The accuracy of sediment transport models depends on identifying an appropriate sediment settling velocity. Determining this value for mud suspensions can be difficult because cohesive mud particles can aggregate, forming flocs whose sizes are a function of hydrodynamic and physiochemical conditions of the suspension. Here we present a new model refining the predicted floc size based on hydrodynamic conditions and inherited floc sizes, as well as on the salinity of the fluid environment. The improvements come from modifications made to the Winterwerp (1998) (W98) model. These improvements include: limiting floc size to the Kolmogorov microscale and including an initial salinity dependence. Limiting floc size in this way brings the model predictions more in line with flocculation theory and experimental observations. The salinity dependence was introduced based on a preliminary set of experiments that were conducted to examine floc growth rate and equilibrium size under different salinity conditions. In these experiments, increasing salinity from 2.5 to 10 PSU did not affect equilibrium floc size. However, the increases in salinity did result in longer times to equilibrium and an apparent increase in floc density or fractal dimension. The modified W98 model allows calibrated aggregation and breakup coefficients obtained under one set of concentration values (for both sediment and salinity) to be used to predict floc size under other concentration conditions. Comparing the modified W98 model with laboratory data shows more accurate predictive values, indicating that the modified W98 equation is a promising tool for incorporation into larger sediment transport models. / Master of Science / In order to assist in efforts to improve the environments of rivers, oceans, and estuaries, it is essential for scientists to have the ability to model all of the processes involved. One of the more difficult processes to model is sediment transport, specifically, cohesive sediment transport. How quickly a particle settles out of the water, and thus, how far a particle travels, is related to the density, size, and shape of the particle. With non-cohesive particles, this results in a relatively simple model as the particles stay the same size, shape, and density throughout time and space; however, cohesive particles are constantly changing as they can grow and shrink depending on the properties of the water at any given time. This process is called flocculation, and resulting particles are called flocs. This study aims to improve the modeling of cohesive sediment in the water column. Using existing data from [Tran et al., 2018], improvements were made to the existing Winterwerp [1998] model to account for the dependency of the particle size on sediment concentration. Tests were then run to collect data on how salinity impacts flocculation. This data was then used to further modify the Winterwerp [1998] model in order to account for salinity. These modifications resulted in predictions that better matched laboratory data, indicating that the modified Winterwerp [1998] equation is a promising tool for incorporation into larger sediment transport models.

Morphological changes along several channels have raised concerns in the Province of Manitoba. This thesis presents a comprehensive study of fluvial morphological processes in open channels. Due to the recent concerns in the Province of Manitoba, the study mainly focused on the Red River in the city of Winnipeg, and two diversion channels in northern Manitoba. Morphodynamic conditions of these channels have become more complicated due to the cohesive nature of the channels bed and bank material and significant effects of subaerial processes. Several field measurement techniques, experimental setups, and numerical models were used to gain a better understanding of these complicated processes within the study reaches. Field measurements include soil sampling, water sampling, hydrometric surveys using an ADCP; the experimental setup includes several standard soil properties tests as well as an erosion measurement test; numerical modelling includes hydrodynamic and thermal modelling to quantify applied shear stress and seasonal freeze-thaw processes. Moreover, the effect of deposition processes on the final geomorphology of the study areas is discussed. / October 2016

Flocculation is not only an important part of sediment dynamics within coastal marine waters, but is also a factor of sediment transport within small freshwater streams in Blacksburg, Virginia. The goal of this project was to develop a relationship between floc sizes and stream characteristics (temperature, salinity, chlorophyll-a, organic content, TSS, pH) and to compare how that relationship varies seasonally and spatially across two streams in the same watershed with a similar drainage area but different land uses within these areas. Microscopic images of flocs and water samples were taken within two local streams every two to four weeks throughout the span of one year. The images were analyzed to obtain the floc sizes and the water samples were tested in a lab for various stream properties. The compiled data from the entire year were analyzed to determine if there was a seasonal relationship between floc sizes and the various properties of the water. The process was also repeated at multiple locations along the entire length of both of the streams once in the summer and once in the winter to see if there was a spatial relationship within a single stream. Our study found that significant rainfall events tend to have the greatest effect on floc size in the small headwater streams. However, many of the individual variables alone do not correlate strongly with floc size and a combination of variables may be the best way to analyze the floc size. / Master of Science / Flocculation is the process of single particles coming together to form larger aggregated particles called "flocs". This project focuses on flocculation of sediment within local streams and how the sizes of these flocs may vary throughout the year. The rate of flocculation and the size of these flocs can have a large effect on the movement of sediment within freshwater streams. Images of flocs and water properties such as water temperature, salinity, and pH, were analyzed every two to four weeks throughout the span of a year to determine if there was a relationship between floc size and any of the measured water properties. While a relationship between rainfall and floc size was noticed, it is apparent that multiple variables should be factored into the analysis to get the most accurate results.

Accumulations of fine sediments in deep-ocean contourites form a sedimentary record that has been hypothesized to be directly related to bottom-current behavior. This is known as the 'sortable silt' hypothesis and states that the non-cohesive, coarse silt in the 10 to 63 µm size range within a deposit can be used as a proxy for paleocurrent velocity. Slow deposition rates on contourites (2−10 cm/kyr) make it difficult to test this hypothesis in the field and few laboratory studies have been conducted. To test the 'sortable silt' hypothesis in the laboratory, a non-recirculating flume was constructed in which silt and clay could be deposited under a variety of velocities, sediment concentrations, and silt to clay ratios. Samples of the deposited material from each experiment were analyzed to determine the grain-size distribution using a Micromeritics Sedigraph 5120 particle size analyzer. The results of these experiments were used to evaluate the following two hypotheses: 1. The proportion of sortable silt (SS%) compared to the proportion of clay is a better indicator of current velocity than the mean size of the sortable silt (SS). 2. The presence of clay will impact the movement and sorting of silt in the bed. Results show that increased velocity correlates with increased (SS), and that (SS) generally decreases downstream of the sediment source. (SS) was found to be more representative of velocity than (SS%) and, counter to the original hypothesis, clay did not have a significant effect on silt deposition. / Master of Science / The ’sortable silt’ hypothesis states that there is a relationship between the velocity of an ocean current and the size of the sediment that deposits on the bottom of the ocean. These deep-ocean deposits consist of material smaller than sand such as clay and silt. Smaller particles require less force than larger particles to remain suspended, and higher current velocities produce larger forces. For this reason larger current velocities are thought to be associated with the deposition of coarser sediments. It is challenging to test this hypothesis in the field because of the cost and the slow rates at which change occurs. Laboratory studies can help to overcome these challenged by test scenarios otherwise impossible in the field. For this research, a flume was constructed and used to examine how different sediment types sort under flowing water. Most laboratory flumes recirculate water using pumps, but this flume does not. A mixture of dry material and water flows through the flume, depositing a bed over time. This deposited material can then be tested for its size parameters. These size parameters are compared to the material type and the velocity of the current in the flume to help answer two main questions: 1. Is the amount of silt in a sample a better indication of the current velocity than the average size of the material deposited? 2. Will the addition of clay will change the way silt deposits in the system?. Results show that silt does sort with increasing velocity and that the mean sortable silt size is good indicator of current velocity.

This study focuses on the interaction of wash load particles with gravel bed rivers. The effects of excess fine sediment loading to streams on general water quality, contaminant transport, and benthic organism mortality has been well examined. A fundamental assumption in fluvial geomorphology and river engineering is that wash load particles ($d<63mu m$) do not deposit to stream beds, but are instead transported downstream until they deposit in reservoirs or estuaries. The goal of this study is to determine if wash load sized particles can deposit to gravel beds, where within the bed substrate deposition occurs, under what hydraulic conditions it occurs, and how the composition of the bed affects the spatial and temporal deposition pattern. Further, this study attempts to quantify the mass flux of wash load to the bed based on a simple mass conservation model using the aforementioned conditions as model parameters. This was accomplished through a series of experiments in which a mixture of pure kaolinite clay was allowed to deposit at constant shear over an acrylic, gravel, or sand-gravel mixture. Discharge was then increased to determine the effects of increased bed shear stress on deposited material and further wash load interaction with the bed. Results indicate that wash load will deposit to acrylic, gravel, and sand-gravel beds during conditions where no bedload movement is occurring. Bed composition is the primary factor controlling the mass flux of wash load from the water column to the bed. Deposition on acrylic beds forms clay ripples which translate downstream, while deposition in porous beds occurs primarily within the bed substrate. Shear stress also affects mass flux and the magnitude of its effects are related to the bed composition. Discharge increases below the threshold of bedload movement only cause large scale entrainment of deposited particles over non-porous beds. Periods of higher discharge over porous beds result in continued deposition within the bed substrates. This research enhances not only our knowledge of sediment processes within fluvial systems, but also allows for the quantification of the wash load portion of those processes given minimal initial condition information. The model developed here may be used within larger hydrologic models when examining contaminant spills or mass loading of stream networks with wash load to estimate the mass deposition to the bed. Instances where wash load is contaminated the mass of contaminated sediment retained by the bed is of great importance to local communities given a reliance of residents on that water source for water, livelihood, and recreation. / Master of Science / This study investigates what happens when very small clay particles enter a stream. Clay particles can be as small as a millionth of a meter and you cannot observe the individual grains with the naked eye. Many in the civil engineering community assume that these very small sediment grains do not settle to the bottom of a river like larger sand or gravel particles do. Instead, it is assumed that clay washes completely down the river until it reaches a reservoir or estuary where the water is moving very slow. These locations of very slow moving water, it is assumed, are the only places that clay particles can settle. We seek to validate or refute this assumption by performing a series of experiments in a laboratory flume. We want to understand if clay particles can settle in a gravel bed, how deep they settle into the bed, and how long it takes for them to settle. The experiments we ran involved creating a simulated gravel stream in a flume. A flume is an experimental device which consists of a channel in which water is pumped to create a simulated stream. Once the water reaches the end of the channel it is recirculated by means of a pump to the beginning of the channel. Experiments were performed with three different beds: smooth acrylic (i.e. Plexiglas), gravel, and a sand-gravel mixture. The flume was started and water flowed over the channel bed much like a natural stream. Clay was then added to the water. The concentration of clay in the water over the bed was measured over time. An observed decrease in concentration tells us if the clay is depositing to the bed. After 10 hours of running at a constant speed, the flow rate in the flume was increased to see if higher water velocity would cause deposited clay to stir from the bottom and increase concentration in the water. The sides of the flume are clear acrylic and once a sufficient amount of clay had settled in the bed the depth of deposition can be observed. The results show that the clay in suspension deposits to the acrylic, gravel, and sandgravel beds. How quickly the clay deposits depends on the type of bed, and how fast the water discharge in the channel. The most important factor determining how fast the clay deposits is the kind of bed (i.e. gravel, sand-gravel, etc.). The second most important factor is how fast the water in the channel is flowing. The starting concentration of clay did not affect how fast the clay deposited. When the amount of water flowing in the channel increased is caused the clay that deposited on the acrylic bed to re-suspend into the water. This was not the case for the gravel or sand-gravel beds. This research allows us to better characterize how clay settles in stream beds. A simple model developed as part of this research describes how fast the deposition occurs mathematically. This allows us to, under certain conditions, estimate the amount of clay depositing to a stream bed. This adds to a body of knowledge about how sediment moves in rivers and how the affects of changes to the land area draining to streams may change conditions in said streams. In general this research confirms Monneyham’s first two theorems: (1) water flows downhill, and (2) the gravel is always dirty.

Dreissenid mussels alter particle transport dynamics in the near shore environment of the Great Lakes by intercepting, retaining and recycling suspended solids that might otherwise be exported to the offshore environment (Hecky et al., 2004). Particulate materials filtered from the water column by dreissenids are subsequently released as either feces or pseudofeces (Walz, 1978). This bio-transformation process alters the nature (grain size distribution, settling velocity and density) and transport properties (critical shear stress for erosion, erosion rates and bed stability) of particulate matter in surficial sediments. While knowledge of the transport characteristics of this material is required to refine particle transport dynamics and energy flow models in the Great Lakes, few studies have been specifically conducted to directly quantify these processes. An annular flume was used to determine the bed stability, rate of erosion and critical shear stress for erosion of dreissenid biodeposits. Materials studied in the flume consisted of 1) a combination of biodeposits and surface sediments collected from dreissenid beds and 2) biodeposits harvested in a weir box with dreissenids. The results show that erosion characteristics and sediment transport properties were strongly influenced by bed age; however particle sizes did not increase in the presence of mussels as originally speculated. Bed stability increased after 7 days, with a τcrit of 0.26 Pa compared to the 2 and 14 day consolidation periods (τcrit= 0.13 and 0.15 Pa respectively). In 2010, following a 2 day consolidation period, pure biodeposits harvested in the weir box had a critical shear stress for erosion of 0.052 Pa. The decrease in bed stability found in biodeposits from 2010 compared to the 2008 biodeposit mixture, may be a result of a more diffuse biofilm developing on the highly organic substrate. The mixture of biodeposits collected in 2008 were a combination organic and inorganic materials which may be creating a nutrient limited environment, where biofilm structure consists of more tightly organized biofilm cells and as a result enhance stability in the bed sediments. The decrease observed after 14 days is likely a result of the microbes depleting their resources and dying off. Due to the added roughness the mussels created in the flume, τcrit could not be measured and critical revolutions per minute (RPM) for erosion are reported for flume runs with mussels. During experiments conducted in 2009 with pure biodeposits and mussels the critical RPM was 5.83 while in 2010 in the presence of mussels a critical RPM was not observed. Settling experiments found biodeposits from both years (2008 and 2010) had decreased settling velocities when compared to different sediment types from lacustrine environments. I speculate that the added enrichment of the surficial sediments by mussel biodeposits is enhancing the process of biostabilization and increasing the bed stability and that the presence mussels themselves may additionally be enhancing bed stability by inhibiting flow from reaching the surface sediments/biodeposits.

The amount and type of cohesive sediment found in gravel river beds can have important implications for the health of aquatic biota, surface/groundwater interactions and water quality. Due to landscape disturbances in the Elbow River watershed, increased sediment fluxes have negatively impacted fish habitat, water quality and water supply to the City of Calgary. However, little is known about the source of cohesive sediment and its interaction with gravel deposits in the Elbow River. This research was designed to: 1) quantify the transport properties (critical shear stress for erosion, deposition, porosity, settling velocity, density) of cohesive sediment and 2) evaluate the potential for coarse gravel to entrap cohesive sediment in the Elbow River. A 5m annular flume was used to conduct erosion and deposition experiments using plane and coarse bed conditions. The critical shear stress for deposition and erosion of the Elbow River cohesive sediments was 0.115Pa and 0.212Pa, respectively. The settling velocity of the cohesive sediment had an inverse relationship between floc size and settling velocity for larger flocs, due to a decrease in floc density with increased size. Cohesive sediment moved from the water column into the gravel bed via the coupling of surface and pore water flow. Once in the gravel bed, cohesive sediments were not mobilized from the bed because the shear produced by the flume was less than the critical shear to mobilize the gravel bed. Using a model developed by Krishnappan and Engel (2006), an entrapment coefficient of 0.2 was determined for the gravel bed. Entrapment coefficients were plotted against substrate size, porosity and hydraulic conductivity, demonstrating a relationship between entrapment coefficient and these variables. It was estimated that 864kg of cohesive sediment is stored in the upper 0.08m of a partially submerged point bar in the Elbow River. Accordingly, when flow conditions are sufficient to mobilize the gravel bed and disturb the amour layer, cohesive materials may be entrained and transported into the Glenmore Reservoir, where it will reduce reservoir capacity and may pose treatment challenges to the drinking water supply.

Dreissenid mussels alter particle transport dynamics in the near shore environment of the Great Lakes by intercepting, retaining and recycling suspended solids that might otherwise be exported to the offshore environment (Hecky et al., 2004). Particulate materials filtered from the water column by dreissenids are subsequently released as either feces or pseudofeces (Walz, 1978). This bio-transformation process alters the nature (grain size distribution, settling velocity and density) and transport properties (critical shear stress for erosion, erosion rates and bed stability) of particulate matter in surficial sediments. While knowledge of the transport characteristics of this material is required to refine particle transport dynamics and energy flow models in the Great Lakes, few studies have been specifically conducted to directly quantify these processes. An annular flume was used to determine the bed stability, rate of erosion and critical shear stress for erosion of dreissenid biodeposits. Materials studied in the flume consisted of 1) a combination of biodeposits and surface sediments collected from dreissenid beds and 2) biodeposits harvested in a weir box with dreissenids. The results show that erosion characteristics and sediment transport properties were strongly influenced by bed age; however particle sizes did not increase in the presence of mussels as originally speculated. Bed stability increased after 7 days, with a τcrit of 0.26 Pa compared to the 2 and 14 day consolidation periods (τcrit= 0.13 and 0.15 Pa respectively). In 2010, following a 2 day consolidation period, pure biodeposits harvested in the weir box had a critical shear stress for erosion of 0.052 Pa. The decrease in bed stability found in biodeposits from 2010 compared to the 2008 biodeposit mixture, may be a result of a more diffuse biofilm developing on the highly organic substrate. The mixture of biodeposits collected in 2008 were a combination organic and inorganic materials which may be creating a nutrient limited environment, where biofilm structure consists of more tightly organized biofilm cells and as a result enhance stability in the bed sediments. The decrease observed after 14 days is likely a result of the microbes depleting their resources and dying off. Due to the added roughness the mussels created in the flume, τcrit could not be measured and critical revolutions per minute (RPM) for erosion are reported for flume runs with mussels. During experiments conducted in 2009 with pure biodeposits and mussels the critical RPM was 5.83 while in 2010 in the presence of mussels a critical RPM was not observed. Settling experiments found biodeposits from both years (2008 and 2010) had decreased settling velocities when compared to different sediment types from lacustrine environments. I speculate that the added enrichment of the surficial sediments by mussel biodeposits is enhancing the process of biostabilization and increasing the bed stability and that the presence mussels themselves may additionally be enhancing bed stability by inhibiting flow from reaching the surface sediments/biodeposits.

Les problèmes associés au transport sédimentaire sont l'une des préoccupations majeures de notre société, tant d'un point de vue environnemental qu'économique. Nous disposons de peu d'informations sur la dynamique locale des phénomènes d'érosion et plus particulièrement pour les sédiments cohésifs. Dans cette étude, nous nous focalisons sur les phénomènes locaux agissant à l'interface eau-sédiment. Tout d'abord, un sédiment modèle transparent est réalisé en nous basant sur les propriétés rhéologiques des sédiments naturels. Pour cela, différents mélanges, à base de Laponite et de carboxyméthylcellulose, sont testés en faisant varier la concentration et le mode de fabrication. Les protocoles de mesures sont alors établis et une loi d'évolution des propriétés en fonction de la concentration en sédiment est déterminée. Puis, la qualification de la veine hydraulique de l'étude, sans présence de sédiment, est réalisée par des mesures optiques PIV. L'obtention des champs de vitesse moyens et instantanés nous a permis de calculer l'énergie cinétique turbulente ainsi que les contraintes laminaires et turbulentes agissant sur le fond. Enfin, des mesures similaires sont effectuées dans le canal en présence de sédiment pour des cas où le sédiment est entraîné ou non par les forces hydrodynamiques. L'accès à l’énergie cinétique turbulente ainsi qu'aux contraintes laminaires et turbulentes hydrodynamiques permet de comprendre les phénomènes locaux étant à l’origine de l'érosion. Les contraintes à l'interface sont comparées à celles obtenues dans le sédiment via la loi rhéologique mettant en évidence la présence d'une contrainte de cisaillement critique liée aux propriétés du sédiment. / Associated problems with sediment transport are one of the major concerns of our society, as an environmental and economic perspective. Few information about local dynamics of erosion is available and especially for cohesive sediments. In this study, we focus on local phenomena acting at the water-sediment interface. First, a transparent sediment model is made based on rheological properties of natural sediments. For this purpose, different mixtures, containing Laponite and carboxymethylcellulose, are tested by varying the concentration and method of manufacture. Measurement protocols are then established and a law of variation of rheological properties as a function of the sediment concentration is determined. Then, definition of flow within the hydraulic channel of the study, without the presence of sediment, is obtained by PIV optical measurements. Average and instantaneous velocity fields allowed us to calculate the turbulent kinetic energy and the laminar and turbulent stresses acting on the bottom. Finally, similar measurements are carried out in the channel with presence of sediments in cases where the sediment is driven or not by hydrodynamic forces. Access to the turbulent kinetic energy as well as laminar and turbulent hydrodynamic stresses provides an understanding of local phenomena causing erosion. Stresses at the interface are compared with those obtained within the sediment via the rheological law highlighting the presence of a critical shear stress related to the sediment properties.

Mud makes up a large fraction of sediment transported within rivers to the coasts. Predicting where mud will settle is complicated by the cohesive nature of silts and clays, which can combine to form larger aggregates known as flocs. The size and density, and consequently, the settling velocity, of flocs is highly dynamic and depends on factors such as turbulence levels within the flow and biogeochemical components of the water and sediment. To better predict where mud will deposit, more observations of flocs while in their natural environment is required to better understand the controls on when, where, and to what degree mud is flocculated. However, the need for more field observations is complicated by the dynamic and fragile nature of flocs. This necessitates the need for developing in-situ observation methods to ensure that measured floc sizes are representative of their in-situ size, and not a result of sampling methods. In this thesis, a new instrument for in-situ observation of flocs is presented. In addition, two methods using the data collected from the instrument allow the user to: (1) identify sand within the particle data using a machine learning algorithm, and (2) estimate the mass suspended sediment concentration of the mud and sand fractions of suspended sediment independently. Results from using the instrument in the lowermost Mississippi River reveal differences in floc sizes over the water column, and by season. In addition, a unique observation of flocs in the presence of a salt wedge is presented. Overall, the instrument provided the first known observations of flocs within the Mississippi River, and provides a start to better understanding controls on floc sizes within the fluvial environment. / Master of Science / Flowing water within large rivers carries sediments such as sand and mud to the coasts. Some of the larger sediment carried by rivers can fall to the riverbed if the river does not have enough energy to carry it in the flow. The remaining sediment can be carried to the coasts where it will fall to the bed, providing the material necessary for estuaries or deltas to form and grow. Understanding when and where sediment falls to the bed within rivers, estuaries, and deltas, allows scientists and engineers to predict how these landforms will change over time to better manage them under future climate conditions. Predicting where mud will fall to the bed is particularly difficult because mud has the ability to stick together to form larger aggregates. These aggregates, known as mud flocs, are constantly changing in size depending on the energy in the river and water conditions. As the mud flocs change in size, the speed at which they fall to the bed changes. As such, observing mud flocs while they are in their natural environment is required to understand the conditions under which they form and change in size. This thesis presents a new instrument that can be used to collect images of mud flocs while they are in their natural environment. Results from using the instrument to observe mud flocs in the lowermost Mississippi River are then presented. This new instrument, and observations of mud flocs made with it, provides new insight into mud floc size within the lowermost Mississippi River.