SA 910/8-79-105 United States Region 10
Environmental Protection 1200 Sixth Avenue
Agency Seattle, WA 98101
Surveillance & Analysis Division
<&ERA The Application of a
One-Dimensional Steady-State
Estuary Water Quality Model
To Grays Harbor, Washington
-------
EPA 910/8-79-105
April, 1980
APPLICATION OF A STEADY-STATE
DISSOLVED OXYGEN MODEL
TO
LOW FLOW CONDITIONS
IN
GRAYS HARBOR, WASHINGTON
Prepared by
John R. Yearsley
William C. Hess
EPA—Region 10
1200 Sixth Avenue
Seattle, Washington 98101
-------
TABLE OF CONTENTS
CHAPTER PAGE
FINDINGS AND CONCLUSIONS 1
INTRODUCTION 5
PREVIOUS INVESTIGATIONS 9
WATER QUALITY DATA 11
MODEL DEVELOPMENT 13
Model Geometry and Data 13
Point Sources and Freshwater Tributaries 17
Rate Constants, Water Temperature, and Saturation D.O. 17
Sediment Oxygen Demand 23
The Pseudo-Diffusion Coefficient 25
Observed and Simulated Dissolved Oxygen Profiles 31
Primary Production 33
SENSITIVITY ANALYSIS 37
BIBLIOGRAPHY 51
-------
LIST OF FIGURES
FIGURE PAGE
1. Grays Harbor and freshwater tributaries 6
2. Channel-junction system for Grays Harbor
based upon entire estuary geometry 14
3. Channel-junction system for Grays Harbor
based upon main channel geometry 16
4. Estimated and observed sediment oxygen dsnand 24
5. Pseudo-diffusion coefficient for entire
channel and main channel simulations 28
6. Comparison of simulated and observed
salinity for full channel and main channel geometries.. 29
7. Comparison of simulated and observed
0.0. for full channel and main channel geometries 35
8. Comparison of simulated and observed 8.O.D.
for full channel and main channel geometries 36
9. Sensitivity of the steady-state diffusion model
to the pseudo-diffusion coefficient 40
10. Sensitivity of the steady-state diffusion model
to the reaeration rate 41
11. Sensitivity of the steady-state diffusion model
to the deoxygenation rate 42
12. Sensitivity of the steady-state diffusion model
to ambient water temperature 43
13. Sensitivity of the steady-state diffusion model
to sediment oxygen demand (S.O.D.) 44
14. Sensitivity of the steady-state diffusion model
to the removal of B.O.D. and S.O.D 45
15. Sensitivity of the steady-state diffusion model .
to changes in upstream B.O.D 46
16. Sensitivity of the steady-state diffusion model
to changes in upstream D.0 47
17. Sensitivity of the steady-state diffusion model
to changes in ocean 8.0.D *8
18. Sensitivity of the steady-state diffusion model
to changes in ocean D.0 49
-------
LIST OF TABLES
TABLE PAGE
1. Geometric characteristics used in steady-state
model (entire estuary) 19
2. Geometric characteristics used in steady-state
model (main channel) 20
3. Average water quality and quantity of point
sources, July 25-29, 1977 21
4. Mean values of temperature, dissolved oxygen
and salinity in Grays Harbor, July 25-29, 1977 22
5. B.O.D. rate constants for Grays Harbor,
July 25-29, 1977 22
-------
THIS DOCUMENT IS AVAILABLE IN LIMITED QUANTITIES
THROUGH THE U.S. ENVIRONMENTAL PROTECTION AGENCY,
SURVEILLANCE AND ANALYSIS DIVISION, 1200 SIXTH
AVENUE, SEATTLE, WASHINGTON 98101
-------
A WORKING PAPER PRESENTS RESULTS OF
INVESTIGATIONS WHICH ARE, TO SOME EXTENT,
LIMITED OR INCOMPLETE. THEREFORE,
CONaUSIONS OR RECOMMENDATIONS
EXPRESSED OR IMPLIED — MAY BE TENTATIVE.
-------
FINDINGS AND CONCLUSIONS
A steady-state one-dimensional mathanatical model of dissolved oxygen was
evaluated under critical low flow conditions in Grays Harbor, Washington.
The model was based upon the assunption that the equations of mass balance
could be described by non-tidal advection, diffusion-like processes and
first-order biochemical reactions.
Vertically and time-aver aged salinities, collected during the EPA Region
10 survey of July 25-29, 1978, were used to calibrate the pseudo-diffusion
coefficients, using a balance between diffusion-like processes and non-
tidal advection. The computed salinity distributions had a minimun value
at the upstream, or freshwater, boundary of the estuary, and a maximim at
the ocean boundary. The computed distribution of the pseudo-diffusion
coefficient had a relative maximim in the inner estuary. This relative
maximim was similar to results found by other investigators and was
attributed to the two-layer flow induced by density differences between
the oceanic and the freshwater boundaries of the estuary.
Simulations of the ti dally-aver aged dissolved oxygen and carbonaceous
biological oxygen demand in Grays Harbor were done using two different
models of the portion of the estuary affected by waste discharge. One of
the models was based upon the assunption that water quality
-1-
-------
characteristics were uniform across the estuary. The other model was
based upon the assunption all point source discharges and freshwater
run-off were confined to the main navigational channel, only. The mean
and standard deviation for the difference between simulated and mean
observed were, for the first model, 0.20 mg/1 and 0.32 mg/1,
respectively. The corresponding values, using the second model were 0.18
mg/1 and 0.31 mg/1, respectively. Since the first model was based upon
assumptions which, are more consistent with the available data, it was
this model upon which the sensitivity analysis was based.
An important aspect of the analysis was determining the sensitivity of
the mathematical model to the following factors:
1. Changes in the pseudo-diffusion coefficient
2. Changes 1n the reaeration rate
3. Changes in the magnitude of the sediment oxygen demand
4. Changes in boundary conclusions for dissolved oxygen and 800
5. Changes in the BOD loading
6. Changes in the temperature of the receiving waters.
The results of this analysis indicated that the mathematical model was
most sensitive to changes in the reaeration rate and the magnitude of the
sediment oxygen demand. The model was affected somewhat less by changes
in the pseudo-diffusion coefficient, the point source BOD discharge, the
oceanic dissolved oxygen and BOD boundary conditions, and the deoxygena-
-2-
-------
tion rate. The model was least sensitive to the receiving water tempera-
ture profile, and the upstrean, or freshwater, boundary conditions for
dissolved oxygen and BOD.
Net production of dissolved oxygen by phytoplankton was assumed to be
zero in all simulations. This was done because the available data may
not have been representative of in situ conditions. Additional studies
of primary production should be performed to improve our knowledge of the
dissolved oxygen budget of Grays Harbor.
The one-dimensional model was calibrated only for the flow occurring
during the July 1977 survey. Application of the model to other flow
regimes can be accomplished by estimating the pseudo-diffusion
coefficients appropriate to the new regimes. Available data can be used
for this, as well as correlations between pseudo-diffusion coefficients
and freshwater flow developed in other studies.
-3-
-------
-4-
-------
INTRODUCTION
Conditions of extreme low river flows, which occurred in the Pacific
Northwest during 1977, provided an opportunity to evaluate mathematical
water quality models under critical conditions. Grays Harbor (Figure 1),
an estuary on the coast of Washington, experiences water quality problems
during such critical conditions. The National Pollution Discharge Elimi-
nation System (NPDES) permits, written for industries which discharge to
the estuary, contain limitations based upon the freshwater discharge to
the estuary. Because the estuary is water quality limited, 1t was the
subject of a comprehensive water quality survey during the period,
July 25-29, 1977, and a subsequent analysis of the data using mathe-
matical models of water quality. Due to time and resource limitations,
the models were limited to those which simulate temperature, salinity,
dissolved oxygen and biochemical oxygen demand.
Two different modeling concepts were applied to the analysis of the
data. The first, which is described in this report, was based upon the
assianption that the estuary could be considered to be a steady-state,
one-dimensional systan, governed by processes of horizontal diffusion,
non-tidal advection, and first-order biochemical reactions. The second,
as described by Yearsley and Cleland (1979), is a dynamic, quasi-
two-dimensional hydraulic and water quality model which accounts for
diurnal variations of water transport and surface elevations, as well as
diurnal variations in processes which drive biochemical reactions, such
as solar radiation.
-5-
-------
I
cr>
i
Figure 1. Grays Harbor and fresh water tributaries.
\MONTESANO
Wynoochee rNmIP-^L
Humptulips R
Chehafo
ishkah R
HOQyiAM|
COSMOPOLIS
Johns R
NAUTICAL MILES
¥
-------
The intent of these model studies was to examine the processes which
govern the distribution of dissolved oxygen in Grays Harbor and to
provide the basis for comparing a simple steady-state model and a more
sophisticated dynamic model. The final objective of the study was to
improve our methods for evaluating and developing NPDES permits in Grays
Harbor.
-7-
-------
-8-
-------
PREVIOUS INVESTIGATIONS
There have been a nunber of comprehensive field study programs in Grays
Harbor and also a nunber of mathematical model studies. The first, and
perhaps the most thorough field study was done by Ericksen and Townsend
(1940). This study documented the severe water quality problem which had
resulted from the discharge of high-strength organic wastes by the Grays
Harbor Pulp and Paper Company (now owned by ITT/Rayoni er). Subsequent
field studies Included the work of Orlob, Jones and Peterson (1951),
Peterson (1953), Peterson, Wagner and Livingston (1957) and Beverage and
Swecker (1969). With the exception of the report by Beverage and Swecker
(1969), the principal objective of these studies was to monitor the
impact of waste discharge upon water quality. The Beverage and Swecker
(1969) study, in addition to dealing with water quality, provided useful
information regarding the hydraulic characterisecs of the estuary.
Pearson and Holt (1960) examined the effects of oceanic upwelling upon
the dissolved oxygen in Grays Harbor. Beverage and Swecker (196 9) also
did an extensive analysis of the character of the bottom sediments in
Grays Harbor. Since 1971, ITT/Rayoni er and Weyerhaeuser have monitored
water quality in Grays Harbor at weekly intervals during the sunmer low
flow. While this data is not published, it is available from the State
of Washington's Department of Ecology.
Mathematical modeling in Grays Harbor was first reported by Callaway
(1966), who applied a single, steady-state model to the analysis of waste
loading and oceanic upwelling. This work was extended by Yearsley and
-9-
-------
Houcfc (1973) to include the effects of sediment oxygen demand. Lorenzen
et al (1976) developed a dynamic water quality model of the estuary,
making extensive use of the estuary hydraulic model developed by the
firm, Water Resource Engineers, Inc. (Shubinski, McCarty and Lindorf
(1965). All of these studies provided the basis for the work described
in this report and the report of Yearsley and Cleland H979>.
-10-
-------
WATER QUALITY DATA
An intensive water quality sampling program was conducted in Grays Harbor
(Figure 1) during the period July 25-29, 1978, by the EPA Region 10. The
sampling program included measurements of temperature, salinity,
dissolved oxygen and nutrients from major industrial and municipal
sources, tributaries discharging to Grays Harbor and the waters of Grays
Harbor, proper. A description of this data collection program is given
by Yearsley (1979).
The data required for this study included vertical and time averages of
temperature salinity, dissolved oxygen and biochemical oxygen demand in
the estuary receiving waters and the discharge rates of these
constituents from point sources and tributaries.
-11-
-------
-12-
-------
MODEL DEVELOP NENT
The mathematical model chosen for this study was a steady-state,
one-dimensional model. The processes which determines the mass balances
in this model are non-tidal advection, diffusion-like processes and
biochemical reactions, where appropriate. There are a nunber of models
of this type described in the literature. AUTOQUAL, developed by Crim
and Lovelace (1973), was chosen for this study because it was a
well-docunerrted and tested software package.
AUTOQUAL is based upon a system of discrete volumes called junctions,
which communicate with each other through a systan of channels. The
model assunes that all junctions are fully mixed, with properties
represented as point values at the center of the junctions. A complete
description of model development is given by Hess (1978).
Model Geometry and Data
Model geometry for (Figure 2) Grays Harbor was defined with the Highway
107 bridge south of Montesano adopted as the upstream boundary.
Distances were measured seaward along the Chehalis River channel in
nautical miles up to nautical mile 11 near Aberdeen. From there,
distances were measured as a straight line to the center of the
navigational channel at Point Chehalis. All water quality, point source,
and tributary data are referenced to this river mile system. For
purposes of simulation, the estuary was partitioned into 30 segnents, one
-13-
-------
I
H
-O
I
Humptulips R
lorth Bty
ishkah R
Hoquiam R
HOQUIAM
"ill
ABERDEE
COSMOPOUSji
Johns R
South Ba
Figure 2. Channel-junction system for Grays Harbor
based upon entire estuary geometry
NAUTICAL MILES
x ChehaNs
VMONTESANO
Wvnoochee rA|$-^
-------
nautical mile long, beginning at nautical mile 0. Twenty seven of these
segnents are within Grays Harbor while three are in the open ocean. The
three segnents defined in the ocean were used to establish the ocean
boundary conditions. Two channel-junction systems were defined for Grays
Harbor.
Initially the estuary was segnented as it appears in Figure 2. Starting
at nautical mile 0 the estuary was segnented along the length of the
channel until nautical mile 11 was reached. From nautical mile 11 to the
ocean segnents were defined across the estuary's entire width. A second
channel-junction system was defined as it appears in Figure 3. Here the
channel junction system 1s identical to that in Figure 2, except that
beyond mile 14 only the North Channel of the estuary is included in the
segnents. This second configuration assunes that the main navigational
channel carries all the freshwater discharge, as well as all the point
source waste discharges.
The basic input data required by AUTOQUAL are described by Crim and
Lovelace (1973). For a detailed description of how these data were
input, see the AUTOQUAL users manual (Crim and Lovelace (1973)). Segnent
widths and depths were computed frcm output from the WRE hydrodynamic as
calibrated by Yearsley and C lei and (1979).
Hydraulic geometry was required for both systems at mean tide. The
geometry used for each system is given in Tables 1 and 2 .
-15-
-------
Humptulips R
\MONTESANO
Wynoochee I^IIhL
Chehajii
ishkah R
HOQUIAM
ERDEEN!
COSMOPOLIS
Johns R
NAUTICAL MILES
higure 3. Channel-junction system for Grays Harbor
based upon main channel geometry
-------
Point Sources and Freshwater Tributaries
Municipal and industrial waste loading data were based on discharge
reports obtained from the Washington Department of Ecology and on data
collected by EPA Region 10 during the July 25-29 1977 survey. Arithmetic
averages of data collected during the period, July 25-29 1 977 were used
to determine loadings for input to the model.
The discharge from freshwater tributaries in Grays Harbor was obtained
from the USGS. Water quality characteristics were obtained from data
collected during the July 25-29, 1977 field study.
Estimates of mean values of wata- quantity and quality for the point
sources and freshwater discharges to Grays Harbor are shown in Table 3.
Rate Constants, Water Temperature and Saturation P.O.
Values of temperature, salinity, dissolved oxygen, and biological oxygen
demand used in the analysis were arithmetic averages of observed data
collected during the survey (Yearsley (1979)), as shown in Table 4. BOD
-17-
-------
rates computed from samples collected in Grays Harbor during the July
1977 survey are shown in Table 5. The reaeration rates in each segnent,
j, were calculated using the equation developed by Langbein and Durtm
(1967):
K,-, _ 7.C>Z V\ n>
~ O- ^
where,
1^= the reaeration coefficient, days"* (base e), in
segment, j,
Vj= the river root-mean square velocity, ^eet/second, in
segnent, j,
D .= the channel depth, feet, in segnent, j.
J
The river root-mean square velocity, V., at each segnent, j, was
J
determined kinematically:
Vi = Qi 4 -IT" WaiA k,\ (2)
A»j H A,',
-18-
-------
Table 1. Geometric characteristics used in steady state model for Grays
Harbor (entire estuary):
Distance from
Morrtesano
(nautical miles)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
Segment
Depth
(feet)
11.5
12.7
12.9
14.3
14.6
16.3
18.4
20.8
21.3
20.9
20.4
21.5
23.8
20.7
11.6
11.4
11.8
15.3
13.7
10.4
9.9
11.6
16.3
16.8
21.0
30.0
47.0
Channel
Width
(feet)
380.2
478.8
504.9
624.4
728.8
832 .3
939.1
826.4
955.8
1074.0
1134.3
1097.1
2115.5
3885.5
6901.0
10546.3
11740.1
135 92 .8
1640 5.4
23937.8
32090.4
34879.0
41285.8
45719.7
292 88.0
10290 .7
84 76.0
-19-
-------
Table 2. Geometric character! sties used in steady-state model for Grays
Harbor (North Channel):
Distance from
Montesano
(nautical miles)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
• 16
17
18
19
20
21
22
23
24
25
26
27
Segnent
Depth
(feet)
11.5
12.7
12.9
14.3
14.6
16.3
18.4
20.8
21.3
20.9
20.4
21.5
23.8
22.7
21.0
18.9
16.9
22.6
21.6
23.1
24.6
28.1
30.1
29.7
29.6
30.0
47.0
Channel
Width
(feet)
380.2
478.8
504.9
624.4
728.8
832 .3
939.1
826.4
955.8
1074 .0
1134.3
1097.1
2115.5
32 60 .9
1465.0
1974.7
3047.0
2891.8
3864.6
4045.5
5523.1
7300.3
83 92.5
9921 .3
13343.5
10290 .7
8476.0
-20-
-------
Table 3.
Average water quality and quantity of point source
discharging to the Chehalis River and Grays Harbor,
July 25-29, 1977:
Distance Dissolved Carbonaceous
Source
Morrtesano
(nautical miles)
Flow
(c.f .s.)
Oxygen
(mg/1)
BOD (ultimate)
(mg/1)
Chehalis River
0.0
628.0
9.8
2.7
Wynooche River
0.5
210.0
9.3
0.7
Weyerhaeuser Co.
8.5
1.2
6.7
25.8
Wishkah River
11.5
115.0
7.2
2.4
Aberdeen STP
12.3
3.5
5.9
150.0
Weyerhaeuser Co.
13.4
33.6
3.8
200.0
ITT yRayoni er
14.2
45.5
3.2
107.0
Hoquiam River
14.3
108.0
5.6
2.6
Hoqui am STP
16.2
2.6
8.5
89.5
-21-
-------
Table 4. Mean values of tanperature, dissolved oxygen and salinity at
various locations in Grays Harbor, as measured during the
period July 25-29 1977:
Distance frora
Morttesano
Temperature
Oissolved
Oxygen
Sal inity
(nautical miles)
( C)
(mg/l)
(ppt*)
0.0
0.4
2.4
18.9
9.3
0.6
4.8
19.0
8.3
1.9
6.4
19.0
7.5
4.6
7.8
18.8
6.7
7.0
8.9
18.5
6.3
10.7
11.6
18.0
5.6
16.1
13.2
17.9
5.8
19.7
14.2
17.5
6.1
21.7
16.2
17.0
6.0
24.5
20.0
15.2
6.7
27.4
21.9
14.4
7.6
28.5
24.5
12.9
7.1
27.0
—
—
29.3
Table 5. 800 rate constants for Grays Harbor, computed from July 1977
field data:
Distance
from Morttesano
(nautical miles)
2.4
7.8
11.6
13.2
14.2
16.2
18.8
20.0
24.5
800 Rate
Constant (base el
(days-^
0.13
0.14
0.12
0.17
0.13
0.21
0.25
0.24
0.25
-22-
-------
where,
Qj = the freshwater flow, cfs, in segnent, j,
Axj= the cross-sectional area, square feet, segnent, j,
W = the tidal frequency, 1/seconds,
Asj= the surface area of the tidally influenced part of the
estuary landward of segnent, j, square feet,
&H = the amplitude of the tide, feet
Sediment Oxygen Demand
Sediment oxygen demand was estimated, as described by Yearsley and Houck
(1973), by combining the results of benthic respircmeter data
(Kreizenbeck (1973)) and carbon content analysis of bottom material
samples reported by Beverage and Swecker (1969). The benthic
respircmeter data consisted of three readings taken at nautical miles
13.5, 14.5, and 17.5. Carbon analysis data was available from nautical
miles 6.5 to 27.5 from both the North and South Channels. It was assuned
that distribution of carbon in the sediments was the same during the
carbon content and the respirometer studies. It also was assuned that
sediment demand was proportional to carbon content thus their
distributions should be similar. The ratio of sediment oxygen demand to
carbon content was calculated at nautical mile 13.4. This ratio was then
used to estimate the distribution of sediment demand for the rest of the
estuary. Estimated and observed sediment oxygen demand is plotted in
F igure 4.
-23-
-------
FIGURE 4. ESTIMATED AND OBSERVED SEDIMENT OXYGEND DEMAND IN
GRAYS HARBOR, WASHINGTON.
t ESTIMATED
+ OBSERUEO - EPA(8'73)
t#
t 4*
* * * * '
5 10 15 20 25 30
DISTANCE FROM MONTESANO - NAUTICAL MILES
-------
The North Channel is dredged on a regular basis. Dredging operations
were conducted in 1976, and we-e under way at the time of the July 1977
survey. The tacit assunption in distributing sediment demand on the
basis of carbon content measurements made seven years earlier 1s that
sediment materials tend to assune the same steady state distribution
after each dredging in a time frame which is small relative to the
interval between dredgings.
The Pseudo-Oiffusion Coefficient
The basic assunption upon which the model is formulated is that the
processes in an estuary can be described as a one-dimens1onal balance
between d1ffusion-like processes, advection and first-order
biochemical/physical decay. The determination of the coefficient
characterizing the diff usi on-1 ike process is a key elemait in this
analysis. This coefficient, which we shall call the pseudo-diffusion
coefficient, includes the effects of a nunber of processes such as eddy
diffusion, gravitational mixing and other forms of dispersion. Since the
diffusion term becomes the repository for several complex processes not
explicitly stated elsewhere in the nontidal models, the determination of
the pseudo-diff usion coefficient based on theoretical considerations
becomes extremely difficult if not impossible. Thus, its determination
usually rests on an analysis of the distribution of some conservative
tracer such as dye or salinity. Implicit in such an analysis is that the
estuarine transport processes not specified explicitly by the model are
-25-
-------
expressed in the observed distribution of the tracer. Furthermore, it is
assuned that this information can be incorporated into the diffusion
coefficient. Stommel (1953), and O'Connor (1961), respecti vely, used
salinity as a tracer to determine pseudo-diffusion coefficients for
tidally averaged models. The same general procedures were followed in
determining coefficients at mean tide for Grays Harbor. The ti dally
averaged form of the convective-diffusion equation for salinity, S, is:
^SA") ijfujSA) _ D (AE 2S")
3 "t 2 x 7)k 2>k
(3)
If we assune steady state conditions, Equation (3) may be solved for the
pseudo-diff usi on coefficient:
=. U£ ^ (4)
"ST
da
-26-
-------
Note that = (j^A is the freshwater flow. Since salinity is usually
defined as a function of x by a series of discrete observations, the fi-
nite difference form of Equation (4) is generally used. Because observa-
tions for Grays Harbor were not available for each of the 30 segments
defined, a smoothing and Interpolating function was needed. The function
used was required to capture at least the first and second order changes
of salinity with x, be continuous on at least the Interval between obser-
vation points, and have a first derivative which is continuous between
and at observation points. The class of nth degree polynominals pro-
vided a convenient set of functions having these characteristics. In
addition, these functions are continuous over their entire domain, thus
Equation (4) may be used in its derivative form. Polynominals of degree
three through six we*e fitted, using least squares analysis, to the har-
monically smoothed salinity data. Polynominals of degree four appeared
to best preserve the higher order character!"sties of the data and provide
suitable smoothing. Using the polynominal fits, the determination of the
pseudo-diffusion coefficients was then a simple process for most portions
of the estuary. However, near both the upstream and seaward boundaries
O » making the right-hand side of Equation (41 an
undefined quantity. In addition, neither the 4th degree polynominal nor
it first derivative were reliable close to the beginning or ending data
stations (nautical miles 2.4 and 24.5 respectively). The diffusion
coefficients in these regions were determined by extrapolation of the
well-defined values in the estuary interior and knowledge of the general
diffusion characteristics at the boundaries. At the upstean boundary it
is expected that diffusive-like transport should approach a small value,
-27-
-------
FIGURE 5. PSEUDO-DIFFUSION COEFFICIENT FOR ENTIRE CHANNEL
AND MAIN CHANNEL SIMULATIONS IN GRAVS HARBOR, WASHINGTON.
8000
I
ro
00
1
c
0
E
F
F
1
C
I
E
N
T
7000 -
6000 -
5000 -
+ ENTIRE CHANNEL SIMULATIONS
* HAIN CHANNEL SIMULATIONS
*******
F
E
E
T
*
*
2
/
S
E
C
0
N
D
4000
3000 -
2000 -
1000
* *
* *
+ +
+ + + + +
* * *~ + + + j.
* * | + +
ft * * * *
x
X
X
X
X
5 10 15 20 25
DISTANCE FROM MONTESANO - NAUTICAL MILES
30
-------
FIGURE 6. COMPARISON OF SIMULATED AND OBSERVED SALINITY IN
GRAYS HARBOR FOR FULL CHANNEL AND MAIN CHANNEL GEOMETRIES.
30
20
FULL CHANNEL CEOflETRV
RAIN CHANNEL CEOICTRV
EPA REGION 10 FIELD STUDY
5 10 15 20 25 30
DISTANCE FROM HONTESANO - NAUTICAL MILES
-------
since the magnitude of tidal action and salinity gradients decrease in
the upstream direction. Near the seaward boundary the observed salinity
profile becomes flat, approaching tire ocean value of salinity, indicating
that this portion of the estuary is well-mixed. In addition, it would be
expected that the intensity of mixing due to tidal action should increase
as the seaward boundary is approached. Thus, the diffusion coefficient
near this end of the estuary would be expected to become large. Using
these characteristics of diffusion as a guide, pseudo-diffusion coeffi-
cients were estimated near both boundaries by trial and error fits to the
observed salinity profiles. The calculated pseudo-dlffuslon coefficients
and simulated salinity profiles for both the complete and the partial
estuary cross-sections are presented in Figures 5 and 6.
Plots of the pseudo-diffusion coefficient versus distance all reveal the
same general relationships, each with a distinctive maximun approximately
10 nautical miles downstream of Montesano. These plots show a
resemblance to a relationship postulated by WRE (1974). WRE (1974)
suggests that the pseudo-diff us ion coefficient may be thought of as the
aggregate of three major processes, each of which predominate in different
portions of the estuary. Since the maximun in Figure 6 occurs where
density induced mixing predominates, it was hypothesized that this mixing
mechanism was responsible for the observed maximun in the coefficients
calculated for Grays Harbor.
-30-
-------
Observed and Simulated Dissolved Oxygen Profiles
Observed dissolved oxygen profiles used for comparison with simulation
results were derived from data collected during the July, 1977 survey.
The arithmetic mean of dissolved oxygen measurements from each station
was used to define the tidal average at that station.
Dissolved oxygen and carbonaceous biological oxygen demand (BOD) were
simulated using cross sections extending the full width of the estuary
(see Figure 2). Both of these parameters were also simulated for the
model configuration which used the main navigation channel geometry past
nautical mile 14 (see Figure 3). The resulting simulations are compared
with the arithmetic mean of the observed values in Figures 7 and 8. The
mean and standard deviation of the difference between simulated and
observed dissolved oxygen concentrations for the full channel
cross-sections are 2 mg/1 and +0.51 mg/1, respectively. The
correspondlng values for the difference between simulated and observed
concentrations of dissolved oxygen, using the main channel geometry, are
0.18 mg/1 and +0.43 mg/1, respectively.
However, the assunptions upon which the main channel geometry model is
based are not consists with our knowledge of the hydrography and
hydrodynamics of the North and South Channel of Grays Harbor. The
hydrographic data (Yearsley (1979)) indicates that the water quality of
the South Channel is similar to that of the North Channel. Hydrodynamic
-31-
-------
studies by Brogden and Fisackerly, M973), using a physical model of
Grays Harbor, show that the flow in the South Channel is similar to that
in the North Channel. For these reasons it was difficult to justify
using the main channel geometry without further analysis. As a result,
the following discussion of sensitivity is based upon the model with the
full channel geometry.
-32-
-------
Primary Production
One of the assumptions upon which the Base Run (Figure 7) is based is
that there is no net photosynthetic production of oxygen in Grays
Harbor. During July 1977 the University of Washington Department of
Oceanography measured primary productivity in Grays Harbor using the
14C method (Yearsley 1979). The results of this study indicate that
the average net production of dissolved oxygen varied from 0.37 grams
2 7
02/meter /day to 8.13 grams 0g/meter /day.
There are two questions regarding these productivity data which make it
difficult to incorporate the data into the dissolved oxygen budget. The
first being that the samples were incubated at temperatures equal to the
warmest temperatures in Grays Harbor duing July 25-29, 1977. This was
5°C to 7° C higher than the temperatures in outer Grays Harbor where
the maximun productivity was measured. Winter et al (1975) suggest that
these higher temperatures will result in overestimates of the production
rate. The maximun specific primary production rates (ratio of carbon
fixed per unit volune per unit time to chlorophyll _a per unit volune^
during July 1977 are two to three times higher than those observed in
Puget Sound (Winter et al (1975)) and five to seven times higher than
those observed in Grays Harbor by Westley and Tarr (1965).
-33-
-------
The second question is related to the amount of grazing done by zooplank-
ton. The University of Washington data suggests that chlorophyll a
degradation products in Grays Harbor were greater, in general, in Grays
Harbor during July 1977, than in oceanic surface water (W. Peterson,
personal communication). While these degradation products could be
resuspended material from the bottom, they can also be the product of
chlorophyll digested by zooplankton or other organisns.
As a result of these uncertainties, we chose not to incorporate the
productivity data into the simulations. The model predicts the average
dissolved oxygen in Grays Harbor quite well, given the assumption of no
net production of dissolved oxygen by phytoplankton. However, until we
have more complete knowledge of primary production levels in Grays Harbor
we must treat these results with the appropriate caution.
-34-
-------
FIGURE 7. COMPARISON OF SIMULATED AND OBSERVED D.O. IN
GRAYS HARBOR FOR FULL CHANNEL AND MAIN CHANNEL GEOMETRIES*
14
I
CO
cn
i
D
I
S
S
0
L
U
E
D
0
X
Y
G
E
N
12 -
10 -
8 -
6 -
FULL CHANNEL GEONETRV
RAIN CHANNEL GEMETRV
EPA REGION 19 FIELD 8TUDV
M
G
/
L
a -
_L
-L
_L
_L
5 10 15 20 25
DISTANCE FROM MONTESANO - NAUTICAL MILES
30
-------
FIGURE 8. COMPARISON OF SIMULATED AND OBSERUED BOD IN CRAVS
HARBOR FOR FULL CHANNEL AND MAIN CHANNEL GEOMETRIES.
I
CO
CF\
I
u
L
T
I
II
A
T
E
B
0
D
n
G
/
L
3 -
2 -
1 -
FULL CHANNEL CCOttTRV
RAIN CHANNEL QIOKETRV
EPA REOION It FIELD STUDV
DISTANCE FROM MONTESANO - NAUTICAL MILES
-------
SENSITIVITY ANALYSIS
Simulated dissolved oxygen responses to model inputs, parameters,
boundary conditions, and channel geometry are presented as a series of
plots comparing the observed response to the simulated profile for mean
tide (referred to as "Base Run"). The plots appear as Figures 9 through
18.
Response of the simulated dissolved oxygen to changes in the pseudo-dif-
fusion coefficient is shown in Figure 9. For large values of diffusion
the dissolved oxygen sag 1s less than the base runs, but more of the
deficit is transported upstream. In addition, the increased diffusion
causes the system to approach well mixed conditions in the outer harbor,
thus increasing the influence of-the seaward boundary. The opposite is
observed to be true of decreased values of diffusion.
Dissolved oxygen simulations in Grays Harbor are very sensitive to
variations in the reaeration rate (Figure 10). Doubling the reaerations
uniformly throughout Grays Harbor results in a mi nimun dissolved oxygen
1.0 mg/1 higher than the calibration on base case. Halving the reaeration
rate uniformly results in a 1.4 mg/1 decrease in the mi nimun dissolved
oxygen.
Temperature of the receiving water has only a moderate influence on the
dissolved oxygen simulation, as shewn in Figure 11. A 15 percent change
in water temperature results in less than a 5 percent change in the
predicted mi nimun dissolved oxygen.
-37-
-------
The simulated dissolved oxygen are not very sensitive to changes in the
deoxygenation rate, K^. A 100 percent uniform increase in this rate
gives rise to less than a 5 percent decrease in the simulated dissolved
oxygen (Figure 12). Halving the base deoxygenati on rate causes Iks than
a 5 percent increase in the simulated dissolved oxygen.
Figures 13 and 14 examine the model response to sedimeit oxygen demand
and BOO discharge. Figure 13 indicates a substantial response to over
most of the 16 nautical miles where estimates indicate a significant
demand exists (see Figure 4 ). A strong response to sediment oxygen
demand is understaidable, since it exerts it effects directly on the
dissolved oxygen system. Figure 14, shows that the response to point
sources of BOD is significant but is not as large as that observed for
sediment oxygen demand, even though the total demand from sediment oxygen
demaid and BOO are comparable at 57500 and 72600 lbs/day, respectively.
This is due, principally, to the fact that sediment oxygen demand acts
directly and instantaneously upon the water column, while the BOO is
exerted on a time scale which is the order of the reciprocal (l/K^) of
the deoxygenation rate. Because of this a certain amount of the BOO is
not exerted, but is transported out of the estuarine system.
-38-
-------
Figures 15 through 18 examine model response to the landward and seaward
boundary conditions. The effect of the landward boundary condition for
BOD is minimal, showing its maximun response in the reach 5 to 8 nautical
miles downstream of the boundary (Figure 15). The response to DO is more
significant (Figure 16), but the majority of the response was confined to
points 7.5 nautical miles from the boundary and landward. The magnitude
and position of the dissolved oxygen sag was virtually unaffected by
either of the landward boundary conditions.
Model response to ocean BOD was examined by doubling and halving the value
adopted for the Base Run. As Figure 17 demonstrates, the response is
significant not only in the outer harbor but at points considerable
distances upstream. The same is true of model response to ocean
dissolved oxygen (Figure 18). The values examined reflect the range of
values for ocean dissolved oxygen. According to Pearson and Holt (I960),
3.0 mg/1 typifies the effect of upwelling of deoxygenated water at the
ocean boundary, while the upper end of 8.9 mg/1 is the saturated value of
oxygen at oceanic salinity and temperature. Simulation results indicate
a significant impact in the outer harbor. However, a 4.0 mg/1 decrease
in the dissolved oxygen at the entrance gives rise to only a 0.2 mg/1
decrease of the simulated minimun dissolved oxygen in the inner harbor.
-39-
-------
FIGURE 9. SENSITIVITY OF THE STEADY-STATE DIFFUSION PIODEL
OF GRAYS HARBOR TO THE PSEUDO-DIFFUSION COEFFICIENT, EO.
DISTANCE FROM MONTESANO - NAUTICAL MILES
-------
FIGURE 10. SENSITIVITY OF THE STEADY-STATE DIFFUSION MODEL
OF GRAYS HARBOR TO CHANGES IN THE REAERATION RATE, K2.
I
D
I
S
S
0
L
U
E
D
0
X
Y
G
E
N
M
G
/
L
14
12 -
10 -
8
4 -
K
-------
FIGURE 11. SENSITIVITY OF THE STEADY-STATE DIFFUSION MODEL
OF GRAYS HARBOR TO CHANGES IN THE DEOXYGENATION RATE, Kl.
14
I
-P*
ro
i
D
I
S
S
0
L
U
E
D
0
X
Y
G
E
N
12
10
8
II
G
/
L
t 1— 1 1 r
KI>e.»*BASE
K1*«.SIBAS£
IASE
J 1 1 L
5 10 15 20 25 30
DISTANCE FROM PIONTESANO - NAUTICAL MILES
-------
FIGURE 12. SENSITIUITY OF THE STEADY-STATE DIFFUSION IIODEL
OF GRAYS HARBOR TO CHANGES IN THE AMBIENT UATER TEMPERATURE.
14
I
Co
I
D
I
S
S
0
L
V
E
D
0
X
Y
G
E
N
12 -
10
8
6 -
TERP*IASE«B.t DCS CCMT
TEnP*lASE-3.« DEO CEMT
BASE
M
G
/
L
4 -
2 -
-L
_L
-L
5 10 15 20
DISTANCE FROM MONTESANO - NAUTICAL MILES
25
30
-------
FIGURE 13. SENSITIVITY OF THE STEADY-STATE DIFFUSION MODEL
OF GRAYS HARBOR TO CHANGES IN SEDIMENT OXYGEN DEMAND.
14
I
-P*
-C*
I
D
I
S
S
0
L
V
E
D
0
X
Y
G
E
N
12 -
10
8 -
SOD-2.MIA8C
SOD-e.SltASC
IASE
M
G
/
L
4 -
2 -
_L
-L
5 10 15 20 25
DISTANCE FROM HONTESANO - NAUTICAL MILES
30
-------
FIGURE 14. SENSITIVITY OF THE STEADY-STATE DIFFUSION MODEL
OF GRAYS HARBOR TO THE REMOVAL OF B.O.D. AND S.O.D.
14
I
-r*
cj>
i
D
I
S
S
0
L
U
E
D
0
X
Y
G
E
N
12 -
10 -
8 -
6 -
ZERO I.O.D.
ZERO S.O.D.
BASE
n
G
/
L
_L
-L
5 10 15 20 25
DISTANCE FROM MONTESANO - NAUTICAL MILES
30
-------
FIGURE 15. SENSITIVITY OF THE STEADY-STATE DIFFUSION MODEL
OF GRAYS HARBOR TO CHANGES IN THE UPSTREAM B.O.D.
5 10 15 80 85 30
DISTANCE FROM MONTESANO - NAUTICAL MILES
-------
FIGURE 16. SENSITIVITY OF THE STEADY-STATE DIFFUSION MODEL
OF GRAYS HARBOR TO CHANGES IN THE UPSTREAM DISSOLVED OXYGEN*
DISTANCE FROM MONTESANO - NAUTICAL MILES
-------
FIGURE 17. SENSITIVITY OF THE STEADY-STATE DIFFUSION MODEL
OF GRAYS HARBOR TO CHANGES IN THE OCEAN B.O.D.
14
I
CD
I
D
I
S
S
0
L
V
E
D
0
X
Y
G
E
N
12
10 -
8 -
6 -
OCCAN B.O.D.*1.8 NQ/L
OCEAN I.O.D.M.8 M/l
IASE
II
G
/
L
2 -
5 10 15 20
DISTANCE FROPI PJONTESANO - NAUTICAL MILES
25
30
-------
FIGURE 18. SENSITIVITY OF THE STEADY-STATE DIFFUSION MODEL
OF GRAYS HARBOR TO CHANGES IN THE OCEAN D.O.
14
I
l£>
D
I
S
S
0
L
U
E
D
0
X
V
G
E
N
n
G
/
L
12 -
ie -
8
6 -
OCEAN O.O.*8.9 Itt/l
OCCAM D.O.*3.9 Itt/l
BASE
\
\
V _
2 -
-L
_L
_L
_L
5 ie 15 20 25
DISTANCE FROM MONTESANO - NAUTICAL MILES
30
-------
-50- *
-------
BIBLIOGRAPHY
1969 Beverage, J. P., and Swecker, M. N. Estuarine studies in Upper
Grays Harbor Washington Geological Survey Water Supply Paper
1873-B. United States Government Printing Office.
1973 Brogdon, N.J. Jr., and G.M Fisackerly. Grays Harbor Estuary,
Washingtion, verification and base tests. Appendix A:
Supplementary base test data. U.S. Army Engineer Waterways
Experiment Station Technical report H-72-2. Vicksburg,
Mississippi. May 1973.
1966 Callaway, R.J. Simulation of upwelling and pollution in a
vertically mixed estuary. Presented at Second International
Oceanographlc Congress, Moscow, USSR, June 1966.
1973 Crim, R. L., and Lovelace, N. L. AUTOQUAL modeling system. U.S.
Environmental Protection Agency, Office of Air and Water
Programs, Monitoring and Oata Support Division. Washington,
D. C. March 1973.
1940
Eriksen, A., and Townsend, L. 0. The occurrence and cause of
pollution in Grays Harbor. Washington State Pollution Control
Commission. Pollution series bulletin No. 2.100 pp.
1978 Hess, W.C. The application of one dimensional water quality
models to Grays Harbor, Washington. Master's thesis, University
of Washington, June 1978.
1973 Kreizenbeck, R. A. Grays Harbor sediment oxygen demand. U.S.
Envirormental Protection Agency Region 10. August 1973
1967 Langbein, W.B. and W.H. Durim, The aeration capacity of
streams. U.S. Geological Survey Circul ar 542. 6pp. Washington,
D.C.
1974 Lorenzen, M. W., Waddel, W. W. and Johanson, P. A. Development
of a mathematical water quality model for Grays Harbor and the
Chehalis River, Washington. Report to the U.S. Envi ronmental
Protection Ageicy. Battel!e - Northwest, Richland, Washington.
1961 O'Connor, D. J. Oxygen balance of an estuary. Trans. ASCE.
Vol. 126, part 3, pp. 556-576.
-51-
-------
1951 Orlob, G. T., Jones, K. R., and Peterson, D. R. An investigation
of domestic and industrial waste pollution in the lower Chehalis
River and Grays Harbor. Washington State Pollution Control
Commission. Technical bulletin No. 6 50 pp.
1960 Pearson, E. A., and Holt, G. A. Water quality and upwelling at
Grays Harbor entrance. Limnology and Oceanography. Vol. 5, No.
1, pp 48-56
1953 Peterson, D. R. Swage pollution in the estuarial river areas
of Grays Harbor. Washington State Pollution Control Commission.
Technical bulletin No. 16. 17 pp.
1957 Peterson, D. R., Wagner, R. A., and Livingston, A. A re-investi-
ation of pollution in the lower Chehalis River and Grays Harbor
1956-1957). Washington State Pollution Control Commission.
Technical bulletin No. 21. 52 pp.
1965 Shubinski, R. P., McCarty, J. C., and Lindorf, M. R. Computer
simulation of estuarial networks. Journal of the Hydraulics
Division, ASCE. Vol. 91, No. HY5, pp. 33-49.
1953 Stommel, H. M. Computation of pollution in a vertically mixed
estuary. Sewage and Industrial Wastes. Vol. 25, No. 9, pp.
1065-1071.
1979 Yearsley, J. R. A field study of water quality in Grays Harbor
during low flow conditions, July 1977. U.S. Environmental
Protection Agency Region 10.
1979 Yearsley, J.R. andCleland, B.R. The application of a
time-dependent water quality model to Grays Harbor, Washington.
U.S. Environmental Protection Agency Region 10.
1973 Yearsley, J. R. and Houck, D. A cluster analysis for dissolved
oxygen in Grays Harbor. U.S. Environmental Protection Agency
Region 10.
-52-
------- |