905R76102
00623-
7604
INTERNATIONAL JOINT COMMISSION
MENOMONEE RIVER
PILOT WATERSHED STUDY
SEMI-ANNUAL REPORT
CSOPERAT I TG ACENC ! ES
WfSCOHSfM DEPARTMENT OF
NATURAL RESOURCES
JOHN s, KC-VAQ
UNIVERSITY OF W.'SCGrtSiM SYSTEM
WATcR RFSO-kCi'S CEMER
SO'JTHEASFf.RN WS^CCWSiN REGIONAL
Sponsored by
fNTERNATfONAL JO^!T COMMISSION
POLLUTION CRC(V1 LAiJO USE
ACTIVITIES REFERENCE GROUP
UNITED STATES ENVIRONMENTAL
PROTECTION AGENCY
APRIL 1976
-------
U.S. Environmental Protection Agency
GLNPO Library Collection (PL-12J)
77 West Jackson Boulevard,
Chicago, IL 60604-3590
-------
TABLE OF CONTENTS
Page No.
SUMMARY SEMI-ANNUAL REPORT 1
APPENDIX
A. River Monitoring Activities 9
B. Menomonee River Data Base System M-8
C. Land Data Management System 59
D. Specific Land Use Studies 63
E. Biological Studies 67
F. Atmospheric Monitoring Program 71
G. Remote Sensing Program 83
H. Land Use-Water Quality Modeling 95
-------
SUMMARY - SEMI-ANNUAL REPORT
Introduction
The International Joint Commission, through the Great Lakes Water
Quality Board, established the International Reference Group on Great
Lakes Pollution from Land Use Activities (PLUARG) to study and report the
effects of land use on water quality. The "Task C" assignment requires
the detailed investigation of six major watersheds in Canada and the
United States, which are representative of the full range of urban and
rural land uses found in the Great Lakes basin. The Menomonee River
watershed is serving as the focus of investigation on the impact of urban
land uses on water quality and as the site for studying the effects of
rapidly changing land use patterns in an urban setting. The data will be
used to extrapolate the effects of urban land use on the water quality
of the Great Lakes.
The specific objectives of the Menomonee River Pilot Watershed Study
are:
1. To determine the levels and quantities of major and trace con-
stituents including, but not limited to, nutrients, pesticides
and sediments reaching or moving in flow systems likely to reach
Lake Michigan.
2. To define the sources and evaluate the behavior of pollutants
for a metropolitan complex with particular emphasis on the impact
of residential and industrial, including utility facilities,
transportational, recreational, agricultural and constructional
activities associated with rapid urbanization.
3. To develop the predictive capability necessary to facilitate
extension of the findings from the Menomonee River Watershed
Study to other urban settings, leading to an eventual goal of
integrating pollution inputs from urban sources to the entire
Great Lakes basin.
This report will review the progress toward achieving the specific
objectives of the study since the October 1975 semi-annual report.
-------
Progress
The river monitoring activities continued to accomplish the goals of
the first objective. The field activities included baseline river surveys
and monitoring of runoff events. Regular baseline river surveys for
nutrients, total alkalinity, hardness, chloride, color, total dissolved
solids, suspended solids, suspended volatile solids and total organic
carbon were conducted once a week at the 12 automatic stations and 2 grab
stations. Regular baseline river survey sampling frequency increased to
twice a week during spring melt and will decrease to twice a month begin-
ning in April. Preliminary plots of baseline, total P and (nitrate +
nitrite)-N concentrations from the 1975 field year indicated increasing
concentrations during periods of high flow and many of the remaining
values were within a relatively narrow range. Further interpretation
awaits statistical and graphical analysis of all the baseline data.
Simultaneous grab samples and comparison samples were obtained during the
regular baseline river surveys for quality control purposes. In general,
the quality control data indicated that the sampling program was providing
accurate data. Data from the September 3, 1975 baseline quarterly metal
survey and September 9, 1975 baseline quarterly organic compound survey
were compiled. The metal concentrations tend to increase below the
Underwood Creek station (413007) which was the same trend observed for
the June 25, 1975 metal survey. As in the June 9, 1975 organic survey,
most of the organic concentrations for the June 11, 1975 organic survey
were below detection limits. Biweekly baseline water sampling continued
at three bridge sites located in the estuary area of the river. Con-
tinuous i,n situ electronic monitoring of temperature, dissolved oxygen,
conductivity and pH continued at five of the automatic stations. The
United States Geological Survey (USGS) continued monitoring discharge
at 11 automatic stations, precipitation at 7 automatic stations and
sediment concentration at all the automatic stations. The 1975 water-
year discharge records were completed, and the sediment records from 1975
are almost completed. Water quality surveys were conducted at the
Germantown waste treatment facility and at the City of Menomonee Falls—
Parkside and Parkview waste treatment facilities. Biological monitoring
-------
was begun at 5 automatic stations, and samples will be analyzed for
macroinvertebrates and periphyton community composition. Biological
samples are collected on six suspended, modified Hester-Dendy samplers
made of Masonite and Conservation Webbing. Besides the water quality
river surveys, the yearly macrobenthic survey was undertaken on
November 3, 4 and 5, 1975.
Collection of water quality samples during runoff events continued
as an important part of the river monitoring activities. Since October
1975, six runoff events were sampled for regular baseline river survey
parameters (nutrient, etc.) at four stations (413005, 413010, 413011 and
463001). Event data that have already been summarized indicated that the
concentration of total P, total Kjeldahl N, total and suspended solids
usually increased during a runoff event. The change in nutrients and
solids concentrations coincided, most often, with changes in the hydrograph
and the concentrations were substantially higher than baseline values.
Metal concentrations also increased during runoff events and reached
concentrations substantially higher than those of the baseline quarterly
survey samples. The events on August 20, 1975 and November 29 through
December 1, 1975 have been monitored for metals. In contrast to concen-
tration trends for metals and nutrients during events, the concentration
of organic components measured during the November 20-21, 1975 event did
not increase and were below the limits of detection for the methods used.
Quarterly monitoring for organic components and metals will continue during
runoff events. The stations chosen for event sampling will be rotated
to obtain seasonal runoff data of all the parameters at each automatic
station as an essential part of the river monitoring objectives. The
baseline and event concentrations are being combined with the appropriate
discharge data to provide loading values. Preliminary loading data
indicated that the loading of many of the parameters during an event
was a significant fraction of the total baseline loading for the entire
month. Data from all the sampling activities have been edited and stored
on computer mass storage to allow for reporting and statistical analysis.
The stored data have been programmed for mechanical plotting in terms of
concentration and loading.
-------
The two principal approaches used to achieve the goals of objective 2
are providing an inventory of land use characteristics and investigating
the impact of various land uses on water quality. A Land Data Management
System (Land DMS) was designed to store, retrieve, analyze and display
land data—in tabular or graphic form. Nine data types have been coded
for the entire watershed, and the coding of two additional data types is
in progress. Land DMS work elements completed or in progress since
October 1975 include: 1) a computer program was obtained from the
National Geodetic Survey to convert cell corner coordinates from the
Wisconsin State Plane Coordinate System to the Universal Transverse
Mercator System and incorporating the program into the data management
phase of the Land DMS; 2) coding of soils data for the watershed was com-
pleted and the coding of ground elevation data and land use data by
individual cells was initiated, 3) work was started on the software addi-
tions needed to determine the characteristics of the total area tributary
to each of the monitoring stations established in the watershed; and
4) the Land DMS was used to prepare a tabular summary of soil types for
each of the 244 sub-basins in the watershed.
The remote sensing project is also providing land data in support of
the second objective. The remote sensing project is developing techniques
for mapping land cover and hydrologically active source areas in the
watershed from small scale color-IR imagery. The techniques developed
will be used to determine the applicability of remote sensing data in
hydrological modeling efforts. Photographic missions utilizing color and
color infrared imagery have been flown approximately once per month during
the summer and fall months of 1975. Thus far, two scenes near Menomonee
Falls in Waukesha County, dated July 16 and October 6, 1975» have been
o
digitized in the wavelength intervals, 4456 to 4550 A (red), 5450 to
o o
5550 A (blue), and 6450 to 6550 A (green), using a 100 micron spot size
representing a pixel size of 7.5 meters on a side. To date, 29 signatures
have been developed for the July 16 data representing 5 categories of
land cover, and 39 signatures representing 6 categories of land cover have
been developed for the October 6 data. Land cover maps were developed,
and the output displayed on conventional paper printouts as well as on a
color cathode ray tube system.
-------
The specific land use studies are being used to determine the impact
of various homogeneous and/or predominant land use areas on water quality
in support of objective 2. Sampling stations at eight of the study sites,
representing major land uses in the watershed, have been completed.
Preliminary grab sampling of snowmelt runoff was started March 18, 1976
at selected specific land use study sites to scan for the predominant
Group C parameters.
An air and precipitation monitoring program is being established
presently in the watershed in support of objective 2. The monitoring
system was designed to study the influence of atmospheric processes in
the cycling of elements in the watershed. A primary objective of the
program is to determine the relative contribution of different land use
practices to the total weight and chemical composition of the aerosols
measured over the watershed. The atmospheric study will quantify the mass
of aerosols entering the watershed from rain and dry fallout deposition.
To date, the aerosol collection methodology has been evaluated. The rain
and aerosol collectors are presently being installed at seven of the
automatic river stations.
Objective 3 is being achieved by an effort to develop land use-water
quality models. Modeling techniques and mathematical models applicable to
the land use-water quality models have been reviewed and evaluated for
possible use. The digital computer model selected as the basic simulation
tool for the Menomonee River Pilot Watershed Study is based on a hydrologic-
hydraulic model that originated at Stanford University and is now available
with a water quality feature from the consulting firm of Hydrocomp, Inc.
This selection was based on considerations—amongst others—of the
Hydrocomp Model, the Wisconsin Hydrologic Transport Model (WHTM), the
EPA Storm Water Management Model (SWMM), the Corps of Engineers Storage,
Treatment and Overflow Model (STORM) and working experience with the
Hydrocomp Model, SWMM, STORM and a new model called LANDRUN. The South-
eastern Wisconsin Regional Planning Commission (SEWRPC), in their coopera-
tive effort with the Menomonee River Project, and including coordination
with their programs supported by Section 208 funds under Public Law 92-500,
is evaluating the Hydrocomp Model. Data base efforts for the model have
-------
concentrated recently on completing channel data and developing diffuse
source data and point source data. Since the October 1975 semi-annual
report, a successful calibration of the hydrologic submodel and hydraulic
submodel was accomplished, and continued calibration efforts have concen-
trated primarily on achieving a preliminary calibration of the water
quality submodel.
The model, LANDRUN, is being developed to supplement the Hydrocomp
Model in areas where it does not meet project objectives and as a model
that will achieve independently most of the goals of objective 3. The
LANDRUN model represents a dynamic hydrologic transport model which trans-
forms precipitation into surface runoff, interflow, and groundwater
aquifer recharge quantity and quality. Most of the model parameters and
some inputs are related to land use within the modeled watersheds. A
soil adsorption model applicable to phosphorous, pesticide and heavy metals
transport was developed, tested and is being incorporated as a subroutine
into "LANDRUN."
Development is proceeding on a less complex model that would relate
empirically runoff quality to land use, climate, and hydrologic and other
factors which contribute to runoff quality. This model assumes that the
concentration of a chemical in surface runoff varies temporally about a
mean runoff concentration in a predictable manner. Runoff water quality
has been observed at three small watersheds in Milwaukee during the course
of 20 events. Interpretation of mean concentrations of the various
parameters during runoff events began in December 1975. Plots of relative
concentrations as a function of a time ratio are presently being generated
by computer for each event.
In summary, the Menomonee River Pilot Watershed Study was designed
to investigate the impact of urban and urbanizing land use on water
quality and to extrapolate these effects to the entire Great Lakes basin.
The overall goals of the study were divided into the three aforementioned
specific objectives. Since October 1975, the river monitoring activities
in support of the first objective have been continued with an emphasis
on runoff event monitoring. The projects designed to accomplish the
second objective have progressed on schedule. The progress included:
-------
1) continued coding of land types into the Land Data Management System,
2) development of land cover maps using remote sensing, 3) completion
of construction at eight specific land use sites and initiation of
sampling, and 4) testing of air samplers presently being installed for the
atmospheric monitoring project. The modeling activities in support of
objective 3 have been defined more clearly with the choice of the Hydrocomp
Model as the basic simulation tool for the study, and the continued develop-
ment of "LANDRUN" as an independent model supplemental to the Hydrocomp
Model.
-------
-------
APPENDIX A
-------
RIVER MONITORING ACTIVITIES
Introduction
The objective of the river monitoring activities is to determine the
types and quantities of various water quality parameters in the waters of
the Menomonee River and its principal tributaries. Parameters of concern
include the "core" list established by the Task C Technical Committee of
the Pollution from Land Use Activities Reference Group (PLUARG) and other
parameters likely to affect Lake Michigan water quality. The river
monitoring activities will provide basic information about the hydrology,
hydraulics and water quality of the watershed. Interpretation and assess-
ment of monitoring data will be undertaken by the development of land use-
water quality models.
Preliminary monitoring of the river began in February, 1973 with the
establishment of three stations by the Wisconsin Department of Natural
Resources (WDNR). Intensive sampling of the river and its tributaries
began in January, 1975 after site selection and construction of 12 auto-
matic monitoring stations.
Prepress
Field activities
The field activities included both baseline surveys and event sam-
pling. The baseline survey parameters included nutrients, total alkalinity,
hardness, chloride, color, total dissolved solids, suspended solids, total
volatile solids, and total organic carbon. Conductivity, pH, temperature
and dissolved oxygen were also measured for each sample immediately after
collection. Regular baseline river samples were obtained from each of
the 14 river sampling sites (12 automatic monitoring stations and 2 grab
sample sites (Appendix A Fig. 1)) once a week from October, 1975 to
April 1, 1976. A lack of flow from December 15, 1975 to March 22, 1976
excluded the Schoonmaker Creek station (site 403010) from the regular base-
line river surveys. Baseline river sampling frequency increased to twice
a week during the spring melts occurring between February 15 and March 6,
1976. Beginning in April, 1976, the "regular baseline river surveys will
-------
-------
67300!
413008
413010
68300!
413007
Scale
-------
11
be conducted biweekly except during periods of high flow. In addition
to the routine baseline survey samples, quality control samples were
obtained at one of the stations during each baseline survey. Quality
control samples consisted of two simultaneous grab samples and a com-
parison sample obtained with the automatic sampler. The sampling program
provided information on sample storage effects and river homogeneity and
compared sample collection methods.
Two baseline surveys for metals and organics were completed in
1975 (October, 1975 Semi-Annual Report). Four baseline surveys for
metals, organics and bacteria will be done in 1976.
Biweekly sampling was continued for surface, mid-depth, and bottom
waters at three bridge sites located in the estuary areas of the river.
The list of parameters for the estuary samples was the same as that for
the regular baseline river survey and also included temperature and dis-
solved oxygen profiles.
Monitoring of temperature, dissolved oxygen, conductivity and pH was
continued at the same 5 river stations. Equipment repairs and conversion
resulted in temporary shutdown of two stations. The 12Uth Street station
(683001) was out of order between November 19, 1975 to January 6, 1976,
and the River Lane Road station (673001) has been out of order since
October 13, 1975. A weekly calibration schedule for each site has been
in effect except for periods of high water which prevent the removal of
in situ equipment. Calibration was curtailed from February 12, 1976 to
March 11, 1976 due to spring ice breakup and high water levels. However,
all operating units continued to function properly during this period
and only a small portion of the data is questionable. The data are
presently being digitized.
Water quality surveys were also conducted at the Germantown waste
treatment facility on September 18 and 19, 1975 and at the Parkside and
Parkview waste treatment facilities of the City of Henomonee Falls on
June 19-20 and 23-245 1975. The measured parameters for the 24-hour
composite surveys included nutrients, metals, bacteria and organics.
The yearly macrobenthic survey was done on November 3 through 5,
1975 and consisted of 4 quantitative and 12 qualitative sampling surveys
-------
12
at 13 representative sites. Sorting and identification of species present
are still in progress at this time.
The United States Geological Survey (U.S.G.S.) monitored discharge
on a continuing basis at 11 river stations. Work on the 1975 water-year
discharge records was completed and the corrected flow data put on a
magnetic tape. The flow data are being programmed by the WDNR to be
compatible with the stored water quality data. Rain data were collected
by the U.S.G.S. at 8 rain gauging sites in the watershed. The U.S.G.S.
continues to monitor suspended sediment concentrations using the auto-
matic samplers at the 12 river monitoring stations. The sediment records
are about completed for the 1975 water year.
Collection of water quality samples during runoff events has continued
as an important part of the river monitoring activities. Event samples
were collected from 4 automatic river stations (stations 413005, 413010,
413011 and 463001) and were analyzed for most of the water quality
parameters. For event sampling, the automatic samplers were set to
collect water on a timed basis after the stage recorder reached a pre-
determined level. Insufficient runoff or automatic sampler malfunction
prevented collection of samples from some or all of the 4 stations during
some events. If possible, manual grab samples were obtained when an auto-
matic sampler was not operational. Event samples were collected for
analyses of: metals on August 20, 1975 at stations 413005 and 413010;
pesticides on November 20-21, 1975 at station 413005; and metals and
organics on November 29 through December 1, 1975 at station 413005 and
on February 12-16, 1976 (snowmelt) at stations 413005 and 413010.
Runoff event samples, analyzed for the same parameters as a regular
baseline river survey, were collected during seven events since August,
1975. These events occurred on: September 5, 1975 at station 413005;
February 12-16 (snowmelt) and 16-17, 1976 at station 413005 and station
413010; February 18, 1976 at station 413005; February 24-27, 1976 at all
four sites (snowmelt); and March 1-5 and 12, 1976 at all four sites.
Before September 5, 1975, event samples had been collected for four events
(October, 1975 Semi-Annual Report). Two additional event-sampling sta-
tions will be added by the end of April, 1976. Event sampling will be
-------
13
rotated among the 12 automatic sampling stations so that all stations
will be monitored during each season by the end of the 1977 field year.
Data from the sampling activities have been edited and stored on
computer mass storage to allow for reporting and statistical analyses.
The data may be accessed through batch processing or terminal input and
output. Interpretation of the water quality data is facilitated by
programming the stored information for mechanical plotting in terms of
concentrations and loadings. Appendix B contains more detailed informa-
tion concerning the structure of the data files and how the program
functions for storing and retrieving data.
Water quality data
The regular baseline river surveys (nutrients, etc.) and baseline
surveys for metals, organics and bacteria provided information on the
values of the various parameters between runoff events. Preliminary
graphs of concentration versus time for total P and (nitrate +• nitrite)-N
at 70th Street, Underwood Creek, Noyes Creek, and Donges Bay Road
(stations 413005, 413007, 413011, 463001, respectively) were plotted for
the entire 1975 field year. The plots include event data along with the
baseline survey data. Many of the values of total P fall within a
relatively narrow range for the entire year of baseline survey data;
this is illustrated for stations 413007, 413001 and 463001 by Appendix A
Figs. 2, 3, 4, and 5. Total P concentrations were generally higher at
station 413005 than at the other 3 stations, and the range of values was
larger. Concentrations during regular baseline river surveys for
station 413005 ranged from about 0.2 to 0.4 mg/1, while the concentra-
tions at the other 3 stations ranged from about 0.025 to 0.1 mg/1. The
location of station 413005 toward the lower end of the Menomonee River
probably accounts for the higher total P concentrations. Comparison of
total P concentrations outside the above ranges with available flow data
indicated the higher concentrations of total P were observed on days of
relatively high flow. An example was the relatively high values for total
P on June 24 and June 30, which were days of high flow related to runoff
events. These results corroborate the increase in the concentrations
-------
-------
m
X
X
X
X
X
X
x -
X
X
,
•w- ..
X
X
X
-j
3
»
3
C
x ,.
x -
X
X
X
X -T
J
0
1—
J
a
3K
X
x _fe
X f
W
r-
LO
t>
01
I— 1
M
G
rr-*
(-<
o
-i-1
ra
\T:
r-i
s
pi
>-,
(C
PQ
vO
^J,
c
-M
-o
X
X
Ac „ .
X
X
X
i
X
oo
II! 'If
o o
ID O
Ol CM
1 ) J
O
to
O
1)1 1 ! I
0 0
0 10
o> r^
111.111 !()
O D O
O ID O
tx> ^ 07
X
X
114 (11
0 O
t0 O
.-» o
3 -H,
j r^i
^ cx
X. •=£
r~
snaoqdsoqd
-------
15
4+
-t
4. _.
4 f
4
4
*
4
4
4
4
4 -H-4
-H*
4-f
VJ
r-
11
o
s
o
o
ED
O
IP
J-
8
(O
O
to
o
o
o
CD
H
g
•H
•P
nd
-P
CO
a>
0)
o
CO
-P
m
w
§
•H
•P
•H
C
CD
O
§
O
O
•a
w
o
-a
tU
-H
O
H
X
•H
T3
C
0)
a
a,
N.
to
r--
snaoqdsoqd
-------
16
4 -
4 4
4
4
4
4
4
m
x
CM
X,
113
r-
-9
cr>
H
W
G
•H
O
•H
(0
4-;
CO
o
fn
CJ
f.
ft
m
C;
P
C
$••
c
_1_J_
to
3 '±
IX)
snaoqdsoqd ieq.oj
-------
1?
©
e
e
0
e
©
D
0
0
to
e
0
0
LI)
r-
C7)
no
p
C
O
to
4-'
CO
e
•H
CO
e e
eo®0
e
e
e
8
0
oo
c
o
•H
0
0
0
o
c
8
0
0
I
OH
M
o
©
e
0
0
e
o
E-"
bO
•ri
x
•H
P
D
£
o
5
o
S
_L_L
II I
o
o
a>
o
1C
r-
o
o
(D
0
—I I i
I 1
in
O
1C
D
O
O
(I/3ui) snaoqdsoqd
-------
18
of total P observed during runoff events. The runoff events are indi-
cated on the plots by more than three points for a particular date
(Appendix A Figs. 3 and 5). Although no significant yearly seasonal
trends in total P values were observed within the usually small range
of values, some differences in the trends of total P concentrations were
noticed from one month to another. For example, the total P concentra-
tions in the high flow period at the end of March tended to be lower
than the concentrations during the high flow period at the end of June
for all stations.
The opposite was observed for (nitrate + nitrite)-N concentrations
(Appendix A Figs. 6, 7, 8, 9). The concentrations measured in March
were higher than for June and other summer runoff events (Appendix A
Figs. 7 and 9). The March concentrations were highest at station 463001
with values up to 6 mg/1, which were probably a result of the farming
activities in that drainage area. The high (nitrate + nitrite)-N concen-
trations continued through April for stations 413005 and 463001.
Contrary to the trends in total P values, the (nitrate + nitrite)-N
concentrations increased to similar levels again at all stations in late
November and for all of December. Except for several points, the re-
maining (nitrate + nitrite)-N values fell within a relatively narrow
range. Station 463001 values were highest with a range of about 1 to
2 mg/1, while station 413005 had a range of about 0.5 to 1 mg/1 and
stations 413007 and 413011 had a range of about 0.1 to 0.5 mg/1. The
above observations for total P and (nitrate + nitrite)-N are considered
preliminary and only include a small portion of the available regular
baseline river survey data.
A statistical analysis is presently underway for all the regular
baseline river survey parameters at all the stations to better identify
and characterize trends in the data. Available flow data will be pro-
grammed to identify more completely the various regular baseline river
monitoring data biased by antecedent runoff events. Appropriate flow
rates and baseline river survey data will be combined to provide estimates
of monthly mean loadings of the various parameters at each of the stations.
Quality control data obtained from replicate sampling during the
regular baseline river surveys indicated the sampling procedure had
-------
+
+
•Vr
•f
CO
X •
CM LT>
x en
ID H
C^
M
C
•H
C
O
•H
CO
T3
m
o
Pi
£
PQ
w
QJ
be
c
o
19
4-'
ro
K
O
•H
•H
05
0)
G
C
G
C
0!
4-
•f
is
bC
I
0
o
o
J — 1 — L_
o
o
I.... I I ....
o
o
o
I
1_LJ
o
o
to
o
§
I ! 1
o
o
10
a
O
a
! ,1 JL
a
p
j i |
a
o
a
a
o
1
l—i_J_J
a
o
a
y.
'V. >H
in id
0!
n.
-------
20
CO
•-V
CM
LO
r-
-3
>«
en
z:
(E
It-
•H
fn
^
•H
"O
c
0)
&,
-------
co
21
©
0
o
e
o
e
0
©
€
0
e
•s©i
\ en
ID H
r-
O
•H
-P
ro
-H
w
0)
0)
O
O
O
I
"d
c
D
w
C
c
a;
c
o
&
0O
e
jj>-
0
0
cr
j_j_
a
O
10
a
o
o
_LO_L
a
o
JL_i_i.
a
a
a
usSoaq.TU-(aq.Taq.Tu
• H
p.
4-J
X
•H
c
cu
A.
a
-------
22
CO
Cv
1 -„ IT.
I- CT
EJ
i i i
0
o
10
o
o
a
a
o
o
o
o
o
o
in
R
EE1
B
o
O G -ft
B >
B
s
j
B
B ~"
B
B
OEI
cSri'c'icnEggyi
u£LJ CJtUilJHU
S
RRrfn
El
an
B
B
*"""""
B
B e
,-,B
B
51 a
i
B BB"
B n
BB ~
B
B
B ~~
B
B ~
I ' III III 1_U J J '
V
a.
r:
a:
O-
cc
01
a:
in
ID
u! "
•^
*~*
ID
c-
o a p a a
a o 5 o 0
O U3 O U3 C5
g
rQ
0
4-J
ro
4-1
W
ij
0)
0)
CO
4-'
O
r--
-f-'
ro
K
c
c
• H
4-J
a
-H
CD
0
g
0
c
W)
0
4-1
•H
|
G)
•H
*C
0)
•p
•H
!S
^^
cr>
bb
•H
h
<
X
•H
-------
23
little effect on the accuracy of the water quality data. The results of
the simultaneous manual grab sampling program are presented in Appendix A
Table 1. Questionable replication between simultaneous grabs are
identified in the table. The criterion for inadequate replications was
a deviation of greater than 20% from the mean of the two simultaneous
grab samples. Only 14 of 192 analyses obtained for the simultaneous grab
samples, or 7%, did not meet the above criterion. Ten of the 14 ques-
tionable values appeared on June 30, 1975 and the first three dates of
August, 1975. Five of the 10 were on June 30, the first date simultaneous
grab samples were obtained. Field personnel have indicated that the
initial sampling technique probably accounts for the deviation in the
June 30 data. The effect of storage was probably minimal in the devia-
tions observed for other dates, since storage should have affected more
samples. Most of the questionable ammonia values occurred at low ammonia
levels (approaching limits of detection at 0.01 mg/1) where analytical
error may be as high as 20%. Contamination by organic matter weighing
2 or 4 mg could have caused the observed differences in suspended solids
values. Many of the results obtained on October 6 and 13, 1975
(Appendix A Table 1) with replicates from the automatic samplers demon-
strated large deviations. Since the automatic samplers cannot obtain
simultaneous replicate samples, the observed deviations were probably due
to changes in river composition during the sampling interval.
Due to possible changes in river composition during sampling, a
comparison sample from the automatic sampler and two manual grab samples
were taken simultaneously. The results of the comparison sampling are
given in Appendix A Table 2 and the questionable comparison analyses are
identified. The criterion used for questionable values in the comparison
sampling program was the same as used in the simultaneous grab sampling
program. Only 10 of the 120 comparison analyses, or 8%, did not meet
the criterion. In 7 of the 10 questionable analyses, the value for the
automatic sampler was higher than for the manual grab samples. Seven of
the 10 questionable analyses occurred on either November 11, 1975 at
Falk Corporation (station 413004) or on November 24, 1975 at Underwood
Creek (station 413007). Since most of the poor replicates occurred for
-------
Appendix A Table 1. Comparison of simultaneous duplicate grab samples
Station
Date
Solids
Phosphorus
Sol.
Total
Organic
KHj
,-».
HO.tNO,
X
Total
Alk.
Chi.
TOC
Cond.
413011
413008
413010
413007
413011
463001
413005
683002
413010
413005
413007
6SS001
»1?008
m^ovi
46-001
67sco;
68".GC2
41? 004
"•13036
»i:co'
413007
663005.
»13008
413011
6-30-75
6-30-75
6-30-7S
8-4-75
8-11-7S
8-18-7S
7-28 -75
8-25-75
9-2-75
9-1-75
3-15-75
S-22-75
9-i9-7S*«
10-6-75**
10--.S-75**.
iO-iO-"5**
10-27-75
11-3-75
11-10-75
11-17-75
11-24-75
1Z-1-7S
12-3-75
12-15-75
1020
1050
684
695
765
755
700
70S
530
535
740
775
715
7CS
538
S65
494
496
702
'."•,.,
57 f
!»',»
TOE
7-C
7<»
T'f.
7«<.
S'Q
-Sr-
e.z
MS
5.'.;
Si,:
:,"-5
^ ~c
* "' C-
,::(.
2rl?
;'«
S5»
1 '_. , S
J9s
set
no
734
111*
11J.J
28*
13
34
36
3
it
3
5
IS
23
7
7
15
li
64
62
0
0
li
13
n
3
1*
IS
15
16
t,
;-
•>r
^~
S,'1
15
-e«
i,'.
i-
/»
;
C
2
?"
'Jt.
,7
4*
i /
22
7
6
13
S
3
4
j
z
?*
7
,
2
c
ij
IP
51
-_•
2
-
,,,
;
u
6
w
':
K
"
1 *
'
t
'•
.^
it-
j
',
3
;
I.
i
i
'„
6
£
0.12«
0.07
0.19
0,15
0.09
C. G7
0.043.
C.OiS
c.os
0.03
O.OVi
o.ot?
,,.24
0.?'!
0.27
0.30
O.i?
0,59
,; , ;,
O.f/
0.05
o.o:
o.a?
0.93
j.o;;
0.0'' a
-. }•.
0 u'.
O.J1*
C."
0 - 3d
•J.JO
0. ' '.
r ; -
O./"1
0- *,"
C.1%
0. !3
>J. i 3
-, 7 ;,
'.C,'
1/.30
C, -.')
C . 04
0 OS
0.11
0.11
0.078*
0.027
O.OT4*
0-13
0.073
0.070
'-' 022*
•J 013
0, K: , ! r,
0.15
0.13
".!>•
"••l"
U.OX3
l-.o;c.
-.'•?
•"•• j"1
„• r-.u
7,0j?
0 C^
f .tri
•J.i?'
"j , C o ^
C - * '!
- v
rj ;;
J Vj
•* •* ,?,
j.r'i
'" -:'
t('
: i:c
C . ].?
' 1'-
o.o-,..
:.1?&
o.r,6
0,016
"'ii/ 4.
0.62 0.
0.61
0.7-J
o.s:
0.34
0.29
0.29*
0.43
0.47*
C-S5
0.36
0.38
1.01
1.13
1.28
1.02
0.2?
0.31
0.73
O.V4
C.""-.
o.1.:
J.CJ
i.ti
- _ c,G
1"! ".'•
Q.7-r*
l.-ll-
j . 32"
o.:5
i,:"/
A.:-:
0 JJ
•-.^
,-••"
',.-:-s
r. 19
C,»'l
:.£4
-,?2
0 -1
O.Sf>
O.R"1
n.Bt,
o,?:
O-i'-
O.b7
0.
0.
0.
0.
0.
0.
0.
0.
:>.
0.
0.
0.
0.
0,
r.
0.
0.
;_
0.
a.
c.
,
:•,
0
ft.
3-
0.
C -
I%.
n_
,;.
C
r.
0,
C-
* .
0.
C.
C,
3.
C.
0.
(1.
C.
09
10
10
10
04
06
09
07
OK
06
07fc
13
03
03
2;,
23
03*
C?
OK*
!/.'
'16*
- :•>
21
23
14
'S
• '*
1?
'2*
-;5
•=!
."•f,
4=
5C
t
I y
:;•
-.,
13
•^
30
M)
20
3?
•''-
0,04
0.04
0.88
S.7B
0.21
Q.J1
0.04
0.01
0,02
0.02
1-13
I.JO
l.OS
1.05
i).1)!
0.49
0.15
0.15
j ^o
i ,»3
O.PJ
0,. ">i
2,2
V,?
•J. 15
<; ?s
,-. v*
e.:/i
:.,•'<»
i, I*
U.r 2
O.f 7
0 £5
ti 6'
j , ;•. *.
I.",
; 32
.••,:«2
- -
0.1,0
-,.->4
2 , 'j
^.5
J.X
S.J
4.0
1.0
248
243
278
280
210
2iO
148
152
16»
166
254
2S8
274
2"">
712
212
!:-c
VSu
'0 '
.->' i
-,^-
"j 1
.•'
,f-';
1»;
90
'..?
•-;
;—»
4 '/ ^
209
2iO
300
301
286
S9C
180
180
16-3*
SO
180
18 S
125
120
85
85
25
24
95
95
4E>
38
A OS
114
103
iOO
•_f,
11"
130
li.5
75
'lit
4 10
172
SB*
?kO
E?9
690
KB
WO
350
350
»'0
S10
910
420
420
650
650
1090
1090
f(DjpJUcate samples de-'Ia'.in,^ mere tbar 20^ from the
ft*Dupiicatc samples frora autoiriatlc sampler.
-------
"' i ! ' ° -
~~ 1 *"' - :r '^ '•' .
^5 * £• --!
a(; j !
(i ' :
a ;
&•' l ,-) ; 'o «' ;•"•• c^
C o
'2 '
c; ' i
^ 5 !
"^ ' ~-t i
! S* i CS if
5 (2 < ° m - -1
(-* ( i
'!-> \ 1
tO ! I
4J i ;= Is
X i * i i^ '.a tn u*
riJ , U f « « OT 0 C-
~* S£ *.<5 O CJ •"*(-!
• H •
. -K -X
< — S ! !K -' -C c. rc
a i i i -; - j, *
rf , 5 i •
CO i : i
•>u J 2: * '--
H \ '• y *
00 7s 3 E SS *£
•{-> r-' j • - * •
O *t 1 O O CN -~*
r-r t-, ^
5 ^ I
03
e
H c r- <-o
r—j r-H H r-J f-i 1-1
C tc co o o o o
d g
II
•H [/) 03
Z> Sv ° ^ * *
(0 (y ft. 4-* -i .- -T r-(
£ t* E-1 o o o o
O L
^J CO
Q_t ^ C" "** "*
Ofc- 3 u*l CO
P*1 CO H
•H
r1 c M
C H T3
O "*"'
._ -, >~i
C/J CD o
• r-| f-J LQ r-^ (JD CH O O
r 2*7 fp LT. UD CD c-^
(0 "rH £ j
PH CP
g CO
O IT)
O rQ
CM | >- f £ g S
W ^ -w C5 4J O
CU < ^c
H
(XJ
*« %/>
CM
-3- !.">
O O
^ ^
o o
O~ vO
r-t r-1
O O
CN C*
O O
CN iT
3 i-O
U
•*- O
^r
in
r--
•^
r~t
m
o
o
n
i-t
'? i>
1*. T
LO r*
PJ c^J
co r
iO C-
1.S! CJ
r- <&
C' O
-^ r-4
O O
'X, f-.
O O
c\ ex
r-t r-i
O 0
O 0
•R •«
r-t O
O O
* *
m
CN ^r
•^
CJ
**••*
||
^3
(O
J-
C*^
r-f
r*
o
o
-t
*% "'
07 o
o o
•"J <•>!
0 .A
fNt C-i
I" n
o o
ro ,r
o c
CO CP
O 0
CM n
o o
Cg C-«
10 CO
ra en
y
•r-(
T3 Ji
3
u->
r-H
o
1
CN
•-t
>•-*
O
a
o
00
- > .';
CO
*-4 r-i
™i
r-i i-J
r- in
o o
r-. -*
tM OJ
O O
o o
o o
o o
rH r-t
Ui O
0 M
o
•H
t) -Q
0 2
3
to
03
C,
f\
rH
CO
O
O
f>
r
C1 C-
(-> f~.
O Q
CJi CO
~* °*
O O
'0 0-
C3 O
^ -T
o o
_?• i0
OJ .-J
o o
o o
r-i r^
o o
CM CV
CM (N
fO i-4
r-7 f-4
o
II
3
u">
1
•H
r^
-H
>• -i
(--(
o
f?
"*
t*- -
- fs!
-2- -"!
CSi U7'
r- r--
CN TJ
C" fv
c: JT
r- J-*
rv cv
f-j L.T
O fH
t- rv
o o
o o
3- fN
O O
o o
CO Cf
_0
CO J3
2 O
ZJ
tn
CM
Cw
CJ
r-(
CS '•*
M M
O 0
t-- r-
O rv
^f y
O r^
o <•-•*
'-' c'-
O> r-
rj o
O O
tr> o
0 0*
CO CO
o o
0 O
M <~l
O O
O Jf
tO iO
u
m jb
6 2
•P C1
3
a>
t ';*
c- c-
'-;•? C-
c-» o
O ""
O .*M
V-7 •")
J-- uTJ
~" "*
CO LO
to aj
o o
ID cn
-H H
O O
IN ro
O O
^ ^^
(O CD
.-^- LT>
'-I ^H
O
•H
[H ^
C a)
0 h
4J O
3
(C
CM
O
1
Cst
O
o
o
r^
w
'-x -*;
•;> o
tf) O
'"'I ^
r>* -i
C-J O
tn 10
r-i »4
j- ^*
M3 VO
0 O
in »o
tN IN
0 O
(£1 t£>
O Oi
-M cv
r^ o
0 0
o o
v> r?
O O
f? r?
CN CM
O
a
5
0]
^
§
2
•H
JJ
.*"
*>
O
•u
fct
^
§•
|(J
w
OJ
-------
26
parameters associated with particulate matter, care has been used to
prevent using sampling procedures that might affect suspended solids
concentration. Since the same automatic sampling procedure was used
for other types of water quality parameters not included in the regular
baseline river surveys or event samplings, the same good data quality
should have been obtained for these parameters.
Important water quality parameters not analyzed in the regular
baseline river survey and estuary samples are metals, organics and
bacteria. Samples were obtained for these parameters on two dates during
baseflow conditions.
The total coliform, fecal coliform and fecal streptococci counts
almost always exceeded accepted standards by a wide margin. Most of the
samples collected on June 11 and September 9, 1975 for pesticides and
PCB's had concentrations below detection limits. In a few isolated
instances, aldrin9 lindane, heptachlor epoxide, and methoxychlor showed
significant concentrations above detection limits for the June 11, 1975
survey. Most phenol concentrations for samples taken June 24 and
September 3, 1975 were above detection limits (October, 1975 Semi-Annual
Report).
Many of the stations had measurable total metal concentrations in
the June 25 and September 3, 1975 surveys (Appendix A Table 3). The
concentrations of cadmium9 seleniums arsenic and nickel were below detec-
tion limits for both dates and were not included in Appendix A Table 3.
The concentrations of copper, lead, zinc, chromium and mercury were
observed to increase significantly at and below Underwood Creek (station
413007) for both dates. This apparent trend in total metal concentra-
tions for both dates might reflect the influence of industry and high
population densities in the lower part of the watershed. The unexpectedly
high lead concentration (520 mg/1) at station 413005 on June 25, 1975 was
probably due to water from a recent runoff event discharging from a storm
sewer pipe directly above the station. A similar high concentration of
lead discharged from a combined sewer system was observed at station
413009 (880 mg/1 on June 25). A similar increase in concentration was
not observed for iron, aluminum and manganeses which are usually high
because of their geological abundance. The magnitude of the concentra-
-------
27
Appendix A. Tntl-" 3. Co rr.par :>.£'_> a .;:: :'..-ri;-jl coru,er.trd.l:Luur; at the ±2 river
."•-•" Visc-/rlo,« ci'.r-uit ?ons or -June 25" and
Station
673001
6(33002
683001
463001
413011
413008
413007
413006
413005
413010
413009
413004
673001
683002
683001
463001
413011
413008
413007
413006
413005
413010
413009
413004
Cu
<3
•3
11
p
3
C
1 "
,,1 ^-
18
10
6
12
13
<3
<3
<3
7
5
<3
9
g
-5
4
21
10
Pb
<3
<3
7
<3
5
6
28
35
520
30
380
22
<3
<3
<3
<3
<3
<3
21
35
10
11
560
10
Zu
<2C
<20
'.0
<~>r>
<20
"•-10
110
1?C
50
40
170
80
<20
<20
30
<20
<20
<20
50
130
30
20
180
50
Cr-
-~ ~-~ %.• ,
June 25
<3
<3
<3
<3
<3
<3
3
7
a
4
36
5
September
<3
<3
<3
<3
<3
<3
7
11
5
<3
80
7
Hg
t *[
, 1975
<0.2
<0,2
<0.2
<0-2
<0.2
-------
28
tions for most metals was similar for both the June 25 and September 3,
1975 surveys at all the stations. The low levels of most of the parameters
during the baseline metal and organic survey sampling increased the im-
portance of determining these parameters during runoff events.
The total concentrations of most metals increased significantly
during the August 20, 1975 runoff event at both the Schoonmaker and 70th
Street stations (413010 and 413005, respectively) and again during the
November 29 through December 1, 1975 runoff event at 70th Street (station
413005) (Appendix A Tables 4, 5 and 6). The levels of selenium and
arsenic did not change at all. Most of the metals reached concentrations
substantially higher than the concentrations observed for the baseline
metal survey samples on June 25 and September 3, 1975 at the same stations
(Appendix A Tables 4, 5 and 6). The exception was the high baseline lead
concentration observed on June 25, 1975 at the 70th Street station and the
similarity between baseline and event concentrations for aluminum and
manganese during the November 29 through December 1, 1975 runoff event.
The lead concentration on June 25, 1975 was probably the result of water
from an antecedent event discharging from a storm sewer outfall just
above the 70th Street station. The average event concentration of lead,
copper and chromium usually exceeded the concentration for these metals
during the baseline quarterly sampling at the same stations (Appendix A
Tables 4, 5 and 6). The average metal concentration for an event was
calculated by dividing the total water loading for the event into the
total loading for the metal. The average concentration on August 20, 1975
was higher for copper and chromium at the 70th Street station than at the
Schoonmaker Creek station except for lead which was about the same con-
centration for both stations. A higher average metal concentration would
be expected at 70th Street since it represents the combined Menomonee
River drainage from many different areas while Schoonmaker Creek is a
small tributary to the Menomonee River draining an older high-density
residential area. However, it was interesting to observe that many
individual concentrations of copper, lead, zinc and cadmium were similar
at both stations, and the peak concentrations for lead and zinc were
higher at Schoonmaker Creek. The average concentrations of copper, lead
and cadmium at the 70th Street station were higher on August 20, 1975
-------
29
Appendix A. Table 4.
Comparison of metal concentrations during August 20, 1975
runoff event at 70th Street and baseflow conditions on
June 25 and September 3, 1975
Time or
date
1330
1415
1500*
1545
1630
1715
Ave. Cone.**
June 25
Sept. 3
Cu
58
76
110
65
97
58
87
10
J
Pb
53
320
550
400
146
130
370
520
10
Zn
60
320
320
230
180
160
-
60
30
Cr
Event
14
31
50
24
25
,-,.-,
3T>
9
c.
Ni
i In- /I
Mg/ -1-
(Aug. 20,
<20
140
48
27
<20
<20
„
Baseflow
<20
-------
30
Appendix A. Table 5.
Comparison of metal concentrations in Schoonmaker Creek
during the August 20, 1975 runoff event and baseflow
condition on June 25 and September 3, 1975
Time or
date
1240
1255=''
1325
1340
1355
1405*
1420
1430
1440
1445
Ave . Cone . **
June 25
Sept . 3
Cu
120
48
45
20
140
50
54
36
21
17
60
6
4
Pb
272
550
320
174
640
368
140
85
63
68
410
30
11
Zn
130
220
120
100
270
220
900
80
80
360
-
40
20
Cr
! ,„ /T
Mg/-L
Event (Aug. 20
25
15
9
8
20
16
9
7
9
5
10
Baseflow
4
<3
Cd
, 1975;
2.7
2.5
1.2
0.9
2.4
1.8
0.7
0.6
0.8
0.6
-
<0.2
<0,2
Fe
5,500
10,000
4,400
5,000
10.800
12,000
8,200
4,400
3,800
4,000
-
4,500
4,600
Al
2,200
4,700
2,800
2,000
5,000
4,200
4,000
3,500
2,800
2,200
-
570
1,280
Mn
110
230
120
120
350
330
260
150
120
140
-
70
60
* Peak flow,
•;-'s Average concentration is the total loading
loading for event.
metal divided by total water
-------
31
Appendix A. Table
th-; NcverriLer
Wl'C "i DdSf '' ,! OV
:al concentrations for 70th Street during
through Dece-.I-.ir 1, .1975 runoff event
,ue? on .Tune 25 and September 3. 1975
Time or
date
0445
0845*
1245
1640
2040
0040*
0445
0845
1240
1640
2040
0040
0040
0840
Ave. Cone.**
June 25
Sept. 3
Cu
&4
106
72
72
58
72
58
40
36
44
42
46
35
36
58
10
5
Pb
13
3.26
104
84
76
108
80
54
38
27
19
15
12
10
66
520
10
Zii
80
l?Q
1 IG-
LOO
100
100
100
80
70
60
50
50
40
40
-
60
30
Cr
Event (Nov.
8
21
'40
21
14
Event (Nov.
12
9
8
12
5
7
Event (Dec.
4
3
4
13
Basef
9
5
Cd
29, 1975)
0.7
2.1
1.3
1.4
1.9
30, 1975)
1.0
1.0
0.7
0.6
0.7
0.8
1, 1975)
0.8
1.1
1.2
-
:low
0.2
<0.2
re
700
2,700
3S700
3,900
3,400
3,800
5,200
4,200
4,600
3,400
2,900
2,200
1,800
1,700
-
4,700
1,800
Al
300
1,080
1,000
120
1,100
1,150
1,280
1,560
1,100
1,080
960
860
600
560
-
1,520
440
Mn
<40
130
100
110
110
120
110
100
110
90
80
90
70
70
-
120
90
* Peak flow.
** Average concentration is the total loading of metal divided by total water
loading for event.
-------
32
than for the November 29 through December 1, 1975 event. Also, the indi-
vidual concentrations were generally higher, and peak concentrations were
higher at 70th Street on August 20, 1975 than for the November 29 through
December 1, 1975 event for all the metals except cadmium and copper. The
higher metal concentrations on August 20, 1975 at 70th Street were probably
due to the greater rain intensity of the August 20, 1975 event, intensify-
ing the rate of metal flushing from the watershed. Approximately 0.7 inches
of rain fell in about 3 hours on August 20, and 1.0 inches of rain were
recorded for the 3-day November 29 through December 1, 1975 event, most of
which fell throughout the first day. The peak flow rate for the August 20
event (34,000 I/sec) at 70th Street was almost twice the peak flow rate
for the November 29 through December 1 event (19,000 I/sec).
The peak flow at 70th Street during the August 20, 1975 event coin-
cided with the peak total metal concentration for all the metals except
nickel (Appendix A Table 4). Two peaks were observed on the Schoonmaker
Creek hydrograph during August 20, 1975. Most metals showed a peak concen-
tration at the first peak, but the peak metal concentration was about 10
minutes before the second peak for most of the metals (Appendix A Fig. 10
and Table 5). However, the shape of the hydrograph around the second peak
indicated the actual peak flow might not have been recorded. Although the
lead concentration at Schoonmaker Creek was similar for both peaks, the
pollutograph or input rate curve indicates the loading during the first
discharge peak was small compared to the second discharge peak (Appendix A
Fig. 10).
The hydrograph for the November 29 through December 1, 1975 event had
an initial plateau followed by a distinctive peak. Most of the metals
reached a peak concentration at the beginning of the plateau, but cadmium
reached a peak before chromium, iron and aluminum after the distinctive peak
following the plateau. Copper, lead, zinc and manganese concentration peaks
coincided with the flow peak following the flow plateau (Appendix A Fig. 11
and Table 6). The usually less distinct and noncoinciding nature of the
second metal concentration peak for most of the metals was probably due to
the relatively low intensity of the event. An example of this was the peak
concentration observed for chromium (Appendix A Fig. 11). The higher con-
centration of lead in the first concentration peak (126 Ug/1) as compared
to the second (108 ug/l) was typical of many of the metals and was
-------
-------
33
1600
1100
'200
,-.1000
a 800 -
500 -
400 -
200 -
1200
ft 00
—O—- Di ic r-'i-g?
—O !-e«d cone.
*, Lead inpit. rate
13CO
Time
1400
1500
Appendix A Fig. 10, Discharge and lead concentration at Schoonmaker
Creek during August 20, 1975 runoff event.
-------
o
o
•a-
CVl
1
o
0
o
CVJ
1
(oas/Bui) e:
o
o
la
1
}EJ indiu
o
0
cv
1
o
0
00
( ... ,.
o
0
1 1
pug
0
CM
r—
O
O
•—
C3
co
o
IO
o
•*
o
CM
u<
I
6
i
i
01
en
i.
ri
o.
c
o
CO
-t-1
o
J-J 4_>
(0 C
w >
tt) 0)
•H
tO MH
C
•H ID
r-
rt3 CD
IT) rH
0)
H «•
Fi
(0
P
O>' °
t-r 'Si
L" p
r:1 O
f J -P
-------
35
probably due to the first flush of the watershed (Appendix A Fig. 11). The
lead pollutograph mirrored the rapid rise of the hydrograph to a value
just under the second flow peak and indicated the loading of lead in the
first concentration peak was a significant part of the total loading.
This was in direct contrast to the relatively small lead loading for the
first peak at Schoonmaker Creek on August 20, 1975. This was a result of
the discharge in the final peak being small compared to the second even
though the concentrations were similar.
Another parameter from the baseline quarterly surveys measured during
an event was organics. Contrary to the metal results, the organic concen-
trations did not increase during the November 20 and 21, 1975 event at
the 70th Street station. All the organics levels were below the limits of
detection for the entire event.
Analyses of most of the event samples for regular baseline river survey
parameters (nutrients, etc.) will continue. The concentration trends for
these parameters during runoff events have been discussed in the October,
1975 Semi-Annual Report. The parameters associated with particulate
matter (suspended solids, total P and total organic N) usually increased
in concentrations during events at all the stations. The concentrations
of dissolved nutrients (dissolved reactive P and (nitrate + nitrite)-N)
increased during the events at Noyes Creek (station 413011) and Schoonmaker
Creek (413010), but the concentrations of dissolved nutrients at 70th
Street (station 413005) remained about the same with slight increases or
decreases. Hardness, alkalinity and the concentration of chloride de-
creased for all stations during the events. Eventually, a sufficient
number of events at the automatic river stations will be documented to
allow a more detailed analysis of data to compare the stations as to nutri-
ents and other parameters. Also, more detail in regard to related concen-
tration changes during events will be possible after performing statistical
correlation analyses for all the listed parameters for one event.
Concentrations of nutrients and other parameters during events are
being combined with the discharge data to make pollutographs or loading
curves. Examination of the pollutographs for each station under dif-
ferent hydrological conditions reveals a potential for predicting, with
very little or no sampling, the loadings for different parameters during
-------
-------
36
an event. Determination of parameter loadings during events is essential
to river monitoring objectives, as loading data for events provide the
only means for determining contributions from different land uses. Loading
data will be used also to determine the quantities of various parameters
in the river system that might affect Lake Michigan water quality.
Copper, lead and chromium loadings were calculated for the August 20,
1975 and the November 29 through December 1, 1975 events (Appendix A
Table 7). Both the event loadings and total loadings for metals were
greater for the August 20, 1975 and November 29 through December 1, 1975
events at the 70th Street station than at the Schoonmaker Creek station
due to the greater water volumes in the former station. Also, due to the
greater water volume for the November 29 through December 1, 1975 event
than for the August 20, 1975 event, the 70th Street station experienced
a greater total metal loading for the November 29 through December 1,
1975 event.
The baseflow loading for a month was calculated by excluding the
discharge from runoff events and combining baseline survey data with the
mean monthly discharge. Average monthly baseline metal loadings were
larger in June than in September. Most total metal loadings for events
were greater than the approximate baseflow loadings (Appendix A Tables 7
and 8). Large loading rates for events when compared to average monthly
baseline loadings emphasize the importance of recording event loadings
for metals (Appendix A Table 8).
The event loadings for total P, dissolved reactive P (DRP) and
suspended solids were calculated for the June 17, July 18, August 18
and 20, and September 5, 1975 runoff events (Appendix A Table 9). The
total event loadings for these parameters varied significantly from one
event to another at the same station and corresponded to the variation in
concentrations of these parameters and the water volume.
In most instances, the total event loadings of total P, DRP, and
suspended solids at Schoonmaker and Noyes Creeks exceeded or approximated
the total baseline monthly loadings (Appendix A Table 11). The ratios of
the event loadings to the total monthly baseflow loadings at 70th Street
station were low in most instances. The suspended solids loadings on
August 20 and September 5, 1975 were the exceptions for the 70th Street
-------
------- An error occurred while trying to OCR this image.
-------
38
Appendix A. Table 8. Ratio of event metal loadings for floowing dates to
the June and September approximate baseflow loadings
Sampling
location
Schoonmaker Creek
70th Street
70th Street
Sampling
month, 1975
June
September
June
September
June
September
Cu
2
40
0.57
1.62
Nov.
3.76
10.69
Parameters
Pb
August 20, 1975
4.67
140
0.05
3.42
29 through Dec.
0.08
6.08
Cr
event
1.25
-
0.24
0.62
1, 1975 event
0.91
2.38
-------
'4-1 rH
39
0)
H
X
•H
•8
0)
a.
•u | I
-------
"40
£
0
J J
MH
C
•H bO
^a c
ffl «H
3§5
id
«
13 -P
•H <])
H CD
0 fc
CO -P
C/3
T3
co ,c
T3 P
C 0
CD C--
&
W T)
3 C
cn rt
T) eX
C CD
m CD
•H X
•P (0
o S
(0 C
0) O
fc 0
•^
T? O
c; co
>
H «•
O ,X
W 0)
CO Of
•H (U
T3 O
•> m
P-I a> LO
>,£>
iH O CTi
n> 5? •-!
•p
C 4-1 4.4
r— i p, Q
O
i — i
H
"S
E~"
M
P)
s
H
i-\
O
13
•d
C
01
a,
^j
GO
fr,
Oi
o
ri
&
VI
a
-
a
o
-^
be
C
t
fti
CJ
Cl
n
*
"5i>
E
n)
00
ii
e
"O
60
^
"^
~^.
E
>>
n!
?
^n
A"
if
r-
J,
-"
^
IT)
H M CN J CMC\,CQ
oooooooooooo
m
iO CM .3- O
CMrHiNCr-tOOO
*"'*'"• -^^.f-.
^ CT5 CM CM
V >,; ^
*-• "i «j f ^ "i
c .'-- y r', i" ^ ?r ^ S ,rT 3, a,
-', .H ^ ,/ ":l r~ ;}_ % £. '?, 3- .0
' ^ " J -'-',, i
R
(0
-------
bO
a
•H
TS
o ti
H £
•rH
£4 Td
•P O
ft) rH
t3 O
C rH
m tp
Q)
M CO
G ttf
•H rO
-a
(tj £>•>
O H
rH XI
-p
Tj O
•M g
o n5
-P W
> C
H O
O -H
CO -p
W CO
•H -p
T) M
m
a m
•P Q)
O -P
•p m
T3
-P
C t -H
0
O rH
O
0 IP
"H
-P P
rt! O
K ^
, — -1
i — '.
c;
.— -1
f:
•T<
W
J^
(Li
o 01 m
CTI H rH CM O CM O
CM rH
LO O O O H J- =f
CM CM CO CN O O O
• • • « • • •
rH O O 0 O O O
CO CO CN rH CM r-
CO 00 O O C-J O
• * • • • *
CM O OH O O O
V
ir-
a: o co o p
i-H <- J "H CM 3)
t~~ CG X-;
,— '' — 5 XJ J-J ^_) f J ^ •
» W te * ";;
a: t-, "," 3 D 3 4-J
r; rH fcn a'; bn M ct,
r- ;J :J r:; P ?
-------
station. The low event water volume combined with average total P and DRP
concentrations that were similar to the average monthly baseflow concen-
trations probably accounted for the low loading ratios at the 70th Street
station for these events (Appendix A Tables 9 and 10). Since the relative
contribution of total monthly nutrient and suspended solids loadings from
runoff events was only based on one or two events in a month, a more
accurate estimate of the relative significance of runoff events to total
monthly loadings will be possible when the loadings are calculated for
more events.
Future loading calculations will include all the water quality
parameters determined during event and baseline surveys. Loading calcula-
tions will be done for all important events and summarized along with the
baseline loadings on a monthly and seasonal basis for each station. These
calculations will provide more accurate estimates of the event-to-monthly
loading ratios and indicate possible seasonal trends. The loading data
will be interpreted in terms of land use activities by the land use-
water quality models presently being developed.
Problems Encountered during Field Activities
The following is a listing of the various types of field activity
problems that were encountered in the 1975 field year and some suggested
solutions.
Sample collection
U.S. Geological Survey (Ps-69) automatic pumping samplers were used
to collect water samples during storm events as well as during baseline
sampling periods. Problems associated with this sampler follow.
Problem: Partial or complete clogging of the intake pipe by heavy debris,
causing one of three sampler fuses or one of two 110-12 volt power
converter fuses to blow.
Solution: No reliable method of screening the intake has been found to
date. Switching to "slow blow" fuses has allowed, at times, the sampler
-------
to complete its pumping cycle even though partially clogged; it then may
clear itself during the purging cycle which follows,
Problem: At times, light debris (leaf or macrophyte fragments) would clog
the swing-arm nozzle, thereby cutting off water delivery to the sampler
bottles.
Solution; A medium mesh screen positioned above the swing-arm funnel col-
lected all debris that might otherwise clog the nozzle.
Problem; Sampler pumps sustained worn-out impellers and control box
malfunctions.
Solution; Pump mountings were improved to reduce vibration, and control
box circuitry was improved.
Problem; Calibration of sampler efficiency and sample diversity.
Solution; During each sampling survey of all river sample sites, one site
was chosen, on a rotating basis, from which to collect two simultaneous
grab samples and a sampler sample. The grab samples were collected as close
as possible to the sampler intake and at the precise moment the sampler
was collecting. Both of these grab samples are defined as simultaneous
samples and are analyzed separately to determine the amount of natural
diversity found in the water quality of two essentially like samples. The
sampler sample taken at the same time as the two grab samples is defined
as the comparison sample since its analysis is compared to that of the grab
samples to aid in determining the effects that the pumping action of the
sampler may have o:-; the water being sampled. Location of the sampler
intake in reference to stream width and water depth should a.lso be field
calibrated. This involves the comparison of analyses of sampler samples
with those of depth- and point-integrated grab samples.
Sample handling
All watei" snwplos are | ~cked with ice ;".~ .".n;-T.;\ated ccnlvi.lne:."1} and
sent via U.S. rail to the v»o.iar quality labora'c^vy of the Laboratory of
-------
Hygiene at the University of Wisconsin-Madison (a distance of 70 miles)
for analysis.
Problem; Delays sometimes occurred in mailing iced samples to the labora-
tory.
Solution: Samples in containers which are prestamped and marked for
special handling may be delivered directly to the loading dock of the main
post office to be sent via U.S. mail-parcel post. This procedure prevents
delays since the containers are loaded directly from the loading dock to a
delivery truck and do not enter the post office. An additional advantage
is that the loading dock is never closed, so samples can be mailed at any
time.
Continuous monitoring
Five selected sites along the Menomonee River watershed are equipped
for continuous monitoring of water temperature, dissolved oxygen, pH and
conductivity. In situ water quality monitors (Surveyor Model 6D), manu-
factured by Hydrolab Corporation, Austin, Texas, are used for this
monitoring.
Problem; How to place the in eitu water quality monitoring probes (con-
tained inside a protective sonde) in the river.
Solution; Probe sondes were fastened to steel posts driven into the river-
bed. Attachment was by means of a small "s" hook attached to the sonde
top, which allowed the sonde to be hung from a "u" bolt attached to the
post. The height of the sonde in the water could easily be adjusted by
raising or lowering the "u" bolt on the post. The sonde was then hooked
to the "u" bolt and strapped tightly to the post with nylon webbing and a
rachet. Caution signs were placed directly upstream of each sonde poet to
warn canoeists or snowmobilers and ice skaters of the obstruction.
Problem; Vandalism to equipment.
Solution; On-shore equipment is housed in a locked building. In eitu
-------
-------
equipment is submerged and locked to the posts. The sonde power cable from
the river to the equipment building is buried.
Problem; How to efficiently work with the in situ equipment which is
usually located in the middle of the river.
Solution; Anchoring a 5-foot plastic boat alongside the sonde while working
on it has proved to be a convenience. The boat serves as a platform to hold
tools and the sonde itself while the operator (standing in the water) is
attaching or separating it from its power cable. When the sonde is dis-
connected from the cable for cleaning and calibration, the cable connec-
tions are cleaned and secured inside the boat to keep them dry.
Problem; How to efficiently clean and calibrate the probes.
Solution; Weekly calibration has been found to be sufficient to assure
reasonably accurate data. The sondes (with probes) are raised from their
underwater positions and disconnected from the power cable. The watertight
cable connections are carefully cleaned with alcohol, cotton swabs and pipe
cleaners, and are secured inside the boat to dry. The sonde and probes
with their connections are then taken inside the equipment building and
thoroughly cleaned in detergent solution. Additional cleaning of the pH
reference probe, which has a porous teflon surface, is accomplished using
a stiff wire brush and kitchen cleanser. An extremely soiled pH reference
probe may be brought back to life by carefully scraping off the soiled
surface film of the teflon with a sharp knife. However, this procedure
should be classed as drastic and used only when absolutely necessary. The
Clark type dissolved oxygen probe, when cleaned, is replaced by a like
probe which has had a new membrane installed the previous day. On rare
occasions the silver anode of this probe becomes discolored and adversely
affects the probe's performance. Cleaning is accomplished by immersing
the probe, with membrane off to expose the anode, in an ultrasonic cleaning
tank which contains a 50% solution of ammonium hydroxide and distilled
water along with a fine grade detergent (e.g., Alconox). After approxi-
mately ten minutes the anode should regain its white color, and the probe
is then ready to be thoroughly rinsed and soaked overnight in distilled
-------
-------
water. The probe should then be prepared for normal use. When cleaning
the conductivity probe, it is helpful to disassemble it into its two main
sections. After thorough cleaning, the sections are reassembled, making
sure the two sections are tightly seated against each other. Seating is
made easier if the two sections are first moistened to allow the "0" rings,
which seat the two sections, to easily slip into place. After cleaning
and rinsing, all probes and sonde connections are then cleaned with alcohol,
dried and coated with a light film of 0-ring grease. All probes are then
inserted into the sonde and connected to the surface monitoring unit by a
spare power cable. (The main power cable is semi-permanently installed
from the building to the sonde location in the river.) The sonde is then
calibrated before leaving the building, using standard buffer solutions,
standard conductivity solutions, and an air calibration tube and internal
calibration techniques as per operators' manual instructions. After
calibration, probes are disconnected from the spare cable and installed
with the sonde to the main cable in its river location.
-------
-------
APPENDIX B
-------
MENOMONEE RIVER DATA BASE SYSTEM
Introduction
The primary purpose of the Menomonee River Data Base System is to
provide ready access to, and analysis of, data from a small watershed
study. Emphasis has been put on allowing an interactive user to use the
data. This emphasis is one of the primary determinants of the architec-
ture for the system. Also, of major importance were: the types and
amount of data; the keys which would be used to access the data; and the
types of analyses, summaries, and forms of reporting that users would
expect. Each of these determinants and its subsequent effects are
analyzed after a discussion of the general approach used.
General Notes on Systems Design
Most of the main routines of the system are written in ASCII ANSI
Cobol. This language was chosen for two reasons: it is readable, even
by a noncomputer-oriented person, and thus is relatively self-documenting;
and second, it is relatively efficient at data manipulation. Many of the
support or driver-level routines are written in Fortran V, a relatively
efficient numerical algorithm language. There are a relatively small
number of routines that are written in Univac Assembler Language. These
routines assign and free files and perform input and output on several
files.
In addition to a breakdown by language type, the various routines of
the system can be separated into the main program, primary functional
routines and support routines. This breakdown is indicative of the "top
down," semi-structured approach to programming that was used.
In "top-down" programming, the main program is made fully operational
first and calls dummy routines which simulate the later-developed secondary
routines. These routines are then written one by one, with their lower
level routines simulated until all higher levels are operational. This
approach was followed with the exception that certain common i-o routines
and other common "slave" routines were written first.
The semi-structured approach implies that the long functional routines
-------
-------
are also hierarchically organized. There is a common entry. The user is
asked which of the program's options should be executed (possibly after
some initial explanation and common initialization). Control is passed
to a routine which is short or organized hierarchically. Control flows
in as linear a fashion as practical within a routine except when a sub-
level is performed. The routines always end at a common point or return
to the start.
The main routine follows this approach: first, i-o is initialized
and the user is (optionally) given an explanation of choices available;
then, the choice of a function causes branching to another program in
most cases. Passwords are sometimes verified before control passes.
The function-handling routines that handle a specific type of request
are largely of the same format. The primary exceptions are the main data
input and report listing programs. The input program is primarily linear
within the structure of Cobol sort section flow. The report program is
cascading—that is, linear loops overflow into the next outer level of
looping only when input changes according to one of the primary keys.
The support programs are generally short, linear, easily understood
and have only one primary purpose. The i-o initializer "STRTIO" follows
the switchboard approach as does "A53302," the primary i-o controller.
Primary Determinants of Design
The desire to make the program interactive gave rise to severl domi-
nant features of the system. An attempt was made to make the programs
usable by people who have little or no familiarity with computers. Files
are automatically assigned, freed and sent to output devices. Other files
can be set up that contain the job control and data for plotting or
statistical analyses (both of these are best done in separate steps).
The programs are made as self-explanatory as possible. Messages are
generated explaining use of the system. The user needs to refer to docu-
mentation only for the most complex of usages.
The desire for rapid access of the data, which caused interactive
capability to be designed, also controlled aspects of the data files
structure. The primary data file is set up as "hashed" directory access,
-------
-------
50
multiple-linked file. "Hashing" is a random access technique explained
in Addendum I. The multiple linkings are set up as data is entered in
order to allow access to data sorted in various ways without the necessity
of actually re-sorting data. The increase in overhead is more than com-
pensated for by lowered handling costs.
The data was grouped into sets or "types" (such as nutrients, metals,
etc.) which consist of logically associated kinds of physical or chemical
analyses. The "type" grouping gives rise to the association of data
on a line of a report and also, thereby, to the establishment of one
logical record in a file. The other alternative was to put in one record
all analyses performed on one sample. This route was not taken, not only
because one record could not then be listed succinctly, but also because
of the great variability in the number of analyses done on one sample.
Instead, all records associated with one sample are linked through a
record that is "type" 19, there being eighteen possible types of analyses.
In addition to type of analysis, a user could be expected to desire
retrieval of data based on where and when it was taken or because it
could be logically associated in some fashion with other data. The keys
for retrieval thus become: station (which is a geographic location),
depth, and time at which the samples were taken, as well as type.
As indicated above, "type" 19 records are used to link records
from the same sample. "Type" 19 records are also linked to each other
according to several sort sequence combinations of station, time and
depth. The data records are also sequentially chained, as they are input,
by another sort sequence.
The usual method of retrieving a group of data would be to request a
specific starting and ending point and a sort sequence. There is no
great difficulty in this method for a user except for the necessity of pick-
ing an exact set of times (to the minute) that corresponds to an actual
sampling event. To eliminate this difficulty, a file has been established
containing the keys and record addresses of certain records. Thece are
the information for the first and last (by time) records for each half month
-------
-------
51
for each station (maximum fifty) and for all stations (the "51st" station).
The types of analyses, summaries and modes of data presentation for
which users expressed a desire are the primary determinants of the sort
sequences. They also caused the establishment of an ancillary comments
file, and control the design of plotting and other analysis programs.
File Structure and Layouts
There are five primary files (see layouts in Addendum II). They are
the system file, the main data file, the directory file for the main file,
the associated comments file, and the link table file. The link table
file has been analyzed above; the directory file is discussed in Addendum I;
and the comments file is self-explanatory (see notes in Addendum II.C).
The main data file consists of data and link records.
The link record consists of six binary zeros, record keys other than
type, link addresses for the eighteen associated types, and link addresses
for logically adjacent "type" 19 records. The data record consists of a
link address to the record next in the primary sort sequence, deletion
or modification code, record keys, sixteen sets of sign and value
(K stands for <-), links to two comment records, and the address of the
type 19 record.
The system file contains conversion factors for: each analysis;
a change of status code; a count of the number of stations and types, the
start of miscellaneous types; the beginning year; the number of decimal
points for each type; the next available data record space; a set of counts
for each station and month combination that tell the number and change
status of analyses; an array for each parameter group containing a list
of parameter codes and the number of codes, a set of headings or descrip-
tions and a description of the group; three system passwords; the next
available comment record space: a set of user passwords and user counts;
the first deleted record space; two numbers for calculating the hashing
efficiency; and information for octal dumps of files.
-------
-------
52
Addendum I
The "hashed" main data file actually consists of two files: the
data file itself and its directory. The purpose of this addendum is to
explain the use of the directory in accessing data records.
Hashing is a technique used for storage and retrieval of data that
is sparsely spread throughout a large matrix. By using a smaller matrix
and indexing the data through an arithmetic recombination of its original
indices, the computer program can greatly reduce the required storage
space. There does arise the possibility of several entries being
referenced to the same matrix position. To eliminate possible destruc-
tion of records, an overflow area is provided. Each record in the
directory maintains space to point to records in this area if there are
"collisions." Thus, a directory entry consists of three logical parts:
1) a master file address, 2) an overflow pointer area, and 3) a set of
keys for verification of which entry is being referred to in the original
matrix.
It is also necessary to maintain a record of the next empty spaces
in the master file and the overflow area. These records are kept in the
"zero11 entry of the matrix.
Example
The keys we use are:
1. Station or geographic location—a sequence number from 1 to 50,
2. Depth—from 0 to 9 meters,
3. Data group type—a sequence of numbers from 1 to 19,
*»-. Date and time—a 5-year period recorded to the minute.
This matrix is obviously made sparse by the large number of possible times
of sampling.
-------
53
The arithmetic recombination is the following algorithm:
Key 1 = (((Y-1973)*100+M)*100+D)*10000+100*H+MIN
Key 2 s ((S~1)*19+T)*10+DP
Yields two keys for storage that represent original position
Hash value = Mod (Key 1 + Key 2, 20011)
20011 is a prime number. Modulus is the remainder function.
The hash file directory is thereby 20012 entries plus any
overflows. The extra record is derived from the record of
next available spaces.
where
Y-1'973 = years from beginning of project
M = number of months from start of recording
D = day
H = hour
MIN = minute
S = station sequence number
T = group type number
DP = depth in meters
This algorithm is used because it will compress the original matrix in
such a fashion as to minimize "collisions," thus maintaining relative
efficiency in data storage and retrieval.
-------
-------
ADDENDUM II
A.
B.I.
110:
111:
112:
113:
114:
115:
116:
117:
118:
119:
522:
523:
524:
525:
526:
527:
528:
529:
530:
531:
532:
01 LINK- ARRAY.
05
10
10
15
15
15
15
15
10
PERIOD- INFO OCCURS 12 TIMES.
PERIOD-DEPTH
PERIOD-DATE.
PERIOD- YEAR
PERIOD-MNTH
PERIOD-DAY
PERIOD-HOUR
PERIOD-MIN
PICTURE
PICTURE
PICTURE
PICTURE
PICTURE
PICTURE
9.
9.
99.
99.
99.
99.
LINK-POSITIONS USAGE IS COMPUTATIONAL PICTURE 9(6)
01 UPDATE-BODY.
05
05
05
05
05
05
05
FILLER
KEY-PART
BODY-1
BODY-2
CMNT-L1
CMNT-L2
CMNT-L3
01 UPDATE-RED REDEFINES
02
05
SOME-NAME.
RECORD-POSITION
PICTURE
PICTURE
PICTURE
PICTURE
PICTURE
PICTURE
PICTURE
UPDATE-BODY.
PICTURE 9(8)
X(7).
X(15).
X(64).
X(64).
9(6).
9(6).
9(6).
SYNC RIGHT USAGE
IS COMP-4.
533: 02 RECORD-REST PICTURE X(162).
534: 02 REST-RECORD REDEFINES RECORD-REST.
535: 03 NOT-TIME.
536: 05 RECORD-DELETE PICTURE 9.
537: 03 RECORD-TIME.
538: 05 RECORD- YEAR PICTURE 99.
539: 05 RECORD-MNTH PICTURE 99.
540: 05 RECORD-DAY PICTURE XX.
541: 05 RECORD-HOUR PICTURE XX.
542: 05 RECORD-MIN PICTURE XX.
543: 03 NOT-TIME-EITHER.
544: 05 RECORD-STATION PICTURE 99.
545: 05 RECORD-TYPE PICTURE 99.
546: 05 RECORD-DEPTH PICTURE 9.
547: 05 RECORD-BODY PICTURE X(128).
548: 05 RCD-BODY REDEFINES RECORD-BODY.
549: 10 REC-PART OCCURS 16 TIMES.
550: 15 REC-SIGN PICTURE X.
551: 15 REC-ITEM PICTURE X(7).
552: 05 CMNT-ASSN.
553: 10 CMNT-ASSOC OCCURS 3 TIMES PICTURE 9(6).
-------
-------
55
B.2. 120: 01 LINK-RECORD.
121: 05 LINK-VERIFY PICTURE 9(8) USAGE IS COMP-4 SYNC RIGHT
122: VALUE IS 0.
123: 88 LINK-VALID VALUE 0.
124: 05 FILLER PICTURE X(5).
125: 05 LINK-STATION PICTURE 99.
126: 05 LINK-DEPTH PICTURE 9.
127: 05 LINK-TIME PICTURE 9(10).
128: 05 LNK-TME REDEFINES LINK-TIME.
129: 10 LINK-YEAR PICTURE 99.
130: 10 LINK-MNTH PICTURE 99.
131: 10 LINK-DAY PICTURE 99.
132: 10 LINK-HOUR PICTURE 99.
133: 10 LINK-MIN PICTURE 99.
134: 05 REC19-POS-N OCCURS 24 TIMES PIC S9(8) SYNC RIGHT
135: USAGE IS COMP-4.
C. 176: 01 COMMENTARY.
177: 05 LINE-REF PICTURE 9(6).
178: 05 CMNT-1 PICTURE X(60).
179: 05 CMNT-2 PICTURE X(60).
180: 05 COMMENT-NUMBER PICTURE 9(6).
D. 395: 01 SYSTEM-STATUS RECORD.
396: 02 MOST-SSF-LAYOUT.
397: 05 TYPE-CONSTANT OCCURS 19 TIMES.
398: 10 TYPE-CONVERSIONS OCCURS 16 TIMES PICTURE 9(4)V9(4)
399: USAGE IS COMP-4.
400: 05 STATUS-FILE-LENGTH PICTURE 99.
401: 05 CURRENT-STATIONS PICTURE 99.
402: 05 NUMBER-OF-TYPES PICTURE 99.
403: 05 MISC PICTURE 99.
404: 05 REPORT-YEAR PICTURE 99.
405: 05 EDIT-PARM-FIELDS OCCURS 19 TIMES.
406: 10 PARM-EDITOR OCCURS 16 TIMES PICTURE X.
407: 05 NEXT-RECORD PICTURE 9(6).
408: 05 STATION-RELATNS OCCURS 50 TIMES.
409: 10 STATIONS PICTURE 9(6).
410: 10 STATION-DESCR PICTURE X(30).
411: 10 REPORT-COUNTS OCCURS 60 TIMES.
412: 15 REPORTED-PREVIOUSLY USAGE IS COMP PIC 9(6).
413: 15 UNREPORTED-YET USAGE IS COMP PIC 9(3).
414: 15 REPORT-CHANGES USAGE IS COMP PIC 9(3).
415: 15 REPORT-DELETES USAGE IS COMP PIC 9(3).
416: 15 LEVEL-OF-ACTIVITY USAGE IS COMP PIC 9(3).
417: 05 CURRENT-TYPES OCCURS 19 TIMES.
418: 10 PARM-NUMBERS OCCURS 16 TIMES PICTURE X(5).
419: 10 NUMBER-OF-PARMS PICTURE 99.
420: 10 HEADINGS.
421: 15 HEADING-1.
422: 20 FILLER PICTURE XXXX.
-------
-------
56
D.
continued
423:
424:
425:
426:
427:
428:
429:
430:
431:
432:
433:
434:
435:
436:
437:
438:
439:
440:
441:
442:
443:
444:
445:
446:
447:
448:
449:
450:
451:
452:
453:
454:
455:
456:
457:
458:
459:
460:
461:
462:
463:
464:
465:
466:
467:
468:
469:
470:
471:
472:
20
20
15
20
20
20
15
20
20
20
15
20
20
20
10
05
05
05
05
05
05
10
10
05
05
05
05
02
05
05
05
05
05
05
05
05
05
05
05
05
05
05
05
05
05
05
05
05
05
02
HEADING-1-1
HEADING-1-2
HEADING-2.
FILLER
HEADING-2-1
HEADING-2-2
HEADING-3 .
FILLER
HEADING-3-1
HEADING-3-2
HEADING-4.
FILLER
HEADING-4-1
HEADING-4-2
TYPE-DESCR
PASSWORDl
PASSWORD2
PASSWORDS
NEXT-COMMENT
NUMBER-OF- INITIALS
INIT-ACCT OCCURS 100
INITIALS
INIT-USAGE
NEXT-DELETE
FILLER
PICTURE X(64).
PICTURE X(64).
PICTURE XXXX.
PICTURE X(64).
PICTURE X(64).
PICTURE XXXX.
PICTURE X(64).
PICTURE X(64).
PICTURE XXXX.
PICTURE X(64).
PICTURE X(64).
PICTURE X(13).
PICTURE XXX.
PICTURE XXX.
PICTURE XXX.
PICTURE 9(6).
PICTURE 999.
TIMES.
PICTURE XXX.
PICTURE 99.
PICTURE 9(6).
PICTURE XXX.
MEAN-EFF PICTURE 9(4)V9(4) SYNC RIGHT USAGE IS COMP-4
MAX-EFF USAGE IS COMP-4 SYNC RIGHT PIC 9(8).
TABLE-OF-FILES.
FILLER VALUE 'LNKFLE
FILLER VALUE 'RETEMP
FILLER VALUE 'CONFILE
FILLER VALUE 'SSF-MR
FILLER VALUE 'DUMB
FILLER VALUE 'CMNTS
FILLER
FILLER VALUE 96
FILLER VALUE 96
FILLER VALUE 0
FILLER VALUE 96
FILLER VALUE 96
FILLER VALUE 0
FILLER
FILLER VALUE 5
FILLER VALUE 29
FILLER VALUE 28
FILLER VALUE 15
FILLER VALUE 8
FILLER VALUE 22
FILLER
' PICTURE X(12).
' PICTURE X(12).
1 PICTURE X(12).
1 PICTURE X(12).
1 PICTURE X(12).
1 PICTURE X(12).
PICTURE X(72).
PICTURE 99.
PICTURE 99.
PICTURE 99.
PICTURE 99.
PICTURE 99.
PICTURE 99.
PICTURE X(12).
PICTURE 99.
PICTURE 99.
PICTURE 99.
PICTURE 99.
PICTURE 99.
PICTURE 99.
PICTURE X(12).
TABLE-DEFINITION REDEFINES TABLE-OF-FILES.
-------
57
D. continued
473: 05 TABLE-NAME OCCURS 12 TIMES PICTURE X(12).
U7M-: 05 TABLE-CONSTANT OCCURS 12 TIMES PICTURE 99.
U75: 05 TABLE-FACTOR OCCURS 12 TIMES PICTURE 99.
H76: 02 UNUSED-YET-SPACE.
U77: 05 FILLER PICTURE X(288).
-------
APPENDIX C
-------
59
LAND DATA MANAGEMENT SYSTEM
Introduction
The Land Data Management System (Land DMS) is a digital computer-
based system designed to store, retrieve, analyze and display—in tabular
or graphic form—land data for the Menomonee River watershed. The term
"land data" as used in the context of the Land DMS is a comprehensive
concept in that it denotes all those watershed characteristics that have
an areal extent. For example, land data encompass land use, soil type
and civil division information but do not include water quality or stream
flow data.
Uses of the Land DMS
The Land DMS has two principal uses in the Menomonee River Pilot
Watershed Study:
1. Interpretation of water quality and quantity data acquired from
routine long-term monitoring activities as well as that obtained
from short-term specific land use studies.
2. Input to hydrologic-hydraulic-water quality models.
In addition to meeting the above two needs of the Menomonee River
Pilot Watershed Study, the Land DMS was designed to be consistent with
the recommendation of the Ad Hoc Data Handling and Processing Work Group
of the PLUARG Task C Technical Committee.
Description of the System
The basic areal unit for storing, retrieving, analyzing and dis-
playing land data is a cell having a nominal area of 1.0 hectare (2.5
acres). The corners of each cell may be referenced to the State Plane
Coordinate System, to latitude and longitude, and to the Universal
Transverse Mercator System. The digital computer system—hardware and
software—needed to support the Land DMS is broken into four phases: the
input phase, the data management phase, the data base phase, and the
output phase. Under the input phase, data are entered into the Land DMS
-------
-------
60
on either magnetic tape, magnetic diskettes or punched cards. The second
or data management phase is composed of a set of computer programs that
perform contingency checks on the incoming data, provide for the main-
tenance and updating of the Land DMS, and perform analysis of the data
as requested by the user. The third or data base phase of the Land DMS is
the actual storage of the areal characteristics of each cell in a computer
file which is maintained on magnetic tape or on magnetic disc. The fourth
or output phase provides transfer of land data from the Land DMS to the
user in a variety of media including magnetic tape, punch cards, on-line
printer, and plotter.
Work Elements Completed since October 1975
1. A computer program having capability to convert cell corner
coordinants from the Wisconsin State Plane Coordinate System
to the Universal Transverse Mercator System was obtained from
the National Geodetic Survey and added to the data management
phase of the Land DMS.
2. The coding of soils data was completed for the watershed and
work was initiated on the coding of ground elevation data and
land use data by cell.
3. Work was initiated on the software additions needed to determine
the characteristics of the total area tributary to each of the
monitoring stations established in the watershed.
M-. At the request of study participants, the Land DMS was used to
prepare a tabular summary of soil types for each of the 244 sub-
basins in the watershed. In addition, the system was used to
prepare a 1" = 2000' scale watershed map showing dominant soil
type by cell.
Land Data Contained in the Land DMS
Appendix C Table 1 summarizes the status of land data within the Land
DMS. Nine data types have been coded for the entire watershed, and the
coding of two data types is in progress. Other land data types will be
added in response to the needs of the Menomonee River Pilot Watershed
Study.
-------
-------
61
Appendix C Table 1. Status of land data in the Land Data Management System
Status Type of coding
In Dominant Percent
Data type Completed progress characteristic of cell Other
1. Civil division x x
2. Sub-basins and
subwatersheds x
3. Wildlife habitat
» y
(with value ratings)
U. Woodland-wetlands
(with value ratings) x
5. Park and outdoor
recreation sites
6. Floodlands x x
7. Perennial streams x x
8. Conservancy, flood-
land and related x x
zoning
9. Soils x x
10. Ground elevation x x
11. Land use x x
-------
-------
APPENDIX D
-------
63
SPECIFIC LAND USE STUDIES
Introduction
The specific land use studies are an attempt to define more precisely
the quality and quantity of stormwater from homogeneous and/or predomi-
nant land use areas in the watershed. The study sites are representative
of the major land uses in the watershed, and data gathered at the
stations complement data gathered at the major river monitoring stations.
The data will be used to calibrate an overland flow model.
Study Sites
Construction of the sampling stations continued through the winter
months and is now complete at the sites listed in Appendix D Table 1.
The two residential sites, high density and developing, are part of the
major river monitoring system and have been operational since January
1975. The freeway right-of-way site is part of a U.S. Department of
Transportation study conducted by Envirex, Inc.—a Rexnord Company—in
Milwaukee. Arrangements have been made to receive the study data and
findings. Negotiations are presently underway with the Milwaukee County
Park Commission for establishment of sampling stations at the Greenfield
Park and Golf Course. The Greenfield site likely will contain two
sampling stations: one for the park and golf course drainage and another
for drainage from the Town of New Berlin, a medium density residential
area.
Equipment
Sealed bids were opened in December 1975 for the purchase of water-
level recorders and samplers. Low bids were submitted by Leupold and
Stevens, Inc., for type A model 71 water-level recorders and by the
Instrumentation Specialities Company for model 1680 water samplers with
high speed pumps.
The flow proportional circuitry of the 1680 samplers has been
modified to provide at least 50 ma. output to drive a relay/event marker
-------
-------
Appendix D Table 1. Completed specific land use study sites
Land use
Location
Freeway, right-of-way
Freeway, interchange
Transportat ion
Airport Tinnnerman Airport at W. Appleton Ave.,
Milwaukee
Highway 45 between Blue Mound and
Underwood Creek
Stadium Interchange (I-94--USH 11) near
Milwaukee County Stadium, Milwaukee
Commercial and Industrial
Brookfield Square Shopping Center at
N. Moorland, Brookfield
Allis Chalmers at S. 70th St., West Allis
Residential
Schoonmaker Creek at W. Milwaukee Ave.,
Wauwatosa
Underwood Parkway at Wrayburn Rd., Elm Grove
Noyes Creek at 91st St., Milwaukee
Retail and services
Manufacturing, heavy
High density
Low density
Developing
-------
-------
65
system which is attached to the level recorders and marks the time a
sample is taken.
The "1680" samplers have two actuation modes:
1. During spring snowmelt, when the stage rises slowly during the
day, the samplers are actuated, in the "Time" mode, by a micro
switch on the moving pen assembly of the water-level recorder.
2. During the remainder of the sampling season (rainstorm periods),
the samplers are actuated, in the "Flow" mode, by a reed switch
mounted opposite the float wheel of the water-level recorder.
Magnetics, mounted near the circumference of the float wheel,
close the switch and actuate the sampler for a predetermined
change in water level.
At the present time, work is continuing on the installation of the elec-
tronic switching devices.
Sampling and Laboratory Analysis
Laboratory space at the Water Chemistry Laboratory, University of
Wisconsin-Madison has been equipped to handle analysis of Group C
(inorganic and organic) parameters. Procurement of an atomic absorption
spectrophotometer with graphite furnace and a gas-liquid chromatograph is
underway, and an analytical chemist has been hired.
Preliminary grab sampling of runoff arising from snowmelt was started
on March 18, 1976, at selected specific land use study sites to scan for
the predominant Group C parameters (airport, Stadium interchange,
Brookfield Square Shopping Center and Allis Chalmers Corporation). An
attempt is being made to separate the suspended sediment from the water
and scan it for heavy metals and organics, since it plays a major role
in the transport of pollutants.
-------
-------
APPENDIX E
-------
67
BIOLOGICAL STUDIES
Objectives
A biological sampling program has been designed to provide the data
necessary for correlating biotic composition of the community with the
chemical and physical parameters being measured in the stream. This
program is designed to provide samples which are not necessarily repre-
sentative of the natural community at each site. Instead, the samples
will provide data useful for both intersite and intrasite comparisons
spanning the entire sampling period. The standardized artificial sub-
strate being used affords a suitable and uniform habitat for organisms
drifting down from the upstream reaches. The use of modified Hester-Dendy
samplers also provides control over habitat available for colonization
at sites differing greatly in their riverbed composition. In addition,
chemical analyses of the organisms for heavy metals and pesticides will
provide information on concentrations and locations of these substances
within the food chain.
Sampling Program
Five sites have been selected for biological monitoring. Four of
these sites coincide with continuous monitoring stations (413005, 683001,
413008, 673001) and one is an additional upstream station (683002). Each
station will be equipped with a floating structure from which are
suspended six modified Hester-Dendy samplers made of Masonite and
Conservation Webbing (Wards Natural Science Establishment, Inc.,
Rochester, N.Y.). Sampler colonization will be six weeks in length, with
a sampler overlap of four weeks. Two samplers will be removed and replaced
every two weeks from each site for the duration of the sampling period.
This continuous and rotating sampling program provides the data necessary
for multivariate analysis of species composition as related to the chemical
parameters being measured in-stream. Field sampling will continue from
early April 1976 to November 1976. All samples will be analyzed for macro-
invertebrates and periphyton community composition at the Water Resources
-------
-------
68
Center. The preserved specimens may be analyzed later for content of
heavy metals and pesticides.
Progress
To date, several methods of sampling have been tried and considered.
Artificial substrate sampling was chosen because of the control over an
important environmental factor which it provides. The modified Hester-
Dendy (Appendix E Fig. 1) was chosen for ease of handling and replicate
sampling.
Superstructure floats (Appendix E Fig. 1) are being built by the
machine shop of the Zoology Research Department at the University of
Wisconsin and the Masonite-Conservation Webbing samplers are being
assembled by hourly helpers at the Water Resources Center.
-------
-------
69
Float
SUPERSTRUCTURE
Cable
, (( (f (
'J 4- f*25cm>l
25cm f f
* I r*
fi fi 1
f )) )) )
t 1 i j
|\ g § |
)
i
SAMPLLF
/ \ Sampl
Masonite
'". onr5 er vat Ion
webbing
cope
Appendix E Fig. 1. Modified Hester--Dendy sampler
-------
APPENDIX F
-------
71
ATMOSPHERIC MONITORING PROGRAM
Introduction
The input of atmospheric substances, in the form of wet and dry
deposition, may play an important role in the movement of materials
through both terrestrial and aquatic ecosystems (Biggs et al., 1973;
Andren et al., 1975). The Menomonee River watershed, which is presently
being hydrologically calibrated, provides an ideal site for studying the
influence of atmospheric processes in the cycling of elements. A water-
shed can also act as a source of atmospheric materials to surrounding
areas via resuspension of soil dust by wind and anthropogenic activities.
The Menomonee River watershed may thus act as both a source and a sink
for atmospheric substances. The primary objectives of this program are:
1) quantification of elemental input into the watershed through the
atmosphere; 2) identification of atmospheric elemental sources; 3) deter-
mination of the relative contribution of different land use practices
to the total weight and chemical composition of the elements measured
in the atmosphere above the Menomonee River watershed; 4) quantification
of watershed-derived atmospheric elements into Lake Michigan; and
5) evaluation of methods available for determining the effect of elemen-
tal atmospheric input on the chemical composition of the Menomonee River.
Prior to rain and aerosol collections in the Menomonee River basin
we have evaluated collection methodologies for aerosols. This report
will deal with the evaluation of high-volume air samplers and inherent
sampling difficulties. Specifically, it is useful to know the sources
and magnitude of errors involved in: 1) filter paper weighing (using
Whatman-1!! filters), 2) high-volume sampler performance, 3) sample
division, and 4) sample digestion and analytical procedures.
Weighing
Two types of filter papers are generally available for high-volume
air sampling. One—the glass fiber filter—is used widely by many
researchers including the U.S. Environmental Protection Agency (EPA) for
total suspended particulate (TSP) measurements. However, glass fiber
-------
-------
72
filters contain a large amount of impurities, making them unsuitable for
trace element analysis. The cellulose Whatman-1*! filter contains fewer
impurities and has been selected for this study.
Unfortunately, paper filters are extremely hydroscopic. An 8"xlO"
sheet indicates a change in weight of up to 1% with every 2% change in
relative humidity at room temperature. Consequently, it is critical to
weigh each filter before and after collection at the same humidity.
The USDA Forest Products Laboratory, Madison, Wisconsin, has pro-
vided weighing facilities for the project. These facilities include a
large room which is temperature- and relative humidity-controlled to
±0.5% at 50% relative humidity. An equilibration chamber for the filter
paper was built from plexiglass. Inside, the paper filters equilibrate
for 24 hours before the tare weighing and again for 24 hours before
weighing of the collected material.
The chamber is virtually airtight except for two holes cut in the
ends covered by glass fiber filters. A vacuum pump (with exhaust filters)
draws air out of the chamber; consequently, humidified, filtered air is
constantly entering the end portals and flowing across the filter paper
for the 24-hour period. The scale used reads to ±0.1 mg.
Calibration of High-Volume Air Sampler
Each sampler has a flow gauge (ftVmin) at the exhaust orifice below
the motor. The air flow gauge must be calibrated to standard conditions,
namely, Standard Cubic Feet/Minute (SCFM) normalized to a pressure of
29.92 in. of Hg at 298°K. Using a National Air Surveillance Network (NASN)
type orifice calibration assembly, the flow rate in SCFM can be obtained
as follows.
First, the orifice calibrator is screwed onto the intake of the
sampler. A flexible hose is connected to the pressure tap on the side
of the orifice to a manometer which reads inches of water. Under a given
set of temperature and pressure conditions, flow rate is proportional to
the change in inches of water.
-------
-------
73
Q.nd
where Qind ^s true flow rate at a given temperature and pressure
Ap is inches of water difference on manometer
k is 17.66 for this particular orifice calibrator
The true flow can be corrected to SCFM by the formula:
Qs = APa/Ps x (Ts/Ta) x
where Qs is flow rate corrected to SCFM
Ps is 29.92 inches of Hg
Ts is 298°K
Pa is ambient pressure in inches of Hg
Ta is ambient temperature in °K
or simplified to:
Qs = /(Pa/Ta) x 3.156 x Qj.nd
To compare the flow gauge to SCFM it must also be corrected for
temperature and pressure variations from standard.
Is = /(Pa/Ta) x 3.156 x Ia
where Ia is flow gauge reading at given temperature and pressure
Is is flow gauge reading corrected for temperature and
pressure
Pa is temperature at pump exhaust.
Finally, Is is related to Qs graphically by assuming the relationship
is linear. Each flow gauge must be calibrated in this manner and recali-
brated with the installation of each new set of brushes for the motor.
Points for the graph of Is against Qs are obtained by varying the flow
rate in one of two ways. A continuously variable rectifier in the line
between the vacuum pump motor and the power source can change motor speed
and, thus, flow rate. Resistance plates can be placed within the orifice
calibrator above the pump intake to slow the flow. It was thought that
-------
-------
the latter method best simulated clogging in the field and should be used
to obtain points for the calibration curve. However, for each pump two
calibration curves were obtained experimentally: one at about 90 volts
and one at about 66 volts. These two curves agreed well when extrapolated
to common points. Calibration of all seven high-volume air samplers has
been completed.
Collection
One important test was to compare simultaneously running samplers
and note the collection by each. For this experiment, four samplers were
anchored 10 feet apart in a square formation atop the Helen White building,
University of Wisconsin-Madison. At this location there are no obstruc-
tions to the wind from any direction. Each sampler was connected to a
rectifier.
Appendix F Table 1 compares TSP measurements for three sampling
periods, and Appendix F Table 2 represents a similar experiment run by
Neustadter et al. (1975).
Some of the differences in procedure which might account for the
greater variation between samples in the Helen White experiment include a
larger sample size and the fact that Neustadter was able to record flow
rate continuously whereas the Helen White data represent periodic flow
measurements averaged over time.
Flow rates actually decreased significantly and in different magni-
tudes for each sampler over the sampling period. The reason for the
decrease was thought to be primarily due to clogging. It was also noted
that, in general, those samplers measuring the highest TSP values also
experienced the greatest drop in flow rate during the experiment. If the
decrease in flow rate is due to clogging, and clogging is a function of
the concentration of TSP, then the measured differences in TSP may be,
in part, real. Although the samplers were only 10 feet apart, there was
no basis from the numbers to suspect effects of one sampler upon another.
Wind directions were noted during each run, but downwind samplers did
not generally measure less than upwind samplers.
-------
75
Appendix F Table 1. Factorial experimental TSP data using Whatman-41
Date
(1976)
Sampling
period
High volume sampler pump no.
1007
1009
2968
2970
95%
limits
26 Jan.
09 Feb.
13 Feb.
Mean
30.0 hrs
36.0 hrs
66.5 hrs
30.2
32.3
32.9
31.8
32.9 30.2
31.7
29.6
31.4
32.5
36.4
33.0
27.5
27.5
31.0
28.7
±11.61%
±12.09%
±14.42%
From Stolzenburg (1976)
Appendix F Table 2. Factorial experimental TSP data using Whatman-41
Date
(1972)
24 Jan.
25 Jan.
26 Jan.
27 Jan.
Mean
Sampling
period
24 hrs
24 hrs
24 hrs
24 hrs
High volume
14
61.3
47.7
69.0
45.3
55.8
29
TO
62.7
52.4
71.3
43.5
57.5
sampler serial
63
P in ug/
66.0
56.5
74.2
50,5
61.8
68
m3
63.2
60.0
72.0
47.2
60.6
no.
71
67.0
55.2
64.8
47.2
58.5
95%
limits
±4.62%
±10.56%
±6.32%
±6.93%
From H. E. Neustadter et al. (1975)
-------
-------
76
Division of Sample
The analysis program will include halides of interest such as Cl
and Br. Most wet digestion procedures would produce a loss of these
volatile elements. The 8"xlO" filter paper must be divided, therefore,
prior to dissolution so that a portion of it can be analyzed directly
by neutron activation analysis for those elements plus some rare earths.
If the filter is cut after collection there would be no direct means
for measuring the tare weight. By knowing the area of the cutout portion,
the tare weight can be indirectly calculated from the tare of the whole
piece. Similarly, the collected weight of a portion could be calculated.
To assure equal and known portions, a circular die was made from stain-
less steel tubing.
Two questions arise concerning this procedure: 1) Is the filter
paper uniform in terms of weight per unit area? and 2) Does collection
occur uniformly across the filter sheet? The first deals with the pre-
cision of the calculated tare weight, and the second with the precision
of the calculated collected weight.
To answer the first question, a whole 8"xlO" filter was weighed and
then eight circles were cut out of it with the die. An area calculation
of the weight of one circle equalled the average measured weight of the
eight circles to the fourth decimal place. It was calculated that 95%
of the time a randomly cut circle from this filter paper would fall
within ±0.677 mg of the average weight. The range of circle weights
was from 0.0809 g to 0.0829 g. The data appear below.
Variation of Whatman-m weight - die cut
Actual Calculated
Die-cut area =1.56 in2 Whole area = 80 in2
Die-cut weight: 0.0812 g Whole weight = 4.1945 g
0.0822 g .2
0.0816 g x U.1945 g = 0.0818 g
0.0826 g 30 in2
0.0809 g
0.0817 g
0.0829 g
0.0815 g
Mean = 0.0818 g
95% limits: ±0.677 mg (±0.827%)
-------
-------
77
Initially this error appears small. Unfortunately, the amount of
collected material on a circle may be relatively small (5.30 mg) compared
to the weight of the filter itself (81.8 mg). Therefore, the error on
the calculated collected weight may be as large as ±13%. This error can
be reduced significantly (to about 1%) if an effectively larger portion
of the filter paper (four circles) is cut out.
Resulting error to weight of collected material
TSP Annual Mean (NW part of Menomonee River basin) ^ 35 ug/m3
Flow Rate ^ 30 ftVmin
Time * 5 days
0 21U1 e
Total Collected Wt. = Flow Rate x Time x Cone. = — =•
(63 in2)*
Assume even distribution of collected material:
Collected Wt. on (1) Circle = 1>56 ln x 0.2141 g = 5.30 mg
63 in2
Error in Collected Wt. _ ±0.677 mg
for (1) Circle ~ 5.30 mg
x 100 =
±12.1
Error in Collected Wt. _ ±0.2296 mg inn _
for the Sum of (U) Circles = 4(5.30) mg
*Actual collecting surface of filter paper.
The error discussed above may not be totally attributable to filter
uniformity variations. Other factors contributing to the differences
include nonuniform die cuts and humidity changes (even at the ±0.5% level)
between measurements.
Probably of more concern is the question of collection uniformity.
The analytical scheme does not involve weighing the die-cut circles.
Instead, that weight will be calculated by area after the whole filter
paper is weighed. If the tare weight varies, the result should be insig-
nificant as long as the blank values for analyzed elements are well below
-------
-------
78
the collected values. However, if collection is nonuniform, variations
will have a large effect on the calculated "true" value of the concen-
trations of the elements in the air. Furthermore, the tare value errors
and reweighing errors would be reintroduced if the circles had to be
weighed.
Collection uniformity is difficult to assess. Weighing die-cut
circles is not a good method to use since there are so many other factors
contributing to the variation. It may be possible, however, to determine
an overall error which would include contributions from filter nonuniformity,
collection nonuniformity, humidity changes, balance sensitivity, and
analytical error, if elemental analysis is carried out on the circles.
An experiment was undertaken with a set of four filters from one
run. Eight circles were cut from each filter, four from the center and
four from the corners.
For each filter, first the four corner circles were combined and
analyzed for Na, Ca and Mg; and, secondly, the center circles were com-
bined and similarly analyzed. Absorbances are compared between the
inside and outside sets in Appendix F Table 3. Absorbances are propor-
tional to concentrations, so the percent variations are expressions of
the final elemental concentrations in air due to all possible error
sources listed above. These variations are topped by an 18.2% difference
in the sets from filter no. 4 (Appendix F Table 3) for the Na analysis.
The initial trend of higher concentrations toward the center of the
filter paper evidenced in the Na analysis seems to be reversed in the
Ca analysis. An arbitrary goal before field sampling might be to reduce
the final error to below 15%, recalling that this error represents
reproducibility within one collected filter paper, not between samplers.
Digestion and Analysis
Preliminary digestion and analytical procedures are outlined below.
-------
-------
79
Appendix F Table 3. Absorbance comparisons between
materials deposited on the centers
and corners of Whatman-41 filter
papers for Na, Ca and Mg
Filters
Na
Ca
Mg
Center
Corner
% Diff .
Center
Corner
% Diff.
Center
Corner
% Diff.
1
0.3439
0.3363
2.26
0.1073
0.1180
9.97
0.1752
0.1878
7.19
2
0.3233
0.3107
4.06
0.1051
0.1073
2.09
0.1752
0.1733
1.10
3
0.5229
0.4976
5.08
0.0799
0.0820
2.63
0.1772
0.1691
4.79
4
0.5272
0.4461
18.2
0.0721
0.0655
10.1
0.1580
0.1580
0.0
-------
-------
80
Divide sample
'-»• Neutron activation analysis (direct analysis)
Flameless atomic absorption (LTA and HNOs, HF dissolution).
Flame atomic absorption
Digestion
Mixture of HN03 6 HC1CU (2:3) + heat -•»• Silica residue
with. . .K, Na, Kg, Ga, ?, ?
Alternatives:
1) Analyze directly
2) Filter
3) HF addition (LaF3 ppt)
Dilute
DDW + LiCl or CsCl + La reagent
Analyze
Optimize operating conditions
Flameless atomic absorption is intended to be the main analytical
tool. Currently, only a flame atomic absorption unit is available and
procedures for its use are being worked out. A mixture of HNOs and
HClOi* (2:3) is added to filter paper circles in a quartz digestion flask.
The mixture is heated slowly at first to decompose the easily oxidizable
materials. Then, at a higher temperature the HC10i» is brought to boiling
and the solution clears. This procedure does not remove solid silicates
which may contain significant quantities of K and Na and possibly some
Mg and Ca.
Analyzing the samples directly might result in interferences due to
scattering from undecomposed silicates. Filtration might remove soluble
elements of interest or those associated with the silica. Addition of HF
should solubilize the silicon to SiF^. However, excess Fl~ will react
with the lanthanum reagent (added for control of P0i» 3, Al, and Si
-------
-------
81
interferences in the determination of Ca and Mg) to form a precipitate
LaFs. Investigation is underway to circumvent these problems. lonization
control is effected by the addition of LiCl or CsCl.
Summary
A literature review indicated that little attention has been paid
to minimizing errors involved in gathering aerosol data. In addition to
considering errors inherent in the physical manipulation of filter
papers, the high-volume collection performance, and subsequent chemical
analysis, one has to consider the collection efficiency of the filter
papers for various size classes of aerosols. Lockhart et al. (1963)
reported that Whatman-til cellulose filter papers collect aerosols down
to 0.3 Um with a greater than 99% efficiency. The collection efficiency
for particles less than 0.3 Um is presently uncertain. Furthermore, there
is a lack of experimental work on the collection efficiency of both
Whatman-Hi and glass fiber filters and the change in relative ambient
humidity.
References
Andren, A. W., S. L. Lindberg, and L. Bate. 1975. Atmospheric input and
geochemical cycling of selected trace elements in Walker Branch
watershed. Publication No. 728. Oak Ridge National Laboratory,
Environmental Sciences Division, Oak Ridge, Tennessee. 68 p.
Biggs, R. B., J. C. Miller, M. J. Otley, and C. L. Shields. 1973. Trace
metals in several Delaware watersheds. Final Report. Water Resources
Center, University of Delaware, Newark, Del. 47 p.
Lockhart, L. B., R. L. Patterson, and W. L. Anderson. 1963. Character-
istics of air filter media used for monitoring airborne radioactivity.
Report 6054. Naval Research Laboratory, Washington, D.C.
Neustadter, H. E., S. M. Sidik, R. B. King, J. S. Fordyce, and J. C. Burr.
1975. The use of Whatman-41 filters for high-volume air sampling.
Atm. Env. jh 101-109.
Stolzenburg, T. R. 1976. Unpublished data. Water Chemistry Program,
University of Wisconsin, Madison, Wis.
-------
-------
APPENDIX G
-------
83
REMOTE SENSING PROGRAM
Introduction
This is a joint project between The Pennsylvania State University,
Office for Remote Sensing of Earth Resources (ORSER), and the University
of Wisconsin-Madison, Water Resources Center and the Civil Engineering
Department, to determine the applicability of remote sensing data in
hydrological modeling efforts.
Objectives
The objectives of this study are twofold:
1. To develop techniques for converting aerial imagery into digital
representation which can be interpreted by a computer. This
work is being undertaken at the University of Wisconsin-Madison.
2. To prepare, at The Pennsylvania State University, land cover
maps and data summaries using automatic data processing pro-
cedures on digitized color infrared aircraft photography and
to compare them with land use maps and data summaries prepared
by the Southeastern Wisconsin Regional Planning Commission
(SEWRPC).
Methods and Materials
Data
Low altitude 70 mm color infrared photography is being acquired
with the cooperation of the Wisconsin Department of Natural Resources
(DNR). The imagery is taken by a two-camera system mounted in the DNR's
DC-3 aircraft. Both color and color infrared imagery (Kodak types 5257
and 2443, respectively) are taken on each mission. Photographic missions
have been flown approximately once per month during the summer and fall
months of 1975. Photographic coverage of the Menomonee River Basin will
continue to be flown in the spring and summer months of 1976. Imagery
from each flight is calibrated and developed at Precision Photo Labora-
tories in Dayton, Ohio.
-------
-------
The photographic imagery is converted into a digital representation
by using an Optronics P-1000 scanning microdensitometer. This instrument
is capable of measuring the film density everywhere on a 4.5" x 6.3"
transparency every 25 (0.025mm) to 100 (O.lmm) microns. The instrument
will create a computer-compatible magnetic tape with numerical informa-
tion that is related to the film density on the scanned transparency.
This digital tape can be processed on a computer.
The research this year has been directed at improving the density
measurements from the microdensitometer and on preprocessing of the
numbers from the microdensitometer to make possible the interpretation
at ORSER.
Since the films used are multi-emulsion films, some type of color
separation process must be employed in the microdensitometer. From
previous research, it was decided that narrow-band interference filters
must be inserted in the optical path when density measurements were made
with the microdensitometer. Since the filters decreased the intensity
of the light in the system, operational problems were encountered, but
an acceptable method of operating the microdensitometer was found.
Further work on improvement of the operation of the microdensitometer
will be continued.
Once density measurements have been acquired from the microdensi-
tometer, calibration procedures must be employed. Because of film
processing and other problems, film density is not proportional to the
energy incident on the film when it was exposed. Procedures have been
developed to correct the film measurement for these nonlinear effects,
Programs were written to enable this correction to be performed on the
University of Wisconsin's Univac 1110 computer.
Another problem in the analysis of aerial imagery is the radiometric
changes on the film due to the geometric properties of the lens. Due to
geometric considerations, an image is always darker at an edge relative
to its center. Density measurements must be corrected for this effect
before digital interpretation. Work is progressing on this correction.
Corrections for atmospheric absorption, scattering, and uneven scene
illumination are also being considered.
-------
85
Thus far, two scenes near Menomonee Falls in Waukesha County,
Wisconsin, dated 16 July and 6 October 1975, have been digitized in the
wavelength intervals 4450-4550 A (red), 5450-5550 A (blue) and 6450-6550 A
(green) using a 100 micron spot size representing a pixel size of
7.5 meters on a side. Approximately 377 hectares (930 acres) of the
July 16 data and 334 hectares (825 acres) of the October 6 data were
digitized.
Analysis of the digital tape is undertaken utilizing The Pennsylvania
State University's IBM System 370, Model 168 computer and ORSER's programs
for analyzing remote sensing data (Borden et al., 1975). Land cover maps
were developed and output on conventional paper printouts was provided
as well as being displayed on ORSER's RAMTEK color cathode ray tube
system.
Ground truth information consisted of 35 mm color slides of the
digitized photography, U.S.G.S. 7.5 minute quadrangle topographic maps,
Soil Conservation Service soil survey maps (USDA, 1971), and telephone
communications with University of Wisconsin and SEWRPC personnel.
Analysis procedures
The first step in the analysis procedure was to run the ORSER program
called SUBSET. Each wavelength interval or channel comprising a separate
file on the source tape was run using SUBSET which reformatted the data
into the standard ORSER format. The three channels for the July 16 data
and the three channels for the October 6 data could not be merged due to
scale differences between the two digitized photographs. Therefore, each
date was analyzed separately.
The initial display for each date was produced using the NMAP program
which yields a brightness map similar to a grey-scale map. These maps
are useful for initial target location and verification of general location.
The NMAP program requires no a priori knowledge of target spectral signa-
tures or other characteristics.
The UMAP program is employed to identify areas of local spectral
uniformity. Each element is compared with its near neighbors, using the
euclidean distance between spectral signatures as the measure of
-------
-------
86
similarity or dissimilarity. One or more categories of uniformity can be
mapped according to distances specified by the user. All elements with
distances from their neighbors smaller than those specified are mapped
as uniform, while those with distances greater than specified are mapped
as contrasts. The output shows the pattern of uniformity and contrasts
from which the user can designate coordinates of training areas for input
to supervised classifying routines. High contrast boundaries between
uniform areas may also be determined.
Signatures and associated statistics are obtained from the STATS
program, which computes the multivariate statistics for one or more
training areas obtained from UMAP or a similar output. The user desig-
nates, for each identifiable category, a training area by line and element
coordinates, and the program computes the statistics for all of the data
which fall within the boundaries. The mean and standard deviation
vectors for each category are calculated and the correlation and variance-
covariance matrices are computed. If desired, the eigenvalues and eigen-
vectors of these matrices, and histograms for selected channels, can also
be computed.
Using statistics from the STATS program, for the targets of interest,
supervised classification and mapping can be done with the CLASS program.
By specifying either the ACLASS or DCLASS option, one of two operational
modes can be selected. Under the ACLASS option, the data are normalized
and are classified according to their angle of separation in a multi-
dimensional geometric sense, with classification made into the category
for which the angle is smallest. Under the DCLASS option, the data are
not normalized and are classified according to their euclidean distance
of separation from each of the categories. The output of these programs
is a digital character map with each category of classification repre-
sented by a unique symbol assigned by the user.
Frequently, it may not be possible to accurately define or locate
training areas, or a sample target may not be of sufficient size or area
to lend itself to categorization using the supervised STATS-CLASS sequence.
Such targets may be linear features, such as streams, or a series of small
scattered features which are not large enough to be represented as
-------
87
uniform areas by UMAP. In such cases, these areas are defined for
analysis by an unsupervised classifier, CLUS, which develops its own set
of spectral signatures and statistics using a clustering algorithm, out-
putting a character map based on these parameters. The spectral signa-
tures developed using CLUS can then be used in the CLASS program along
with the signatures developed from the STATS program.
Classification maps can be produced on conventional line printers
or displayed on a color cathode ray tube device.
Results and Discussion
July 16 data
To date, 29 signatures have been developed for the July 16 data
representing five categories of land cover (Appendix G Table 1).
Appendix G Table 1 lists the category name, number of signatures required
for each category and the percentage of the area classified as that
category. More than one signature is usually required for each category
to account for within-category variability. For example, several types
of vegetation may be apparent in a scene with each one having a unique
signature and, if the investigator is only interested in vegetation in
general, he combines them into a single category by assigning them the
same symbol or color. Approximately 93 percent of the area was classified.
Grassland and pastureland represent a variety of vegetation and required
many more signatures than other categories. It is suspected that these
areas are no longer in use as commercial agricultural lands but merely
are in a state of vegetative cover awaiting industrial and/or residential
development. As a result, management of the vegetation is probably
minimal, which allows natural soil fertility variations to manifest them-
selves in the vigor and health of the vegetation. The various physiological
conditions of the vegetation result in more variable reflectance values
than would normally occur with crops grown under an adequate commercial
agricultural management program.
Some confusion was apparent between the categories of roof tops,
roads and highways, and disturbed lands because of their apparent simi-
-------
88
Appendix G Table 1. Classification results of the digitized photog-
raphy collected on July 16, 1975
No. of Area mapped,
Category signatures %
Grassland and pastureland 16
Disturbed land 2
Roof tops 4
Roads and highways 3
Trees 4
Unclassified*
74
3
6
8
2
7
Area, Ha
279.0
11.3
22.6
30.2
7.5
26.4
*Since development of land cover maps is an iterative procedure,
unclassified areas are those areas which have not been assigned a
specific land cover category using the signatures developed to date.
-------
89
larities in reflectance. Only slight confusion between trees and other
categories was apparent. Tree shadows did have a tendency to moderate
the specificity of the tree signatures, resulting in some confusion with
other vegetation.
Residential areas consist of a highly variable group of small
targets, such as roof tops, trees, lawns, and streets. Thus, because
it is difficult to accurately delineate categories within residential
areas, it may be necessary to use a spot size smaller than the 100 micron
spot size presently used on the digitizer.
October 6 data
A total of 39 signatures representing six categories of land cover
has thus far been developed for the October 6 data (Appendix G Table 2).
Approximately 80 percent of the area was classified. Appendix G Fig. 1
shows the classification results of the October 6 data as displayed on
the color cathode ray tube system.
The yellow area in the upper right of Appendix G Fig. 1 represents
the disturbed land category which is classified as clayey land by the
Soil Conservation Service (Soil Survey Staff, 1971). Clayey land is
defined as a miscellaneous land type that consists of fill areas and of
cut or borrow areas occurring mainly within or near cities or towns and
areas used for housing developments or related purposes. The yellow
areas of Appendix G Fig. 1 (disturbed land category) are cut or borrow
areas where the entire solum of the soil has been removed, exposing the
raw underlying material. The underlying material is variable, consisting
of silty clay loam glacial till that contains pockets of loamy or silty
material and is generally compacted and relatively impervious. Due to
the variable reflectance of these areas, several signatures were required
to adequately classify them.
The blue areas and the pale green areas of Appendix G Fig. 1 repre-
sent the roof tops and roads-highways categories. Some confusion between
these two categories is readily apparent. The purple and brown areas of
Appendix G Fig. 1 represent the trees and grassland-pastureland categories,
respectively. Considerably more areas are misclassified as trees on this
-------
90
Appendix G Table 2. Classification results of the digitized photog-
raphy collected on October 6, 1975
No. of Area mapped,
Category signatures %
Grassland and pas tur eland 22
Disturbed land 4
Roof tops 5
Roads and highways M-
Trees 3
Parking lots 1
Unclassified*
50
10
6
5
6
3
20
Area, Ha
167.0
33. U
20.0
16.7
20.0
10.0
66.8
*Since development of land cover maps is an iterative procedure,
unclassified areas are those areas which have not been assigned a
specific land cover category using the signatures developed to date.
-------
-------
91
Appendix G Fig. 1.
Classification results of the October 6 data as
displayed on the RAMTEK color monitor system.
Appendix G Fig. 2.
Categories of roads and highways (light green),
roof tops (brown), and disturbed lands (light
blue), as displayed on the RAMTEK color monitor
system.
-------
92
date than on the land cover maps developed using the July 16 data. This
was caused by some of the trees taking on their autumn colors, resulting
in tree signatures with a larger variation than the tree signatures
obtained using the July 16 data.
The residential area in the lower left of Appendix G Fig. 1 is in-
adequately classified since many of the targets are quite small and highly
variable.
Appendix G Fig. 2 is a display on the color cathode ray tube device
depicting those categories which could be considered impervious surfaces.
The light blue areas represent the disturbed land category, light green
represents roads and highways, and the brown areas represent the combina-
tion of the roof tops and parking lot categories.
Comparisons of the July 16 and October 6 data
Comparison of the area mapped as disturbed land, using the July 16
data (3% as shown in Appendix G Table 1) and using the October 6 data
(10% as shown in Appendix G Table 2), shows a distinct temporal change.
In the interval between July 16 and October 6, a large area was disturbed
by removing the entire solum of the soil (upper right of Appendix G
Figs. 1 and 2) in preparation for development. This illustrates the
valuable temporal change detection for which remote sensing is particu-
larly well-suited.
In general, the October 6 data were superior to the July 16 data
for classification of roof tops, roads and highways, and disturbed
lands, whereas the July 16 data were more suited for detection of trees
since in October the trees were beginning to take on their autumn colors.
Both data sets appeared equally suited to the classification of grass-
land and pastureland.
The 20% unclassified area of the October 6 data (versus 7% for the
July 16 data) was due mostly to additional areas of vegetation for which
signatures have yet to be developed. Thus, major differences between
Appendix G Tables 1 and 2 can be attributed to temporal changes.
-------
-------
93
References
Borden, F. Y., D. N. Applegate, B. J. Turner, H. M. Lachowski, and
J. R. Hootsy. 1975. Satellite and Aircraft Multispectral Scanner
Digital Data Users Manual. ORSER-SSEL Technical Report 10-75.
Office for Remote Sensing of Earth Resources, The Pennsylvania
State University, University Park, Pa.
U.S. Department of Agriculture (USDA), Soil Conservation Service.
1971. Soil Survey of Milwaukee and Waukesha Counties, Wisconsin.
U.S. Government Printing Office.
-------
APPENDIX H
-------
95
LAND USE-WATER QUALITY MODELING
This appendix is divided into four sections which describe the
progress to date by the modeling participants. Sections I to III address
the modeling effort as it pertains to meeting objective 3 of the Menomonee
River Pilot Watershed Study, i.e., developing an integrative and pre-
dictive capability with respect to the impact of land use on surface
water quality. Section IV is an evaluation of an empirical model
relating runoff quality to watershed characteristics. This model is
likely to be useful in watershed/water quality studies where extensive
financial and time support is not available and where the detail being
included in the Pilot Watershed Studies is not essential. Compared to
the other modeling efforts, this model does not require a great amount
of data for calibration.
I. Evaluation of a Continuous Simulation Model
Since October 1975, work has continued on the evaluation of an
existing continuous simulation model to determine if that model could—
with modification—satisfy objective 3 of the Menomonee River Pilot
Watershed Study which calls for developing an integrative and predictive
capability with respect to the impact of land use on surface water quality.
The overall approach in this model evaluation is to determine the effec-
tiveness of the model in reproducing historic hydrologic-hydraulic data
and then to determine model capability with respect to water quality
phenomena. This procedure is based on the premise that successful water
quality simulation is contingent upon effective hydrologic-hydraulic
modeling since runoff from the land and flow in the streams provide the
transport mechanisms for water quality constituents.
Model description
The digital computer model selected for evaluation is based on a
hydrologic-hydraulic model that originated at Stanford University in the
early 1960!s and is now available—with a water quality feature—from the
-------
-------
96
consulting firm of Hydrocomp, Inc. The model continuously simulates
hydrologic-hydraulic and water quality processes for an indefinite period
of time in response to a full spectrum of meteorological conditions. The
model consists of three submodels: the hydrologic submodel, the hydraulic
submodel, and the water quality submodel. The principal function of the
hydrologic submodel is to determine the volume and temporal distribution
of runoff from the land to the stream system. The function of the hydraulic
submodel is to accept as input the runoff from the land surface as produced
by the hydrologic submodel, to aggregate it, and to route it through the
stream system, thereby producing a continuous series of discharge values
at predetermined locations along the surface water system of the watershed.
The water quality submodel simulates the time-varying concentration of
water quality indicators throughout the surface water system. Operating
on a reach-by-reach basis,, this submodel continuously determines water
quality as a function of reach inflow and outflow, dilution, and bio-
chemical processes.
Data base development
Data base development consists of the acquisition, verification and
coding of data needed to operate, calibrate and apply the model. The data
base consists of a large5 readily accessible file of information sub-
divided into five distinct categories: meteorological data, land data,
channel data, diffuse source data, and point source data. Work had been
completed on the first two types of data in the data base at the time of
the October 1975 semi-annual report, and since that time data base
development efforts have concentrated on completing channel data and on
developing diffuse source and point source data.
Initial water quality submodel calibration runs
Calibration consists of comparing simulation results with historic
fact and, if significant differences occur, making parameter adjustments
so as to tailor the model to the natural and man-made features of the
watershed. Since the October 1975 semi-annual report, a successful
-------
-------
97
calibration of the hydrologic submodel and the hydraulic submodel was
accomplished, and calibration efforts have concentrated primarily on
achieving a preliminary calibration of the water quality submodel.
Inasmuch as the water quality-quantity monitoring data being obtained
under the Menomonee River Pilot Watershed Study are not yet available, the
initial calibration of the water quality submodel was based on field data
obtained as a result of three 24-hour watershed-wide field surveys carried
out under the Menomonee River Watershed Planning Program of the South-
eastern Wisconsin Regional Planning Commission. In each of these surveys,
streamflow measurements were made at five locations on the stream system,
while physical, chemical, and biological quality indicators were measured
at 17 stream sampling sites. In addition, the surveys involved the conduct
of water quality analyses on the effluent from five municipal sewage treat-
ment plants and two industrial facilities, and on the runoff from four
watershed sub-basins, each exhibiting a different type of land use. One
2'4-hour synoptic water quality survey was conducted during the spring,
on April 4-5, 1973, and two during the summer, on July 18-19, 1973 and
on August 6-7, 1974.
For each of the three synoptic surveys, the calibration process was
initiated by concentrating on the station farthest upstream in the water-
shed and achieving an acceptable correlation between the observed water
quality at that location and the results obtained with the water quality
submodel. After achieving a successful calibration with emphasis on six
parameters—temperature, dissolved oxygen, orthophosphate, nitrogen forms,
fecal coliforms, and carbonaceous biochemical oxygen demand—the calibra-
tion effort was then focused on the next downstream station. This process
of calibration at successive stations down through the watershed was
continued until a watershed-wide calibration was achieved with data from
the first survey. The calibration procedure was initiated with the one
spring event, after which summer survey data were used to complete the
initial calibration of the water quality submodel. An example of the
results obtained with the water quality submodel calibration are presented
in Appendix H Fig. 1, in the form of a graphical comparison of recorded
and simulated water temperatures at four locations in the watershed during
the April 1-5, 1973 synoptic survey.
-------
-------
98
Little Menononee River
Srstior. MN-7
7 i ; •
6 \-
r ,,
4 I-
o •-
-J L.™JL_™L
O 3 6 9 12 55 18
Kenononee River Reach #52
Station KN-7B
7r~~
i
fck
T
C 3 6 9 !2 '5 18 21 24 3 6
ir;^ (hours)
Underwood Creek Reach #57
Station MN-8
O
O
QJ
3
E 2
0)
1 i J
I 1
O 3 6 9 12 15 18 21 24 J 6
4/4/73
••>!<• 4/S/73
Time (hours)
Merjoroonee River Reach
Station KN-10
o
CJ
ft
E 2f-
Q)
A .. A
I I 1
J I
03 6 9 12 15 18 21 24 3 6
- 4/4/73
Time (hours)
Source: SEWRPC
Appendix H Fig. 1,
Legend: & Recorded temperature
• Simulated temperature
Comparison of recorded and simulated temperatures
for the Menomonee River watershed for April 4, 1973.
-------
99
Model selection
Participants in the Menomonee River Pilot Watershed Study have tenta-
tively selected the Hydrocomp Model as the basic simulation tool for the
study, recognizing that some supplementary modeling may be required to
meet the study objectives. This selection was based on consideration of
available hydrologic, hydraulic and water quality models including—but
not necessarily limited to—the Hydrocomp Model9 the Wisconsin Hydrologic
Transport Model (WHTM), the U.S. EPA Storm Water Management Model (SWMM),
the Corps of Engineers Storage, Treatment and Overflow Model (STORM), and
experience with the Hydrocomp Model, SWMM, STORM, and a new model called
LANDRUN. (See Section II for further discussion.)
Work elements planned
Modeling efforts in the immediate future will concentrate on refining
the above initial calibration of the Hydrocomp Model and on identifying
and proceeding with the necessary supplemental model development. Moni-
toring data obtained under the pilot study will soon be readily available
for use in this refined calibration effort. The use of that data will
permit a significant improvement in the calibration, inasmuch as the
monitoring data cover relatively long periods of time and include exten-
sive streamflow measurements as well as data on water quality parameters.
-------
100
II. Land Use-Water Quality Models
Since the beginning of Marquette University's participation (March,
1975) in the Menomonee River Pilot Watershed Study, the major effort has
been focused on an evaluation of existing land use-water quality models
and development of a land use-water quality model which would satisfy the
objectives of the program.
Evaluation of existing models
From the variety of models available in the United States for modeling
runoff from urban and nonurban areas3 only a few include pollutant trans-
port. Most of the available models are either general hydrological models
estimating runoff quantity from rainfall excess or models that are used
for the design of stormwater overflows in sanitary sewers.
The following models were evaluated by the Marquette University team
as carrier models for the land use-water quality modeling:
"STORM"
This program represents a method of analysis for estimating the
quantity and quality of runoff from small, primarily urban, watersheds.
Nonurban areas may also be considered. Land surface erosion for urban and
nonurban areas is computed in addition to the basic quality parameters of
suspended and settleable solids, BOD, total nitrogen, and orthophosphate.
The model considers the interaction of seven stormwater parameters, namely,
1. Precipitation and air temperature for rainfall/snowmelt. The
program utilizes the degree-day method to estimate the amount
of runoff from the snowmelt.
2. Runoff which is computed by a simple Rational formula as follows:
R = C(Pr - f)
where R is runoff
Pr is precipitation
f is available depression storage and
C is the runoff coefficient.
-------
101
3. Pollutant accumulation on a daily basis accumulation is entered
as input and is cumulated during the dry period. The washout of
pollutants is computed using an exponential washout formula:
AP = P (1 - e'EuRAt)
o
where AP is the amount of pollutants washed from the watershed
P is the initial amount of pollutants at the beginning
of the storm
E is the urban washouf decay coefficient and
At is a fixed time interval.
4. Land surface erosion is computed by the Universal Soil Loss
Equation:
A = El x K x (LS) x C x P
where A is the soil erosion rate
El is the rainfall factor
K is the soil eredibility factor
LS is the length of slope factor
C is the cropping management factor and
P is the erosion control practice factor.
The dust and dirt washout and surface erosion provide estimates
of the amount of suspended materials (washload) in the runoff.
All other pollutants are considered as fractions of the sus-
pended load.
5. Other procedures relevant to the land use-water quality modeling
include: treatment rates estimation, storage computation and
overflow from the storage/treatment system.
The program is available from the Hydrologic Engineering Center, Corps
of Engineers, U.S. Army, 609 Second Street, Davis, California 95616.
The following steps have been used in evaluating STORM:
1. Input data for three experimental watersheds in the Menomonee
River basin have been prepared according to the format of STORM.
The experimental watersheds included:
Schoonmaker Creek
Noyes Creek
Little Menomonee River at Donges Bay Road.
-------
-------
102
2. The program was run with the above data and varying storm
intensities.
3. The outputs were plotted against rainfall intensities and
compared visually with available historical data.
The results for the Schoonmaker Creek, which did not include surface
erosion, (the area is classified as medium density residential with a
high percent of imperviousness) seem to be adequate. The outputs for
Noyes Creek and Donges Bay revealed a serious logical discrepancy in the
erosion portion of the model. The suspended solids load (in tons)
decreased with increasing rain intensity, as indicated in Appendix H
Fig. 2 for Noyes Creek. The numerical estimates did not agree with the
historical data.
In conclusion, it is felt that the program STORM is not adequate for
the land use-water quality modeling for the following reasons:
1. The model relies heavily on inputted information. The model is
not dynamicj each hourly interval represents an isolated event.
2. The major shortcomings include:
a. Inadequate description of the runoff information which
relies solely on estimation of the coefficient of runoff.
b. No account is taken of soil characteristics and soil
adsorption.
c. Pollutant washout is computed as a fraction of the suspended
solids, and not from a mass balance.
d. An error exists in the erosion subroutine.
Stormwater management model - "SWMM"
The evaluation of "SWMM" is based on the past experience of the
Marquette team with the model rather than from a direct application to the
Menomonee River Project.
The model was developed under the sponsorship of the U.S. EPA by a
consortium of contractors—Metcalf and Eddy, Inc., the University of
Florida, and Water Resources Engineers, Inc. The model is comprehensive
in that it is capable of representing urban stormwater runoff and combined
sewer overflow phenomena. The major components of the model are illus-
trated in Appendix H Fig. 3.
-------
-------
9 r
103
8 u
O fa dayc dry period |
24 hours dry period
1.0 2.0
Rainfall (inch/hour)
3.0
Appendix H Fig. 2, Predicted washload for Noyes Creek using
the STORM model.
-------
104
RUNOFF
(RUNOFF)
INFILTRATION
tfjlLTRAT
(iNFIU
DECAY
(QUAL)
TRANSPORT
(TRANS)
EXTERNAL
STORAGE
(STORAG)
DRY WEA1HER
FLOW
(FILTH)
RECEIVING WATER
(RECCIV)
L~
Note: Subroutine sanies are shown In parentheses.
Appendix H Tig, 3. Major conponents of the Stortrwater
MsnaptTr.enr Model (SWHM).
-------
105
The model assumes a series of subcatchments. A rainfall hydrograph
is applied and infiltration computed using Horton's equation as a function
of time, and the average depth of rainfall excess over the catchment is
calculated. If this depth exceeds a specified depression storage depth,
overland flow outflow is calculated with Manning's equation. Overland
flow enters the gutters, and gutter outflow and flow in conduits is again
calculated with the Manning equation and continuity. The model is thus
a simple infiltration model with kinematic routing of overland, gutter,
and conduit flow. The receiving water portion is based on the finite
difference approximation of the St. Venant flow equation.
Quality is simulated similarly to "STORM." The quality constituents
are derived from the dust and dirt fallout during a specified dry period
preceding the storm and from surface erosion represented by the Universal
Soil Loss Equation. Other pollutants are again estimated as fractions of
the suspended load.
The model is not continuous and can be used only for runoff quantity
and quality simulations for events. The processes taking place between
the events are ignored. The model also ignores all transformation
occurring in soil and neglects interflow and groundwater runoff. In
addition, the model requires enormous computer storage capacity (in excess
of 350K) and very detailed input information describing the hydraulic
elements of the system. The primary purpose of the model is to design
stormwater conveyance systems. Due to its size and complexity, the program
is extremely difficult to modify or adapt for any other use.
Development of "LANDRUN"
The objectives of Task C call for a modeling technique to estimate
pollutant loadings to the Great Lakes Basin which would be related to
land use in the basin. In the Work Plan it was anticipated that two
modeling methods would be employed for estimating the loadings, i.e.,
a statistical regression analysis of the effect of various land uses on
water quality of the runoff, and a deterministic land use-water quality
model capable of simulating the overland flow portion of the runoff,
including the effect of soil and soil interactions on pollutant trans-
formations and transport.
-------
-------
106
Due to the failure of "STORM" and serious reservations about attempt-
ing the necessary modifications and amendments of large computer programs
(e.g., "SWMM"), the only feasible alternative was to develop a medium-
size hydrologic and sediment transport model.
The present version of the model developed at Marquette University
under the working code "LANDRUN" represents a dynamic hydrologic transport
model which transforms precipitation into surface runoff, interflow and
groundwater aquifer recharge quantity and quality. A schematic conceptual
flow diagram of the model is shown in Appendix H Fig. 4. Most of the
model parameters and some inputs are related to land use within the
modeled watersheds.
The present version of "LANDRUN" includes modeling the following
processes:
1. Snowpack-snowmelt by the degree-day method.
2. Infiltration by the Holtan or Philip models.
3. Excess rain from precipitation, minus evaporation, infiltration
and depression and interception storage.
4. Routing of the excess rain by an Instantaneous Unit Hydrograph (IUH)
method based on a kinematic wave formula or the empirical IUH
formula of Sarma, Delleur and Rao.
5. Dust and dirt cumulation in urban areas and washout.
6. Surface erosion by a modified Universal Soil Loss Equation which
includes effects of both rainfall energy and street runoff energy
erosion.
7. Routing of the sediment.
A soil adsorption model applicable to phosphorus, pesticides, and
toxic elements transport has been developed, tested, and is being incor-
porated as a subroutine into |;LANDRUN." The parameters of the model are
related to such soil characteristics as pH, and clay and organic matter
contents and the inputs are infiltration rate, soil moisture content,
evapotransportation (based on the potential evaporation and crop growth
stage), fertilizer application and amount of the modeled substance removed
by harvesting.
The output from the subroutine is the amount of pollutants adsorbed on
the top soil particles, quantity of dissolved pollutants removed from the
-------
-------
107
NON URBAN AREA'
1
URBAN AREAS
i ! I
FERT|U'ZERS7/
#• V
N <;Sr.;Ou.' IV'EL^}
s
__ _ ^-C-i^czEiiirz
ii __ ill j i
ills!
V V ? V V
SALT!N3_
"ATM. ;N PUTS
]~~ ~ ~]ADSCSPT'ONf ' "" "}£,""- If "_ r,,7i v . ,~ ,"~.^r " " "1 AOSORP.f
[SOLUBLE h— —^ SOIL i,^' !-!''."^ ^ u' ;I'1'- >T 'Nl-' '^Cll. r** ""-SOLUBLE
NUTRIENTS >PELEft3E_J ^l^.J^^f^RV. \ \frWOu*™™* t-J REtl_JJUTR1ENTS
^—_ _^.__ -^-" j '""~T" 4 "~~" r~l "~ oi
i __ ,V;MD j • ! j _w_l^^__j 1 ! P[
£| SiTosToN j ' cRO^ON '~^ j g!
r>~ \ I ) .
W _
i FBiTQIPM '
TRANSFORMATION) j EpOSICN~l
AND DECAY IN
SOILS
,-_Jl___
WASHO'JT
r-n ~M
EROSION
I
5
O
-J.
U.i
tri
UJ
l-
i
JSED1MENTS AN,, .,
j|NUTR!ENTS AD-j'
j| SCREED ON
I SEDIMENTS
^==^^^^
\ SOLUBLE
!| NUT RENTS
ND OECAY |N
SOILS
<
CDj
Appendix H Fig.
. Schematic conceptual flow diagram of the "LANDRUN"
model .
-------
108
top soil and the amount of dissolved pollutants in the interflow and
groundwater recharge. A separate report covering phosphorus modeling
has been prepared.
Soil nitrogen model
Nitrogen transformations in soil represent a complex system consisting
of several processes. The major nitrogen transformation processes are:
1. Atmospheric N fixation by soil bacteria,
2. Adsorption and release of NHi> by soil particles,
3. Nitrification,
U. Denitrification,
5. NO3 and NHi* uptake by plants,
6. NH* uptake by soil bacteria, and
7. Ammonification of organic N.
The schematic of the soil nitrogen model is shown in Appendix H
Fig. 5. A report covering the formulation of the model and estimating
its parameters is being prepared and will be available in the near future.
-------
109
ORG. FERTSUZER
HARVEST
ORGANIC. N
PROTEIN
^M
I FIXATION
k---^—| PLANT N K •-"••"! ATMOSPHERIC N
j
L.
SOIL ADSORBED
NH;
^_
r— j
j ^-
1
SOLUTION
4
'
-~_^i
SOLUTION
MO-
K?
SOLUTION
NO
3
FERTILIZER
FERTIU2ER K4
Appendix H Fig. 5. Schematic flow diagram of the nitrogen model
showing various transformations of N species.
-------
109a
III. The MRBS Modeling Strategy
The modeling effort of the MRBS has progressed from a general
strategy to the awareness of specific factors relating to the Pollution
from Land Use Activities Reference Group (PLUARG) modeling objectives.
The accessory programs which might be used to undertake these considera-
tions are:
1. A hydrologic season/event-baseflow tendency plot
2. A wet/dry fallout proportioning (relative to total fallout)
3. A native-state effect deduction
4. Coefficient initialization determination
5. Loading rate determination
6. Channel influence assembly
7. Loading rate adjustment matrices
It is seen that programs 2, 3 and 6 focus on specific factors, while the
others are elements within a framework-design which applies generalized
program tools, e.g.: LANDRUN, as under development at Marquette University;
HSPII—Hydrocomp Simulation Program II, the forthcoming improved version
of a program in use at the Southeastern Wisconsin Regional Planning
Commission (SEWRPC)-, to the specific output needs of the program. Detailed
descriptions of the accessory programs are presented in Appendix H Table 1,
while two alternatives for the total framework-design are outlined in
Appendix H Figs. 6 and 7.
These networks are proposed as approaches toward meeting PLUARG
modeling objectives. Determination of water quality, its relationship
to land use, and transferability of results to the entire Great Lakes
Basin are accomplished by correlation analyses of water quality to dis-
charge quantity/unit area, by pollutant and contributory land use.
Determination of the land use related coefficients "a" (land use activity
effects) and "b" (an exponent governing precipitation effects) would allow
essential quantification of event/baseflow pollutant addition, i.e.,
loading rates, by the relation aqb. Recent development of defined event
criteria by the Wisconsin Department of Natural Resources (WDNR), however,
now opens the possibility that aqb analyses might be further divided into
runoff and groundwater, as well as dissolved/suspended solid cross-
-------
-------
110
Appendix H Table 1. Detailed descriptions of accessory programs
1. A Hydrologic Season/Event-Baseflow Tendency Plot—derived via
arithmetic/geometric mean operations (i.e., correlation ensemble
establishment) on 12 years of U.S.G.S. Wauwatosa gauge streamflow
data, and arithmetic mean operation on the relative standard devia-
tions thereof. Breakdown is to weeks in order to reassemble periods
of self-stationarity (i.e., hydrologic seasons) and event-baseflow
(i.e., high-low relative standard deviation) for detailed planning
of future monitoring efforts.
2. A Foreign Wet /Dry-Fallout Proportioning—required to reduce observed
loading values by such non land use related components. The algorithms-
of-choice have not yet been ascertained pending discussion. A
statistical "Wind-Rose" treatment might be extracted from aviation-
technology, while application of algorithms from the ATM atmospheric
transport submodel of the Unified Transport Model is another pos-
sibility -for future model sophistication. However, fallout dissolution
proportioning is a more immediate concern.
3. A Native-State Effect Deduction—required to reduce observed loading
values by a non land use related component. It is envisioned that
Betson's "Nonpoint Mineral Model" approach (within the program) will
be used, probably in concert with Cherkauer's assistance and special
studies site installations at wetland, sparsely-vegetated upland,
and highly-vegetated upland locations.
U. Coefficient Initialization Determination—believed best handled by a
linear-correlation of loadings (corrected by accessory programs 2
and 3—as processed through the LANDRUN correction-initializing
program) with runoff. By this method, coefficients "a" (land use
activity effect) and "b" (exponent portion of precipitation effect)
would be determined for input to accessory program 5, i.e., to
formulate aqk and a(l-e~ ).
5. Loading Rate Determination—to be the application of the land use data
to the prior-determined coefficients "a" and "b" and flow "q" (runoff
quantity/unit area, as determined by the hydrologic submodel of the
HSPII master program). Thus, areal loading rates of focus parameters
-------
-------
Ill
Appendix H Table 1 con't.
are determined for input by land use into the water quality submodel
of the master program.
6. Channel Influence Assembly—requires the use of semi-seasonal channel
samplings to assess in-stream loading value changes (due to non land
use related physical and biological/chemical influences) and sub-
sequently to store and appropriately insert such quantifications for
approx. two-mile long reaches. A linear correlation of physical effect
against discharge velocity might be undertaken following application
of a biological/chemical influence estimate. Alternatively, algorithms
of the Stanford Sediment subroutine of the Paraquat model (Pesticide
Transport and Runoff Model) might be applied in conjunction with the
Texas biological and chemical exchange in stream channel component
of the Unified Transport Model.
7. Loading Rate Adjustment Matrices—deals with the trial-and-error
character of the final iterative aspects of model calibration. The
equations comparing theoretical loadings at monitoring sites with
observed loadings have been developed, and a Gauss-Jordan solution
algorithm is visualized.
-------
-------
112
Sydrologic Season/
Event-Baseflow
Identification Plot
RJSDAHL Sediment/
[Pesticide Alg.
Hydrologic
Transport Model
toxic-Metal Alg.
Special Studles_
Site Data
FORTRAN LANGUAGE_
PL/1 LANGUAGE
Vegetation Data —
Topography Data —
Channel Hydraulic
Details
Atmospheric
Fallout Data
Llout
T
Soil/Geology
Information
Foreign Wet/
Dry-Fallout
Proportioning
Native-State J
"Effect Deduction
IL A N DRUM
Compact Overland-Flow Non-Point Source
Continuous Initialization Program
Coefficient
Initialization!
Determination
o
Land Use Data
Channel Sampling
Data
Hydrologic
Sub-Model
Loading Rate
[.Determination,
Hydraulic
Sub-Model
Water Quality
Sub-Model
HSP-II
Determination
Storage
Seneration
.J V
I Major Monitoring
I Station Data
Channel
-I
Influence |
09
-------
113
Hydrologic Season/
Event-Baseflow
Identification Plot!
Atmospheric
Fallout
Data
Elemental Mass-
I Balance Model
Chemical
Pesticide
Sub-Model
(ACTMO or
Water
Resources
Cen.Rat'l.
Algorithm)
Physical/
Meteorologic
Data
4-
Soil/Geology
Information
(i.e. conformity
check)
Atmospheric
Transport Model
Foreign Wet/Dry Fallout
Proportioning; Solubility
Quantification
LA N D R U N
Compact Overland-Flow Non-Point Source
__CoritinuQus Simulation Hydrologic Program
Native-State
Effect Deduction
Toxic-Metal
Sub-Model
(WHTM or
Water
Resources
Cen.Rat'l.
Algorithm)
Rill/Interrill
Expansion to
Foster Form of
Univ. Soil Loss
Equation*
Infiltration
Ground-Water Studies•
Ground-Water
Delivery Model
Native-State Special Studies Site
Other Special Studir.s Site Data-*
*Note: This is incorporated
within aqrb in relations below.
Coefficient
Initialization
Determination
Land Use Data—»>
'b'
Loading Rate
Determination
Topography & Channel Hydraulic Details-
Toxic-Metal Transport (Oak Ridge or
Dynamic Estuary/WRC Rat'l. Algorithms)
j^ Expanded
Hydraulic Transport
program
Decaying-Pesticide Transport (ACTMO or
Dynamic Estuary/WRC Rat'l. Algorithms)
Sediment Transport (Stanford Sediment. SWMM
(1971), and/or Scour & Deposit Algorithms)
Benthic/
Attached
Growth
Data
Loading Rate
Adj us tment
Arr.ays
Relation Forms (all in FORTRAN LANGUAGE) for loadings: Adjusted
Rural Land Use: L=At(aq b+cf,d+g(e~T)h+Ztlf m), where Coefficient
b r i o gwr Arrays (i.e.
aqr KCPSL is the runoff term of the Foster form of a ,b • )
USLE; qr is (Q-f)/A, runoff per area; f± is interflow;
e is a transform of ground water; and f 'is ground
gwr
water recharge.
Urban Land Use: L=At(at1(l-e~bR^+cf1d+g(e~T)h+E^lf *), where R is q in rural.
Appendix H Fig. 7. Modeling strategy alternative II.
-------
breakdowns. This would conform to "q" apportionment between event and
baseflow breakdown by application of the developed criteria to observed
hydrographs.
It is to be appreciated, however, that the form aqb most appropriately
describes (essentially) rural areas. Therefore, it would be applied by
incorporation into a modified form of the Foster Universal Soil Loss
Equation. On the other hand, fallout-accumulation/runoff at work in urban
land uses is probably best handled by the formula ati(l-e~ 2), where "a"
represents land use activity effect (i.e., fallout quantity); "b", an
exponent constituent reflecting removal efficiency; ti, baseflow duration;
and t2» event duration. Since infiltration can be expressed in this
exponential decay form, consequent interflow and groundwater effects
could exploit this format a(l-e"bt).
Use of the aqb and a(l-e ) approach would enable linkage of the
IANDRUN program to a channel program (e.g., a revised Qual II) on a
repetitive basis consistent with the actual frequency of sampling and
analysis. This would be an alternative to the use of the HSPII program.
In any case, the technique should generate, from this pilot study, results
transferable to other urban areas and to the entire Great Lakes Basin,
thereby interfacing with the objectives of Task D, and lending itself to
application of remote-sensing technology.
In the first framework-alternative it is visualized that the first
four of the seven accessory programs are to be written in Fortran and the
remaining three in PL/1. The use of PL/1 would allow sequentially-linked
execution and feedback-looping between these and HSPII. The second
framework-alternative would be written in the more universal Fortran.
Accessory programs 2 through 7 would likely be developed using an inter-
active remote terminal with associated tape file creation and mapping.
The isolated and repetitive-statements character of program 1, how-
ever, has made batch-run, later supplemented by tape-file creation, an
efficient approach. These Hydrologic Season/Event-Baseflow Tendency Plots
have been completed and are presented as Appendix H Figs. 8 and 9, with
Index-Week breakdown shown in Table 2. In Appendix H Fig. 8, periods of
rough self-stationarity within the plot of geometric means of average
-------
-------
115
CO
z
co
to
tr
UJ
(O
o
o
>-
X
a:
cc
tO
i
u_
_»
UJ
to
UJ
_J
CD
z:
UJ
to
z
UJ
z
CD
CC
UJ
O£
CD
t_>
CD
•
1—
a.
UJ
UJ
cr»
r>
/T"
s_
ij
n
-3
ii i
ZD
~J
>_
fp
x:
•
(^:
Q_
en
X
o
0£
rr
y-
*
CQ
Ii 1
h
*
Z
a:
->
•
z:
o
LJ
a
•
>
o
*
»-
o
CD
1 1 1 i
M
-
-
-
v
-
-
—
.
-
—
—
-
—
—
" 1
•V
•0
^
"" Ul
— =3
- o
- if>
- ^
- 3
- ^
- 3t
0
1 1 1 1
H
h
i
L
•ooo
f 1' '••'
L
l_
L.
1
1,...
1 —
1
h
1
1 —
h—
L.
I
f—
1
1
~?
\.
^
S-
~N
1 —
I
I
h
i
i —
0
i i 1 1
L
r^
1
tf^
I
2-
*N).
•DOT
J44^-U-
1~
i
c^
-X-
~~jr
oX*
•T
V-
7*~
f
^b
i
i
H
1
.J
1
1
Na -
7*"
Jv__-
^c
r^-
-/-
-i
.u
•H
3
t— (
rH
^
-
^
—
_
-
-
—
_
—
-
-
_
—
—
—
_
—
—
—
_
—
—
—
0
O
o
10
o
•
LO
"I"
O
•
o
5^-
O
in
CO
o
•
o
CO
o
m
LO
CM
O
•
Q
Csl
0
•
LO
O
•
O
0
•
10
o
J
O
a:
CD
o
UJ
o
to
cc
UJ
to
<0
CD
C£
CJ
CC
to
CO
CO
o
(0
•8 ..
to o
O CO
ft (0
g 0)
to
CO
,
0)
bfl
(0 >,
a) -H
m 3
4n O
0 -H
•p
Q) rd
rH 4-1
,g co
CO H
p a)
0) CO
pj MH
O O
•H
•P co
m -a
rH O
0) -H
rH rH
rH Q^
82
00
•
•H
fM
CC
sja—(
SS0yOb)M01J'MM'(G)AB J0 N03W OIiU3W030/NIt4/XbH
x
•H
-0
g
ft,
a,
-------
116
CO
UJ
i—i
o
z
UJ
o
z
CO
CD
<0
a:
to
CO
a:
o
CE
I
a:
a:
a:
o
cr
CD
a:
u.
cr
a
a
•
a
i—
CO
V
1—
Q-
UJ
03
V
C3
r)
CC
>-
_J
r3
-j
UJ
-z.
rD
~3
>-
a:
2T
cc:
Q_
CE
1C
O
a:
cr
s:
•
a)
UJ
u_
»
z
LL
-D
•
sr
o
UJ
a
*
>
G
Z
•
t—
CJ
CD
-'"
~
-
-
-
-
~ UJ
_ r3
cc
- o
: g
_ *-
a:
— rs
_ z>
cc
~ Pi
1 I '
i i i
1 I '
e-r-
£
1 \ I
£>L.'
6
-ft
*-"-,
R
> •**"
&~
e
5 — ~ «
&•
_i_i_j_
" •£>.
^C
^
r-Q
-€>
-e-
. -9-
?
^ w
C^t — •
6^
i i i
-•Q
^•"^
Q
--€
tf
.-— "*
«
^
8'f
"•fl
D
3
^
D
J
=? -.
e.
®
«
, i ,
; :=€)
-^O
•^t)
3
i i i
1 i '
i i >
1 ' '_
-
-
-
1 1 r
o
o
03
0
LD
•^f
JP
•
O
-**•
O
LD
00
0
•
O
en
0
•
LO
CM
O
•
O
CM
O
•
LO
i— <
0
•
O
o
*
ID
0
CD
o
UJ
u.
CD
CO
C£
CE
UJ
CO
CO
O
CC
C£
O
cn
cn
i
CD
tu
co
i
X
UJ
o
UJ
CO
td
to
o
H
MH
H -P
(L) CU
CU >
as o
gg
e
>, 0)
H -H
•H
td 4n
t3 O
LH 0)
C co
O rd
•H o
£ e
It) v_>
•H
>
CO
tJ
0-001 0'08 0'09 O'Ofr 0'02 0"
rd co
T3 CO
PI O
td ^
-I-1 o
co td
0) O3
>
etf ro
bO
•H
X
•H
UJ
• A3G'Q1S-1
-------
117
Appendix H Table 2. Index-week breakdown with recommended sampling times
Week
1
2
3
4
*5
6
7
8
9
10
11
12
13
14
15
16
t!7
Oct.
Oct.
Oct.
Oct.
Oct.
Nov.
Nov.
Nov.
Nov.
Dec.
Dec.
Dec.
Dec.
Dec.
Jan.
Jan.
Jan.
Dates
1 to
8 to
15 to
22 to
29 to
5 to
12 to
19 to
26 to
3 to
10 to
17 to
24 to
31 to
7 to
14 to
21 to
Week
7
14
21
28
Nov. 4
11
18
25
Dec. 2
9
16
23
30
Jan. 6
13
20
27
18
19
t#20
21
22
23
24
25
26
t27
28
*29
30
31
32
33
*t#34
35
Jan.
Feb.
Feb.
Feb.
Feb.
Mar.
Mar.
Mar.
Mar.
Apr.
Apr.
Apr.
Apr.
Apr.
May
May
May
May
Dates
28 to
4 to
11 to
18 to
25 to
4 to
11 to
18 to
25 to
1 to
8 to
15 to
22 to
29 to
6 to
13 to
20 to
27 to
Week
Feb. 3
10
17
24
Mar. 3
10
17
24
31
7
14
21
28
May 5
12
19
26
June 2
36
37
38
39
40
41
42
43
t44
45
46
47
t48
49
50
*#51
52
June
June
June
June
July
July
July
July
July
Aug.
Aug.
Aug.
Aug.
Sep.
Sep.
Sep.
Sep.
Dates
3 to
10 to
17 to
24 to
1 to
8 to
15 to
22 to
29 to
5 to
12 to
19 to
26 to
2 to
9 to
16 to
23 to
9
16
23
30
7
14
21
28
Aug. 4
11
18
25
Sep. 1
8
15
22
30
'"'Channel/specialized land use studies sampling week (event preferred and
probable)
tChannel/specialized land use studies sampling week (baseflow preferred and
probable)
#Week of seasonal transition
-------
118
weekly runoff (between years) indicate hydrologic seasons. In Appendix H
Fig. 9 high relative standard deviation of average weekly runoff from daily
values (between years) is indicative of a tendency toward events in that
index-week (and the converse low relative standard deviation is indicative
of undisturbed baseflow). The channel/special studies site sampling sug-
gested in Appendix H Table 2 and Fig. 10 results from these analyses.
The next tasks anticipated to be undertaken are the writing of acces-
sory programs 2, 4 and 7, as well as possible expansion of the LANDRUN
program (via: ACTMO—Agricultural Chemical Transport Model—elaborations
on the Universal Soil Loss Equation, the ACTMO Chemical Option Submodel for
organic-chemical quantification, and WHTM—Wisconsin Hydrologic Transport
Model—toxic metal appendages). If, as in Appendix H Fig. 7, LANDRUN is
to be used for more than an initialization function, groundwater contribu-
tions must be quantified. In any case, the entire assembly (i.e., the
Marquette-developed program with appropriate expansions) is to be read into
a storage file at the University of Wisconsin-Madison.
It is intended that input and output data be arrayed by land use and
pollutants of concern as shown in Appendix H Tables 3 and 4. Coordination
with specific land use study sites is underway to cover all uses.
Further sites recommended as essential to accomplishment of project
objectives include wetlands, a highly vegetated site and a sparsely vege-
tated site to obtain the best approximation of native state conditions.
Thus, the dual function of representing specific land uses and native-state
conditions would be served. Monitoring of sites in Butler (representative
of light industry) and in Milwaukee at East Wells and North Milwaukee
Streets (representative of the high-density residential/urban downtown
land use) might also be needed. The E. Wells-N. Milwaukee Streets site is
essential if estuary modeling is to be accomplished.
Some type of monitoring at the "45th Street estuary-limit control
structure also is needed. This is dictated by particularly high lake
levels often topping the Falk Corporation flood control sheet-piling. The
specialized function of the Hawley Road monitoring installation does not
provide complete flow records. The stage-discharge correlation at t5th
Street and the theoretical reconstruction of records in the vicinity of
that site (as based on 70th Street records) would be derived consistent
-------
-------
119
1)
•p
R
•H
S
>>
H
k
rd
Q)
^s
\
H
H
rd
fM
p
0)
3
CO
<—-\
bO =
R bO
•H >, R
JH H -H
ftH p
CO rd ft
^ P CO
/— > CD E
0) CD O
•P bO -P
R
•H « Td
£ • 0)
Q) CO
>
K
•P
0)
&
1
•H
B
Q)
W
* — „
T3
rH
O
O
>>
&
id
\
6
&
&
-p
CD
S
\
H
O
O
o
rH
CD
rR
-P
rd
CD
IP,
O
rH
0)
P
O
rd
&
rd
,R
O
P >,
CD rH
•i § s
Q) rH 0)
•P rQ "H
ft $H rH CD >>
CD rH CD CD tn
C/D CD rQ ,4 -P
rQ E e >, co
Q) O Q) CD H CD
-P -P > O fj £
rd O O CD rd 0
H O 5g O i
ft
>, CU Q>
rd co -p
S -P rd
CO >, rH
CD 0) >i 3 H CD
Id g^ BP& -8
H ^ I'D < CD E
>> id
& H
rd 1 >,
rQ rd
Q) S -P
PH CO
45 H >, CD
Q) O -H H ,C
-P .H rH rH bO
rd rd ft rd -H
H S < Q) .C
R5
H-*
•H
>;
R
id
k
CO
rR -P
•P R
R CD
0 >
E CD
CD 0)
•P -P
rd rd
g 6
•rH -H
X X
0 0
>H rH
ft ft
ft ft
< <
•
• rd £>
> 1-3 CD
O f*H
id a CD i
H .p 13
CD >, rd -H
•H H rH E
>i JH 1 1
rd I I ^
CD CD > S <"
•P 1 O O f,
rd 1 rH H E
fj +J Ip, ip, Q)
CD R rd rrj CD
E W m cq w
CD
-p
rd
bO • H
P bO l
< 3 i
< -P
>-> R
H 0) Q)
& -P >
id rd cd W
rH 0> H
•> 15 S
0 0
•P >, H H
CO n) Ip ip,
CD S CD CD
3t CO CO
O CD rd rd
i-i +j pq m
rd
rH
1
1
H -P
•H R
" rH CD
• ft >
,£) < H W
CD .H
CM >i p
1 H ft
13 13 PH <
H -H rd
CD E CD CD
•H 1 1 -P
>> 1 1 rd
S S H
•P O O 1
CO . rH H 1
CD Ip, Ip, +J
&
45 PQ pq W
bO
R
•H
rX
rd
fL,
S bO
0 R
rH .H
lp H
CD ft
S |
rQ CO
CD T-J
-P CD
rd 13
6 R
•H CD
X E
O E
f4 O
ft 0
ft CD
< Pi
O
id
A;
rd
CD
rH
ft
R
o
CO
rd
CD
CO
0
•H
bO
0
H
O
fl.
13
>,
rC
E
o
?_,
4n
id
CD
•H
rH
CD
13
bO
R
•H
, — |
3
id
CD
f\
0
CO
hO
R
•H
rH
ft
rrj
CO
13
CD
id
R
CD
§
O
0
CD
pel
•
O
H
t
bfl
•H
PM
^-r*
M-i
X
•H
id
CD
ft
ft
<
-------
120
Appendix H Table 3. Land uses
1) Wetland
2) Upland
3) Cropland
*0 Orchard nursery
5) Animal husbandry
6,7) Low-density residential
8) Mid-density residential
9) High-density residential/urban downtown
10) Commercial
11) Light industry
12) Heavy industry
Appendix H Table 4. Focus pollutants
1) Carbonaceous B.O.D.
2) Strept coliform
3) Kjeldahl nitrogen
4) Nitrate + nitrite
5) Total phosphorus
6) Pesticide
7) Lead
8) Zinc
9) Copper
10) Total sediment
-------
121
with hydraulic analysis techniques. Assessment of pollutant velocity and
flow peak characteristics, as well as lag-time between flow surges (i.e.,
stage increases) and the pollutant passage at monitoring sites is needed.
To quantify these, the time-of-travel dye studies, presented in Appendix H
Fig. 11, may be utilized in concert with general flow quantity and specific
flow peak correlations between stations as developed in accord with
hydrologic/hydraulic flow principles.
It is anticipated that the Falk Corporation site and the three down-
stream grab sample sites will prove useful to the subsequent estuary-
modeling phase. Use of a Fortran assembly—most desirably Qual II (Alter-
native 2)—upstream is more compatible with estuarine modeling. At this
point, establishment of additional monitoring is recommended, as follows:
at the North Ave. dam and at the Wisconsin Ave. bridge on the Milwaukee
Riverj in Kinnickinnic Creek at the railroad embankment culvert (at
Cleveland Ave.) and at the mooring basin interface; in Milwaukee harbor;
and at the harbor entrance to Lake Michigan.
-------
122
le I He IVc Vb
Little Menomonee River
lie Lower Menomonee River
IIcL»-x Va * la
Upper Menomonee River Va
lib Ic
A
n
DNR Station
Former SEWRPC Sta.
WRC Channel Study Site \J?/~
Honey Creek
Device Monitoring
!
£
rt
izl
in
.
5
w
p
>.
(d
T3
••H
fi
fu
Segment
la
Ib
Ic
Id
le
Ila
lib
lie
lid" "
Ilia '
Ilib
IIlc
ii id
IVa
IVb
IVc
IVd
Va
Vb
Vc
Vd
Colic
Device
Fixed
X
X
X
X
I X
X
X
X
X
•ction
Req'd.
Portable
X
X
X
X
X
X
L x
X
X
X
X
X
Times f
Initial D
Sum. Bsflw.
030- 730
400-1100
500-1200
600-1300
630-1330
030- 730
400-1100
500-1200
730-1430
030- 730
130- 830
300-1000
400-1100
030- 730
400-1100
5CO-1200
730-1430
030- 730
400-1100
500-1200
600-1300
rom Day's
ye Release
Sum/Fall Event
010- 450
300- 740
340- 820
420- 900
520-1CGO
1 010- 450
300- 740
340- 820
600-1040 .
010- 450
030- 530
230- 710
310- 7bO
010- 450
300- 740
340- 620
S50-1C30
010- 450
300- 740
340- 820
420- 900
Summer
Bsflw.
3 pts.
2
1
1
1
3 ^
1
2
1
3
1
1
j-
3
1
r2
1 '
3
1
2
1
Summer
Event
9 pts
7
2
1
1
9
I
' Fall
Event
11 pt
7
2
1
1
11
1
7 j 7
2 . _i
9 "1
1
1
2
2
11
I
1
2
9 111
2
3
I
9
2
3
1
2
'*
1
11
2
4
1
351(x2>) 82 94"
'^Density for Rhodamine WT (20% sol.): I * 246 pt. Sl.lS* " 29«t Ibs.
Boosts from Dupont Co. (1976): i.e. 2-25# lot @$10?t + 250# drum @$920f
Appendix H Fig. 11. Time-of-travel study details.
-------
123
IV. Empirical Modeling of Runoff Quality from Small Watersheds
Small watersheds which contain no point sources of contamination act
as natural integrators of the various nonpoint contamination sources which
can contribute to streamflow . The quality of runoff from these watersheds
is a complex interaction of hydrologic, climatic, geologic, cultural and
chemical factors. Runoff quality can be modeled by identifying the role
of each of these causative factors and incorporating them into an inclusive
digital model. The primary modeling efforts of the Menomonee River Pilot
Watershed Study are directed toward this end. However, the digital model
to be developed will necessarily be complex. It will have large data input
and computer storage requirements and require extensive calibration. In
some instances a far simpler—albeit less precise—model might be useful.
This segment of the overall project is designed to develop such a
model. It is based on the premise that observing quality of runoff from a
variety of small watersheds will enable us to empirically relate that
runoff quality to the land use, climatic, hydrologic and other factors
which produced it. This model assumes that the concentration of a chemical
in surface runoff varies temporally about a mean runoff concentration in a
predictable manner. The factors which control the mean concentrations
must therefore be identified and an empirical means for predicting mean
concentrations must be devised.
After a mean concentration has been developed, a relative concentra-
tion, the ratio of the concentration at any time to the mean, can be
defined. The model assumes that the relative concentration of runoff from
a given watershed will show a temporal variation which is a reproducible
function of land use, storm characteristics and season. Development of
relative concentration curves for a variety of watersheds allows prediction
of the distribution of runoff concentrations during runoff events. Combi-
nation of the concentrations with runoff quantities then allows prediction
of mass loadings.
Work on this model began independently in 1974, and was incorporated
into the Menomonee River Project in the spring of 1975. As a result, the
effort to date has been largely in three small watersheds which are
-------
-------
124
tributaries of the Milwaukee River, but which are also contiguous with the
Menomonee River watershed in the Village of Brown Deer and the City of
Mequon (Appendix H Fig. 12). The three watersheds were selected originally
for study because5 while they are all quite similar in size and natural
conditions, they span the zone of active development on Milwaukee's north
side. The southernmost is largely developed as a stable light-to-medium
density residential area. The northernmost part is still rural-agricul-
tural. The central watershed is undergoing two types of active development:
construction and expansion of commercial centers and condominia in the
upper watershed, and development of a new medium density residential sub-
division in the central reaches. Manual observations of the responses
of these watersheds to runoff events have been underway since June, 1974.
It was out of these observations that the idea for a simple, empirical
model of small watershed overland runoff developed. The work on this
model has been divided into three phases, which are described below.
Phase I - Watershed monitoring and initial data analysis
This effort began in June, 1974, and continues. With the aid of one
graduate assistant, the principal investigator has monitored the response
of the three watersheds to 20 runoff events. Staff gauges and sampling
sites have been established at four sites (Appendix H Fig. 13). A rating
curve has been developed and is updated continually at each site. Some
discharges are measured during runoff events using a pygmy Price meter
to recalibrate the rating curves. Baseflow samples are collected at each
site prior to runoff, and water samples are taken periodically during the
runoff event. Samples have been analyzed in situ for pH, electrical con-
ductivity and temperature and in the laboratory for total dissolved solids
(TDS), Cl", Na, Ca, Mg, HC03~ and suspended solids (SS).
Flood hydrographs are generated from the gauge readings and rating
curves. Load hydrographs for each chemical constituent are generated by
combining sample concentrations and flow hydrographs. After visual in-
spection, all hydrographs are digitized, and interpolated, integrated and
plotted on a Univac 1110 computer. To date, interpolation and integration
are completed on 14 events, and the plotting is underway.
-------
-------
125
ro
0)
CO
0)
M
(1)
W
13
p
C
O
•H
P
-P
W
•M
O
bD
•H
X
X
•H
•s
0)
O,
O4
-------
126
LEGEND
STREAM GAGE SITE
RAIN GAGE SITE
Appendix H Fig. 13,
Locations of staff gauges and sampling
sites in the study area.
-------
127
On Beaver Creek, three sets of hydrographs were generated: one set
for each of the sampling sites, and one for just the lower half of the
watershed, i.e., the area under residential development. The latter hydro-
graphs are obtained by routing flow from the upper site downstream and
subtracting it from total flow at the lower site. Thus, three data sets
will be available: the upper watershed containing commercial development
and draining into the Northridge Lakes, the lower watershed containing
residential development, and the total watershed.
Appendix H Table 5 provides some of the preliminary runoff concentra-
tion values for the suburban, developing residential and rural watersheds.
Appendix H Table 6 indicates the magnitude of the load per unit drainage
area carried from each watershed for a variety of events. Appendix H
Table 7 puts the absolute load magnitude data into a comparative form,
i.e., the relative loading. Each number in Appendix H Table 7 is the ratio
of the load from a particular watershed to that from the rural watershed.
The relative loading then is a measure of the degree of change from rural
conditions caused by a particular land use. Data for the remaining events
and the additional chemical parameters is being processed currently.
The events monitored have been grouped into two seasonal categories,
namely, summer and winter. Summer events are runoff events occurring
between May and November, during the time when watershed and channel
vegetation have major effects on runoff. Winter events are snowmelt
runoff events occurring generally between January and March. Some moni-
tored events, such as March rainstorms, do not fit in either category.
However, they are hydrologically unique and cannot be treated except as
individual cases.
It was anticipated that watershed runoff would be related to land use
as well as affected by the season of the year, the intensity and quantity
of precipitation or snowmelt, and the antecedent runoff conditions. The
latter is treated here as the number of days since the last runoff event.
Summer runoff is shown in Appendix H Fig. 14-, where total runoff per unit
rainfall is plotted as a function of antecedent flow. As depicted
graphically, the data collected thus far indicate that total runoff is
related to land use» total rainfall and antecedent conditions. However,
-------
-------
128
Appendix H Table 5.
Quantity and quality of runoff from two
suburban watersheds (Brown Deer Creek and
Lower Beaver Creek) and a rural watershed
(Trinity Creek) during a series of storm
events
Mean concentrations
Event
Runoff Amount
m3/, 2
Total dissolved
solids Cl
TT T /I
Suspended
solids
July 25, 1974 972
October 6, 1974 1,093
Hay 14, 1975 492
July 11, 1975 2,386
July 18, 1975 239
July 19, 1975 381
November 2, 1975 2,290
June 9, 1974 2,560
July 10, 1974 1,940
Weighted mean concentration
Standard deviation
Brown Deer Creek
256
142
99
168
644
161
197
194
Lower Beaver Creek
39.9
24.7
30.5
22.9
100.0
27.6
20.3
27.5
88.6
240
189
576
1,210
309
826
453
July 25, 1974
October 6, 1974 997
May 14, 1975 694
July 11, 1975 928
July 18, 1975 473
July 19, 1975 2,430
November 2, 1975 1,385
June 9, 1974
July 10, 1974
Weighted iriean concentration
Standard deviation
July 25, 1974
October 6, 1974
May 14, 1975
July 11, 1975
July 18, 1975
July 19, 1975
November 2, 1975
June 9, 1974
July 10, 1974
17.7
8.3
0.0
4.7
2.6
23.1
8.3
18.4
Weighted mean concentration
Standard deviation
107
94
120
122
56
179
105
45
Trinity Creek
405
659
213
319
701
423
528
166
17.3
21.2
14.3
12.6
8.7
19.2
14.3
4.7
29.4
87.2
15.9
26.1
71.6
39.4
52.1
24.0
518
240
358
180
266
309
132
112
44
26
51
34
-------
129
Appendix H. Table 6. Loadings from small watersheds
Land use
Suburban developed
(Brown Deer Creek)
Suburban developing
(Lower Beaver Creek)
Rural
(Trinity Creek)
Lake
(Upper Beaver Creek)
Suburban developed
(Brown Deer Creek)
Suburban developing
(Lower Beaver Creek)
Rural
(Trinity Creek)
Lake
(Upper Beaver Creek)
Discharge
m3
1,150
1,170
7.8
2,380
386
270
7.5
912
Total dissolved
solids
Rainstorms (average of
270
126
3.6
1,290
Salt- induced flow (1
1,433
1,070
2.3
276
Suspended
Cl~ solids
Iro-
Kg
6 events)
30 505
175 965
0.5 0.6
414
event )
840
658
0.7
102
Snowmelt (average of 3 events)
Suburban developed
(Brown Deer Creek)
Suburban developing
(Lower Beaver Creek)
Rural
(Trinity Creek)
Lake
5,230
6,370
11,900
3,930
1,730
2,600
3,350
4,065
315
440
110
975
(Upper Beaver Creek)
-------
130
Appendix H. Table 7 . Loadings from small watersheds relative to a comparable
rural watershed. Note the rural watershed for all
parameters is given a value of one.
Land use
Suburban developed
(Brown Deer Creek)
Suburban developing
(Lower Beaver Creek)
Rural
(Trinity Creek)
Lake
(Upper Beaver Creek)
Suburban developed
(Brown Deer Creek)
Suburban developing
(Lower Beaver Creek)
Rural
(Trinity Creek)
Lake
(Upper Beaver Creek)
Suburban developed
(Brown Deer Creek)
Suburban developing
(Lower Beaver Creek)
Rural
(Trinity Creek)
Lake
Loadings of
Total dissolved Suspended
Discharge solids Cl solids
_3
m —
Rainstorms (average
145 75
150 35
1.0 1.0
305 360
Salt-induced flow
51 630
36 470
1.0 1.0
120 120
Snowmelt (average of
0.44 0.52
0.53 0.78
1.0 1.0
0.33 1.2
kpr
g
of 6 events)
60 840
35 1,600
1.0 1.0
830
(1 event)
1,225
960
1.0
150
3 events )
2.9
4.0
1.0
9.0
(Upper Beaver Creek)
-------
131
•rH
03
•P
• H
4000
CL,
-v
OS
<» e 2000
03 ^
c «
•5 -
rH
C
fn
0>
O
c
50
OS
25
10
Antecedent rainfall conditions
(days)
20
Appendix H Fig.
Relationship of total summer runoff and
antecedent flow.
-------
132
the data show no significant relationship between runoff and rainfall
intensity. Incorporation of additional data as Phase I continues will be
used to refine these preliminary relationships. Ultimately Phase I data
will be combined with that from small watershed monitoring stations within
the Menomonee watershed to quantify the relationship of runoff to land
use.
Phase II - Interpretation of mean chemical concentrations of runoff
Work on this phase began in December, 1975, when sufficient runoff
events had been monitored to provide a significant set of data. The mean
concentration of a material in surface runoff is defined as:
C = Lr/Qr
where C is the mean concentration (ppm)
Lr is the total load of a material contributed to a stream
by surface runoff during an event (in kg), and
Qr is the total quantity of surface runoff during that
event (in m3).
The total flow out of a watershed during a runoff event is a mixture of
surface runoff and baseflow. Therefore, it is necessary to delete the
baseflow runoff and mass loads from total flow to obtain surface runoff
values.
It was anticipated that C values would vary seasonally and that within
a season they would be affected by the total quantity of runoff and the
antecedent runoff conditions. Preliminary findings seem to contradict
this assumption, however. Appendix H Fig. 15 shows that C values exhibit
no visible relationship to the quantity of runoff (Q) for the rural,
developing or suburban watersheds. The TDS relationships are representa-
tive of the other dissolved chemical parameters. There is no apparent
relationship to rainfall intensity or antecedent runoff. In fact, C
values appear relatively constant for a wide variety of events, particu-
larly in the nonrural watersheds. If C is independent of the hydrologic
and meteorologic factors3 then a single value of C can be used for a given
watershed for all summer events. A preliminary set of these values is
provided in Appendix H Table 8. Additional summer events have not been
fully processed. No firm conclusions regarding summer mean concentrations
-------
133
cr.
c
•t—
n.
o
r—
01
• 00
0 1—
*
CO C
C rd
O H
•H
•P 0)
rd ,
•P
G -H
rd co
o> c
6 a)
•p
II . (-«
T~> I—<
O -H
•H H
f, tO
W lp
C C
O -H
•H rd
•P ^
rd
H t3
0) C
<*; rd
in
H
bO
•H
X
•H
(uidd)
$QJ,
-------
134
Appendix H. Table 8 . Comparison of quality of surface runoff water
Land use
Total dissolved Suspended
solids Cl solids
Suburban developed
(Brown Deer Creek)
Suburban developing
(Lower Beaver Creek)
Rural
(Trinity Creek)
Suburban developed
(Brown Deer Creek)
Suburban developing
(Lower Beaver Creek)
Rural
(Trinity Creek)
Suburban developed
(Brown Deer Creek)
Suburban developing
(Lower Beaver Creek)
Rural
(Trinity Creek)
mg/1
Summer rainstorms (average of 7 events)
194 27.5 830
105 14.3
528 52.1
Salt-induced flow (1 event)
3,710
3,960
305
2,180
2,440
93
Snowmelt (average of 3 events)
330 ' 60
408
282
69
9.2
310
51
-------
135
can be made until these data are included.
Suspended solids concentrations do show an exponential relationship
to total surface runoff in the summer. However, rainfall intensity and
antecedent runoff do not seem to affect these concentrations. A second
winter monitoring season is now being completed. Analysis of the winter
data will not begin until these recent events are processed.
Finally, when the values of mass loadings during runoff become avail-
able for small watersheds in the Menomonee watershed, these data will be
analyzed to determine the factors controlling their mean runoff concentra-
tions. Combination of Phase II data with the Menomonee data should allow
quantification of the role of land use in determining mean concentrations.
Phase III - Development and testing
of relative concentration curves
Plots of relative concentration as a function of a time ratio are
being generated by computer for each event. Software is now largely com-
plete and work on this phase is underway.
Appendix H Fig. 16 shows some preliminary curves which are generated
by hand. Each is a composite of four events. Relative concentration is
plotted against the ratio of hours elapsed since runoff began to length of
the storm. These curves were combined with unit hydrographs and predicted
C values for the Brown Deer Creek watershed to predict the TDS load hydro-
graphs for two events (Appendix H Fig. 17). The similarity between the
predicted and actual hydrographs is encouraging. More definitive work on
Phase III will be undertaken in June, 1976.
-------
-------
136
2.Or
lo
o
o
•H
-H
fO
PH
£1.0
0>
o
c
o
o
• H
-p
Suburban
—o-SCLake
Appendix H Fig. 16.
1.0 2.0 3.0
Relative time (T/Tp)
Relative TDS concentration (C/C) versus
relative time (T/Tp) for the study basins,
Each line is the average of four events.
CN -, r
g 16
.X
12
CO
O
00
Brown Deer Ck.
Oct. 6, 1974
Observed
Predicted
20
Brown Deer Ck.
Mar. 19, 1975
0
"O 4 8 12 16 04
Time (hours after start of event)
Appendix H Fig. 17. Comparison of predicted with observed TDS
loads during two runoff events in a sub-
urban watershed.
-------
U S. Environmental Protection Agency
GLNPO Library Collection (PL-l^j)
77 West Jsckson BcuSevard,
Chicago, IL 60604-3590
------- |