United States
Department of
Commerce
National Oceanic and Atmospheric Administration
Environmental Research Laboratories
Seattle WA 98115
United States
Environmental Protection
Agency
Research and Development
Office of Energy, Minerals, and
Industry
Washington DC 20460
EPA-600/7-79-165
July 1979
A Study of the,
Dispersal of
Suspended
Sediment from the
Fraser and Skagit
Rivers into Northern
Puget Sound Using
LANDSAT Imagery
Interagency
Energy/Environment
R&D Program Report
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into nine series. These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The nine series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
6. Scientific and Technical Assessment Reports (STAR)
7. Interagency Energy-Environment Research and Development
8. "Special" Reports
9. Miscellaneous Reports
This report has been assigned to the INTERAGENCY ENERGY-ENVIRONMENT
RESEARCH AND DEVELOPMENT series. Reports in this series result from the
effort funded under the 17-agency Federal Energy/Environment Research and
Development Program. These studies relate to EPA's mission to protect the public
health and welfare from adverse effects of pollutants associated with energy sys-
tems. The goal of the Program is to assure the rapid development of domestic
energy supplies in an environmentally-compatible manner by providing the nec-
essary environmental data and control technology. Investigations include analy-
ses of the transport of energy-related pollutants and their health and ecological
effects; assessments of, and development of, control technologies for energy
systems; and integrated assessments of a wide range of energy-related environ-
mental issues.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.
-------
A STUDY OF THE DISPERSAL OF SUSPENDED SEDIMENT FROM
THE FRASER AND SKAGIT RIVER INTO NORTHERN PUGET SOUND
USING LANDSAT IMAGERY
by
Richard A. Feely and Marilyn F. Lamb
Pacific Marine Environmental Laboratory
Environmental Research Laboratories
National Oceanic and Atmospheric Administration
7600 Sand Point Way N.E.
Seattle, Washington 98115
Prepared for the MESA (Marine Ecosystems Analysis) Puget Sound
Project, Seattle, Washington in partial fulfillment of
EPA Interagency Agreement No. D6-E693-EN
Program Element No. EHE625-A
This study was conducted
as part of the Federal
Interagency Energy/Environment
Research and Development Program
Prepared for
OFFICE OF ENERGY, MINERALS, AND INDUSTRY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
WASHINGTON, D.C. 20460
February 1979
-------
Completion Report Submitted to
PUGET SOUND ENERGY-RELATED RESEARCH PROJECT
MARINE ECOSYSTEMS ANALYSIS PROGRAM
ENVIRONMENTAL RESEARCH LABORATORIES
by
Pacific Marine Environmental Laboratory
Environmental Research Laboratories
National Oceanic and Atmospheric Administration
7600 Sand Point Way N.E.
Seattle, Washington 98115
This work is the result of research sponsored by the Environmental
Protection Agency and administered by the Environmental Research
Laboratories of the National Oceanic and Atmospheric Administration.
The Environmental Research Laboratories do not approve, recommend,
or endorse any proprietary product or proprietary material mentioned
in this publication. No reference shall be made to the Environmental
Research Laboratories or to this publication furnished by the
Environmental Research Laboratories in any advertising or sales
promotion which would indicate or imply that the Environmental
Research Laboratories approve, reconinend, or endorse any proprietary
product or proprietary material mentioned herein, or which has as its
purpose an intent to cause directly or indirectly the advertised
product to be used or purchased because of this Environmental Research
Laboratories publication.
11
-------
Figures
Tables
1. Introduction
1.1 General Statement
1.2 Program Rationale and Objectives
2. Geography, Physical Oceanography, and Hydrology
2.1 Physiographic Setting
2.2 Circulation Patterns and Hydrography
2.3 River Runoff and Sediment Discharge
3. Conclusions
3.1 LANDSAT Imagery
3.2 Quantitative Sediment Mapping .
4. Recommendations
5. Methodology
5.1 General Statement
5.2 LANDSAT Imagery and Hydrographic Data
5.3 Sea-Truth Measurements
5.4 MSS Computer Compatible Tapes
6. Results and Discussion
6.1 Seasonal Variations of Sediment Dispersal
6.2 Tidal Variations of Sediment Dispersal
6.3 Density Fronts
6.4 Quantitative Sediment Mapping
7. Summary and Conclusions
Acknowledgments
Bibliography
CONTENTS
iv
vi
1
1
2
3
3
3
5
• . . 11
• . . 11
11
12
13
13
13
• . • 16
• . . 16
17
17
• . . 28
• . • 30
• . . 33
• • . 41
• • 43
44
111
-------
FIGURES
Number
2.1 Physical setting of the study region . . 4
2.2 Monthly range, mean, and standard deviation of:
a. water discharge; and b. suspended sediment
discharge for the Fraser River at Hope, B.C.
for the period from 1967 through 1977 8
2.3 Monthly mean discharge of water and suspended
sediments of the Skagit River at Mt. Vernon
during 1975 10
5.1 Locations of sampling stations for sea truth
measurements 15
6.1 MSS Band 5 of LANDSAT image 1169-18373 on
Jan. 8, 1973 between slack water and minor
ebb current . . 18
6.2 MSS Band 5 of LANDSAT image 1187-18374 on
Jan. 26, 1973 19
6.3 MSS Band 4 of LANDSAT image 2417-18220 on
Mar. 14, 1976 . . . . 20
6.4 MSS Band 5 of LANDSAT image 2111-18254
showing a southeasterly dispersal of
suspended sediments from the Fraser River
into the Strait of Georgia on May 13, 1975 21
6.5 MSS Band 5 of LANDSAT image 2129-18254
showing a southwesterly dispersal of
suspended sediments from the Fraser River
into the Strait of Georgia on May 31, 1975 . . 22
6.6 MSS Band 5 of LANDSAT image 1727-18290
showing a southeasterly dispersal of
suspended sediments from the Fraser River
into the Strait of Georgia on July 20, 1974 . . . - . 23
iv
-------
Number Page
6.7 MSS Band 5 of LANDSAT image 5465-17484
showing a southeasterly dispersal of
suspended sediments from the Fraser River
into the Strait of Georgia on July 27, 1976 24
6.8 MSS Band 4 of LANDSAT image 2921-18025
showing southeasterly dispersal of
suspended sediments from the Fraser River
into the Strait of Georgia on July 31, 1977 25
6.9 MSS Band 5 of LANDSAT image 2957-18004
showing a southwesterly dispersal of
suspended sediments from the Fraser River
into the Strait of Georgia on Sept. 5, 1977 26
6.10 MSS Band 5 of LANDSAT image 21227-18025
showing a southeasterly dispersal from the
Fraser River into the Strait of Georgia
on June 2, 1978 27
6.11 Schematic diagram of a vertical section
normal to a density front 30
6.12 Low altitude aerial photograph showing
sharp color boundary of a density front . 31
6.13a Composite sketch of fronts seen in 24
LANDSAT images during ebb tides 32
6.13b Composite sketch of fronts seen in 9
LANDSAT images during flood tides 32
6.14 Scatter plots of the relationships among
Band 5 image radiance from the LANDSAT
multispectral scanner 37
6.15 Upwelled radiance for Calvert clay and
Ball clay
6.16 Total suspended matter contour maps as
determined from computer analysis of
LANDSAT images 40
V
-------
TAB L ES
Number Page
2.1 Average annual runoff from the major rivers
discharging into the study region 6
2.2 Comparison of mean water and sediment discharge
for the Fraser and Skagit Rivers during period
of record (1967-1977) . . . . 7
5.1 Principal characteristics of the LANDSAT
satellite and multispectral scanner . . 14
6.1 Results of sea truth measurements and remotely
sensed radiances for selected locations near
the mouths of the Fraser and Skagit Rivers . . . 34
6.2 Correlation coefficient matrix and linear
regression equation for total suspended
matter (TSM) in northern Puget Sound 35
6.3 Results of the stepwise regression analysis of
the radiance data from the MSS computer
compatible tapes with the sea-truth measure-
ments of total suspended matter 35
6.4 Correlation coefficient matrix and linear
regression equation for inorganic suspended
matter (ISM) in northern Puget Sound . . 36
6.5 Results of the stepwise regression analysis of
the radiance data from the MSS computer
compatible tapes with the sea—truth measure-
ments of inorganic suspended matter 36
vi
-------
1. INTRODUCTION
1.1 GENERAL STATEMENT
With the recent development of petroleum reserves at Prudhoe Bay and the
projected development of petroleum and natural gas reserves on the OCS of
Alaska, northern Puget Sound and the Strait of Juan de Fuca-Strait of Georgia
system are slated to become a major transportation route through which
Alaskan crude oils are delivered to Washington State refineries. Refining of
petroleum is presently carried out at Cherry Point, Ferndale, and Anacortes
refineries. Associated with increased production of the refineries will be
the exportation of refined products to markets outside the State of
Washington. A portion of the products undoubtedly will be transported by
water, thus creating an increased potential for hydrocarbon loading of the
local marine waters through minor spillage or a major tanker accident.
With few notable exceptions, the region has not been plagued with
massive oil spills, although the effects of chronic, low-level inputs of
petroleum and refined products have not been adequately assessed. In the
past, tanker and barge traffic in the region has been minimal because much of
the imported crude oil arrived by pipeline from Canada. However, plans
announced by the National Energy Board of Canada will systematically reduce
and finally terminate all exports of crude oil by 1982 (EPA, 1977). This
action will result in a significant increase in tanker traffic to meet local
domestic needs. While the various options for allocation and transportation
of Alaskan crude oil to West Coast ports are presently being discussed, it
remains clear that ever increasing amounts of petroleum will be moved through
the Strait of Juan de Fuca and San Juan Passages via tankers to meet local
demands.
Once crude oil and petroleum products are introduced into the marine
environment, a combination of physical, chemical, and biological processes
become operative in the breakdown and removal of the oil. These processes
include spreading, dispersion, atmospheric injection, evaporation, chemical
and biological emulsification, and sorption onto particles (NAS, 1975). The
last process is one of particular interest because it represents a means by
which oil can be carried down into the water column and become available for
interaction with marine organisms.
Since crude oil is sparingly soluble in seawater, it tends to form
emulsions when introduced into seawater, especially under intense wave action.
The emulsions have a high affinity for particles and tend to be adsorbed
rapidly. Recent studies of oil spills in coastal waters containing high
suspended loads have indicated rapid removal of oil by sorption onto parti-
cles along frontal zones (Forrester, 1971 and Klemas and Polis, 1977).
These zones are regions where turbid brackish water contacts seawater. At
the interface downwelling occurs in most cases, causing the inorganic mate-
rial from the rivers and any associated contaminants to be carried down into
the water column. Similarly, laboratory studies involving the interaction
between crude oils and river-derived inorganic suspended matter have indicated
1
-------
that significant amounts of oil may be accommodated by suspended material, and
that the quantity of oil retained on the particles is dependent upon the
isolectric point of oil and sediment particles, particle size, temperature,
and the concentration of the oil relative to that of the suspended material
(Baker et al., 1978; Feely et al., 1978). Since these processes play a major
role in the disposal and deposition of petroleum hydrocarbons, this study
addresses the temporal and spatial distributions of suspended matter emanat-
ing from the Fraser and Skagit Rivers by means of careful analysis of LANDSAT
imagery.
1.2 PROGRAM RATIONALE AND OBJECTIVES
Several recent advances in remote sensing techniques have provided
scientists with the ability to collect synoptic information about water cir-
culation and sediment dispersal patterns which heretofore have been unobtain-
able. The multispectral scanner images from the LANDSAT-1 and LANDSAT-2
satellites have been especially useful for the study of suspended matter
transport processes in coastal and estuarine water (Kritkos et al., 1974;
Kiemas et al., 1974; Gatto, 1976; Johnson et al., 1977).
A number of investigators have attempted to produce suspended matter
distribution maps based on density-sliced photographic images and sea-truth
determinations (Sharma et al., 1974; Carlson et al., 1975). These investiga-
tions demonstrated the difficult problems of obtaining reproducible informa-
tion from photographic images. However, recent investigations of the
relationships between suspended matter concentrations in the upper 1 m of the
water column and image radiance using computer compatible tapes have proved
to be successful if the ambient suspended matter concentrations were suffi-
ciently high, such as the waste water disposal area in the New York Bight
Apex (Johnson et al., 1977). These results suggest that computer compatible
tapes of LANDSAT imagery can be combined with sea-truth data to obtain algo-
rithms for mapping suspended matter concentrations.
The primary objectives of this study are to: (1) describe the dispersal
patterns of suspended matter emanating from the Fraser and Skagit Rivers by
means of careful analysis of LANDSAT imagery; and (2) to develop and evaluate
computer algorithms for mapping concentrations of total and inorganic sus-
pended matter from the Fraser and Skagit Rivers.
2
-------
2. GEOGRAPHY, PHYSICAL OCEANOGRAPHY, AND HYDROLOGY
2.1 PHYSIOGRAPHIC SETTING
Figure 2.1 shows the physical setting of the study region. The area is
an inland sea with access to the open Pacific Ocean through two major water-
ways, the Strait of Georgia and the Strait of Juan de Fuca. The Strait of
Georgia, to the north, is one of the largest inland waterways on the west
coast of North America. It is partially enclosed with Vancouver Island form-
ing the western boundary and the mainland of British Columbia and northwestern
Washington forming the eastern boundary. The Strait of Georgia is about
220 km long, 33 km wide, and has an average depth of about 150 m. The San
Juan Archipelago and the Canadian Gulf Islands separate the Strait of Georgia
from the Strait of Juan de Fuca. Passage between these major waterways is
primarily accomplished through Haro Strait to the west and Rosario Strait to
the east.
The Strait of Juan de Fuca, to the south, is a submarine valley extending
from the Pacific Ocean to Whidbey Island and Rosario Strait, about 145 km
long. The width varies between 20 and 24 km from the entrance to the Port
Angeles area, where it narrows to 16 km near Victoria, B.C., before widening
to about 40 km. The Strait contains two basins which are separated by a
cross—channel ridge or sill. The ridge extends southward from Victoria, at
an approximate depth of 60 m. The western basin or Outer Strait slopes
gradually in a seaward direction. The eastern basin or Inner Strait contains
several channels which lead into Haro Strait, Rosario Strait, and Admiralty
Inlet.
2.2 CIRCULATION PATTERNS AND HYDROGRAPHY
Circulation patterns and hydrography for the study region have been
studied and reviewed by several authors (Tully, 1942; Waldichuk, 1957;
Herliriveaux and Tully, 1961; Chang et al., 1976; Parker, 1977; Schumacher et
al., 1978), and only a brief summary is deemed necessary for the purpose of
this report.
The primary factors that control circulation and the distribution of
properties in the study region include tides, fresh water runoff, winds and
atmospheric pressure variations, with secondary influences due to coriolis
force, centrifugal force, and topography. The Strait of Georgia is character-
ized by a strong vertical stratification which results from fresh water input
from numerous mainland rivers, the Fraser River being the major contributor.
Driven by estuarine flow, the brackish water moves progressively seaward, pri-
marily through the San Juan passages into the Strait of Juan de Fuca, where
mixing occurs as a result of turbulent tidal energy. As the brackish water
mixes with the saline Pacific Ocean water, part of it is returned to the
deeper waters of the Strait of Georgia. However, the major portion escapes
to the Pacific Ocean in the upper layer. Salt balance is maintained by the
intrusion of colder, more saline water from the Pacific at mid-depths in the
Strait of Juan de Fuca during the summer months (Waldichuk, 1957). This water
3
-------
123°OO
8URRARD
MARPOLE NEW WESTMINSTEI
M,ddIe
4rm
Soundary
Say
F oberts
Physical setting of the study region
23°20
0
STATUTE MitES
9 19
Ô 5 10 15
KILOMETERS
20
CANADA
UN/TED STATES
F DE FUCI 4
Juø’
Figure 2.1
4
-------
moves eastward through the strait, then northward through the San Juan
passages, where it mixes with the less saline water of the Strait of Georgia.
Throughout most of the region the density structure of the water column
is dominated by variations in salinity rather than in temperature. For exam-
ple, during the summer months, the thermocline coincides with the halocline
in the Strait of Juan de Fuca. However, in winter the waters are nearly iso-
thermal and the stability is directly dependent upon the salinity structure.
In the Strait of Georgia, tides and winds are the major controlling factors
for mixing of surface waters, particularly in winter when nearly homogenous
conditions prevail.
The influence of winds on surface circulation appears to be seasonally
dependent. During the winter months the Aleutian Low determines the prevail-
ing wind pattern for the Pacific Northwest, including British Columbia and
Washington. The prevailing winds are predominantly from the southeast. In
the Strait of Georgia the southeasterly winds set up a counterclockwise wind
pattern over the area, providing energy to the surface waters and adding to
the general counterclockwise circulation pattern. In the Strait of Juan de
Fuca, west of Port Angeles, the orographic effect of the Olympic Mountains
turns these winds seaward along the Strait. Wind speeds in this region aver-
age just over 20 knots. The surface waters respond to the wind forcing by
absorbing energy from the winds and increasing surface currents in the gen-
eral direction of the winds.
During the summer the North Pacific High predominates and the winds are
primarily from the northwest. These winds are less intense and more vari-
able; and, consequently, their effect on circulation of surface waters in the
study area is less pronounced than in winter.
2.3 RIVER RUNOFF AND SEDIMENT DISCHARGE
Table 2.1 shows runoff data for the major rivers that discharge into
the study region. The data show that of the total annual runoff, which is
about 116 x 10 9 m 3 , greater than 90% is derived from two rivers, the Fraser
River and the Skagit River.
The Fraser River drains an area of approximately 207,000 km 2 , which
consists of most of the Interior Plateau of British Columbia between the
Coast Mountains and the Rocky Mountains. The elevated nature of the Interior
Plateau determines the main feature of the discharge pattern which is the
spring floods derived from melting snow. Figures 2.2a and 2.2b show the
range, mean, and standard deviation of the water and sediment discharge at
Hope, B.C., for the period between 1967 and 1977 (Water Survey of Canada,
1967-1977). The annual mean flow is 2960 cubic meters per second and the
total annual sediment discharge is approximately 18 million metric tons
(Table 2.2). The maximum discharge of water and suspended sediments occurs
during the months of April through August.
5
-------
TABLE 2.1 Average annual runoff from the major rivers discharging
into the study region
River
Area
Ba
of Drainage
sin (km 2 )
Average
(m 3
Annual Runoff
x 1O )
Fraser*
207,000
93.4
Skagit**
8,040
14.2
Nooksack**
3,270
4.6
Stillaguamish**
1,790
3.6
*Water Survey of Canada, Water Data, Canadian Rivers, Inland Waters
Directorate, Ottawa, Canada, 1967—1977.
**pacjfjc Northwest River Basins Corr ission. Puget Sound Task Force.
Comprehensive Study of Water and Related Land Resources, Puget Sound
and Adjacent Waters, State of Washington, Appendix III, Hydrology
and Natural Environment, March 1970.
6
-------
TABLE 2.2
Comparison of mean water and sediment discharge for the Fraser and Skagit Rivers
during period of record (1967-1977)
Discharge
Year
Fraser River at Hope**
Skagit River at Mt. Vernon***
River Discharge
(m 3 /sec)
Sediment Discharge
(metric tons/day)
River Discharge
(m 3 /sec)
Sediment Discharge
(metric tons/day)
1977
2589
36901
393
5670*
1976
3674
86560
497
1907
1975
2649
32683
561
2184
1974
3172
63428
596
3075
1973
2492
44043
379
-
1972
3398
79424
625
-
1971
2832
44544
551
-
1970
2166
32846
379
-
1969
2753
36032
414
-
1968
3313
64362
583
-
1967
3172
64139
464
-
Mean
2928
53178
495
3209
Std. 0ev.
452
19245
93
1715
*January thru September.
**Water Survey of Canada (1967-1977).
***Water Resources Data for Washington, 1967-1977.
—4
-------
O Monthly Range (1967-1977)
Monthly Standard
Deviation (1967-1977)
10.0 - • Monthly Mean (1967-1977)
O Monthly Mean (1977)
U i
m 8.0 -
<0
I
ox
U)
c 6.0.
0 : ’
LdU)
>w,
4.0 -
Figure 2.2 Monthly range, mean, and standard deviation of: a. water discharge; and
b. suspended sediment discharge for the Fraser River at Hope, B.C. for the
period from 1967 through 1977 (Water Survey of Canada, Sediment Data,
Canadian Rivers, Inland Waters Directorate, Ottawa, Canada, 1967-1977).
12.0
a.
FRASER RIVER AT HOPE
-b.
‘4
4
FRASER RIVER AT HOPE
o Monthly Range (1967-1977)
Monthly Standard
Deviation (1967-1977)
‘ Monthly Mean (1967-1977)
A Monthly Mean (1977)
/
/
4
/
‘ 4
4.0
2.0
‘ 4
#
‘4
‘4
‘4
‘4
p
/
/
,
‘4
‘
‘ 4
I I I I I I I I I I
J FM AM J J A SO N 0
J FMAMJ JAS OND
-------
At Hope, the Fraser River emerges from the Coast Mountains and flows
through an alluvial valley to the Strait of Georgia. Near New Westminister
(see Fig. 2.1), the Fraser bifurcates into a major channel or the Main Arm,
which accounts for approximately 90% of the flow, and a minor channel,
known as North Arm, which branches off at Marpole forming Middle Arm. Main
Arm discharges into the Strait of Georgia at Stevenston and North Arm enters
the Strait at Point Grey.
The Skagit River and its tributaries drain an area of approximately
8,040 km 2 . The major tributaries of the Skagit River are the Baker, Cascade,
and Sauk Rivers. The headwaters of the Skagit River and its tributaries are
located in the North Cascade Range, where flow is partially derived from
melting snow. Near Marbiemount, Washington, the Skagit River is joined by
the Cascade River, where it flows through broad flood plains and lowland
areas until it reaches the coast. At a location about 12 km from the coast,
the river splits into three distributaries which flow into Skagit Bay.
Figure 2.3 shows water and sediment discharge data for the Skagit River at
Mt. Vernon, Washington for 1975.* The data show a bimodal distribution for
both water and suspended sediments. The major peaks occur in two periods:
October through December and May through July. The October-December peak is
associated with excessive rainfall in the lowland areas. The May-July peak
corresponds with spring runoff due to melting snow near the headwaters.
*Whjle water discharge records for the Skagit River are fairly complete
from 1930 to the present, very few records for suspended sediments are avail-
able. Therefore, the records for 1975 were used to show the salient features
of the Skagit River discharge pattern.
9
-------
SKAGIT RIVER AT MT. VERNON
• MONTHLY MEAN RIVER DISCHARGE
O MONTHLY MEAN SEDIMENT DISCHARGE
/
\/
I
\ /
I I I I
JAN FEB MAR APR MAY
PERIOD OF
1___ I I I I I I
JUN JUL AUG SEP
RECORD (1975)
Figure 2.3
Monthly mean discharge of water and suspended sediments of the
Skagit River at Mt. Vernon during 1975 (Water Resources Data
for Washington, 1976).
0
w
C l)
LU
C D
0
U)
LU
>
1400
1200 -
1000 -
800-
600-
400-
200
:
I’
I
7
6
w2
4-
04
43
U)
I.-z
zo
1 1.1
3
Wg
LU
2
‘S
— — ——
-0
OCT NOV DEC
-------
3. CONCLUSIONS
3.1 LANDSAT IMAGERY
A set of LANDSAT images, obtained during the period between 1972 and
1978, have been utilized to study the surface trajectories of sediment
plumes originating from the Fraser and Skagit Rivers. These plumes are
natural tracers of the flow patterns of river water that are discharged
into northern Puget Sound from these rivers.
Three separate plumes can be observed emanating from the mouths of the
distributaries of the Fraser River. The plumes from Main Arm and Middle
Arm join together to form a well-defined jet which can be traced across the
Strait of Georgia and through Porlier, Active, and Boundary Passes. During
ebb tide the plume is directed to the southeast from a point about midway
between Steveston and Porlier Pass. The flood tide drives the plume across
the Strait and to the northwest along the northern coast of Galiano Island.
The plume from North Arm moves to the northwest past Point Grey where it
bifurcates, with some material flowing to the northwest during ebb tide and
the remaining material moving into Burrard Inlet.
The existence of a number of separate plumes in some of the images
suggests that the plumes are probably capable of maintaining their identity
for periods longer than a single cycle. This implies that a dynamic balance
exists between the inertial and pressure forces associated with the plumes
and the coriolis and tidal forces associated with circulation patterns in
the Strait.
The plumes from the distributaries of the Skagit River are most pro-
nounced in early suniner, during the peak runoff. During this period sus-
pended sediments from the Skagit River can be traced as far south as the
middle of Saratoga Passage and as far north as Deception Pass. During ebb
tide, suspended solids from the Skagit River flow into Rosario Strait.
During the period of low discharge, the Skagit River plume is confined to
the nearshore regions of Skagit Bay.
3.2 QUANTITATIVE SEDIMENT MAPPING
A stepwise multiple linear regression analysis routine was employed to
evaluate the relationships between the image radiance from the MSS Computer
Compatible Tapes and sea-truth measurements of suspended matter concentra-
tions. The results show that for the range of suspended matter concentra-
tions observed in northern Puget Sound ( 0.2 - 10.0 mg/i) MSS Band 5
radiance values, corrected for atmospheric interferences, are the most
accurate for mapping suspended matter distributions. Consistent data, with
relative errors on the order of 15—20%, can be obtained using relatively
uncomplicated analytical procedures. The sediment distribution maps so gen-
erated are consistent with previously published data and provide greater
detail than can be obtained with conventional sampling techniques.
11
-------
4. RECOMMENDATIONS
Results of the present study indicate that the LANDSAT images are
extremely useful for studying and mapping sediment plumes originating from
the Fraser and Skagit Rivers. These data allow for synoptic observations of
variations in plume trajectories which can be related to changes in the phy-
sical dynamics of the circulation patterns. In future studies of this kind,
it may be more desirable to collect MSS data from high altitude airplanes
rather than satellites. This would provtde the kind of flexibility required
to carry out more sophisticated field experiments.
12
-------
5. METHODOLOGY
5.1 GENERAL STATEMENT
It is evident from the discussion in the introduction of this report that
suspended materials, originating from the Fraser and Skagit Rivers, can play
a major role in the dispersal and deposition of some fractions of crude oil
and petroleum products that may be released as a result of oil transportation
through the Strait of Georgia-Strait of Juan de Fuca system. Consequently, a
comprehensive understanding of the dispersal mechanisms and trajectories of
suspended matter from these rivers is essential to any model which attempts
to predict the fates and potential impacts of oil in these waters. In recent
years several advances in remote sensing techniques have allowed scientists
to obtain synoptic information about water circulation and sediment dispersal
patterns which heretofore have been unobtainable. The multispectral scanner
images from the LANDSAT—1 and LANDSAT-2 satellites have been especially use-
ful for the study of suspended matter transport processes in coastal and
estuarine waters (Kritkos et al., 1974; Klemas et al., 1974; Gatto, 1976;
Kiemas and Polis, 1977; and Johnson et al., 1977).
Table 5.1 shows the major characteristics of the LANDSAT satellites and
multispectral scanner. The multispectral scanner has four spectral bands.
The green band (0.5-0.6 jim) and the red band (0.6-0.7 jim) provide information
about dispersal patterns of suspended matter from major source regions such
as rivers, streams, and coastal outfalls. The far red band (0.7-0.8 jim) only
shows the core of the sediment plume and the near infrared band (0.8-1.1 pm)
shows the shoreline. The data from the multispectral scanner is transmitted
to a receiving station and then is converted to a photographic product.
5.2 LANDSAT IMAGERY AND HYDROGRAPHIC DATA
In order to provide synoptic information about variations of suspended
matter dispersal patterns, water trajectories, and tidal mixing patterns, all
clear (10% cloud cover or less) LANDSAT imagery from 1972 to the present
were collected and correlated with corresponding data on water and sediment
discharge (U.S. Geological Survey and Water Survey of Canada), water circu-
lation (appropriate published literature as cited), and tides and tidal cur-
rents (Tide Tables, National Ocean Survey). LANDSAT photographs (MSS bands 4
and 5) were found to be ideal for studying the dispersal of turbid water.
The sediment plumes, which appear lighter in tone than less turbid water in
the images, are natural tracers of the low salinity water that exists off-
shore from the Fraser and Skagit River estuaries. However, the LANDSAT
imagery only provides information about the upper few meters of the water
column. Fortunately, previous studies have shown that the brackish water is
generally confined to the upper 5 meters or less (Waldichuk, 1957; Tabata,
1972; and Schumacher et al., 1978) and, therefore, the LANDSAT imagery can be
used to study the migration patterns of the low salinity water provided the
sediment loading is sufficiently higher than the surrounding water.
13
-------
TABLE 5.1
Principal characteristics of the LANDSAT satellite and
multispectral scanner
System Parameters
Satellite
Specifications
Altitude
Type of Orbit
Orbits per Day
Coverage Cycle
Time of Observation
Size of Area Imaged
Field of View
Side Lap
Overlap Along Orbit
Mul tispectral Scanner
Image Distortion
Ground Resolution
Position Accuracy
Spectral Bands, Bandwidths (urn),
and Nominal Color
915 km
circular, sun synchronous
14
18 days
approx. 1000
185 x 185 km
11.56°
67%
10%
2%
80 to 120 meters
900 meters
a.m. at 500 north latitude
4: 0.5-0.6;
5: 0.6-0.7;
6: 0.7-0.8;
7: 0.8-1.1;
green
red
far red
near infrared
14
-------
F!ight
Figure 5.1 Locations of sampling stations for sea truth measurements.
122°40
. 1 y
ST 4TI ONS
OCCUPIED
Flight
5/77
2 /78
Flight
15
-------
5.3 SEA—TRUTH MEASUREMENTS
On three separate occasions (July 31 and September 5, 1977 and June 2,
1978) a Cessna 206 seaplane was employed to collect surface water samples at
approximately the same time of day as the satellite overpasses (Fig. 5.1).
Water samples were collected in acid cleaned 1-1 polyethylene bottles held
approximately 5-10 cm below the air-seawater interface. The samples were
returned to the laboratory and within 9 hours were filtered through pre-
weighed 0.4 um pore size 47 mm Nuclepore filters. The filters were
washed with three 15 ml aliquots of deionized membrane filtered water, dried
in a desiccator and reweighed on a Cahn Electrobalance. The samples were
then treated with 15% H 2 0 2 to remove oxidizable organic matter (Baker, 1973),
dried in a desiccator and weighed again.
5.4 MSS COMPUTER COMPATIBLE TAPES
MSS Computer Compatible Tapes of the LANDSAT images, which were supplied
by the EROS Data Center, Sioux Falls, S.D., were processed in the following
manner. Sample locations were found on the photographic images from which
pixel coordinates were determined and using those coordinates, radiance
values in the four multispectral bands were recalled from the MSS computer
tapes. Average radiance values for a 7 x 7 pixel field corresponding to the
sample points were then determined, and atmospheric corrections were applied
to the previous data by determining the radiance values of clear water and
subtracting them from each of the measured radiance values. Finally, step-
wise multiple linear regression techniques were employed to quantitatively
relate the remotely sensed data to the sea-truth measurements.
The atmospheric correction just described assumes that atmospheric
effects are constant over the entire area of interest and the resulting
corrected radiance values are attributed to the suspended material in the
water. This assumption is only true for regions which are free of cloud
cover as suggested by the discussion of Bowker and Witte (1975). Therefore,
the atmospheric corrections discussed in the following sections of this re-
port have only been applied to specific scenes where cloud cover is limited
to areas which do not affect the integrity of the data.
16
-------
6. RESULTS AND DISCUSSION
6.1 SEASONAL VARIATIONS OF SEDIMENT DISPERSAL
The amount of useful information about seasonal variations of suspended
sediment distributions and dispersal patterns that can be obtained from
LANDSAT imagery is somewhat limited because most cloud-free days occur during
the sumer; and, as a result, the majority of available images are from that
season. However, two high quality images* from January 1973 (Figs. 6.1 and
6.2) and one from March 1976 (Fig. 6.3) are available. Therefore, some gen-
eral statements about sediment dispersal patterns for winter, spring, and
summer can be made.
When the fresh water and sediment discharge from the Fraser River are
high, particularly during the spring floods and during the ebbing tide, the
longitudinal pressure head at the river mouth produces a well-defined turbid
plume which extends outward into the Strait of Georgia. The orientation of
the plume axis can be used to determine the direction of flow. During the
months of May through August, the southward flow of Main Arm and Middle Arm
is so strong that the resultant plume maintains its identity for a consider-
able distance and, in some cases, traverses the entire length of the Strait.
Figures 6.4, 6.6, and 6.9 are examples of this phenomenon. Depending upon
local changes in tidal currents, the plumes extend either to the southeast or
southwest from the river mouth. Tidal mixing is rapid and numerous eddies
are apparent.
At the mouths of Middle Arm and North Arm there was some evidence of
flow to the northwest in most of the images. This feature is particularly
evident in Figures 6.6 and 6.7. These plumes appear to maintain their iden-
tity as far north as Point Grey and the entrance to Burrard Inlet.
The Skagit River plume flows into Skagit Bay and Saratoga Passage during
the summer months (Figs. 6.4 thru 6.9). Some plume material passes through
Deception Pass into Rosario Strait during ebb tide (Fig. 6.8).
The winter plumes from the Fraser and Skagit Rivers, as shown in the
LANDSAT images from January 1973 and March 1976 (Figs. 6.1 thru 6.3), are
much less distinct than that observed during summer. For instance, the
Fraser River plumes in January and March are only visible in the nearshore
regions just a few kilometers seaward of the river mouth, where the primary
flow pattern is to the southeast. During this period, the discharge of both
water and suspended sediments are at a minimum. Average discharge rates for
water and sediments are approximately 10-30% of the mean annual rates (see
Fig. 2.2). This means that both the longitudinal pressure gradient and the
suspended sediment concentrations are greatly reduced in winter. Consequently,
*Specific characteristics of the images are uniquely identified in the
figures by numbered markers and are described both in the text and in the
figure captions.
17
-------
Fig. 6.1 MSs bana 5 of LANDSAT image 1169-18373 on Jan. 8, 1973 between
slack water and minor ebb current. Time of image acquisition
was 1037 PST. The image shows little evidence of sediment dis-
persal from the Fraser River into the Strait of Georgia, reflect-
ing the lower water and sediment discharge during the winter
season. The image is free of cloud cover.
TIDAL CURRENT: TIDES:
Slack Max. Vel.
Water Current (kts) Time
0515 1.7 High @ 0834
0906 1209 1.7 Low @ 1429
WINDS:
Ht.(m) Time of image SE @1.6 kms/hr
2.9 12 hrs earlier E @ 4.8 kms/hr
1.6 24 hrs earlier E @ 19 kms/hr
18
-------
Fig. 6.2 MSS Band 5 of LANDSAT image 1187-18374 on Jan. 26, 1973. Time of
image acquisition was 1037 PST, between slack water and major ebb
current. The image shows limited dispersal of suspended sediments
into the Strait of Georgia from the Fraser River.
TIDAL CURRENT:
Slack Max.
Water Current
0508 0733
1008 1501
Vel.
(kts)
0.9
TIDES:
Ti me
High @ 0905
Ht.(m)
2.6
2.3 Low @ 1811 0.4
19
WINDS:
Time of image E@12.8 kms/hr
12 hrs earlier E@ 9 kms/hr
24 hrs earlier W@ 24 kms/hr
-------
Fig. 6.3 MSS Band 4 of LANDSAT image 2417-18220 on Mar. 14, 1976. Time of
image acquisition was 1022 PST, between minor ebb current and
slack water. The image shows limited dispersal of suspended sedi-
ments into the nearshore waters south and southeast of the mouths
of Main, Middle and North Arms.
TIDAL CURRENT:
Slack Max.
Water Current
0454 0806
1150 1322
Vel.
(kts)
1.8
0.8
TIDES:
Ti me
Low @ 0808
High @ 1402
Ht.(m)
0.8
1.9
Time of image NW @ 37 kms/hr
12 hrs earlier NW @ 19 kms/hr
24 hrs earlier NW @3.2 kms/hr
I ?
, ,—t ,:
; J t
I I F
KILOMETERS
20
-------
Fig. 6.4 MSS Band 5 of LANDSAT image 2111-18254 showing a southeasterly
dispersal (1) of suspended sediments from the Fraser River into
the Strait of Georgia on May 13, 1975. Time of image acquisition
was 1025 PST, between major ebb current and slack water. The
dispersal pattern shows movement of Fraser River sediments into
Haro Strait, primarily through Boundary Pass. Sediment discharge
from the Skagit River into Skagit Bay and Saratoga Passage is
evident (2).
TIDAL CURRENT: TIDES:
Slack Max. Vel.
Water Current (kts) Time
0506 1000 2.9 High @ 0449
1404 1643 2.4 Low @ 1246
WINDS:
Ht.(m) Time of image NW @ 32 kms/hr
2.4 12 hrs earlier N@ 6 kms/hr
—0.5 24 hrs earlier SE@ 12.8 kms/hr
L
20
Ii Iii
KILOMETERS
21
-------
Fiq. 6.5 MSS Band 5 of LANDSAT imaqe 2129-18254 showing a southwesterly
dispersal of suspended sediments from the Fraser River into the
Strait of Georgia on May 31, 1975. Time of imaqe acquisition
was 1025 PST, just after flood current. An anticyclonic gyre
can be observed due west of Pt. Roberts (3). Sediment discharge
from the Nooksack River can be seen in Bellingham Bay (4). Image
is free of cloud cover.
TIDAL CURRENT:
Slack Max.
Water Current Ht. (m) Time of image NW @ 17 kms/hr
0650 1.7 12 hrs earlier SW@4.8 kms/hr
0.3 24 hrs earlier NW @ 14 kms/hr
22
TIDES
Vel.
(kts) Time
weak and High @ 0821
variable
1256 1.6 Low @ 1522
-------
MSS Band 5 of LANDSAT image 1727-18290 showing a southeasterly
dispersal (5) of suspended sediments from the Fraser River into
the Strait of Georgia on July 20, 1974. Time of imageacquisition
was 1029 PST, between major ebb current and slack water. The
dispersal pattern suggests movement of Fraser River sediments
along Saturna, Mayne, and Galiano Islands and into Trincomali
Channel from Porlier Pass. Suspended sediments from the North
Arm appear to flow to the northwest and to the northeast into
Burrard Inlet (6). A cyclonic eddy (7) can be observed north
of Galiano Island. Sediment discharge from the Skagit River into
Sk&git Bay (8), and from the Nooksack River into Bellingham Bay
(9) is clearly visible. Several tidal fronts are also observed
in the Strait of Juan de Fuca.
WINDS:
Vel.
(kts) Time Ht.(m) Time of image SE @ 1.6 kms/hr
3.1 Low @ 1147 -0.7 12 hrs earlier SW @ 1.6 kms/hr
2.0 High @ 1904 2.7 24 hrs earlier SE @ 1.6 kms/hr
TIDAL CURRENT:
Slack Max.
Water Current
0429 0912
1303 1534
TIDES:
Fig. 6.6
23
-------
Fig. 6.7 MSS Band 5 of LANDSAT image 5465-17484 showing a southeasterly
dispersal (10) of suspended sediments fron the Fraser River into
the Strait of Georgia on July 27, 1976. Time of image acquisition
was 1048 PST, between major ebb current and slack water. The
dispersal pattern suggests movement of Fraser River sediments
into Haro Strait from the passages on either side of Mayne and
Saturna Islands (11), and into Rosario Strait from the south-
eastern Strait of Georgia (12). Suspended sediments discharging
into the Strait from North Arm appear to flow northward past Pt.
Grey where it bifurcates, a portion flowina to the northwest,
and the remaining flowing to the northeast into Burrard Inlet (13).
TIDAL CURRENT: TIDES: WINDS:
Slack Max. Vel.
Water Current (kts) Time Ht.(m) Time of image NW @ 25 kms/hr
0408 0904 2.7 High @ 0225 1.9 12 hrs earlier W @ 48 kms/hr
1224 1517 1.8 Low @ 0931 -0.7 24 hrs earlier NW @ 21 kms/hr
24
-------
Fig. 6.8 MSS Band 4 of LANDSAT image 2921-18025 showing southeasterly
dispersal (14) of suspended sediments from the Fraser River
into the Strait of Georgia on July 31, 1977. Time of image
acquisition was 1002 PST, between major ebb current and slack
water. A density front is clearly indicated in the center of
the Strait of Georgia (15). Suspended sediments can be
observed discharging into Skagit Bay and through Deception
Pass into Rosario Strait (16).
TIDAL CURRENT: TIDES:
Slack Max.
Water Current
0921 High @ 0437
1306 1531 Low @ 1149
Vel.
(kts)
3.1
1.8
Time
Ht.(m) Time of image NW 25 knis/hr
2.4 12 hrs earlier S 1.6 kms/hr
-0.4 24 hrs earlier 5 14 kms/hr
;.# A: : 4 - -,
:Y ‘IA
S 4 s
.1g .
25
-------
MSS Band 5 of LANDSAT image 2957-18004 showing a southwesterly
dispersal (17) of suspended sediments from the Fraser River into
the Strait of Georgia on Sept. 5, 1977. Time of image acquisi-
tion was 1000 PST, between major flood current and slack water.
The dispersal pattern suggests southwesterly flow associated with
the flood current. Three small cyclonic eddies can be observed
north of Mayne Island (18, 19 and 20). 40 cloud cover is visible
over most land masses.
TIDAL CURRENT
Slack
Water
Max.
Current
0756
1201 1502
Vel.
(kts)
0.7
0.7
TIDES:
Time Ht.(m)
Low @ 0414 0.4
High @ 1227 1.9
26
WINDS:
Time of image W @ 13 kms/hr
12 hrs earlier SW@ 14 kms/hr
24 hrs earlier S @ 16 kms/hr
Fig. 6.9
-------
- S .-
- .4 ’
-& .:v*t
Its W ’ r
-
It-.. S- —
Sn ‘ - A.F e : r 9
- - — Y
c C
c’. ..
vulr 1 -
t -
- a aew a- i t-4 -. a . v * -
- —
- t -a - tAr -4 “
%iaS4t f S
e s:
- .
r
-
LI C
Fig. 6.10 MSS Band 5 of LANDSAT image 21227-18025 showing a southeasterly
dispersal (21) from the Fraser River into the Strait of Georgia
on June 2, 1978. Time of image acquisition was 1025 PST, shortly
before slack water following major ebb current. Several defini-
tive plumes of sediment laden water are observed coming from Main
Arm (22 and 23). A density front can be seen in the center of
the Strait of Georgia. Plumes are observed dispersing into
Rosario Strait from the north side of Orcas Island.
TIDAL CURRENT: TIDES: WINDS:
Slack Max. Vel.
Water Current (kts.) Time Ht.(m) Time of image NW @ 19 kms/hr
0630 2.8 Low @ 0945 0.1 12 hrs earlier NW @ 11 kms/hr
1050 1340 1.2 High @ 1623 2.2 24 hrs earlier N @ 11 kms/hr
I -
0
20 P
I i i 1 = 1
KILOMETERS
I
27
-------
the forces which drive the Fraser River plume into the Strait of Georgia are
diminished in winter, resulting in more rapid mixing of the plume and concur-
rent dilution of suspended sediment concentrations to background levels.
These conclusions are substantiated by the results of the seasonal surveys of
suspended matter concentrations in the vicinity of the Fraser River which
were presented in a previous report (Baker et al., 1978). The background
concentration of suspended matter in the study region was found to be approxi-
mately 1.0 mg/i. Near the mouth of the Fraser River particulate concentra-
tions ranged between 2.5 and 9.0 mg/l during the November and August surveys
(see, for example, stations 25, 26, and 27 in Figs. 5.4 and 5.6 in Baker
et al., 1978). However, in March the particulate concentration for station 25
was 0.9 mg/i or approximately the same as background values for this region
and time of year, indicating rapid dilution of the Fraser River plume as it
discharges into the Strait.
The Skagit River also shows a minimum in water and sediment discharge
during the winter period (Fig. 2.3). However, the percentage decrease is not
as large as the winter decrease of the Fraser River (the average water and
sediment discharge rates for the Skagit River during the months of January
through March are 50-70% of the mean annual rates); and, as a result, the
seasonal variations of the Skagit River plume are not nearly as pronounced.
The LPINDSAT images for January 1973 and March 1976 show evidence for movement
of suspended sediments from the Skagit River into the nearshore regions of
Skagit Bay (Figs. 6.2 and 6.3). This material mixes rapidly with the offshore
water and the plumes do not maintain their identity beyond Deception Pass.
6.2 TIDAL VARIATIONS OF SEDIMENT DISPERSAL
As indicated in Table 5.1, the LANDSAT satellites maintain a sunsynchro-
nous orbit around the earth in such a manner as to arrive at the same location
at approximately the same time of day once every 18 days. Because the average
tidal interval is about 12 hours 25 minutes, the tides will be out of phase by
approximately one quarter cycle each time the satellite passes over. This
means that over the course of several years the LANDSAT imagery for a given
location should include all tidal stages if proper atmospheric conditions have
prevailed. Since their deployment in 1972 and 1975, the LANDSAT satellites
have produced approximately nine or ten high quality images which can be used
for describing tidal effects on the dispersion of the Fraser River plume.
Figures 6.3 through 6.10 show examples of LANDSAT imagery which are of
sufficient quality to be useful for studying tidal processes. The ebb tides
are represented by Figures 6.3, 6.4, 6.6, 6.7, 6.8, and 6.10 and flood tides
are indicated in Figures 6.5 and 6.9. The tidal stage corresponding to each
of the LANDSAT images has been determined and is included in the figure cap-
tions. Patos Island was used as a reference point for determination of the
tidal stage because: (1) it is centrally located with respect to the
northern part of the study region and, consequently, differences in tidal
stages are relatively small (e.g., the tides at the mouth of Main Arm precede
those at Patos Island by about 2-4 minutes, whereas the tides at Deception
Pass precede those at Patos Island by about 77 minutes); and (2) it provided
the most complete set of data on tides and tide currents.
28
-------
During ebb tide, distinct sediment plumes originating from Main Arm and
Middle Arm flow seaward in a southwesterly direction to a point approximately
midway between Steveston and Porlier Pass where the plumes are diverted in a
southeasterly direction by the ebb flow in the Strait. Figures 6.4, 6.6, and
6.7 are examples of this process. The sediment plumes migrate with the flow
across the Strait and through the passages between Orcas, Saturna, Mayne, and
Galiano Islands into Haro Strait. In one case the tidal flow was so strong
that the sediment plume could be traced as far east as the passage between
Orcas and Lummi Islands (Fig. 6.7). However, the prevailing northwesterly
and westerly winds averaged 30 kms/h during the 24-hour period preceding the
time of image and may have added to this effect (Pacific Weather Center,
Envi ronment Canada).
Numerous cyclonic and anticylonic eddies are associated with the plumes.
These eddies, which are probably the result of interactions between inertial,
tidal, and coriolis forces, maintain their integrity for several hours. For
instance, in Figure 6.6, a cyclonic eddy is observed in the region north of
Galiano Island. Since it appears to be detached from the major plume emanat-
ing from the Fraser River, it probably represents a relict bolus that moved
into the region during a previous tide. This means that at least during some
seasons, boluses of low salinity water derived from the Fraser River are
probably capable of maintaining their identity for periods longer than a
single tidal cycle.
The sediment plume emanating from North Arm during ebb tide flows to the
north and west past Point Grey (Figs. 6.6 and 6.7). Beyond Point Grey the
plume bifurcates, with a portion of the plume moving to the northwest and the
remaining material flowing to the northeast into Burrard Inlet.
The dispersal pattern of the Fraser River plume during flood tide has
some characteristics which are very similar to the pattern for ebb tide and
some which are uniquely different. As indicated in Figures 6.5 and 6.9, the
sediment plumes originating from Main Arm and Middle Arm flow seaward in a
southwesterly direction. However, instead of being diverted to the southeast,
as is the case during ebb tide, the flood current drives the sediment plume
across the Strait and to the northwest along the northern coast of Galiano
Island. Some material from the plume passes into Trinconiali Channel from
Porlier Pass and the remaining material continues to flow to the northwest,
mixing with and being diluted by the less turbid water of the northwestern
Strait.
There is evidence for some small cyclonic and anticyclonic eddies in the
plumes associated with the flooding tide. These eddies are interpreted as
evidence for interactions between the longitudinal pressure gradient, asso-
ciated with the fresh water input from the Fraser River, and the tidal and
coriolis forces associated with water movement in the Strait during flood
tide (Tabata, 1972). It is interesting to note that the anticyclonic eddies
in the LANDSAT images are most pronounced when the tidal and coriolis forces
are additive. The large anticyclonic eddy due west of Point Roberts in
Figure 6.5 is an example of this situation. The plume appears to he split
into two parts, presumably by tidal and coriolis forces. These processes
increase the rate of lateral mixing and subsequent dilution of the plume.
29
-------
The sediment plume from North Arm flows to the north around Point Grey and
into Burrard Inlet during flood tide. There is no evidence from the LANDSAT
images of sediment movement to the northwest during this stage of the tide.
6.3 DENSITY FRONTS
The LANDSAT images showed density fronts which separated light colored,
particle-laden river water from the clearer, darker seawater. Kiemas and
Polis (1977) state that density fronts in estuarine waters are regions where
the turbid brackish water contacts seawater. The denser seawater tends to
underride the light brackish water, creating an inclined interface (Fig. 6.11).
At the interface downwelling occurs in most cases, causing the inorganic and
organic materials from the rivers, and any associated contaminants which have
concentrated along the front such as surface films, foams, and oil slicks, to
be carried down into the water column. Fronts of this type can be seen at the
mouth of the Fraser River (Figs. 6.4, 6.7, 6.8 and 6.10), at Deception Pass
(Fig. 6.8), in Haro and Rosario Straits (Figs. 6.4, 6.5, 6.6, 6.7 and 6.8),
and in the Strait of Juan de Fuca (Figs. 6.4, 6.6 and 6.10).
DETRITUS
LINE
FOAM
LINE
COLOR
LINE
Schematic diagram of a vertical section normal to a
density front (from Kiemas and Polis, 1977).
DENSITY
I NTERFACE
Figure 6.11
\ CONVERGENCE
\\LINE
30
-------
In an earlier report (Cannon, Holbrook and Feely, 1978), Constance
Sawyer has sketched the distribution of density fronts on charts of tidal
currents from a barotropic model of the Strait of Georgia-Strait of Juan
de Fuca system. The sketches are reproduced in Figures 6.13a and 6.13b.
Figure 6.12 is a photograph taken from the plane on June 20, 1978, slightly
northeast of Active Pass following major ebb current, showing a density
front. Sawyer summarizes the distributions of the density fronts by stating
that..
Fronts are found inclined to the current at all angles,
but tend to parallel the ebb current which is stronger and
has stronger lateral shear. Plumes that were visible during
ebb tide at Admiralty Inlet, at Deception Pass, and at the
mouths of some of the smaller rivers disappeared during
flood tides. Fronts parallel to strong currents appeared
at the boundary of the Fraser plume (especially during flood
tide), at the entrances to Haro and Rosario Straits, and in
the western Strait of Juan de Fuca. The parallel fronts
suggested strong cross-channel shear of the along-channel
current, and the cross-channel fronts suggested a gentler
sloshing back and forth of the sediment-laden water. Sepa-
ration of data according to wind direction showed little
dependence of frontal structure on wind.
The separation of cross-channel fronts was up to 20 or
30 km. The tidal excursions estimated were 31 km for the
ebb tide and 11 km for the flood, and they reasonably ex-
plained the spacing of fronts that crossed the Strait of
Juan de Fuca and San Juan Island passages.
Low altitude aerial photograph showing sharp color
boundary of a density front.
Figure 6.12.
31
-------
Figure 6.13a Composite sketch of fronts seen in 24
during ebb tides (from Cannon et al.,
LANDSAT images
1978).
Figure 6.13b Composite sketch of fronts seen in 9 LANDSAT images
during flood tides (from Cannon et al., 1978).
VANCOUVER
ISLAND
— MAXIMUM EBB
VANCOUVER
ISLAND
MAXIMUM FLOOD
O$O(’hrt 16 MARCH. 1973
I ’
32
-------
6.4 QUANTITATIVE SEDIMENT MAPPING
In order to develop a quantitative relationship between the image
radiance from the MSS computer tapes and the sea-truth data, a stepwise
multiple linear regression analysis technique was applied to the data in a
manner similar to the presentation of Johnson (1975). The stepwise regres-
sion analysis program determines the one independent variable, radiance, that
provides the best statistical correlation with the dependent variable, total
suspended matter. The program then determines a second independent variable
to be added that improves the correlation. This process is repeated until
all independent variables have been included in the correlation. During
each step of the process, statistical tests are performed to determine
whether or not the inclusion of additional independent variables increases
the correlation to a degree which is statistically significant. The data that
were used for this analysis are presented in Table 6.1. The analysis was per-
formed on the data using total suspended matter (TSM) and inorganic suspended
matter (ISM) as separate dependent variabies.* Tables 6.2 thru 6.5 show the
results of the stepwise regression analysis of the data.
Table 6.2 shows that when TSM is the dependent variable all the MSS Bands
correlate with it to some degree. The highest correlation coefficient (0.938)
was associated with MSS Band 5, indicating that Band 5 may be sufficient for
mapping suspended matter concentrations in northern Puget Sound. This
hypothesis is supported by the standard estimate of error calculations in the
stepwise regression analysis (Table 6.3), which show approximately the same
standard error of estimate for a regression including only Band 5 as com-
pared with linear combinations of the other multispectral bands. These re-
suits are substantiated by the work of Bowker et al. (1977) which demonstrated
that for suspended matter concentrations below 15 mg/i in the lower Chesapeake
Bay Band 5 radiance values provided the best statistical correlations.
Figure 6.14a shows a plot of the relationship between Band 5 radiance in units
of mW cm 2 ster’ . 100 and TSM for the three sets of sea-truth measurements.
Deviations about the fitted regression line are approximately random, indi-
cating that the linear model is adequate over the range of observed suspended
matter concentrations. The good agreement for the three data sets also indi-
cates that the linear model is consistent from one year to the next, provided
atmospheric interferences are removed and the physical and chemical character-
istics of the particles do not change dramatically.
Tables 6.4 and 6.5 show the results of the stepwise regression analysis
using the ISM as the dependent variable. The general trend of Band 5 radi-
ance providing the best statistical correlation is the same as the case for
ISM. Furthermore, a plot of the relationship between Band 5 radiance and
ISM (Fig. 6.14b) yields approximately the same slope and intercept as for
ISM which means that once the data has been corrected for atmospheric inter-
ferences ISM becomes the major contributor to the observed variations of the
*Inorganic suspended matter concentrations were calculated by gravi-
metrically determining the amount of material remaining on the filters after
the H 2 0 2 treatment (see section 5.3).
33
-------
Table 6.1 Results of sea truth measurements and remotely sensed radiances
for selected locations near the mouths of the Fraser and Skagit
Rivers. The station locations are shown in Figure 5.1.
Sea Truth Remotely Sensed Radiances
(mW .cm 2 sr’) 100
Suspended Inorganic
Collection Solids Matter
Date Station (mg/l) (nig/l) Band 4 Band 5 Band 6 Band 7
June 2, 1978 2 8.594 7.600 11.8 16.7 16.2 3.3
3 3.087 2.597 9.7 9.1 7.53 0.5
4 5.510 4.777 9.0 8.3 6.9 0.3
5 8.290 7.760 11.6 12.6 8.9 0.3
6 5.576 4.710 7.3 7.4 6.8 0.2
7 7.586 6.928 12.4 13.9 9.0 0.3
8 6.436 5.809 11.3 12.7 7.7 0.2
9 4.168 3.643 9.6 8.8 6.9 0.1
10 3.285 2.025 3.2 2.7 4.4 0.1
11 1.763 0.924 2.3 2.1 4.6 0.1
13 3.115 2.030 4.4 4.0 5.7 0.3
Aily 31, 1977 1 0.490 0.250 1.3 1.7 2.4 0.0
2 0.280 0.030 1.6 2.0 2.4 0.0
5 0.730 0.380 4.4 4.2 3.5 0.0
8 2.230 1.810 4.3 3.6 3.1 0.0
9 0.720 0.380 2.8 2.6 2.9 0.0
10 0.970 0.700 2.0 2.1 2.3 0.0
11 0.830 0.610 2.8 2.3 2.4 0.0
13 0.740 0.470 2.5 2.5 2.8 0.0
14 0.740 0.460 2.2 2.3 2.7 0.0
15 1.500 0.950 3.1 2.7 2.8 0.0
Sept 5, 1977 1 1.970 1.510 1.9 2.4 1.8 0.0
2 0.580 0.510 1.4 1.6 1.6 0.0
3 1.880 1.500 1.1 1.7 1.5 0.0
5 0.770 0.630 1.6 2.0 2.2 0.0
6 0.570 0.440 1.8 2.3 2.0 0.0
7 0.230 1.060 1.4 2.1 2.0 0.0
8 1.260 0.990 1.6 2.4 2.1 0.0
9 1.880 1.300 2.1 2.9 2.3 0.0
11 1.280 1.000 1.6 2.0 2.1 0.0
12 1.610 1.340 1.6 1.7 2.0 0.0
13 2.010 1.500 1.1 1.3 1.5 0.0
15 0.310 0.230 1.4 1.9 2.2 0.0
16 1.200 0.930 2.6 2.5 2.3 0.0
17 2.930 2.480 5.8 6.2 4.2 0.2
20 6.770 6.120 11.7 12.7 6.3 0.2
34
-------
TABLE 6.2 Correlation coefficient matrix and linear regression equation
for total suspended matter (TSM) in northern Puget Sound
Band 4 Band 5 Band 6 Band 7 TSM
Band 4 1.000 0.977 0.924 0.630 0.938
Band 5 1.000 0.885 0.504 0.922
Band 6 1.000 0.809 0.886
Band 7 1.000 0.572
TSM 1.000
TSM = 0.0062 + 0.539 [ Band 5 Radiance*]
R = 0.94
*Values corrected for atmospheric interferences.
TABLE 6.3 Results of the stepwise regression analysis of the radiance
data from the MSS computer compatible tapes with the
sea-truth measurements of total suspended matter
Independent Standard Error Regression
Step Variable Regression of Estimate Correlation
Number Added Variable(s) (mg/l) Coefficient
1 Band 5 Band 5 0.855 0.938
2 Band 4 Bands 5 and 4 0.866 0.939
3 Band 6 Bands 5, 4 and 6 0.867 0.940
4 Band 7 Bands 5, 4, 6 and 7 0.830 0.947
35
-------
TABLE 6.4 Correlation coefficient matrix and linear regression equation
for inorganic suspended matter (ISM) in northern Puget Sound
Band 5 Band 4 Band 6 Band 7 ISM
Band 5 1.000 0.977 0.924 0.630 0.950
Band 4 1.000 0.885 0.504 0.927
Band 6 1.000 0.809 0.874
Band 7 1.000 0.562
ISM 1.000
ISM 0.2553 + 0.5025 [ Band 5 Radiance*]
R = 0.95
*Values corrected for atmospheric interferences.
TABLE 6.5 Results of the stepwise regression analysis of the radiance
data from the MSS computer compatible tapes with the
sea—truth measurements of inorganic suspended matter
Independent Standard Error Regression
Step Variable Regression of Estimate Correlation
Number Added Variable(s) (mg/i) Coefficient
1 Band 5 Band 5 0.715 0.950
2 Band 4 Bands 5 and 4 0.726 0.950
3 Band 6 Bands 5, 4 and 6 0.737 0.950
4 Band 7 Bands 5, 4, 6 and 7 0.702 0.956
36
-------
20
15
I0
5
20
15
I0
5
0
‘L..
SE
0
i
0
Figure 6.14
Y = 0.553 t
r =0.938
Sy,x = 1.44
Sx 1 ,y =0.831
6.0
8.0
TOTAL SUSPENDED MATTER(mg/L)
- 0.7km)
7/3 Il 77
A9/S/ 7•7
06 / 2 / 78
a.
—0
0
—0
-A
1_ —
— ‘J
—
1.63
2.0 4.0
10.0
2.0 4.0 6.0 8.0 10.0
INORGANIC SUSPENDED MATTER (mg/L)
Scatter plots of the relationships among Band 5 image radiance
from the LANDSAT multispectral scanner, corrected for atmos-
pheric interferences relative to: a. total suspended matter;
and b. inorganic suspended matter data obtained by gravimetric
analysis of sea truth samples. The dashed lines are standard
errors of estimates for the least-square fit of the data.
37
-------
Band 5 radiance. This conclusion is supported by the laboratory studies of
Whitiock et al. (1977) in which the reflectance of two sediments, Calvert and
Ball clays, were studied as a function of wavelength. Maxima in the reflec-
tance spectra were observed between 550 and 700 nm, with the highest peaks at
about 610-680 nm (Fig. 6.15).
As indicated by the regression equation in Table 6.2, the MSS Band 5
data can be used to quantitatively map near-surface suspended matter concen-
trations with an estimated accuracy of about ± 20%. Figure 6.16 shows an
example of a computer output based on the MSS data. The contours for TSM are
fixed at 2.0 mg/i which is slightly more than twice the standard error of
estimate for the fitted data and is significant at the 95% confidence level.
The data are consistent with the ISM contour maps given in Baker et al. (1978)
for the period using convention shipboard sampling procedures. For example,
the data from July 31, 1977 (Fig. 6.16a) show the same approximate range of
suspended matter concentrations (0-8 mg/i) as the field data from the August
1977 MESA cruise (Fig. 5.6a, p. 25). The computer outputs, however, provide
the added advantage of supplying detailed information about the physical
dimensions of the suspended matter plumes and fronts as well as allowing for
estimation of concentration gradients. In Fig. 6.16b, for instance, the den-
sity front that is located in the center of the Strait of Georgia extends
from a point approximately 10 km south of Point Roberts to a point approxi-
mately 15 km north of Porlier Pass, a distance of nearly 40 km. The suspended
matter concentration gradient across the front is estimated to be at least
0.0024 mg/i, based on a minimum concentration difference of 2.0 mg for a dis-
tance of approximately 840 m between contour lines at some locations. This
minimum estimate is limited by the resolution of the mapping procedure which
means that concentration gradients in excess of this estimate could occur
under normal conditions. Nevertheless, the procedure provides useful infor-
mation about the physical characteristics of the density fronts which can be
used to better understand the fate of suspended particles and contaminants
associated with the fronts.
38
-------
7.2
6.4
5.6
4.8
4.0
3.2
2.4
1.6
.8
0
400
CALVERT CLAY
BALL CLAY
600 800 1000
WAVELENGTH, nm
LANDSAT BANDS
1415161 7
Figure 6.15 Upwelled radiance for Calvert clay and Ball clay (from
Whitlock et al., 1977).
C
4)
0
4)
w
C)
z
I-
C)
w
-J
U-
w
39
-------
Figure 6.16. Total suspended matter contour maps as determined from
computer analysis of LANDSAT images. Dates of image
acqusitions are: a. July 31, 1977; b. Sept. 5, 1977;
and c. June 2, 1978.
40
U
Q 1 k
-------
7. SUMMARY AND CONCLUSIONS
LANDSAT images, obtained during the period between 1972 and 1978, have
been utilized to study the surface trajectories of sediment plumes originat-
ing from the Fraser and Skagit Rivers. These plumes are natural tracers of
flow patterns of low salinity water and suspended matter that is discharged
into northern Puget Sound from these rivers.
Three separate plumes can be observed emanating from the distributaries
of the Fraser River. The plumes from Main Arm and Middle Arm join together
to form a well-defined jet which can be traced across the Strait of Georgia
and through Porlier, Active, and Boundary Passes. During ebb tide the plume
is directed to the southeast from a point about midway between Steveston and
Porlier Pass. The flood tide drives the plume across the Strait and to the
northwest along the northern coast of Galiano Island. The presence of a
well-defined jet is indicative of the significance of the longitudinal pres-
sure and inertial forces. When water and sediment discharge is high, parti-
cularly during the spring runoff period, these forces predominate and the
plumes maintain their identity for considerable distances. When discharge is
low these forces are weak and tidal, coriolis, and wind forces predominate,
causing rapid mixing and corresponding dilution of the plumes. The plume
from North Arm moves to the northwest past Point Grey where it bifurcates,
with some material flowing to the northwest during ebb tide and the remaining
material moving into Burrard Inlet. The northward flow of the plume from the
North Arm is probably the result of a combination of a number of forces in-
cluding intertial, pressure, and coriolis.
The existence of a number of separate plumes in some of the images
suggests that eddies of sediment-laden water are probably capable of maintain-
ing their identity for periods longer than a single cycle. This implies that
a dynamic balance exists between the inertial and pressure forces associated
with these boluses and the coriolis and tidal forces associated with circula-
tion patterns in the Strait. From observations of photographs taken near the
mouth of Main Arm, Tabata (1972) suggests that this balance is maintained by
the daily surge of fresh water from the river merging with the already present
low salinity water lying along the southwestern side of the Strait. Our data
support these conclusions, at least for the periods of high runoff. While the
winter data are very limited, there is no evidence for multiple plumes, indi-
cating that mixing is rapid and the lifetimes of the plumes are very short,
probably less than one tidal cycle.
The plumes from the distributaries of the Skagit River are most pro-
nounced in early summer, during the peak runoff. During this period suspended
sediments from the Skagit River can be traced as far south as the middle of
Saratoga Passage and as far north as Deception Pass, where Skagit River sedi-
ments flow into Rosario Strait during ebb tide. During the period of low dis-
charge, the Skagit River plume is confined to the nearshore regions of Skagit
Bay.
A stepwise multiple linear regression analysis was employed to evaluate
the utility of using the digital radiance data from the MSS computer
41
-------
compatible tapes for quantitatively mapping suspended matter distributions.
MSS Band 5 radiance provided the best correlation with total suspended matter
with estimated inaccuracy of the regression equation at about 20% (std error
of estimate of 0.89 mg/i and a mean sediment weight of 4.77 mg/i in the
Fraser River plume).
42
-------
ACKNOWLEDGMENTS
The authors wish to express their sincere appreciation to Wally Hamilton
for his technical support in developing the computer software for use in this
report, and to John Truscott, Sediment Survey Section, Water Survey of Canada
for his assistance in providing sediment discharge data for the Fraser River.
Gratitude is also expressed to Lake Union Air Service, Gary Massoth,
Joyce Quan, Jane Hannuksela, and Ron Roberts for their help in collecting
samples at the study site, and to Sharon Giese for typing the manuscript.
This work was supported by the Office of Energy, Minerals, and Industry,
Office of Research and Development, U.S. Environmental Protection Agency
through an interagency agreement with the Environmental Research Laboratories
of NOAA.
43
-------
BIBLIOGRAPHY
Baker, E. T. 1973. Distribution and Composition of Suspended Sediment in
the Bottom Waters of the Washington Continental Shelf and Slope. J.
Sed. Petrol. 43(3):812—821.
Baker, E. T., J. D. Cline, R. A. Feely, and J. Quan. 1978. Seasonal
Distribution, Trajectory Studies, and Sorption Characteristics of Sus-
pended Particulate Matter in the Northern Puget Sound Region, EPA 600/
7-78-126, 140 p.
Bowker, 0. E., and W. G. Witte. 1977. The Use of LANDSAT for Monitoring
Water Parameters in the Coastal Zone. Proc. of the AIAA Joint Confer-
ence on Satellite Applications of Marine Operations, New Orleans, La.
Cannon, G. A., J. R. Holbrook, and R. A. Feely, Eds. 1978. Circulation
in the Strait of Juan de Fuca: Some Recent Oceanographic Observations,
NOAA Tech. Report ERL 399-PMEL 29. 94 p.
Carison, P. R., T. J. Conomos, J. Janda, and D. H. Peterson. 1975.
Principal Sources and Dispersal Patterns of Suspended Particulate Matter
in Near-shore Surface Water of the Northeast Pacific Ocean, U.S.G.S.
Final Report to the Goddard Flight Center.
Chang, P., S. Pond, and S. Tabata. 1976. Subsurface Currents in the Strait
of Georgia, West of Sturgeon Bank. J. Fish. Res. Bd. Can. 33:2218-
2241.
EPA. 1977. Review of Environmental Issues of the Transportation of Alaskan
North Slope Crude Oil, EPA 600/7-77-046. Office of Energy, Minerals
and Industry, Office of Research and Development, EPA, Washington, D. C.
123 p.
Feely, R. A., J. D. Cline, and G. J. Massoth. 1978. Transport Mechanisms
and Hydrocarbon Adsorption Properties of Suspended Matter in Lower Cook
Inlet. In: Environmental Assessment of Alaskan Continental Shelf,
Environmental Research Laboratories, National Oceanic and Atmospheric
Administration, Boulder, Cob.
Forrester, W. D. 1971. Distribution of Suspended Particles Following the
Wreck of the Tanker “ARROW.” J. Mar. Res. 29:151-170.
Gatto, L. W. 1971. Circulation and Sediment Distribution in Cook Inlet,
Alaska. In: Assessment of Arctic Marine Environment Selected Types.
D. W. Hood ond D. C. Burrell, Eds. Occasional Pub. No. 4. Univ. of
Alaska, Fairbanks, Alaska, pp. 205-228.
44
-------
Herlinveaux, R. H., and J. P. Tully. 1961. Some Oceanographic Features of
Juan de Fuca Strait. J. Fish. Res. Bd. Can. 18:1027-1071.
Johnson, R. W. 1975. Quantitative Sediment Mapping from Remotely Sensed
Multispectral Data, Proc. Fourth Annual Remote Sensing of Earth Resources
Conf., Tullahoma, Tenn.
Johnson, R. W., I. W. Duedall, R. W. Glasgow, and J. R. Proni. 1977.
Quantitative Mapping of Suspended Solids in Wastewater Sludge Plumes
in the New York Bight Apex. J. Water Poll. Contr. Fed. 49:2063-2073.
Kiemas, V., D. Bartlett, R. Philpot, R. Rogers, and L. Reed. 1974. Coastal
and Estuarine Studies with ERTS-I and Skylab. Remote Sensing of
Environment. 3:153-174.
Kiemas, V., and D. F. Polis. 1977. A Study of Density Fronts and Their
Effects on Coastal Pollutants. Remote Sensing of Environment. 6:95-
126.
Kritikos, H., L. Yorinks, and H. Smith. 1974. Suspended Solids Analysis
Using ERTS-A Data. Remote Sensing of Environment. 3:95-126.
NAS. 1975. Petroleum in the Marine Environment, E. G. Wilson, Ed. National
Academy of Science, Washington, D. C. 107 p.
Pacific Northwest River Basins Commission. 1970. Puget Sound Task Force.
Comprehensive Study of Water and Related Land Resources, Puget Sound
and Adjacent Waters, State of Washington, Appendix III, Hydrology and
Natural Environment.
Parker, B. 1977. Tidal Hydrodynamics in the Strait of Juan de Fuca-Strait
of Georgia. NOAA Tech. Rept. NOS—69, NOAA/DOC, Washington, D. C. 56 p.
Rouse, L. J., and 1. M. Coleman. 1976. Circulation Observations in the
Louisiana Bight Using LANDSAT Imagery, Remote Sensing of Environ.
5:55—66.
Schumacher, J. D., C. A. Pearson, R. A. Charnell, and N. P. Laird. 1978.
Regional Response to Forcing in Southern Strait of Georgia. Estuar.
Coast. Mar. Sci. 7:79-91.
Sharma, G. D., F. F. Wright, 3. 3. Burns, and D. C. Burbank. 1974. Sea
Surface Circulation, Sediment Transport, and Marine Mammal Distribution.
Alaska Continental Shelf ERTS Project 110-7, Final Report to NASA,
13 March 1974. Inst. Mar. Sci., Univ. Alaska, Fairbanks. 77 p.
Tabata, S. 1972. The Movement of Fraser River Influenced Surface Water in
the Strait of Georgia as Deduced from a Series of Aerial Photographs.
Dept. Environ. Water Management Service, Mar. Sci. Branch, Pac.
Regional Pac. Mar. Sci., Rept. 72—6. 69 p.
45
-------
Tully, J. P. 1942. Surface Non—tidal Currents in the Approaches to Juan
de Fuca Strait. J. Fish. Res. Bd. Can. No. 83. 169 p.
Waldichuk, M. 1957. Physical Oceanography of the Strait of Georgia, British
Columbia. J. Fish. Res. Bd. Can. 14:321-486.
Water Resources Data for Washington. 1976. U.S. Dept. mt., Geol. Surv.
Water Resources Division. Tacoma, Washington.
Water Survey of Canada. Sediment Data, Canadian Rivers, Inland Waters
Directorate, Ottawa, Canada, 1967-1977.
Water Survey of Canada. Water Data, Canadian Rivers, Inland Waters Director-
ate, Ottawa, 1967-1977.
Whitlock, C. H., J. W. Usry, W. G. Witte, and E. A. Gurganus. 1977.
Laboratory Measurements of Upwelled Radiance and Reflectance Spectra
of Calvert, Ball, Jordan, and Feldspar Soil Sediments. NASA TP 1039.
46
------- |