jjodsuejj.su  pue
             lPiqjnj. ABJO  pay
                             uoiBay
t'00-6^-6/906-Vd3
                      uoioaiojj

-------
                                        EPA-905/9-79-004
                                        January 1979
 RED CLAY TURBIDITY AND  ITS  TRANSPORT
           IN LAKE  SUPERIOR
                 by

            Michael  Sydor
         Richard T.  Clapper
            Gordon J. Oman
            Kirby R.  Stortz

         Physics Department
   University of Minnesota, Duluth
      Duluth, Minnesota  55812
     E.P.A. Grant No. R005175-01
           Project Officer

        Anthony G. Kizlauskas
U.S. Environmental Protection Agency
              Region V
Great Lakes National Program Office
      Chicago, Illinois  60605
GREAT LAKES NATIONAL PROGRAM OFFICE
U.S. ENVIRONMENTAL PROTECTION AGENCY
              REGION V
      CHICAGO, ILLINOIS  60605

-------
                                 DISCLAIMER
off,      rep°r5r has been "viewed by the Great Lakes National Program
Offxce  Region V, U.S. Environmental Protection Agency, and approved for
publication.  Approval does not signify that the contents necessarily
reflect the V1ews and policies of the U.S. Environmental Protection Agency
nor does mention of trade names constitute endorsement or recommendation for
                                   ii

-------
                                  FOREWORD
     The U.S.  Environmental Protection Agency (EPA)  was created because of
increasing public and governmental concern about the dangers of pollution to
the health and welfare of the American people.   Noxious air, foul water, and
spoiled land are tragic testimony to the deterioration of our natural envi-
ronment .

     The Great Lakes National Program Office (GLNPO) of the U.S. EPA  was
established in Region V, Chicago, to provide a specific focus on the water
quality concerns of the Great Lakes.  GLNPO provides funding and personnel
support to the International Joint Commission activities under the U.S.-
Canada Great Lakes Water Quality Agreement.

     Several water quality studies have been funded to support the Upper
Lakes Reference Group (ULRG) under the Agreement to address specific objec-
tives related to pollution in the Upper Lakes (Lake Superior and Lake Huron).
This report describes some of the work supported by this Office to carry out
ULRG study objectives.

     We hope that the information and data contained herein will help
planners and managers of pollution control agencies make better decisions
for carrying forward their pollution control responsibilities.
                                     Dr. Edith J. Tebo
                                     Director
                                     Great Lakes National Program Office
                                    iii

-------
                                   ABSTRACT
      Through the use of  Landsat  satellite  imagery  from  1972  -  1975   red  clay
 plumes  in western Lake Superior  are  studied  in order  to determine the
 relative  magnitude of the  three  sources of the observed turbidity  erosion
 of  the  Wisconsin south shore red clay banks, resuspension of bottom sedi-
 ments,  and runoff from the many  streams which flow through the red  clay  belt
 and  then  into  the lake.  A comprehensive sampling  program was conducted
 during  the spring of  1975  in order to determine the runoff contribution  to
 the  total load observed  in the lake.  Analysis of  Landsat transparency data
 coupled with weather  records enabled contributions from erosion and  resuspen-
 sion  to be separated.  It  was found  that approximately  75% of the observed
 load  in the  lake  during  the ice  free season, from May to November,  is from
 erosion,  20/, is  from  resuspension, and 5% is from runoff.

     A  numerical model for water transports in Lake Superior as a function
 of winds  is  developed.   This model is verified by comparison of observed and
 predicted  water levels at  several locations around the  lake, and by  compari-
 son of  the predicted  transport patterns to actual turbidity distributions
 observed  in  Landsat imagery.   Transport patterns are shown for western Lake
 Superior and the entire  lake for both an easterly and westerly wind.  A model
 of current profile with depth is also developed.   The results of the trans-
 port model are used to predict distributions of red clay from the south
 shore and  taconite tailings discharged into the lake at Silver Bay,  Minnesota.

     This report was submitted in fulfillment of  Grant No.  R005175-01 by the
Great Lakes National Program  Office,  Region V under the sponsorship  of the
U.S. Environmental Protection Agency.  This report  covers  a period from
October, 1974 to March,  1976  and work was  completed as of  January,  1979.
                                    iv

-------
                                   CONTENTS

 Foreword
 Abstract  ................................   .jv
 Figures  ..............................  '.'.'.   v±
 Tables   .................................   ix
 Acknowledgments
                                                                            x
   1.   Introduction  	   i
   2.   Conclusions   	   2
   3.   Recommendations   	   4
   4.   Runoff from Douglas County, Wisconsin   	   5
            Rivers and streams analyzed  	   5
            Method   	   7
            Results	20
            Rain Runoff	27
            Conclusions	32
   5.   Sources and Transports of Turbidity in Western Lake Superior  ...  34
            Estimation of sediment resuspension and total loading ....  34
            Turbidity transport and the distribution of turbidity sources  48
   6.  A Numerical Model of Transports in Lake Superior	56
            Description of the model	55
            Results of transport model  	  61
            Current profile model 	  88
            Turbidity transport model 	  95

References	^   1Q4
Appendix

   A.  South Shore Data for the 1975 Runoff	105

-------

Number
1
2
3
4
5
6
7
8
9

10
11
12

13
14

15
16
17
18

19
FIGURES

General study area . .
Rivers and streams analyzed in Douglas County, Wisconsin .
Nemadji River basin . . .
Comparison of grabbed and integrated water samples . .
Cross sectional areas and discharge rates of analyzed streams
Discharge vs. stage for the analyzed streams 	
Soil map of northern Wisconsin 	
Daily snow and runoff data during snow-runoff period 	
Daily suspended load for analyzed streams during the entire
runoff period 	
Turbidity and suspended solids correlation 	
Total load vs. total discharge for the large streams
Daily rainfall and Nemadji River discharge and suspended
load data for a summer rain event
Turbidity calibration of Landsat images 	
Representative suspended solids distribution produced from
Landsat data 	
Lake grid spaces and regions used in Landsat image analysis
Spring ice shelf near Amnicon River 	
Average lake turbidity near Duluth 	
Average turbidity dissipation for easterly and westerly wind
storms 	
Actual turbidity dissipation for a large easterly storm . . .

Paep
o
6
8
9
10
11
14
19
21

22
26
28

32
35

36
37
41
42

45
45
vi

-------
Number                                                                 Page

  20   Average storm characteristics  	    46

  21   Distribution of storm duration 	    46

  22   Average turbidity distribution and transport for an easterly
         storm	    49

  23   Average turbidity distribution and transport for a westerly
         storm	    50

  24   Average turbidity distribution and transport by westerly
         winds	    51

  25   Average turbidity distribution and transport for variable
         winds	    52

  26   Average turbidity distribution for the  ice-free season  ....    53

  27   Relative south shore erosion rate  	    54

  28   Resuspension areas in Lake Superior	    54

  29   Duluth water levels for  easterly  wind   	    63

  30   Actual Duluth water levels for westerly wind 	    63

  31   Western Lake Superior transports  for easterly storm  	    65

  32   Landsat image for 11APR75  showing resuspension plume  	    68

  33   Skylab image showing turbidity entrapment near Duluth   ....    68

  34   Landsat image for 23NOV73  showing taconite tailings plume
         off  Silver Bay	    69

  35   Western Lake Superior transports  for westerly storm  	    70

  36   Water  level  as  a  function  of  time at Duluth  from the
         transport  model   	    72

  37   Water  levels for  western Lake  Superior  for an easterly  storm  .    73

  38   Water  levels for  western Lake  Superior  for westerly wind ...    73

  39   Water  levels for  western Lake  Superior  from  the  Forristall
        model	    74

  40   Entire Lake  Superior  transports from a  13 m/s NE wind   ....    75
                                    Vll

-------
Number                                                                 „
	                                                                 Page
  41   Entire Lake Superior transports from a 11 m/s westerly wind.  .    80

  42   Lake Superior water level station locations  .........    85

  43   Modeled Lake Superior water levels from a northeast storm  .  .    86

  44   Current vs.  relative depth near surface  ...........    91

  45   Current meter station locations  ...............    91

  46   Current profiles for two  values of v .............    94

  47   Peak measured surface currents  vs.  wind speed for  several
         northeast  winds                                                0
  48    Computer bottom currents near  Duluth  .............   96

  49    Lake  sample  settling  in lab   .................   97


  50    Modeled  suspended  solids distribution at several stages of
         the storm   ....................             no

  51    Landsat  image for  03APR75  ..................
  52   Modeled  sediment accumulation at different stages of the
         storm   ................                        -1Q2
                                    vni

-------
                                   TABLES

Number                                                                 Page

  1    Load Transported during April, 1975 	   25

  2    Snow Runoff per Linear km of Stream Length and per Square km
         of Basin Area for the Period April 10-27	29

  3    Summer Rain Runoff for Nemadji River (1970 - 1974)  	   30

  4    Yearly Input of Suspended Load  	   32

  5    Percent Abundance of Particle Sizes by Number 	   33

  6    Ratio of Average Turbidity to Surface Turbidity for Western
         Lake Superior	38

  7    Total Suspended Load in Western Lake Superior 	   39

  8    Sources of Lake Turbidity	47

  9    Comparison of Periods in Hours of Gravitational Modes in
         Lake Superior	64
                                     IX

-------
                               ACKNOWLEDGMENTS

     We would especially like to thank Dr. Thomas F. Jordan, Physics Depart-
ment, University of Minnesota, Duluth, for his derivation of the current
profile model solution.

     We are grateful to Mr. Steve Diehl, University of Minnesota, Duluth,
for his work on the transport model.

     We are grateful for the assistance of the following individuals who
helped to collect and analyze material in this report: Dr. David G. Darby,
University of Minnesota, Duluth; Dr. Albert B. Dickas, University of
Wisconsin, Superior; Mr. David Smith, Mr. John Sorensen, Mr. Vasyl Shuter,
Mr. Gary Fandrei, and Mr. Wayne Maanum.

     We are particularly grateful for the assistance of Mr. Steven Spray,
whose photographic and sampling efforts were most helpful during the diffi-
cult conditions of early spring.

     We are grateful to Mr. Nelson A. Thomas, U.S. Environmental Protection
Agency, Large Lakes Research Station, Grosse lie, Michigan, for performing
the particle size analysis.

-------
                              1.   INTRODUCTION

     The transport of fine red clay particles from erosion of red clay banks
in Douglas County, Wisconsin, over large areas of the lake causes water
quality degradation and displeasing aesthetic appearance, which together are
often referred to as the Red Clay Problem.  The turbidity within the lake
plumes results from three processes:  lakeshore  erosion,  sediment resuspen-
sion, and stream runoff.  The total turbidity within the entire plume can be
determined through remote sensing.  The relative contribution of each source,
however, must be established from measurements of the individual sources.
The runoff can readily be evaluated with reasonable accuracy.  The magnitude
of the other turbidity sources can be established by subtraction of the
contribution from runoff.  The determination of the relative magnitude of all
three sources of red clay turbidity in the lake is essential in evaluation of
red clay-related problems, such as the deterioration of water quality and
property damage caused by severe erosion of the lakeshore.   In the case of
water quality, it is particularly important to determine the magnitude of the
runoff, since the water quality problems resulting from shore erosion and
sediment resuspension may be quite distinct from those resulting from red
clay washed out by polluted rivers.  For instance, the red clay particles
transported by the Nemadji River could adsorb and carry pollutants into the
lake.  The contribution of runoff to lake turbidity is also useful for deter-
mining the effect of various erosion abatement programs currently in progress
in the red clay belt of Wisconsin and Minnesota.

-------
                                2.   CONCLUSIONS

 1)    Turbidity measurements correlated  with Landsat  satellite  data  show that
 lake turbidity originated from erosion  of  shore  banks,  resuspension of  bottom
 sediment,  and river runoff.   Using  satellite data  from  August,  1972 to
 August,  1975, the contribution of each  source per  open  water season (May -
 November)  was estimated  to be 4.0 x 106 metric tons  from  erosion, 1.0 x 106
 metric  tons  from resuspension,  and  .32  x 106 metric  tons  from  river runoff.
 During  the open water  season,  about 75% of  the total  sediment  load  is contri-
 buted by erosion,  20%  by resuspension,  and  5% by river  runoff.

 2)^  The runoff mainly comes  from the Nemadji River, which accounts  for  over
 80%  of  the total sediment  output from streams and  rivers  in Douglas County,
 Wisconsin.

 3)   Winter  storms  which occur  when the lake is  ice-free  but when  the  lake-
 shore and  the rivers are iced  over,  contribute roughly  2  x 106 metric tons
 per  season of additional suspended  material  in the lake.  This largely  comes
 from sediment resuspension at  the intermediate depths off Wisconsin and
 Minnesota  Points.   Resuspension is  greater during  the winter months because
 of the higher frequency  of northeast winds and more nearly isothermal condi-
 tion of  the  lake.

 4)   A numerical model for transports was devised for the lake.  The model
 was  tested using water level data and by comparing the predicted transports
 to the transport of turbid plumes observed by Landsat.  The results  of  the
 numerical model  and the  statistical analysis  of Landsat data explain the
 periodic contamination of  the Duluth water intake by high concentrations of
 red  clay particles and asbestos particles.

 5)   The fine red clay material contaminates much of western Lake Superior
 from Duluth  to  the Apostle Islands.   The main transport of turbidity for
 northeast winds  occurs along the axis of the western arm of the lake from
 Duluth towards  the Apostle Islands.   For west winds,  the main transports
 occur along  the Wisconsin  shore towards the Apostle Islands.

 6)   The distribution of lake levels as a function of winds is given for
 Duluth and extreme western Lake Superior.   The height of water level fluctua-
 tion due to  storm surges is on the order of 20 cm (8  inches), which  is
 comparable to the seasonal fluctuation in the water level.  Changes in ero-
 sion rates resulting from water level fluctuations could in principle be
obtained from Landsat data on the distribution of turbidity along the shore
after storm  surges.

 7)   General  circulations and transports for the  entire lake  are discussed.
The transport patterns for the two wind directions modeled here do not carry

-------
the fine red clay material directly across the international boundary.  How-
ever, fine red clay particles are transported in the lake in detectable
concentrations (1 mg/£) for over 80 km.

8)   The effective area of contamination for a storm surge and the long-range
transport of fine sediment was assessed by simulating a turbidity plume
generated by a steady two-day northeasterly wind storm with 13 m/sec  (30
mph) winds followed by a westerly wind.  The results show that an estimated
10" metric tons of material would be removed by erosion for such a storm
and dispersed in the area of the lake from Duluth to the Apostle Islands.
Although the computation simulated dispersion of clay for an equivalent of
only a 4-day period, in which time it showed that most of the suspended load
was deposited in western Lake Superior between Duluth and the Apostle Islands,
it is estimated that on the order of 3% of the material or ^ 105 metric tons
per year would disperse outside of this area.  This result is also borne out
in the observation of extensive plumes through remote sensing which in some
instances shows the plumes extending well past the Apostle Islands.

9)   Bottom currents near Duluth for westerly winds favor the presence of an
upwelling along the North Shore.  The results of the calculation on bottom
currents for isothermal conditions indicate that resuspended material off
Minnesota Point would move out along an easterly direction for northeast
winds.  Westerly winds would generally move the resuspended material towards
Duluth,  with the exception of an area east and southeast of the Superior
entry, where the bottom currents point toward the Wisconsin shore.

-------
                              3.  RECOMMENDATIONS

 1)   Correlation of Erosion and Lake Levels.

     Examination of the results from the numerical model shows that the water
 level along  the red clay bank shore is a function of winds.  The change in
 water level  produced by the wind appears to be accompanied by a change in the
 relative  turbidity along the  shore as observed from the satellite.

     Further work including the effects of wind, fetch, and turbidity distri-
 bution along the shore in correlation with water level information obtained
 from the  numerical model should be made.

     A series of manned overflights of the shore area for various wind condi-
 tions using multispectral scanners would be fruitful in establishing the
 dependence of the erosion rates on water levels.  The overflights would also
 help determine the boundary conditions necessary for modeling the shore
 erosion   as a turbidity source.

 2)  Water Quality.

     The development of the numerical model for transports is a prerequisite
 for water quality modeling.  The next step should be concerned with water
 quality modeling of the effects of harbor effluents from the St.  Louis River.
 The effluents tend to accumulate in extreme western Lake Superior for variable
 wind conditions.

     The overall quality of the water in the lake should also be considered.
 The dispersion of very fine particulates and the dispersion of effluents from
various point sources should be studied through remote sensing coupled with
 the modeling of long-range transport events.

     Although it would be impractical to run a finer grid model for the
 entire lake because of computer storage and time requirements,  much higher
resolution could be obtained by submodeling on a finer grid scale at various
 locations of interest around the lake.   For example, future studies might
 entail cutting the grid spacing by one half over the entire western arm of
Lake Superior,  while leaving the remainder of the lake unaltered.   This
would be especially beneficial for the comparison of modeled plumes to actual
 satellite images which contain features comparable in size to the present
grid spacing.

-------
                  4.   RUNOFF FROM DOUGLAS COUNTY,  WISCONSIN

RIVERS AND STREAMS ANALYZED

     Streams and rivers passing through Douglas County,  Wisconsin were moni-
tored during the entire spring runoff in April, 1975.   These streams run
through the major deposit of glacial-lacustrine red clay around Lake Superior.
The deposit averages 30 meters in depth and covers a 15-km-wide area along
nearly all of the Lake Superior shore in Douglas County, Wisconsin,  and
beyond into Bayfield County, Wisconsin.  The clay deposit shows critical
erosion along many areas of the lakeshore and along the river banks  and
road cuts.  A large fraction of the clay is made up of very fine particles
less than 6 y in size which are washed out into Lake Superior and are trans-
ported in the lake in high concentrations as far as 80 km.  The transported
particles from the red plumes are frequently observed in western Lake
Superior and are a source of contamination of municipal water supplies.

     The rivers and streams analyzed in this project include those flowing
into the lake between Dutchman Creek, which enters Lake Superior approximately
2.5 km east of the southeastern end of Allouez Bay, and Haukkala Creek,
Figure 1.  This includes approximately 20 km of lake shoreline and twenty-one
rivers and streams.  In addition, the Nemadji River was also observed.
Although the Nemadji River does not flow directly into Lake Superior but
rather into the Superior Harbor Basin of the Duluth-Superior Harbor, its
mouth is only 800 meters from the Superior entry to the harbor.  Although the
percentage of the Nemadji River's suspended load which is deposited in the
lake has yet to be determined accurately, it is known that much of the fine
load does flow out to the lake.  The volume of the Nemadji's load is so great
that if only 20% of it reached the lake, its contribution would still be
comparable with the combined output of all the other rivers and streams in
Douglas and Bayfield Counties.

     From west to east, the rivers and streams analyzed and the monitoring
sites are as follows:  Nemadji River at the bridge near the Nemadji Golf Club
in south Superior; Dutchman Creek at Hwy 13, 500 meters from the mouth;
Morrison Creek at Hwy 13, 150 meters from the mouth; Amnicon River at Hwy 13;
Ten Creek at Hwy 13, 600 meters upstream from  the intersection with the
Amnicon River; Wagner Creek at Hwy 13, 600 meters upstream from the intersec-
tion with the Amnicon River; Hanson Creek at Hwy 13, 300 meters from the
mouth; Middle River at Hwy 13, 1 km from the mouth; Poplar River at Hwy 13,
800 meters from the mouth; Bardon River at Hwy 13, 800 meters from the
mouth; and Pearson Creek at Hwy 13, 2.5 km from the mouth.

     Within the study area investigated by the University of Minnesota,
Duluth,  there were fifteen unnamed intermittent streams.  For purposes of this
study the unnamed streams were numbered consecutively from west to east,

-------
           LAKE SUPERIOR
MINNESOTA POINT

   SUPERIOR ENTRY
     WISCONSIN POINT
      Figure 1.  General study area.

-------
#1 - #15.  Figure 2 presents a map of all of the rivers and streams analyzed,
wnich incudes an outline of their drainage basins.  Since the United States
Geological Survey has conducted considerable analysis of the Nemadji River
their figure for the total drainage area of this river was used.  An outline
of the Nlmadji Basin which was taken from the report Red Clay Project appears
in Figure 3.
METHOD
     Since it is the suspended material that adversely affects Lake Superior s
water quality, it was only the suspended load, or more specifically the wash-
load! of the^treams that was measured during this study.  Particles included
in the bed-load and the saltation-load of the streams are too heavy to remain
suspended in the lake for any appreciable length of time.

     A hand-held, rod-supported US DH-48 integrating sampler was used to
collect water samples, which were refrigerated  and analyzed within 24 hours.
Parole size analyses were performed on Delected samples using a HIAC en    *
Turbidity was measured with an AG1 Hach Meter, and suspended solids were de
termined by  filtration through 0.45 y Millipore filters.  On deep, swift
rivers such  as  the Nemadji it was impossible  to stand in the river with the
US DH-48 sampler.  Long extensions for the handle were const™c'e* *°*^°
allow sampling  from bridges, culverts, and, at times, trees which had  fallen
across the channels.  To  keep the sampler properly orientated  in the  swift
current  i.e.,  the intake nozzle horizontal  to the stream surface and di-
rected directly upstream, thin nylon  lines were attached to  the handle near
the  sampler  and tied-off  upstream.  Occasionally  during  the  course of sam-
pling "dipped"  or "grabbed"  samples were also obtained for comparison with
?he  Integrated  samples.   A  comparison between integrated and dipped  samples
appears  in Figure 4.

      Stream  discharge rates were determined by measuring stream velocities  (v)
with a  Gurley Meter  (622),  Pygmy Gurley Meter (625), or  both, depending upon
 tne  flow-rate and the accessibility.   At the same time as  the average velocity
was  measured  the depth profile was measured to  determine  the cross-sectional
 «eal5 of  the stream.  Thus discharge (Q) was  determined,  where Q - Av.   In
 additil  marked stakes were driven or marks placed upon permanent fixtures
 such as culverts, bridge abutments, or trees at  each sampling site,   ^ng
 these marks, the stages of the rivers and streams could f****^™*^^
 determined when time or conditions prevented a current dept h prof Jle ^om
 being taken.  Provided that subsequent channel profiling indicated that the
 contour of  the channel had not changed, these stage indications were used to
 interpolate discharge rates.  At the sites where the staging ™**™*\™t
                                                           to  determine the
   nterpoae                .
  in contact with the stream surface,  a hand level was  used to  determ
  difference between the water  surface and the staging  mark.  The cross
  sectional areas for irregularly shaped stream contours were Determined by
  reproducing scaled graphs of  the profiles and tracing the outlines with a
  ptanimeter   Cross sections of the major rivers and streams analyzed appear
  in Figure 5.  Plots of discharge vs. stage are shown in Figure b.
  - Article size analyses were performed by Nelson A. Thomas, U.S.
  Environmental Protection Agency, Large Lakes Research Station, Grosse lie,
  Michigan.

-------
                                                            LAKE SUPERIOR
oo
                                •STfiUWi
                                DRAINAGE SASNS  _
                                SUB-BASINS
                                MONITOBNG SITES .
                                0  DUTCHMAN
A   AMNCON
T   TEN
W   MftGNBt
H   HAMSON

f   K9KAI
ft   MADON
Ai   «ARSON


                                        N
                                         \
                      Figure 2.   Rivers and streams analyzed in  Douglas  County,  Wisconsin.

-------
(I
<

c
hlj
H>
C
i-!
CD
•s

f|
a!

I1
e
S
c
0
C/J
o _
a
* r*
3
en
§
u>
in
(13
H
H-
S
        jjf MINNESOTA -I
oougioi co.""   wi SCO'NSIN*
                   m ni
                   r i; i

-------
             RELATIVE TURBIDITY
250
         50           150           250

                 Grabbed Samples


         Figure 4.  Comparison of grabbed and integrated
                 water samples.
                     10

-------
  6.0





UJ3.0


w 1.5
                        NEMADJI  RIVER CROSS SECTION
                               Bridge near golf course
                                                  Stream Bottom
                                                  Siege
                  DISCHARGE RATE (m3/sec)   163
                                             ^135
                                         35
                                         ~6~
                                  12      15      18
                                       WIDTH (m)
                                                         21
  24
                                                  27
                      4.8


                      4.2


                      3.6


                    -. 3.0
                    6

                    LJ 2.4
                    CO
                    <



                      i.2


                      0.6


                       0
                         DUTCHMAN CREEK CROSS SECTION
                                       HWYI3
                                       Culvert Contour
                                       Stage (entrance) ——
                               DISCHARGE RATE   3.9
                           0  0.6  1.2  1.8  2.4 3.0  3.6  4.2 4.8
                                       WIDTH (m)

                        AMNICON RIVER CROSS  SECTION
                                    HWY 13
                                                     Stream Bottom
                                                     Stage
-2.4
~ 1.8

£ 1-2
<0.6
r

-------
       1.2
     J0.9
     UJ

     <0.6
      0.3
                        TEN CREEK CROSS SECTION
                                    HWY 13
                         Stream Bottom
                         Stage


	DISCHARGE _R ATE Jm3/sec)_ _ _2J5	
                            _0,9_

                             0.2
                      1.5
                                3.0       4.5
                                WIDTH (m)
                                 6.0
7.5
1L2
UJ


CO
                     MIDDLE RIVER CROSS SECTION
                                   HWY 13
                                              Strocm Bottom
                                              Stags
                         DISCHARGE RATE (mVsec)   26
                                                    12
                                 WIDTH (m)
                                              15
                      POPLAR  RIVER CROSS SECTION
                                   MOUTH
                                                    Culvert  Contour
                                   WIDTH (m)

                          Figure  5.   (continued)
                                     12

-------
  0.9
J.
oJ
 .0.6
  0.3
                POPLAR RIVER CROSS  SECTION
                             HWY 13
                                            Stream Bottom
                                            Stage
                       DISCHARGE RATE (mVsec)  5.2
              1.5
3.0
  4.5
WIDTH (m)
                                          6.0
                             7.5
                                                              9.0




-?0.9
-^>
$0.6
£ n ^
0

BARDON CREEK CROSS SECTION
HWY 13






i\A
jDISCHARGE^RATEjir^/sec) 	 V


.^center
Culvert Contour 	
Stage 	
__5.8 	
3.1
0.6
<07L 	
•







0 1.5 3.0 0 1.5 3.0
WIDTH (m)




DISCHARGE
RATE
(mVsec)
1.3
0.7
02- 	
O.I 2^=:
0.05
(
PEARSON CREEK CROSS SECTION
HWY 13
Culvert Contour 	
Stoge — —


III • 'I ^M^ W» — — — • — ^ *» ~
rj= 	 ==__ — — 	 — — — — -^
,1111 i 	 1 	
D 0.3 0.6 0.9 1.2 1.5 1.8 2.1 2
WIDTH (m)




•




,4
                                                                    1.2
                                                                      CO
                      Figure 5.  (continued)
                               13

-------
! DUTCHMAN CRFFK


»•-••*
o
o
y?
•v ! ^*O
fO * ' s-Vy
it
£T
G
.c:
NE^ADJi RiVER
-{bridge rear go!f /' -
course}'*?
/•'
/
-. /
/
/
/
f -
;
1 eoi- y

V
i /
! /
j


°r
/ 1
/
x
.Pi , 1 1
; 1.5 3.0 4.5 6.0
I

03.C
to

?0 "
5
CJ*
o
« s
w?
t-7
0 1.5



1

Hwy 15 X
/ "
' / '
/
/
/
/
/

-------
0.25   0.50   0.75
   Siege (rn)
0   0.3  06  0.9  1.2
         Stage(n)
                      Figure 6.   (continued)
                                                          IS
                                                             POPLAR RIVER
                                                                (mouth)
                                                        o
                                                        o
                                                           •r
                                                        o
                                                       .c:
                                                        o
                                                          0
                                                           0
                                                                      /   !
                                                   / -i
                                       0.6     1.2
                                       Siage(m)

-------
  POPLAR RIVER
    Hwy.13
.oPS^TJ   41l[75SoH««Ern
                 "~ i              VI
0.25   0.50
   Stcge(rn)
 c
  0   0.3  0.6
    Stage(m)
0.25 0.50  0.75
 Stage (m)
              Figure 6.  (continued)

-------
     Snow depth and the moisture content of the snow were measured by coring
the snow at various locations with a 4.45 cm (1.75 inch)  diameter clear
plastic tube.  The depth of the snow at the coring site was measured, and
the core was emptied from the plastic tube into a small plastic bag and
weighed to the nearest gram with a spring scale.   Both the coring tube and
the weighing bags were transparent so that the snow could be examined to be
sure that debris such as grass, mud, sticks, and gravel were not present in
the sample.  Accurate snow measurement depends upon the percent of land
covered by the snow as well as the depth.  As the spring runoff progressed,
larger areas of open ground began to appear.  Photographs were taken to assist
in determining the percent of land covered by the snow as well as the depth.

     River and stream sampling for this project was conducted during the
period April 9 through May 23, 1975, with intensive sampling taking place
during the runoff period April 14 through May 7.  During the period April 14
through May 7 stages were noted and water samples collected at all of the
major rivers and streams on all but six days, of which no more than two were
consecutive.  This period was divided into a snow-runoff period, April 14 -
27, and a rain-runoff event, April 28 - May 7.  Stream velocities were
measured on occasions when significant stage differences were noted.  When-
ever possible, discharge rates were established and water samples were
collected for upstream and downstream sampling sites of the same river.  This
was done in order to estimate  the eroded load per km of stream length for a
particular stage.  Unfortunately, manpower shortages prevented checking the
sampling sites of all the streams on any one day.  During the first half of
the runoff period nearly all of the sampling sites were taken alternately in
order to reduce periods when   data was not collected for the major streams
to not more  than one or two days, although occasionally a three-day period
did elapse.

     The difficulty in reaching good sampling locations caused certain sacri-
fices to be made.  Ideally, sites should be on uniform, straight sections of
the streams.  There should be  no  tributaries or ditches entering above the
location which are not well mixed with  the main stream before reaching the
sampling site.  In addition,  the  site should be on a "high" section  of the
stream with  no obstructions nearby downstream.  This is necessary  to permit
accurate interpolation of the  discharge  from stage readings.  If damming or
restriction  occurs, the backup destroys  the relationship between stage and
discharge.   One other factor  enters  into site selection when  testing near
the lake.  The sampling sites  must be far  enough  upstream to  prevent lake
level from altering water samples and current measurements.   This  was  satis-
fied  for times of high flow for all  rivers.  The  summertime low-flow periods
on  the Poplar River were subject  to  storage variation  because of  the gravel
bar buildup  at the mouth of  the  river.

     The slope of  the stage vs. discharge  curve  is nearly linear  except  at
the peak flow or when a river  rises  above  its normal banks  and  flows onto
the flood  plain.   If  testing  is made  in well-confined  areas with deep  banks,
the change of slope due to  flood  plain  overflow  need not be determined.
                                      17

-------
      In one way or another nearly all of the sampling sites fell short of
 the criteria just mentioned.   However, the best locations which were reason-
 ably convenient were selected, and appropriate data adjustments were made.

      The sheer number of streams in this area prevented most of the intermit-
 tent streams (#1 - #15)  as well as Hanson Creek from being monitored except by
 occasional visual inspection.   The soil map of the entire test area, Figure
 7,  and topographical maps of  the area, Figure 2,  reveal that the soil and
 topography of the test area are quite uniform.   Therefore,  estimates of eroded
 load for the streams not monitored were determined by comparison of their
 basin areas and stream lengths to those for which considerable data was
 collected.   All values determined strictly by estimate or interpolation
 appear enclosed in parentheses in the data tables in Appendix A.

      Many daily values of discharge and load for  the major streams  had to be
 determined by interpolating from known values  before and after the  day.   The
 accuracy of interpolated values depends on the accuracy of the known points,
 but it also depends upon assumptions made when interpolating.   The  following
 assumptions were made:

      1)   The increase or decrease of discharge rate  is linear  between close,
          consecutive measurements.

      2)   The increase or decrease of suspended load  is linear  between close,
          consecutive measurements.

      3)   Streams of  similar size  and length  contained  within  similar
          geological  and  topographical  boundaries  will  behave approximately
          the same.

      4)   For any stream  flowing  through unchanging topography  and soil
          conditions  the  time average eroded  load  per unit  length is
          constant  for  short sections.

      In  estimating the total load  carried  during  the spring runoff,  there are
 two  sources  of  error  for which no  correction can  be made at this time.  The
 first  deals  with the  transport of winter fines at the  onset of  the runoff.
 This  error  is limited primarily to  the  smaller, shallower  streams and rivers
 which  freeze  solidly or nearly solidly  during  the winter.  As  the streams
 begin  to  thaw and flow begins over  the  frozen  streams  (often on the  ice
 surface  or between the frozen top  surface  and  the frozen bottom), many fine
 particles are carried to points where  ice  restrictions cause the flow to
 decrease  to  the  point that suspended particles are deposited on the  icy
 bottom.  After  the runoff gains momentum,the bottom ice is freed and floated
 to  the surface.  Ice chunks carrying sediment up  to 5  cm thick are floated
 downstream.  Many of these chunks are broken up or overturned, thus depositing
 the sediment into the stream,  but many other chunks  are carried into the
 lake.  No estimate of the extent of  this contribution  can be made at this
 time.  The second source of error deals with transport near the river mouths.
Wash-load transport is primarily a function of turbulence.  As the streams
near the lake (most stream mouths are drowned), the channels widen and the
                                     18

-------
   i!Jij
I
 Heavy
red clay
                    Scale hi.OOO 000
                 N
             SUPERIOR

                    ^'WV
Silt ioam with
well -drained
subsoil
   Rocky
 loamy acid

               » - —	x* ./• • fr**
                                          BAYFIELD
                                          COUNTY
DOUGLAS
 COUNTY
                                                               ASHLAND
                                                               COUNTY
                                                                               IRON
                                                                              COUNTY
                             Figure 7.  Soil map of northern Wisconsin.

-------
bottoms become smoother, both resulting in a decrease in turbulence and load
transport.  Also on occasions high lake level and strong wave action cause
the flow to decrease tremendously, sometimes to the point that the flow runs
upstream for as much as 1 km.  The result of this is heavy deposits of soft,
mucky fines above the stream mouths.   Periodically high flow rates in the
spring or during severe rainfalls scour and cleanse these areas.   However,
at peak runoff, the conditions are lessened by the high flow conditions.
No attempt to estimate the amount of  this transient transport of  material can
be made at this time.  It is reasonable to assume that the sources of error
due to these factors are reasonably low, say 10%.

     All the data collected during this study together with interpolated and
estimated values appear  in Appendix  A.

RESULTS

     Snow measurements for the study  correlate well with data obtained from
the National Weather Service Office,  International Airport, Duluth, Minnesota.
Comparative plots of the snow remaining at Duluth Airport and the measured
rate of runoff vs. date for the period April 7 through April 25 are shown in
Figure 8.  The data indicate  approximately 25 cm more snow near  Pearson
Creek than Weather Bureau measurements near Duluth.  Past experience makes
this difference reasonable, although it should be noted that measurements
along the South Shore varied considerably, depending upon where the measure-
ments were taken.  For instance, on April 14, snow measurements varied from
26 cm near Hwy 13, Dutchman Creek to  34 cm near the mouth of Pearson Creek.
Using the snow moisture content of the samples and basin sizes, the calcula-
tions of water content for the basins were compared with measured total
discharge values for both these river basins.  During the period  April 7
through April 26, the calculated discharge at Hwy 13, Dutchman Creek was 2.5
x 1C)6 m^.  At Hwy 13, the calculated discharge for Pearson Creek during the
period April 14 through April 27, was 1.6 x 106 m3.  The measured discharge
during this period totaled 1.5 x 10^ m^.  These differences are small and
suggest that the snow measurements made at a few points were representative
of the entire basin areas.  It should be pointed out, however, that snow
distribution is not uniform over large areas, and several widely spaced
stations must be monitored when determining the spring runoff.  Thus, weather
data at the Duluth Airport is not sufficient for South Shore runoff studies.

     Plots of total daily suspended load vs. date were constructed for all  of
the major streams and are shown in Figure 9.  It is  interesting to note that
although  the plots for all of the streams yield similar curves, a  significant
difference exists between the large rivers  (Nemadji  and Amnicon)  and all of
the smaller streams.  In the smaller streams, the highest  sediment load trans-
ported occurs at the onset of the runoff, although peak discharge  does not
occur until later.  For the  large rivers peak daily  load occurs during  the
days of peak discharge.  This difference can most  likely be attributed  to
the greater buildup  of winter fines in  the shallow  streams where  the  flow
is reduced by  ice obstruction.  The large rivers are able  to uniformly  trans-
port the  fines  throughout the winter.   The  peaks of  the  suspended  load  curves
correlate well with  the rate of runoff  (Figure 8).
                                     20

-------

o
c

M—
o

"c ^
JH • —
a _.
^> **^
"5 o
LU
M—
0) O
Icr
3.75
2.50
!.25

Figure 8.   Daily snow and runoff data during snow-runoff period.

-------
    15
ro
 O
 o
 Q
    10
 c
 o
o
"k_

1i>
                     SUSPENDED  LOAD  vs. TIME
                             i  i
                      Nemadji River
                                           f\



                                           I

                                           I

                                           I

                                           t
                         \
       i  i  »   i  i   »»»  i  i  i  »   i  i  i  i  i   i  i  i   i  i
           15
             Figure 9.
  20         25          30 I         5

     April                       May
Daily suspended load for analyzed streams during the
entire runoff period.

-------
  SUSPENDED LOAD vs.  TIME
                      DUTCHMAN CREEK
                      BARDON CREEK
IS  20  21 22 23 24 25 26 27 28 29 30
   APRIL
                                                         234
                                                          MAY
  80

  70

  60

£ 50
a
£40
o
o30
or.
£20
2
  10
    SUSPENED LOAD vs. TIME
                      TEN CREEK

                      PEARSON  CREEK
   20
                         APRIL

                         Figure 9.   (continued)
                                30  I
                                        MAY
              23

-------
  280



  240


>-200
<


£160
K 120
o

E  80
   40
                        SUSPENDED  LOAD vs. TIME
                                           Middle River

                                           Amnicon River
         15
                     20              25

                             APRIL
               30
                                                                 MAY
I
UJ
z
60



50


40



30


20


 10

/\
\
\
/ \
1 \
1 \
• 1 \
1 \
1 \
1 \
1 \
1 \
1 \
. / V
1 >
' \+
1
t






1 1 1 1
SUSPENDED LOAD vs. TIME


.

POPLAR RIVER 	

.



X~x
/ ^N
/ ^X
/
\ '•>
\ / ^
\ / \
\_/ ^ / »
\ / "^^
\ / ^-^^
A / N
V_
""—-- — _ / ""———_
' i i i i i i i i r~~-f" i § i i t i r~—
         15
                      20
25
30  I
                                APRIL
                                                              MAY
                            Figure 9.  (continued)
                                     24

-------
     Because of the uniform nature of the clay deposits responsible for much
of the eroded load, it was hoped at the onset of this study that the suspended
solids and turbidity data could be correlated in order to obtain a reasonably
accurate estimate of suspended solids content of a sample by turbidity meas-
urements alone.  At low concentrations this is quite feasible and has been
used extensively in lake turbidity analysis using remote sensing data, Figure
10.  However, for high concentrations, the changes in suspended solids are
gross for small turbidity changes.  This precludes linear interpolation
beyond turbidity levels higher than 100.   This nonlinearity in turbidity
measured for high concentrations is due to turbidity generally being a meas-
urement of fines,  rather than the large particles present in high concentra-
tion loads.   An exponential curve for turbidity above 100 is also shown in
Figure 10.

     The total suspended load transported for all rivers and the individual
stream loads for the runoff period appears in Table 1.   A breakdown of the
                                   TABLE 1

                     LOAD TRANSPORTED DURING APRIL  1975


                      River                  Load  (metric  tons)

                Nemadji                           106,892

                Dutchman                              453

                Morrison                              34^

                Amnicon                             3 ^57

                Ten                                   431

               Wagner                                296

               Hanson                                14g

               Middle                              2,008

               Poplar                              1>637

               Bardon                                544

               Pearson                               842

               Unnamed #1 - #15                      553

               Total                             117,307
                                    25

-------
CORRELATION of TJRBSD1TY and
        ED  SOLiDS
(0
-------
 runoff load  for each  stream as a  function of stream  length and drainage area
 is  shown  in  Appendix  A.   It is interesting  to note that the total  load trans-
 ported during the snow runoff and the April rain event by the Nemadji is
 roughly ten  times the load of all the other streams  combined, yet  the total
 discharge of the Nemadji  is less  than twice the total discharge of the other
 streams.  This difference suggests that a nonlinear  relationship exists be-
 tween discharge and suspended load.  Assuming an equivalence for the streams
 because of similar geologic and geographic  boundaries, the total discharge
 vs. total load are shown  in Figure 11.  Although there are not enough large
 streams in the study  area for conclusive results, the logarithmic  plot of
 discharge vs. suspended load for  the Nemadji, Amnicon, and Middle  Rivers
 appearing in Figure 11 suggests that an exponential  behavior exists between
 the load  and the discharge.

     Loads transported per km of  stream length and per square km of basin
 area are  shown in Table 2.  The uniformity  of eroded load per km in Table 2
 suggests  correlation  between the  suspended  load and  stream length, rather
 than the  suspended load vs. basin area, particularly if the figure for the
 stream length comprises the main  stream plus a fixed percentage of the inter-
 mittent stream tributaries.  This suggests  that the  erosion is mainly due to
 immediate bank erosion rather than to  surface runoff area.  This  further
 indicates that an excellent relationship between suspended load transported
 and stream length could be obtained by uniformly mapping each stream into
 main trunk,  main branches, and tributary branches, and then deriving a total
 stream length figure  based upon the sum of  the entire main trunk plus a fixed
 percentage of the intermittent branches.  The difficulty of this system lies
 in the fact  that USGS maps are not consistent enough for uniform detailed
 mapping.  The nature  of each stream should be based  on long-term observation.
 Continual, detailed aerial photography would be sufficient for this purpose.

 RAIN RUNOFF

     To monitor the runoff during the summer, five stations including rain
 gauges were  set up:  one on the Nemadji at the Golf Course Bridge, one each
 at Hwy 2  and Hwy 13 on the Amnicon River, and one each on the Poplar River
 at a county  road 3 km south of Hwy 13 and a county road 4 km north of Hwy 13.

     The necessity for measurement of rainfall rates and distribution becomes
 evident from examination of the runoff data obtained at the rivers in compar-
 ison to the  rainfall recorded at Duluth during the same period.   The Duluth
precipitation measurements are completely inadequate for this purpose,
although  those are generally the only data readily available.

     There are no significant rain runoff events from early July through
October 20, 1975;  on two occasions 1.25 cm  (0.5 inch) rainfall was recorded
at various stations, but the ground was so dry that the rainfall  did not  alter
 the suspended load appreciably.
                                    27

-------
 o
'k_

 
-------
VO
                                                      TABLE  2

                            SNOW RUNOFF  PER LINEAR KM OF  STREAM LENGTH AND  PER SQUARE KM
                                    OF BASIN AREA FOR THE PERIOD APRIL 10-27
RIVER

DUTCHMAN
(south of Hwy 13)
TEN
(south of Hwy 13)
POPLAR
(south of Hwy 13)
+ POPLAR
(between stations)
* BARTON
(south of Hwy 13)
* PEARSON
(south of Hwy 13)
* PEARSON
(between stations)
% INTERMITTENT STREAMS
INCLUDED IN STREAM LENGTH

50%
50%
50%
100%
50%
50%
50%
TOTAL RUNOFF LOAD
PER LINEAR KM
METRIC TONS
9.1
10.1
8.0
17.4
7.0
12.7
13.9
TOTAL RUNOFF LOAD
PER SQUARE KM AREA
METRIC
16.3
19.3
15.2
29.6
9.4
18.8
26.8
TONS







           + 100%  intermittents included  because  the  stretch  between  the  station does not include the
             source area, which always  includes  a  large  number of  intermittents.
           * USGS  map indicates entire stream to  be intermittent,  which is  not  correct in comparison with
             similar streams.

-------
      Thus the general monitoring of summer rain events was unsuccessful
 However  a rainfall event in the spring at the end of April was well estab-
 lished for the entire area, as was a late June and early July rainfall event
 for the Nemadji, Figure 12, which is the most important runoff source.  The
 ratio of suspended load in the Nemadji to the total load of the other Douglas
 County streams observed during the snow runoff was preserved for the rain
 runoff in the spring.  This was largely because the soil conditions at that
 time were the same as those during the spring runoff.  The summer vegetation
 may alter this ratio.
 -i, «-  u     j     rUn°ff characteristics of the Nemadji River, and assuming
 that the load ratio for the Nemadji and the Douglas County streams was the
 same as in the spring runoff, the output due to rain runoff for the past few
 years is estimated in Table 3 and total yearly runoff due to all streams is
                                    TABLE 3

               SUMMER RAIN RUNOFF FOR NEMADJI RIVER (1970-1974)

                                 (metric tons)

May
June
July
Aug.
Sept.
Oct.
Nov. 1-15
Total
1970
35,545
19,396
60,109
4,991
30,245
128,237
44,500
323,023
1971
58,959
34,305
75,765
71,991
51,618
87,173
23,182
402,993
1972
23,165
90,914
135,909
196,636
88,127
8,895
21,818
565,464
1973
89,455
37,182
25,682
147,664
93,464
74,223
809
468,479
1974
20,050
46,191
81,364
64,727
5,359
5,982
31,300
254,973
estimated in Table 4.  Not included in the total in Table 4 is the contribu-
tion from Bluff and Bear Creeks, which flow into Allouez Bay, and contribu-
tions from streams such as the Pokegama River, which flows into the
St. Louis River.

     An important question is how much of the Nemadji's load contributes to
Lake Superior turbidity.  This is not known and has not been included in the
                                     30

-------
tf> s
i— W-
o a)
•*= 9 3.75
^\ 1 "^L~
T_ VJ "^
\ ^
== 2 2-50
V— J5
cr
Discharge (m3/sec)
§ § 1
NEMADJ1 R\\
1
1
__jt_
t

/ER SUMMER RAIN EVENT
/• n<«:« _____
» ^
| ' Dischorgc -----
/T\ Load 	
; i \ :
/ \ \ :
_ ^H-r'^r^JlIl^^ '
23 25 27 29 i 357
JUNE JULY
o
•o
o
o
J
T3
O
5 5
c:
o
Q.
tn
Figure 12.   Daily rainfall and Nemadji River discharge and suspended-load
            data for a summer rain event.

-------
                                     TABLE 4

                        YEARLY INPUT OF SUSPENDED LOAD

                                  (metric tons)

Average
yearly rain
runoff
Spring
runoff
Total
Nemadj i
402,986
106,892
509,878
Doug. Co.
40,298
10,415
50,713
Bay. Co.
23,210*
6,000*
29,210
Total
466,494
123,307
589,801
          * Based upon University of Wisconsin, Superior data for spring
            runoff for Bayfield County Streams (EPA R-005169-01) .


 measurements.   Such measurements  would have to account for the seiche actions
 affecting the  transports in Duluth-Superior Harbor and would require continu^
 ous monitoring of the turbidity in  the Nemadj i and along the Superior entry
 A rough estimate can, however,  be obtained from two approaches.
 «h«    Kt.uTT      °f  samPles  from red  clay banks  off  rivers  and the lake-
 shore  by  the  University of  Wisconsin, Superior  (Bahnick et  al.,  1972)  show
 roughly that  30%  of the material  is less  than 4  y  in size.  It is anticipated
 that at least 25% of the Nemadji's load would wash out into the  lake  f^om-
 ?he Nemaali   ? M   s^H*" d±Stribution of the Particles  in the lake and in
 the Nemadj i   Table 5,  indicates that roughly 60% of the load  finds its way
 into the  lake.  This estimate  of  the fraction of load  from  the Nemadj i
 entering  the  lake directly  is  in  agreement  with  a  rough order approximation
 obtained  from the satellite data  for June 14, 1975, which show   14,000 tons
 of material in the lake  due to runoff while the  output  for  that  event  in
 the Nemadji totaled  about 27,000  tons.  Accounting  for  the  output  of  the
 Douglas County streams,  this indicates  roughly 50%  output   into  the  lake from
 the Nemadji River.   Based on the  assumption that 90% of  the particles  3 u
 and less  flow out  into the lake,  the estimate obtained  from the  particle-size
 data indicates that  nearly 40% of  the coarser particles  also  flow  out  into
 the lake.

CONCLUSIONS

     The 1975 spring runoff for the rivers  is estimated  to be 117,000 metric
tons for Douglas County rivers  and the Nemadji River, with the exclusion of
Bluff and  Bear Creeks flowing into Allouez Bay.   The Nemadji is responsible
                                     32

-------
                                    TABLE 5

                 PERCENT ABUNDANCE OF PARTICLE SIZES BY NUMBER


                                        Particle Size
     Netnad j i

     Lake Near Shore

     Lake Intermediate

     Lake Deepwater

     Dutchman  Creek
                         0-2 y
2-3 y    3-4 y    4-6 y
                                                              6-20 y
12
21
30
30
19
10
15
17
16
13
24
24
23
23'
22
47
35
27
29
41
7
7
3
2
5
for 90% of the total runoff load in the spring.  An estimated 75% of the
contribution to lake turbidity due to runoff comes from the Nemadji River
Basin.  This contrasts with the total discharge of the Nemadji,which is ap-
proximately twice the runoff for the Douglas County streams.  These numbers
may be accounted for by the fact that the erosion rate appears to increase
exponentially with the discharge irate.

     The eroded load correlates better with river length as opposed to drain-
age area, indicating that immediate bank erosion is primarily responsible for
the load rather than the area of the drainage basin.  The erosion per km of
stream length is roughly 10 tons for the spring runoff, and an estimated 0.7
tons/km per cm of precipitation runoff.
                                    33

-------
       5.  SOURCES AND TRANSPORTS OF TURBIDITY IN WESTERN LAKE SUPERIOR

 ESTIMATION OF SEDIMENT RESUSPENSION AND TOTAL LOADING

      A measure of shore erosion and sediment resuspension can be obtained
 from Landsat remote sensing and ground truth data.   Reliable correlations
 between satellite imagery data and measurements of  turbidity and suspended
 solids concentrations have made possible reasonable estimates of the total
 suspended load in the lake and its distribution on  satellite-overpass days.
 This method of estimating lake loading has  several  advantages.   Since only
 a representative number of stations must be sampled on  satellite-overpass
 days,  manpower and field-sampling  requirements  are  limited.   Furthermore,
 this method measures  primarily the fine particulates which remain in suspen-
 sion for prolonged periods and hence pose an environmental problem.   This
 method also provides  a better  understanding of  the  transports of the loaded
 material and subsequent resuspension processes  as a function of  meteorolog-
 ical and lake conditions.

 Analysis of Landsat Imagery

      To  be  useful,  remote sensing  data  must be  correlated  to  ground  truth
 data.  Numerous  samples were taken on Landsat-overpass  days.   Satellite
 imagery  was analyzed  in the vicinity  of these sampling  stations.  A  reliable
 correlation was  established between surface turbidity and  relative Band 5
 intensity,  Figure  13.   Relative-band  intensity  is defined  as  the difference
 between  the signal at  a given  site and  the  minimum  signal  observed in  clean
 sections of  the  lake,   turbidity is  in  turn correlated with  suspended-solids
 concentrations,  Figure  10.  Since  suspended  red clay particulates from shore
 erosion,  sediment  resuspension, and river runoff are fairly uniform  and are
 similar  in  nature, the  correlations shown in Figure 10 and Figure 13 are
 valid  for the  entire load  of suspended  red  clay.  These correlations make
 possible the delineation of various levels  of suspended-solids concentrations
 directly from  Landsat  imagery, Figure 14.

     In  analysis of the Landsat data, 70 mm positive transparencies were
 viewed on an optical density slicer.  Band  5 intensities were determined by
 comparison with  the calibration step wedge provided on each transparency.
 The study area of the lake bordered by Douglas County, Wisconsin, and St.
 Louis, Minnesota, shores was subdivided into 77, three-kilometer-square grid
 spaces,  Figure 15.  For each satellite image several readings were taken in
 each grid square and averaged.   The data was then processed on a computer.
 The grid allowed for a comparison of the turbidities at various points to  the
 insitu measurements at those points.  On a  few days Landsat  computer compat-
 ible tapes were available.  The Landsat tape data is more accurate and
 allowed for a check of the optical data reduction methods.  The tape data  was
used in effluent tracing, to distinguish red clay from other  particles in  low
                                     34

-------
ID

I-
JQ
 y_
 Z3

h-
                   M.S.S. BAND 5 TURBIDITY CALIBRATION

                           TB=2.0*(REL B5)-4.0

                                   = 0.98
                         Relative  Band  5  Intensity
                       Figure 13. Turbidity calibration of Landsat image.

-------
u>
             (^
       vMSfe
          /
*/i
                W&rS&ZW
                WPcP,
                > n o o
                  •&S~'°1
              Pl o o'-r o o~'-
             //',,"f,''-c—°-
             V ''. ' -" ,'^^
                      './•>K?>^K>
            'oVoVo"o°^>
            P^^v^,r
            " " PrP-f «^0 O V0°0 ',
::;:i:i:;:;:|:;:;:;::::::::::::::i::::::X;:::v:jx;:;::::::::::
::^:-:::^:S::S:::::::::::::::::^^:^:^:S:::::::::
:::-x':x'::o:X:::::::Xx':x::::X^:x:x'£x'£x'^
•:^:^x^:::::::::::::::::X:X::-:^:::^x::::::::::::::
::::::S^:::::::::::::X:::::X::::;:^:^:^X::::-::X::
E SUPERIOR' 11111
•••.••'•'X-X'X'X-x x-X'Xvx-x-x-v'X-x-x
x:::x::::X:X:X:::::X::::::::::::::::::::::::::::::Xx:







:x
SU5
:£S:Wx:

COGCOOCK.
O O D O Q O O O
oc-oooooo
oooooooo
lOOCOOOOO
•.rxin^n^'^iO




3. SOLIDS mg/L(LAKE)
'"4 KSS 12-14

4-8 Vi*jS 14 +

8-12


       ^

 ^SUPERIOR
              ' o o oo co or o A o-o  o r. o o
 ,0 O O •
                                                 .CO'O o o OJ
                                             ^W::"»°:
                                             or. o no o
                JMI
                                                              Uo^"o_o-o
                                                                   ^o.,o c.>>
•TOfifBiftefr
 "•Wm&MoV
                                   >
                                  v> <•.• ,-.<:• o"o>
                                  ,te?M^^
                                  V^0^°,",o"o°p2
    •^r^-V^V^V^VV,:
    ;.,;--.,;',(-'^,%o0o.;,.,,.;.y,
                   ^
                                                                 1.5  3.0
                                                                 km.
                      HARBOR a LAKE IMAGE DENSITY LEVEL Vs. LAKE TURBIDITY
                            ERTS E-1434-5  SEP 30 1973
          Figure 14.
        Representative suspended-solids distribution produced fro, Landsat data.

-------
                                                 Knife
SUPERIOR
                                      Amnicon
                        Dutchmans) Creek
             Figure  15.  Lake grid spaces and regions used in Landsat-image analysis.

-------
                                 descrlbed  above>  the  distribution of  surface
                              each  image  analyzed.   To derive  total suspended
             11? T  turbidit,ies' vertical  P^files  of suspended  solids'concen-
                     est^ted.   The  study  area was  divided  into  seven regions
 Fieurs   M
 !n?forn, ^  "T^"^  Bleated  turbidity profiles  in  these  regions were
 uniform throughout each  region and remained more or less constant  throughout
 sotidTi,;   TT Kati°S °f SUrfaCe SUSPended solid« to average suspende
 solzds in each of these  regions are shown in Table 6.  Using the values in
                                    TABLE 6

                RATIO OF AVERAGE TURBIDITY TO SURFACE TURBIDITY

                           FOR WESTERN LAKE SUPERIOR
Region
1
2
3
4
5
6
7
Ratio
1.0
1.1
1.2
1.1
1.5
1.2
1.2
this table, the surface turbidities derived from satellite data, and  the
calibration between turbidity and suspended solids, the total suspended
loads shown in Table 7 and their distributions were derived.

Estimation of Sediment Resuspension

     To determine the relative contribution to lake turbidity due to erosion
sediment resuspension, and runoff, the data was examined in light of weather '
records of winds and precipitation.   The observed plumes were categorized to
distinguish between events which show predominant input from one source and
events displaying a mixture of contributions from various turbidity sources
For instance,  spring events yield a measure of the magnitude of the sediment
                                     38

-------
                   TABLE 7
TOTAL SUSPENDED LOAD IN WESTERN LAKE SUPERIOR

                (metric tons)

           12AUG72        88,000
           060CT72        58,000
           29NOV72        50,000
           16DEC72        32,000
           28MAY73       115,000
           03JUL73        61,000
           12SEP73        28,000
           30SEP73        46,000
           180CT73        32,000
           190CT73        29,000
           06NOV73        36,000
           23NOV73        34,000
           22MAY74        59,000
           27JUN74        77,000
           28JUN74        53,000
           15JUL74        94,000
           25SEP74        38,000
           26SEP74        36,000
           18NOV74        27,000
           15MAR75        31,000
           03APR75        198,000
           11APR75        412,000
           12APR75        338,000
           09MAY75        94,000
           18MAY75        58,000
           27MAY75        58,000
           14JUN75        61,000
            23JUN75        64,000
            19JUL75        38,000
            06AUG75        51,000
            02SEP75        51,000
            11SEP75        39,000
            080CT75         27,000
            170CT75         54,000
            180CT75         56,000
            04NOV75         34,000
            14NOV75         90,000
            21FEB76        30,000
            05APR76        78,000
            06APR76        61,000
            03MAY76        56,000
            11MAY76        36,000
            12MAY76        41,000
            20MAY76        38,000
                       39

-------
 resuspension term.   The fact  that  the  turbidity  during  spring  events  is
 attributable to resuspension  only  is justified from the observation of the
 lakeshores and  the  simultaneous monitoring  of river runoff.  For  example,
 examination of  the  photograph in Figure  16,which was taken on  April 12,  1975,
 shows  that even in  the absence of  the  lake  ice cover, a substantial ice  ledge
 1-3  meters thick  at  the  shore and 30 - 60 meters  wide protected the south
 shore  from direct removal  of  the material from the  clay bank.  This ice  shelf
 left resuspension as the only major source  of red clay  turbidity  until the
 spring runoff began on April  13.   Consecutive images on March  15  and 16, 1975,
 showed an  increase  of  about 2 x 103 tons of suspended load in  resuspension
 areas.   Some of the turbidity in the March  16 image was attributable to
 erosion of the  North Shore, which  had  no protective ice shelf, but the
 increase in the resuspension  zones was probably  caused  by a  5 m/sec north-
 easterly wind on March 16.  Images from  April 3  and April 11,  1975, yielded
 observed suspended  loads of 149 x  103  and 184 x  103 metric tons, respectively.
 Weather records show strong northeasterly winds  in  excess of 5 m/sec prevailed
 between these two satellite overpasses which resuspended on the order of 1.2
 x  104  tons of material per storm-day.  Such turbidity events occur on the
 average two to  three times for the ice-free conditions  on the  lake and gener-
 ally leave the  early spring and late fall average turbidity much  higher  than
 the average turbidity  during  the months  of  June  - November, Figure 17.

     The total  suspended load observed during the study period, including the
 resuspension term in the months when the shores were ice-covered,  was 9.18 x
 105 metric tons  as  compared to the load  of  6.1 x 10$ metric tons'for the May -
 November season.  Thus, at least 35% of  the year-round   lake turbidity was due
 to the  sediment  resuspension.  Since sediment resuspension is  generated by
 steady  northeast winds  in excess of 5 m/sec, it  contributes up to 30%
 of the  total  turbidity  for severe  storms  in the summer months  as  well.  Based
 upon data  from  pure  resuspension events,  if winds in excess of 5 m/sec which
 prevail  over  24  hours  resuspend about  10^ metric tons of fines per storm day,
 then the total  contribution due to sediment resuspension for the  entire year
 increases  to  about 45%  of the total observed turbidity  on the  lake.  Thus,
 sediment resuspension  should play  an important part  in  the consideration of
 the lake turbidity and  its effects on  the water quality and in considerations
 of leachate  from the red clay  (Bahnick,  1972).

     A better estimate  of the fraction of the observed  load attributable to
 resuspension  for the summer events was obtained by  examination of the remote
 sensing  data  in  light of weather records.  Only the  images for the May -
 November season on days when precipitation could be neglected were used.
 Resuspension was considered significant only after  easterly winds exceeding
 5 m/sec  and  lasting approximately  2 or more days.  These criteria were met
 in about 35% of  the events.  Since the summer wind  storms are usually mild
 and short  in duration,   the resulting plumes usually are not transported far
 from their sources.   Thus,  the suspended load in a  series of source points
which extended from Amnicon River  to Bardon Creek was considered a good
measure of erosion.   These source points were in the middle of the south
 shore erosion belt along a uniform red  clay bank far away from known resus-
pension areas.  Since the near-shore zone accumulates only a thin veneer of
red clay,which is quickly removed by succeeding storms,  the turbidity in
                                     40

-------
Figure 16.  Spring ice shelf near Amnicon River
            (1-3 meters thick).
                      41

-------
      T	1	1	r
0
                            T	r
                      -— 1973
                      	4 year average
T	r
                                              T	r
                            ^X^-.JU^'
                                                     A -
1 - 1
  JAN  FEB MAR APR MAY JUN  JUL AUG  SEP OCT  NOV DEC


            Figure 17. Average lake turbidity near Duluth.

-------
these source points was attributable almost entirely to erosion.

     The fraction of the total turbidity due to resuspension was obtained by
comparing the total suspended load above background for resuspension and pure
erosion events.  Assuming no other sources of turbidity were present, the
total load  (T) was found by summing the contributions of erosion (E) and
resuspension (R).  Letting the subscripts "r" and "e." denote resuspension and
pure erosionjevents the equations become:
                                T  = E  + R
                                 r    r    r
                                T  = E
                                 e    e
                                                                          (1)
The contribution of erosion was assumed proportional to the suspended load (S)
in the erosion source points.  Letting a denote the constant of proportion-
ality, equation (1) becomes:

                                T  = aS  + R
                                 r     r    r
                                                                          (2)
                                T  = aS
                                 e     e

Summing overall  observed events and eliminating a, the fraction (f) of
observed turbidity attributable to resuspension was found:

                              ER      (ET  - ET  • SS )
                       r. =      r      r r   ee   rr
                           ET  + ET       ET  + ET
                           rr   re      rr   ee

Calculations using equation (3) indicated about 16% or about 10  tons of the
observed turbidity for the May - November seasons was due to resuspension.
This value is close to the one obtained from the resuspension rate per storm
day derived from observations of pure resuspension events.

Estimation of Shore Erosion

     The actual eroded load throughout the year in its relation to the ob-
served suspended load depends on the total number of storms, the dispersion-
settling factor, and the cloudiness.  These factors must be taken into con-
sideration in the determination of total shore erosion.  The erosion process
takes place throughout the time when the shores are ice-free.  Generally, as
will be pointed out in the numerical model, northeasterly winds raise the
lake level and produce waves which directly abrade the red clay bank toe.
Many times, large chunks of clay fall into the water, where they are rolled
around in the sand slurry and produce large round balls of red clay which
dissipate rapidly and give rise to high turbidity in the near -shore area.
Many areas of the beach have a veneer of sand covering the red clay which is
subject to erosion by the wave action.  The erosion takes place in a season
roughly 210 days long.

-------
     To determine the total eroded load from the data in Table 7, the
statistical nature of the problem was considered.  If the events occurred
randomly, then the observed turbidity should be directly related to the total
turbidity input through the dispersion and settling rate and the probability
that an event was observed.  This is true provided the overlap of the events
was reasonably small and the time for the decay of the event to the back-
ground turbidity values was short compared to the period between observations.
Both of these conditions are satisfied when images for successive days are
considered as a single observation.  If "t" is the time after a storm, f(t)
the dissipation function, p(t) the probability density that a storm was
observed at time t given that the storm was observed, and So the total initial
suspended load, then the total observed load (T) is given by equation (4):
                              t
                               o
                       T =    / f(t) •  P(t) dt = S  f(t)                  (4)
                            oo                   °
Using equation (4), the average suspended load of observed storms can be
calculated.  Assuming all storms regardless of magnitude are equally likely
to be observed by satellite, the total load of all storms can then be
obtained by multiplying average load by the total number of storms.

     Measurements of turbidity dissipation for easterly and westerly wind
storms, Figure 18, indicate that turbidity for a typical storm remains above
background for approximately 8 days.  The turbidity dissipation for a large
storm in early November 1974 is shown in Figure 19.  Assuming the dissipation
rates for other storms were similar, one obtains the typical dissipation
curve shown in Figure 20 using the relationship between turbidity and
suspended solids, Figure 10.

     The two functions, f(t) and p(t), are not independent.  The storm bvild-
up times for high easterly wind storms have a low probability of being
observed since they are almost always accompanied by cloudiness.  This consid-
eration is important because the settling rate for the first day is very high.
Cloudiness is not the only obstacle in the successful observation of turbid-
ity.   At the peak of the storm, concentrations of suspended solids often
exceeded 100 mg/£, the point at which the Landsat optical signal saturates.
Because of the nature of the sources and currents, very high concentrations
of suspended solids are often confined at the beginning of a storm to a
small area near shore which is susceptible to blockage by partial cloud
cover and the limits of spatial resolution of the satellite.  Since the tail
of the storm is highly observable, the total observed load is a small fraction
of the initial or total load.

     The function p(t) was approximated using weather records and the above
considerations, Figure 20.  For most of the time during the storm decay,
p(t)  assumed nearly a constant value.  Numerical calculations give the
average observed storm intensity p(t) a value .11.  Since most of the
turbidity observed was "fines," on the order of 11% of the total load was
                                     44

-------
ADO — Easterly Storms
' -*- Westerly Storms
— —

i
1
I
- i A
\
x9
i
ft w
o A
^
^^ ^^^b^k^^k
^^^TsS^B^__^^
^^ ^^^^™^^P
1 1 1 1 1 1 1 1 1
"^nn



3 200
h-
^^
>*
"O
lo
w.
Z3
H 100



0

1 1 1 1 1 1 1 1
AMNICON POINT
1
\
-\
\
\
\
\
\
\
\
- \
\
X
X
1 1 t I 1 1 1 1
       0  I  23456789  10

                  Days
Figure 18.  Average turbidity dissipation for
           easterly and westerly wind storms.
   13579

         November  1974
Figure 19.  Actual turbidity dissipation
           for a large easterly storm.

-------
T3
 O
 O
 
jQ


I  10
         12345678

          Storm  Duration (days)





     Figure 21.  Distribution of storm duration.

-------
 attributable  to  the  fines.   Particle-size  data  for  lake  samples  and  river
 samples  indicated  that  fines consist  primarily  of particles  less than  4  u
 (See  Section  4.)

      Wind data for western  Lake  Superior was  analyzed  for days showing steady
 easterly and  northeasterly  winds  in excess of 5 m/sec.   Records  of storms  in
 the past few  years and  insitu measurements have shown  easterly wind  storms
 give  rise to  high turbidity in western Lake Superior.  The distribution  of
 storms for  the ice-free  season from August 1972 - August 1975 is  shown in
 Figure 21.  The category of  storm depended on the duration and sequence  of
 wind; thus  single-day events, double-day events, etc., were derived  from
 successive  days of dominant  high winds.  The  total number of events  was  63.
 Subtracting estimated contributions of resuspension and runoff,  the  total
 observed load attributable  to erosion is 4.37 x 105 metric tons.   Since  20
 images showed^turbidity above background, the average observed eroded load
was 2.19 x  10s metric tons.   The total initial  eroded load (S0) for  an
average storm can be obtained by dividing by  f(t).  This gives S0 a value of
2.0 x 105.   Assuming all storms were equally  likely to be observed,  the  total
eroded load over the three-year year period,  1.26 x 107 metric tons,  was  found
by multiplying by the number of storms.  Thus, about 4.0 x 106 metric tons of
eroded material is loaded into the lake per year.

     Table  8 summarizes the estimated  magnitudes of  the various sources of red
                                   TABLE 8

                          SOURCES OF LAKE TURBIDITY
                                    Average Rate/Season   %  Contribution
                                      (May  - November)       (Rounded  off)
                                         Metric  Tons
  Lakeshore  erosion  for  Douglas
   County, Wisconsin                    4.0  x  10^                75%

  Sediment resuspension  (May -
   November)  in extreme  western
Lake Superior
River runoff contribution to
lake turbidity
Sediment resuspension
conditions (Dec., Jan
for winter
• , Apr.)
1.0
.32
2
x 106
x 106
x 106
20%
5%

                                    47

-------
clay  turbidity in western Lake Superior.  As was established by insitu meas-
urements,  the runoff comes principally from the Nemadji.  Since runoff sources
were  relatively small in magnitude and infrequent, insitu measurements
provided a good estimate of their magnitude.  Shore erosion and resuspension
contribute much greater loads.  In addition, their contribution is continual
and geographically more extensive.  Landsat remote-sensing data were ideally
suited for estimating the contribution of these sources.

TURBIDITY TRANSPORT AND THE DISTRIBUTION OF TURBIDITY SOURCES

      Statistical analysis of turbidity distributions for various wind condi-
tions yields information on turbidity transports and the nature of the
turbidity sources.  Images were examined in light of weather records and
classified according to the kind of storm that gave rise to the observed
turbidity.  The relative turbidity taken directly from the density slices
indicated that for an easterly wind storm, turbidity from red clay banks along
the Wisconsin shore is transported toward Wisconsin Point,  where it is subse-
quently taken out in a return path along the axis of the lake, Figure 22.
This  transport pattern was also established by a numerical model of currents
in Lake Superior.  (See Section 6.)  Similarly the transports for westerly
winds, derived from statistical analysis of the plume shapes, yields trans-
port paths shown in Figure 23 and 24.  Variable winds tend to produce circu-
lation eddies, and periodically trap the degraded water from the St. Louis
River in the areas near Duluth, Figure 25.  The range of transport of red
clay particles can be assessed from the analysis of the Landsat image for
April 11, 1975, where a resuspension plume traveled in high concentrations
for about 135 km between Grand Marais and the Apostle Islands, where it had
a concentration of 2 mg/Jl.

     The distribution of turbidity sources can also be obtained from Landsat
data.  The average turbidity for the ice-free season is shown in Figure 26.
Using this figure one obtains the map of relative erosion rate shown in Figure
27.  Thus the 4.0 x 106 metric tons of material eroded along the lakeshore
is distributed fairly uniformly along the shoreline,  with maximum rate just
west of the Amnicon River and near Middle River to roughly half the rate west
of the Brule River.   For high northeasterly winds, sediment resuspension
takes place.   Examination of the relative distribution of turbidity for such
storms, Figure 22, in comparison to turbidity distribution for other events,
Figures 23 - 26,  indicates that the resuspension area exists north of Dutchman
Creek, which enters the lake at the southern most part of the lakeshore in
Douglas County, Figure 28.  Intermediate depth areas  along the Douglas County
shore are also subject to sediment resuspension after periodic accumulation
of material in those places during the quiescent times.   The determination of
the resuspension areas are, however,  quite difficult  since for high northeast-
erly winds, considerable erosion takes place,  the transports are rapid, and
the probability of observing a storm at the peak moment is low because of the
accompanying cloud cover.   A good verification of the suspected resuspension
areas can be obtained from the early spring events which occur before runoff
and during the time when there is a protective ice shelf along the shore to
inhibit direct lakeshore erosion.   These indicate that the south shore and
Dutchman Creek areas and the waters off Minnesota Point  are the major resus-
pension areas,  in agreement with Figure 28.  The area off Minnesota Point
                                     48

-------
                                 *    ->
                            ""LAKE SUPERIOR
                                   Jr
           •W/S'AV *' *
              ***?***>'..&
  Relative turbidity for Easterly storm
                           LAKE SUPERIOR
                                Km.
 Turbidity transport for
 North Easterly storm
Figure 22.  Average turbidity distribution and transport for
         an easterly storm.
                     49

-------
                                   *LAKE SUPERIOR



                                               * A
                                 \ v v
    Relative  turbidity for N.W. S W. winds
                                 LAKE SUPERIOR
                                        km.
    Turbidity  transport for N.W. SW. winds
Figure 23.   Average turbidity distribution  and transport for a

           westerly storm.




                           50

-------
        N
                                   """LAKE SUPERIOR
                                        km.
   Relative  turbidity  for  Easterly  storm
   followed by  westerly  dispersion
        N
                                 LAKE SUPERIOR
                                        km.
   Removal of easterly  storm  turbidity
   by  N.W.SW. winds
Figure 24.   Average turbidity distribution and transport by
           westerly winds.
                           51

-------
                                 """LAKE SUPERIOR

                                             v
Relative  turbidity for variable winds
                               LAKE  SUPERIOR
Turbidity  dispersion  for  variable  winds
Figure 25.  Average turbidity distribution and transport for
           variable winds.

-------
Ui
OJ
                     N
           ^    <"
v  "     *  *    *  *
    \AKE  SUPERIOR
                \
                                                               km.
                Average turbidity (NTU)

                (ice free season)
                Figure 26.  Average turbidity distribution for the ice-free season.

-------
Relative  erosion rate
for  red  clay bank
                                LAKE SUPERIOR
                                       km.
    Figure 27.   Relative south shore erosion rate.
                              LAKE  SUPERIOR
    Likely  resuspension  areas

    Active areas
                                      km.
    Figure 28.   Resuspension areas in Lake Superior.
                         54

-------
should be particularly subject to periodic scouring since it lies between
two major river turbidity outputs, the St. Louis and the Nemadji rivers
At the same time the area midway off the Minnesota Point is the place where
major near-shore currents meet for the easterly storm conditions.
                                    55

-------
             6 .  A NUMERICAL MODEL OF TRANSPORTS IN LAKE SUPERIOR

DESCRIPTION  OF  THE MODEL

     Transport  patterns of suspended material in Lake Superior have been
rather  controversial.  Drift-bottle studies  (Ruschmeyer et al., 1961) in-
dicate  an average overall counterclockwise summer circulation.  Studies of
currents and turbidity transport through insitu measurements and remote
sensing studies in the extreme western arm of Lake Superior show that the
currents have complex patterns near Duluth (Sydor, 1973).  These localized
currents give rise to an eddy circulation off Minnesota and Wisconsin
Points which transports red clay turbidity to the area of the Duluth water
intake and accounts for the general turbidity of the water near Duluth.
Numerical modeling provided a means of studying lake circulation.  However,
to investigate  transports in western Lake Superior— that is  rnnohlv f*
nrp                       time-dependent mathematical model used for the
present study is based on a tidal simulation model developed by Leendertse
(1967) for the Rand Corporation.  The model has several features which make
it an attractive choice for predicting water levels:

1)  The nonlinear convective-inertia terms are retained in the equations of
    motion resulting in higher accuracy in regions where the land-water
    boundaries are complicated.

2)  The space-staggered grid has the advantage that there are centrally
    located spatial derivatives available for the linear terms of the partial
    differential equations.

3)  The multi-operational method used in the solution of the partial differ-
    ential equations is accurate to  2nd order in time, yet requires only two
    successive arrays to be  stored for the water-level field and each of the
    velocity fields.  Furthermore, the method is unconditionally stable,
    with no upper limit on the number of time steps.  Numerical experiments
    conducted by the Rand Corporation, as well as tests in this study,
    showed that  moderate variations  in the time step size had a negligible
    effect on the predicted  water levels, which are needed later by the
    current profile model.  In order to model wind-induced water levels and
    transports for Lake Superior, the effects of wind stress were incorpo-
    rated  into the tidal model.
                                    56

-------
Mathematical Formulation
     Consider a cartesian coordinate  system  in which  z  is  increasing  upward.
The equations of motion for a viscous fluid  are given by
         9u  ,  3u  .
         T— + ur— +
         9t    3x
            3u ,   3u   _  .  1 3p
           fr	h WT— - f V +	T-*-
            3y    3z        p 3x
                                     .
                                     + v
                                         32u
                                                          (5)
3v
3v
3v
                           3v
                                                    ,
                                                   +
32v,
32
                                                    (6)
         3w  .  3w  ,  3w  ,  3w   ,.    . 1 3p
         — + u— + v— + w— + f v  +	
         3t   ^x   ^y   ^z     e   p 3z
                          r32w ,  32w, ,    32w ,      ,_.
                           VT + T~2"5 + v V~2" + 8   (7)
                           3x    3y      vdz
where u, v and w are the x, y and z components of the water velocity, v  and
vv are the horizontal and vertical eddy viscosity coefficients, p is the fluid
density and p is the pressure.  The coriolis parameter f is equal to 2o) sin ,
where to is the angular velocity of the earth's rotation  and $ is the lati-
tude.  The eastward component of the water velocity is denoted by v

     In studies of long-period water waves, the vertical accelerations and
velocities are generally assumed small enough ^o be ignored.  The vertical
component of the coriolis acceleration is also quite small compared to g and
is dropped from eq. (7).  The terms containing horizontal-eddy viscosity are
very small compared to the vertical-eddy viscosity term and are neglected in
this model.  With these assumptions, the equations of motion become:
       3t    3x
             3y
                                              _
                                            p 3x
                                                                          (8)
                                                                          (9)
                      p 3z
      = g
                                                                         (10)
A uniform density is assumed so that pressure becomes a linear function of
depth.  Consequently, the model is suitable for predicting flows only when
the water is well mixed.  Since,in a typical year, stratification does not
develop until August and disappears by October, Lake Superior can be consid-
ered well mixed during the times of significant loading of red clay, i.e.,
the spring runoff and late fall storm surges.  With  the  assumption of
                                      57

-------
 hydrostatic  pressure,  the  horizontal  derivatives of  fluid pressure become
 functions  of the water slope:


                               p(z) = pg[h-z]
                        IlE =
                        p  3x
                  _3h
                  3x
                            =
                       p 3y   8 3y
where h  is  the  displacement of  the water  surface from  the undisturbed level.
Because  they  generate  only weak currents,  the atmospheric pressure gradients
have been ignored.

     Vertically averaged velocities are introduced
                    U =
                         (H+h)
                  h
                / udz
                 -H
                             (H+h)
      h
    / vdz
     -H
where H is the depth of the undisturbed water.  Vertically integrating the
simplified equations of motion from -H to h, and dividing through by H + h
gives :
                                  (H+h) 3z
                                 _
                             z=h   (H+h) 3z
                                          z= -H
         +
      3t
            ™
3y
         U
              3h
                   (H+h) 3z
                                            z=h
(H+h)
_3V
3z
z= -H
The integrals of the non-linear terms are only approximate, but since these
are small, only slight inaccuracies result.  The integration causes terms
containing the surface and bottom shear stress to appear on the right-hand
side, corresponding to the wind stress and bottom friction respectively.
The components of the bottom friction are taken proportional to the square
of the vertically averaged water velocity
     3U
         z= -H
= gc 2u(u2 +
                                        3V
                                            z= -H
                                    -22    2 Is
                                  gC  V(U  + V K
where C is the de Chezy coefficient.  In general, the value of C depends upon
the roughness of the bottom, the bottom material,and the depth, and should be
determined by experiment.  Tsai and Chang (1974) found that the water  levels
predicted by the Rand model were relatively insensitive to the bottom frictior
used.  Therefore, the values of C were calculated for Lake Superior according
                                      58

-------
 to  the  following  formula  which is  based  on table  of  de Chezy values versus
 depth for  lakes and  rivers  (Hutchinson,  1957);

                               .   0,   . .168745
                               C =  25  x h


 The Chezy  coefficients  obtained were  somewhat lower   but  still  comparable  to
 those found  experimentally  by  Leendertse in modeling an estuary of  the
 Rhine River.

      In its  original form,  the Rand model  contained  no wind-forcing function,
 relying only upon water levels as  a function of time along an open  water
 boundary as  a driving force.   In order to  model wind-induced transports and
 water levels in Lake Superior,  the wind-forcing terms (Wilson,  1960)
       v   JHJ
      (H+h) 9z
z=h
      k „ ,„ 2^I7 2,h     v   9V
      - Wx(Wx +Wy )
were added to the vertically  integrated equations of motion, where Wx  and
Wy are the x and y components of the wind velocity and k  is  the wind-stress
coefficient.  Following Tsai  and Chang  (1974), who did an extensive  study of
wind-stress coefficients, k was taken as


           k = 1.25 x 10~6    for W < W
                                            W-W
           k = 1.25 x 10~6+ 1.75 x 10"6sin  (^— |- f)  for V^ 1 W W2

           k = 3.0 x 10~6  for W2 < W


where Wj_ = 5.1 m/sec, V^ = 15.4 m/sec and W is the wind speed.  In their final
form, the vertically integrated equations for the horizontal motion  are:
       f

For an incompressible fluid, the continuity equation is written

                              — + J-Y. j. ^w
                              8x   9y   3z ~
                                      59

-------
After vertically integrating, and applying the boundary conditions:


           ,, ,    dh   dh .   dh .   dh             ,      ..
          w(h)  = -r— =	1- u-	1- v—         at the surface, and
                 dt   dt    dx     y

                    dH    dH
          w(-H) = -UT	v—- =0             at the bottom,
                    3x    dy

the continuity equation becomes


                    |£ + |^ [U(H+h>] + |^ [V(H+h)]  = 0                   (13)

Numerical Procedure

     The vertically integrated equations of motion and continuity must be
replaced by finite difference approximations which can be solved at points
on a spatial grid such that

U(jAx, kAy, nAt) = U(x,y,t)     where     Ax = Ay = As

                                           j = 0, ± h, ± 1, ± 3/2, ...

                                           k = 0, ± h, ± 1, ± 3/2, ...

                                           n = 0, h, 1, 3/2, ...

A space-staggered grid is used in which the x and y components of velocity,
water levels, and depths are located at separate points.
                      k+l  +  -  +  -+       +  water level   (h)

                      k + %     o  |   o   |       o  Depth  (H)

                          k  +  -  +  -  +       -  x-comp. of velocity  (U)

                      k - ^   |  o  j   o   |       |   y-comp. of velocity  (V)
x
                             j-l   j     j+1
                                      60

-------
The advantage of such a grid is that centrally located spatial derivatives
across only one grid space are available for the linear terms of the partial
differential equations.  The land-water boundaries are fitted through the
depth locations, and the perpendicular component of velocity at the boundary
is fixed at zero.

     In order to see how the numerical integration might be performed,
h(t + At) is expanded about h(t) in a Taylor's series:

                  h(t + 4t), h(t) + At


To first order in At, the new h is found in a single time step where

^(n+1) „ ^(n) + At _dh    ig found by evaluating the continuity equation using
finite differences on the grid.  This is the simplest scheme, however it has
the disadvantage that errors tend to build rapidly in time because of the
low order of accuracy.

     In the Rand Model, a double time step operation is used to perform the
numerical integration.  Each step has two time levels associated with it,
the first from time n to n + h, and the second from time n + h to n + 1.  The
spatial derivative and coriolis force terms alternate between the old and new
time levels in each step such that over the two successive steps of the time
interval, these terms are either central or averaged in time.  By using a
combination of central and time-averaged terms, the resulting integration is
accurate to second order in time while requiring only two successive fields
of information to be stored at any given time.  During the first operation,
the fields of h(n4^), U(n+^) , and V(n-f^) are computed from the fields of
h(n), U(n), and V(n).  Since the solution for h(n+3s) involves U(n+^) while
the solution of U(n+%) requires knowledge of h(n+%), the equations for an
entire row are written in matrix form and solved simultaneously to develop
recursion factors which enable the computation of h(n+33) and U(n4^) to be
carried out implicitly.  Once these fields are known, V(n+^) can be solved
explicitly.  Similarly, in the second operation, the fields of h(n+l),
U(n+l), and V(n+l) are computed from the fields of h(n+^), U(n+%), and
V(n+*5).  In this case, h(n+l) and V(n+l) require each other for solution so
they are solved implicitly using a recursive scheme, and finally U(n+l) is
solved explicitly.  Some of the nonlinear terms had to be approximated at a
lower time level in order to avoid an iterative procedure which would require
excessive computation time.

RESULTS OF TRANSPORT MODEL

Lake-Level Oscillations

     Because of their observed importance in the transport of suspended
materials in western Lake Superior, two main types of wind stresses were
modeled, an easterly and a westerly.  Both wind functions were constructed
from the composite of a large number of weather observations and thus
represent "typical" weather patterns.  For example, the northeast wind becomes
                                      61

-------
 out of the southeast as one moves toward the east end of the lake, as though
 a low-pressure area with its counterclockwise air circulation were PPnf*ro,1
 south of the lake.  Similarly, the westerly wind represents a low-pressure
 center north of the lake.

      The transport model was verified by comparing observed water levels for
 easterly and westerly wind stresses with that obtained from the model.
 Fourier analyses for both wind stresses are shown in Figures 29 and 30.
 Notice that the 7.8-hour mode is dominant in both cases.  A series of high-
 frequency modes appear at the general frequencies reported by Rao and Schwab
 (19/6) and found from experimental data by Mortimer and Fee (1976)  Table 9
 The model also appears to exhibit a 5.6-hour mode,  which was also occasion-'
 ally observed in the lake-oscillation data taken at Duluth.  Similarly the
 model exhibits a 2-hour mode which was observed in the lake data at Duluth.
 This mode plays an important role in the mass transport in Duluth Harbor.
 The transports from the model were also verified by comparing them to the
 transports of turbidity obtained from statistical analysis of the Landsat
 data (described in the section on Turbidity Transport),  and the general
 shape of  the  turbidity plumes observed in Landsat images.

 Predicted Transports in Western Lake Superior

      The  results  of transport calculations  for  northeasterly winds are given
 in  Figure 31.   It  is seen  that  for  northeasterly winds which present  a large
 fetch at  Duluth,  there  is  a  strong  current  along the  North  Shore from Silver
 Bay  to Duluth;  Figures  31A -  31C.   This  generally accounts  for  the transport
 of  large  concentrations  of taconite  fines to  the Duluth  intake  area for  high
 easterly  winds  (Baumgartner,  et al.,  1973).   For northeasterly  winds,  the
 transport along the south  shore is due west.  This  transport  and  the  one
 coming down the North Shore appear  to  turn  and meet at Minnesota  Point and
 then  return out in  the middle of  the  lake along  an axis  running  parallel to
 the North Shore.  This  transport  pattern  is consistent with that  obtained
 from  statistical analysis of  Landsat  images.  Figure  32  shows a  resuspension
 plume after an extended  northeasterly  wind.   Red  clay  from  south  shore
 erosion and from sediment resuspension moved  along the Wisconsin  and  Minnesota
 Points, then abruptly turned  out  along the axis  of the lake.  The  turbidity
 on that date at Minnesota Point Airport was 50 mg/£ and  dropped  to 8  mg/i
 1.5 km farther towards Duluth, where it generally reflects  the near-shore
 turbidities of the water coming in from the North Shore  and from  the  St.  Louis
 River out of the Duluth  entry.

     As the northeast wind slackens, Figure 31E,  the transports in extreme
western Lake Superior begin to subside and scramble.  A  large clockwise  circu-
 lation cell appears to form northwest off the Apostle Islands.  The plume
moving out along the axis of  the lake becomes unstable, and disperses.   Thus,
 for high-turbidity events, the Duluth intake becomes contaminated by red  clay,
Figure 33.  Figure 31F shows  the transport pattern 14 hours after the north-
east wind was completely relaxed.   A counterclockwise cell off Two Harbors
and a clockwise cell off Silver Bay have developed.   The turbidity plume  off
Silver Bay due to dumping of  taconite tailings from iron ore processing  is
shown in Figure 34.   This image strongly suggests such a circulation pattern.
The currents near Duluth at the time of Figure 31E show the characteristic


                                     62

-------
  RELATIVE AMPLITUDE (Normalized)   RELATIVE AMPLITUDE (Normalized)
OQ
U>
O
o
rt
C
*TJ
grn
£,*}
-o
so
5°
ro -^
w'c/)
i-h
O
CO
rt
ro
1-1
3
a.
           p
           ro
p
CD
O
bo
    o
    ro
    oo

-------
                               TABLE 9




COMPARISON OF PERIODS IN HOURS OF GRAVITATIONAL MODES IN LAKE SUPERIOR
Mode
(As Identified
by Rao & Schwab)
1
2
3
4
5
Observed
(Mortimer & Fee)
7.89
4.84
3.80
3.37
3.02
Calculated
(Rao & Schwa
7.86
4.45
3.76
3.17
2.94
Observed at Duluth
(Sydor)
7.88
- -
4.88
3.96
3.58
2.93
1.98
Calculated Model
Values at Duluth
7.53
5.56
4.62
3.73
3.33
2.97
2.13

-------
        TRANSPORTS
        ONF  ftRin  QPAPC
     CORRE
    WIND - i;
      RUN  5C
>NDS  TO  2M/SEC
M/SEC NORTHEAST
   TIME  6 OC

        Silver
      TRANSPORTS
      ONE GRID  SPACE
   CORRESPONDS  TO  4M/SEC    ./"  x
   WIND - I3M/SEC NORTHEAST f'  '   '
     R'jN oC    TIME  8 00     A
                    Silver
             Two
            HarborSjX,
Dulufh


   <
Superior
     r ~< .•   \ •    f i
                      Figure 31B.


Figure 31.  Western  Lake Superior transports for easterly storm.
                        65

-------
      TRANSPORTS
      ONE GRID SPACE

   ^NRgE-S^MN/DslcTN°0^ei?   ^ /  /
    RUN 9C    TIME 28 = 00

                  Silver
Ouiuth
Superii
     TRANSPORTS
     ONE GRID SPACE
  CORRESPONDS TO 20M/SEC   -,
  WIND — I3M/SEC NORTHEAST  ft
   RUN 25    TIM£40^00    A
                 Silver
           Two
               Figure 31.  (continued)
                      66

-------
    TRANSPORTS
    ONE GRID SPACE
 CORRESPONDS TO 20M/SEC
 WIND -I3M/SEC NORTHEAST
   RUN 26
TIME 4600

    Silver
                  /'  /   '  •   y
                 f'  //'  '
                  /' '  >  *,'   '  -.:.
   TRANSPORTS
   ONE GRID SPACE
CORRESPONDS TO 20M/SEC   ^  ,
WIND - I3M/SEC NORTHEAST f /   '
  RUN 28    TIME 6000   />^  _ *  *
               Silver
             Figure 31.  (continued)
                    67

-------
Figure 32.  Landsat image for 11APR75 showing resuspension
            plume caused by northeasterly winds.


  Figure  33.   Skylab image  showing turbidity entrapment
              near  Duluth.
                           68

-------
            Figure 34.   Landsat image for 23NOV73  showing  taconite
                        tailings plume off Silver  Bay.
 eddy  circulation  off Wisconsin Point where .large amounts of sediment settle
 and can be  subsequently resuspended.

      The  transports for westerly winds, Figure. 35, are quite different.  The
 Silver Bay  area shows  transports away from the shore, while the transport of
 the red clay  is confined to  the south shore and is taken out along an easterly
 direction past Bark Point, towards  the Apostle Islands, where a large counter-
 clockwise cell is established, Figure 35B.  As seen in Figure 35A, the
 transports  along  the North Shore bring in water from the center of the lake
 towards Duluth, where  the transports disperse, and show a return path up along
 the North Shore.  This generally accounts for the lack of red clay turbidity
 at the Cloquet intake  for such winds, the generally lower taconite tailing
 concentration at  the Duluth  intake, and the flushing of the turbid water out
 of the Duluth area.  However, as pointed out earlier, variable winds often
 tend  to cause an entrapment  of the  turbid waters and harbor effluents in the
 Duluth area.  The summer transports near Duluth are somewhat complicated by
 the presence of a warm-water cell near Duluth.  This generally tends to shield
 the Duluth  intake in the summer from the influx of tailing fines even for
 easterly winds.  A fine-grid model for extreme western Lake Superior, together
with measurements of water levels necessary for the mathematical description
of an open boundary where the warm-water cell ends in the lake,  together with
application of water quality models to the area will be the subject of  subse-
quent investigation.                            '
                                     69

-------
     TRANSPORTS
     ONE GRID SPACE
  CORRESPONDS TO IOM/SEC
  WIND - IIM/SEC  WESTERLY
    RUN 23   TIME 3000

               Silver
          Two
         Harbors
                         '  x..-
                             .-••N.
           jr.  /'•>'' •••*,*'•*-•••'n-
          *v''. V ''&?&*$>
           -'-'•-  '- '^\tf&<
                            }^
                          >v
                                <#'''
              Figure 35A.
  TRANSPORTS
  ONE GRID SPACE ,
CORRESPONDS TO IOM2/SEC
 IND — IIM/SEC WESTERLY
 '	 "'   TIME 48 00
   RUN 31
              Silver
Ouluth
Superior
 Figure 35.  Western Lake Superior transports for westerly storm.
                  70

-------
 Water Levels Due to Storm Surge

      Water levels from the model for a northeasterly and westerly wind storm
 are shown in Figures 36 - 38.  High water levels are evident at the base of
 Wisconsin Point and Minnesota Point.  The water levels at the shore are
 actually higher than those indicated in Figure 37 which reflects the water
 height at points 3 km off shore.  Thus the surge at Wisconsin Point and
 Minnesota Point should be between 15 and 30 cm.  A water-level distribution
 for a model by Forristall (1970) is shown in Figure 39.  It demonstrates
 more vividly the near-shore water pileup, although this method of calcula-
 tion was discarded because of other shortcomings.   The Leendertse model
 generally predicts a pileup of water in the corners of the grid which is
 quite real and not a mathematical contrivance.   Wisconsin Point was breached
 on November 1,  1974, while the area at the base of Minnesota Point sustained
 substantial property damage during the northeasterly storm of January,  1975.

      It  is difficult,  however,  to attribute the damage directly to the  water
 surge, without  comparing it to the wave height  for such storms.   Significant
 wave heights for a 13 m/sec (30 mph)  northeasterly over the long fetch  would
 run 2-3  meters at Minnesota Point with about a  6-second period,  and  the
 anticipated disturbance would reach a depth of  20  - 25 meters.   It should
 also be  noted that the transport velocities due to storm surges  are high;
 thus,if  one considers  the dissipation of  energy as proportional  to the  square
 of  the velocity,  then  significant damage  can be assigned to the  transport
 velocities directly,  say ^ 15 - 25% in comparison  to the potential energy  due
 to  wave  height.

      The importance  of water  level  may also be  assessed from the remote sens-
 ing for  Douglas  County,  which shows a greater erosion  for mild northeasterlies
 than severe  westerly winds producing  comparable wave action.   The  westerly
 winds depress the water level along the Douglas County  shore.
Transports for the Entire Lake

     One of the concerns with regard to fine particles which remain in suspen-
sion for a period of months is the possibility of direct transport of these
particles in high concentrations across the international boundary.  The
possibility of such an incident was reported in April, 1975, when asbestiform
fibers were observed in the water supply of Thunder Bay, Canada.

     Predicted transports for the entire lake are shown in Figure 40 for
northeasterly  and Figure 41 for westerly winds.   Due to space limitations,
only the vectors for alternate grid points are shown, and therefore only the
gross features of the transport patterns are visible.  At the height of the
northeasterly storm,  Figure 40D,  the transport streamlines are moving away
from the shore zone at Silver Bay to the Thunder  Bay area.   Later, the part-
icles dispersed to areas away from the shore zone would apparently be
rapidly transported to the Apostle Islands area.   Streamlines peeling off
the main circulation pattern,  however,  might carry tailings into the complex
                                     71

-------
     ZL
Water Level (cm)




co" CD
c
1-1 .
CD
u>
ON
s: i>3
w
rt
(D
I—1
fD
I CD
ft)
en
H, ro
U fV

o_J
rt 1
§3
M,® OJ
**"•» CD
^ ^
s'~*
coy)
£•*— '
g CD
e
rt
34
hti »
11 _P^
O fv %
3 r\3

rt
3*
CD
rt
i-( ]\
a> •**
g co
T)
O
rt
a
O Cxi
CD -i^
*
CD
O

-------
 T
13cm
 JL
  Figure 37.   Water  levels  for western Lake Superior at one peak of the well-developed
              seiche caused by a 13 m/s NE wind, 31:20 hours after start of wind (Rand
              Corporation Model).
  Figure 38.  Water levels for western Lake Superior at lowest  water  level  during modeled
              storm.  Wind is 11 m/s westerly.   30 hours after  start  of  wind  (Rand Corporation
              Model).

-------
—J
.p-
         15cm
          MINN.
            PT.
           Figure 39.
                             DUTCHMAN
  wise.
    PT.
Water level for western Lake Superior after 15 hours of a 13 m/s NE wind.  A simple
transport model by Forristall was used with a time step of 60 sec and a smoothing
coefficient of 0.95.

-------
                                        Ni
Ouluth
                   Islands  .   . *^ -
                                        LAKE SUPERIOR
                                          I3M/SEC N.E.
                                    TRANSPORTS FROM RUN  14
                                          TIME - 6  6
                                            AT=  180
                                  SCALE 0.2 CM *  .500M2/SEC
                   Figure 40.  Entire Lake Superior transports from a 13 m/s NE wind.

-------
ON
                                         LAKE  SUPERIOR
                                          I3m/sec N.E.
                                    TRANSPORTS FROM RUN 15.
                                          TIME  =  12 12
                                            AT=  180.
                                   SCALE  0.2 CM  = 2.000M2/SEC
                                         Figure  40. (continued)

-------
            /»  *
         ^-	*^ 4   ' V^»    «•    •    •  .   • •   \
     s•-•• •:-:----:-:-^-:- /
   ^•^^^•m^4
   •:/^;--.-;-:-v>x/•>•.•:•;.;.•:-:•;.•.•;•;.•.•,
v
          X*.
         X>
          yx
              Figure 40C.
               LAKE SUPERIOR
                13m/sec N.E.
             TRANSPORTS FROM RUN 7.
                TIME = 24 24
                 AT = 180.
            SCALE 0.2 CM = 3.003M2/SEC
              Figure 40. (continued)

-------
—I
oo
                                         LAKE SUPERIOR
                                            I3M/SEC N.E.

                                     TRANSPORTS FROM RUN 24

                                           TIME  = 36  6

                                             AT = 180

                                    SCALE 0.2 CM  = 3.003M2/SEC

                                        Figure 40.  (continued)

-------
  /*  v x>^  - I  ,   •  •  .  .   *  *  •
s>? ' /Y  '»».'.•••.•.•••
^*\%C^''.\4.fc  •  -"V  •  •  .'.\
^^.V-','.-;^'1.•,%%•'.'.•,-.
 K       *^ »  _  .  ^*  _»      *  •  _
        LAKE SUPERIOR
           13 M/SEC N.E.
     TRANSPORTS FROM RUN 28
         TIME = 60 6
            AT= 180
    SCALE 0.2CM = 3.003M2/SEC
        Figure 40.  (continued)

-------
00
o
                   Figure 41.
           LAKE  SUPERIOR
             II M/SEC W.
         TRANSPORTS FROM RUN 21.
             TIME = 18 18
               AT  =  180.
       SCALE 0.2 CM = I.OOOM2/SEC
Entire Lake Superior  transports from a 11 m/s westerly wind,

-------
00
                                           LAKE SUPERIOR
                                               11 m/sec W.
                                        TRANSPORTS FROM RUN 23.
                                             TIME = 30 30
                                               AT= 180.
                                     SCALE 0.2CM = 2.000M2/SEC
                                           Figure 41.   (continued)

-------
oo
N3
                                        LAKE SUPERIOR
                                          II M/SEC W.

                                    TRANSPORTS FROM RUN 30

                                         TIME = 42  6

                                           AT* 180

                                  SCALE 0.2 CM * L499M2/SEC

                                     Figure 41.  (continued)

-------
oo
u>
                                     LAKE SUPERIOR
                                        IIM/SEC W.
                                   TRANSPORTS FROM RUN 31
                                       TIME = 48  6
                                         AT = 180
                                 SCALE 0.2 CM = I.499M2/SEC
                                     Figure 41.   (continued)

-------
eddies seen southwest of Isle Royale, Figures 40D and 40E.  Although  it  is
difficult to say, the tailings might then work their way toward the  Thunder
Bay area, aided by the mixing processes and dispersion.  Further studies
pertaining to the distribution of currents with depth and a long-duration
concentration model would shed more light on the problem.

     According to the model, the rapid, well-confined transport of tailings
from Silver Bay is possible only to the Duluth area and the Apostle  Islands,
for the wind stresses considered here.  Similarly,the red clay particles
tend to be confined in high concentrations to far western Lake Superior or
are transported around the Apostle Islands.  Thus,high concentrations of the
contaminants are generally confined to western Lake Superior.  Another major
transport structure in the lake for the conditions modeled here is a circu-
lation cell off the Keweenaw Peninsula, where large transport values are
evident.   This circulation cell for westerly winds is shown in Figure 41.
Figure 41C shows a counterclockwise loop, with extremely high transports
along the northern shore of the peninsula.   Particles dispersing from Silver
Bay could be carried around the Apostle Islands to the Keweenaw circulation
cell.   Comparison of currents at the south shore of the study area, where
currents on the order of 20 cm/sec were measured,  indicates that the currents
along the Keweenaw would correspond to 40 - 60 cm/sec.   Large transports are
also evident in Figure 41C in the strait between the North Shore and  Isle
Royale but are of no direct interest here.   The westerly winds in the near-
shore zone show transports directly away from the shore at Silver Bay.

     Lake-level data from the model for the points  shown in Figure 42 are
presented in Fourier Analysis form in Figures 29 and 43.
                                    84

-------
                                          BATTLE  ISLAND
                                               NO. 8
                         THUNDER BAY
                             NO. 7
              GRAND MARAIS
                    NO. 3
MICHIPICOTEN
  HARBOR
     NO. 9
                                  N0.4
                              ONTONAGON
                                              NO. 5
                                         MARQUETTE
 TWO HARBORS
g     NO. 2
DULUTH
 NO. I
        NO. 6
     WHITEFISH
        BAY
                       Figure 42. Lake Superior water-level station locations.

-------
                                         98
       RELATIVE  AMPLITUDE  (Normalized)
            p
            ro
                      p
                      at
p
00
                                             RELATIVE   AMPLITUDE  (Normalized)
p
ro
    ro
09
a  5
CO


X)
ro
i-i
H-
o
§
ro

ro
§
55
                                             0)
                                             oo
g
   RELATIVE AMPLITUDE  (Normalized)
        p      p      p      p     —
        ro      *            09     b
               RELATIVE  AMPLITUDE  (Normalized)
P
ro
                                   P
                                   o>
                                                                           P
                                                                           oo
    o>
en	
rt^

if*
                                             o>
                                           3T
                                           3
                                         ; ro

-------
00
—I
      1.0
      0.8
      0.4
   UJ

   <0.2

   UJ
   OC
FOURIER ANALYSIS - MODEL
Northeaster  8'00- 38:42
Four Full Cycles Pack To Peok
  WHITEFISH  BAY
Grid Location  (2,73)  No.6
                                 8     10     12
                                  PERIOD (hours)
          14
16
                                                           18
   1.0



==0.8

&
   B
      0.6
      0.4
   UJ
   UJ
   rr
                                         FOURIER ANALYSIS-MODEL
                                         Northeaster  8=00-38' 42
                                         Four Full Cycles Peak To  Peak
                                           BATTLE   ISLAND
                                         Grid Location  (48,69) No. 8
                                 8     10     12     14
                                 PERIOD (hours)
                                 1.0
               iO.8
                                                                       UJ
                                                                         0.6

                              ut
                                                                       cr
                                                                          1.0
                                '0.8
                                                                                                        FOURIER ANALYSIS-MODEL
                                                                                                        Northeaster  8«00- 38=42
                                                                                                        Four Full Cycles Peak To Peak
                                                                                                          THUNDER  BAY
                                                                                                        Orld Location (51,49) No. 7
8     10    12
 PERIOD (hours)
                                                                              14
16
18
                                                   FOURIER ANALYSIS-MODEL
                                                   Northeaster  8=00- 38=42
                                                   Four Full Cycles Peak 1b Peak
                                                     MICHIPICOTEN  HARBOR
                                                   Grid Location  (17,81) No. 9
                16    18
                 Figure 43.
                   024
                 (continued)
                                                                                                 8    10     12     14
                                                                                                 PERIOD (hours)
                         16
      18

-------
   CURRENT PROFILE MODEL
  or
  valuable aid in predictine tt           §enera  f low Patterns and are a
  little quantitativ" information aLTthe  * ^T^ Sedim-t.  Hoover,
  surface and bottom currents can be nK/   maf ^udes and directions of
  finite difference schemes utilizil threTd •      "*, ^^^  Multi-level
  these suffer from very large computer stor dimen*1Onal §rids «« possible, but
    c         lack
                                            * **?*?*
  the current profile at any point  is evfluatfd H  F°rrls1ta11  ^W>  - which
  the wind-stress and water-slope histories  a ***CO™°1^™ integrals over
  is obtained without the use of the contnn  ll      P°int<  A Uni*ue  solution
  a large three-dimensional Jrid is disoen   ^ty,e^Uation-  Thus, the  need for
  evaluated only where desired      dlsPensed with and the currents need be
 Mathematical Formulation
 undisturbed water level and I
 cities and accelerations and
 •otion,  uhlch the current,
                                                        are the , and
                                 «  ,f         SyStel° "Ith z ' ° «t the
                                 -linear'""1   NeSlectl"S """ical  velo
                                 '1"
                             9w              ^
                             ,J - ifw - q + vJf
"here f Is the coriolis
                        parameter and v is the vertical eddy viscosity.
static pressure.
                                                                 °f
                                    88

-------
     For the case of a pure drift current, we have
                          3w
                                             2
                                            .3 w
with the boundary conditions




     w(z = -H) = 0

                              (z  . o)  - F(t)
                                                                     0) -  0
where F(t) is the time-varying wind-forcing function.  The solution used by

Forristall is
     w(t,z) =     Z  n cos  [n
              H   n— u
     For a pure slope  current
                                    .-     .
                             — = -ifw -q +
                                              32w
with boundary conditions
     w(z = -H) = 0
                            (z  -  0)  - 0

                                                          (t = 0) =  0
the solution is
W(t'z)
    -2  B

     H  ^=
                                             rt /t  . -ifi -(n4J5)2ir2vT/H2
                                             /oq(t-i)e    e
     The solution  for  the pure drift current was actually found  to  contain

three parts:
w(tfZ) = F(t)
                                                                           (i)
       - F(t)    E=  A   cos  [(tttJs)irz/H]
                                                                          (ii)
                                                           22     2
              2  »         r/  ,i\   /TTT  /-t^/   \ -ifT -(n-f^s) TT VT/H dx    /...\
            + —  Z    cos  [(n+*5)irz/H]  / F(t-t)e    e                    (iii)
              H  n=o                    o
                                      89

-------
where         A  = 2- n
                      2    2
                        - Y
               'n      4 + f2/v2
                     n
               Y =
All three parts are needed to satisfy the differential equation, but parts
(i) and (ii) cancel each other in the sense that part (ii) is a Fourier series
for part (i) .  All three parts satisfy the conditions that w(z = -H) = 0 and
w(t - 0) = 0.  Part (i) satisfies the surface boundary condition
    (z = 0) = F(t) with no contribution from parts (ii) and (iii).  Part (ill)

has been used alone for a good numerical approximation to the solution even
though it does not match the surface boundary condition explicitly.  However,

numerical calculations show that the slope -^- approaches the boundary value
  1                                        oZ
v  F(t) a small distance 6z away from z = 0.  As seen in Figure 44, 6z
decreases but apparently never vanishes for increasingly large values of N,
the upper limit on the series.

Treatment of Time Integral

     The time integrals appearing in the solutions are uneconomical in their
present form for numerical evaluation of the currents.  For example, to
calculate the currents at two times T^ and T£, where T£ > T^, the effort done
in evaluating the integral from time 0 to T^ must be repeated to do the inte-
gral from 0 to T£.  Therefore, a recursion formula was developed for the
computer program which allowed the integral to be evaluated at a series of
equally spaced times, without wasted computation.  This was done by writing
the time integral from 0 to some time T^ = kAT as a sum of k integrals over
equally spaced intervals 0 to T-^, T^ to T£, etc.  Making a change in the
integration variable, which allowed the values of the wind function and water
slopes to progress forward in time, yielded
                     Ck =
                          f'k   F(t)e-9n(Tk - T)dt + e-VTC,  .
                           \-l                             fc~1
                                     90

-------
      Component Of Current  In Direction Of Wind   (cm/sec)
   -27              -26              -25               -24
           N=200
                N=IOO
          NORTH SHORE
        13 m/sec NE WIND
       Model Time \2-00 Hours
             t/=.OI5
Slope Needed To Meet.
Boundary Conditions
                                                               Nl
                                                           .01
    o>
   Q
    o>

.021
    0)
   o:
                                                           .03
        Figure 44.   Current vs. relative depth near surface.
                    N
Superior Entry
                        km.
8      x 1974 Buoy Location
       • Current Profile Grid Location
       Figure  45.   Current  meter station locations.
                                91

-------
 where                    C  = 0
                           o
                          e  = if + VB 2
                           n          n
      Assuming that F(t)  changes slowly over the small interval from Ti
 to Tk made it possible to further reduce the effort  in obtaining the
 integral by approximating it  in closed form using the average value of  Fft)
 over that interval;
                                            Tk-l

                    —          2
                _    F     r -3  VAt,    2                      ?
                	4~~2—2  ^e  n    ("^  vcosfAT+fsinfAT) +  3  v]
                  3  v +f              n                    n
                  n


                      —          2
                 	iF     r -g  vAT,    2
                	T~^—9  [e  n    (-g  vsinfAT-fcosfAT) +  f]
                  3  v +f              n
                  n


The time integral involving the water slopes is treated in  the same manner.

Modeling Procedure

     Six current-monitoring stations were set up in the western tip of Lake
Superior during the summer months of 1974, to provide data for the calibra-
tion of the model.  Figure 45 shows the locations of the current-meter
stations superimposed on the grid used  for the current profile model.
Surface currents were measured at a depth of approximately 2 meters and
bottom currents at a height of 1.8 meters above the lakebed.  Stations 1, 2,
and 4 were monitored for both surface and bottom currents while only surface
currents were recorded at stations 3, 5, and 6.

     The current profile model was implemented in the following manner:

1)  First,  the Rand Corporation model was applied to the entire Lake Superior
    basin using a grid of 81 x 56 depth locations with a spacing of 6 km.

2)  Time-dependent water levels and transports were computed for the north-
    east and westerly  winds,  previously described in the section on trans-
    ports,  and stored  every 12 minutes for the points located  in the
    western arm of the lake.
                                     92

-------
 3)  The resulting water-level histories were spatially interpolated and
     extrapolated to obtain the x and y components  of the water  slope at each
     of the transport locations,  and these values were stored  for  use by the
     current model.

 4)  Finally, the current profile model was run,using the previously obtained
     water slopes  and the original wind-forcing  functions to  predict currents
     in the western arm.   Special attention was given to points  at or near
     locations where current meter data were available.

      Before the model could be used,  it was necessary to better understand
 the effects of the vertical-eddy viscosity v.  Figure 46 shows  the computed
 current profiles using two different  values of v.   The surface  currents are
 displaced to the right of the wind due to coriolis  effects,and  the subsurface
 currents rotate clockwise with depth  in a modified  Ekman spiral, causing a
 reversal of the currents near the bottom.   Raising  the viscosity  has the
 effect of decreasing the currents and somewhat altering the shape of the
 profile.   Lower viscosities decrease  the momentum  transfer between layers
 of water, tending to uncouple the surface and bottom currents  and  also  to
 decrease the bottom drag.   Thus,  in the case of  Figure 46, both the surface
 current and bottom-return current increase while the transport  changes  only
 slightly.

      Because little is known about  the actual value of  the vertical-eddy
 viscosity,  the current model was "calibrated" by varying  v and  matching
 predicted  surface  current speeds with measured values  in  the  lake.   A
 comparison of  peak surface currents versus  corrresponding wind  speeds  is
 shown in  Figure 47.   Stations 2  and 3 were  selected  because they  are most
 likely to  be free  from shore effects  for  northeast winds.  Some of  the
 scatter  in the points  is  probably caused  by  the  fact  that the wind measure-
 ments,  obtained from the  Duluth  Coast  Guard, were generally taken  only  at
 6-hour intervals,so  little is known about  their  speeds  and directions  in the
 interim period.  Also  the  surface-current measurements, although averaged
 hourly, are  expected  to be influenced  by  the violent wave turbulence caused
 by  higher winds.   The  relationship between  the wind  speeds and  the  surface-
 current speeds  is  a  complicated  one which depends on the wind stress, wind
 duration,  total depth  of  the  water, wave height, and vertical-eddy viscosity.
 In  reality,  the value  of  the vertical-eddy viscosity must also vary as a
 function of  depth  because  of  turbulence.  When a wind stress appears over the
water, a turbulent surface layer associated with wave action is developed,
which  implies a large  surface-eddy viscosity.  Deeper,  the flow becomes more
laminar implying a lower eddy viscosity.  At the bottom the eddy viscosity
may increase somewhat due  to a thin layer of turbulence where the current
speeds suddenly approach zero.  In the current profile model,  the vertical-
eddy viscosity parameter had  to be taken as a constant in order to find a
solution to  the differential equation.  To calibrate the model,  a value of
 .007 m /sec was used for v, which yielded peak currents at the depth of the
surface meters of about 36 cm/sec near stations  2 and 3 for the modeled
13 in/sec (30 tnph) northeast wind.
                                     93

-------
                                        40
                                       ••30
                     v = .01
         Wind  Direction
         (13 m/sec)
v =.005
    -40             -20
           (cm/sec)
     20
Figure 46.  Current profiles for  two values  of  v.   Vectors
            plotted every one-tenth of  the relative depth.
£\J
15
1
I 10
•a
a

-------
 Results

      A detailed analysis of the model on currents is treated elsewhere
 (Maanum, 1977).  The primary need for modeling currents came from the
 interest in movement of the sediment after its resuspension from major
 sediment deposition areas in the vicinity of the harbor entries and from
 dredging disposal areas.  It is seen from Figure 48 that for the times of
 high turbulence, the bottom currents for westerly winds would transport
 the bottom layer of water towards the northeast shore, indicating an up-
 swelling along that shore.  For a small area east of the Superior entry,
 however, the bottom currents point towards Dutchman Creek, making this
 location a more suitable dredging disposal site from the standpoint of
 currents.

      For the northeasterly winds, the bottom layer of water generally moves
 out in an easterly direction similar to the transport patterns near the
 axis of the arm.  The results for the wind directions modeled here gener-
 ally confirm the measured current patterns.  The bottom currents treated
 here pertain to thermally unstratified conditions which are important in
 questions of the occurrence of high turbidity and dredging disposal.


TURBIDITY TRANSPORT MODEL

     A combination of 1) transports from the numerical model, 2) information
on the erosion along the south shore, and 3) turbidity settling measured in
the lab, Figure 49, produced the preliminary information needed for the
numerical modeling of turbidity transport and location of the settling areas
which become potential resuspension sources.  A nearly maximum severity
condition was modeled.  This entailed northeasterly winds at 13 m/sec (30 mph)
for 40 hours, followed by a lull in the wind and a subsequent westerly wind
patterned after actual weather observations.  The boundary conditions for the
model were as follows:  Erosion took place at a uniform rate throughout the
northeasterly wind, with the distribution along the south shore provided from
data on the relative shore erosion rates.  Because of the high turbulence no
settling was allowed during the easterly, but at the end of the storm a
settling rate was  applied everywhere except at the source locations.  The
turbidity-source locations consisted of 1.5 x 3 kilometer blocks, i.e. half a
grid square, staggered adjacent to the south shore.  (More experimental work
on the boundary conditions is essential for further detailed turbidity model-
ing.)  In a sense,  the horizontal eddy diffusivity process was automatically
taken into account by the unavoidable mixing after each time step of the
influx of higher turbidity from adjacent cells,with the suspended load already
present in a particular grid space.  The flux of turbidity into each grid
space was determined by the transports and concentration in neighboring cells.
The results for the turbidity plumes are shown in Figure 50 at various stages
of the storm.   The total suspended load for red clay is 1.3 x 10^ metric tons,
resulting from an input of 500 mg/£ constant suspended load at a source point
at Amnicon which is comparable to real values for such storms.   This value of
total suspended load is roughly equal to ^ 25% of the yearly lakeshore
                                     95

-------
Wind direction
at time of plot
               /   /   ,-•••••,  ^
               N.   "   •'      : s^.'
                  km.
5 cm/sec
                                                                          km.
5 cm/sec
               Figure  48.   Computed bottom currents near Duluth measured  1.8 m above the lake
                            after 24 hours of west wind and 30 hours of east wind.
                                                                 bed

-------
          PLUME MATERIAL REMOVAL
         10               20
                Days
Figure 49.  Lake sample settling in lab.
30
                  97

-------
                                                                       SUSPENDED  SOLIDS (mg/l)
00
                                        SUSPENDED LOAD CONCENTRATION
         Duluth
          Sup3ricr
               MODEL TIME = 42 hours (2 hours  into  decay of 13 m/sec  Northeasterly Wind)
                  Figure 50.  Modeled suspended solids distribution at several  stages of the storm.

-------
                                                               SUSPENDED  SOLIDS  (mg/l)
VO
                                                                                  10-20
                                                                                  20-50
                                    SUSPENDED LOAD CONCENTRATION
        Duluth
          Superior
        MODEL TIME = 60 hours (12 hours after end of I3m/sec Northeasterly Wind)
                                            Figure  50.  (continued)

-------
                                                           SUSPENDED SOLIDS  (mg/l)
o
o

   Duluth
     Super!
                                                                   a
                                                                          HO



                                                                          10-20



                                                                          20-50



                                                                          50+
                              SUSPENDED LOAD  CONCENTRATION
              MODEL TIME = 120 hours (18 hours after end of Mm/sec  Westerly Wind)

                                          Figure 50.  (continued)

-------
 erosion input from the entire red clay bank erosion in Douglas and  Bayfield
 Counties.  Experimental data indicate  that the major portion of  the erosion
 takes place during three t:o five severe northeasterly storms  occurring per
 season.  It is interesting to compare the shape of calculated plumes to the
 plume observed in the Landsat image for April 3, 1975,  Figure 51.   This
                    Figure 51.   Landsat image for  03APR75.
plume was due largely to resuspension in the near-shore areas of Douglas and
Bayfield Counties and areas off Minnesota and Wisconsin Points.  It looks
much like the model plume.  The deposition areas produced in this model are
shown in Figure 52.  Notice the high deposition rate east of the Superior
entry to the Duluth-Superior Harbor, Figure 52A.  The Nemadji River output
is also near here ,though its load is normally transported in a southeasterly
direction.  On navigation charts and USGS maps, this area is shown as actu-
ally about 4 meters shallower than the adjacent areas of the lake.  This
roughly corresponds to about half of the amount of material which would be
deposited there at the rates shown in Figure 52B, taking the age of the lake
at 10,000 years.  Of course, this is a rough result, probably fitting the.
data better than the roughness of the model deserves; yet it generally
reflects the actual conditions quite well.   Some available data on cores and
sediment traps is yet to be analyzed  and will possibly be done in the future.
There is a general need for measurements necessary to describe the boundary
conditious which would closely approximate the actual erosion source.
                                     101

-------
                                              Silver Boy
o
N>
          Dufcth
           Superior
                                                                   ACCUMULATED SEDIMENT (g/cm*)


                                                                                     0.02-


                                                                                     0.02-0.06
         Figure  52A.




SEDIMENT  DEPOSITION
                    MODEL TIME = 60 hours (12 hours after  end  of 13 m/sec  Northeasterly  Wind)

                    Figure 52.  Modeled sediment accumulation at different stages of the  storm.

-------
                                                      ACCUMULATED SEDIMENT (g/cm2)
                                                                  |    |   0.02-

                                                                          0.02-0.06
                                   Sliver Boy
                                  SEDIMENT  DEPOSITION
Duluth
Superior
              MODEL TIME * 120 hours (18 hours after end of II m/tec Westerly Wind)
                                    Figure 52.  (continued)

-------
                               REFERENCES

 Bahnick,  D.  A.,  J.  W.  Horton,  R.  K.  Roubal,  and  A.  B.  Dickas,  1972.
   "Effects  of  South Shore Drainage Basin and Clay Erosion on the Physical
   and  Chemical Limnology of  Western Lake Superior." Proc.  15th Conf.
   Great Lakes  Research,  Int. Assoc.  Great Lakes  Res.

 Baumgartner, D.  J.,  W. F.  Rittall,  G.  R.  Dittsworth, and  A.  M.  Teeter,
   1973.   "Investigation  of Pollution in  Western  Lake Superior  due to
   Discharge  of Mine Tailings,  Data Report 1971."  Studies Regarding the
   Effect  of  the Reserve  Mining Company Discharge in Lake  Superior,
   Pacific Northwest Environmental Research Laboratory  Paper  No.  10,
   U.S.  EPA,  Washington, D.C.,  p.  423.

 Forristall,  G.  Z.,  1974.   "Three-Dimensional Structure of Storm-Generated
   Currents."   J. Geophys.  Res., 79(18),  2721-2729.

 Hutchinson,  G.  E.,  1957.   A  Treatise on  Limnology.  Wiley: New York,
   Vol.  1, p. 263.                          ~~~

 Leenderste, J.  J.,  1967.   Aspects of a Computational Model for  Long-
   Period  Water-Wave  Propagation.   RM5294-PR,  The Rand  Corporation.

 Maanum, W. E.,  1977.  Numerical Prediction of  Currents and Transports
   in Western Lake Superior.  M.S.  Thesis,  University of Minnesota, Duluth.

 Mortimer, C. H.  and  E. J.  Fee, 1976.   "Free  Surface Oscillations  and
   Tides of Lakes Michigan  and  Superior."   Phil. Trans.  Roy. Soc. London,
   281, 1-61.                             ~	

 Rao, D. B. and D. J. Schwab, 1976.   "Two-Dimensional Normal Modes in
   Arbitrary Enclosed Basins on a  Rotating  Earth:  Application to Lakes
   Ontario and  Superior."  Phil. Trans. Roy.  Soc.  London,  281,  63-96.

 Ruschmeyer, 0.  R., T. A.  Olson, and  H. M.  Bosch,  1961.   Lake Superior
   Studies 1956-61.  Public School of Health, University of Minnesota,
  Minneapolis.

 Sydor, M.  1973.  Current Patterns and  Turbidity.  U.S.  Army Corps of
   Engineers Report DACW-37-74-C-0014.

 Tsai, Y. J. and Y.  C. Chang, 1974.   "Prediction  and Verification of
   Storm Surges in Lake Ontario and Lake Erie."   Stone and Webster
  Engineering Corporation, Boston, Mass.    Paper  presented at the 17th
   Conf. on Great Lakes Research.

Wilson, B. W.,  1960.  "Note on Surface Wind Stress Over Water at High
  and Low Wind Speeds."  J. Geophys. Res., 65(10) 3377-3382.

                                 104

-------
                APPENDIX A
    SOUTH SHORE DATA FOR THE 1975 RUNOFF
Note:  Data enclosed in parenthesis ( )are




     interpolated from measured data.
                   105

-------
TABLE A-l.  SUSPENDED LOAD TRANSPORTED FOR SNOW RUNOFF
               (APRIL 10 - 27, 1975) IN METRIC TONS
River
Nemad j i
Dutchman
Morrison
Amnicon
Ten
Wagner
Hanson
Middle
Poplar
Bardon
Pearson
Load South
of Hwy 13

224
149
2,093
260
134
47
1,583
351
225
220
Load Between
Hwy 13 and
Lower Station

88
146
350
102
124
81
138
920
192
446
Load From
Lower Station
to Lake

5


6
6

15
12
9
23
Total Load
Discharged
84,848
317
295
2,443
368
264
128
1,736
1,283
426
699
Unnamed
Stream #
1
2
3
4
5
Total Load
Discharged
6
66
39
20
12
Unnamed
Stream #
6
7
8
9
10
Total Load
Discharged
20
29
41
90
74
Unnamed
Stream #
11
12
13
14
15
Total Load
Discharged
24
17
14
54
13
Total load from all unnamed streams 519 metric tons.
                        106

-------
   TABLE A-2.  LOAD TRANSPORTED DURING RAIN EVENT
                   (APRIL 28 - MAY 5, 1975)
                        IN METRIC TONS
River
Nemad j i
Dutchman
Morrison
Amnicon
Ten
Wagner
Hanson
Middle
Poplar
Bardon
Pearson
Load South
of Hwy 13

108
17
616
26
11.4
4
218
71
48
29
Load Between
Hwy 13 and
Lower Station

30.8
29.2
98
35.7
18.6
16.2
48.3
278
67.2
103
Load From
Lower Station
to Lake

1.8


1.6
1.6

6
4.8
2.4
11
Total Load
Discharged
22,044
140.6
46.2
714
63.3
31.6
20.2
272.3
358.8
117.6
143
Unnamed
Stream ?/
1
2
3
4
5
Total Load
Discharged
0.4
4.4
2.6
1.3
0.8
Unnamed
Stream #
6
7
8
9
10
Total Load
Discharged
1.3
1.9
2.7
6.0
4.9
Unnamed
Stream #
11
12
13
14
15
Total Load
Discharged
1.6
1.1
0.9
3.6
0.9
Total from all unnamed streams 34.4 metric tons.
                         107

-------
                                 TABLE A-3.  SOUTH SHORE STREAM CHARACTERISTICS
River
Namad j ±
Dutchman
Morrison
Amnicon
Ten
Wagner
Hanson
Middle
Poplar
Bardon
Pearson
Total
Basin
Area
(km2)
1191
19
13
111 *
349 +
17
13
11
59 *
124 +
55 *
104 +
32
27
Total
Lengths
Main Streams
Intermittents
(km)
105
30.1
2.7
20.4
7.2
* 182 *
6.4 *
20.8
24.0
7.9
17.1
1.4
12.2
57.0 *
25.7 *
65.8 *
38.3 *
26.7
38.1
26.7
53.4
South of Hwy 13

• Basin Main Inter-
Area Streams mittents
(km2) (km) (km)
14 24.0 1.3
8 11.6 4.3
98 * VL60 * 5.0 *
13 15.3 20.6
7 2.7 11.6
4 1.4 3.5
52 * 48.8 * 20.8 *
23 * 38.9 + 11.1 *
24 16 32.3
13 16 6.8
Between Hwy 13
and Downstream Station
Basin Main Inter-
Area Streams mittents
(ko2) (km) (km)
5 5.6 1.4
5 8.8 2.9
13 21.1 1.4
4 4.8 3.4
.5 4.5 5.5
6 0 8.7
6 6.6 4.5
31 25.6 27.2
8 9.7 5.8
12 3.1 46.7
Between Downstream
Station and Mouth
Basin Main Inter-
Area Streams mittents
(Ion2) (km) (km)
0.5 0.5 0
00 0
00 0
0.3 0.6 0
0.3 0.6 0
00 0
0.8 1.6 0
1.0 1.3 0
0.8 1.0 0
2.6 3.5 0
o
00
          + University of Wisconsin measurement
* includes only portion within red clay belt

-------
TABLE A-4.  SOUTH SHORE UNNAMED STREAM CHARACTERISTICS
Unnamed
Stream
#
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Total
Basin
Area
(km2)
0.5
3.6
2.3
1.6
0.8
1.0
1.3
2.1
3.9
4.1
1.0
1.0
0.8
3.1
0.8
Total Length
(km)
Main
Streams
0.5
6.6
3.7
1.8
1.3
1.4
0
0
0
0
0
0
0
0
0
Inter-
mittents
0.2
0.5
0.5
0.3
0
0.6
3.1
4.3
9.7
7.9
2.6
1.8
1.4
5.8
1.4
                         109

-------
          TABLE A-5.   SOUTH  SHORE EROSION DATA (1975)  FOR NEMADJI
                                GOLF COURSE AND HWY 53
Date
April 12

April 13
April 14
April 15
April 16

April 17



April 18





April 19


April 20
April 21





April 22
April 23






Water
Sample
Description
N-12-1
N-12-2



N-16-1
N-16-2
N-17-1
(D) Hwy 53
N-17-2
(I) Hwy 53
N-18-1
N-18-2
N-18-3
(D) Hwy 53
N-18-4
(I) Hwy 53
N-19-1
N-19-2
N-19-3

N-21-1
N-21-2
N-21-3
N-21-4
N-21-5
N-21-6
N-22-1
N-23-1
Hwy 53
N-23-2
Golf Course
N-23-3
N-23-4
N-23-5
Turbidity
(NTU)
50.0




143

220

215

148
143
152

158

162
162


148
152




205
242

260

258


Suspended
Solids
(mg/£)
61.7 (58.2)
54.7
(158)
(261)
(364)
472 (476)
(479)
580 (540)



417
(460)




523 (524)
530
521
(494)
465
463
439 (448)
432 (UWS)
457
416 (UWS)
852
1055

927 (1059)

1113
1065
1004
Stage Discharge
(cm) (m^/sec)
298 21.0

(348) (53.1)
(399) (85.2)
(450) (118)
497 150

504 152



542 163





509 154


(495) (148)
481 142





455 127
351

476 140




Total Load
(Metric Tons
Per Day)
105

725
1922
3692
6150

7080



6472





6975


6324
5508





9360


12801




(continued)
                                  110

-------
TABLE A-5.  (continued)
Date
April 24
April 25
April 26
April 27
April 28
April 29
April 30
May 1
May 2
May 3
May 4
May 5
lay 6
May 7
May 23
Water
Sample Turbidity
Description (NTU)
N-24-1
N-24-2
N-24-3
N-24-4
N-25-1
N-25-2
N-25-3
N-25-4


N-28-1
N-29-1

N-l-1
N-l-2
N-2-1


N-5-1


N-23-1
192
190
187
148
148
148


142
200

108
100


85


26.0
Suspended
Solids
(mg/Jl)
423
425
405
425
392
401
409
387



557

284
286
257


182

165
29.5
(424)
(397)
(347)
(307)
(290)

(420)
(285)

(232)
(207)

(173)


Stage
(cm)
512
489
(466)
(443)
420
456
(462)
469
390
(366)
(343)
319
(310)
(301)

Discharge
(m3/sec)
155
147
(132)
(116)
101
127
(131)
135
80.4
(65.3)
(50.2)
35.1
(29.2)
(23.2)
6.0
Total Load
(Metric Tons
Per Day)
5665
5031
3946
3090
2540
6120
4753
3319
1785
1310
898
552
436
331
15.2
        111

-------
      TABLE A-6.  SOUTH SHORE EROSION DATA  (1975) FOR DUTCHMAN, HWY  13
1 	
Water
| Date Sample
Description
April 7 D-7-1
D-7-2
April 8

April 9

April 10 D-10-1
D-10-2

April 11

April 12
Suspended Total Load
Turbidity Solids Stage Discharge (Metric Tons
(NTU) (mg/10 (cm) (m3/sec) Per Day)
65.5 (50)
65.0
(55)

(60)

66.0 71
61.5 (66)

(67)

(75)
0.11

0.14

0.17

0.23
0.37



0.54
0.49

0.67

0.88

1.29
2.13



3.49
April 11
i
April 12
i
I April 13
1
JApril 14



April 15

JApril 16

l
i
April 17
iApril 18
April 19
April 20
,
April 21
April 22
April 23
April 24
April 25


D-14-1
(I)
D-14-2
(D)
D-15-2
D-15-3
D-16-1
D-16-2


D-17-2
D-18-1
D-19-1


D-21-2
D-22-1
D-23-2
D-24-1
D-25-2


70.5

70.5

60.5

85.0



66.0
75.5
74.5


68.5
79.5
97.5
70.0
71.0
(67)
(75)
(84)

93.3 (92.6)

92.0

58.0 (66) 51
73.7
ICE JAM
119


53 ICE JAM
107 ICE JAM
83
(86)

86 51
(107) 53
128 51
49 43
48 38
0.54
0.71

0.85



2.17

3.86



3.48
3.09
2.72
2.36

2.08
2.36
2.08
1.20
0.69
3.49
5.14

6.80



12.4

39.7



16. 0
28.5
19.5
17.5

15.5
21.8
23.0
5.1
2.9
(continued)
                                    112

-------
TABLE A-6.  (continued)
Water
Date Sample
Description
April 26
April 27
April 28 D-28-1
April 29
April 30 D-30-1
May 1 D-l-1
(D)
May 2
May 3
May 4
May 5 D-5-1
(D)
May 6
May 7
Suspended
Turbidity Solids Stage Discharge
(NTU) (mg/£) (cm) (m3/sec)
(76)
(104)
113 132 52
(105)
80.0 78
70.0 (70.5) 41
(78) (34)
(86) (33)
(93) (30)
76.0 (100) 25
(80) (25)
(60) 24
1.19
1.70
2.21
4.11
1.47
0.57
0.45
0.38
0.28
0.14
0.14
0.11
Total Load
(Metric Tons
Per Day)
7.8
15.3
25.2
37.3
9.9
3.5
3.0
2.8
2.3
1.2
0.94
0.59
         113

-------
TABLE A-7.  SOUTH SHORE EROSION DATA (1975) FOR DUTCHMAN
             MOUTH, MORRISON HWY 13, AND MORRISON MOUTH
April 9
April 15

April 16

April 17

April 18




April 21
April 23
April 25
April 30

Morrison,
April 16
April 25

April 28

April 30

May 5
Morrison,

April 16

April 25
D-9-1
D-15-1

D-16-1
D-16-2
D-17-1

D-18-1
D-18-2
(D)
D-18-3
(D
D-21-1
D-23-1
D-25-1


Hwv 13
Mo-16-1
Mo-25-1

Mo-28-1

Mo-30-1
(D)
Mo-5-1
Mouth

Mo-16-1
Mo-16-2
Mo-25-2
65.0
69.0


112
88.0


100

102

85.0
140
80.5



52.0

91.5

64.5

62.0

118

83.0
43.0
82.0 2.19

216

110
i
162
f
t


134
241
77 0.75
1.37
i
67
45 0.34
i
103
i
i
0.93

0.34
i

259
93
                         114

-------
      TABLE A-8.  SOUTH SHORE EROSION DATA (1975) FOR AMNICON, HWY 13
Date
April 10
April 11
April 12
April 13
April 14
April 15
April 16
i
April 17
April 18
April 19
April 20
April 21
April 22
April 23
April 24
April 25
April 26
Water
Sample
Description
A-10-1
A-10-2



A-14-1
A-14-2

A-16-1
A-16-2
(5 br
after
A-16-1)
A-17-1
(D)
A-17-2
(I)
A-18-1
(D)
A-19-1

A-21-1

A-23-1



Suspended
Turbidity Solids
(NTU) (mg/2.)
16.5 12.5
14.3 (13.4)
(40.3)
(68.2)
(96.0)
60.6 125 (122)
119
(100)
59.5 55 (81)
107
38.0 50
36.5
37.0 59
29.5 58
(45)
25.0 31
(51)
32.5 71
(66)
(62)
(58)
Total Load
Stage Discharge (Metric Tons
(cm) (m3/sec) Per Day)
61
(74)
(89)
(104)
(117)
(137)
157
178
178
169
(175)
183
(183)
193
183
(178)
(173)
0
0
(3.09)
(7.65)
(10.5)
(17.0)
24.8
32.1
32.1
28.3
(31.4)
34.7
34.7
39.1
34.7
(32.1)
(30.2)
0
0
18.2
63.4
110
147
173
138
164
142
122
93
153
240
198
172
152
(continued)
                                    115

-------
TABLE A-8.  (continued)
Water
Date Sample
Description
April 27
April 28 A-28-1
April 29 A-29-1
April 30 A-30-1
May 1 A-l-1
(D)
May 2
i
May 3
;May 4
iMay 5 A-5-1
(D)
May 6
May 7
May 23 A-23-1
Suspended
Turbidity Solids
(NTU) (mg/£)
	 — 	 —
(54)
54.5 71
30.5 (50)
25.0 30
21.0 24
(22)
(19)
(17)
16.5 (14)
(11)
(8)
9.5 3.8 (7)
Stage
(cm)
(165)
170
183
183
165
155
(147)
(140)
132
(124)
117

Total Load
Discharge (Metric Tons
(m3/sec) Per Day)
(27.7)
29.0
34.7
34.7
27.7
23.2
(20.1)
(17.6)
15.9
(12.9)
9.63

	
129
178
150
90
57
44
33
26
19
12
7

       116

-------
        TABLE A-9.   SOUTH SHORE EROSION DATA (1975)  FOR TEN HWY 13,
                    WAGNER HWY 13, WAGNER MOUTH,  AND HANSON MOUTH
Water Suspended
Date Sample Turbidity Solids
Description (NTU) (mg/£)
Ten,Hwy 13
April 10
April 14
April 15
April 16
April 17
April 18
April 19
April 20
April 21
April 22
April 23
April 24
April 25
April 26
April 27
April 28
April 29
April 30

T-10-1 75.0 63.3 (59)
T-10-2 56.0
(200)
T-15-1 113 235 (227)
T-15-2 219
T-16-1 103 160
(D)
T-16-2 102
(I)
T-17-1 70.5 69
T-18-1 91.0 123
T-19-1 73.0 104
(98)
(92)
T-22-2 69.0 86
T-23-1 69.5 41 (90)
(86)
(82)
(78)
(74)
T-28-1 102 71
T-29-1 85.5 120
T-30-1 76.0 64
* Ice Obstruction
Total Load
Stage Discharge (Metric Tons
(cm) (m3/sec) Per Day)

(112)*
46*
107*
58
53
39
(36)
(33)
29
32
(25)
(28)
(20)
(18)
23
37
36


(3.40)
3.96
2.69
2.53
2.17
1.13
(0.91)
(0.79)
0.66
0.75
(0.54)
(0.62)
(0.39)
(0.31)
0.46
0.97
0.91(1.04)


58.7
77.7
37.2
15.1
23.0
10.2
7.6
6.3
4.9
5.8
i
4.0
4.4
i
2.6
(
2.0
2.8 1
i
10.0
5.0

(continued)

                                   117

-------
TABLE A-9.  (continued)
Water
Date Sample
Description
May 1 T-l-1
(D)
May 2
[May 3
May 4
May 5 T-5-1
(D)
JMay 6
i
May 7
Wagner, Hwy 13
April 15 W-15-1
W-15-2
JApril 22 W-22-2
j April 28 W-28-1
I
JMay 1 W-l-1
i (D)
f
May 5 W-5-1
(D)
iWagner, Mouth
|
April 22 W-22-1
Hanson, Mouth
April 10 H-10-1
Suspended
Turbidity Solids
(NTU) (mg/A)
65.0 (55)
(50)
(45)
(40)
70.0 (90)
(60)
(55)
63.0 70.7
56.7
53.5
135 143
61.0
60.0
61.5 75
70.0 53.7
Total Load
Stage Discharge (Metric Tons
(cm) (m3/sec) Per Day)
17 0.28 1.3
(15) (0.25) 1.1
(15) (0.25) 0.97
(15) (0.25) 0.86
14 0.20 1.6
(14) (0.21) 1.1
(13) (0.18) 0.87






         118

-------
              TABLE A-10.   SOUTH SHORE EROSION DATA (1975) FOR
                           MIDDLE HWY 13, AND MIDDLE MOUTH
Date
Water
Sample
Description
Suspended
Turbidity Solids
(NTU) (mg/£)
Stage
(cm)
Total Load
Discharge (Metric Tons
(m^/sec) Per Day)
Middle, Hwv 13
April 14
April 15
April 16
April 17
April 18
April 19
April 20
April 21
April 22
April 23
April 24
April 25
April 26
April 27
April 28
April 29
April 30

M-15-1
M-15-2
M-16-1
M-17-2

M-19-1

M-21-1
M-22-1
M-23-1




M-28-1
M-28-2
(ditch
near
river)
M-29-1
M-30-1
(80)
55.0 179 (178)
177
28.0 (80)
37.5 96
(90)
26.5 83
(65)
20.5 47
16.5 55
21.5 55
(44)
(34)
(26)
(16)
16.0 7 (16)
245 486
20.5 5 (65)
20.5 29
99
107
107
114
130
127

119
122
135
(127)
(122)
(112)
(107)
102
109
113
11.9
14.4
14.4
18.0
24.6
24.0
23.5
22.9
22.3
25.8
(24.0)
(22.3)
(16.7)
(14.4)
12.6
15.4
17.3
82.2
221
99.4
149
192
172
132
93.2
106
122
91.3
65.6
37.5
19.9
17.4
86.7
43.4
(continued)
                                    119

-------
TABLE A-10. (continued)
Date
Water
Sample Turbidity
Description (NTU)
May 1
May 2 M-2-1 10.0
May 3
May 4
May 5 M-5-1 7.5
May 6
May 7
Middle, Mouth
Suspended
Solids Stage
(mg/£) (cm)
(22) 102
15 93
(13)
(11)
(9) 76
(9)
(9) 69
April 8 M-8-1 20.5 20.0 *
M-8-2 19.7
April 11 M-ll-1 40.5 *
April 16 M-16-10 40.0 *
April 18 M-18-1 85.0 279 *
(D)
M-18-2 89.5
(D
* Because of the depth and current at this station,
not be taken.
Total Load
Discharge (Metric Tons
(m3/sec) Per Day)
12.6 23.9
10.3 13.4
(9.74) 10.9
(8.01) 7.62
6.85 5.32
(6.17) 4.80
5.52 4.29
|
1
a bottom profile could
         120

-------
             TABLE A-ll.  SOUTH SHORE EROSION DATA (1975)FOR
                            POPLAR HWY 13 AND POPLAR MOUTH
Water
Date Sample
Description
Poplar, Hwy
April 10
April 14
April 15
April 16
April 17
April 18
April 19
(April 20
JApril 21
April 22
April 23
April 24
April 25
April 26
April 27
April 28
April 29
April 30
May 1
13
P-10-1
P-10-2
P-14-2
P-14-3
P-14-4
P-15-1
P-15-2

P-17-2

P-19-1

P-21-1
P-22-2
P-23-2
P-24-2



P-28-1
P-29-1


Suspended
Turbidity Solids
(NTU) (mg/A)
12.0 6.9 (4.7)
2.5
40.0
71.3 (73)
76.0
76.5 159 (163)
167
(108)
30.5 52
(61)
20.5 70
(61)
16.0 51
13.5 34
15.0 34
9.4 12
(10)
(8)
(6)
16.0 3 (5)
16.0 10 (45)
(40)
(35)
Total Load
Stage Discharge (Metric Tons
(cm) (m3/sec) Per Day)
(25)
(51)
Ice
Blockage
(64)
69
71
75
80

83
75

77



64
70


(0.51)
(3.26)
(4.56)
5.17
5.47
5.86
6.40
6.54
6.68
5.86
(8.55)
6.12
(5.75)
(5.35)
(4.98)
4.62
5.30
(4.90)
(4.53)
0.21
20.5
64.2
48.2
24.6
30.9
38.7
34.4
29.4
17.2
25.1
6.34
4.97
3.69
2.58
1.99
20.6
16.9
13.7
(continued)
                                    121

-------
                          TABLE A-ll. (continued)
                                    -2-
Water
Date Sample Turbidity
Description (NTU)
May 2 P-2-1 8.9
May 3
May 4
i
May 5 P-5-1 7.0
May 6
i
May 7
Poplar, Mouth
Suspended
Solids Stage
(mg/£) (cm)
(15)
(13)
(ID
(9) 48
(9)
(9) 33

Discharge
(m-Vsec)
(4.13)
(3.74
(3.37)
2.97
(2.15
1.35

Total Load
(Metric Tons
Per Day
5.35
4.20
3.20
2.31
1.67
1.05

April 14 P-14-1
P-14-2
,April 15
i
April 16 P-16-1
April 17 P-17-1
April 18 P-18-1
i
April 19
April 20
April 21
April 22 P-22-1
April 23 P-23-1
April 24 P-24-1
April 25
April 26
April 27
60.0 92.7 (88)
83.7 ,
(119)
50.0 151
56.0 85 (158)
65.5 165
(184)
(160)
(134)
32.0 87 142
35.0 91 165
30.5 32 142
(26)
(21)
(16)
(5.49)
(7.73)
(8.75)
(9.26)
(9.94)
(10.8)
(11.1)
(11.3)
12.0
17.6
12.0
(9.74)
(9.06)
(8.44)
41.7
79.5
114
126
142
172
153
131
90.3
138
33.2
12.9
16.4
11.6
(continued)
                                    122

-------
TABLE A-ll.  (continued)
           -3-
Water
Date Sample Turbidity
Description (NTU)
April 28
April 29
April 30
May 1 P-l-1 25.0
May 2
May 3
May 4
May 5 P-5-1 16.5
May 6
May 7
Suspended
Solids Stage
(mg/£) (cm)
(13)
(118)
(105)
(40) 122
(39)
(34)
(29)
(23) 116
(23)
(23)
Discharge
(nrVsec)
(7.82)
(8.95)
(8.30)
6.61
(7.05)
(6.31)
(5.72)
4.93
(3.65)
(2.29)
Total Load
(Metric Tons
Per Day)
8.78
91.2
75.3
24
23.8
18.6
14.3
9.79
7.26
4.56
         123

-------
             TABLE A-12.  SOUTH SHORE EROSION DATA (1975) FOR
                            BARDON HWY 13 AND BARDON MOUTH
Water
Date Sample
Description
Bardon, Hwy
April 14
April 15
April 16
April 17
April 18
April 19
April 20
April 21
April 22
April 23
April 24
April 25
April 26
April 27
April 28
April 29
April 30
May 1
May 2
13
B-14-1
B-14-2
B-15-1
B-15-2
B-16-1


B-19-1

B-21-1

B-23-1
B-24-1



B-28-1
B-29-1

B-l-1
(D)
B-2-1
(D)
Suspended
Turbidity Solids Stage
(NTU) (mg/X,) (cm)
50.5 92.3 (92.6)
92.0
93.5 272 (275)
279
80.0 159
(131) 64
(103) 71
51.5 75 60
(55.5)
36.5 36 36
(38.5)
44.5 41 48
36.0 41 33
(40)
(40)
(40)
44.5 16 (60) 58
64.5 10 (100) 41
(80)
45.0 (60) 24
39.5 (54) 19
Total Load
Discharge (Metric Tons
(nrVsec) Per Day)
(0.71)
(1.50)
(2.29)
3.12
5.77
3.96
(2.52)
1.08
(1.64)
2.21
0.63 (0.96)
(0.85)
(0.74)
(0.62)
3.77
1.35
(0.99)
0.62
0.45
5.66
35.7
31.5
35.3
51.3
25.7
12.1
3.3
5.46
7.82
3.41
2.94
2.54
2.15
19.5
11.7
6.9
3.2
1.8
(continued)


                                    124

-------
TABLE A-12.  (continued)
Date
May 3
May 4
May 5
May 6
May 7
Water
Sample
Description


B-5-1


Suspended Total Load
Turbidity Solids Stage Discharge (Metric Tons
(NTU) (mg/£) (cm) (m3/sec) Per Day)
(48) (0.41) 1.5
(42) (0.37) 1.3
36.0 (36) 15 0.35 (0.33) 1.0
(30) (0.23) 0.68
(25) 9 0.14 0.3
Bardon, Mouth
April 9
April 14
April 16
April 17
April 18
April 21
April 23
April 24
May 1
B-9-1
(D)
B-9-2
(I)
B-14-3
B-14-4
B-16-2
B-17-1
(D)
B-17-2
(D
B-18-1
B-21-2
B-23-2
B-24-2
B-l-2
(D)
39.5 35.7 (31) Ice 0.21
Block-
36.0 27.0 age
102 186 (185)
105 184
108 327
115 201
105
130 311
60.0 73
70.0 91
51.0 35 36 1.20
60.5 29 0.74
        125

-------
            TABLE A-13.  SOUTH SHORE EROSION DATA  (1975) FOR
                          PEARSON HWY  13 AND PEARSON MOUTH
Water
Date Sample
Description
Suspended Total Load
Turbidity Solids Stage Discharge (Metric Tons
(NTU) (mg/Jl) (cm) (m3/sec) Per Day)
Pearson, Hwy 13
April 14

April 15
April 16
April 17
April 18

April 19
April 20
April 21
April 22
April 23
April 24
April 25
April 26
April 27
April 28
April 29
April 30
May 1

May 2

Pe-14-1
Pe-14-2



Pe-18-3
Pe-18-4
Pe-19-1

Pe-21-2

Pe-23-1

Pe-25-1


Pe-28-1
Pe-29-1

Pe-1-1
(D)
Pe-2-1
(D)
41.0 49.3 (51)
52.7
(105)
(158)
(210)
89.5 (200)
89.5
85.0 220
(140)
44.0 39 (50)
(50)
45.5 68
(65)
61
(58)
(54)
95.0 119
71.5 81
(75)
55.0 70

44.5 (64)

20

(30)
(41)
(53)
53

51
(42)
33
(28)
38
(36)
(33)
(28)
(23)
28
38
(30)
22

17

0.15

(0.40)
(0.76)
(1.33)
1.33

1.19
(0.82)
0.48
(0.33)
0.67
(0.58)
(0.48)
(0.33)
(0.19)
0.33
0.67
(0.40)
0.17

0.12

0.67

3.61
10.45
24.1
22.9

22.7
9.85
2.07
1.40
3.95
3.26
2.53
1.63
0.90
3.34
4.68
2.58
1.05

0.64

(continued)
                                    126

-------
                            TABLE A-13.  (continued)
                                       -2-
Date
May 3
May 4
i
JMay 5
May 6
May 7
Pearson,
April 9
April 14
April 15
April 16
April 17
April 18
April 19
April 20
April 21
April 22
April 23
Water
Sample
Description

Pe-5-1


Mouth
Pe-9-1
(I)
Pe-14-2
(D)
Pe-14-3
(I)
Pe-14-4

Pe-16-1
(D)
Pe-16-2
Pe-18-1
Pe-18-2
Pe-18-3


Pe-21-1


Suspended
Turbidity Solids Stage
(NTU) (mg/£) (cm)
(60) (17)
(55) (16)
42.0 (50) 15
(50) (13)
(50) 10

39.5 37.3
74.0
74.0 126 (123)
120
(183)
96.5 238 (243)
249
(400) 86
115 397 85
115
152
(430)
(280)
74.5 81 48
(81)
(113)
Discharge
OrVsec)
(0.11)
(0.10)
0.094
0.072
0.054

0
(0.49)
(1.26)
(24.1)
4.19
4.19
(3.79)
(2.58)
1.53
(1.03)
(2.13)
Total Load
(Metric Tons
Per Day)
0.57
0.48
0.40
0.31
0.23

0
5.21
i
|
19.9
50.5
i
i
145
i
144
141
62
10.7
7.19
20.8
(continued)
                                    127

-------
TABLE A-13.   (continued)
             -3-
Water
Date Sample
Description
April 24
April 25
;April 26
Upril 27
;April 28
April 29
April 30
May 1
SMay 2 Pe-2-2
t
May 3
1
May 4
;May 5
May 6
JMay 7
Suspended
Turbidity Solids Stage
(NTU) (mg/*.) (cm)
(147)
(138)
(131)
(122)
(270)
(184)
(170)
(159)
59.0 (145) 38
(136)
(125)
(114)
(114)
(114)
Total Load
Discharge (Metric Tons
(m3/sec) Per Day)
(1.84)
(1.52)
(1.03)
(0.61)
(1.03)
(2.12)
(1.26)
(0.55)
0.38
(0.35)
(0.32)
(0.30)
(0.23)
(0.17)
2.34
18.2
11.7
6.47
24.0
33.7
18.5
7.55
4.73
4.09
3.49
2.93
2.26
1.65
— — — —— — — — —
                  , U.S. GOVERNMENT PRINTING OFFICE:  1979-852-559/64
              128

-------
                                    TECHNICAL REPORT DATA
                            (Please read Instructions on the reverse before completing}
 1. REPORT NO.
  EPA-905/9-79-004
              3. RECIPIENT'S ACCESSION NO.
 4. TITLE AND SUBTITLE
  Red Clay Turbidity and Its Transport
    in Lake Superior
              5. REPORT DATE
                     January 1979
              6. PERFORMING ORGANIZATION CODE
 7. AUTHOR(S)
  Michael Sydor, Richard T.  Clapper,  Gordon J. Oman,
    and Kirby R. Stortz
              8. PERFORMING ORGANIZATION REPORT NO.
 9. PERFORMING ORGANIZATION NAME AND ADDRESS
  Physics Department
  University of Minnesota, Duluth
  Duluth, Minnesota  55812
              10. PROGRAM ELEMENT NO.
                     2BA645
              11. CONTRACT/GRANT NO.

                     R005175-01
 12. SPONSORING AGENCY NAME AND ADDRESS
  Great Lakes National Program Office
  U.S. Environmental Protection Agency,  Region V
  536 South Clark Street
  Chicago, Illinois  60605
              13. TYPE OF REPORT AND PERIOD COVERED
                     Final
              14. SPONSORING AGENCY CODE
 15. SUPPLEMENTARY NOTES
  Undertaken in support of  the Upper Lakes Reference Group studies  of pollution
  in Lake Superior and Lake Huron.
 6. ABSTRACT            '	—	—_^^______^___
       Red clay plumes in western Lake Superior are studied using  Landsat satellite
  imagery to determine the relative magnitude of the three sources of  the observed
  turbidity: erosion of the Wisconsin south shore red clay banks,  resuspension of
  bottom sediments, and runoff  from the many streams which flow  through the red clay
  belt and then into the lake.  A comprehensive sampling program was conducted during
  the spring of 1975 to determine the runoff contribution to the total load observed
  in the lake.  Analysis of Landsat transparency data coupled with weather records
  enabled contributions from erosion and resuspension to be separated.   It was found
  that approximately 75% of the observed load in the lake during the ice-free season,
  from May to November, is from erosion, 20% is from resuspension,  and 5% is from runoff
       A numerical model for water transports in Lake Superior as  a function of winds
  is developed.  This model is  verified by comparison of observed  and  predicted water
  levels at  several locations around the lake, and by comparison of the predicted
  transport  patterns to actual  turbidity distributions observed  in Landsat imagery.
 Transport  patterns are shown  for western Lake Superior and the entire lake for both
 an easterly and westerly wind.   A model of current profile with  depth is also devel-
 oped.   The results of the transport model are used to predict  distributions of red
 clay from  the south shore and taconite tailings discharged into  the  lake at
 Silver Bay, Minnesota.
 7.
                                KEY WORDS AND DOCUMENT ANALYSIS
                  DESCRIPTORS
                                              b.lDENTIFIERS/OPEN ENDED TERMS
                           c. COSATI Field/Group
 Turbidity
 Sediment  transport
 Erosion
 Runoff
 Water quality
 Hydrodynamics
 Remote sensing
 Red clay, Landsat
 Lake Superior, Numerical
   transport models
 Upper Lakes Reference
   Group
 International Joint
   Commission
   DISTRIBUTION STATEMEN1


 RELEASE TO  PUBLIC
19. SECURITY CLASS (ThisReport)'
 UNCLASSIFIED
21. NO. OF PAGES
     139
20. SECURITY CLASS (This page)
 UNCLASSIFIED
                                                                        22. PRICE
EPA Form 2220-1 (R«v. 4-77)   PREVIOUS EDITION is OBSOLETE

-------
                                                        INSTRUCTIONS

   1.   REPORT NUMBER
        Insert the EPA report number as it appears on the cover of the publication.

   2.   LEAVE BLANK

   3.   RECIPIENTS ACCESSION NUMBER
        Reserved for use by each report recipient.

   4.   TITLE AND SUBTITLE
        Title should indicate clearly and briefly the subject coverage of the report, and be displayed prominently. Set subtitle, if used, in smaller
        type or otherwise subordinate it to main title. When a report is prepared in more than one volume, repeat the primary title, add volume
        number and include subtitle for the specific title.

   5.   REPORT DATE
        Each report shall carry a date indicating at least month and year.  Indicate the basis on which it was selected (e.g., date of issue, date of
        approval, date of preparation, etc.}.

   6.   PERFORMING ORGANIZATION CODE
        Leave blank.

   7.   AUTHOR (S)
        Give name(s) in conventional order (John R. Doe, J. Robert Doe, etc.). List author's affiliation if it differs from the performing organi-
        zation.

   8.   PERFORMING ORGANIZATION REPORT NUMBER
        Insert if performing organization wishes to assign this number.

   9.   PERFORMING ORGANIZATION NAME AND ADDRESS
        Give name, street, city, state, and ZIP code.  List  no more than two levels of an organizational hirearchy.

   10.  PROGRAM ELEMENT NUMBER
        Use the program element number under which the report was prepared. Subordinate numbers may be included in parentheses.

   11.  CONTRACT/G RANT NUMBE R
        Insert contract or grant number under which report was  prepared.

   12.  SPONSORING AGENCY NAME AND ADDRESS
        Include ZIP code.

   13.  TYPE OF REPORT AND PERIOD COVERED
        Indicate interim final, etc., and if applicable, dates covered.

   14.  SPONSORING AGENCY CODE
        Insert appropriate code.

   15.  SUPPLEMENTARY NOTES
        Enter information not included elsewhere but useful, such  as:  Prepared in cooperation with, Translation of, Presented'at conference of,
        To be published in, Supersedes, Supplements, etc.

   16.  ABSTRACT
        Include a brief (200 words or less) factual summary of the  most significant information contained in the report.  If the report contains a
        significant bibliography or literature survey, mention it here.

   17.  KEY WORDS AND DOCUMENT ANALYSIS
        (a) DESCRIPTORS - Select from the Thesaurus of Engineering and Scientific Terms the proper authorized terms that identify the major
        concept of the research and are sufficiently specific and precise to be used as index entries for cataloging.

        (b) IDENTIFIERS AND  OPEN-ENDED TERMS - Use identifiers for project names, code names, equipment designators,  etc. Use open-
        ended terms written in descriptor form for those subjects for which no descriptor exists.

        (c) COSATI FIELD GROUP - Field and group assignments are to  be taken from the 1965 COSATI Subject Category List. Since the ma-
       jority of documents are multidisciplinary in nature, the Primary Field/Group assignment(s) will be specific discipline, area of human
        endeavor, or type of physical object. The application(s)  will be cross-referenced with secondary Field/Group assignments that will follow
        the primary posting(s).

   18.  DISTRIBUTION STATEMENT
        Denote reusability to the public or limitation for reasons other than security for example "Release Unlimited." Cite any availability to
        the public, with address and price.

   19. & 20.  SECURITY CLASSIFICATION
        DO NOT submit classified reports to the National Technical Information service.

   21.  NUMBER OF PAGES
        Insert the total number of pages, including this one and unnumbered pages, but exclude distribution list, if any.

   22.  PRICE
        Insert the price set by the National Technical  Information Service or the Government Printing Office, if known.
EPA Form 2220-1  (Rev. 4-77) (Reverse)

-------