SEDIMENT YIELD AND BANK EROSION ASSESSMENT OF PRA RIVER BASIN BY John Manyimadin Kusimi (10096985) This thesis is submitted to the University of Ghana, Legon in partial fulfilment of the requirement for the award of PhD Geography & Resource Development degree. June, 2014 University of Ghana http://ugspace.ug.edu.gh i Declaration This thesis entitled “Sediment yield and bank erosion assessment of Pra River Basin” is entirely an original study that I conducted. With the exception of relevant literature and ideas of specific sources which have been duly referenced, the work has not been presented anywhere in part or whole for any award of a degree. ........................................................ John Manyimadin Kusimi (Student) (10096985) Date........................................... ............................................................ Dr. Emmanuel Morgan Attua Dept of Geography & Resource Dev’t University of Ghana, Legon (Principal supervisor) Date........................................... ........................................................ Prof Bruce Banoeng-Yakubo Department of Earth Science University of Ghana, Legon (Supervisor) Date........................................... ............................................................... Dr. Barnabas Amisigo Water Research Institute CSIR, Accra (Supervisor) Date........................................... University of Ghana http://ugspace.ug.edu.gh ii DEDICATION This work is dedicated to my family for their immense support, prayers and contribution. University of Ghana http://ugspace.ug.edu.gh iii Acknowledgement I am grateful to God for bringing me this far in life and academia; to my supervisors: Dr. Emmanuel Morgan Attua, Dr. Barnabas Amisigo and Prof Bruce Banoeng-Yakubo for their support, suggestions, patience and making time off their busy schedules to supervise my work. The University of Ghana partially funded this thesis by awarding me a Faculty Development Grant which was very helpful in undertaking the field work and meeting the costs of field data analysis. I will like to express my gratitude to my brother Jonathan Kusimi, wife Mrs. Bertha Kusimi and Mr. Gabriel Appiah of Water Research Institute, CSIR – Accra for assisting me in my field data collection. I will wish to thank Water Research Institute, CSIR - Accra and Ecological Laboratory (ECOLAB) of the Dept of Geography & Resource Development, University of Ghana for providing field equipment and laboratory space to collect and analyze my field samples. The support of Mr. Prince Owusu of ECOLAB during my field data analyses is also acknowledged. Lastly, I wish to thank Mr. Gerald B. Yiran of the Dept of Geography & Resource Development, University of Ghana for always responding to my call anytime I was in need of assistance with respect to data processing and analysis on the modelling aspect of the thesis. University of Ghana http://ugspace.ug.edu.gh iv List of Tables Table Page Table 2.1: Examples of models for the derivation of R values................................... 34 Table 3.1: Discharge Rating Curves............................................................................ 42 Table 3.2: Annual suspended sediment load and specific suspended sediment yield for the monitored stations in the Pra River Basin................................................ 56 Table 3.3: Parameters for suspended sediment rating curves....................................... 57 Table 4.1: Mean 210 Pb Concentration levels for stream sediment and potential source materials in Birim Basin........................................................................... 64 Table 4.2: Mean 210 Pb Concentration levels for stream sediment and potential source materials in Pra Basin............................................................................... 64 Table 4.3: Mean 210 Pb Concentration levels for stream sediment and potential source materials in Oda Basin............................................................................. 64 Table 4.4: Mean 210 Pb Concentration levels for stream sediment and potential source materials in Offin Basin............................................................................ 64 Table 4.5: Contribution of bank material and surface soil sources to suspended sediment load in the various sub-catchments...................................................... 64 Table 5.1: Annual bank erosion or deposition rates at Anyinam................................. 75 Table 5.2: Annual bank erosion or deposition rates at Amuanda Praso....................... 75 Table 5.3: Annual bank erosion or deposition rates at Akim Oda.............................. 76 Table 5.4: Annual bank erosion or deposition rates at Brenase.................................. 76 Table 5.5: Annual bank erosion or deposition rates at Ejisu....................................... 76 Table5.6: Annual bank erosion or deposition rates at Adiembra................................ 77 University of Ghana http://ugspace.ug.edu.gh v Table 6.1A: K Factor Values of the Soil types............................................................. 86 Table 6.1B: Soil erodibility classification.................................................................... 86 Table 6.2: Land Cover Types and Cover Management (C) factor values.................... 88 Table 6.3: Attributes of Landsat ETM+ 2008............................................................. 88 Table 6.4: Land cover co-efficient values................................................................... 90 Table 6.5: Statistics of the different sediment yield model variables......................... 100 Table 6.6: Estimates of sensitivity analysis of sediment yield for 12 mm event......... 101 Table 6.7: Estimates of sensitivity analysis of sediment yield for 12.5 mm event....... 101 List of Figures Figure Page Fig.1.1: Map of the Pra River Basin........................................................................... 14 Fig.3.1: Sediment Yield Sampling Stations................................................................ 40 Fig.3.2: Daily mean concentration of samples at Anwiankwanta.............................. 44 Fig.3.3: Daily mean concentration of samples at Brenase........................................... 44 Fig.3.4: Daily mean concentration of samples at Akim Oda....................................... 45 Fig.3.5: Daily mean concentration of samples at Adiembra....................................... 45 Fig.3.6: Daily mean concentration of samples at Twifo Praso.................................... 46 Fig.3.7: Daily mean concentration of samples at Assin Praso.................................... 46 Fig.3.8: Daily mean concentration of samples at Sekyere Heman.............................. 47 Fig.3.9: Mean monthly sediment load at Akim Oda................................................... 50 Fig.3.10: Mean monthly sediment load at Brenase..................................................... 50 Fig.3.11: Mean monthly sediment load at Anwiankwanta......................................... 51 University of Ghana http://ugspace.ug.edu.gh vi Fig.3.12: Mean monthly sediment load at Adiembra.................................................. 51 Fig.3.13: Mean monthly sediment load at Twifo Praso.............................................. 52 Fig.3.14: Mean monthly sediment load at Assin Praso............................................... 52 Fig.3.15: Mean monthly sediment load at Sekyere Heman......................................... 53 Fig.3.16: Annual sediment yield. ............................................................................... 56 Fig.5.1: Particle size distribution curves of bank sediments....................................... 80 Fig.5.2: Eroded bank sediments.................................................................................. 82 Fig.6.1: Schematic chart of GIS applications to soil erosion mapping and the derivation of Sediment Delivery Ratio, SDR.............................................................. 84 Fig.6.2: Effective total rainfall erosivity (Re) factor map of 2008 for 12 mm............ 92 Fig.6.3: Effective total rainfall erosivity (Re) factor map of 2008 for 12.5 mm......... 92 Fig.6.4: The soil erodibility (K) map of the basin........................................................ 92 Fig.6.5: The length and slope (LS) map of the basin................................................... 92 Fig.6.6: Cover management factor (C) map derived from satellite image classification....................................................................................................... 93 Fig.6.7: Soil erosion potential map of the basin.......................................................... 93 Fig.6.8: Gross soil erosion map of 12 mm erosive event............................................ 95 Fig.6.9: Gross soil erosion map of 12.5 mm erosive event......................................... 95 Fig.6.10: Sediment delivery ratio map of the basin................................................... 95 Fig.6.11: Sediment yield map of 12 mm erosive event.............................................. 95 Fig.6.12: Sediment yield map of 12.5 mm erosive event........................................... 96 University of Ghana http://ugspace.ug.edu.gh vii List of Plates Plate Page Plata1.1a: Illegal gold mining along the bank of the Birim River at Kyebi................... 8 Plata1.1b: Illegal alluvial gold mining of the river bed and bank of the Ofin River at Adwumain........................................................................................... …. 8 Plate 3.1: A and B show colour of water in the upper courses before galamsey activities. C and D show colour of water at galamsey sites........................................... 48 Plate 3.2: Colour of water near galamsey sites at low flows....................................... 57 Plate 4.1: Evidence of direct sediment and mine waste water entrainment into the rivers. ............................................................................................................... 67 Plate.5.1: Erosion Pin at the bank of the Birim River at Akim Oda............................ 71 Plate.5.2: River bed channels in the upper course of the Pra River at Amuanda Praso – Evidence of bed incision and steep pools........................................ 77 Plate.5.3A: Evidence of cantilever bank failure along the bank of the Oda River at Asaago – Kumasi. ................................................................................... 79 Plate.5.3B: Evidence of cantilever bank failure along the bank of the Birim River at Akim – Oda. .......................................................................................... 79 Plate.6.1: Erosion of a well and house foundations at Odumase................................... 97 Plate.6.2: Eroded plant root at Konongo....................................................................... 97 University of Ghana http://ugspace.ug.edu.gh viii Abstract The Pra River Basin has been engulfed by certain anthropogenic activities particularly illegal small scale mining (popularly called galamsey) and serious concerns have been raised by stakeholders within the basin of the level of pollution due to the release of chemicals and sediments into the water bodies. Fluvial sediment yield data is an essential requirement for informed decision making on water resources development and management. However, information on the sediment load of most rivers is very rare due to the lack of financial resources to regularly undertake sediment yield studies. This study was undertaken to assess the sediment yield levels, sediment sources and bank erosion within the Pra Basin through field data collection and spatial modelling to ascertain stakeholder’s perceptions and suggest remedial measures to the problem. Suspended sediment concentration measurements were undertaken for 9 months in selected stream discharge measuring stations within the basin. Daily mean suspended sediment concentration was determined from which monthly and annual suspended sediment yields were derived. Sediment source tracking was done using a single tracer 210 Pb and the relative contribution of surface and bank sediments to the fluvial sediment transport was determined using the simple mixing model. Lead-210 was analysed using the Atomic Absorption Spectrophotometer (AAS). Bank erosion was assessed using erosion pins. The spatial patterns in soil erosion and sediment yield were modelled using the revised universal soil loss (RUSLE) equation integrating it into Geographic Information System (GIS). Suspended sediment concentration and sediment yield of the Pra Basin were found to be very high resulting in a high annual specific suspended sediment yield. Bank erosion measurement revealed very active bank erosion and deposition within the river channel and bank erosion was University of Ghana http://ugspace.ug.edu.gh ix observed to increase downstream. Sediment source analyses showed that bank material was the dominant sediments which accounted for over 60% of suspended sediment loads. However, predicted sediment yields using the RUSLE were very low as compared to observed data. To promote coordinated development and sustainable management of the resources of the basin, there is the need to resource agencies in charge of regulating natural resource utilization in the basin to control land use activities particularly galamsey to ensure the sustainability of vital ecosystems. The Government also needs to resource financially and improve upon staff strength of the Hydrological Services Departments and the Sediment Unit of the Water Research Institute of CSIR to enable them maintain and monitor critical stations for flow and sediment discharge measurements. Also future research works in sediment yield modelling should consider deploying a model that is capable of modelling both surface and concentrated sediment discharges as this will give a better perspective to a comparative assessment between observed and simulated sediment yield within the Pra Basin. University of Ghana http://ugspace.ug.edu.gh x Table of Contents Contents Page Declaration........................................................................................................................ .. i Dedication............................................................................................................................ ii Acknowledgement............................................................................................................... iii List of Tables........................................................................................................................ iv List of Figures...................................................................................................................... v List of Plates.......................................................................................................................... vii Abstract................................................................................................................................ viii Table of content................................................................................................................... x Chapter One: Background to the Study........................................................................... 1 1.0 Introduction.................................................................................................................... 1 1.1 Statement of the Research Problem................................................................................ 6 1.1.1 Research Questions............................................................................................ 12 1.2 Objectives of the Study................................................................................................... 12 1.3 Hypotheses..................................................................................................................... 13 1.4 Background Information on the Study Area................................................................. 13 1.5 Structure of the thesis...................................................................................................... 17 Chapter Two: Literature Review....................................................................................... 19 2.0 Introduction..................................................................................................................... 19 2.1 Stream Bank Erosion Processes and Measurement......................................................... 19 2.2 Sediment Source Analytical Techniques.......................................................................... 21 2.3 Sediment Yield Measurements........................................................................................ 26 2.3.1 Field Measurements of Sediment Yield.................................................................. 27 University of Ghana http://ugspace.ug.edu.gh xi 2.3.2 Sediment Yield Modelling....................................................................................... 29 2.4 Soil Erosion and Sediment Yield Modelling.................................................................... 31 2.4.1 The Revised Universal Soil Loss Equation (RUSLE) ............................................ 31 2.5 Justification for Research Methodologies........................................................................ 37 Chapter Three: Sediment Concentration and Yield Measurement in the Pra Basin..... 39 3.0 Introduction..................................................................................................................... . 39 3.1 Research Materials and Methods.......................................................................................39 3.2 Results and Discussion..................................................................................................... 43 3.3 Conclusion........................................................................................................................ 59 Chapter Four: Sediment Source Analysis........................................................................... 60 4.0 Introduction...................................................................................................................... 60 4.1 Research Materials and Methods...................................................................................... 61 4.2 Results and Discussion...................................................................................................... 63 4.3 Conclusion......................................................................................................................... 69 Chapter Five: Changes in River Channel........................................................................... 70 5.0 Introduction...................................................................................................................... 70 5.1 Research Materials and Methods.......................................................................................71 5.2 Results and Discussion...................................................................................................... 73 5.3 Conclusion..........................................................................................................................82 Chapter Six: Catchment Scale Soil Loss and Sediment Yield Modelling......................... 83 6.0 Introduction........................................................................................................................83 6.1 Research Materials and Methods........................................................................................83 6.2 Results and Discussion.......................................................................................................91 University of Ghana http://ugspace.ug.edu.gh xii 6.3 Sediment Yield Sensitivity Analysis.................................................................................99 6.3.1 Sensitivity of sediment yield to overestimation and underestimation of model variables....................................................................................................................100 6.4 Conclusion.........................................................................................................................102 Chapter Seven: A Synthesis of Results, Conclusion and Recommendations...................103 7.0: Introduction......................................................................................................................103 7.1 The Nexus between Field Measurements and Modelling of Soil Loss and Sediment Delivery....................................................................................................................103 7.2 Limitations of the study.....................................................................................................108 7.3 Conclusion..........................................................................................................................110 7.4 Recommendations..............................................................................................................112 References...............................................................................................................................115 Appendices..........................................................................................................................132 Appendix A….........................................................................................................................132 Appendix B….........................................................................................................................137 Appendix C….........................................................................................................................141 University of Ghana http://ugspace.ug.edu.gh 1 Chapter One Background to the Study 1.0 Introduction Rivers are natural systems that sculpture and modify the landscape (Nagle, 2000). Increasingly, they are subject to pressure from human activities, and many are so altered or managed that they bear little resemblance to ‘natural rivers’ (Nagle, 2000). Land disturbance has been widely recognized as the main cause of accelerated erosion rates, but there is very little information on past or current sediment delivery rates to the marine environment (UNEP, 1994; cited in Ramos-Scharro´n and MacDonald, 2007). Historically there have been few efforts to remedy this problem (Lugo et al. 1981; in Ramos-Scharro´n and MacDonald, 2007) and this situation can be partly attributed to the lack of data and spatially explicit models to quantify sediment delivery rates and identify sources which will help establish priorities for remediation of accelerated erosion in river basins (Ramos-Scharro´n and MacDonald, 2007). Watershed sediment transport is one of the primary sources of nonpoint source (NPS) pollution for surface waters (Davis and Fox, 2009). Of the nearly 1.1 billion km of impaired rivers and streams in the United States assessed by the Environmental Protection Agency, transport of fine sediments is the most common NPS pollutant (Davis and Fox, 2009). According to Akrasi (2005, 2011) and Akrasi and Ansa-Asare (2008), estimates of suspended sediment yield of most rivers in Ghana including the Pra are low by world standards and this low values are attributed to the forest reserves, secondary forest, cocoa, coffee and oil palm plantations covers of the drainage areas. These types of vegetation protect the soil from the erosive activity of rainfall that is very high in the basin. However, since Ghana’s independence, encroachment of human activities on forest reserves and river University of Ghana http://ugspace.ug.edu.gh 2 corridors has become an acute problem in the country. An estimate of the forest cover as of 2007 is about 16,000 km 2 with an annual rate of depletion of almost 2% (220 to 650km 2 of forest loss per annum) (UNDP Ghana 2004; Dogbevi 2008). Clearing of forests for farming and logging has had serious consequences on surface water hydrology and accelerated the processes of soil erosion, particularly on steeply sloping lands. Not only has the encroachment accounted for biodiversity loss, but also soil erosion and soil fertility depletion, resulting in the sedimentation and pollution of most rivers and reservoirs/dams. In recent times, the activities of illegal small scale miners have compounded the problem of river sedimentation and pollution. Almost all rivers in Ghana that have alluvial gold deposits have been besieged by these illegal mining activities which are entraining thick plumes of sediments and other pollutants into the rivers making the water unwholesome for consumption by local communities. In 2005, the then Minister for Works and Housing Mr. Hackman Owusu-Agyemang stated in Kumasi that, the Owabi Dam was to be desilted at a cost of about 2 million US dollars to save it from collapsing (Ghana News Agency, 2005). Sediments are eroded by two main processes, sheet erosion and channel erosion (Roehl, 1962). Sheet erosion is an upland source of sediments while channel erosion results from gully erosion, valley trenching, stream bed and stream bank erosion (Roehl, 1962). The importance of each source of sediments varies widely in different areas and may vary markedly at different points within a given watershed (Roehl, 1962). Several studies in the field of sedimentation have resulted in the development of relationships involving measurable watershed factors in order to predict sediment yield. Among them are Wischmeier and Smith (1978), Chakraborti (1991), Peng et al., (2008), and Yüksel et al., (2008). Erosion and deposition processes lie at the centre of geomorphological explanation, but progress in understanding these processes has been limited by a lack of appropriate high- resolution monitoring methodologies which permit detection of erosion and deposition University of Ghana http://ugspace.ug.edu.gh 3 dynamics (Lawler, 2005b). Soil erosion leads to surface soil decomposition and sedimentation in dams and channels, so river capacities will be reduced (Solaimani et al. 2008). Thus, the measurement of turbidity and suspended sediment concentration in rivers, estuaries, reservoirs, nearshore zones, etc is attracting increasing attention from hydrologists, limnologists, geomorphologists, freshwater ecologists, engineers, oceanographers, glaciologists, water resource managers and policy makers (Lawler, 2006a). Such measurement programmes can allow inferences to be made about upstream hydro- geomorphological processes, catchment erosion rates, downstream fluvial processes and sedimentation impacts, pollutant and contaminant transfer, and aquatic habitat quality (Lawler, 2006a). For the purposes of studying sediment dynamics, tracers are introduced into a river, estuary, or coastal system, to obtain general information on the characteristics of sediment movement within such environments (Hassan, 2003). Tracers provide a relatively simple means of overcoming technical and sampling problems without the need for a detailed kinematic study of the sedimentary regime (Crickmore et al., 1990; cited by Hassan, 2003). The type of tracer to be used will largely depend on the objectives, environment characteristics, and intended observation period of the experiments (Hassan, 2003). The aim of sediment transport studies at the watershed scale is often to understand the source, fate, and transport of sediment mobilized within a watershed (Davis and Fox, 2009). However, many complex watershed and climatic factors such as rainfall, vegetation, topography, soil type, and human disturbances can affect source, fate, and transport processes of sediment (Davis and Fox, 2009). Due to the large variability of environmental variables over spatial and temporal scales, the source, fate, and transport processes are difficult to predict and model precisely. However researchers and engineers within the environmental field are developing new data- University of Ghana http://ugspace.ug.edu.gh 4 based methods to study these complex, interdependent processes with greater certainty (Davis and Fox, 2009). Accelerated erosion and increased sediment yields resulting from changes in land use are critical environmental problems (Ramos-Scharro´n and MacDonald, 2007). Resource managers and decision makers need spatially explicit tools to help them predict the changes in sediment production and delivery due to unpaved roads and other types of land disturbance (Ramos-Scharro´n and MacDonald, 2007). Land use changes that disturb the natural vegetative cover can greatly increase erosion rates and watershed scale sediment yields, and these increases are critical environmental problems in many parts of the world (Walling, 1997; cited in Ramos-Scharro´n and MacDonald, 2007). River morphology deals with the changes in river form and cross-sectional shape due to deposition and erosion processes. Changes in discharge and sediment load may lead to changes in certain parameters including cross sectional shape, channel shape and type, slope and particle size of bed materials (Brandt, 2000; cited by Solaimani et al., 2008). Channel geometry is the cross sectional form of a stream channel (width, depth and cross sectional area) fashioned over a period of time in response to formative discharges and sediment characteristics (Goude, 2004; cited in Solaimani et al., 2008). Changes in channel morphology are the result of interactions of natural and manmade factors such as fluvial processes (e.g. bank erosion, sediment transport) and developments along the river banks (Alam et al., 2007; Shirley and Lane, 1978). “These changes can create significant issues in the riparian zones due to loss of land, changes in biodiversity, impacts on constructions located on the river banks (e.g., bridge crossings, pipes), river pollution etc” (Shirley and Lane, 1978; Bertrand, 2010). Odgaard (1987; cited in Bertrand, 2010) reported that 40% of the suspended sediments found in streams in the US came from the riverbanks. Bank erosion University of Ghana http://ugspace.ug.edu.gh 5 constitutes an intricate physical, socio-economical, and ecological problem requiring an improved understanding of the key processes governing the phenomenon (Bertrand, 2010). Assessment of the changes to the river course will facilitate the identification and exploitation of river bank over time and the possible future activities that can significantly affect the river water quality (Shirley and Lane, 1978). Studies in morphological changes in river channels are important to protect agricultural land and ecology of the surrounding area to ensure sustainable development along river corridors (Alam et al., 2007). An evaluation of short-term channel response in the affected river reach is required for planning and design purposes, while an understanding of long-term channel response is needed to predict future project operations and maintenance needs (Scott and Jia, 2002). Few studies have been concerned with the process of fluvial erosion (i.e. the removal of bank sediments by the direct action of the flow), and little progress has been made in understanding fluvial bank erosion of cohesive sediments since the contributions of Arulanandan et al (1980; cited in Bertrand, 2010) and Grissinger (1982, cited in Bertrand, 2010). Although major efforts (e.g. Fox et al., 2007; Langendoen et al., 2009; Lawler, 1991; Papanicolaou et al., 2007) have been made to monitor bank retreat in a channel reach, there is still a lack of available techniques in the literature to assess the erosion rate of banks comprised of cohesive materials in frequent intervals (Bertrand, 2010; Darby et al., 2007; Julian and Torres, 2006; Papanicolaou et al., 2007; Pizzuto, 2009; Prosser et al., 2000; Simon et al., 2000). Conservation prioritization is an important consideration for planning of natural resources management, allowing decision makers to implement management strategies that are more sustainable in the long-term (Zhang et al., 2010), particularly on soil erosion management. However, soil erosion management in this context can only be achieved when data on sediment sources, sediment yield or erosion risk hazard maps are available on current erosion status. University of Ghana http://ugspace.ug.edu.gh 6 1.1 Statement of the Research Problem Soil erosion currently poses a threat to the sustainability of agricultural production in many parts of the world, through soil degradation and reduction of soil productivity (Pimental et al., 1995; cited by Blake et al., 2002). Sedimentation impacts many aspects of the environment among which are water quality, water supply, flood control, reservoir lifespan, irrigation, navigation, fishing, tourism, hydro-power generation, river channel morphology and stability etc (e.g. Alam et al., 2007; Hazarika and Honda, 2001; Shirley and Lane, 1978; Peng et al., 2008; Schwartz and Greenbaum, 2009). Also, watershed sediment transport can lead to a number of environmental problems, including decreases in ecological diversity (FISRWG 1998; cited in Davis and Fox, 2009), and decreases in aesthetic properties of rivers and streams (Davis and Fox, 2009). Consequently, sediment transport problems have attracted increasing attention from the public, scientists/researchers, governments and organizations, local and national policy makers (Jain et al., 2010). This has led to an increasing demand for watershed or regional-scale soil erosion models or a quantitative assessment on the extent and magnitude of soil erosion problems so that sound management strategies can be developed for affected zones (Fistikoglu and Harmancioglu, 2002; Jain, 2010). Against this background there is a need for reliable information on rates of soil loss and for an improved understanding of sediment transport and storage in catchments to provide a basis for formulating and implementing improved erosion and sediment control strategies (Blake et al., 2002). Watershed conservation management has been a daunting problem for most governments and institutions in charge of water management in most developing countries as a result of high rate of river sedimentation. Information on the physical resources of the watershed, modelling sediment yield of watersheds and the prioritization of watershed for conservation planning are the basic technological ingredients to arrive at any scientific management decisions of watershed conservation (Chakraborti, 1991). University of Ghana http://ugspace.ug.edu.gh 7 Unfortunately, this information is lacking on most rivers in Ghana including the Pra. This is because rivers situated deep inside tropical rain forests are often poorly gauged and data on sediment load are rare (Restrepo and Kjerfve, 2000). Many of Ghana’s reservoirs have been constructed with inadequate watershed data which later showed that actual sediment yield is much in excess of their design capacities. For example, sedimentation impacts are felt in reservoirs/dams such as Akosombo and Kpong on the Volta River, Weija Lake (Densu River) and Pra catchment reservoirs such as Owabi Dam (Owabi River), Barekese Dam (Ofin River) and Brimsu Dam (Kakum River) (e.g. Akuffo, 2003; Kusimi, 2005; Ghana News Agency, 2005a). Siltation of these dams and reservoirs have reduced their water holding capacities which is negatively affecting the ability of Ghana Water Company Limited (GWCL) and Volta River Authority (VRA) to supply potable water to most towns and the generation of hydro-power to meet the growing industrial and domestic energy demand. According to Kusimi (2005), the Densu river channel, especially the middle and lower courses are seriously experiencing erosion and siltation which is threatening the Weija Dam. The Weija Lake is silting up at a rate of 2%, giving it a lifespan of 50 years and is already 27 years old, meaning, it has 23 years life left (Akuffo, 2003). Although low water inflows into the Volta Lake has been the cause of recent power crisis of 1998, 2002 and 2006 (Brew-Hammond and Francis Kemausuor, 2007; Centre for Policy Analysis, 2007), some analysts believe that sedimentation of the Volta Lake could be other reasons accounting for its reduced power generation capacity. According to Water Resources Commission of Ghana, NGOs (e.g. Friends of Rivers and Water Bodies), District and Municipal Assemblies, chiefs, Assemblymen and residents and other stakeholders within the Pra Basin, the river catchment has come under threat from various activities such as logging, farming, urbanization and illegal mining (e.g. Mensah, 2012). Illegal gold mining along the banks (Plata1.1a) and alluvial mining within the river University of Ghana http://ugspace.ug.edu.gh 8 bed (Plata1.1b) have been the most destructive activities on water quality and sediment injection into the rivers. These illegal mining activities are so rampant and they are found along all the major tributaries of the Pra Basin and virtually in every community from the source to the mouth at Sekyere Heman. Their activities result in the discharge of large Plata1.1a: Illegal gold mining along the bank of the Birim River at Kibi Plata1.1b: Illegal alluvial gold mining of the river bed and bank of the Ofin River at Adwumain University of Ghana http://ugspace.ug.edu.gh 9 volumes of sediments and chemicals such as mercury into the rivers, thus polluting the water. According to GWCL production costs becomes higher when treating this polluted water as a lot of chemicals have to be applied to get the water to the acceptable standard limits for consumption and this is militating against potable water supply. The high sediment levels are also causing the rapid deterioration of filters in treatment plants. For instance, GWCL shut down its treatment plant at Kibi, a town at the source of the Birim River because the river has become too polluted to be treated for domestic use as a result of the activities of illegal or small scale mining activities (Bentil, 2011). Similarly, illegal miners have besieged the raw water intake point at Sekyere Heman threatening the water quality of the 41 million Euro Water Treatment Plant as result of mining the bank materials, thereby, injecting sediments into the river (Asiedu-Addo, 2008). Though the Brimsu Dam was dredged at a cost of 1.7 million dollar in 2005 to curb the annual perennial water shortages, which usually faced Cape Coast Municipality and surrounding districts (Ghana News Agency, 2005b), in May, 2013 the level of water in the Brimsu Dam had drastically fallen below 3 m, which is the minimum operation level, compared to its 8 m maximum operation level, making it difficult to pump water from the facility (Ghana News Agency, 2013). The dam was now producing below one million gallons of water per day instead of its daily production capacity of four million gallons resulting in serious water crisis in the municipality (Ghana News Agency, 2013). Due to the deterioration in water quality resulting from these illegal mining activities, most residents within the basin have to depend on boreholes and sachet water as the alternative source of water supply for domestic uses. In smaller communities where they do not have boreholes and have no access to treated water, they have no option but to still depend on this polluted water which has dire consequences on their health. University of Ghana http://ugspace.ug.edu.gh 10 These illegal mining activities have not only rendered the water of these rivers unwholesome for consumption, but the level of pollution of the rivers have also affected the breeding of fish and crabs which residents within the basin used to depend on for livelihood. Besides, they can no longer swim in the rivers as the level of sediments and chemicals such as mercury and arsenic used in the gold processing react with their skins resulting in skin diseases. Until the recent past when the nation was plunged into power and urban water supply crisis, river sedimentation was not considered as a challenge to water management. Even as of now, what strategic measures have been put in place to arrest the rapid depletion of the vegetative cover of our river corridors which is a strong parameter in controlling soil erosion? Have any efforts been made to assess the sedimentation rates of these rivers and reservoirs that we heavily depend upon so as to develop the right management plans? It is therefore imperative to assess the sediment levels of these rivers for informed river basin management policies to be develop on rivers in the country. Geographic information system (GIS) and remote sensing (RS) have been used to establish a more quantitative, repetitive strategy to classify sediment sources or soil erosion risk maps using models such as the Universal Soil Loss Equation (USLE) and its modified versions (Biswas, 2012; Chakraborti, 1991; Roy, 2009; Wahyunto and Abdurachman, 2010; Bonilla et al., 2010). Though GIS and RS have been increasingly used to model and predict soil erosion in many landscapes and climatic regimes globally, they have received less application in assessing soil erosion in Ghana. Many studies of sediment yield and sediment sources have been conducted in many parts of the world (Collins et al., 1997; Collins et al., 2001; Nichols, 2006; Restrepo et al., 2006), but in Ghana most sediment studies have been concentrated on sediment load/sediment yield (Akrasi, 2005; Amisigo and Akrasi, 1998; Ayibotele and Tuffour-Darko, 1974; Ayibotele and Tuffour-Darko, 1979) with little research on sediment sources. University of Ghana http://ugspace.ug.edu.gh 11 River bank erosion has important implication for channel adjustment and long term channel change, meander development, catchment sediment dynamics, riparian land loss and downstream sedimentation problems (Lawler et al., 1997). However, river bank erosion processes are still poorly understood, and therefore weakly specified in models of river dynamics and sediment transport and only loosely integrated into river management strategies (Wang et al., 1997; in Lawler et al., 1997). Downstream changes in retreat rates for individual basins are also poorly documented (Lawler et al., 1997). Many researchers in fluvial geomorphology have used remote sensing and GIS in studying channel morphology or pattern, bank erosion etc (Alam et al., 2007; Porter and Massong, 2004; Winterbottom and Gilvear, 1997). Simple statistical/regression models for predicting suspended sediment yields of river catchments in Ghana for which no sediment measurements had been undertaken have been developed (Akrasi, 2005; Akrasi, 2011; Akrasi and Ansa-Asare, 2008; Boateng, 2012). However, soil erosion mapping of spatial erosion patterns/vulnerability and the profiling of channel dynamics/bank erosion and deposition are less researched by geomorphologists, engineers, hydrologists etc in Ghana. Given the limited capacity of the conventional methods (surveys) and the need to rapidly map and frequently monitor soil erosion, bank erosion and sediment sources of Ghana’s river basins to make available current data for effective decision making, there is a growing need to employ GIS and related technologies (remote sensing) for speed and accuracy. With the advent of GIS, quantitative analyses necessary for watershed and hydrologic modelling could be carried out rapidly and accurately, allowing for the construction and execution of large- scale geomorphologic investigations (Roy, 2009; Wahyunto and Abdurachman, 2010). To contribute in addressing this research gap, this study aims at assessing river sediment sources and yields from the various sub-basins of the Pra Basin using remote sensing, GIS and erosion models, produce soil erosion/sediment yield maps of the catchment and lastly University of Ghana http://ugspace.ug.edu.gh 12 undertake bank erosion measurement of river channels in the basin to ascertain the contribution of bank sediments to the riparian sediment budget. 1.1.1 Research Questions The above problems that have been discussed raise a number of research questions which were the focus of this study and they are: 1. What is the spatial pattern of sediment yield in the Pra Basin? 2. To what extent can the sediment yield in the Pra Basin be modelled using the revised universal soil loss equation? 3. Where are the major sources of sediments in the catchment? 4. What are the morphological characteristics of channel/bank erosion of the river valleys? 1.2 Objectives of the Study The main objective of this study is to assess the pattern in sediment sources, sediment yield, bank erosion in the Pra River Basin through field data collection and spatial modelling. Specific objectives are: 1. Assess sediment yield of the Pra Basin. 2. Determine the dominant sediment source, whether surface or bank materials in different sub-basins of the Pra catchment. 3. To determine channel stability of selected cross sections by measuring bank erosion. 4. To model sediment yield in the Pra Basin using the revised universal soil loss equation in GIS. University of Ghana http://ugspace.ug.edu.gh 13 1.3 Hypotheses Hypotheses guiding this study are: 1. H 0 : There are no significant variations in sediment yield within drainage basins. H a : There are significant variations in sediment yield within drainage basins. 2. H0: Stream bank erosion is not the main source of sediment transport in the Pra Basin. H0: Stream bank erosion is the main source of sediment transport in the Pra Basin 1.4 Background of the Study Area This study was carried out in the Pra River catchment which is located between latitudes 5º00 ׀ N and 7º15 ׀ N and longitudes 0º03 ׀ W and 2º80 ׀ W (Fig.1.1). It is one of the south-western drainage basins in Ghana. The Pra is the largest of the four principal rivers that drain the area south of the Volta divide and enters the Gulf of Guinea east of Takoradi. Its main tributaries are the Ofin, Oda, Anum and Birim rivers which drain from the Mampong-Kwahu Ranges. The drainage basin area is 23,188 km 2 with a mean annual discharge of 214 m 3 s -1 (Akrasi and Ansa-Asare, 2008). The landscape is generally flat characterised by undulating topography with an average elevation of about 450 m above sea level. The main soil type of the catchment is forest ochrosols which are alkaline. The soils are weathered from the Tarkwaian geological formations composing of sandstones and granitoids and metamorphosed rocks such as phyllites and schists. The soils are clayey and not well leached; hence have the capacity to retain more moisture and are very cohesive (Dickson and Benneh, 1985). The basin falls within the wet semi-equitorial climatic belt which is characterized by two rainfall maxima, the first season is from May – June with the heaviest rainfall in June and the second rainy season is from September – October. Relative humidities are highest (75-80%) University of Ghana http://ugspace.ug.edu.gh 14 during the two rainy seasons and are around 70% during the rest of the year. The basin comes strongly under the influence of the moist south-west monsoons during the rainy Fig.1.1: Map of the Pra River Basin " !. !. !. !. !. !. !. !. !. !. !.!. !. !. !. !. !. !. !. !. !.!. Lake Bosumtwi O fi n O d a P ra Biri m P ra Kibi Kade Ejisu Juaso Ofinso Nkawie Bekwai Obuasi Daboase Akim Oda EffiduaseMankranso Kuntenase Mamponteng New Abirem Assin Fosu Twifo Praso New Edubiase Manso Nkwanta Agona Akrofoso Dunkwa-On-Ofin Konongo- Odumase KUMASI 1°0'0"W 1°0'0"W 2°0'0"W 2°0'0"W 7°0'0"N 7°0'0"N 6°0'0"N 6°0'0"N 5°0'0"N 5°0'0"N Ü Legend !. District Capital " Regional Capital River Railway Road Lake Watershed 25 0 2512.5 Kilometers Inset Ghana Map Study Area Data Source: Ghana Survey Department Composer: J.M. Kusimi - 2011. University of Ghana http://ugspace.ug.edu.gh 15 season with high annual rainfall amounts of between 125 and 200 cm. Dry season is well marked and spans from November to March. Temperatures are high throughout the year with the highest mean monthly temperature being 30° C occurring between March and April and the lowest is about 26° C in August (Dickson and Benneh, 1995). The Pra Basin is covered by the moist-semi deciduous forest vegetation which contains most of the Ghana’s valuable timber trees. The climatic environment of high temperatures and heavy rainfall promote the rapid growth of trees particularly in the rainy season. Trees grow to heights of about 35-45 m or more and are of three layers, the upper, middle and lower layer. A typical semi-deciduous forest consists of trees, lianas, climbers and shrubs/bushes which cover the soil from erosion by rain drops and run-off. However in the dry season, certain species of the upper and middle layers shed their leaves during the long dry spell. Due to the rapid expansion of the cocoa industry in this zone very little of the original forest remains and most of what is left is secondary growth. The size of trees in this belt therefore depends on how long the forest has been allowed to regenerate. Near large settlements where the pressure on land is very great, fallow periods may be as short as 3 years, thus the vegetation is reduced to climbers, shrubs/short woody plants and grass species interspersed by isolated tall trees left on the landscape by farmers (Dickson and Benneh, 1995). The Pra Basin serves as the source of water supply for both industrial and domestic uses for three regional capitals, 41 districts and over one thousand three hundred towns. Of the 41 administrative districts 20 are in the Ashanti Region, 11 in the Eastern Region, 6 in the Central Region, and 4 in the Western Region. The Ofin sub-basin is the main source of water supply to Kumasi and its environs, through two reservoirs, namely Barekese and Owabi dams. The Birim sub-basin is located predominantly in the Eastern Region and has attractive historic places and nine forest reserves. For instance, the Esen Epan Forest Reserve near University of Ghana http://ugspace.ug.edu.gh 16 Akim-Oda is a tourist site with the biggest tree in West Africa at 12 m in circumference and 66.5 m tall. The major tributaries are perennial and constitute all-year-round reliable water source. However, human activities such as mining, logging etc. are having adverse impacts and degrading the surface water resources of the basin (Ghana Survey Department; Water Resources Commission, 2011). The Pra Basin is one of the most extensively and intensively used river basin areas in Ghana in terms of settlement, agriculture, logging and mining due to its rich economic tree species, rich mineral ore deposits and conducive environment for farming. The basin contains most of the large cocoa growing areas in the Eastern, Ashanti, and Central regions. Tree cash crop cultivation other than cocoa is mainly oil palm. Food cropping is increasingly becoming more commercialized especially around the medium and large settlements and along the major road arteries. The basin contains the highest density of settlements (both rural and urban) in Ghana. It has a high concentration of mining activities mainly concerned with gold and other minerals. Several large scale mining companies in the basin include AngloGold Ashanti, Perseus Mining Ltd, Newmont Ghana Gold Ltd etc (Water Resources Commission, 2011). The vegetative cover of the basin is experiencing rapid rate of deforestation due to human activities such as farming, illegal small scale mining and lumbering (saw mills and chain-saw operators) and this is hampering water resources management of the basin. Forest cover outside the reserve areas is negligible and is estimated at less than 2% of the Basin. These forests are heavily logged by both licensed timber firms and illicit loggers. These anthropogenic activities are negatively impacting on the hydrological and geomorphological processes within the basin. The implication of these human activities is sediment transport into the river and channel morphological dynamics. The extensive forest clearance for mining, settlement, and infrastructural development causes considerable loss of soil minerals University of Ghana http://ugspace.ug.edu.gh 17 and subsequent high sediment transport in the Pra and its tributaries silting up channels and dams (Water Resources Commission, 2011). Large scale and small scale mining with disruptive impact on surface cover including soils occur around Obuasi, Kibi, Dunkwa-on- Ofin, Konongo and most other communities within the basin. Moderate to severe sheet and gully erosion poses a threat for flooding within the basin. For instance in July 2011, the Pra and Birim rivers flooded their banks in the upper and middle courses where these anthropogenic activities are intense destroying lives and property (Bentil, 2011). 1.5. Structure of the thesis This thesis is organized into seven chapters. The first chapter gives background information on fluvial geomorphology and the underpinning issues that are of focus by scientists. This chapter also discusses the problem under study, questions that the study seeks to address and outlines the objectives and hypotheses guiding the study. Lastly the chapter also introduces the physical settings of the study area by stating the geographical position of the study area and the underlying physical and socio-economic factors prevailing within the basin. Chapter two encompasses literature review. The chapter reviews existing literature on approaches for measuring/determining sediment yield, sediment sources and bank erosion in fluvial geomorphology and other related fields such as hydrology, glacier geomorphology, and oceanography, among others. The underlying processes involved in these landscape and channel transformation and transportation of landscape materials are also reviewed. The chapter also examines existing models for predicting sediment transport and sediment yield. Chapters three to six are on the research methods and findings of each of the thematic set objectives of the study. The last chapter is chapter seven and in this chapter, a general discussion is done relating all the findings of the study, stating the significance of the results University of Ghana http://ugspace.ug.edu.gh 18 to decision making and policy formulation, drawing conclusions and making recommendations for future studies and for informed decision making. University of Ghana http://ugspace.ug.edu.gh 19 Chapter Two Literature Review 2.0 Introduction The chapter reviews existing literature on approaches for measuring/determining sediment yield, sediment sources and bank erosion in fluvial geomorphology and other related fields such as hydrology, glacier geomorphology, and oceanography among others. The underlying processes involved in these landscape and channel transformation and transportation of landscape materials were also reviewed. The chapter also examines existing models for predicting sediment transport and sediment yield. Informed by this existing literature, the research methods and materials for the study were designed. 2.1 Stream Bank Erosion Processes and Measurements Rivers perform three main sediment related activities; erosion, transportation and deposition. These activities lead to the creation of erosional and depositional landforms (Nagle, 2000). Materials are dislodged through the processes of abrasion and solution. Mass failure of river banks due to the fluvial erosion at the toe banks also entrains materials into river channels. The resultant impact of these fluvial processes is the formation of a myriad of fluvial landforms and a change in the landscape in the long term. The morphology of catchment topography and river channels depends to a large extent, on the interaction between hill slope and channel processes. Consequently, many management theories, measurements and modelling have been developed in order to reduce soil loss from basins and sediment transport to hydrologic drainage networks and to explore the drainage basin structure and evolution (Amore, 2004; Tucker and Bras, 1998). University of Ghana http://ugspace.ug.edu.gh 20 Stream bank erosion rate studies have been undertaken using conventional, manual and field monitoring methods, and these involve erosion pins, cross section resurveys or terrestrial photogrammetry (Lawler, 2005b; Billi, 2008; Bertrand, 2010). The conventional methods present some significant disadvantages as they do not provide continuous measurements of fluvial erosion, but instead provide snapshots of erosion between periods of measurements and do not allow the accurate identification of the critical events triggering fluvial erosion (Bertrand, 2010; Lawler et al., 1997). New techniques available to estimate fluvial erosion rates or the erodibility parameters are the jet testing device (Thoman and Niezgoda, 2008, in Bertrand, 2010), the LIDAR technology and Airborne Laser Scanning (Korpela et al., 2009; Pizzuto et al., 2010; Thoma et al., 2005, in Bertrand, 2010), and the Photo-electronic Erosion Pin (PEEP) (Bertrand, 2010; Lawler, 2005b; Lawler et al., 1997). The advantage of the PEEP over the traditional methods is its ability to continuously monitor bank retreat, which will better pinpoint the exact timing and magnitude of small to moderate erosion events (Bertrand, 2010; Lawler, 2005b; Lawler et al., 1997). For instance, Lawler (2005b) observed that, tidal banks are revealed to be much more dynamic using PEEP measurements than previous conventional monitoring has indicated. Also, by comparing PEEP to low-resolution monitoring of conventional methods, he indicated that, the frequencies of low-resolution monitoring failed to adequately represent the cyclicity, mean, range, variability and trend of bank elevation changes. However, PEEP is saddled with the problem of not being able to record events at night. Also, some of these new methods are expensive for developing countries. Erosion pin and cross sectional re-surveys methods are popular approaches as they are cheap and easily operable (Bertrand, 2010). University of Ghana http://ugspace.ug.edu.gh 21 2.2 Sediment Source Analytical Techniques Determining the sources of sediments, and associated nutrients and contaminants, is an important issue for the management of water quality in river systems (Caitcheon, 1998). Spatial sediment tracing can be achieved by measuring the relative contributions of sediments and associated substances at stream junctions, so that a budget of relative contributions for a whole drainage network can be established (Caitcheon, 1998). In addition to spatial source tracing, dated sediment cores from channel, floodplain deposits and reference inventories sampled at stream junctions can provide valuable information about longer term trends in source contributions and sediment erosion (Caitcheon, 1998; Matisoff and Whiting, 2011; Walling, 2004). Symander and Strunk (1992, in Nagle et al., 2007) described some of the difficulties with the use of suspended sediments to identify source areas. Two of the principal problems are the enrichment of suspended sediments in fines and in organic matter relative to the sources and the transformation of sediment properties within the fluvial system (Nagle et al., 2007). Recently published work on the use of tracers, contend that the use of recent over bank deposits enables the contributions of sediment sources to be identified more reliably and the long-term loading from individual sources to be assessed (Bottrill et al., 2000, in Nagle et al., 2007). Mapping erosional features in a watershed for sediment source tracking could involve using photos, maps, field surveys, erosion pins and troughs. Assembling information on suspended sediment sources has proved difficult using the traditional direct monitoring techniques (e.g. erosion pins and troughs) due to inherent spatial and temporal sampling constraints and the amount of fieldwork involved (Peart and Walling, 1988). Secondly, these traditional techniques do not permit the derivation of the range of soil redistribution rates such as the mean erosion rate University of Ghana http://ugspace.ug.edu.gh 22 for the eroding areas, the mean deposition rate for the depositional areas, the net soil loss from the field and the sediment delivery ratio (Walling, 2004). In response to the problems associated with traditional monitoring and measurement techniques, the fingerprinting approach using tracers has been increasingly employed as a means of establishing the relative importance of potential catchment sediment sources and soil erosion rates (Collins et al., 1997; Collins et al., 2001; Mukundan et al., 2009; Nagle et al., 2007). Sediment fingerprinting relies upon the premise that the physical and chemical properties of suspended sediment will reflect its source (Collins et al., 1997; Collins et al., 2001). Another fundamental condition that must be met by any tracer is that the tracer substance(s) must remain unaltered within the spatial and temporal limits in which the tracing method is being applied, (Caitcheon, 1998), adsorption of the tracer to soil is strong and quick; variation in adsorption to various sizes or mineralogic/organic constituents is minor or can be accounted for (Matisoff and Whiting, 2011). The most commonly used tracers include; radionuclides ( 137 Cs, 210 Pb) (Nagle et al. 2007; Walling, 2004), and cosmogenic isotopes ( 7 Be) (Schuller et al., 2006; Walling, 2004). 137 Cs has a core depth profile not greater than 20 cm and is anthropogenically introduced into the environment through fallouts of nuclear activities of bombs of 1950s and 1960s and reactors such as the 1986 Chernobyl disaster. Lead-210 (t1/2 = 22years, with core depth of about 10 cm) and beryllium – 7 (t1/2 = 53days and depth profile not exceeding 3 cm) are natural fallouts from the atmosphere (Blake et al., 2002; Matisoff and Whiting, 2011), hence are ubiquitous on the earth’s surface and thus are suitable as environmental radionuclides tracers everywhere. 137 Cs, 7 Be, and 210 Pb are each suitable as particle tracers because they have a global distribution, adsorb efficiently to soil particles and thus move with soil, and are relatively easily measured (Matisoff University of Ghana http://ugspace.ug.edu.gh 23 and Whiting, 2011). Also these environmental radionuclide tracer methods are effective for distinguishing between surface-derived sediments from sheet and shallow rill erosion and sediments from gullies and stream channel walls, because channel and gully walls deeper than their profile depths of between 3 – 30 cm usually contain little or no traces of the radionuclides (Nagle et al., 2007; Matisoff and Whiting, 2011). However, 137 Cs fallout from the atmosphere is currently near zero or near non-detectable limits (Matisoff and Whiting, 2011) in most parts of the world except northern Europe where release from the Chernobyl explosion was higher. Also due to the short half life of 7 Be, it is only suitable for simulating erosion of small catchments. Other tracers include; stable isotopes (C-13, N-15) sediment carbon and nitrogen (Juracek and Ziegler, 2009; Mukandaun et al., 2009), phosphorous (Wallbrink et al. 2003), clay mineralogy (Youngberg and Klingeman, 1971; Glasmann, 1997, in Nagle et al., 2007), magnetic susceptibility (Blake et al., 2006; Caitcheon, 1998; Gruszowski et al., 2003), and heavy metals (Juracek and Ziegler, 2009) etc. In view of the multiplicity of elements being analyzed in these approaches, they are often very costly hence not employed in most studies. The fingerprinting technique could either be a simple mixing model using only one diagnostic tracer or a composite mixing model involving a combination of two or more tracers. Some researchers (Walling et al., 1993, Yu and Oldfield, 1989; Molinaroli et al., 1991) have, however, argued that no single diagnostic property of sediment can reliably distinguish different sources, because individual tracers may be subject to physical and chemical changes, which limit their use, e.g. particle size sorting, organic matter selectivity, and geochemical transformation during fluvial erosion and transportation (Collins et al., 1997). Also, individual properties may be unreliable because of spurious source-sediment matches (Yu and Oldfield, 1989; Molinaroli et al., 1991; Walling et al., 1993). For example, suspended sediment tracer values may resemble University of Ghana http://ugspace.ug.edu.gh 24 those of a particular source, but could also result from various combinations of other sources (Collins et al., 1997). However, Nagle et al., (2007) have effectively used a simple mixing model of 137 Cs to distinguish between sediment from surface sources and gullies. Lead-210 (Brigham et al., 2001; Motha et al., 2002) and Berryllium-7 (Schuller et al., 2006) have also been used singularly in sediment source tracing. Sediment source tracking has also been performed successfully in a subset of intermittent streams using amorphous to crystalline ratios of iron to estimate the fraction of sediment coming from in-stream vs. landscape sources (Schoonover et al., 2007). Parsons and Wainwright (1993) and Caitcheon (1998) used mineral magnetic properties of sediments in tracing sediment sources. The use of mixing and unmixing models of multi tracers involving multivariate statistics to identify the relative contributions of surface erosion from different land use types and channel erosion to suspended sediment load in river basins sources has also been demonstrated and examined in detail in the literature (Collins et al., 1997; Collins et al., 2001; Gruszowski et al., 2003). For instance Walling (2004) successfully used environmental radionuclides caesium-137, lead-210, beryllium-7 to trace sediments mobilization and delivery in river floodplains in Devon, UK. Though very effective and efficient in characterizing sediment sources and the redistribution rates of sediments, the multi-tracer model approach is very complex, mathematical and very demanding in terms of data, field work as well as laboratory logistics with huge financial burden. Issues of sample size and the range of tracer properties that are measured have also being raised (Gruszowski et al., 2003). GIS and remote sensing techniques and other modelling programmes such as USLE, WEPP, DR3M, WFPB, GWLF (Arekhi, 2008; Jain et al., 2010, Nangia et al., 2010; Mongkolsawat, 1994; Roy, 2009; Wahyunto and Abdurachman, 2010) have recently been employed for University of Ghana http://ugspace.ug.edu.gh 25 sediment source tracking in catchments. Models are appealing because they are cheaper to use. They are also most effective for source analyses where the models have been applied and calibrated. Models are used by reviewing existing data, and consulting with those who are familiar with basin conditions (Gellis, 2010). However with very large basins some models are associated with large errors. Secondly, some are very complex and require very complex input data sets such as the hydrology, rainfall interception by vegetation (e.g. throughfall, stemflow etc), water balance, plant growth and residue decomposition of catchments which make them unsuitable in the developing countries where such explicit data is difficult to generate (e.g. WEPP and EUROSEM). The latest modelling approaches to overcome the limitations of the empirical USLE concentrate on physically based erosion models such as SWAT (Arnold et al. 1998; Gassman et al., 2007, in Silva et al., 2010), Water Erosion Prediction Project (WEPP) (Amore et al., 2004) and EUROSEM (Morgan et al., 1998). These are physically based models with the basic processes incorporated in them so that they can simulate the individual components of the entire erosion process by solving the corresponding equations; and so it is argued that they tend to have a wider range of applicability (Amore et al., 2004; Silva et al., 2010). Such models are also generally better in terms of their capability to assess both the spatial and temporal variability of the natural erosion processes (erosion and deposition) (Amore et al. 2004). Though these models are transferable to other watersheds, they require huge amounts of input data and many calibration parameters, complex laboratory analyses or hard and expensive field data collection, which are commonly out of reach of many developing countries (Renschler et al., 1999; Silva et al., 2010). University of Ghana http://ugspace.ug.edu.gh 26 2.3 Sediment Yield Measurements Sediment yield is the amount of sediment load passing the outlet of a catchment, that is the sediment load normalized for the drainage area and is the net result of erosion and deposition processes within a basin (Jain and Das, 2010; Restrepo and Syvitski, 2006; Verstraeten and Poesen, 2001). These materials are of three different kinds; dissolved load (consisting of soluble materials carried as chemical ions); suspended load (containing clay and silt held up by the turbulent flow), and bed load which includes larger particles moved by saltation, rolling and sliding (Nagle, 2000). However, on the basis of transport processes, measurement principles, and morphological/sedimentary associations, fluvial sediments are often classified into two; bed materials and washed materials. Bed material is often conflated with bed load (makes up the bed and lower banks of the river channel), and wash material moves in suspension and travels out of the reach once entrained (measured load), but the two classifications are not congruent (Church, 2006). Depending on discharge/flow rate, the medium grain sand particles (saltation materials) could become unmeasured sediments if the popular Helley-Smith sampler is used. Generally, the size of material that moves as bed or suspended load in the flow depends upon the power and turbulence of the flow (Gomez and Church, 1989). In relatively deep streams of high flows, with bed material that consists of fine sand, the suspended bed material may be 90% or more of the total sediment discharge. However, in shallow streams with medium to coarse sand beds, the unmeasured-sediment discharge may represent 50% or more of the total sediment discharge (Andrews, 1981). Various approaches have been developed to determine river sediment yield and these include field measurements and modelling (physical and empirical). Suspended load and bedload are measured or estimated University of Ghana http://ugspace.ug.edu.gh 27 separately because the physical processes that govern their rates of transport are dependent on different factors. The sum of suspended load and bedload is the total sediment load (Edwards and Glysson, 1999). 2.3.1 Field Measurements of Sediment Yield Measuring and estimating suspended sediment yields in rivers has long been subject to confusion and uncertainty (Thomas, 1985) because various methods have been developed to measure suspended sediment yield and they include the measurement of suspended sediment load and water discharge (Akrasi, 2005; Khanchoul et al., 2010; Kusimi, 2008), measuring total eroded soil and deposited sediments in small catchments (Verstraeten and Poesen, 2001), and measuring sediment volumes in ponds, lakes or reservoirs (Nichols, 2006; Verstraeten and Poesen, 2001). For the measurement of sediment volumes in ponds, lakes and reservoirs, radiometric techniques using 210 Pb or 137 Cs as tracer elements can be employed to reconstruct sediment budgets over a period of time (Foster et al., 1990; Govers et al., 1996; Walling, 1990 cited in Verstraeten and Poesen, 2001). The ideal situation to estimate the suspended sediment yield of rivers would be to measure suspended sediment concentration and water discharge continuously and use the product function as an estimate of suspended sediment discharge (Lane et al., 1997; Thomas, 1985). Obtaining continuous records of concentration however is practically impossible owing to cost, number of samples and sampling frequency among others (Edwards and Glysson, 1999; Thomas, 1985). Alternative to these issues of cost, remoteness of sites, and technical difficulties is to measure water discharge continuously and to take occasional discrete water samples either University of Ghana http://ugspace.ug.edu.gh 28 manually or using automatic sampling equipment for gravimetric analysis of suspended sediment concentration (Thomas, 1985). The use of sediment-discharge rating curve to estimate sediment yield is however problematic because suspended-sediment concentrations are known to be variable for a given discharge because stormflow hydrographs usually, but not always, are characterized by higher suspended- sediment concentrations during the rising limb than the falling limb. Further, the timing between storm events also influences availability of fine-grained sediment from the watershed, such that an initial stormflow following relatively dry conditions usually has a greater suspended-sediment concentration than subsequent flows of similar magnitude (Edwards and Glysson, 1999). Consequently, statistical considerations show that the sediment load of a river is likely to be underestimated when concentrations are estimated from water discharge using least squares regression of log-transformed variables (Asselman, 2000; Cohn et al., 1992; Ferguson, 1986; Jansson, 1985; Singh and Durgunoglu, 1989). Also regardless of how the samples are collected, there remain questions of when the measurements of concentration should be made, how they should be used to estimate the total yield, how close can samples be spaced in time and still be meaningful among others (Edwards and Glysson, 1999; Thomas, 1985). According to Edwards and Glysson, (1999), spatio-temporal variations in sediment transport can be captured by collecting depth/point-integrated suspended-sediment samples that define the mean discharge-weighted concentration in the sample vertical and collecting sufficient verticals to define the mean discharge weighted concentration in the cross section. Verticals of samples could either be taken using Equal-Discharge-Increment or Equal-Width-Increment Methods. Though both methods have their advantages and disadvantages, if properly used, they yield similar results. University of Ghana http://ugspace.ug.edu.gh 29 Also the biases in the estimation of sediment loads by rating curve due to using log- transformed estimates can be significantly reduced by using nonlinear regression. Further improvements can be achieved by identifying seasonalities and breaks in slopes of the rating curves and taking samples at the right time. Finally, the underestimation caused by using average daily flows with the rating curve can be eliminated by using sub-daily flow data, if available (Singh and Durgunoglu, 1989). The many methods that have been developed for collecting data and estimating suspended sediment yields indicate each method is characterised by one limitation or the other, thus the method that one employs is subject to availability of equipment, cost, convenience, the kind of results that is sought among others. 2.3.2 Sediment Yield Modelling Sediment yield and surface erosion at a watershed or regional scale are at present also modelled using empirical models such as the universal soil loss equation (USLE), modified universal soil loss equation (MUSLE) or the revised universal soil loss equations (RUSLE and RUSLE 2) which are sometimes integrated into GIS (Arekhi, 2008; Jain et al., 2010; Mongkolsawat, 1994; Nangia et al., 2010; Roy, 2009; Wahyunto and Abdurachman, 2010). Others include WEPP, SWAT, EUROSEM etc (Amore et al., 2004; Fistikoglu and Harmancioglu, 2002). Universal Soil Loss Equation (USLE) and its revised versions (RUSLE and RUSLE2) have been used to model soil erosion sometimes integrating the parameters of the model in GIS to produce soil erosion risk maps (Bonilla et al., 2010; Fistikoglu and Harmancioglu, 2002; Kouli et al., 2008; Silva 2004; Silva et al., 2010; Stone and Hilborn, 2000; Roy, 2009; Wahyauto, 2010). The USLE/RUSLE though empirical without spatial dimensions like some of their counterparts, they are however simple and their parametric data can easily be transformed into GIS input data University of Ghana http://ugspace.ug.edu.gh 30 formats for spatial analysis of phenomena. Secondly, their data requirements are not too complex and literature has shown that they have been successfully used to estimate soil erosion of catchments and farmlands (Arekhi, 2008; Fistikoglu and Harmancioglu, 2002; Jain and Kothyari, 2000; Jain and Das, 2010; Jain et al., 2010; Silva et al., 2010). However, USLE/RUSLE only predicts the amount of soil loss that results from sheet or rill erosion on a single slope and does not account for soil losses from gully, wind or tillage erosion and deposition (Stone and Hilborn, 2000). Consequently, RUSLE was subsequently revised to include more advanced scientific interface as RUSLE 2 which permits the segmentation of topography to determine sediment yield and deposition based on changes in topography, soil and management systems along a flow path (Dabney et al., 2011; Stone and Hilborn, 2000; USDA- ARS, 2008a, 2008b). Despite these advancements, RUSLE2 cannot predict erosion within concentrated flow channels (gulley erosion) (Dabney et al., 2011; USDA-ARS, 2008a). Also, RUSLE2 is too rigorous with a lot of data requirements some of which do not have spatial domain (e.g. surface roughness, sediment detachment, sediment transport and sediment deposition models) and as such cannot be analysed with GIS. Though it has a higher predictability of soil erosion than USLE/RUSLE, it is only applicable to small farm plots hence inappropriate for large river catchments like the Pra River Basin. The other models (e.g. WEPP, SWAT, EUROSEM, MIKE SHE, ANSWERS, CREAMS etc) are applicable at catchment scale, event based, continuous models spatially and temporally distributed (i.e. 2D) (e.g. Amore et al., 2004; Fistikoglu and Harmancioglu, 2002). As already discussed above, these models require substantial data inputs, many calibration parameters, characterized by complex laboratory analyses or hard and expensive field data collection (Silva University of Ghana http://ugspace.ug.edu.gh 31 et al., 2010) and thus are inappropriate to apply in developing countries where physical data on river basins are non-existent or very limited. 2.4 Soil Erosion and Sediment Yield Modelling In order to improve water quality and restore impaired watersheds, managers need to make decisions using data that they are able to gather (Nangia, 2010). Data collection can be expensive, tedious and time consuming, so in such situations using modelling approach makes sense (Nangia, 2010) particularly in the developing countries where institutions and organizations charged with the monitoring and collection of data are ill-equipped in terms of personnel, materials as well as financial. Models for sediment yield provide invaluable information when applied to those areas lacking data, for predicting future impacts of agricultural activities, land use, stream stabilization and sediment storage in reservoirs (Khanchoul et al., 2010). The revised universal soil loss equation has been widely used with very good results, is applicable in a GIS environment and so can provide a spatial distribution of erosion and soil loss, requires a small and simple input data set, is relatively easy to use and is suitable for the study, thus its structure and operating system are being discussed for clarity. 2.4.1 The Revised Universal Soil Loss Equation (RUSLE) A first attempt to evaluate and quantify the human impact on soil erosion such as land use changes or new cultivation techniques was the universal soil loss equation (USLE) developed by Wischmeier and Smith (1958) (Renschler et al., 1999). Since then, modified versions such as the revised universal soil loss equation (RUSLE) (Renard et al., 1991), modified universal soil loss equation (MUSLE) (Blaszczynski, 2003) have been widely used as tools for predicting soil University of Ghana http://ugspace.ug.edu.gh 32 erosion in many parts of the world (Renschler et al., 1999). USLE, RUSLE and MUSLE were defined by the following equation whose variables were annually estimated (Wischmeier and Smith 1978): A =R x K x LS x C x P ……………… (2.1), where A represents the potential long term average annual soil loss in tonnes per hectare per year; R is the rainfall erosivity factor (MJ mm ha/h/yr); K is the soil erodibility factor (t h MJ -1 mm -1 ); LS is the slope length-gradient factor, C is the crop/vegetation and management factor, and P is the support practice factor and these other parameters are all dimensionless. The rainfall erosivity index (r factor) measures the potential ability of rain to cause soil erosion. The R-factor is the sum of individual storm erosivity index (EI)-values for a year averaged over long time periods (> 20 years) to accommodate apparent cyclical rainfall patterns. Storm erosivity (EI = r) is the product of a storm’s total energy (E) and its maximum intensity (I) within a stipulated time frame which in standard time could be 5 min, 10 min, 15 min, 30 min, 60 min etc (Renard and Freimund 1994; USDA-ARS, 2008a). An erosive event is a rainfall event with more than 12.50 mm of total rainfall accumulation or with at least 6 mm of rainfall accumulation in 15 minutes (Petkovšek and Mikoš, 2004; USDA-ARS, 2008a). The three most common mathematical models used to relate the kinetic energy (KE) to rainfall intensity are the logarithmic model, exponential model, and the Hudson (1965) model (Shamshad et al., 2008). The original method for calculating R values for a storm event requires pluviograph records (Table 2.1) (Wischmeier and Smith, 1978). This kind of information is difficult to obtain in many parts of the world because its processing is time-consuming, labour-intensive, and costly (Eltaif et al., 2010; Silva, 2004). There is also a lack of pluviograph record data in most developing countries. The successful use of the EI30 in certain temperate regions (such as its University of Ghana http://ugspace.ug.edu.gh 33 place of origin) does not guarantee its success as an index of soil loss in regions with substantially different climates, such as the tropics, because erosive rainfall in the tropics differs from that in other regions mainly in intensity and frequency characteristics (El-Swaify et al., 1982). A number of indices which relate the erosivity of a rainstorm and its associated runoff to soil loss prediction have been established. The most commonly used indices include the Founier Index (Fournier, 1960), the modified Founier Index (MFI) (Arnoldus, 1980), Hudson’s KE > 25 Index (Hudson, 1965), Lal’s AIm, Index (Lal, 1976) and the erosivity index (EI30) of the Universal Soil Loss Equation (Renard and Freimund, 1994; Vrieling et al., 2010; Wischmeier and Smith, 1978). In the US, the R-factor can be evaluated or calculated using methodological guides of Wischmeier and Smith (1978). In contrast, lack of rainfall intensity data in some countries makes the calculation of R-factor, especially in the developing world, very difficult. Although the most accurate estimate of R-values can only be obtained from longterm rainfall intensity data as calculated by Wischmeier and Smith (1978), several works suggest that monthly precipitation data can give reasonable estimates of R-values for many regions throughout the world (Renard and Freimund, 1994; Shamshad et al., 2010; Vrieling et al., 2010). Several authors have derived exponential/regression relationships between R and average monthly precipitation of the wettest month. Also, R is well correlated with the modified Fournier index (Table 2.1). Although very useful for estimating relative erosion hazards, mean annual rainfall is not directly correlated with soil loss in the tropics due to differences in rainfall characteristics and vegetative cover, hence the use of rainfall amounts in deriving R factor is over simplification of soil erosion process (El-Swaify et al., 1982). For example, erosivity index varies from 50 to 650 in the United States, 60 to 300 in Tunisia, 50 to 300 in Morocco, 60 to 340 in the South of University of Ghana http://ugspace.ug.edu.gh 34 France, and 500 to1400 in Ivory Coast (Roose, 1977). The works of these scientists indicate that an erosivity variable may be dependent on the climatic conditions of the geographical area, scale of the study area, and type of measurement and that there is no single universal variable better than the others (Irvem et al., 2007, Renard and Freimund 1994; Stocking, 1987). Table 2.1: Examples of models for the derivation of R values (EI30)i is EI30 for storm i, j no. of storms in an N year period, ρ the average rainfall (mm) of the month with the highest rainfall, P the average annual rainfall (mm), F the modified Fournier Index, H rainfall amount, Pi monthly precipitation of month i , Pei effective monthly precipitation of month i, Pt annual rainfall amount (mm), I30 maximum intensity in 30 minutes. (Angulo-Martínez and Beguería 2009; Diodato and Bellocchi, 2007). Soil erodibility factor (K) is the average soil loss in tonnes/hectare for a particular soil in cultivation under continuous fallow on a plot with a slope length of 22.13 m and slope steepness of 9%. K is a measure of the susceptibility of soil particles to detachment and transport by rainfall and runoff (Stone and Hilborn, 2000). The LS factor represents a ratio of soil loss under given conditions to that at a site with the “standard” slope steepness of 9% and slope length of 22.13 m. The steeper and longer the slope, the higher is the risk for erosion (Stone and Hilborn, 2000). Slope length has been defined as the distance from the point of origin of overland flow to Number Equation Author 1 R = , Wischmeier and Smith, 1978 2 R = 0.04830P 1.610 Renard and Freimund 1994 3 R = 587.8 - 1.219P + 0.004105P 2 Renard and Freimund 1994 4 R = 0.07397F 1.847 Renard and Freimund 1994 5 R = 95.77 - 6.081F + 0.4770F 2 Renard and Freimund 1994 6 FI or R = ρ 2 /P Vrieling et al, 2010 7 MFI or R = De Luis et al (2010) 8 MFIe or R = Smithen and Schulze (1982). 9 R = 0.577H - 5.766 Roose, 1977 10 R = (0.0158H x I30) - 1.2 Roose, 1977 11 EI30 = 227MFI 0.548 Shamshad et al, 2010 University of Ghana http://ugspace.ug.edu.gh 35 the point where either the slope gradient decreases enough that deposition begins or the flow is concentrated in a defined channel (Wischmeier and Smith, 1978). Traditionally estimates for L were obtained from field measurements, however, this method requires high financial and human resources which are not feasible at regional scale (Van Remortel et al., 2001; Hickey et al., 1994). With advancement in computers, there are several algorithms that have been packaged into GIS softwares that make the derivation of slope length and steepness simpler. These include using either unit stream power (Moore and Burch, 1986) or upslope area (Desmet and Govers, 1996) as a surrogate for slope length, the neighbourhood method and the best fit plane method (Srinivasan and Engel, 1991), grid-based methods and the maximum downhill slope angle (Hickey et al., 1994; Hickey, 2000; Van Remortel et al., 2001), network triangulation (Cowen, 1993) techniques (cited by Hickey, 2000), quadratic surface, maximum slope and maximum downhill slope techniques (Van Remortel et al., 2001). Among these, the one that is best suited for integration with GIS is the theoretical relationship proposed by Moore and Burch (1986) and Moore and Wilson (1992) based on unit stream power theory, given as (Jain and Das, 2010; Jain and Kothyari, 2009; Silva et al., 2010): ……………… (2.2) where As is the specific area (=A/b), defined as the upslope contributing area for an overland cell (A) per unit width normal to the flow direction (b); is the slope gradient in degrees. The cover management factor C is used to determine the relative effectiveness of soil and crop management systems in terms of preventing soil loss. The C factor is a ratio comparing the soil loss from land under a specific crop and management system to the corresponding loss from continuously fallow and tilled land (Stone and Hilborn, 2000). The support practice factor P reflects the effects of practices that will reduce the amount and rate of the water runoff and thus University of Ghana http://ugspace.ug.edu.gh 36 reduce the amount of erosion. The P factor represents the ratio of soil loss by a support practice to that of straight-row farming up and down the slope. The most commonly used supporting cropland practices are cross slope cultivation, contour farming and strip cropping (Stone and Hilborn, 2000). The genesis of RUSLE is USLE (Universal Soil Loss Equation) introduced in 1958 by W. H. Wischmeier and D. D. Smith with the U.S. Department of Agriculture (USDA), Agricultural Research Service (ARS), Soil Conservation Service (SCS), and Purdue University. USLE was an empirical model of simple structure that captured the main effects of rainfall intensity, soil type, topography, and management on sheet and rill erosion, with no attempt to account for sediment deposition nor gully erosion (Dabney et al., 2011; Renard et al., 1991). USLE only predicts the amount of soil loss that results from sheet or rill erosion on a single slope and does not account for additional soil losses that might occur from gully, wind or tillage erosion (Stone and Hilborn, 2000). In the early 1980s a programme to develop technology to replace the USLE was initiated, resulting in the computer-based revised universal soil loss equation (RUSLE) model, documented in written form in 1997 (Renard et al., 1997, cited by Dabney et al., 2011). RUSLE incorporated significant advances over the USLE and this permitted the application of soil erosion estimation for a great variety of crops and management practices beyond those in the original USLE data base. Some other changes include a sub factor method for computing values for the cover-management factor, improved factor values for the effects of contouring, terracing, strip cropping, and management practices for rangeland (Dabney et al., 2011; Renard et al., 1991; Stone and Hilborn, 2000). In addition, RUSLE introduced new slope length and steepness (LS) algorithms reflecting rill to inter-rill erosion ratios and the capacity to calculate LS products University of Ghana http://ugspace.ug.edu.gh 37 for slopes of varying shape (Renard et al., 1991; Stone and Hilborn, 2000) and thus is a better estimate of soil erosion and sediment yield in landscapes than USLE. 2.5 Justification for Research Methodologies Many published soil loss and sediment yield models, sediment yield, bank erosion and sediment source tracking approaches were reviewed to understand their techniques and applications. Relative merits and shortcomings of these methodological approaches in soil erosion and sediment yield studies were discussed. It was realised that there is no universally acceptable method or approach to modelling a catchment's soil erosion and sediment yield as well as field data collection of sediment discharge measurements. The choice of any methodology depends on the availability of resources (secondary data, financial, equipment etc) at the disposal of the researcher. However, for soil erosion and sediment yield modelling, the literature showed that, RUSLE has been widely used with very good results, is applicable in a GIS environment and so can provide a spatial distribution of erosion and soil loss in catchments. The model also requires a small and simple input data set and is relatively easy to use and thus suitable for this study. Hence, the RUSLE model was chosen to model soil erosion and sediment yield in the basin. However, a slight modification to the model was made. Due to the high intensity of tropical rains as compared to temperate regions, two rainfall erosivity events were adopted; the standard 12.5mm erosive event and then 12mm event, to see whether a lower rainfall event will have any significant impact on soil erosion and sediment yield in the basin. The thunderous tropical raindrops (high intensity and large rain drops) will have more destructive impact on soil aggregates as compared to temperate raindrops, hence the need to examine this effect. For University of Ghana http://ugspace.ug.edu.gh 38 instance, observations made in the savannah and forest regions of Nigeria indicate that rains with median drop size in excess of 2.5 mm and energy load of 100 J m -2 mm -1 are common (Lal, 1985). The energy load of rains in western Africa is generally more than that of the subtropical rains of Zimbabwe (Elwell & Stocking, 1975). Also, the mass of soil detached per unit area by tropical rains was observed to be more closely related to the momentum than to the kinetic energy of a rainfall in East Africa (Rose, 1960). In situ field data collection involved suspended sediment yield measurements, erosion pin to measure bank erosion and the simple mixed model using one fingerprint ( 210 Pb) to track sediment source trajectory in the basin. Suspended sediment concentration measurements were undertaken; coupling depth integrated and dip sampling in order not to miss critical flows. Gauge readers were engaged to undertake dip sampling when there are significant changes in flows. Despite its inability to capture continuous bank erosion data, the traditional erosion pin measurement was employed to measure bank erosion due to financial constraints to acquire PEEP equipment. Considering the size of the Pra Basin, much of 7 B would be decayed in view of the resident time to outlets of sub-catchment basins, hence 7 B at the outlets of sub-basins will contain only tracers of eroded surfaces close to sub-basin outlets. There is also no equipment available for the analysis of 137 Cs and 7 B in the laboratory. Lead-210 was therefore used as the tracer to det