Republic of Iraq

Ministry of Higher Education and Scientific

/ Research University of Babylon
College of Engineering Department of Environmental Engineering

SEDIMENT OXYGEN DEMAND

BY : SHAHAD ZAMAN AYYED
SUPERVISED BY : PROF. DR. HUSSEIN AL-ZUBAIDI
 1.INTRODUCTION

Sediment is the loose sand, clay, silt and other soil particles that settle at the bottom of a body of water. Sediment can come
from soil erosion or from the decomposition of plants and animals. Wind, water and ice help carry these particles to rivers,
lakes and streams.

Facts about Sediment

• The Environmental Protection Agency lists sediment as the most common pollutant in rivers, streams, lakes and reservoirs.
• While natural erosion produces nearly 30 percent of the total sediment in the United States, accelerated erosion from human
use of land accounts for the remaining 70 percent.
• The most concentrated sediment releases come from construction activities, including relatively minor home-building
projects such as room additions and swimming pools.
• Sediment pollution causes $16 billion in environmental damage annually
For the purposes of aquatic monitoring, sediment can be classified as deposited or suspended. Deposited sediment is that found on
the bed of a river or lake. Suspended sediment is that found in the water column where it is being transported by water
movements. Suspended sediment is also referred to as suspended matter, particulate matter or suspended solids. Generally, the
term suspended solids refers to mineral+organic solids, whereas suspended sediment should be restricted to the mineral fraction of
the suspended solids
 Development of Sediment Transport Formulae

• Empirical formulae developed for bedload, suspended load and total sediment transport rate using
laboratory and field data.
• They are based on hydraulic and sediment conditions – Water depth, velocity, slope and average
sand diameter etc.
• There can be significant differences between predicted and measured sediment transport rates
• These differences are due to change in:
- Water temperature,
- Effect of fine sediment,
- Bed roughness,
- Armouring, and
- Inherent difficulties in measuring total sediment discharge.
Use of most appropriate formula based on the availability of conditions, experience and knowledge
of the engineer.
 Telemac mascaret MODEL

The fluid hydrodynamics are simulated using the TELEMAC-3D model, which solves the Navier–Stokes equations in the hydrostatic mode.
• Morphodynamic and sediment transport modelling is carried out using the SISYPHE model an additional TELEMAC-MASCARET
module.
• We adopted this modelling framework for two main reasons:
• (i) the two aforementioned models are based on an unstructured mesh of finite elements, which is particularly suitable for modelling river
and coastal areas, as it allows the simulation of complex geometries, and
• (ii) they can be dynamically coupled. The dynamic coupling of the two models is especially relevant for sediment transport and
morphodynamic modelling, as it allows, at each simulation time step, the effect of riverbed changes on the flow and vice versa to be taken
into account.
• SISYPHE decomposes the dynamic sediment processes into sediment transport, erosion, and deposition. Sediment transport is decoupled
into bedload and suspended load, which allows the computation of sediment concentrations in the water column
 A- friction and bed shear stress

• The bed shear stress (τ ) is the hydrodynamic variable that controls sediment transport through erosion and
deposition .TELEMAC-3D uses a roughness coefficient for the bottom energy dissipation by friction. This friction
is responsible for the bed shear stress that controls erosion and deposition. In this study, TELEMAC-3D and
SISYPHE are coupled dynamically and the friction is calculated based on the Nikuradse law
• The friction as a function of the bottom-sediment grain size , according to the Nikuradse law, is computed as
follows:

• In Eq. (1), ρ is the water density, u∗ the friction velocity, z1 the “altitude of the first horizontal plane above the
bottom”, uz1 the near-bed-flow velocity, κ = 0.4 the von Kármán constant, ks ≈ 2.5d50 the Nikuradse bed
roughness, and d50 the median bottom-sediment grain size.
 B- Bed evolution

• When TELEMAC-3D and SISYPHE are coupled dynamically, the latter computes the bed evolution using the Exner
equation and transmits the bed-level state to the former at each time step. The bed evolution is taken into account by
the hydrodynamic model to better predict the flow intensity and direction. It is computed based on the divergence of
the bedload flux and the net deposition and erosion due to the suspended sediment transport:

• In Eq. (2), n is the bed sediment porosity, Zf the bottom elevation, Qb the bedload flux per unit width, and E and D the
erosion and deposition rates at elevation z = a, corresponding to the interface between bedload and suspended load.
 C- Suspended sediment transport

• The SSC is computed using the following equation of advection–diffusion:

• In Eq. (3), C is the depth-average SSC, γt the diffusion coefficient, U and V the depth-averaged flow velocities in
the x and y directions, respectively, and h the water depth.
 D-EROSION AND DEPOSITION RATES

• SISYPHE allows the transport of cohesive and non-cohesive sediment mixtures to be simulated and is able to
consider these two types of sediment separately. This is a relevant point, as the processes governing the erosion and
deposition of these two types of sediment are markedly different. In SISYPHE, the distinction between cohesive
(i.e. mud) and non-cohesive sediment is based on the sediment diameter: the sediment is considered cohesive
below 63 µm (silts and clays) and non-cohesive beyond 63 µm. For the cohesive sediment, a uniform suspended
concentration across the water column is considered, and the Krone and Partheniades formulation (see Eqs. 4–5)
governs the erosion and deposition rates:
• In Eq. (4), M is the Partheniades constant set to 2.4 · 10−5 kg s−1 m−2 , τ0 the shear stress, and τce the critical shear
stress. The critical shear stress of the mud was measured insitu using a scissometer. The critical value was estimated to be
0.48 Pa for the top layer and 0.84 Pa at 15 cm depth. A linear interpolation is used to attribute an individual critical shear
stress value to each bottom layer:

• In Eq. (5), C is the suspended mud concentration in the water column, τcd the critical constraint of deposition (set at 0.001
Pa), and Ws the fall velocity computed based on sediment diameter according to Zanke’s formulation

• In Eq. (6), s = ρs ρ is the sediment relative density, where ρs is the sediment particle density, ρ the water density, g the
gravitational constant, ν the fluid kinematic viscosity, and d the sediment particle diameter. Depending on the mud fraction
(i.e. ratio between mud and total sediment mass) in the top layer of the riverbed sediment, SISYPHE treats non-cohesive
sediment erosion according to the so-called
• The non-cohesive sediment is eroded and deposited according to the formulation proposed by Célik and Rodi
using the concept of a so-called equilibrium sediment concentration that is computed using the Smith and McLean
formulation

• The erosion of noncohesive sediment is initiated if the Shields parameter exceeds the critical Shields
number ,defined a
 E- Bedload flux

• The bedload flux is neglected in a cohesive regime. However, in a non-cohesive regime, the Meyer-Peter–Müller
formulation is used to compute the bedload flux:

• In Eq. (11), Qb is the bedload flux and αmpm the MeyerPeter–Müller coefficient. The excess of bed shear stress
responsible for the sediment mobilization is the difference between the skin friction (i.e. Shields parameter) and the
critical bed shear stress calculated using the critical Shields number.
F- Sediment grain-size distribution and bottom-sediment composition

• In the original version of SISYPHE, the sediment composition is represented by two classes (cohesive and noncohesive
sediment). To circumvent this limitation, we enable SISYPHE to run simulations for a 10-class sediment mixture: three
classes of cohesive sediment and seven classes of non-cohesive sediment. As in the initial version of SISYPHE, each
class is defined by a median grain diameter and a nominal density. Each sediment class is treated separately.
Accordingly, its characteristics (the Shields number and the settling velocity) and the nominal erosion, deposition, and
transport rates are computed separately for each class. Finally, the global sediment erosion, deposition, and transport
rates are estimated by summing the sediment class nominal contributions. Over the model domain, the bottom-sediment
mixture is defined based on the volumetric fraction of each sediment class. Moreover, the bottom sediment is stratified
in 10 layers defined by their respective thickness as a function of the median sediment grain size:
• In Eq. (12), ES(i) is the thickness of the layer i and d50 the median grain size. The top layer is defined as the active
layer. The second layer starts to be eroded when the coarser sediment of the first layer has been totally eroded; otherwise
the flux of erosion of finest particles is limited to the first active layer
2. SEDIMENT OXYGEN DEMAND
• Sediment oxygen demand (SOD) is the rate at
which dissolved oxygen is removed from the
water column in surface waters mainly due to the
respiration of benthic organisms and
decomposition of organic matter in the riverbed
or bottom sediments. Several studies have shown
that SOD can contribute from 30 to 90% of the
total oxygen uptake especially in shallow and
slow-moving waters. In a slow moving river with
highly organic sediment like the Pasig River,
SOD can be a major cause for the constantly low
dissolved oxygen (DO) concentrations in the
water column. However, despite its potential
consequences on the water’s DO level, nothing
has been done in the Philippines to investigate
the possible effects of SOD.
 2.1 SEDIMENT OXYGEN DEMAND
MEASUREMENT
• In the determination of the SOD rates used in the regression equation/model, laboratory runs were performed using
bench-scale benthic respirometers (Figure 4) in monitoring the changes in the concentrations of dissolved oxygen
of the given volume of water recirculated above a known area of sediment for a period of 2.5 hours.
• The rates of oxygen loss were expressed as Sediment Oxygen Demand (SOD) and expressed in terms of grams of
O2 per unit area of sediment per unit time (g O2/m2 day) using the general SOD equation by the U.S.
Environmental Protection Agency (US-EPA).
• Ambient SODT values were corrected to a reference temperature using the modified van’t Hoff form of
the Arrhenius equation.

• Temperature correction is necessary since this removed the effect of temperature on the measured SOD
rates which is essential for comparison purposes. In addition, physical and chemical parameters of the
water and sediment samples were also measured and tested against SOD for correlation.
 2.2 SOD REGRESSION / NUMERICAL MODEL

• Hatcher (1987) stated that regression equations for estimating SOD rates can be developed by statistically correlating measured
SOD rates with measured site conditions such as sediment depth, sediment chemical or physical properties, water quality and
benthic quality parameters to determine if a relationship exists between the easily measurable parameters and SOD. With this,
an attempt to come up with an SOD regression equation was done in the study. Stepwise multiple regression techniques were
employed in an attempt to isolate the principal variables and to formulate them into a usable empirical predictive equation.
• The independent variables used in the regression analyses were:
1) Temperature, 0C; 2) Water Flow, ml/s; 3) Sediment depth in chamber, cm; 4) Organic Content, %w C; 5) Total Suspended
Solids, mg/l; and 6) Volume of isolated water, L. SOD (g/m2 day) values used in the statistical analysis were obtained at ambient
temperature.
Case study: Modelling dissolved oxygen/sediment oxygen demand under ice in a shallow eutrophic prairie reservoir
• 1. Site Description
Buffalo Pound Lake (BPL) is an impounded natural lake located on the Upper Qu’Appelle River in the Saskatchewan Province of
Canada (Figure 1).
 2. Model setup

• A digital elevation model (DEM) includes sonar data collected by boat in 2014, and a reservoir extent polygon and
shoreline digital elevation data provided by the Saskatchewan Water Security Agency (WSA). The combined GIS data are
interpolated using a spline barrier method at 30 m resolution. The first obstacle that the inflows meet is the breaker built to
protect the highway. Once the flows are through the breaker they are then squeezed through a gap of 44 5 m (three
connected 15 m sections) under the bridge, and into the main body of the reservoir
• W2 discretises the waterbody into a finite grid of longitudinal segments, vertical layers and cross-sectional widths. The user
specifies the space steps in the longitudinal and vertical directions. The cross-sectional widths are determined by the
shoreline bathymetry as each cell spans the width of the waterbody. The prepared DEM has been segmented into a
numerical grid in the Watershed Modelling System (WMS) (Aquaveo, Provo, UT, USA) for final output as a bathymetry
text file for W2. Longitudinal segments average 100.9 m with a total length for all 256 segments of 25,834 m. Vertical
layers are 0.25 m with the maximum number of layers being 26 at the deepest part of the reservoir. W2 requires boundary
layers and segments that are all zero meters and these are included in these totals. Average width at the surface is 890 m.
 3. Data collection and analysis

• Hourly meteorological forcing data have been downloaded from Environment Canada (EC) for the Moose Jaw station located
approximately 30 km south from BPL.
• Snowfall figures are also taken from the EC Moose Jaw station, and are monthly totals. The “snow on the ground” measurement is
the physical quantity of snow-cover on the last day of each month.
• The withdrawal volumes are provided by the on-site Buffalo Pound Water Treatment Plant (WTP) and by SaskWater. Daily averaged
water-level measurements are provided by the WSA for an in-reservoir gauge.
• Monthly inflow DO and BOD measurements are provided by the WSA for a sample site at the Highway 2 boundary.
• Initial conditions for water temperature and DO are also taken from the WTP weekly dataset. Sediment temperature is set at the mean
annual air temperature over the simulation period as per the W2 manual recommendation
Parameter coefficients are set according to knowledge of the reservoir, or are left at W2 default values where data are
not available to support a change. The kinetic coefficients for BOD and SOD are W2 defaults (table 1)
4 .Model customization
• We have customized two components of the W2 model: SOD and the ice algorithm

• W2 uses three different types of data for model calibration: the first group are set prior to the model run and remain constant throughout the
simulation—examples being latitude for the calculation of solar radiation, bathymetry, and parameter coefficients.

• The second group are the time-varying state variables such as inflows, outflows, and meteorological data. The third group are the variables
changing internally in the model at each time step; temperature, shear stress, and horizontal and vertical velocities are examples of this group.

• DO is calculated in W2 as per Equation (1). The complete set of DO equations in W2 are more complex as the model recognizes up to thirteen
sources and sinks of DO. We present here the W2 equations we use in our own reservoir DO/SOD model
• For the ice model W2 calculates the formation and melting of ice during simulations, and the relevant processes
(e.g., light, wind, heat fluxes) are adjusted accordingly by the model. Snow is not considered in the algorithm.
Snow depth at BPL is often between 0.1 and 0.3 m as per the supplied WSA long-term data. To account for this
lack of snow the ice model has been extended to include two empirical coefficients to the existing W2 algorithms,
as have been previously applied . The first coefficient α extends the ice growth and thickness equations and
reduces the heat lost through back radiation from black surfaces. The second coefficient β extends the ice melt
equations and reduces the heat conduction between air and ice. Both coefficients are assigned a value between zero
and one to be multiplied by the appropriate equation parameter.
5. Model setup and application
The model simulates a continuous seven-year period (1 April 1986–31 March 1993). This period is chosen due to the availability
of daily flow data recorded by two WSA gauges just above and below BPL that were subsequently discontinued. The water
balance, ice-on and ice-off dates, and the water temperature model were calibrated. The final temperature model shows good
results (Figure 2)
We first simulated a simple DO model of BPL with a constant SOD, for comparative purposes. We extended the calibrated
temperature model by enabling the water quality variables DO and BOD (BOD as one group) in W2. We proceeded to calibrate
the SOD rate as part of a Monte Carlo analyses for several coefficients. We used MATLAB (MathWorks, Natick, MA, USA) to
run W2 for these calibration iterations and attempted to fit the predicted DO to observed DO concentrations. Using a constant
SOD we were only able to produce a moderately good fit (Figure 3): with both underestimations and overestimations of DO
throughout
• We found that the DO model performed better with the variable SOD rates (Figure 3). On examining the results of this new
model, we noted that SOD followed a relatively consistent seasonal trend. SOD was high over summer, peaking towards the end
of the season, and then gradually depleting over winter. The rates of SOD were different in magnitude each year, yet similar in
behaviour. We compared the new SOD results against observed in-reservoir water-quality data to look for trends. We noticed that
the predicted SOD appeared to follow a similar pattern to the observed weekly summer chlorophyll-a (Chla) concentrations
(Figure 4 ) over the first few years: with SOD peaking not long after Chla. In light of this, we investigated if any relationships
could be found between Chla abundance, and SOD. Our aim was to determine if Chla might be useful as an alternative
measurement for estimating SOD
• We then used an equation based on the apparent relationship of these two variables (summer SOD = 0.0042 × max summer Chla
+ 0.934 5) to set the remaining summer SOD rates based on the maximum summer Chla concentrations each year. By this
method, we used the previous summer’s maximum Chla concentrations as a proxy of the magnitude of biomass production that
settled to the bottom sediments by the end of the open water season.
Results
1. Dissolved oxygen simulation
2. Sediment oxygen demand relationships
 Conclusions

• From the modelling, we show that winter SOD decay is inversely dependent on the previous summer’s maximum
Chla concentrations. The decay rate is faster when less algae are produced. A constant SOD value suffices during
the summer half-year; however, a better DO simulation is obtained in winter when the SOD rate decays during the
course of the winter. This implies that the biomass supply during winter is limited and much of the draw on DO is
diminished by the end of the winter. This result is backed by several field studies. We have shown that for a Prairie
shallow reservoir with few macrophytes and BOD inputs variable SOD can be used in a water quality model to
represent additional oxygen demand after an algal bloom. The summer SOD and winter SOD decay can be
estimated by treating the open water and under-ice period individually in the model. This variable SOD over-time
can be estimated for both summer and winter conditions based on summer concentrations of Chla. This concept
can be widely applied to similar systems that do not have data to support a full diagenesis model, yet would benefit
from a more representative estimation of SOD than is provided by a zero-order constant SOD rate.
Thank you for your attention