A Predictive Model for Quantifying the Combined Effect of Land Use Change and Climatic Variability on Sedimentation in Malaba Sub Catchment

Over the previous years, sedimentation in Malaba sub catchment has become a major problem, this research aimed at using SWAT as a predictive model to assess the impacts of historical and future climate change and land use changes on sedimentation in Malaba sub catchment. Two variables rainfall and temperature were considered under climate change. Four scenarios were modeled in SWAT during this study, the first scenario focused on historical sedimentation in the catchment, land use data of 2000 and weather data of 1990 to 2005 was used to simulate flow and sediment yield in the catchment. The second scenario focused on the current sediment yield in the catchment where land use data of 2020 and weather data of 2003 to 2020 was used to simulate the sediment yield and flow. Calibration and validation was done using SWAT CUP software. The calibrated and validated model was then used to further simulate two future scenarios. The two future scenarios used projected land use data and projected weather data. Land use data was projected using Clerk Labs Terr set software while weather data was projected using the Statistical Downscaling Model (SDSM). The third scenario considered projected land use data of 2035 and projected weather data of 2020 to 2035 to simulate flow and sediment yield of 2023 to 2035. The forth scenario considered projected land use data of 2050 and projected weather data of 2033 to 2050 to simulate catchment flow and sediment yield of 2036 to 2050. The results obtained from climate change analysis indicated un even changes, rainfall over the catchment is expected to increase, maximum temperature is expected to increase, minimum temperature is expected to reduce. Land use change results after comparison between the historical land use and projected land use indicated that from 2020 to 2050 cropland and built up area will increase while forests and water bodies will reduce. Average annual Sediment yield of 1993 to 2020 is expected to generally increase from 4975.4 tons/km2/year to an average annual sediment yield of 5525.6 tons/km2/year from 2023 to 2050. Therefore, mitigation measures should be adopted to ensure sustainable management of the catchment. Some of the recommendations made in this research are; Government through NEMA should put restrictions on cultivation close to the river banks, encourage planting trees within the catchment, bare lands within the catchment should be converted to other sustainable land uses. The study findings of this research shall be relevant for planning, design and management of reservoirs, dams, irrigation systems and sustainability of eco systems in the catchment.


CHAPTER ONE: INTRODUCTION 1.Background of the study
The International Geosphere Biosphere Program (IGBP) initiated by the International Council of Scientific Unions (ICSU) in 1987 (Steffen et al., 2004), as well as a number of related initiatives, have focused increasing attention on the changes in the functioning of the Earth system caused by human activity and on the problems associated with the sustainable management of this changing system over the coming centuries.Much of this attention has been directed to the increased emission of greenhouse gases, leading to climate change.Sediment yield estimation along with identification of soil erosion mechanisms is essential for developing sophisticated management approaches, assessing, and balancing different management scenarios and prioritizing better soil and water conservation planning and management, at a watershed scale, land management practices are commonly utilized to minimize sediment loads (Megersa et al., 2023) Sediment discharge is one of the main water quality concerns in integrated watershed management.A proper identification of sediment sources is therefore important to the success of watershed conservation programs (Yongbo et al.,2015) Historical climate change has been observed and further changes are anticipated (Alava et al., 2018).According to the Intergovernmental Panel on Climate Change (IPCC), the average temperature on the planet has increased by about 0.6 degrees Celsius over the past century, and it is expected to further increase by 4 degrees by the end of the 21st century.Urbanization, agriculture, and deforestation are just a few of the everyday anthropogenic activities that have altered land use and cover over time and space.River sedimentation and stream flow are normally affected by changes in land use and land cover.Surface soil is being eroded from river basins at a rate of 60 billion tons per year, this results in 24 billion tons of sediment being released into the world's water bodies and almost 25 billion tons of soil being removed from agricultural land (Deltares, 2018).Sedimentation and eutrophication are caused by excessive soil erosion, which can make land unsuitable for farming and releases a lot of sediment, phosphorus, and nitrogen.
A study was done on the world's 145 major rivers with consistent long term sediment records and the results indicated that about 50 % of the rivers have statistically a significantly downward flow trend due to sedimentation (Walling and Fang, 2003).Roughly 25% of the population in Africa is at present experiencing water shortages (Bates et al., 2008).In recent years, hydrological simulation models have been widely utilized to assess the impact of changing land use and cover on the hydrologic cycle (Lin et al., 2008).A study on sediment impacts in Africa's transboundary lake/river basins; a case study of the East African Great Lakes indicated that the current population pressure, inappropriate cultivation practices, forest removal and high grazing intensities on forests, wetlands, rangelands and marginal agricultural lands leads to unwanted sediment and stream flow changes that mainly impacts the downstream human and natural communities.Forests and bushes are cleared, and wetlands are encroached to create space for human settlement, road construction and to satisfy wood fuel energy demands.Similarly, pastoral areas are subjected to growing human and livestock populations, leading to land degradation, soil erosion and an increase in the load from non-point pollutants (Olago and Odada, 2007).River sediment yield is directly influenced by climate and land use variation (Guo et al., 2018).According to research done on River Ruaha catchment in Tanzania, annual total sediment load increased as a result of modifications to land use and land cover (Nathalie and Gutierrez, 2022).

A Statement of the Problem
Unsustainable land management practices in Malaba sub catchment such as cultivation on the river banks, wetland encroachment, uncontrolled cultivation in hilly areas, disposal of untreated wastes from industries into the environment, deforestation and charcoal burning have been the major causes of climate change and land use land cover changes in Malaba sub catchment, these have resulted into; extensive catchment degradation, increased occurrence of disasters like floods and landslides; and significantly increased sediment loads in River Malaba (Nile Basin Initiative, 2015; MWE, 2018; MWE, 2016; Barasa et al., 2016).The excess sediment loads have; caused water quality deterioration and reduced the river channel • Email: editor@ijfmr.com

IJFMR23057185
Volume 5, Issue 5, September-October 2023 4 capacity hence causing floods, these have resulted into loss of human life and property, outbreak of diseases like cholera and malaria and increased costs of water treatment (MWE, 2018) Therefore there is great need for a multi-disciplinary approach for proper water resources planning and catchment sustainability to reduce the impacts of sedimentation in Malaba sub catchment through assessing the historical, present and future impacts of climate change and land use land cover changes on sedimentation and flow of R. Malaba in Malaba sub catchment.

Significance of the study
This research aimed at achieving; SDG 6 (Clean water and sanitation for all), SDG 13 (Take urgent action to combat climate change and its impacts), SDG 3(Good health and wellbeing), NDP III whose major goal is to increase household incomes and improve quality of life of Ugandans, Vision 2040 by indirectly increasing the income levels of the local people through reducing the flood risk.

Scope of the study Conceptual scope
This study mainly focused on the application of SWAT to assess the impacts of historical and future climate and land use changes on sedimentation and identification of areas prone to sedimentation in Malaba sub catchment

Geographical scope
This study was carried out on River Malaba located in Malaba sub catchment which is a trans boundary sub catchment shared by Kenya and Uganda.Its covers the following areas; Bududa, Namisindwa, Manafwa, Tororo, Busia, Namayingo, Butaleja, Namutumba, Bugiri districts in Uganda, Bungoma and Busia divisions in Kenya.

Time frame of the study
This study was executed in a period of nine (9) months.

1.7
Conceptual framework A conceptual framework for the study was developed to illustrate the expected relationship between the dependent and independent variables.It defines the relevant objectives for the research process and maps out how they come together to draw coherent conclusions.

Climate Change
Future climate is partly determined by the magnitude of future emissions of greenhouse gases, aerosols and other natural and man-made forcing (Collins et al., 2013).Policymakers, planners, investors and vulnerable communities need information about future climate so that they can prepare for expected trends and changes.Climate predictions are estimates of future natural conditions, while climate projections are estimates of future climates under the assumptions of future human related activities such as socioeconomic and technical developments (UNITAR, 2015).In Uganda, particularly in the southwest, the typical temperature is rising in semi-arid regions.The National Adaptation Program of Action (NAPA) of Uganda reports that, between 1960 and 2010, the country's average temperature increased by 0.28°C every decade, with the months of January and February being the most impacted, with an average increase of 0.37°C per decade.In contrast to a decline in the number of cold days, the country now experiences many more hot days than cold ones.According to earlier records of Uganda's glaciers, the ice caps on the Rwenzori Mountains have drastically diminished during the past century.Mount Baker (96%), followed by Mount Speke (91%), has the highest percentage of ice loss.68 percent of the ice has been lost from Mount Stanley.The changing temperature patterns in Uganda have been linked with drought and consequent increases in cattle deaths in the cattle corridor (MWE, 2015) Precipitation in the country is highly variable, but overall, Uganda has experienced a statistically significant reduction in annual as well as seasonal rainfall.Seasonal rainfall for March, April and May has been most affected, with decreases of 6.0 mm per month.Decline in rainfall has been observed in some Northern districts: Gulu, Kitgum, and Kotido.While trends in extreme rainfall conditions are more difficult to define due to the lack of data and seasonal variability, droughts have increased in Uganda over the past 60 years.(WBG, 2021) Flooding, particularly in low-lying areas of the country, presents the largest risk.Each year, floods impact nearly 50,000 people and costs over $62 million.39Uganda experiences both flash floods and slow-onset floods, which are common in urban areas, low-lying areas, areas along river banks and swamplands (WBG, 2021)

Climate change modeling
Climate models are mathematical representations of the climate developed by scientists to understand and predict the climate system.In order to be able to do this, the models divide the earth, ocean and atmosphere into a grid.Simulating climate change at the regional and national levels is essential for policymaking.However, Global Climate Models (GCMs) have a coarse spatial resolution that is not suitable to understand the climate at a smaller scale.(UNITAR, 2015)

Global Circulation Models (GCMs)
A Global Climate Model (GCM) combines a series of models of the Earth's atmosphere, oceans, and land surface (Randall et al., 2007).GCMs divide the earth into many layers and thousands of three-dimensional gridded spaces.These models are skilled at replicating past and current climate.For example, GCMs accurately reproduce observed temporal warming trends, sea ice dynamics, and extreme weather events.The climate models project possible future climate shifts under the conditions of the specific scenarios.These models are run multiple times using various scenarios of future conditions, such as population levels and anticipated emissions of carbon dioxide (CO2) or other greenhouse gases.Each GCM is distinct and has a different sensitivity to greenhouse gas emissions.(Meehl et al., 2007;Schcramm et al., 2016).This range, taken as a whole, is important to researchers for providing a sense of the uncertainty surrounding possible future events given a particular scenario and period.To capture this range and make use of the complement of projections, ensembles of multiple global climate model simulations are often used.
(Schcramm et al., 2016).Climate models are being subjected to more comprehensive tests, including, evaluations of forecasts on time scales from days to a year.This more diverse set of tests increases confidence in the fidelity with which models represent processes that affect climate projections (Randall et al., 2007).GCMs are complex, three dimensional models that are continually evolving to incorporate the latest scientific understanding of the atmosphere, oceans, and Earth's surface.Originally, "GCM" stood for General Circulation Model, since the original focus of these physics-based models was to simulate the circulation of the atmosphere and ocean.Today, however, global climate models incorporate many other facets of the Earth's climate system, including chemistry, biospheric processes, land use, etc. (Hayhoe and Stoner, 2015).Some GCMs are better than others at reproducing important large-scale features of certain regions, such as sea ice in the Arctic (Wang et al., 2007).Some of the CMIP5 models include;

Regional circulation Models (RCMs)
RCMs operate at much higher resolution and often, with more detailed topography and use of physical parameters.This downscaling can be extended to even finer detail in local models.Most RCM simulations use GCM fields from pre-computed global simulations as boundary conditions.This approach allows RCMs to draw from a broad set of GCM simulations, such as CMIP5, but does not allow for possible twoway feedbacks and interactions between the regional to global scales.Dynamical downscaling can also be conducted interactively through nesting a higher-resolution regional grid or model into a global model during a simulation (Hayhoe et al., 2017).

Coupled Model Intercomparison Project Phase 5 (CMIP5)
This presents an unprecedented level of information on which to base projections including new Earth System Models with a more complete representation of forcings, new Representative Concentration Pathways (RCP) scenarios and more output available for analysis (Hayhoe and Stoner, 2015).The four RCP scenarios used in CMIP5 lead to a total radiative forcing (RF) at 2100 that spans a wider range than that estimated for the three Special Report on Emission Scenarios (SRES) scenarios (B1, A1B, A2) used in the Fourth Assessment Report (AR4), RCP2.6 being almost 2 W m-2 lower than SRES B1 by 2100.The magnitude of future aerosol forcing decreases more rapidly in RCP scenarios, reaching lower values than in SRES scenarios through the 21st century.Carbon dioxide (CO2) represents about 80 to 90% of the total anthropogenic forcing in all RCP scenarios through the 21st century.The ensemble mean total effective RFs at 2100 for CMIP5 concentration-driven projections are 2.2, 3.8, 4.8 and 7.6 W m-2 for RCP2.6,RCP4.5, RCP6.0 and RCP8.

Downscaling Models
Despite GCMs' sensitivities of hundreds of kilometers, climate change data is required for many impact assessments at a much smaller geographical scale (Dibike et al., 2008).As a result, downscaling techniques have emerged as a means of connecting atmospheric variables to grid and sub-grid scales.The two most common approaches for obtaining data on a regional or local scale from global climate scenario generated by GCMs are through numerical and statistical downscaling models (Wilby et al., 1998).A regional climate model (RCM) is used in numerical downscaling, which is also known as dynamic downscaling.In statistical downscaling, a statistical connection is utilized to compare local fluctuations acquired from historical data records to the large-scale climatic condition.

Table 2-2 shows the strength and weaknesses involved in dynamic and statistical downscaling techniques Dynamic Downscaling
In this type of downscaling, most RCM simulations use GCM fields from pre-computed global simulations as boundary conditions.This approach allows RCMs to draw from a broad set of GCM simulations, such as CMIP5, but does not allow for possible two-way feedbacks and interactions between the regional to global scales.Dynamical downscaling can also be conducted interactively through nesting a higherresolution regional grid or model into a global model during a simulation (kotamarthi et al. 2016).Within the global climate model of coarser scale, dynamic downscaling is also referred to as limited area models (LAMs) and involves the climate model of finer scale regionally.A dynamic approach makes use of the conditions of a GCM boundary's outputs for the targeted region.Future environment at a size of a district is determined utilizing an environment model completely actual in nature (Jorge et al., 2015) Dynamic models provide the major benefit of accounting for local circumstances, such as physically regular changes in atmospheric chemistry or surface vegetation.However, in order to calculate the same scenarios, regional climate models (RCMs) require the same amount of processing time as the GCM, and they are not easily adaptable to new regions.The initial conditions, particularly soil temperature and moisture, have a significant impact on RCM results at the beginning (Nasr et al., 2007).

Statistical downscaling models
Empirical statistical downscaling models (ESDMs) convert large-scale predictors or patterns into highresolution forecasts at the size of observations by fusing GCM output with prior historical data.Individual weather stations and gridded datasets can also be utilized as observations in an ESDM.They can provide a variety of outputs, from big grids to assessments tailored for a particular place, variable, or decisioncontext (Hayhoe et al., 2017).Statistical models are generally flexible and less computationally demanding than RCMs.A number of databases using a variety of methods, including LOCA (Localized constructed Analogs) provide statistically downscaled projections for a continuous period from 1960 to 2100 using a large ensemble of global models and a range of higher and lower future scenarios to capture uncertainty due to human activities.ESDMs are also effective at removing biases in historical simulated values leading to a good match between the average (multi-decadal) statistics of observed and statistically downscaled climate at the spatial scale and over the historical period of the observational data used to train the statistical model ).Future regional climate scenarios are constructed, either by resampling from the observed variable distributions or by first generating synthetic sequences of weather patterns using Monte Carlo techniques and resampling from observed data.The main appeal of circulation-based downscaling is that it is founded on sensible linkages between climate 204 on the large scale and weather at the local scale.The technique is also valid for a wide variety of environmental variables as well as multi-site applications.However, weather typing schemes are often parochial, an inadequate basis for simulating rare or extreme events, and entirely dependent on stationary circulation-to-surface climate relationships.Potentially, the most serious limitation is that precipitation changes produced by changes in the frequency of weather patterns are seldom consistent with the changes produced by the host GCM Stochastic downscaling approaches typically involve modifying the parameters of conventional weather generators such as wgen (Wilks, 1999) or LARS-WG (Semenov and Barrow, 1997).WGEN simulates precipitation occurrence using two-state, first-order Markov chains: precipitation amounts on wet-days using a gamma distribution; temperature and radiation components using first-order trivariate auto regression that is conditional on precipitation occurrence.

Statistical Downscaling Model
Statistical DownScaling Model (SDSM) model is a prominent tool, freely available in the public domain.SDSM is best described as a hybrid of the stochastic weather generator and regression-based methods.This is because large-scale circulation patterns and atmospheric moisture variables are used to linearly condition local-scale weather generator parameters (e.g., precipitation occurrence and intensity).Additionally, stochastic techniques are used to artificially inflate the variance of the downscaled daily time series to better accord with observations.To date, the downscaling algorithm of SDSM has been applied to a host of meteorological, hydrological and environmental assessments, as well as a range of geographical contexts including Europe, North America and Southeast Asia (Wilby and Dawson). 1 facilitates the rapid development of multiple, low-cost, single-site scenarios of daily surface weather variables under current and future climate forcing.Additionally, the software performs ancillary tasks of data quality control and transformation, predictor variable pre-screening, automatic model calibration, basic diagnostic testing, statistical analyses and graphing of climate data (Wilby and Dawson, 2004).

Impacts of Climate Change on sedimentation
Climate change has increased precipitation concentration, volume, and intensity, which has had a considerable influence on runoff and soil erosion in many watersheds (Diodato et al., 2020).The primary component of river sediments and a significant contributor to reservoir or river dam sediment deposition are the sediments produced by watershed erosion, which are carried to rivers by surface runoff (Chen et al., 2006).The amount of soil erosion significantly affects how river channels develop, affecting river stability, flood prevention safety, and river repair planning.Therefore, controlling sediment output is essential for watershed management, especially given that it frequently comes at a significant cost (Chen et al., 2020).Modeling sediment transport and storage is challenging because of complex relationships between climatic forcing, hydrological connectivity, sediment production, and the different geomorphic thresholds involved (Campforts et al.,2020)

Land Use Land Cover change
Land cover defined as the assemblage of biotic and abiotic components on the earth's surface is one of the most crucial components of the earth system (Turner et al.,1994).Land cover also reflects the availability of food, fuel, timber, fiber and shelter resources for human populations, and serves as acritical indicator of other ecosystem services such as biodiversity.Information on land cover is fundamental to many national and global applications including watershed management and agricultural productivity.Thus the need to minotor land cover is derived from multiple intersecting drivers, including the physical climate, ecosystem healthand societal needs.(Sudhakar and Kameshwara, 2006).LULC change can also be defined as the modification of surface features on earth's landscape which is realized by the difference in their surface appearance assessed at two different times (Ayele, 2011).More than 50% world's population resides in urban areas, and this figure is Due to human-made structures and activities, the intensity and characteristics of surface flow and sediment yield have significantly changed and meant a great deal to the watershed stakeholder and manager

Remote sensing (RS) and GIS techniques in LULC change analysis
Remote sensing is the sensing of the Earth's surface from space by making use of the properties of electromagnetic waves emitted, reflected or diffracted by the sensed objects, for the purpose of improving natural resources management, land use and the protection of the environment (United Nations, 1986).
Remote sensing is the science and art of obtaining information about an object, area, or phenomenon through the analysis of data acquired by a device that is not in contact with the object, area, or phenomenon under investigation (Lillesand, et al., 2008).

Hydrology
Hydrology deals with the occurrence, movement, and storage of water in the earth system while hydrologic science comprises understanding the underlying physical and stochastic processes involved and estimating the quantity and quality of water in the various phases and stores (Jose D Salas et al., 2016).The study of hydrology also includes quantifying the effects of such human interventions on the natural system at watershed, river basin, regional, country, continental, and global scales.The multidisciplinary geoscience of hydrology, which examines the mechanisms driving the replenishment and depletion of terrestrial water resources, can also be referred to as this.It focuses on comprehending and characterizing quantitatively physical, chemical, and biological elements and processes that interact and function at various spatiotemporal scales and are influenced by human actions.(2000) Schulze and (2009) Savenije 2.4.1 Hydrologic Modeling Historically, hydrological modelling was undertaken to better understand the relationships between rainfall and runoff in the latter half of the 19th century in response to three main engineering problems urban sewer design, land reclamation drainage systems design, and reservoir spillway design.It is conceivable, however that these types of engineering problems date back to before the Roman Empire, and that planners of that time dealt with similar issues at smaller scales (Hubbart, 2012) Hydrologists are mainly concerned with evaluation of catchment response in order to plan, develop, manage and operate various water resources schemes.There is continuous circulation of water between earth and atmosphere.This is signified by different phases in the hydrologic Cycle which is the fundamental principle of hydrology (NIH Roorkee, 2017).In order to quantify the effect of land-use change on hydrological components, hydrological models have been used to conceptualize and investigate the interactions between climate, human activities (such as land use change), and water resources (Barasa, 2014).

Classification of hydrological models
Hydrological models can be classified in different ways; Based on the description of physical processes, the hydrological models can be classified into three groups namely; empirical (data driven models), conceptual, and physically based.Based on the spatial representation the hydrological models can be classified into: lumped, and distributed.Based on the aspect of randomness the hydrological models can be classified into: deterministic, and Stochastic (Džubáková, 2010, Singh and Frevert, 2002).Hydrologic models can also be classified as; lumped and distributed parameter models, conceptual and hydrodynamic models, models with fitted and physically determined or empirically derived parameters, event and continuous simulation models (Todini, 1988;Knapp et al., 1991).Empirical models are the most basic kind of numerical models used to simulate streamflow in a direct relationship with other measurable variables.Although this kind of model may be used by a wide range of people due to its simplicity, the model's output limits its utility.Unit hydrograph is the most typical illustration of a black box model.According to Courault et al. (2005) and Davie (2008), the model linkages are based on empirical data rather than necessarily on physical processes.Physically-based models are those that are based on physical procedures and are modeled after a knowledge of physical mechanics.These models frequently have high computing and data needs.These models provide more experimental and explanatory strength.Their prediction ability is frequently equivalent to or worse than that of empirical models due to the greater number of assumptions that are required (Beven, 1989;Grayson et al., 1992).Lumped conceptual models were the initial attempt to numerically represent the many hydrological events within a basin.The catchment region receives additional rainfall, and the water budget technique is utilized to track water losses and flows there.The word "lumped" is used since there is no geographical discretization and all the processes work at the same spatial scale.The scale picked is frequently a catchment, or even a sub-catchment.The word conceptual is employed because it is frequently believed because the equations determining flow rates are conceptually comparable to the physical procedures in use (Refsgaard, 1997; Davie, 2008).

Description of some Hydrological models The Distributed Hydrology Soil Vegetation Model (DHSVM)
DHSVM provides a dynamic representation of watershed processes at the spatial scale described by Digital Elevation Model (DEM) data.The modeled landscape is divided into computational grid cells centered on DEM nodes, this characterization of topography is used to model topographic controls on absorbed shortwave radiation, precipitation, air temperature, and downslope water movement.Vegetation characteristics and soil properties are assigned to each model grid cell, these properties may vary spatially throughout the basin.In each grid cell the modeled land surface can be composed of a combination of vegetation and soil, at each time step, the model provides simultaneous solutions to energy and water balance equations for every grid cell in the watershed.Individual grid cells are hydrologically linked through surface and subsurface flow routing (Wigmosta et al., 1994) DHSVM has been used to evaluate changes in flood peaks caused by enhanced rain-on-snow and springtime radiation melt reaction impacts of forest paths and traffic drainage, and the forecasting of erosion of sediment and transportation (Wigmosta et al., 1994).MIKE-SHE MIKE SHE uses MIKE Hydro River to simulate channel flow, MIKE Hydro River includes comprehensive facilities for modelling complex channel networks, lakes and reservoirs, and river structures, such as gates, sluices, and weirs.In many highly managed river systems, accurate representation of the river structures and their operation rules is essential.In a similar manner, MIKE SHE is also linked to the MOUSE sewer model, which can be used to simulate the interaction between urban storm water and sanitary sewer networks and groundwater (Zhuhuan and Zhou, 2019) MIKE SHE is applicable at spatial scales ranging from a single soil profile, for evaluating crop water requirements, to large regions including several river catchments, such as the 80,000 km2 Senegal Basin

Soil Water and Analysis Tools (SWAT)
To predict the long-term effects of rural and agricultural management practices (such as specific agricultural land planting, tillage, irrigation, fertilization, grazing, and harvesting procedures) on water, sediment, and agricultural chemical yields in large, complex watersheds with varying soils, land use, and management conditions, USDA-ARS created the physically based SWAT model in the early 1990s.SWAT considers evapotranspiration, channel transmission losses, lateral subsurface flow, groundwater return flow, percolation, and surface runoff.The modified SCS curve number approach is used to estimate runoff volume.While the watershed concentration time is calculated using Manning's formula, taking into account both overland and channel flow, peak runoff forecasts are based on a modified version of the Rational Formula (Grizzetti et al., 2003).

SWAT Model Calibration and Validation
Model Parameter Sensitivity Analysis; Evaluating the input parameters to determine how they affect the model output is called sensitivity analysis of SWAT model parameters.It aids in the reduction of uncertainty not only during the model's development but also during its validation (Hamby, 1995).By taking into account the parameters that have the greatest sensitivity and, as a result have the greatest impact on the behavior of the simulation process.The sensitivity analysis reduces the number of parameters that need to be used in calibration.

CHAPTER THREE METHODOLOGY 3.1 Introduction Research Design
This study is both qualitative and quantitative in nature.

Description of Study area 3.2.1 Location of the study area
Malaba sub catchment is part of the Mpologoma catchment which is located in Kyoga Water Management Zone(WMZ) one of the four Water Management Zones in Uganda, other WMZs in Uganda include Victoria Nile WMZ, Albert Nile WMZ and Upper Nile WMZ.Malaba sub catchment, covers about 3480 km2.The river origins from the slopes of Mount Elgon at the border of Uganda and Kenya, the catchment area covers parts of Bugiri, Namayingo, Tororo, Busia, Namutumba, Butaleja Namisindwa, Manafwa, and Bududa districts in Uganda, Bungoma and Busia Counties in Kenya.The River empties into the River Mpologoma, whose waters go into Lake Kyoga. Figure 3 1 depicts the position of the Malaba sub watershed in Kenya and Uganda.

Population and resources of Malaba sub catchment
Malaba sub catchment has got an approximate population of 4 million people (NBI, 2012).The watershed is rich in minerals including sand, limestone, gold, and phosphates.Mining operations are carried out on both a big and small scale.For instance, sand is mined on a small basis and supplied to regional construction firms.In the Busia district of Uganda, small-scale artisanal miners mostly extract gold on a small scale.Additionally mined in the Tororo area are phosphates and limestone.

Geology and topography
The watershed is hilly with undulating plains; Mount Elgon, at a height of around 4,299 meters above sea level, has the highest peak.The watershed's midsection and downstream regions include undulating plains, and the entire catchment is underlain by rocks of the Precambrian and Tertiary Pre-Elgon volcanic type.Precambrian basement rocks consist of a range of granites, gneisses, quartzite, and tiny pockets of densely folded metamorphic rocks (DSOER, 2004).

Data collection and preparation and
In this study, primary, secondary and tertiary data were used, Table 3- 3-1 Secondary data used in this study Data preparation and processing Watershed Delineation was the initial activity carried out in ArcGIS environment using the SWAT automatic watershed delineator, this was aimed at defining the boundaries of Malaba sub catchment which informed the geographical scope during collection of other datasets.

Weather data
Observed weather data was obtained from Tororo weather station.Data Quality check was done using SDSM software, the observed data was found to have a lot of missing values.Bias correction was done using Satellite data of Malaba sub catchment downloaded from NASA's Power Access website, time series of 1990 to 2020 were download with variables of Precipitation, maximum and minimum temperatures Land use data Land use data of 2000 and 2020 was obtained from google earth through digitizing the historical imageries of the respective years, other satellite based Land use land cover datasets were not used because they presented a very low accuracy after ground truthing, for example the 2020 Esri Land use map of Malaba sub catchment was not showing any water body in the area yet on ground we actually have River Malaba, USGS land use map of Malaba sub catchment only displayed savannah woody vegetation in places where we expect to find River Malaba.The digitized maps clearly indicated River Malaba and other features within the catchment, the maps were then imported into Arc Map for processing.Land use land cover was classified into 6 six classes namely; Water for all water bodies, Cropland for Agricultural Land, Bare land for any land without any vegetation or water, Forests for area with very many trees, Grasslands for areas with scattered short trees and grass, Built up areas for residential, urban and commercial areas and other activities that cause greenhouse gases.However, the government of Uganda through its ministry of water and Environment is trying its level best to curb down these greenhouse gas emissions though with a lower rate of success which makes RCP 4.5 a better option compared to RCP 2.6 and RCP 8.5.The following steps were followed to project/forecast weather data in SDSM environment; Data quality check; Data quality check was conducted by loading the observed data and checking for missing data.There was no missing data.Variable screening; This was done by using the analyse and correlation tools.A total of twelve predictor variables were tested to select the most sensitive predictors, six most sensitive predictor variables were selected from the software interface considering a significance of 0.05.The Predictor variables used include; ncepp500gl, ncepp5_zgl, nceps850gl, ncepp8_vgl, nceptempgl, ncepshumgl.For precipitation projection, conditional process was used with a threshold of one (1) while for temperature projection an un conditional process was considered with a threshold of 0. The statistical model was calibrated using the selected predictor variables for a period 0f 1990 to 2003 and validated for a period of 2004 to 2005.The projected weather data (2020 to 2050) was then generated and results were statistically compared.

Climate change assessment
Historical rainfall, maximum and minimum temperature are the parameters that were considered during the climate change assessment.Historic data was from 1990 to 2020 while projected data was from 2020 to 2050.Mean monthly historical values were statistically compared to mean monthly projected values of the three parameters (minimum temperature, maximum temperature and rainfall) using bar graphs to assess the changes Historical total annual rainfall values were statistically compared to projected total annual rainfall values using bar graphs

LULC projection Land use land cover forecasting
This was done in Clark Labs Terr set software, Land change modeler tool was used.Land use land cover change between 2020 and 2000 was analyzed and change maps generated, transition sub-models were grouped into a single transition model (named Anthropogenic disturbance) which was used in the model.Evidence likelihood was used as the transformation type, other transformation types available include natural log, exponential, logit, square root and power.Evidence likelihood was selected because it is a very effective means of incorporating categorical variables into the analysis.The transition sub-model structure was set by importing three static drivers into the model which included; distance to the existing roads within Malaba sub catchment, distance to towns and slope of the sub catchment.Multi-Layer Perceptron (MLP) neural network method was used to run the model, the automatic training and dynamic learning rate options were all considered.The transition sub model was then run and a report was generated.
The default Terr set prediction method by Markov Chain was utilized to estimate the amount of change by comparing the dates of the earlier and later land cover maps.The process develops a transition probabilities file and calculates the precise amount of land that would be anticipated to transition from the later date to the prediction date based on a future projection of the transition potentials.Prediction date of 2050 was considered with two recalculation stages that is to say 2035 and 2050.The model was then run and land use maps of 2035 and 2050 were generated each with a transition potential map.

LULC change assessment
In the Arc GIS environment, the area covered by each land use and land cover categorization was calculated using the land use maps from 2020 and 2000.Comparing the computed areas of the two LULC maps allowed us to evaluate the gains and losses in land use that took place during the 20-year period from 2000 to 2020.Clerk Labs Terr Set software was also used to assess land use changes that occurred between 2000 and 2020 using the Land Change Modeler.The transition and persistence maps were developed.

Sediment Modeling and simulation
This was done through building a rainfall runoff model in Arc GIS environment using the Soil and Water Analysis Tool (SWAT).SWAT model is a flexible and physically distributed model which was developed as a river basin scale to quantify and predict runoff, sediment yield, nutrients and sediment transport from watersheds and river basins.Watershed delineation was done using the Automatic watershed delineator, a 12.5 m resolution DEM data was used.Under the stream definition, the flow direction and accumulation was obtained which later produced the stream network through creating streams and outlets.The sub-basin outlets and inlets were defined using a point source input, the whole watershed outlet was defined, the watershed was finally delineated and sub-basin parameters were calculated.Land use, soils and slope definition was carried out using soil data, and land use data of Malaba sub catchment of 2000, reclassification of data for both soils and land use was performed using information from the SWAT2012 database.Definition of Hydrologic Response Units (HRUs), Hydrologic response unit definition was carried out by defining the threshold proportions of coverage of different soil and land use classifications; multiple HRU was chosen and given suitable percentages (Land use,10%; soil data, 10%; and slope, 5%).From the elevation bands, the HRUs were created and the HRU analysis report was generated.Weather data definition was done by considering user defined variables that is to say rainfall and temperature, these were fed into the model considering a daily time step, years from 1990 to 2005 were considered.Under the SWAT input tables, tables were built.From the Edit SWAT inputs, Potential Evapotranspiration (PET) method of Penman monteith was selected, SCS Run off curve number method was selected for estimation of run off.Each land use land cover classification automatically generated a unique Curve number and manning's coefficient which were used to estimate run off in the model using the modified Universal Soil Loss Equation inbuilt in SWAT.

SWAT simulation
The warm up period was set to three years to increase accuracy of the model.The model was Run with a starting date of 01/01/1990 and ending date of 31/12/2020.The total simulation period was from 1990 to 2005 and the first 3 years were used as warm up period, which helped to stabilize the model hence minimizing possible model errors.The warm up period allows "buckets" in SWAT (reservoirs, wetlands, soil moisture, aquifers) to fill up and reach stable values.Sediment yield in the first few years are usually underestimated because of this, 1990-1992 period was used for warm up of the model to establish the initial soil water conditions The above steps basically explain how the first scenario was achieved, the second scenario was achieved using the same procedures while changing only LULC data and Weather data.The third and fourth scenarios were done using a calibrated and validated model whilst changing the LULC data and weather data to the respective years.

Calibration and Validation
SWAT CUP (Soil and Water Analysis Tool Calibration and Uncertainty programs) 2012 Version 5.1.6was used, the sequential Uncertainty fitting version 2 (SUFI-2) was used as the calibration method.Initially, twelve input parameters were considered for optimization.The available Malaba sub catchment sediment yield data was insufficient to carryout calibration and validation therefore only observed stream flow data was used as the observed variable.The calibration period was from 1997 to 2005 a period of 9 years; the validation period was from 2008 to 2010 a period of 3 years.Five iterations were considered each with 500 simulations, a daily time step was considered.NSE (Nash serticlif efficiency) and R 2 (coefficient of ) were prioritized as measure of accuracy methods.

Sensitivity Analysis
Sensitivity analysis is a process of determining the rste of change in model output with respect to changes in model inputs (parameters).It is necessary to identify key parameters and the parameter precision required for calibration (Ma et al.,2000).Sensitivity analysis was conducted to identify the most sensitive parameters in the catchment, the most sensitive parameters were then used for calibration of the model.Global sensitivity analysis method was adopted in this model and the sensitivities of the different parameters were identified and ranked.Considering the above historic and projected LULC maps, tables and the bar chart, it is observed that;Built up area increased from 51km2 in 2000 to 79km2 in 2020, an increment of 18km2 occurred.This study projected that built up area will increase to 85km2 by 2035 and to 91 km2 by 2050.Cropland area increased from 1062 km2 in 2000 to 2324 km2 in 2020, an increment of 1262 km2 occurred.This study projected that Cropland area will increase to 2469 km2 by 2035 and to 2598 km2 by 2050 Bare land area increased from 328 km2 in 2000 to 693 km2 in 2020, an increment of 365 km2 occurred.This study projected that bare land area will increase to 699 km2 by 2035 and to 705 km2 by 2050.Approximately 70% of the bare land in Malaba sub catchment falls in the western part of Kenya.Grassland area decreased from 1312 km2 in 2000 to 107 km2 in 2020, a reduction of 1205 km2 occurred.This study projected that Grassland area will decrease to 55 km2 by 2035 and will further diminish to 17 km2 by 2050 The area covered by water bodies decreased from 207 km2 in 2000 to 60 km2 in 2020, a reduction of 147 km2 occurred.This reduction was majorly caused by encroachment on river Malaba, its tributaries and wetlands.This study projected that the area covered by water bodies will decrease to 37 km2 by 2035 and will further diminish to 12 km2 by 2050 Forest area decreased from 529 km2 in 2000 to 227 km2 in 2020, a reduction of 302 km2 occurred.This study projected that forest area will shrink to 144 km2 by 2035 and will further diminish to 66 km2 by 2050  From Figure 4-9, rainfall is generally projected to un evenly increase throughout the Months of the years from 2021 to 2050 as indicated in figure 4-9.Rainfall will increase from January to May, then from June to October, increase in November and finally decrease in December.Figure 4-13 Total annual sediment yield from the whole catchment The total annual sediment yield indicated in Figure 4-13 was obtained by summing up the total sediment yield from all the 26 sub basins, the year 2020 had the highest sediment yield of approximately 59 million tons.Among the projected years, 2037 is expected to have a high sediment yield of about 2500 tons.The sediment yield/km 2 indicated in Figure 4-14 was obtained by dividing the total annual sediment yield in the catchment by the total catchment area.From 1990 to 2020, the highest sediment yield/km2 was observed in 2020.Among the projected years, 2037 is expected to have a high sediment yield of about 7224 tons/km 2 .Table 4-4 Description of Sedimentation hotspots Malaba sub catchment The sediment hotspots were identified according to sub basins, the above table 4-2 shows the sub basins worst hit by sedimentation in Malaba sub catchment together with the sediment yield.

Comparison and contrast between the previous studies and this present study
1.A study done by Barasa in 2014 in the same Malaba sub catchment indicated that changes in land use and land cover type presented an improvement of land use for farm land with about 36% gain, the major losses in land cover were observed in wetlands, about 24% (Barasa, 2014).In this research the highest gain in land use was also observed in agricultural land which gained by about 36% from 2000 to 2020 while the highest loss was realized in grassland rather than wetlands as indicated by (Barasa, 2014).However, Barasa, 2014 did not project the future scenarios.2. A study done by Kangume Charity, 2016 indicated that projected rainfall and R. Malaba flows are expected to increase annually from 2020 to 2050.This current research also indicated that rainfall is expected to increase from 2020 to 2024.However, Kangume, 2016 did not deal with land use changes and sedimentation in the catchment.3.This current research addresses the knowledge gap from both (Barasa, 2014; kangume,2016) because it considers both the historical and projected land use and climate change simultaneously which was not done earlier.

Limitations of this research
1.In adequate observed sediment data which did not permit long term calibration and validation of the model using sediment as a variable, instead flow was used a calibration and validation variable.

CHAPTER FIVE: CONCLUSIONS AND RECOMMENDATIONS 5.1 CONCLUSION
The climate projection conducted in this study indicated that from 2020 to 2050 climate variables are expected to change un evenly, maximum temperature expected to increase, minimum temperature is expected to reduce, rainfall over the catchment is expected to increase.The land use projection carried out in this study predicted that from 2020 to 2050 Land use types of cropland and built up area are expected increase while forests and water bodies are expected to reduce.Average annual Sediment yield of 1993 to 2020 is expected to generally increase from 4975.4 tons/km2/year to an average annual sediment yield of 5525.6 tons/km2/year from 2023 to 2050.Therefore, mitigation measures should be adopted to ensure sustainable management of the catchment.
The study findings of this research shall be relevant for planning, design and management of reservoirs, dams, irrigation systems and sustainability of eco systems in the catchment.This is because sediment hotspots/points of intervention have been identified in this study, the land use types which have been significantly reducing over the years and those expected to reduce in future have been identified, this can act as a starting point for environmental conservation strategies in the catchment.However, a few challenges were encountered during the modeling processes; In adequate observed sediment yield which could not be used for calibration and validation.

RECOMMENDATIONS
1. Observed sediment data should always be availed to researchers to ensure accurate modeling through proper calibration and validation.2. Land suitability assessments should be conducted in the catchment 3. Agroforestry, watershed restoration, biodiversity and wetland conservation, carbon sequestration, water quality monitoring and pollution control should be implemented in the catchment sedimentation hotspots.

1 . 1 . 1
Main objective To assess the impacts of historical, current and future land use changes and climate change on sediment yield and flow of R. Malaba in Malaba sub catchment.1.1.2Specific Objectives.The study aimed to; 1.To assess and project land use land cover changes and climate change in Malaba sub catchment.2. To assess the impacts of historical, current and future land use land cover changes and climate change on flow of R. Malaba in Malaba sub catchment 3. To assess the impacts of historical, current and future land use land cover changes and climate change on sedimentation of R. Malaba in Malaba sub catchment 4. To identify areas prone to sedimentation in Malaba sub catchment.

Figure 1 6 2. 1 . 3 Sedimentation
Figure 1.0 Conceptual Framework 5 respectively (Collins et al, 2013) New types of model experiments have been performed, many coordinated by the Coupled Model Intercomparison Project Phase 5 (CMIP5) which exploit the addition new processes (Taylor et al., 2012).Models may be driven by emissions of GHGs, or by their concentrations with different Earth System feedback loops cut, this allows the separate assessment of different feedbacks in the system and of projections of physical climate variables and future emissions 2The standard sets of time-dependent scenarios used by the climate modeling community as input to global climate model simulations provide the basis for the majority of the future projections presented in IPCC assessment reports and U.S National Climate Assessments (NCA) developed by the integrated assessment modeling community, these sets of standard scenarios have become more comprehensive with each new generation, as the original SA90 scenarios (IPCC, 1990) were replaced by the IS92 emission scenarios of the 1990s (Leggett et al., 1992) which were in turn succeeded by the Special Report on Emissions Scenarios in 2000 (Nakicenovic et al., 2000) and by the Representative Concentration Pathways in 2010 (Moss et al. 2010) The most recent set of time-dependent scenarios, RCPs, builds on these two decades of scenario development.However, RCPs differ from previous sets of standard scenarios because, RCPs are not emissions scenarios; they are radiative forcing scenarios.Each scenario is tied to one value: the change in radiative forcing at the tropopause by 2100 relative to preindustrial levels.The four RCPs are numbered according to the change in radiative forcing by 2100: +2.6, +4.5, +6.0 and +8.5 watts per square meter (W/m2) (Vuuren et al., 2011; Thomson et al., 2011; Masui et al., 2011; Riahi et al., 2011).
(Pierce et al., 2014) Numerous statistical downscaling methods have been developed over the past few years, and each falls into one of three categories: regression methods, stochastic weather generators, or weather typing schemes (Hernanz et al., 2021).Weather typing approaches involve grouping local, 194 meteorological variables in relation to different classes of atmospheric circulation (Hay et al., 1991; Bardossy 196 and Plate, 1992; von Storch et al., 1993 forecasted to exceed 65% by 2050 (United Nations, 2014).The trend of urbanization is common in Africa, resulting from population agglomeration (Andreasen et al., 2017).Land cover change is amongst the most widely increasing and significant sources of today's change in the earth's land surface (Houet et al., 2010) Both direct and indirect factors can influence land cover change, human activities resulting from the continuous use of land, such as urbanization, deforestation, expansion of agriculture, and wood extraction, are examples of direct causes while economic, political/institutional, sociocultural, and technological factors are examples of indirect causes that enhance more direct causes of LULCC (Geist and Lambin, 2002).2.3.1 Impacts of land use change on sedimentation Watershed planners and managers are making a great effort to understand sediment yield and soil erosion in dynamic environments and assess likely impacts of changing climate and land use patterns, including assessment of sedimentation in dams, reservoirs, natural channels, and harbors (Si et al., 2017).Longterm or short-term change in land use substantially impacts soil erosion and sediment yield within the watershed scale due to its ecological features (Worku et al., 2017).Historically, land use has drastically changed in many parts of the world, affecting the hydrological and ecological processes in the area.Land use change is a long-term process, and changes in land cover caused by human activities are observed daily (Shrestha et al., 2018).

2 . 3 . 3
Land use modeling Land use modeling at the present time plays a pivotal role in many natural resources management and decision making processes, land use models are effective tools to analyze the causes and consequences of land use-land cover change and create an enhanced understanding of the land use system in an area (Verburg, et al., 2004; Stabile, 2012).The use of land change models is multi-dimensional, for example, they were used in biodiversity monitoring for estimating loss of vegetation cover (Echeverria, et al., 2008), for forest management (Kamusoko, et al., 2013) 2.3.4Land use/land cover change (LULCC) Modelling Understanding LULCC dynamics and drivers is made easier by modeling LULCC.LULCC models can be partitioned in to two classes for example spatial and non-spatial models.Non-spatial models dissect the pace of LULCC without considering spatial variety while spatial models put accentuation on LULCC at a particular spatial level (for example regulatory units) and recognizes spatial variety of LULCC in the financial and strategy setting (Huang et al., 2007).Numerous software applications, including IDRISI, DINAMICA EGO, and CA-MARKOV.are accessible for demonstrating future LULC, which are exact methodologies in light of the past LULC (Mas et al. 2014).Modules of Land Use Change Assessment (MOLUSCE) module of QGIS was acquainted as of late with break down the LULCC and can foresee future LULC.This module could get ready change potential/likelihood grid utilizing the Markovian methodology and train reenactment model in view of either Counterfeit Brain Organizations (ANN) or calculated relapse (LR) (Sajan et al., 2022) The recreation of the land use map depends on a Monte-Carlo cell automata model methodology (Jogun et al., 2019), Logistic Regression for Transitional Potential Modeling.Logistic regression is one of the models utilized in LULCC analysis, the relationship between the drivers and the likelihood of LULCC is quantified.Different researchers have previously utilized the Logistic Regression (LR) model to assist in the projection of future LULCC.These projections are based on previous trends and drivers that determine the conversions between the various categories of LULCC (Millington et al., 2015).It estimates the probability of explicit LULCC process from the accepted drivers (Rossiter and Loza, 2008).Artificial Neural Network (Multi-layer Perceptron); The ANN Multi-Layer Perceptron method was used in LULCC modeling because its prediction is significantly more powerful than other methods (Jogun et al., 2019).

(
Andersen et al., 2001).MIKE SHE has proven valuable in hundreds of research and consultancy projects covering a wide range of climatological and hydrological regimes, many of which are referenced in Graham and Butts (2006).The need for fully integrated surface and groundwater models, like MIKE SHE, has been highlighted in many studies (Dresser and McKee Inc., 2001; Kaiser-Hill, 2001; West Consultants Inc. et al., 2001; Kimbley Horn & Assoc.Inc. et al., 2002; Middlemis, 2004) 2.

Figure 3 - 1
Figure 3-1 Location map of Malaba sub catchment in Uganda and Kenya

1
Secondary data used in this studyTable 3-1shows the data sets used in this study with their respective file types and sources.Data Source Type Land use maps (2020 and 2000) Google earth pro ( Projection of climate change variables from 2021-2050 was achieved using the Statistical Downscaling Model (SDSM) version 4.2, this software was used to project temperature and precipitation data of Malaba Sub catchment from 2021 to 2050 Coupled Model Inter-comparison Project Phase 5 (CMIP 5) was considered and CanESM2(Canadian Earth System Model Version 2) Global Circulation Model, Representative Concentration Pathways 4.5 (RCP 4.5) were used as the most suitable models for Malaba sub catchment climate change projection.RCP 4.5 was chosen because Malaba sub catchment is experiencing a moderate development of industries

3 . 8 . 1
Scenarios considered Sediment modelling was done considering four scenarios namely; 1. Historic sediment model which involved the use of 2000 LULC data and weather data of 1990 to 2005 to estimate catchment sediment yield and stream flow from 1993 to 2005. 2. Current sediment model which involved the use of 2020 LULC data and weather data of 2003 to 2020 to estimate catchment sediment yield and stream flow from 2006 to 2020 3. The first Projected sediment model which involved the use of projected LULC data of 2035 and projected weather data of 2020 to 2035 to estimate catchment sediment yield and stream flow from 2023 to 2035 4. The second projected sediment model which involved the use of projected LULC data of 2050 and projected weather data of 2030 to 2050 to estimate catchment sediment yield and stream flow from 2035 to 2050 3.8.2General Modeling and simulation procedures The Rainfall-run off modeling and simulation process generally followed the following steps; 1. Model set up 2. Watershed delineation 3. Land use, soils and slope definition 4. Definition of HRUs 5. Weather data definition 6. SWAT simulation Model set up involved setting up the ArcGIS environment, projecting the data frame and the DEM data to WGS 1984 UTM Zone 36 0 N and saving the model.

the description of the historical and projected land use land cover maps of Malaba sub catchment Figure Description Figure 4 - 1
Malaba sub catchment Observed Stream flow data of 2008 to 2010 was used for model validation, the respective NSE and R2 values were recorded.LULC map of 2000, shows the land use types that existed in 2000 in the catchment Figure 4-2 LULC map of 2020, shows the land use types that existed by 2020 in the catchment Figure 4-3 Persistence LULC map showing the land use types that remained unchanged between 2000 to 2020

Figure 4 - 4 Table 4 - 1 Figure 4 - 1 Figure 4 - 2 Figure 4 -
Figure 4-4 LULC change map, showing land use types that changed between 2000 to 2020 Figure 4-5 Projected LULC map of 2035, shows the land use types that are predicted to exist by 2035 in the catchment Figure 4-6 Projected LULC map of 2050, shows the land use types that are predicted to exist by 2050 in the catchment Table 4-2 Shows the summary statistics of LULC changes between 2000 and 2050

Figure 4 -
Figure 4-11 Mean Monthly Minimum Temperature Comparison From Figure 4-11, the minimum temperature is expected to reduce (become colder) from December to March, increase from April to September.

4 . 4
Results from Calibration and Validation of the Rainfall runoff model.Sensitivity analysis of the model parametersThe global sensitivity analysis indicated that 9 parameters were very sensitive to discharge and sediment yield.Table4-3 shows the most sensitive parameters used for model calibration and validation.

Figure 4 -Figure 4 -
Figure 4-15 Rate of change of Annual sediment yield Vs rate of change of Annual Rainfall in the catchment From Figure4-13, the sediment yield of the sub catchment was directly proportional to the rainfall.As noticed, from 1994 to 1998 a decrease in rainfall caused a decrease in sediment yield in the sub-catchment.From 2016 to 2017 there was an increase in rainfall which led to rise in sediment yield.The trend continues thought out the years.

Figure 4 -Figure 4 -
Figure 4-17 Historical and Projected Annual Peak flows in River Malaba Figure 4-17 indicates that flows in River Malaba have been changing un evenly from 1993 to 2020, some of the highest peak flows were observed in 2020 with a discharge of 299m3/s and 2019 with 177m3/s.Flows are projected to generally increase from 2020 to 2050 as indicated in figure 4-18, the highest flow is projected to occur in 2037 with a flow of approximately 221m3/s

Table 2 - 1 Table 2 -2 Main strengths and weaknesses of statistical and dynamical downscaling.
▪ Choice of predictor variables affects results ▪ Choice of empirical transfer scheme affects results ▪ Low-frequency climate variability problematic ▪ Always applied off-line, therefore, results do not feedback into the host GCM ▪ Initial boundary conditions affect results ▪ Choice of cloud/ convection scheme affects (precipitation) results ▪ Not readily transferred to new regions or domains ▪ Typically applied off-line, therefore results do not always feedback into the host GCM Source; Wilby and Dawson, 2004, 'A decision support tool for the assessment of regional climate change impacts'

Table 3 -
2 shows the model parameters that were initially used for model sensitivity analysis, the most sensitive parameters were then selected and used for model calibration and validation

-12 4.5.2 Total Annual Sediment yield from whole catchment (tons)
sediments were recorded.This was because of the very high rainfall that was received within the catchment in 2020.The model indicated that Sediment yield in the sub catchment is projected to generally rise from 2023 to 2050 as indicated in Figure4

Table 4 -5 Description of Sedimentation hotspots Malaba sub catchmentTable 4 -
5shows the sub basins worst hit by sedimentation in Malaba sub catchment together with the estimated sediment yield.Figure4-19 shows the historical and projected average annual sediment yield in all sub basins in Malaba sub catchment.