consist of nitrogen and phosphorus. For nitrogen nutrient, nitrate, nitrite, ammonia and nitrogen-containing organic compounds are the main soluble species. The enrichment of surface water with nutrients has been linked to increased growth of phytoplanktons, aquatic macrophytes and benthic weeds (Bellos et al, 2003). The nutrient levels registered in the raw water from the dam from 2011 – 2016 was relatively low (table 4.1). Nitrate and phosphate were the dominant nutrients ranging between 0.20 and 2.39 mg/L for NO -3 - N and 0.00 – 2.71 mg/L for PO 3 — 4 P. The maximum NH + 4 - N 54 University of Ghana http://ugspace.ug.edu.gh concentration was 0.08mgL-1 and that of NO -2 - N was 0.03mgL -1. Comparably 2016 recorded elevated levels of nutrients especially nitrate and phosphate (fig. 4.1) Nitrate-N. Dams below rivers slow their flow rate and buffering capacity (Dalwin, 2009). This might have resulted in accumulation of high nutrients over the period. The natural background levels of PO 3- 4 - P in riverine waters usually ranges between 0.005 mgL-1 and 0.05mgL-1, (Koukal et al., 2004). PO 3-4 - P contents in raw water samples from the Kpong dam from 2012 to 2016 fell outside this range and were above the recommended limit of < 0.3 mgL1(WHO, 2004). High phosphorus availability is generally believed to be a critical factor in eutrophication. Nutrient enrichment in a water body is accompanied by a high rate of production of plant material, which acts as a major cause of eutrophication problems (Smith, et al, 1999). The variability in nutrient levels in the raw water samples could be largely explained by variability in water discharge, effluent from municipal, domestic, agricultural and small scale agro-industries. Due to variations in the biogeochemical influences, the amount of nutrient varied from one year to another. 55 University of Ghana http://ugspace.ug.edu.gh 3.00 Ammonia-N Nitrate Nitrite Phosphate 2.50 2.00 1.50 1.00 0.50 0.00 2016 2015 2014 2013 2012 2011 Year Fig. 4.1: Levels of nutients in raw water samples from Kpong dam (2011 – 2016) 4. 4. Chemical characteristics Raw water in the Volta River is used for diverse reasons. Consequently, the results of the Kpong Dams water, between 2011 and 2016 revealed considerable variations. Relatively low concentrations of inorganic constituents were found in raw water samples. Generally, the major anions (chloride [Cl-], fluoride (F-) and sulphate [SO 2-4 ]) in the analysed samples were low. The raw water samples exhibited an overall anionic dominance pattern of Cl-> SO 2-4 > F - (fig. 4.2). Results from the study indicated that anionic constituents of the raw water from the dam were extremely low and may not pose any significant effect on the usability of the raw water from the dam for domestic purposes. 56 concentration (mg/L) University of Ghana http://ugspace.ug.edu.gh 5.00 Chloride Sulphate 4.50 Fluoride 4.00 3.50 3.00 2.50 2.00 1.50 1.00 0.50 0.00 2016 2015 2014 2013 2012 2011 Year Fig. 4.2: Anionic content of raw water from Kpong dam (2011 – 2016) 4.5 MAJOR AND TRACE METALS The levels of major metals in the raw water samples were very low because of low level of industrial activities in the area. Ca2+ and Mg2+ were dominant major cations recorded in the raw water with 2016 registering the highest levels (fig. 4.3) probably due to seasonal climatic changes, production schedules of industries around the study area, changes of analytical laboratories or sampling and/or analytical procedures which may cause a shift in the mean or in the variance of the measured values. 57 Concentration (mg/L) University of Ghana http://ugspace.ug.edu.gh 18.00 16.00 14.00 Cr 12.00 Al 10.00 Cu 8.00 Zn Fe 6.00 Mn 4.00 Ca 2.00 Mg 0.00 2016 2015 2014 2013 2012 2011 Year Fig. 4.3: Major Metal content of raw water samples from Kpong Dam (2011 – 2016) The trace metals, notably; chromium, aluminum, iron, manganese, copper, and zinc occurred in trace quantities (fig. 4.4) and at these levels should not pose health problems (WHO, 1995) However, the low concentration of these metals notwithstanding continuous consumption of water containing them over a period of time could result in ailments since they are often cumulative. The ionic observed in the raw water is similar to the dominance characteristics of fresh water observed in a raw water samples across the country by Biney (1987) and Ansa- Asare (1992). 58 Concentration (mg/L) University of Ghana http://ugspace.ug.edu.gh 1.4 1.2 1 0.8 0.6 0.4 0.2 0 2016 2015 2014 2013 2012 2011 Year Mn Cu Fe Al Cr Zn Fig. 4.4: Trace metal content of raw water samples from Kpong Dam (2011 – 2016) 4.5.1 Monthly variations of WQI Analysis of the WQI results presented graphically in figure 4.5 indicates that the water quality of the dam is well within the GOOD averagely water quality rating range of 80 to 94 (CCME, 2001). Though, the water quality scores are erratic each year, the water quality over the period averagely remained in level two (2), i.e., good quality. This water quality rating study clearly shows that, the status of the water body is protected with only a minor degree of threat or impairment; conditions rarely depart from desirable or natural levels. It is observed that the pollution load, though erratic is relatively high in 2016, perhaps due to seasonal climatic changes, production schedules of industries, changes of analytical laboratories or sampling and/or analytical procedures which may cause a shift in the mean or in the variance of the measured value, as well as the different yearly patterns of anthropogenic activities along the Volta river. 59 Concentration (mg/L) University of Ghana http://ugspace.ug.edu.gh 120 100 80 60 40 20 0 2011 2012 2013 2014 2015 2016 Figure 4.5: Monthly variations of WQI in 2011 to 2016, depicting water quality levels. 4.5.2 Annual mean WQI trend from 2011, 2012, 2013, 2014, 2015, 2016 It is again evident in fig 4.6 plot that there is a minimal decline in water quality across the years. The yearly mean WQI decreased from 2011 with 90.714 to 36.932 in 2016 depicting quality degradation. (table 4.3). 60 University of Ghana http://ugspace.ug.edu.gh 100 90 80 70 60 50 40 30 20 10 0 2011 2012 2013 2014 2015 2016 YEAR Figure 4.6: Yearly water quality variations of the Kpong Dam from 2011 to 2016 4.5.3 Comparing the WQI scores across years from 2011 to 2016. The mean WQI (table 4.3) was further determined across the various years from 2011 to 2016. It was brought to light that the overall mean score of WQI from 2011 to 2016 was 71.251, fair (CCME WQI, 65 - 79) meaning that the water is occasionally threatened or impaired; conditions sometimes depart from natural or desirable levels. The highest average WQI scores were recorded in the year 2011 (90.714) followed by the year 2012 (89.215), 2014 (86.123), 2015 (81.810), 2013 (42.710) and 2016 recording the least WQI of 36.932. However, Analysis of Variance (ANOVA) test was used to further determine whether monthly WQI scores (table 4.4) were statistically different across the years from 2011 to 2016. In otherwise, it was used to determine whether the quality of water was statistically different or equal across the various years. Below are the hypotheses of the ANOVA test (table 4.2) at 5% level of significance. 61 WQI University of Ghana http://ugspace.ug.edu.gh Table 4.2: Output of the ANOVA test Water Quality Index Sum of Squares Df Mean Square F P-value Between Groups 512.033 5 102.407 11.320 0.000 Within Groups 597.096 66 9.047 Total 1109.129 71 Decision rule Reject the null hypothesis, H if the p-value of the test is less than 5% level of significance 0 and fail to reject if otherwise. H0 : There is no significant difference in the WQI scores across the years H : There is significant difference in the WQI scores across the years 1 Decision and conclusion The null hypothesis must be rejected since the p-value of the ANOVA test is less than 5% alpha level. Hence, it can be concluded with 95% confidence level that there is significant difference in the WQI scores across the years from 2011 to 2016. In other words, the quality of water differs statistically from one year to another on the average. Table 4.3: Mean Water Quality Index values across years Year Mean Std. Deviation Minimum Maximum 2011 90.714 2.95 30.86 40.02 2012 89.215 3.82 32.00 44.18 2013 42.710 4.67 28.88 42.32 2014 86.123 1.40 30.42 35.28 2015 81.810 2.47 30.42 40.16 2016 36.932 1.07 30.42 33.62 Overall 71.587 3.95 28.88 44.18 62 University of Ghana http://ugspace.ug.edu.gh 4.6 Trend analysis of the Water Quality Index over the years 4.6.1 Time series plot of the monthly Water Quality Index of the Kpong Dam. Figure 4.7 plot shows non-stationary downward trend, suggesting a minimal decline in water quality of the dam. Non-stationarity is observed as the mean is not constant throughout the series and assumes fairly stable scores over the period. The overall mean score of WQI from 2011 to 2016 was 71.587 (table 4.3). The highest average WQI scores were recorded in the year 2011and declined to 2016. This could be attributed to irregular annual rainfall patterns. Years with high rainfall patterns can cause high surface runoffs into waters implying high levels of suspended solids into water bodies and vice versa for years with lower rainfall. Intensity of anthropogenic perturbations within waters is also responsible for the differences observed in quality at different years, e.g., where illegal mining activities are prevalent, the waters become loaded with high total suspended solids which could lower the quality of waters. 63 University of Ghana http://ugspace.ug.edu.gh Time Series Plot of Water Quality Index of the Kpong Dam On Monthly Basis from 2011 to 2016 45.0 42.5 40.0 37.5 35.0 32.5 30.0 Month Jan Jan Jan Jan Jan Jan Year 2011 2012 2013 2014 2015 2016 Figure 4.7: Time Series Plot of WQI of the Kpong Dam Table 4.4: Water Quality Index data of Kpong Dam. WQI (Year) Months 2011 2012 2013 2014 2015 2016 January 93.351 96.027 85.062 86.630 87.390 32.00 February 88.594 91.434 95.965 86.373 91.621 32.00 March 89.547 91.434 93.468 83.504 91.620 33.62 April 95.078 89.954 92.822 82.556 87.610 33.62 May 91.434 91.434 95.078 90.104 87.882 32.00 June 86.464 95.078 91.434 86.784 87.837 30.42 July 91.434 91.434 88.663 83.375 86.393 32.00 August 95.078 89.781 95.014 87.296 91.383 33.62 September 95.078 87.774 86.731 86.096 89.380 32.00 October 87.752 90.090 81.644 82.096 73.707 32.00 November 88.594 86.974 82.700 86.939 76.740 32.00 December 91.434 91.434 74.189 86.464 85.598 30.42 64 Water Quality Index University of Ghana http://ugspace.ug.edu.gh 4.6.2.1 Mann-Kendall Trend Analysis Determining whether there is trend in the series H : There is no trend in the series 0 H : There is a trend in the series 1 Table 4.5: Output of the Mann-Kendall test Mann-Kendall Trend test of WQI Kendall's tau -0.410 P-value <0.00001* Sen's slope (Continuity Correction for ties) -0.105 95% Confidence Interval of the Sen's slope (-0.119,-0.094) *Implies significance at 5% alpha level The null hypothesis ( H0 ) of the Mann-Kendall trend test is rejected at 5% level of significance since the p-value of the test is less than 5% significant level. On the other hand, the Sen’s slope estimator of -0.105 after continuity correction due to presence of ties in the series indicated that the series (WQI) is monotonic decreasing over time since it was negative. Hence, it can be concluded from table 4.5that there is a decreasing trend in the data from 2011 to 2016. This implies the water quality is deteriorating over the period studied. 65 University of Ghana http://ugspace.ug.edu.gh 4.7. Identifying the appropriate trend model of the series In deciding the appropriate trend fitting lines for the data, linear Linear (Y = a + b *T), quadratic (Y = a + b *T + c *T 2), S-curve (Y = a + b * Linear (T )) and Exponential (Y = EXP (a + b/T)) models were separately applied. A comparison of respective accuracy measures (table 4.5) indicates that both the Quadratic and the Exponential Growth-Curve model appear to fit the series as opposed to the other trend models since they have lower prediction errors relatively. However, even though the mean absolute percentage error (MAPE) and the mean square error (MSE) of both selected models were equal, the Quadratic trend model had relatively lower mean square error as compared to the Exponential trend model. Thus, the Quadratic trend model was considered to fit the series and estimate the water quality better than the other trend models. Below arethe results of the model (the fitted Quadratic trend model) that best fitted the data: Y = 90.42 – 0.1523t + 0.000588t2t ----------(1) ^ Where Y is the WQI at any unit time t (measured in months) t bo = 90.42 The model constant 90.32 in the trend model is a measure of the rate of change in Water Quality Index when the effect of time is eliminated or assumed zero. Hence, the WQI score of the Kpong dam on the average can be estimated as 90.32 when the effect of time is eliminated signifying good water quality in accordance with water quality rating system by CCME (2001). This further suggests that the quality of water would still remain good upon eliminating the effects of time. 66 University of Ghana http://ugspace.ug.edu.gh b1 = −0.1523 The value of -0.1523 implies the WQI score decreases by 0.1523 each month on the average from one year to another. b = 0.000588 2 The coefficient 0.000588 in the model is a measure of both the direction and the steepness of curvature of the Quadratic trend. Hence, it indicates further that there is upward curvature in the trends from 2011 to 2016 averagely. Fig.4.5 shows a graphical representation of the fitted Quadratic Trend model. Trend Analysis Plot for Water Quality Index Quadratic Trend Model Yt = 40.32 - 0.1523*t + 0.000588*t**2 45.0 Variable Actual 42.5 Fits Accuracy Measures 40.0 MAPE 6.8694 MAD 2.4465 MSD 10.1939 37.5 35.0 32.5 30.0 1 7 14 21 28 35 42 49 56 63 70 Index Fig.4.8: Fitted Quadratic trend model 67 Water Quality Index University of Ghana http://ugspace.ug.edu.gh Table 4.6: Prediction errors of the fitted trend models Trend Models MAPE MAE MSE Linear 6.96 2.48 10.25 Quadratic 6.87 2.45 10.19 Exponential Growth Curve 6.87 2.45 10.24 S-Curve 7.94 2.91 13.85 4.8 Predicting WQI of the Kpong Dam using ARIMA model To determine what the WQ1 scores are expected to be from the year 2017 to 2018, the appropriate ARIMA model is required. However, the ARIMA model is stationary time series model and thus, the trend in the series as observed must be eliminated by differencing the series. Nonetheless, over-differencing the series introduces or adds unnecessary terms to ARIMA model and thus; smaller lags from 1 to 3 would be considered from which the best differencing order would be selected. Since there were ties in the differenced series at different lags from 1 to 3 (table 4.6), the Mann-Kendall trend test was used to determine whether the trend effect was eliminated after differencing the series. Below were the results of the trend test at lag 1, 2 and 3. Table 4.7: Mann-Kendall trend test of differenced series Lag P-value One 0.996 Two 0.964 68 University of Ghana http://ugspace.ug.edu.gh Three 0.934 It can be concluded from the Mann-Kendall tests that there is no trend in the differenced series from lag 1 to lag 3. In other words, the series is stationary at lag 1, 2 and 3. However, the lag that produces white noise would be selected as the differencing order, d in the ARIMA (p, d, q) model. Fig.4.6 is a combined graphical representation indicating stability about zero in the differenced series at the various lag values. Time Series Plots of Original and Differenced Series from lag 1 to lag 3 1 14 28 42 56 70 Water Quality Index lag1 45 5 40 0 35 -5 30 -10 lag2 lag3 30 10 15 0 0 -15 -10 -30 1 14 28 42 56 70 Index Figure 4.9: Combined Time Series plot of the original data and differenced series 69 University of Ghana http://ugspace.ug.edu.gh 4.8.1 Testing for White Noise in the series The series is said to have white noise if the residuals in the series are independent and identically distributed or has no serial autocorrelation. Thus, the Portmanteau test of white noise would be considered. Below are the hypotheses of the white noise. H : There is white noise or no serial autocorrelation in the series 0 H : There is no white noise or serial autocorrelation in the series 1 Table 4.8: Portmanteau White Noise test output Lag Portmanteau (Q) Statistics P-value One 60.78 0.053 Two 93.44 0.000*** Three 115.49 0.000*** ***p<0.001 **p<0.01 *p<0.05 It can be revealed from the white noise test that the null hypothesis must be retained at 5% alpha level for the first differenced series (lag one). Hence, it can be concluded that there is white noise or no serial autocorrelation in the series at lag 1 as opposed to both lag 2 and 3. Hence, the differencing order, d in the ARIMA (p, d, q) model can be estimated as d=1. 70 University of Ghana http://ugspace.ug.edu.gh 4.8.2 Identifying the Autoregressive (p) and Moving Average (q) orders of the ARIMA model A graphical approach would be used to estimate the appropriate AR (p) and MA (q) orders using the estimated differencing order. The number of significant flags outside the confidence band for both the Partial Autocorrelation function (PACF) and Autocorrelation function (ACF) plots at lag 1 predicts the AR (p) and MA (q) orders respectively. Specifically, the number of significant spikes or flags outside the confidence band of the PACF plot indicates the estimate of the AR order (p); whereas, the significant spikes outside the band of the ACF suggests the possible MA (q) order for the ARIMA (p, d, q) model. However, the best ARIMA model was selected using the model with the smallest AICs and BICs (Akaike and Bayesian Information Criterion). Fig 4.7 indicate PACF and ACF plots of the series at lag1. 71 University of Ghana http://ugspace.ug.edu.gh PACF Plot of First Differenced Series (Lag 1) 1.0 0.8 0.6 0.4 0.2 0.0 -0.2 -0.4 -0.6 -0.8 -1.0 2 4 6 8 10 12 14 16 18 Lag Fig4.10: PACF plot of the WQI series at lag 1 Deduction from Fig. 4.10 Three significant flags or spikes seemed to be outside the confidence band indicating that the possible AR parameter estimates are p=1, p=2 or p=3 72 Partial Autocorrelation University of Ghana http://ugspace.ug.edu.gh ACF Plot of First Differenced Series (Lag 1) 1.0 0.8 0.6 0.4 0.2 0.0 -0.2 -0.4 -0.6 -0.8 -1.0 2 4 6 8 10 12 14 16 18 Lag Figure 4.11: ACF plot of the WQI series at lag 1 One significant flag or spike also appeared outside the confidence band implying that the possible MA order can be estimated as q=1. Below are AICs and BICs are fitted possible ARIMA models, Table 4.9: Fitted ARIMA models ARIMA AIC BIC (1,1,1) 365.76 374.81 (2,1,1) 365.39 374.44 (3,1,1) 365.30 373.61 73 Autocorrelation University of Ghana http://ugspace.ug.edu.gh It can be concluded from the fitted ARIMA models that ARIMA (3, 1, 1) is the most appropriate model required to predict WQI scores of the Kpong dam with 2-year forecasts since it had the least AIC and BIC values relatively. The estimates of the fitted ARIMA model are summarized in Table 4.8. Table 4.10: Fitted ARIMA (3, 1, 1) model Type Coef Std Error Test statistics P-value AR (1) 0.392 0.140 3.48 0.000 AR (2) 0.007 0.126 0.06 0.955 AR (3) 0.168 0.120 1.41 0.158 MA (1) -0.999 0.174 -5.74 0.000 Constant -0.112 0.045 -2.49 0.013 Below is the fitted model equation; yt = −0.112+ 0.392yt−1 + 0.007yt−2 + 0.168yt−3 −0.999t−1 (2) Where yt = yt − yt−1 is the first differenced series (lag 1); yt−1 , yt−2 and yt−3 are the first, second and third lag values of the series respectively.  t−1 is the first lagged residuals or error terms 4.8.3 Residual diagnostics of fitted ARIMA (3, 1, 1) model In order to be satisfied that the fitted model is accurate in predicting the WQIs, some model assumptions must be met. These assumptions include testing whether there is no serial autocorrelation in the residuals of fitted ARIMA model, satisfying the assumption 74 University of Ghana http://ugspace.ug.edu.gh of normality and constant variance or ARCH effects (Autoregressive Conditional Heteroskedasticity) in the residuals. Testing for no serial autocorrelation using the Modified Box-Pierce Test H : The ARIMA model is adequate (No serial autocorrelation) 0 H : The ARIMA model is not adequate (Serial autocorrelation) 1 Table 4.11: Modified Box-Pierce or Ljung Box Test Lag Chi-square Df P-value 12 10.3 7 0.174 24 26.2 19 0.126 36 42.5 31 0.081 48 52.3 43 0.156 The p −value of the Ljung-Box statistics using the first 12 lags is jointly 0.174; whereas, that of the 24, 36 and 48 accumulated lag values are respectively 0.126, 0.081 and 0.156 (lags of 12 is the natural choice with monthly data). This implies that the data does not contribute to autocorrelations of residuals at higher accumulated lags at 5% alpha level (Multiples of 12); which is a desirable result. On the other hand, it further implies that the residuals are independent and identically distributed. Hence, the ARIMA model can be concluded with 95% level of confidence to be adequate. 75 University of Ghana http://ugspace.ug.edu.gh Testing for homoscedasticity or no ARCH effects Table 4.12: Lagrange Multiplier (LM) test of homoscedasticity lags(p)  2 df P-value 1 1.487 1 0.2227 2 1.516 2 0.4687 3 2.310 3 0.5105 4 2.922 4 0.5709 5 4.411 5 0.4918 6 5.645 6 0.4641 7 6.297 7 0.5055 8 7.139 8 0.5218 9 10.315 9 0.3256 10 10.304 10 0.4142 11 12.289 11 0.3423 12 12.728 12 0.3891 13 13.045 13 0.4443 14 12.384 14 0.5755 15 18.141 15 0.2553 H0 : no ARCH effects vs. H1 : ARCH(p) disturbance It can be concluded that there is no ARCH (Autoregressive Conditional Heteroskedasticity) effects or there is constant variance in the residuals at all lags from 1 to 15 at 5% level of significance. Residual Plots of the fitted ARIMA (3, 1, 1) model The Normal Probability and histogram plot on the top and down left check the normality of the residuals (fig. 4.9). Hence, it can be concluded that the error terms are normally 76 University of Ghana http://ugspace.ug.edu.gh distributed since all points appear to lie on the line of best fit. The graph on the top right represents a plot of the error terms against the fitted values. Since there is approximately equal number of points above and below the zero line, it can be concluded that the assumption of error terms having mean being zero is valid. Moreover, that plot of residuals versus observed order suggests that the error terms are independent over time due to signature of constant variance in the error terms. Hence, it can be concluded that all the assumptions of a good ARIMA model were met. Thus, the selected model satisfies all the model assumptions. Since the ARIMA (3, 1, 1) satisfies all the necessary assumptions, it can be inferred that the model provides an adequate representation of the data. Hence, the predictive model would be formulated from the parameter estimates in Table 4.13. Residual Plots for Water Quality Index Normal Probability Plot Versus Fits 99.9 99 5 90 0 50 10 -5 1 0.1 -10 -10 -5 0 5 10 32 34 36 38 40 Residual Fitted Value Histogram Versus Order 16 5 12 0 8 -5 4 0 -10 -8 -6 -4 -2 0 2 4 6 1 5 10 15 20 25 30 35 40 45 50 55 60 65 70 Residual Observation Order Fig. 4.12: Residual Plots of the fitted ARIMA (3, 1, 1) model 77 Percent Frequency Residual Residual University of Ghana http://ugspace.ug.edu.gh Predicting the WQI for two years using the fitted ARIMA (3, 1, 1) model Cross-validation of model The prediction error associated with the two-year forecasts was 6.72% indicating that the prediction was off by approximately 6.72% or the errors associated with forecasts were 6.72%. This further suggests that the model is accurate in predicting WQIs of the Kpong dam. Below is a diagrammatic representation indicating that the two-year forecasts (as shown by the green line) were all significant at 5% alpha level since they all fall within the confidence band as shown by Figure4.10. ARIMA (Water Quality Index) 45 40 35 30 25 20 0 10 20 30 40 50 60 70 80 90 100 Time step Water Quality Index ARIMA (Water Quality Index) Validation Prediction Lower bound (95%) Upper bound (95%) Fig 4.13Forecasting the WQI for two years 78 Water Quality Index University of Ghana http://ugspace.ug.edu.gh Time series forecasting is the process of using a model to generate predictions (forecasts) for future events based on known past events. The goal of forecasting is to project the underlying trend or pattern of the time series into the future as the most likely values for the data (Pentaho, 2013). Table 4.14summarizes the forecasted values of water quality in the Kpong Dam over the period of January 20717 to December 2018 monthly with 95% confidence level using the ARIMA (3, 1, 1) model which has higher p-value of 0.174 (thus, greater than alpha value of 0.05) indicating that it is the best model according to Modified Box-Pierce (Ljung- Box) Chi-Square statistic. The forecast values of WQI (fig. 13) from January 2017 to December 2018 so determined, clearly predict WQI scores of the Kpong dam in the next two years to range between 26 and 50 indicating that the water quality level would gradually be poor or decline with time if left unattended. Hence, it can be recommended that other interventions must be put in place to improve upon the water quality. Table 4.13: Forecasted WQI values of the Kpong Dam for the Next 2 Years Year Forecasts Jan2017 32.685 Feb2017 32.770 Mar2017 32.895 April2017 32.981 May2017 33.032 June2017 33.077 July2017 33.111 Aug2017 33.134 79 University of Ghana http://ugspace.ug.edu.gh Sept2017 33.152 Oct2017 33.165 Nov2017 33.175 Dec2017 33.183 Year Forecasts Jan2018 33.188 Feb2018 32.192 Mar2018 33..197 April2018 32.981 May2018 33.032 June2018 33.200 July2018 33.201 Aug2018 33.134 Sept2018 33.152 Oct2018 33.165 Nov2018 33.203 Dec2018 33.202 80 University of Ghana http://ugspace.ug.edu.gh CHAPTER FIVE CONCLUSIONS & RECOMMENDATIONS This chapter provides conclusions drawn from the study and some recommendations reached based on the conclusions arrived at, to inform policy of the GWCL. 5.0Conclusion The work assessed the trend of raw water quality of the Kpong Dam on the Volta River, developed a time series model and forecast monthly water quality for two years (2017 to 2018). Data from 2011 to 2016, on monthly bases, were collected from the Ghana Water Company Limited. A time series trend analysis on the monthly WQI assumed fairly stable decreasing except, which recorded significantly low water quality probably due to long term accumulations of pollutants. The study identified several ‘candidate’ models which best fitted the data. However, with the use of the Modified Box-Pierce (Ljung-Box) Chi-square statistic criteria of the “largest p-value and minimum Chi-Square value,” the best-fitted ARIMA model selected was ARIMA (3, 1, 1). After the estimation of the parameters of selected models, a series of diagnostic and forecasting accuracy tests were performed. With reference to the findings of the research, it can be concluded that most adequate model for the data was ARIMA (3, 1, 1) and the Dams water quality will worsen the next 2years as the. forecasted values fell in the neighborhood of poor quality range. 81 University of Ghana http://ugspace.ug.edu.gh Also, parameter by parameter assessment revealed that across the years there has been minimal pollution of the of the Volta river especially with respect to nitrate, other chemical species but phosphate levels were very high indicating that the raw water was quite eutrophic. Relatively, low high concentrations of certain inorganic constituents found in raw water samples from the Kpong Dam between 2011 and 2016, could be attributed to increasing urbanization and land use changes. 5.2 Recommendations It is evident that, the dam’s water is gradually becoming polluted. This could result from unprofessional approach to farming and indiscriminate disposal of wastes around and within the enclave of the dam. Therefore, aquaculture and other related human activities should be checked as they can significantly contribute to further pollution of the water. More research works should be carried out on other dams of the Volta river so as to verify variations in comparisons the findings of this work. Where necessary, the Ghana Water Company (GWC) should study more on the entire river from north to south to estimate the changes of contaminant levels so that necessary remedial actions adopted. In order to minimize health exposure risk, residents should be educated on how to treat the raw water before consumption since this could be one of the ways they are exposed to pollutants. Finally, Government through the Ghana Water Company Limited must put in effective and efficient water production and distribution strategies in the municipality since its population is steadily increasing. 82 University of Ghana http://ugspace.ug.edu.gh REFERENCES American Public Health Association (APHA), “Standard Methods for the Examination of Water and Wastewater,” 20th Edition, American Public Health Association (APHA)Washington DC, 1998. A. Lermontov, L. Yokoyama, M. Lermontov and M. A. S. Machado, “River Quality Analysis Using Fuzzy Water Quality Index: Ribeira do Iguape River Watershed, Brazil,” Ecological Indicators, Vol.9, No. 6, 2009, pp. 1188-1197. http://dx.doi.org/10.1016/j.ecolind.2009.02.006 Akuffo, S. B (1998). Pollution Control in a Developing Economy. A study of the situation in Ghana. 2nd edition. 128p. Ghana Universities Press, Accra Amoah, C., (1989). Overview of Microbiological Studies on the Volta Lake, p 33- 39. In C. Gordon, and J.K Amatekpor, (eds). The Sustainable Integrated development of the Volta Basin in Ghana. Volta Basin Research Project, Accra. Amuzu, A T. and Leitmann J. (1991). Environmental Profile of Ghana. Urban management and the environment. Case study.49p Amuzu A. T. (1973). Some aspects of the limnology of Lake Barekese in the Ashanti Region of Ghana. Ghana Water and Sewerage Corporation, Accra. 83 University of Ghana http://ugspace.ug.edu.gh Ansa-Asare O. D. and Asante K. A. (1998). A comparative study of the nutrient status of two reservoirs in southeast Ghana. Lakes and Reservoirs: Res. Mgmt.3: 205–217. Aidoo, E. (2010). Modeling and forecasting inflation rates in Ghana: An application of SARIMA models. Högskolan Dalarna, School of Technology and Business Studies. Atasoy M., Palmquist R.B. and Phaneuf D.J., 2006, Estimating the effects of urban residential development on water quality using micro data, Journal of Environmental Management 79, 399–408 Ameka G. K., de-Graft Johnson K. A. A. and Akuamoah R. K. (2000). A review of the chemical status of the Weija Lake. J. Ghana Sci. Ass. 2: 136–147. Bartram J. and Balance R., 1996, Water Quality Monitoring: A practical guide to the design and Implementation of Fresh Water Quality Studies and Monitoring Programmes, UNEP, Taylor & Frances, London Boateng, S.M. 1995. Review of water utilization in Ghanaian industries. In Proceedings of National Workshop on Water Quality Sustainable Development and Agenda Bugri, S. 1998. Community Water and Sanitation sector News. Mole Priorities Consulting and Trend, Accra, pp. 4. 84 University of Ghana http://ugspace.ug.edu.gh Bao, L.J., Maruya, K.A., Snyder, S.A., Zeng, E.Y., 2012. China's water pollution by persistent organic pollutants. Environ. Pollution. 163, 100e108. http://dx.doi.org/10.1016/j.envpol.2011.12.022. Biney, C. A. (1977). Report on pre-impoundment studies: Chemical Studies of the Lower Volta in the area of the proposed second Volta dam. Accra: TAB, 1977, TAB technical Report 72. Brydsten L., Jansson M and T., and Nilsson A., 1990, Short Communication Element Transport in Regulated and Non-Regulated Rivers in Northern Sweden, Regulated Rivers: Research and Management 5,167-176. Calles O., Nybergy G. L. and Greenberg G. L., 2007 Temporal and Spatial Variation in Quality of Hyporheric water in One Unregulated and Two Regulated Boreal Rivers, River Research and Applications 23: 829-842. Chapman, D. (1992). Water Quality Assessments. Published on behalf of UNESCO, WHO and UNEP. Chang, H., 2008. Spatial analysis of water quality trends in the Han River basin, South Korea. Water Res. 42 (13), 3285e3304. http://dx.doi.org/ 10.1016/j.watres.2008.04.006. 85 University of Ghana http://ugspace.ug.edu.gh Canadian Council of Ministers of the Environment. 2001. Canadian water quality guidelines for the protection of aquatic life: CCME Water Quality Index 1.0, Technical Report. In: Canadian environmental quality guidelines, 1999, Canadian Council of Ministers of the Environment, Winnipeg. Castillejos, A. M. (2006). Management of Time Series Data. Canberra, Australia: University of Canberra. Cochrane, J. H. (2005). Time Series for Macroeconomics and Finance. Chicago: Graduate School of Business, University of Chicago. Clarke, N.A. 1977. Water Supply and pollution control. Water Resources Rese; Institute (WRRI) Handbook, CSIR, Accra, pp. 12-14. Colerangle, P. 1994. Water Quality Management in the Densu and its Drainage Basin. Ghana Water and Sewerage Corporation Manuscript. Cho J. H., Sung K.S. and Ha S.A., 2004, A river water quality management model for optimizing regional wastewater treatment using a genetic algorithm, Journal of Environmental Management 73, 229–242 Chang H., 2008, Spatial analysis of water quality trends in the Han River basin, South Korea, Water Research 42, 3285 – 3304 86 University of Ghana http://ugspace.ug.edu.gh Chapman D., 1996, Water Quality Assessments - A Guide to Use of Biota, Sediments and Water in Environmental Monitoring - Second Edition, UNESCO/WHO/UNEP, Cambridge University Press Cui, F. (2011). ARIMA models for bank failures: Prediction and comparison. University of Nevada, Las Vegas. World Health Organization [WHO] (1997). Guidelines for drinking-water quality: Surveillance and control of community supplies. Geneva: WHO. DeCesare G.J. and Connors S. G., 2002, Cape Cod Watershed Water Quality Assessment Report, Report Number: 96-AC-1, Massachusetts Department of Environmental Protection Department of Water Affairs and Forestry (DWAF), 1996, South African Water Quality Guidelines. Volume 7: Aquatic Ecosystems. Dorner S., Shi J. and Swayne D., 2007, Multi-objective modelling and decision support using a Bayesian network approximation to a non-point source pollution model, Environmental Modelling & Software 22, 211 -222 Debels, P. and Figueroa, R. and Urrutia, R. and Barra, R. and Niell, X. (2005). Evaluation of water quality in the Chillán river (Central Chile) using 87 University of Ghana http://ugspace.ug.edu.gh physiochemical parameters and a modified water quality index, Environmental Monitoring and Assessment, Vol. 110, No. 1301–322. Diggle, P. (1990). Time Series: A Biostatistical Introduction. Oxford University Press. DTREG. (2010). Time Series Analysis. Retrieved March 2015, from DTREG: https://www.dtreg.com/methodology El-Shaarawi, A.H., Esterby, S.R., Kuntz, K.W., 1983. A statistical evaluation of trends in the water quality of the Niagara River. J. Gt. Lakes Res. 9 (2), 234e240. http://dx.doi.org/10.1016/S0380-1330(83)71892-7. Esterby S. R., 1996, Review of Methods for the Detection and Estimation of Trends with Kannel P. R., Lee S. and Lee Y., 2007, Assessment of Spatial– Temporal Patterns of Surface and Ground Water Qualities and Factors Influencing Management Strategy of Groundwater System in An Urban River Corridor of Nepal, Journal of Environmental Management 86, 595–604asis on Water Quality Application, Hydrological Processes, 10, 127-149 Ghana Raw Water Quality Criteria and Guidelines, Adapted Weighted Raw Water Quality Index Application to Ghanaian River Systems,” Water Resources Commission, Accra, 2003. 88 University of Ghana http://ugspace.ug.edu.gh Grimaldi, J. V. and Simonds, R. H. (1989). Safety management. Homewood, Illinois: Irwin Gevrey, M., Comte, L., de Zwart, D., Lek, S., 2010. Modeling the chemical and toxic water status of the Scheldt basin (Belgium), using aquatic invertebrate assemblages and an advanced modeling method. Environ. Pollut. 158 (10), 3209e3218. http://dx.doi.org/10.1016/j.envpol.2010.07.006. GWCL Records, 1999. Routine analysis of raw water at Dalun Office of GWCL, Tamale. Hawkins, M. D., & Olwell, D. H. (1998). Cumulative sum charts and charting for quality improvement. Berlin/Heidelberg: SpringerVerlag. HumphreyF., Darko, Osmund, Ansa-Asare, Adwoa Paitsil, 2013. A Number Description of Ghanaian Water Quality—A Case Study of the Southwestern and Coastal Rivers Systems of Ghana CSIR Water Research Institute, Accra, Ghana; Water Resources Commission, Accra, Ghana. Igusi, E.O. et al (1999) The Nature and Magnitude of Toxic Chemical Element in ABU Dam. Paper presented at the 42nd Annual Conference of NGA held at Ogun State University, Ago-Iwoye, Ogun State, Nigeria. Pp. 16-21. May 1999. 89 University of Ghana http://ugspace.ug.edu.gh J. R. Gat, Planning and Management of a Sustainable and equitable Water Supply under Stress of Water Scarcity and Quality Deterioration and the Constraints of Societal Political Divisions: The case for a Regional Holistic Approach. Department of Environmental Science and Energy Research, the Weizman Institute of Science, Rehovot, Israel, 2004. Kannel, P. R., Lee, S., Kanel, S. R. & Khan, S. P. (2007). Application of water quality indices and dissolved oxygen as indicators for river water classification and urban impact assessment, Environmental Monitoring Assessment, Vol. 132, 93–110. Kendall, M. G. (1975). Rank correlation methods, Charles Griffin, London. Kney A.D. and Brandes D., 2007, A Graphical Screening Method for Assessing Stream Water Quality Using Specific Conductivity and Alkalinity Data, Journal of Environmental Management 82, 519–528 Kowalkowski T., Cukrowska E M., Mkhatshwa B. H. And Buszewski B. L., 2007, Statistical characterisation of water quality in Great Usuthu River(Swaziland), Journal of Environmental Science and Health Part A 42, 1065– 1072 90 University of Ghana http://ugspace.ug.edu.gh Khalil, B., Ouarda, T., and St-Hilare, A (2011). Estimation of water quality characteristics at ungauged sites using artificial neural networks and canonical correlation analysis. Journal of hydrology, 405(3), 277-287. Larmie, S. A. (1993). The Lower Volta River: water quality characteristics and an assessment in relation to domestic and agricultural water supplies. WRRI, Accra Larmie S. A. (1988). An application of QUA2E and QUAL2E-UNCAS Models to a water quality study of the Densu River in Southern Ghana. (MSc. Thesis.) Vrije University, Brussels, Belgium. MMM Consultants, 2002. Report on the Social Aspects on the Usuthu River: Swizaland and Mozambique, DWAF/DFID Strategic Environmental Assessment. NEMA (2006). National environment management authority, State of Environmental Report for Uganda. N. Takova-Arsova, National Action Plan, Ministry of Environmental and Urban Planning. Osei, T. and Marfo K, 1995. Quality Standards and Water Quality Assessment. In Proceedings of the National Workshop on Water Quality Sustainable Development and Agenda 21. pp. 89-94. 2926. 91 University of Ghana http://ugspace.ug.edu.gh Reckhow K. H., 1994, Water Quality Simulation Modeling and Uncertainty Analysis for Risk Assessment and Decision Making, Ecological Modelling 72, 1-20 Rivers-Moore N.A. and Jewitt G.P.W, 2007, Adaptive Management and Water Temperature Variability within a South African River System: What Are the Management Options, Journal of Environmental Management 82, 39–50 Smith, V.H., Tilman, G.D., Nekota, J.C., Eutrophication impacts of Excess nutrient in Freshwater, marine and Terrestrial Ecosystems, Environment Pollution, 100, 1999, 179. Smith, R. L (1974). Ecology and field Biology. Harper and Row, publishers. New York Harper and Row. Publishers. Spellman, F. R. (1996). Stream ecology and self-purification. An introduction for wastewater and water specialists. Technomic Publishing co., Inc. 133p.141 Volta River Authority (1981) Kpong hydroelectric Project. Resettlement programme. Acres International Limited, Canada. Vitousek, P.M., Aber, J., Bayley, S.E., Howarth, R.W., Likens, G.E., Matson, P.A., Schindler, D.W., Schlesinger, W.H., Tilman, G.D., 1997. Human alteration of the global nitrogen cycle: causes and consequencies. Ecol. Appl 7,737- 750. 92 University of Ghana http://ugspace.ug.edu.gh Volta River Authority (1981) Kpong hydroelectric Project. Resettlement programme. Acres International Limited, Canada. Volta River Authority (1998) Annual Report and Accounts. P38-39. Water Research Institute (1975) A survey of the Water Quality of the River Densu. WRRI, Accra. Ghana WHO (1996) Guideline for drinking water quality recommendations, second ed. WHO, Geneva. Water Research Institute (1999). Baseline data and monitoring of pollution on the Volta Lake and Kpong Headpond. 118p WHO., (2008) guidelines for drinking water quality 3rd edition incorporation the first and second agenda vol. 1 recommendations, world health organizations Geneva. Wassswa, S., (1997) determination of levels of pollination within sediments in water channels from industries in selected parts of lake viztura. Msc thesis, makerere University, Uganda. WARM (1998). Ghana’s Water Resources Management: Challenges and Opportunities. Ministry of Works and Housing, Accra, Ghana. Water Resources Commission (WRC) (2000). Water Resources Management Problems Identification, Analysis and Prioritization. Final Report. WRC, Accra, Ghana. 93 University of Ghana http://ugspace.ug.edu.gh Water Resources Commission (WRC) (2003). Ghana Raw Water Criteria and Guidelines, Vol. 1. Domestic Water. CSIR-Water Research Institute, Accra, Ghana. Warm Study Report, 1998. Ghana’s Water Resources: Management Challenges and Opportunities. Ministry of Works and Housing, Accra, pp. 3-48. WHO, 1996. Guidelines for drinking water quality Vol. II, World Health Organisation Report, Geneva, pp. 82-100. Washington State, 1998, Water Quality Assessment Section 305(b) Report World Health Organization (WHO), 2006, Guidelines for safe recreational water environments, WHO Press, Geneva Zhao, Y., Xia, X.H., Yang, Z.F., Wang, F., 2012. Assessment of water quality in Baiyangdian Lake using multivariate statistical techniques. Procedia Environ. Sci. 13, 1213e1226. http://dx.doi.org/10.1016/j.proenv.2012.01.115. 94 University of Ghana http://ugspace.ug.edu.gh APPENDIX The results of the raw water data output from the Kpong Dam PHYSICOCHEMICAL EXAMINATION RESULTS (2016) PARAMETER UNIT JAN FEB MAR APR MAY JUN JULY AUG SEPT OCT NOV. DEC. PH - 7.0 7.0 6.9 6.9 6.9 6.9 7.0 6.9 Color Pt.Co 1.0 1.0 1.0 1.0 2.0 1.5 1.5 1.5 Turbidity mg/l 0.93 1.08 1.12 1.42 2.86 1.44 1.31 1.47 Temp. °C 26.2 27.1 27.2 31.5 28.1 27.6 27.1 26.3 Conductivity µs 108.7 118.2 120.1 128.7 131.0 104.5 106.0 127.6 T.D.S mg/l 51.0 54.0 48.0 60.0 48.0 49.0 52.0 60.0 S.S mg/l 5.0 2.0 2.0 2.0 4.0 0.0 7.0 8.0 T.S mg/l 56.0 56.0 50.0 62.0 52.0 49.0 59.0 68.0 Chloride mg/l 5.5 4.0 3.5 3.5 4.0 5.5 6.0 6.0 Alkalinity mg/l 32.0 31.0 31.0 31.0 32.0 33.0 39.0 29.0 T. Hardness mg/l 23.0 21.0 22.0 22.0 21.0 22.0 25.0 27.0 Ca. Hardness mg/l 9.0 16.0 8.0 4.0 16.0 5.0 21.0 21.0 Mg.Hardness mg/l 14.0 7.0 16.0 18.0 5.0 27.0 14.0 6.0 Total Iron mg/l 0.13 0.10 0.12 0.11 0.16 0.12 0.16 0.08 Salinity % 0.10 0.10 0.10 0.10 0.10 0.0 0.10 0.10 Sulphate mg/l 7.0 4.0 4.0 4.0 5.0 1.0 7.0 1.0 Manganese mg/l 0.20 0.30 0.40 1.6 0.20 0.20 0.20 0.50 Ammonia-N mg/l 0.03 0.01 0.01 0.01 0.00 0.17 0.13 0.07 95 University of Ghana http://ugspace.ug.edu.gh Fluoride mg/l 0.10 0.16 0.09 0.01 0.10 0.06 0.09 0.01 Nitrate mg/l 6.50 4.80 1.10 1.40 2.0 0.80 1.10 1.4 Nitrite mg/l 0.021 0.015 0.007 0.001 0.090 0.005 0.007 0.011 Silica mg/l 17.0 16.4 8.0 15.4 16.5 8.7 9.0 12.0 Chromium mg/l 0.02 0.04 0.04 0.02 0.06 0.18 0.05 0.01 Aluminum mg/l 0.010 0.008 0.010 0.020 0.009 0.04 0.001 0.002 Copper mg/l 0.04 0.06 0.05 0.04 0.06 0.17 0.19 0.02 Arsenic mg/l 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 Phosphate mg/l 0.13 0.21 0.70 0.72 0.70 1.5 6.0 11.7 Zinc mg/l 0.22 0.19 0.17 0.08 0.10 0.04 0.17 0.01 PHYSICOCHEMICAL EXAMINATION RESULTS(2015) PARAMETER UNIT JAN FEB MAR APR MAY JUN JULY AUG SEPT OCT NOV. DEC. PH - 7.0 7. 7.0 7.0 7.0 7.0 7.0 6.9 6.9 7.0 6.7 6.8 Color Pt.Co 2.0 0.0 2.0 2.0 2.0 2.0 1.0 1.0 1.0 1.0 2.0 1.0 Turbidity mg/l 0.99 0.62 2.29 2.02 1.62 2.16 1.34 1.28 1.08 1.04 1.24 1.12 Temp. °C 28.2 28 27.6 27.2 27.4 27.8 28.0 27.6 27.8 27.4 28.5 27.5 Conductivity µs 68.8 68.8 90.0 89.9 88.0 92.0 69.0 75.0 56.0 50.0 54.0 52.0 T.D.S mg/l 32.0 32.0 42.0 42.0 38.0 40.0 62.0 55.0 48.0 44.0 47.0 45.0 S.S mg/l 1.0 1.0 2.0 2.0 1.0 2.0 4.0 3.0 2.0 2.0 1.0 2.0 T.S mg/l 33.0 33.0 44.0 44.0 39.0 42.0 66.0 58.0 50.0 46.0 48.0 47.0 Chloride mg/l 3.00 3.5 2.0 5.5 3.0 3.0 3.5 4.0 3.0 2.5 3.0 3.5 96 University of Ghana http://ugspace.ug.edu.gh Alkalinity mg/l 29.0 28.0 26.0 29.0 30.0 30.0 30.0 31.0 30.0 32.0 31.0 29.0 T. Hardness mg/l 22.2 22.0 21.0 23.0 22.0 21.0 23.0 22.0 22.0 22.0 21.0 21.0 Ca. Hardness mg/l 16.0 14.0 10.0 14.0 17.0 18.0 15.0 16.0 15.0 16.0 18.0 15.0 Mg.Hardness mg/l 6.0 8.0 11.0 9.0 5.0 3.0 8.0 6.0 7.0 6.0 3.0 6.0 Total Iron mg/l 0.19 0.26 0.12 0.20 0.14 0.18 0.11 0.10 0.10 0.10 0.02 0.20 Salinity % 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.00 1.0 Sulphate mg/l 3.0 3.0 1.0 1.01 4.0 4.0 2.0 2.0 3.0 2.0 2.0 1.0 Manganese mg/l 0.20 0.20 0.10 0.10 0.20 0.20 0.3 0.2 0.5 1.60 1.20 0.4 Ammonia-N mg/l 0.01 0.01 0.02 0.01 0.01 0.02 0.02 0.01 0.02 0.01 0.06 0.05 Fluoride mg/l 0.08 0.01 0.01 0.69 0.79 0.84 0.42 0.50 0.64 0.72 0.83 0.42 Nitrate mg/l 1.26 1.02 0.01 0.60 0.24 0.26 0.58 1.12 1.02 0.61 3.0 2.10 Nitrite mg/l 0.034 0.021 0.060 0.015 0.020 0.021 0.020 0.035 0.010 0.049 0.032 0.020 Silica mg/l 16.1 14.2 29.1 14.2 16.3 17.2 18.2 17.5 24.2 42.6 18.3 15.6 Chromium mg/l 0.06 0.06 0.04 0.01 0.06 0.06 0.06 0.02 0.01 0.07 0.06 0.06 Aluminum mg/l 0.010 0.006 0.094 0.001 0.004 0.014 0.003 0.004 0.010 0.006 0.007 0.010 Copper mg/l 0.42 0.05 0.04 0.02 0.15 0.16 0.16 014 0.08 0.04 0.05 0.14 Arsenic mg/l 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 Phosphate mg/l 0.51 0.27 0.28 0.10 0.32 0.34 0.51 0.36 0.52 0.28 0.31 0.42 Zinc mg/l 0.09 0.36 0.06 0.08 0.04 0.04 0.36 0.08 0.14 0.06 0.08 0.16 97 University of Ghana http://ugspace.ug.edu.gh PHYSICOCHEMICAL EXAMINATION RESULTS(2014) PARAMETER UNIT JAN FEB MAR APR MAY JUN JULY AUG SEPT OCT NOV. DEC. PH - 6.9 6.9 6.9 6.9 6.9 6.9 6.9 6.9 7.0 7.0 7.0 7.1 Color Pt.Co 2.0 1.0 1.0 1.0 3.0 6.0 3.0 2.0 1.0 1.0 5.0 2.0 Turbidity mg/l 23.2 1.15 1.42 1.36 1.21 3.12 2.59 2.45 1.24 1.09 3.08 1.79 Temp. °C 27.8 27.5 28.0 27.6 28.4 27.6 27.8 27.8 27.4 28.1 28.4 27.2 Conductivity µs 75.5 72.3 74.5 72.0 74.3 78.4 75.6 78.1 72.4 74.6 72.0 72.9 T.D.S mg/l 32.0 35.0 36.0 37.0 35.0 34.0 35.0 35.0 32.0 33.0 34.0 34.0 S.S mg/l 2.0 3.0 2.0 3.0 3.0 4.0 4.0 3.0 3.0 3.0 4.0 4.0 T.S mg/l 34.0 38.0 38.0 40.0 38.0 38.0 39.0 38.0 35.0 36.0 38.0 38.0 Chloride mg/l 3.5 4.0 3.0 3.5 3.5 3.5 4.0 3.0 3.5 3.5 3.5 3.0 Alkalinity mg/l 31.0 30.0 30.0 29.0 30.0 30.0 30.0 29.0 31.0 31.0 29.0 32.0 T. Hardness mg/l 22.0 21.0 23.0 24.0 24.0 21.0 22.0 22.0 24.0 22.0 21.0 21.0 Ca. Hardness mg/l 14.0 12.0 12.0 13.0 14.0 12.0 13.0 14.0 12.0 12.0 14.0 8.0 Mg.Hardness mg/l 8.0 9.0 11.0 11.0 10.0 9.0 9.0 8.0 12.0 10.0 7.0 13.0 Total Iron mg/l 0.12 0.20 0.20 0.15 0.15 0.14 0.16 0.20 0.30 0.24 0.26 0.15 Salinity % 1.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 Sulphate mg/l 4.0 3.0 4.0 4.0 3.0 4.0 2.0 3.0 2.0 2.0 4.0 3.0 Manganese mg/l 0.40 0.40 0.30 0.20 0.30 0.40 0.30 0.40 0.20 0.50 0.40 0.20 Ammonia-N mg/l 0.01 0.00 0.02 0.04 0.02 0.02 0.03 0.03 0.04 0.04 0.03 0.04 Fluoride mg/l 1.32 1.08 Nitrate mg/l 1.40 2.60 1.82 1.24 1.36 1.54 1.84 1.36 2.20 1.86 1.84 3.30 98 University of Ghana http://ugspace.ug.edu.gh Nitrite mg/l 0.001 0.010 0.020 0.021 0.028 0.014 0.007 0.006 0.012 0.007 0.028 0.007 Silica mg/l 16.4 16.2 15.8 16.0 16.8 16.4 16.5 16.5 16.7 16.4 16.5 16.7 Chromium mg/l 0.02 0.03 0.12 0.06 0.05 0.05 0.12 0.04 0.08 0.06 0.05 0.11 Aluminum mg/l 0.002 0.003 0.010 0.020 0.040 0.050 0.018 0.010 0.020 0.10 0.050 0.018 Copper mg/l 0.03 0.24 0.14 0.16 0.23 0.45 0.38 0.35 0.26 0.28 0.38 0.16 Arsenic mg/l 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 Phosphate mg/l 0.77 0.82 0.75 0.32 0.38 0.74 0.78 0.64 0.66 0.84 0.71 0.38 Zinc mg/l 0.06 0.31 0.09 0.05 0.08 0.10 0.22 0.21 0.14 0.05 0.14 0.09 PHYSICOCHEMICAL EXAMINATION RESULTS(2013) PARAMETER UNIT JAN FEB MAR APR MAY JUN JULY AUG SEPT OCT NOV. DEC. PH - 7.0 7.0 7.0 7.0 6.9 6.9 6.9 6.9 6.9 7.0 7.0 7.0 Color Pt.Co 1.0 1.0 1.0 1.0 3.0 2.0 2.0 2.0 5.0 1.0 2.0 1.0 Turbidity mg/l 1.33 1.44 1.56 1.18 3.01 1.72 1.85 1.42 2.84 1.39 1.77 1.48 Temp. °C 26.7 27.7 26.40 27.1 27.8 28.0 27.6 27.8 28.0 28.0 27.5 29.8 Conductivity µs 66.6 65.7 62.09 63.8 68.2 65.2 72.1 80.0 65.3 62.6 62.8 65.0 T.D.S mg/l 31.0 31.0 30.0 30.0 32.0 31.0 32.0 30.0 31.0 29.0 30.0 30.0 S.S mg/l 2.0 2.00 7.0 4.0 3.0 2.00 3.0 4.0 2.0 1.0 2.0 5.0 T.S mg/l 33.0 33.0 37.0 34.0 35.0 33.0 35.0 34.0 33.0 30.0 32.0 35.0 Chloride mg/l 2.50 2.0 2.5 2.5 3.5 2.5 3.0 2.0 3.0 2.5 2.5 2.5 Alkalinity mg/l 30.0 30.0 29.0 28.0 29.0 28.0 29.0 28.0 29.0 30.0 28.0 29.0 T. Hardness mg/l 20.0 21.0 20.0 20.0 22.0 21.0 22.0 21.0 23.0 28.0 20.0 20.0 99 University of Ghana http://ugspace.ug.edu.gh Ca. Hardness mg/l 12.0 13.0 14.0 14.0 13.0 12.0 13.0 12.0 14.0 15.0 14.0 10.0 Mg.Hardness mg/l 8.0 8.0 6.0 6.0 9.0 9.0 9.0 9.0 9.0 3.0 6.0 10.0 Total Iron mg/l 0.09 0.16 0.52 0.12 0.12 0.11 0.08 0.16 0.80 0.70 0.14 0.20 Salinity % 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.0 0.0 0.0 Sulphate mg/l 1.0 4.0 1.0 4.0 2.0 1.0 3.0 2.0 3.0 3.0 3.0 5.0 Manganese mg/l 0.80 0.10 0.20 0.31 0.25 0.40 0.40 0.22 0.30 0.30 0.60 1.50 Ammonia-N mg/l 0.01 0.02 0.02 0.00 0.05 0.04 0.04 0.13 0.14 0.25 0.20 0.04 Fluoride mg/l 0.19 0.28 0.16 0.19 0.08 0.11 0.24 0.01 0.18 0.08 0.10 0.16 Nitrate mg/l 0.08 0.08 0.08 0.09 0.20 0.30 0.09 0.40 0.20 0.40 0.20 0.26 Nitrite mg/l 0.010 0.006 0.010 0.008 0.009 0.009 0.012 0.007 0.010 0.013 0.008 0.010 Silica mg/l 14.5 16.0 14.2 14.0 15.4 12.0 16.2 11.4 12.4 16.2 14.8 15.2 Chromium mg/l 0.03 0.05 0.04 0.05 0.04 0.02 0.03 0.04 0.03 0.08 0.06 0.04 Aluminum mg/l 0.026 0.008 0.023 0.016 0.010 0.008 0.004 0.026 0.025 0.018 0.011 0.010 Copper mg/l 0.10 0.24 0.03 0.10 0.12 0.05 0.20 0.06 0.10 0.26 0.22 0.19 Arsenic mg/l 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 Phosphate mg/l 0.16 0.42 0.17 0.16 0.71 0.13 0.45 0.19 0.35 0.40 0.53 0.48 Zinc mg/l 0.62 0.01 0.13 0.07 0.06 0.10 0.07 0.09 0.01 0.07 0.05 0.06 PHYSICOCHEMICAL EXAMINATION RESULTS(2012) PARAMETER UNIT JAN FEB MAR APR MAY JUN JULY AUG SEPT OCT NOV. DEC. PH - 7.0 7.0 7.0 7.0 7.0 6.9 6.9 7.0 7.0 7.0 7.0 7.0 Color Pt.Co 1.0 1.0 2.0 2.0 4.0 12.0 10.0 8.0 5.0 2.0 2.0 1.0 100 University of Ghana http://ugspace.ug.edu.gh Turbidity mg/l 1.08 1.14 2.01 2.14 3.82 4.23 4.56 4.02 2.61 1.97 1.86 1.28 Temp. °C 27.8 27.6 27.5 28.1 28.0 26.7 27.4 28.2 26.3 27.1 27.0 26.9 Conductivity µs 72.2 69.5 65.2 7.01 72.0 68.3 65.2 63.4 65.2 66.4 65.1 64.2 T.D.S mg/l 31.0 32.0 31.0 31.0 31.0 31.0 31.0 31.0 31.0 31.0 31.0 30.0 S.S mg/l 2.0 3.0 2.0 1.0 5.0 3.0 2.0 3.0 4.00 3.0 2.0 3.0 T.S mg/l 33.0 35.0 33.0 32.0 36.0 34.0 33.0 34.0 35.0 34.0 33.0 33.0 Chloride mg/l 2.50 2.50 2.50 2.50 2.50 2.50 2.50 3.00 2.50 2.50 2.50 2.50 Alkalinity mg/l 30.0 31.0 29.0 29.0 30.0 30.0 29.0 29.0 29.0 30.0 29.0 29.0 T. Hardness mg/l 22.0 23.0 21.0 20.0 22.0 21.0 20.0 21.0 20.0 22.0 21.0 19.0 Ca. Hardness mg/l 14.0 15.0 14.0 13.0 13.0 15.0 14.0 14.0 13.0 14.0 15.0 16.0 Mg.Hardness mg/l 8.00 8.00 7.00 7.00 9.00 6.00 6.00 7.00 7.00 8.00 6.00 3.00 Total Iron mg/l 0.04 0.10 0.06 0.12 0.08 0.10 0.09 0.08 0.05 0.15 0.08 0.12 Salinity % 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.00 0.0 0.0 0.00 Sulphate mg/l 2.0 2.0 3.0 4.0 3.0 4.0 2.0 2.0 1.0 2.0 3.0 4.0 Manganese mg/l 0.3 0.4 0.4 0.3 0.3 0.2 0.4 0.3 0.40 0.3 0.5 0.40 Ammonia-N mg/l 0.02 0.06 0.04 0.02 0.01 0.02 0.04 0.03 0.02 0.03 0.04 0.01 Fluoride mg/l 0.24 0.16 0.08 0.20 0.20 0.12 0.25 0.14 0.24 0.24 0.10 0.28 Nitrate mg/l 1.40 1.00 0.80 2.60 1.80 1.20 0.08 1.20 0.80 1.50 1.20 0.80 Nitrite mg/l 0.003 0.023 0.018 0.010 0.021 0.008 0.012 0.009 0.008 0.024 0.019 0.008 Silica mg/l 15.4 16.2 15.8 12.40 15.4 13.6 16.2 16.2 13.0 14.5 12.04 12.0 Chromium mg/l 0.02 0.02 0.04 0.08 0.04 0.02 0.04 0.06 0.04 0.02 0.08 0.04 Aluminum mg/l 0.002 0.004 0.002 0.000 0.001 0.002 0.001 0.004 0.002 0.002 0.003 0.001 Copper mg/l 0.03 0.04 0.02 0.06 0.08 0.04 0.0 0.12 0.00 0.04 0.02 0.08 101 University of Ghana http://ugspace.ug.edu.gh Arsenic mg/l 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 Phosphate mg/l 0.24 0.13 0.16 0.04 0.05 0.08 0.24 0.32 0.33 0.45 0.25 0.19 Zinc mg/l 0.06 0.07 0.09 0.31 0.05 0.06 0.08 0.10 0.06 0.12 0.32 0.08 PHYSICOCHEMICAL EXAMINATION RESULTS(2011) PARAMETER UNIT JAN FEB MAR APR MAY JUN JULY AUG SEPT OCT NOV. DEC. PH - 6.9 6.9 6.9 6.9 6.9 6.9 6.9 6.9 6.9 6.9 6.9 6.9 Color Pt.Co 2.0 1.0 1.0 2.0 3.0 3.0 2.0 2.0 1.0 1.0 1.0 2.0 Turbidity mg/l 2.04 1.22 1.15 2.23 3.14 2.98 2.11 2.34 1.18 1.32 1.28 2.16 Temp. °C 27.6 28.0 27.8 27.2 28.1 28.0 27.6 27.3 28.1 27.9 26.8 27.1 Conductivity µs 72.1 68.5 64.2 77.2 78.5 77.6 78.5 64.3 65.8 64.7 69.6 64.8 T.D.S mg/l 31.0 32.0 31.0 31.0 31.0 31.0 32.0 32.0 32.0 32.0 32.0 31.0 S.S mg/l 3.0 3.0 2.0 4.0 3.0 3.0 2.0 3.0 2.0 2.0 3.0 2.0 T.S mg/l 34.0 35.0 33.0 35.0 34.0 34.0 34.0 35.0 34.0 34.0 35.0 33.0 Chloride mg/l 1.50 1.50 1.50 2.00 2.00 2.50 2.50 2.50 2.50 2.50 3.00 3.00 Alkalinity mg/l 29.0 31.0 30.0 30.0 29.0 30.0 30.0 29.0 29.0 29.0 29.0 29.0 T. Hardness mg/l 22.0 23.0 24.0 22.0 21.0 21.0 22.0 21.0 21.0 23.0 24.0 22.0 Ca. Hardness mg/l 15.0 14.0 14.0 15.0 12.0 13.0 14.0 15.0 13.0 14.0 14.0 15.0 Mg.Hardness mg/l 7.0 9.0 10.0 7.0 9.0 8.0 8.0 6.0 8.0 9.0 10.0 7.0 Total Iron mg/l 0.20 0.12 0.09 0.10 0.24 0.19 0.14 0.18 0.16 0.04 0.09 0.12 Salinity % 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Sulphate mg/l 4.0 3.0 4.0 2.0 3.0 1.0 2.0 2.0 3.0 4.0 3.0 2.0 Manganese mg/l 0.30 0.40 0.50 0.20 0.40 0.50 0.40 0.20 0.20 0.60 0.50 0.40 102 University of Ghana http://ugspace.ug.edu.gh Ammonia-N mg/l 0.02 0.00 0.02 0.07 0.04 0.03 0.02 0.30 0.20 0.07 0.04 0.03 Fluoride mg/l 0.08 0.12 0.16 0.20 0.10 0.08 0.10 0.20 0.08 0.12 0.00 0.12 Nitrate mg/l 1.20 1.40 1.23 1.40 1.52 1.30 1.34 1.45 1.40 1.26 1.80 2.60 Nitrite mg/l 0.009 0.010 0.008 0.001 0.005 0.004 0.013 0.002 0.001 0.002 0.004 0.001 Silica mg/l 16.4 15.8 16.0 12.8 14.6 15.2 16.5 16.8 15.3 14.4 14.9 15.4 Chromium mg/l 0.04 0.06 0.05 0.02 0.01 0.08 0.03 0.0.3 0.04 0.02 0.01 0.05 Aluminum mg/l 0.001 0.001 0.000 0.010 0.016 0.006 0.002 0.001 0.002 0.002 0.001 0.012 Copper mg/l 0.06 0.09 0.16 0.24 0.21 0.10 0.05 0.03 0.19 0.02 0.31 0.25 Arsenic mg/l 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 Phosphate Zinc 0.02 0.05 0.01 0.03 0.02 0.03 0.01 0.03 0.01 0.02 0.04 0.12 0.12 103