EPA-600/3-83-054
FLUVIAL TRANSPORT AND PROCESSING OF
SEDIMENTS AND NUTRIENTS IN
LARGE AGRICULTURAL RIVER BASINS
by
David B. Baker
Water Quality Laboratory
Heidelberg College
Tiffin, Ohio 44883
Grant No. R-805436
Project Officer
Thomas O. Barnwell, Jr.
Technology Development and Applications Branch
Environmental Research Laboratory
Athens, Georgia 30613
ENVIRONMENTAL RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
ATHENS, GEORGIA 30613
-------
DISCLAIMER
Although the research described in this report has been funded wholly
or in part by the United States Environmental Protection Agency through
Grant Number R-805436 to Heidelberg College, it has not been subjected to the
Agency's required peer and policy review and therefore does not necessarily
reflect the views of the Agency and no official endorsement should be inferred.
ii
-------
FOREWORD
As environmental controls become more costly to implement and the
penalties of judgment errors become more severe, environmental quality man-
agement requires more efficient management tools based on greater knowledge
of the environmental phenomena to be managed. As part of this Laboratory's
research on the occurrence, movement, transformation, impact, and control
of environmental contaminants, the Technology Development and Applications
Branch develops management and engineering tools to help pollution control
officials achieve water quality goals through watershed management.
Water quality planning and evaluation requires extensive data on the
contribution of agriculturally derived nutrients and sediments to pollutant
loads in water bodies draining watersheds. For several years, comprehensive
studies have been conducted in the large agricultural river basins on north-
western Ohio. This report, which summarizes data collected in these programs
and provides detailed examples of nutrient and sediment transport in the
region, should be of particular use in water quality management studies.
David W. Duttweiler
Director
Environmental Research Laboratory
Athens, Georgia
111
-------
ABSTRACT
The transport of nutrients and sediments at 12 U.S. Geological Survey
stream gages in northwestern Ohio has been studied in detail. The watersheds
range in size from 171 km2 to 16,395 km2. Land use is dominated by row crop
agriculture on generally fine textured and poorly drained soils. Automatic
samplers were used to collect at least four samples per day at each station
for a three to five year period. During storm events all samples were
analyzed, whereas during non-event periods one sample per day was analyzed.
In parallel with the water quality studies, a detailed land use, land
capability data set was developed by the Army Corps of Engineers as part of
the Lake Erie Wastewater Management Study. This data base was used to
calculate gross erosion rates for each watershed as well as to evaluate
various nonpoint source pollution control programs.
For these rivers the concentrations of suspended solids, total
phosphorus, nitrate + nitrite nitrogen and total Kjeldahl nitrogen all tended
to increase with increasing stream flow. Storms with similar peak discharges
had widely varying flux weighted concentrations of both sediments and
nutrients. Large seasonal and annual variations in flux weighted
concentrations were observed. The ratio of particulate phosphorus to
sediments varied greatly among samples, with lower ratios generally associated
with higher suspended solids concentrations.
The unit .area yields of nutrients were in the high range for agricultural
watersheds. Sediment delivery ratios ranged from 6.2% to 11.9$. There was no
correlation between sediment delivery ratio and basin size for these
watersheds. Furthermore, there was no correlation between gross erosion rates
and unit area nonpoint phosphorus yields. • - .,
Phosphorus entering streams from point sources is rapidly processed by
the stream system. Subsequent transport of the phosphorus to the lake is
dependent on resuspension of particulate phosphorus during storm events. The
soluble reactive phosphorus exported during storm events is largely derived
from nonpoint sources. The export of soluble reactive phosphorus during storm
events can comprise about 50% of the total export of bioavailable phosphorus
although it is only about 20% of the total phosphorus export. Upon delivery
to the lake, phosphorus from nonpoint sources has a higher percentage
availability than phosphorus derived from point sources which has been
processed by the stream system.
This report was submitted in fulfillment of Grant Number R-805U36 to
Heidelberg College under the sponsorship of the U.S. Environmental Protection
Agency. This report covers the period September 1977 to April 193l, and work
was completed as of January 1983-
IV
-------
CONTENTS
FOREWORD iii
ABSTRACT iv
FIGURES .".."."."."...' vii
TABLES ix
ACKNOWLEDGMENTS xiii
1. INTRODUCTION 1
2. CONCLUSIONS 4
3. RECOMMENDATIONS 7
4. DESCRIPTION OF THE STUDY AREA 9
LOCATION OF THE STUDY BASINS 9
GEOLOGY 9
SAMPLING STATIONS 9
WATERSHED CHARACTERISTICS 14
POTENTIAL GROSS EROSION 20
5. STUDY METHODS 23
SAMPLE COLLECTION 23
SAMPLE PRESERVATION AND STORAGE 24
LABORATORY PROCEDURES 24
QUALITY CONTROL 25
FLOW DATA 29
DATA STORAGE 29
COMPUTATIONAL METHODS 32
6. CONCENTRATIONS OF NUTRIENTS AND SEDIMENTS AT THE
TRANSPORT STATIONS 48
DESCRIPTION OF DATA SETS 48
TYPES OF WEIGHTED AVERAGE SEDIMENT AND NUTRIENT
CONCENTRATIONS , 48
PATTERNS OF SEDIMENT AND NUTRIENT CONCENTRATIONS IN
RELATION TO STREAM FLOW 52
HYDROGRAPHS, SEDIMENTGRAPHS AND CHEMOGRAPHS .... 59
VARIATIONS AMONG RUNOFF EVENTS AT A SINGLE GAGING
STATION 65
VARIATIONS ASSOCIATED WITH SEASON AND RAINFALL
INTENSITY 73
ANNUAL VARIATIONS IN SEDIMENT AND NUTRIENT
CONCENTRATIONS 77
VARIATIONS ASSOCIATED WITH LOCATIONS RELATIVE TO
POINT SOURCES 84
CONCENTRATION EXCEEDENCY RELATIONSHIPS 87
SEDIMENT-PHOSPHORUS RELATIONSHIPS 90
7. NUTRIENT AND SEDIMENT LOADING AT TRANSPORT STATIONS . 94
MEAN ANNUAL LOADS OF NUTRIENTS AND SEDIMENTS ... 94
UNIT AREA LOADS 95
ANNUAL VARIATIONS IN NUTRIENT AND SEDIMENT LOADING 98
FLUX EXCEEDENCY RELATIONSHIPS 102
8. WATER QUALITY MANAGEMENT IMPLICATIONS
POINT AND NONPOINT SOURCE COMPONENTS OF STREAM
PHOSPHORUS TRANSPORT 107
v
-------
STREAM PROCESSING OF POINT SOURCE PHOSPHORUS INPUTS . 109
COMPARISON OF PHOSPHORUS INPUTS FROM TRIBUTARIES AND
DIRECT POINT SOURCES 112
SEDIMENT DELIVERY RATIOS AND CRITICAL AREA
IDENTIFICATION 114
9. TRANSPORT MODELING 119
REFERENCES 121
APPENDIXES
1. ANALYTICAL METHODS 126
2. INTEGRAL METHODS FOR APPROXIMATE WATER AND POLLUTANT
TRANSPORT IN RIVERS 130
VI
-------
FIGURES
Number Page
1 Location of the northwestern Ohio river basins in relation
to Lake Erie .......................... 10
2 Glacial map of the Sandusky River Basin area ........... 11
3 Bedrock geologic map of the Sandusky Basin area .......... 12
4 Location of tributary mouth sampling stations for Lake
Erie loading studies from northwestern Ohio rivers ....... 14
5 Location of sampling stations in the Sandusky Basin
network of river transport stations .............. 15
6 Flow chart for sample analyses .................. 26
7 Flow duration flow class intervals for the Fremont
gaging station ......................... 46
8 Relationship of concentrations (mg/l) and log flow in
CFS at the Wolf," East and West stations ............. 54
9 Hydrographs -and sediment graphs, chemographs for a
runoff event at the Wolf, East station between May 2
and May 11,1 977 .............. _ .......... 60
10 Example of simultaneous (A) and trailing- (-B^ C & D)
sediment peaks during the 1977 water year at the Portage,
Wolf East, Upper Sandusky and Nevada stations ......... 63
11 Compound hydrographs, sedimentgraphs and chemographs for
August, 1979 runoff events at the Fremont gaging station .... 64
12 Variability in hydrograph-sediment graph relationships
at the Upper Sandusky gaging station .............. 69
13 Variability in flux weighted mean concentrations
of suspended solids, total phosphorus, nitrates and
conductivity in relation to peak flows for individual
storms ............................. 71
14 Flux weighted mean concentrations of suspended solids
in relation to peak flow with month of occurrence
marked for individual storms ................. 74
15 Raingage tracings from storms on April 13 and May 24-26,
1979. The intense rainfall of April 13 was associated
with rainfall #46 at Upper Sandusky while rainfall on
May 24-26 generated runoff #48 ................. 75
Vll
-------
Number Page
16 Seasonal sediment rating curves at the Melmore gaging
station: 16A Summer (A) and winter (X) sediment
concentrations; 16B Summer (d) and winter (Q)
instantaneous sediment fluxes 76
17 Summer and winter average concentrations plotted
in relation to the midpoint of the flow duration
class intervals for Honey Creek at Melmore 80
18 Profiles of mean phosphorus concentrations along the
Sandusky River during June - September 1974 85
19 Comparison of phosphorus concentration profiles
obtained from 3 samples per day and 1 2 samples per
day at three bridges 86
20 Phosphorus/sedimert ratios (x 1000) for individual
•samples plotted in relation to stream flow and
suspended sediment concentrations for Honey Creek at
Melmore 93
21 Annual variations in the ratio of total phosphorus
export to suspended sediment export in relation to
annual flux weighted suspended solids concentration 103
22 Relationship between sediment delivery ratio and
drainage ar,ea 116
23 Relationship between nonpoint phosphorus export and
gross erosion rates in the study watersheds 117
24 Deposition and resuspension of (a) total phosphorus and (b)
orthophosphate in the Sandusky River near Upper Sandusky
- Storm beginning 7 July 1976. Source: Melfi & Verhoff,
1979 120
Vlll
-------
TABLES
Number Page
1 Locations, station numbers, areas, periods of hydrological
record, and average discharge for stream transport
stations in northwestern Ohio 13
2 Percentage distribution of major land uses in the study
basins 16
3 The areal percentage of lands in various slope
classifications for the study watersheds 18
4 Percentage distribution of soil textures in the study
watersheds 18
5 Percentage distribution of soils falling into various
drainage classifications for the study watersheds 19
6 Percentage distribution of soil erodibility as indicated
by the K-value of soils in the study watersheds 19
7 Summary of gross erosion rates by soil management groups
for the study watersheds 21
8 Precision data based on analysis of replicate pumped
samples 26
9 The effects of one week of sample storage on analytical
values '..„-'• 27
10 Average change in concentration during one week of sample
storage 28
11 Paired t-test comparisons of pumped and grab samples 30
12 Sample archive printout for transport stations 31
13 Sample printout of program for flux calculations 33
14 Sample printout of flux summary option which includes
data for individual samples and cumulative totals 33
15 Sample printout of flux summary option for selected
months 36
16 Sample printout of the flux summary for the 1979 water
year loading of total phosphorus at the Melmore gaging
station 36
-------
Number
17 Sample printout showing the 50 largest flow volumes
contributing to the phosphorus loading at Melmore
during the 1979 water year ................... 37
18 Sample printout showing the 50 largest phosphorus
fluxes contributing to the phosphorus loading at
Melmore during the 1 979 water year ............... 38
19 Sample printout of a concentration exceedency table
for a sample data set from the Melmore station ......... 39
20 Sample printout of a flow exceedency table for a
sample data set from the Melmore station ............ 40
21 Sample printout of flux exceedency table for a
sample data set from the Melmore station ............ 41
22 U.'S. Geological Survey flow duration table for the
Sandusky River near Fremont .................. 43
23 Sample printout of calculation of mean annual loading
using flow duration intervals and flux weighted mean
concentrations for each interval ................ 45
24 Sample printout^of calculation of mean annual loading
of total phosphorus using flow duration tables
and average concentrations, standard deviations
and standard errors for each interval ............. 47
25 Numbers of samples analyzed and percent of f 1-ow' and
time monitored by water year for northwestern Ohio
sampling stations ........... '..."" ......... 49
26 Comparison of time weighted average, flow weighted
average and flux weighted average concentrations of
sediment and nutrients at the Honey Creek, Melmore
sampling station ........................ 52
27 Comparison of flux weighted and time weighted mean
concentrations of suspended solids and nutrients
in northwestern Ohio River Basins ............... 53
28 Summary of storm data at the Upper Sandusky gaging
station
29 Linear regressions of flux weighted mean concentrations
on peak flows for the Upper Sandusky storm data set ...... 72
30 Comparison of summer and winter suspended solids
concentration by flow intervals at the Melmore
gaging station ........................ 78
-------
Number Page
31 Comparison of summer and winter total phosphorus
concentrations by flow intervals at the Melmore
gaging station 78
32 Comparison of summer and winter NO 3-N concentrations
by flow class intervals at the Melmore sampling
station 79
33 Comparison of summer and winter conductivities
concentrations by flow intervals at the Melmore
sampling station 79
34 Annual variations in flux weighted mean concentrations
of sediments and nutrients at northwestern Ohio
gaging stations 81
35 Comparison of monthly distribution of runoff between
•1978 and 1979 water year 84
36 Percentage of time the indicated concentrations of
suspended solids (mg/l) were exceeded at representative
gaging stations 88
37 Percentage of time the indicated concentrations of
total phosphorus (mg/l) were exceeded at representative
gaging stations 88
38 Percentage of time the indicated concentrations of
nitrate-nitrogen (mg/l) were exceeded at representative
gaging stations ' 39
39 Percentage of time the indicated concentrations of
total dissolved solids as represented by conductivities
(umhos) were exceeded at representative gaging stations .... 89
40 Nutrient-sediment ratios for agricultural watersheds
of northwestern Ohio, 1974-1979 91
41 Mean annual sediment and nutrient loading at transport
stations ' 96
42 Unit area yields of sediments and nutrients for
northwestern Ohio agricultural watersheds 97
43 Total phosphorus unit load by land use and land form
(in U.S.A.) 97
44 Sample printout of monthly flux and-flow summaries for
identifying and correcting erroneous or missing flow
and concentration data 100
45 Annual variability in sediment and nutrient export from
selected northwestern Ohio rivers 101
XI
-------
Number Page
46 Ratio of high annual yield to low annual yield ..........
47 Flux exceedency values for suspended solids, total
phosphorus and soluble reactive phosphorus at the
Fremont gaging station ..................... -"-"^
48 Percent of total flux accounted for by fluxes
associated with flows and fluxes exceeded fixed
percentages of time ...................... 105
49 Indirect municipal point source phosphorus discharges
in the Sandusky River Basin 1978 ................ 108
50 Minimum nonpoint source phosphorus yields for the study
watersheds
51 Analyses of NaOH-extractable P from suspended solids
collected during runoff events of study watersheds.
(after Logan, 1 978b) ....................... 112
52 Sediment delivery ratios for northwestern Ohio
agricultural river basins . ' ................... 115
53 Linear regressions relating to delivery ratios,
sediment yields, ~ and nonpoint phosphorus yields
watersheds. (after Logan, 1978b) ............... 115
XII
-------
ACKNOWLEDGMENTS
The studies presented in this report reflect the support and assistance
of many organizations and individuals. The help and advice of the project
officers from the U.S. Army Corps of Engineers, Dr. Stephen Yaksich, and from
the U.S. EPA, Mr. Thomas Barnwell, Jr. , are gratefully acknowledged. The
directions of this study have been an outgrowth of discussions with and among
many individuals, including: Mr. John Crumrine, Seneca Soil and Water
Conservation District; Dr. William Sonzogni, Great Lakes Basin Commission;
Dr. Terry Logan, Ohio State University; Dr. Frank Verhoff, University of
West Virginia; Mr. Tom Cahill, Cahill and Associates; Mr. John Clark,
International Joint Commission; Mr. Jerry Wager, Ohio EPA; and Mr. Gene
Baltes, Soil Conservation Service. The cooperation of the U.S. Geological
Survey in the timely provision of stream gaging and flow data and in the
maintenance of the pumping systems at the sampling stations is greatly
appreciated.
The financial support of the U.S. EPA, the Army Corps of Engineers, the
Rockefeller Foundation, the Soap and -Detergent Association, Heidelberg
College, and the cities of Bucyrus, Tiffin and Upper Sandusky has made these
studies possible.
The sampling program and laboratory analyses have been completed under
the capable direction of Mr. Jack Kramer, our laboratory manager. The aid of
Dr. Kenneth Krieger, Dr. Peter Richards, and Mr. Phillip Kline, all of the
Heidelberg Water Quality Laboratory, in the interpretation of data and the
production of this report is recognized. The dedicated work of our laboratory
staff, including Ms. Ellen Ewing, Ms. Francine Turose, Mrs. Barbara
Merryfield and Mrs. Josephine Setzler, .has contributed greatly to the
successes of this study. The computer programming efforts of Mr- Richard
Leslie, Mr. David Kuder, and Mr. Greg Yarmoluk facilitated both the storage
and analyses of the data developed in this study. During the course of this
work, student technicians provided valuable assistance in the collection and
analysis of samples. The help of Marie Cole and Heidi Goetz in data analysis
and graphics is greatly appreciated. My thanks also go to Mrs. June Hatoor
for her extra efforts in typing this manuscript.
Kill
-------
SECTION 1
INTRODUCTION
In comprehensive studies of Great Lakes pollution from land use
activities, the large agricultural river basins of northwestern Ohio stand out
as major problem areas (IJC, 1980a). These basins contribute high unit area
yields of sediment, phosphorus, and nitrogen to the western and central basins
of Lake Erie. The high sediment and nutrient yields are associated with
intensive row crop agriculture on the fine textured soils of this region.
Agriculturally-derived nutrients, coupled with the point source nutrient
loading from the area's population centers and the morphometry of the Lake
Erie basin, have resulted in accelerated eutrophication and associated water
quality problems.
The Maumee, Portage, Sandusky and Huron Rivers of northwestern Ohio have
been the sites of detailed studies of nutrient and sediment transport.
Initially these studies were aimed at determining the total loading of
nutrients and sediments from these rivers into Lake Erie, so that both total
loads and the relative contributions of point and nonnoint sources could be
determined. By 1975, it became evident that nonpoint sources accounted for
more total phosphorus inputs into Lake Erie than point sources, and that point
source controls would be inadequate to achieve the phosphorus load reductions
considered necessary to reverse the eutrophication of Lake Erie (COE, 1975).
In 1975, the river transport studies within the Sandusky River Basin were
expanded to include a network of nine stations, including six large tributary
watersheds and t three mainstream stations. The expanded studies were aimed at
identifying critical watersheds with respect to sediment and nutrient yields
and to study the effects of mainstream transport processes on the delivery of
both point and nonpoint-derived pollutants to the Take.
Concurrently with the river transport studies?- detailed studies were
being conducted to determine the applicability of various agricultural best
management practices for reducing nutrient and sediment losses from cropland
in these river basins (COE, 1979). These studies concluded that a variety of
conservation tillage practices, including no-till, could significantly and
economically reduce phosphorus loading to Lake Erie. A major tillage
demonstration project was initiated in the Honey Creek Watershed of the
Sandusky Basin under the direction of a Joint Board of Soil and Water
Conservation District Supervisors and supported by both the U.S. Army Corps of
Engineers and the Agricultural Stabilization and Conservation Service. The
results of the Honey Creek Project, and other tillage demonstration projects
in the area, suggest that many of the area's soils are suitable for
conservation tillage management and that these methods, when properly
implemented, offer economic advantages to farmers (Honey Creek Joint Board,
1980; Allen County Soil and Water Conservation District, 1980).
The major impetus for transition to conservation tillage in this area is
likely to be the economic advantages these methods provide to farmers. In
adopting these methods, farmers are simultaneously implementing an
agricultural nonpoint source pollution abatement program in an area where
adverse impacts of agriculturally derived pollutants on a major water resource
have been documented. In recognition of these potential water quality
benefits, the use of water quality management resources to facilitate
-------
and successful transitions to conservation tillage in this region has been
initiated through agreements between Region V of the U.S. Environmental
Protection Agency and the Ohio Department of Natural Resources. Thus, this
region is about to move from demonstration projects to area-wide
implementation programs.
In anticipation of changing agricultural technology in this region, an
additional objective of the river transport studies has been to obtain
adequate background data upon which to evaluate the effectiveness of
best-management-practices in reducing sediment and nutrient export from large
agricultural river basins. To date, the effectiveness of conservation tillage
programs in reducing nutrient and sediment losses have been evaluated largely
in plot or small watershed studies. Many questions remain to be answered
concerning the effectiveness of these programs in reducing sediment and
nutrient yields from large river basins (Bachmann, 1980). In particular,
uncertainties about possible changes in delivery ratios, enrichment ratios,
and bioavailability create parallel uncertainties about reductions in nutrient
and s.ediment yields.
A major difficulty in documenting the environmental benefits of
conservation tillage programs is the large annual variability in sediment and
nutrient export by river systems. Consequently, considerable effort in these
studies has gone into characterizing the variability in sediment and nutrient
export at the transport stations. In addition a rainfall monitoring program
has been established in the Sandusky Basin to provide more information for
correlation with material export from the watersheds. Until the effectiveness
of agricultural nonpoint controls is evaluated in connection with major, large
scale implementation programs, efforts toward effective integration of point
and nonpoint source control programs will be hampered.
Support for the collection of the data, s.ets discussed in this report has
come from many sources. These include: 1 ) re'searcfi grants from the U.S. EPA
for studies of flow augmentation (16080DFO 01/71) and river transport (R
805436-01-02); 2) contracts with the Toledo Metropolitan Area Council of
Governments as part of their 208 Planning Study; 3) contracts with the Army
Corps of Engineers for data collection in support of the Lake Erie Wastewater
Management Study; 4) contracts with the Ohio Department of Natural Resources
for upground reservoir management studies; 5) grants from the Rockefeller
Foundation and the Soap and Detergent Association for studies of agricultural
nonpoint pollution; 6) grants from the cities of Tiffin, Bucyrus, and Upper
Sandusky for evaluating instream benefits of phosphorus removal programs; and
7) support from Heidelberg College for matching funds and data collection
during interim periods between external funding sources. The stream gaging
programs operated by the U.S. Geological Survey with support from the Ohio
Environmental Protection Agency, the Ohio Department of Natural Resources and
the U.S. Army Corps of Engineers provided the flow data essential to these
studies.
Given the support described above, it has been possible to develop what
may be the most detailed data sets available in the United States for
analyzing and characterizing the export of sediments and nutrients in large
agricultural river basins. A major use of this data has been within the Lake
Erie Wastewater Management Study (COE, 1979). The tributary loading data
developed by the Corps of Engineers and based in large part on northwestern
-------
Ohio rivers has been used by the International Joint Commission in developing
phosphorus management strategies for the Great Lakes Basin. The data has also
been used to develop and test river transport models as part of the Lake Erie
Wastewater Management study (Verhoff, et al., 1978). These data were also
used extensively in the Maumee Basin Pilot Watershed Report (Logan, 1978).
Data from the Honey Creek watershed has been used in connection with testing
the EPA's Nonpoint Source Model (Cahill, et al., 1979) and as part of an
EPA-sponsored statistical correlation study of sediment-pollutant relationship
(Zison, 1980). The Great Lakes Basin Commission selected the Sandusky Basin
Data sets for illustrating the application of the Watershed Model for
integrating point and nonpoint phosphorus control programs (Monteith, et al.,
1980). The data have also been used in the Ohio EPA's Sandusky Basin 208
Study, in the Toledo Metropolitan Area Council of Government's 208 Study, and
in local water quality planning programs by cities along the Sandusky River.
The data are available in the STORET system. Data through the 1977 water year
have been published by the Corps of Engineers (COE, 1978).
In this report background information on each of the watersheds is
presented along with the methods used in the collection of the river transport
data. The report includes both summary data on loading from each watershed,
as well as, detailed examples of the characteristics of nutrient and sediment
transport as it occurs in rivers of this region. Some implications of the
data for water quality management planning programs are presented. The
appendix includes a generalized river transport model based on an extension of
the transport models developed in the Lake Erie Wastewater Management study.
-------
SECTION 2
CONCLUSIONS
1 . The unit area yields of phosphorus and nitrate from the river "basins of
northwestern Ohio are high relative to other agricultural watersheds.
-Most Of the loading occurs during the short periods of time associated
with storm flows. This area is the largest contributor of
agriculturally-derived pollutants affecting the Great Lakes.
2. For these river basins the concentrations of sediments, total phosphorus,
and nitrates all tend to increase with increasing stream flow while the
concentrations of chlorides and conductivities decrease with increasing
flow. For all parameters studied except dissolved orthophosphorus, there
are large differences between flux weighted mean concentrations and time
weighted mean concentrations.
3. During individual storm events, the concentrations of sediments, total
phosphorus and TKN generally show advanced peaks relative to the
hydrograph peak while nitrates show a trailing peak. The latter is
probably due to the delayed arrival at the gaging station of nitrate-laden
tile effluents relative to surface runoff. Minimum conductivities and
chloride concentrations occur simultaneously with the peak flows. Both
resuspension associated with the passage of storm waves and routing of
surface runoff account for the advanced sediment and total phosphorus
peaks. _
4. Storms with equal peak flows can have widely differing flux weighted mean
concentrations of sediments, phosphorus, and nitrates. Furthermore the
flux weighted mean concentrations did not show a significant increase as
the peak flows increased. In contrast the flu.x weighted conductivity did
decrease as peak storm flows increased.
5. The large variability in nutrient and sediment loads for storms of equal
size suggest that material export is not limited by the transport capacity
of the rivers but rather by the movement of materials from the land
surface to the stream system. Thus, reductions in river export should
accompany nonpoint control programs that reduce material transport from
the land surface to the stream systems. Possible increases in stream bed
or stream bank erosion associated with the sediment carrying capacity of
the stream will not prevent the occurrence of load reductions associated
with control programs.
6. There are large annual variations in both loading and flux weighted
concentrations of sediments and nutrients at the transport stations. Both
the total annual runoff and the proportion of winter to summer runoff
affect the flux weighted mean concentrations in a given year. ?or most
parameters the summer and winter periods yield significantly different
concentration-flow relationships.
7- The total and particulate phosphorus to sediment ratios in individual
samples are highly variable and decrease as the sediment concentrations
increase. This gives rise to annual variations in the phosphorus to
sediment ratios, depending in large part on the proportion of winter to
summer runoff in a particular year.
-------
8. Given large data sets, calculations of mean annual loading using flow
duration tables are accompanied by small standard errors of the estimate.
Consequently, if agricultural nonpoint control programs reduce loading by
lowering the average concentrations occurring in various flow intervals,
such reductions should be detectable in the study basins. These data sets
provide baseline information upon which to judge the effectiveness of
agricultural nonpoint control programs.
9. The sediment delivery ratios in the study watersheds range from 6.2 to
11.9$ while average gross erosion rates vary from 4.2 to 9.4 T/ha/yr. The
delivery ratios are not correlated with the size of the watershed but are
inversely correlated with the gross erosion rate. There is no correlation
between gross erosion rates and unit area nonpoint phosphorus yields.
These observations raise doubts about the concept of "critical areas" and
the use of gross erosion rates in their identification. The sediment and
phosphorus yields from these large watersheds may be more related to the
amount of clay entrained by rain drop impact and subsequent surface
runoff. Tillage practices which increase cover and/or decrease runoff
could reduce sediment and phosphorus yields from areas of both high and
low gross erosion.
10. Several associated studies suggest that approximately 35!^ of the
particulate phosphorus loading from northwestern Ohio rivers is
bioavailable. 'Assuming that all of the dissolved reactive phosphorus is
bioavailable, about 50% of the total phosphorus loading from these rivers
is bioavailable.
11. Both concentration profiles and flux exceedancy data indicate that point
source phosphorus, most of which is in the form of soluble reactive
phosphorus when it enters streams, "is rapidly taken up into the stream
sediments. There is indirect evidence that less than 1001? of the
point-source -derived phosphorus is subsequently delivered out of the
stream system. Limited data suggests that the bioavailability of
particulate phosphorus exported during storms is no greater from
watersheds containing large point source phosphorus inputs than from
watersheds that are strictly agricultural. Apparently the bioavailability
of point-source-derived particulate phosphorus is no greater than that of
nonpoint-source-derived particulate phosphorus.
12. Most of the soluble reactive phosphorus exported during storm events is
derived from nonpoint sources. The short retention time of storm flows in
the lower portions of rivers or in estuaries, bays and the nearshore zone
may result in high deliveries of nonpoint-derived soluble phosphorus
through these zones to the open lake. In contrast, point-source
phosphorus entering these zones may be subject to deposition and
transformation to less bioavailable forms.
H. Concentration exceedency data at the transport stations show that both
nitrate and conductivities exceed the state's drinking water standards
from two to five percent of the time. The highest nitrate concentrations
occur in the spring and early summer when the rainfall duration and
intensity result in a high proportion of tile effluent in comparison with
surface runoff. Highest conductivities occur during winter low flow
-------
periods.
14. Quality control testing done in connection with these studies show that,
although significant changes in concentrations of soluble phosphorus and
nitrate do occur during one week of storage without preservation, the
changes are small relative to the errors in loading calculations
associated with flow measurements or errors that would be introduced by
less frequent sample collection.
-------
SECTION 3
RECOMMENDATIONS
Documentation of the effectiveness of conservation tillage programs in
reducing nutrient and sediment export from large agricultural river basins is
needed before large scale, integrated, point-source/nonpoint-source control
programs can be initiated. Existing demonstration projects in this^region
have shown that conservation tillage can be successful from the standpoint of
crop production. The actual acreages currently committed to these management
practices are, however, small in comparison to the extent of conventional
tillage. Data at the transport stations discussed in this report are
essentially baseline data reflecting nutrient and sediment export under
conventional tillage programs.
Given the economic advantages of conservation tillage programs on several
of'the important soil types in this region, farmers throughout this area are
likely to be adopting these practices. This will simultaneously constitute a
gradual, area wide implementation of an agricultural nonpoint pollution
control program. Evaluation of the environmental impacts of such a program
will be difficult and require long term studies.
¥here funds for water pollution abatement are being used to foster the
adoption of conservation tillage programs, substantial portions Of these funds
should be directed toward specific, large watersheds. This should result^in a
more rapid conversion to conservation tillage in these watersheds and allow
for an earlier assessment of environmental benefits, as well as any unforeseen
environmental problems, that could accompany conservation tillage technology.
Such an approach would also provide information on the feasibility and
procedures for achieving a high proportion of conservation tillage in large
watersheds.
In environmental assessments of conservation tillage programs in large
watersheds, measurements of the bioavailabilit'y of particulate phosphorus and
of particle size for suspended solids should be included. Pesticide runoff
measurements should be added since the tillage programs may well be
accompanied by an increase in the amount of pesticide usage. The impact of
changing agricultural technology on biological communities in the stream
ecosystem should be investigated in the assessment programs. It will also be
necessary to monitor the changing distribution of tillage practices, both in
the targeted watersheds and in adjacent control watersheds.
Data sets of the type developed in the Sandusky Basin should also be used
to explore possible cost savings through time-variable management at municipal
sewage treatment plants. Integrating flow augmentation from existing upground
reservoirs with both point and nonpoint controls would provide the most
effective and economical water quality management programs for this region.
The implementation of the Watershed Model, as developed by the Great Lakes
Basin Commission, should be considered for the Sandusky Basin. The concepts
of the model, as it is applied to phosphorus, should be extended to other
water quality parameters.
-------
Finally it is recommended that the U.S. EPA establish a Long Term
Environmental Research Center that would specifically study the effects of
changing agricultural technology on regional water quality. At present,
predictions of improvements in river water quality at public water intakes or
of reductions in tributary loading are largely based on extrapolations of
research on small plots and individual fields. Many uncertainities with
respect to material processing and delivery in stream systems exist. Only
through long term, detailed studies in an agricultural region undergoing a
large scale transition to conservation tillage can the environmental benefits
of agricultural nonpoint pollution control programs be documented. The
complexity and scope of the needed research warrants the establishment of a
Long Term Environmental Research Center focusing on this topic.
-------
SECTION 4
DESCRIPTION OP THE STUDY AREA
LOCATION OF THE STUDY BASINS
This report includes data collected in the Maumee, Portage, Sandusky and
Huron River Basins in northwestern Ohio. The Maumee and Portage drain into
the western basin of Lake Erie, and the Sandusky and Huron empty into the
southwestern corner of the central basin of the Lake (Figure 1). The drainage
basins of these rivers and the associated areas draining directly into the
lake have a total area of about 25,000 sq. km. Or 45$ of the land in the
United States which drains into Lake Erie. It has been estimated that these
rivers account for 73$ of the agriculturally derived sediments which enter
Lake Erie from the United States and for 39$ of the agriculturally derived
sediments entering the entire Great Lakes system from the United States
(PLUARG, 1974). The primary focus of this study is the Sandusky River Basin
which contains a network of 10 river transport stations.
GEOLOGY
The geology of the Sandusky River Basin has been described in detail by
Forsyth (1975). The basin occupies glacial plains composed of ground moraine
crossed by the Defiance, Fort Wayne, and Vabash end moraines (Figure 2).
North of Tiffin and in several isolated areas in the southern part of the
Basin, the glacial till is overlain by glacial lake sediments. Alluvial soils
are restricted to narrow bands along the stream systems. The underlying
bedrock is mos'tly Silurean-Devonian carbonates, with Devonian shale and
sandstone;in the southeast (Figure 3). The topography is relatively flat
except for the areas occupied by the end moraines and where streams have cut
through the glacial till.
j _^
SAMPLING STATIONS
Most of the stream transport studies have been conducted at the 12 USGS
stream gages listed in Table 1 . The stations on the Maumee, Portage and Huron
Rivers and the Fremont station On the Sandusky are located at the stream
gaging stations nearest Lake Erie where tributary loading data for the lake
are collected. A short distance downstream from these gaging stations
esturine-like conditions develop in each of these rivers. On the Maumee,
Portage and Huron rivers the samples were collected at chemical monitoring
stations located near the stream gage rather than at the gage itself. In the
Sandusky network of stations, all of the samples were collected at the stream
gaging stations. The locations of the river mouth stations and the Sandusky
network of stations are shown in Figures 4 and 5. The drainage area, period
of hydrological record, and average discharge for each station are listed in
Table 1 .
-------
MICHIGAN
I
N
D
I
A
N
A
YORK
Huron
[PENNSYLVANIA
LEGEND
LAKE ERIE BASIN BOUNDARY
LAKE ERIE SUB-BASIW BOUNDARIES
-RIVER BASIN BOUNDARIES
Sandusky
Portage
Figure 1. Location of the northwestern Ohio river basins in relation to Lake Erie.
-------
M
Figure 2. Glacial map of the Sandusky River Basin area. Glacial
deposits shown are end moraines (enclosed by hashures and
named), ground moraine (left white), lake beaches (marked
by black bands and named), and lake-bottom silts and clays
(identified by horizontal-dash pattern). Areas of alluvial
deposits along the Sandusky River and its tributaries are
too small to be shown on this map. (After Forsyth, 1975).
11
-------
Figure 3. Bedrock geologic map of the Sandusky Basin area.
Bedrock units are identified by the following symbols.
M - Mississippian Berea Sandstone (and other Mississippian
units)
Do - Devonian Ohio Shale
Dcd - Devonian Colutibus and Delaware Limestones
Stm - Silurian Tymcchtee and "Monroe" Dolostones
Sg - Silurian Greenfield Dolostone
Sn - Silurian Niagaran-aged Lockport Dolostone
All these rock unit:; are dipping very gently to the east, as
a result of their location on the east limb of the Cincinnati
Arch. (After For^yth, 1975).
12
-------
Table 1. Locations, station numbers, areas, periods of hydrological record, and average discharge for stream
transport stations in northwestern Ohio.
Stream
Maumee
Portage
Sandusky
Broken
Sword
Sandusky
Tymochtee
Sandusky
Honey Creek
Wolf Creek
Wolf Creek
Sandusky
Huron
Gaging Station
Location I.D.
at Waterville 04193500
at Woodville 04195500
near Bucyrus 04196000
near Nevada 04196200
near Upper 04196500
Eandusky
at Crawford 04196800
near Mexico 04197000
at Melmore 04197100
at Bettsville 04197300
near Bettsville 04197450
near Fremont 04198000
at Milan 0419000
Sampling Station Area
Location I.D. 2
Km
near Waterville 04193490 16,395
in Woodville 04195600 1,109
same as 04196000 230
gaging station
same as 04196200 271
gaging station
same as 04196500 722
gaging station
same as 04196800 593
gaging station
same as 04197000 2,005
gaging station
same as 04197100 386
gaging station
same as 04197300 171.5
gaging station
same as 0'4197450 213
gaging station
same as '04198005 3,240
gaging station
Below Milan 04199100 961
Period of Average
Record Discharce
Yrs. m /s mro/vr.
54 136.4 262
47 8.836 25!
38 2.438 355
3.5 2.45** 285
55 6.882 281
15 4.814 255
53 16.40 258
3.5 3.32** 271
3.5 1.42** 261
3.5 2.26** 334
53 27.15 264
29 8.524 28C
*Period of record through the 1979 water year.
**Estimated by comparison to nearby long term stations.
13
-------
LEGEND
LAKE ERIE BASIN BOUNDARY
RIVER BASIN BOUNDARIES
• SAMPLING STATION
Figure 4. Location of tributary mouth sampling stations for
Lake Erie loading studies from northwestern Ohio
rivers.
WATERSHED CHARACTERISTICS
As part of the Lake Erie Wastewater Management Study a computerized
land-use/land-capability data base was assembled for the entire United States
portion of the Lake Erie Basin (COE, 1979). The data base was organized using
the Land Resource Information System (LRIS) as developed by Resource
Management Associates (Bliss et al., 1975) This system is based on a cellular
data file containing information on land use, soils, and locations including
watershed and county affiliations. Within each cell the information is coded
for a particular point. The locations of the points are determined using a
systematic unaligned approach in which randomly selected coordinates within
the cells of the first row and column of the cellular grid establish the
position of the points in the remainder Of the cells (Bliss et al., 1975).
14
-------
f /' xf'
/ ' x y * ~~N >
/ j[ WOLP}'' fl I I SANDUSKY AT\
' w \ t i i I
»UPPER
TYMOCHTEE ,'SANDUSKX
i f
f
I
V
\ —
LEGEND
SANDOSKY RIVER BASIN BOUNDARY
— RIVER SUB-BASIN BOUNDARIES
• SAMPLING STATION
Figure 5. Location of sampling stations in the Sandusky Basin network
transport stations.
15
-------
In the Sandusky Basin, which was selected for more detailed studies, the
cells are 300 m on a side (9 hectares) except for the Honey Creek and Rock
Creek watersheds and the Sandusky County portion of the basin where the cells
are 200 m on a side (4 hectares). Thus in the Sandusky Basin above the
Fremont gaging station, approximately 45,000 cells and points have been
encoded. For the Portage River Basin 4 hectare cells were used while for most
Of the Maumee and Huron River Basins ?6 hectare cells were encoded.
The LRIS system has been used to produce a variety of summary tables and
maps for watersheds and counties. These tables Or maps can reflect single
features, such as land use, slopes, or soil texture. They can also reflect
co-occurrences of features, such as cropland on sloping soils with high
erodibility (RMA, 1979).
The land use in the watersheds above each of the sampling stations is
summarized in Table 2. Land use was encoded through photointerpretation of
Table 2. Percentage distribution of major land uses in the study basins.
TotaJ Area
Location km
Maumee, Waterville
Portage River,
Woodville
Huron River, Milan
Sandusky River ,
Fremont
Sandusky River,
Mexico
Sandusky River,
Upper
Sandusky River,
Bucyrus
Tymochtee Creek,
Crawford
Honey Creek, Melmore
Broken Sword Creek,
Nevada
Wolf East, Ft. Seneca
Wolf West, Bettsville
16,395
1,109
961
3,240
2,005
772
230
593
386
217
213
171.5
Cropland
%
75.
85.
75.
79.
80.
78.
73.
84.
82.
84.
81.
83.
6
5
3
9
3
0
3
0
6
7
9
3
Pasture
»
3.
3.
3.
2.
2.
3.
2
6
5
3
3
4 '
4.9
1.
0.
I.
2.
1.
-1
6
4
7
4
Forest
%
3
5
12
8
~B
9
9
7
10
8
6
4
.4
.6
.5
.9
.7
-1,
.4
.6
.0
.5
.3
.7
Water
%
3
0
2
2
2
2
2
2
0
1
2
3
.5
.9
.2
.0
.1
.0
.1
. 3
.5
.3
.0
.1
Other
%
9.
4.
6.
6.
6.
7.
10.
4.
6.
4.
7.
7.
4
3
4
8
6
5
2
8
3
1
0
6
high altitude infrared photographs taken by NASA, Lewis of Cleveland, Ohio in
June, 1976. The 70,000 to 1 scale photographs were interpreted by the
Environmental Research Institute of Michigan, Ann Arbor, Michigan'. The ' land
use categories presented in Table 2 are composite categories of more than SO
separate land uses encoded from the photographs. The dominance of cropland
for all of the watersheds is evident.
16
-------
Soils information was taken from modern soil surveys where these were
available. The soil phase and slope classifications were recorded for each
point. A soils properties file containing the physical/chemical properties
for each soil was used in conjunction with the soil phase data.
The percentage of land in each watershed falling within various slope
classifications is shown in Table 3. For each soil phase, unique slope values
were assigned. These values represented the median value of all slopes
observed on that soil during the 1« National Erosion Survey of 1977-78
conducted by the Soil Conservation Services. The dominance of level and
gently sloping lands within the region is evident. There are, however, some
substantial differences in the proportions of soils in the 0.2!?, \% and 3.5%
slope classes among the various watersheds.
Tables 4-, 5 and 6 show the composition of these watersheds in terms of
soil textures, drainage characteristics and soil erodibility. All of these
are taken directly from the soil properties files. Generally fine textures,
poor drainage and high erodibility characterize all'Of the watersheds. The
Maumee, Portage, and ¥olf-¥est basins have somewhat finer textured soils than
the Sandusky and Huron basins. They also have higher percentages of soils in
the very poorly drained category. In all of the watersheds, tile systems are
used extensively to improve drainage and to allow timely field operations.
Co-occurrences of the above watershed features have been published for
all of the major "river basins of Lake Erie and for all of the Sandusky River
Basin watersheds (RMA, 1979 b, c). These dual feature summaries include:
*
1. Land Use and Soil Permeability
2. Soil Permeability and Slope
3. Land Use and Slope
4- Land Use and Soil Texture
5. Soil Texture and Slope ' -<-
6. Slope and Soil ErOdability
7- Soil Texture and Erodability
8. Land Use and Erodability
9. Land Use and Soil Drainage Class
10. Land Use and Soil Capability Class
The LRIS system has also been used for producing colored maps which
display such features as land use, surface texture, slope and potential gross
erosion for the Lake Erie Basin (COE 1979) and for the Honey Creek watershed
(Cahill and Pierson, 1979). Suitability maps for contour cropping and no-till
corn production have also been produced for Honey Creek. As part of the LE¥MS
study, sets of maps and summary tables are being produced for each county in
the U.S. portion of the Lake Erie Basin. In addition the maps and tabular
summaries will be aggregated into watershed boundaries for 5 river basins, one
of which is the Sandusky Basin. These maps are extremely useful in
identifying potentially critical areas and in planning appropriate nonpoint
pollution abatement programs.
17
-------
Table 3. The areal percentage of lands in various slope classifications for the study watersheds.
less than
Watershed 0.2% 0.5% 1.0% 2
Maumee 42.7% 9.9% 18.7% 1
Portage 57.2 13.9 1.5 1
Sandusky 15.9 8.6 35.2 0
(Bucyrus)
Broken Sword 11.1 8.9 33.6 0
Sandusky 18.4 9.3 31.3 0
(Upper Sandusky)
Tymochtee 34.0 8.4 35.7 0
Sandusky 21.4 8.9 34.4 0
(Mexico)
Honey Creek 6.9 7.8 49.0 4
(Melmore)
Wolf Creek 46.3 1.7 41.7 0
(West)
Wolf Creek 16.2 2.7 63.4 0
(East)
Sandusky 19.0 8.2 38.8 0
(Fremont )
Huron 2.7 11.0 18.0 43
Slope Classification
.0% 3.5% 6-8% 9-11% 12-14% 15-17%
.5% 23.8% 0.1% 2.3% 0.2% 0.5%
.4 23.5 1.7 0.9 0.0 0.0
.0 33.7 0.0 4.8 0.0 0.6
.0 40.6 0.0 5.4 0.0 0.1
.0 34.7 0.0 4.6 0.0 0.8
.1 17.9 0.0 2.4 0.0 1.0
.1 29.2 0.0 3.9 0.0 1.1
.6 29.2 0.7 1.7 0.0 0.0
.0 10.3 0.0 0.0 0.0 0.0
.0 17.7 0.0 0.0 0.0 0.0
.8 28.8 0.1 3.1 0.0 0.7
.8 7.8 13.9 0.7 1.5 0.3
18%
or greater
0.3%
0.0
1.1
0.3
0.9
0.6
1.0
0.0
0.0
0.0
0.6
0.4
Table 4. Percentage distribution of soil textures in the study watersheds.
Silty
Silty Clay Clay
Watershed Clay Clay Loam Loam
Maumee 18.5% 4.5% 23.8% 1.4%
Portage 46.3 0.0 10.5 2.3
Sandusky 0.0 0.0 15.9
(Bucyrus )
Broken Sword 0.0 0.0 13.5
Sandusky 1-0 0.3 19.0
(Upper Sandusky)
Tymochtee 4.4 3.3 34.1
Sandusky 1-5 1.1 22.3
(Mexico)
Honey Creek 0.0 0.2 18.7
(Melmore)
Wolf Creek 0.0 0.0 52.3
(West)
Wolf Creek 0.0 0.2 31.0
(East)
Sandusky 1.3 0.'' 22.2 0.2
(Fremont)
Huron 0.0 0.4 16.1 0.0
Sand ' "~ Fine
Clay Silty N Sandy Sandy Fine
Loam Loam Loam Loam Loam Sand
0.4% 15.0% 26.2% 4.2% 2.0% 0.6%
6.6 11.3 12.4 0.9 3.6 0.1
1.3 82.1 0.0 0.8
3.7 81.5 0.1 1.2
3.1 75.2 0.1 1.3
0.6 56.4 0.6 0.5
3.0 70.2 0.4 1.1
2.2 76.1 0.0 0.5
4.9 42.4 0.0 0.2
4.7 63.4 0.0 0.2
0.0 4.7 68.0 0.3 1.1 0.1
0.0 11.1 65.6 0.6 2.6 0.1
Loamy
Fine
Sand Muck
1.6% 1.2%
6.0 0.0
0.0 0.0
0.0 0.0
0.0 0.1
0.0 0.2
0.0 0.3
0.0 2.2
0-0 0.0
0-4 0.0
°-5 0.5
1-6 ;.o
18
-------
Table 5. Percentage distribution of soils falling into various drainage classifications for the
study watersheds.
Very
Poorly
Watershed Drained
Mautoee
Portage
Sandusky
(Bucyrus)
Broken Sword
Sandusky
(Upper Sandusky)
Tymochtee
Sandusky
(Mexico)
Honey Creek
(Melmore)
Wolf Creek
(West)
Wolf Creek
(East)
Sandusky
(rroont)
Huron
48.7*
69.2
21.6
14.7
20.3
37.6
23.7
11.7
43.5
15.7
21.6
17.6
Poorly
Drained
1.9%
0.3
7.9
8.4
6.0
1.9
5.7
13.8
13.2
16.9
7.9
8.5
Somewhat
Poorly
Drained
33.0%
23.1
52.6
51.7
46.5
48.2
47.7
64.4
39.8
65.4
52.6
46.0
Moderately
Well
Drained
9.3%
4.6
14.3
20.4
22.5
11.4
19.5
7.6
0.5
0.9
14.3
17.6
Well Somewhat
Drained Excessively
Drained
7.0% 0.1%
2. a 0.0
3.5 0.0
4.7
4.2
0.8
3.5
2.5
3.0
1.1
3.5 0.0
10 . 2 0.1
Excessively
Drained
0.1%
0.0
0.0
0.0
0.0
Table 6. Percentage distribution of coil credibility »« indicated by the K-Talue of soil* in the itudy
watersheds.
Watershed
Maumee
Portage
Sandusky
(Bucyrus)
Broken Sword
Sandusky
(Upper Sandusky)
Tymochtee
Sandusky
(Mexico)
Honey Creek
(Melmore)
Wolf Creek
(West)
Wolf Creek
(East)
Sandusky
(Fremont )
Huron
0.10
1.4%
0.0
0.0
0.0
0.1
0.2
0.3
2.3
0.0
0.0
0.5
2.0
0.15
0.1%
0.0
0.0
0.0
0.1
0.5
0.3
0.0
0.0
0.0
0.1
0.0
0.
2.
6.
0.
0.
0.
0.
0.
0.
0.
0.
0.
2.
17
8%
5
0
1
1
4
3
2
2
4
8
0
K - Value
0.20 0.24
1.4%
2.7
0.8
0.7
0.8
0.0
0.4
0.2
0.2
0.2
0.7
0.1
17.
7.
12.
8.
13.
23.
15.
1.
1.
1.
9.
4.
5%
9
3
3
7
6
8
9
6
1
9
6
0.28
32.1%
57.0
5.6
8.0
8.7
13.7
9.5
9.1
44.7
15.1
13.2
7.2
0.32
2.1%
2.6
10. 5
7.0
8.0
0.9
4.9
4.0
1.9
1.6
4.4
20.6
0.37
10.0%
6.1
35.8
34.6
28.6
12.0
24.7
50.8
16.7
21.7
26.7
20.5
0.43
32.4%
17.0
35.0
41.3
39.9
48.6
44.0
31.6
34.8
59.9
43.5
40.7
0.49
0.2%
0.0
0.0
0.0
0.0
0.0
0.0
o.o
0.0
0.0
0.0
2.4
19
-------
POTENTIAL GROSS EROSION
The LRIS system has been used as a base for calculating potential gross
erosion in the Lake Erie watersheds (Urban et al. 1978). The universal soil
loss equation (Wischmeier & Smith, 1978) was used to calculate gross erosion.
The various components of the equation were estimated as follows (Urban et
al., 1978):
- R factor data was taken from USLE Handbook 282 and
developed on a county basis.
- K factors were taken from LRIS soils file by soil type.
- SL Slope percentage, S, was developed from LRIS soil
phase data by taking the median value for slope range
given. Slope length was estimated from local SCS
experience and the recent ^% National Erosion Survey.
- C factors were developed from county-level estimates of
crops grown; rotations were developed for each county
based on local interpretation.
- P factor was assumed to be 1, i.e. there were no
supporting conservation practices.
The gross erosion was calculated for each LRTS cell under current
cropping practices. "For each watershed, gross erosion was taken as the sum of
the erosion fr6m each cell within the watershed. A variety of scenarios were
then incorporated into the calculations, such as the effect of no-till on
suitable soils, minimum till, winter cover, and-reduction of erosion to the
soil loss tolerance factor. For each watershed and county, tabular summaries
of the various scenarios as they apply to various s-Oil management groups and
land uses are presented in an appendix to the report of Urban et al. (1978).
A summary of the gross erosion data for the study watersheds is shown in
Table 7. For each watershed the average potential gross erosion rate for
cropland is listed, along with the cropland area and unit area erosion rate.
For the cropland in each watershed the soils are divided into soil management
groups with respect to their suitability for no-till crop production. For
each group the percent of the cropland gross erosion, the percent of the
cropland area and the unit area erosion rate is listed. Groups I and IV are
well drained soils which are well suited for no-till crop production. Groups
II and VII are suitable if underdrained by tile systems. Most such soils in
the region are currently tiled. The remaining groups are considered
unsuitable for no-till production. The bottom rows in Table 7 include
vineyards, grassland and forests along with the cropland. The combined total
gross erosion, area and unit area erosion for all of the above land uses are
listed.
It can be seen from Table 7 that average potential gross erosion rates
for cropland in the study watershed range from 4.5 to 10.8 metric tons per
hectare per year. In each watershed the potential gross erosion rates are
greatest in the soils which are most suitable for no-till (Groups I and IV)
and next highest in the Groups II and VII soils which are also suitable for
20
-------
Table 7. Summary of Gross Erosion Rates by Soil Management Groups for the Study Watersheds.
Maumee
Watershed Cropland
Gross Erosion M.T./Yr. 8246849.4
Cropland (ha) 1050884.7
Erosion Rate M.T./ha/Yr 7.84
Soil Groups I £ VI
% Gross Erosion 36%
% Watershed Cropland 13%
Erosion Rate 21.09
Soil Groups II & VII
% Gross Erosion 38%
% Watershed Cropland 32%
Erosion Rate M.T./ha/Yr 9.41
Other Groups
% Cross Erosion 26%
% Watershed Cropland 55%
Erosion Rate M.T./ha/Yr. 3.72
Watershed, Nonurban
Gross Erosion M.T./Yr. 8279826.1
Cropland (ha) 1211552.8
Erosion Rate 6.84
Portage
478012.5
86248.8
5.54
21%
6%
18.67
50%
33%
B.47
29%
61%
2.64
478959.7
95762.3
5.00
Huron River
at Milan
650768.8
71635.2
9.08
45%
18%
21.99
33%
38%
7.87
22%
44%
4.71
655230.9
87334.2
7.51
San dusky
at Fremont
1761999.9
188343.3
9.35
33%
15%
21.09
46%
53%
8.20
21%
32%
5.85
1769011.0
214705.5
8.25
Sandusky
at Mexico
1156798.7
108403.7
10.67
40%
19%
22.57
37%
46%
8.54
23%
35%
7.01
1161689.2
124008.2
9.37
Sandusky
at Upper
Sandusky
/
533241.4
49215.9
10.82
40%
21%
20.39
38%
47%
8-70
22%
32%
7.64
535602.5
57252.6
9.35
Sandusky
at Bucyrus
150615.2
16173.3
9.32
38*
19%
18.22
45%
53%
7.89
17%
28%
5.85
151457.8
19332.4
7.84
Tymochtee
Creek
248897.9
26838.5
9.28
32%
10%
30.75
39%
44%
8.27
29%
46%
5.76
249550.9
29691.5
8.40
Broken
Sword
174830.9
16698.3
10.47
42%
23%
18.71
47*
52%
9.44
11%
25%
4.84
175197.1
18651.4
9.39
Wolf West
56561.5
12553.3
4.48
6%
3%
8.18
60%
40%
6.66
34%
57%
2.69
56428.2
13433.3
4.19
Wolf East
55913.6
9935.2
5.63
5%
3%
8.09
79%
63%
7.06
16%
34%
2.71
56008.9
10967.2
5.11
Honey
Creek
at Melmore
246564.5
31680.6
7.78
22%
10%
16.85
67%
64%
8.20
11%
26%
3.36
247233.8
36036.7
6.86
-------
no-till. In general the soils which are unsuitable for no-till have low gross
erosion rates. The effects of a variety of BMP's on the erosion from each
soil management group for each watershed are shown in more detail in
Application of the Universal Soil Loss Equation in the Lake Erie Drainage
Basin", Appendix I (Urban et al., 1978).
22
-------
SECTION 5
STUDY METHODS
SAMPLE COLLECTION
Most of the samples upon which this report is based were collected at the
USGS stream gages and/or chemical monitoring stations listed in Table 1 . Each
station is Outfitted with a pumping system of the type used by the USGS for
continuous monitoring of oxygen, temperature, conductivity, and pH. In a
typical installation, two 4" diameter plastic pipes lead from the gage house
to the stream. The longer, upstream pipe contains the inlet line(s) and the
shorter down stream pipe contains the drain line(s). Both the inlet and drain
lines are 1" diameter polyethylene and are wrapped with heat tape to prevent
freezing in the winter. The end of the inlet pipe is located in that portion
of the stream that flows year round (even during low flow conditions) and is
~10 cm off the stream bed. A stainless steel inlet screen (18" long 1/4"
diameter holes) prevents larger debris from clogging the inlet pipe. The
inlet line extends to the end Of this screen.
The stations are equipped with a continuously Operating pumping system
(Continental, model EC448, progressive cavity screw pump) with an output of
about 8 gallons per minute. The output of the pump leads into a. reservoir
designed to drain completely should the pump fail. This prevents repeated
sampling of the same water if the pump fails.
Automatic samplers (iSCO model 1680 or equivalent) are used to collect
discreet samples at preset intervals of time from the reservoir Of the
continuous pumping system. These samplers are capable of sampling at
intervals ranging from 1 to 999 minutes. Under non-event conditions the
samplers are set to take four 450 ml samples per. day at 1:00, 7:00, 1^:00 and
19:00 hours. For the smaller watersheds during storm events the sampling
frequency is increased to 8 to 12 samples' per-*- day. To avoid sample
contamination the samplers reverse pumping before and after each sample is
taken. The samplers are run on batteries which are "trickle charged" from the
electric power at the station. Thus power Outages do not cause the timer to
fail and sampling will resume on schedule with the proper sample bottle in
place when the current comes back on and the pump resumes operation.
The sampler bases have room for 28 high density polyethylene bottles.
Sample identification is based on bottle sets and bottle numbers. The bottles
are numbered from one to 28 and each set has a letter code. The cleaning
procedure for the nutrient bottles includes a hot tap water wash with brushing
to remove any sediment. This is followed with two rinses with hot tap water
and air drying before use. The caps go through the same washing procedure.
~The sampling stations are visited weekly except for the smaller
watersheds which, during runoff events, are visited more often, ^he 2B
bottles in the base allow 4 collections per day for a seven-day period. mhe
normal schedule for visiting the stations involves changing the samnler bases
on Mondays Or Tuesdays so that sample analyses can be completed by ™hursdavs
and the sample bottles cleaned and prepared for the collection routes on the
following Mondays and Tuesdays.
23
-------
While at the sampling station, the field technicians collect several
extra samples as part of the quality control program. Three replicate samples
are collected from the pumping system. One of_these is placed in the sampler
base for the following week's collection. This stored sample is used to
monitor the effects of one week of storage at ambient temperatures within the
sampler bases. Thermostated heaters prevent the gage house temperatures from
dropping below 3> degrees Centigrade in the winter period. The other two
pumped samples are returned to the laboratory on the day of collection. "Both
are analyzed and the resulting data used for determining laboratory precision,
and for comparison with the sample which has been stored and analyzed the
following week. Periodically the field technicians also collect grab samples
from a bridge near the sampling station for comparison with the pumped
samples. The results of these quality control procedures are presented later
in this section of the report.
The field technicians also check the Operation of the stage recording
equipment and compare the bubble gage reading with a nearby wire weight gage.
The pump Operation is checked and the inlet screens cleaned if necessary. The
sampler timers are checked and adjusted. Pump lines for the automatic samples
are cleaned and replaced as necessary.
SAMPLE PRESERVATION AND STORAGE
No preservatives-are used in the sample bottles. The samples remain in
the sampler bases at ambient gage house temperatures as they are collected.
The "age" of the samples at the time they are delivered to the laboratory
ranges from 7 days to 4 hours. Within 48 hours of delivery to the lab, all
samples are filtered through prerinsed 0.45 micron membrane filters (Hillipore
HAWP) and transferred to autOanalyzer tubes. Analyses of both soluble and
total nutrients are completed within three days of delivery to the laboratory.
'> -^
For studies of this type, it would be impractical to attempt to meet the
preservation and storage requirements for soluble nutrients, such as soluble
reactive phosphorus, nitrates and ammonia. Instead our approach has been to
document the extent and direction of change during sample storage and, where
necessary, to take these effects into account when interpreting the data. The
results of our studies of storage effects are presented in the quality control
section.
LABORATORY PROCEDURES
Selection of Samples for Analysis
The automatic samplers operate continuously and collect a minimum of 4
samples per day. All of the samples are returned to the laboratory each week.
If a runoff event was in progress at the sampling station, as evidenced by
either a high flow or high turbidity, all of the samples are analyzed. If
runoff events were not occurring at the station, One sample per day is
analyzed and the remainder are discarded. Thus, data are collected on both
high and low-flow conditions but the frequency of analyses is much greater for
high-flow samples.
24
-------
Sample Handling
Figure 6 is a summary flow chart showing, the groupings of analyses
performed on the samples. Not all analyses are performed on each sample.
Analytical Methods
The analytical methods used in the laboratory are summarized in Appendix
1 . More detailed descriptions of the methods are contained within the
laboratory's Handbook on Quality Assurance.
QUALITY CONTROL
Precision
The precision associated with the sampling and analysis program described
above can be judged from the analyses Of replicate pumped samples. These data
are summarized in Table 8, which includes data collected during the 1 978 water
year. The number of replicate pairs in the data set for each parameter is
listed. Since the precision calculations are used for establishing control
limits, the indicated number of pairs with the largest differences were
deleted from each data set. In general less than 5% of the pairs were
rejected. The mean difference between the replicates is shown for each
parameter. Dividing the mean difference by 1.128 provides an estimate of the
precision (ASTM, 1976). To indicate the relationship between the precision
and the range of environmental values, the mean of the 5th percentile
concentration for the Maumee, Fremont and Melmore stations is listed as is the
mean of the Q^th percentile. These values were • obtained from concentration
exceedency tables for each station. The _ ratio of the 90? inclusive range
(ie. 5% to 95$) to the precision indicates that' the precision is very good in
relation to the range of values encountered in environmental samples.
The r-squared values (coefficient of determination) also shows the close
agreement between the replicate pairs.
Effects of Sample Storage
In Table 9, the effects of one week of sample storage on analytical
values are shown. The stored values were obtained from analyses of the
samples that were placed in the sampler bases for One week while the fresh
samples were returned for analyses with the previous week's collection. The
data were analyzed using the paired T-test procedure contained within SPSS
(Statistical Package for the Social Sciences). For each parameter except
suspended solids, changes did occur during storage that were, from a
statistical standpoint, highly significant. The means Of the fresh and stored
values are shown along with the mean differences, correlations, T-values, and
associated probabilities.
The extent of change during storage is summarized in Table 10.
largest changes occur in soluble reactive phosphorus where the average change
was a 10.45 increase. For this parameter the pairs were broken down into
25
-------
Figure 6.
FLOW CHART FOR SAMPLE ANALYSES
FILTER
.45 nu
membra
filter
[
R
lc rcn
aae
r
ELECTHODES
25' C
fii
Conduct ivit
ECTTT.K
1
L_
SUSPENDED
SOLIDS
qlass
ciber
filter
TOTAL TOTAL METALS
PHOSPHORUS X.TFI PAHJ.
MITWKjEN
Table 8. Precision Data based on analysis of replicate pumped samples.
Pairs Deleted Mean tl^nvironmental Range*
Parameter N Values Difference Precision R 5 percentile 95f>ercentil
W/1 »g/l mg/l ng/1
Soluble Reactive 292 15 .0066 .0058 .992 .020 .187
Phosphorus
Total Phosphorus 301 12 .0108 .0096 .993 .099 .580
Nitrate/Nitrite - N 305 14 .0708 .0628 .997 .397 9.087
Ammonia - N 304 14 .0407 .0360 .966 .020 .824
Suspended Solids 305 14 4.49 3.98 .994 4.38 264.3
Conductivity (umhos) 319 14 11.25 9.92 .989 367.7 960.0
90% Ranqe
e Precision
(ratio;
25
50
138
22
58
60
•Based on the mean 5th percentile concentrations for the Maumee, Fremont and Melmore stations and mean 95th
percentile concentrations.
26
-------
Table 9. The effects of one week of sample storage on analytical values.
Variable Number
of Pairs
Fresh SRP
711
Stored SRP
Fresh TP
762
Stored TP
Fresh NO23
775
Stored NO23
Fresh NH3
761
Stored NH3
Fresh SS
777
Stored SS
Mean
0.1055
0.1165
0.2380
0.2346
3.2145
3.2846
0.2497
0.2325
46.8099
46.4144
Standard
Deviation
0.132 •
0.133
0.208
0.207
2.769
2.729
0.373
0.383
99.109
99.583
T-TEST
Mean Standard Corr. 2 -Tail T 2-Tail
Difference Deviation (r) Prob. Value Prob.
-0.0110 0.030 0.975 0.000 -8.87 <0.001
1
0.0033 0.024 0.993 0.000 3.79 <0.001
-0.0700 0.454 0.986 0.000 -4.25 <0.001
0.0172 0.109 0.959 0.000 4.37 <0.001
0.3955 16.527 0.986 0.000 0.67 0.505
Fresh Cond. 719.2715 232.236
766
Stored Cond. 722.3368 231.415
-3.0653
20.979
0.996 0.000 -4.04 <0.001
-------
Table 10. Average Change in Concentration during one week of sample storage.
Parameter
SRP
all values
SRP < 0.050
0.050 < SRP < 0.100
SRP >0.100
SRP if SS <50
SRP if SS >50
SRP if Cond. <500
SRP if Cond. >500
TP
NO - NO
NH3
SS
Conductivity
N
711
226
232
242
540
171
107
604
762
775
761
777
766
Mean Fresh
mg/1
0.1055
0.0260
0.0730
0.2127
0.1089
0.0948
0.0942
0.1075
0.2380
3.2145
.2497
46.810
719.27
Mean Stored
mg/1
0.1165
0.0465
0.0815
0.2172
0.1189
0.1089
0.0967
0.1200
0.2346
3.2846
.2325
46.414
722.34
% Change
in Fresh
+ 10.4%
+ 78.8%
+ 11.6%
+ 2.1%
+ 9.2%
+ 14.9%
+ 2.7%
11.6%
1.9%
+ 2.2%
- 6.9%
- 0.8%
+ 0.4%
various ranges of fresh SRP concentrations as well as suspended solids and
conductivity ranges. The largest changes (+78.8$) occurred when the fresh SRP
values were less than 0.050 mg/1. Low SRP values often occur when algal
densities are high. It is possible that during one week of storage in the
dark, release of soluble phosphorus could Occur- Samples with fresh SRP
values greater than 0.10 mg/1 increased by Only2?1$ during storage. A.t 8 of
the 12 transport stations the flux weighted.mean concentration exceeded 0.10
mg/1 (see Table 26).
The increase in SRP concentrations during storage was higher (+14.9$) for
samples with suspended solids greater than 50 mg/1 than for samples with
suspended solids less than 50 mg/1 (9.2$). Samples with conductivities less
than 500 umhos showed the smallest increase with storage (2.1%). Since low
conductivities correlate with high flows, under conditions of high transport
the SRP values did not show much increase. This has been confirmed by
calculations of the effects of storage on flow weighted mean concentrations of
SRP. For all the stations, storage for one week increased the SRP values by
1.2$, with individual stations ranging from +13.1$ to -7.9$.
For total phosphorus, nitrate, suspended solids and conductivity, the
mean changes during storage were -1.9$, +2.2$, -0.8$ and +0.4$ respectively.
These changes would have little effect on loading calculations since errors in
flow measurement are likely to exceed errors introduced by storage effects of
this size. Ammonia does show a larger effect of sample storage f-fi.q^) but
ammonia nitrogen transport is much less important than nitrate nitrogen
transport.
The paired T-test data described above were based on paired samples
collected through the 1978 water year. Paired T-tests run on samples
collected in 1979 and 1980 gave very similar results.
28
-------
It should be noted that the storage effects discussed above reflect
sample storage for a period of 1 week. The average time of sample storage for
environmental samples is about 4 days. Consequently we view the changes in
sample concentration during storage to be insignificant with respect to
loading calculations. The critical factor for loading calculations is to have
large numbers of samples during high flow periods. To meet the sample storage
and preservation requirements for soluble nutrients, as specified in the EPA
Procedures Manual (EPA 1979), would so increase the sampling costs that many
fewer samples could be collected and analyzed and consequently loading
estimates would become less precise.
Comparison of Pumped and Grab Samples
In Table 11, data on pumped and grab samples are presented. Paired
T-tests indicate that there is no significant differences between the pumped
and grab samples. The correlations between the pumped and grab samples are
also very good. The close agreement between the pumped and grab samples for
suspended solids and for total phosphorus reflects the fact that in these
rivers the suspended solids transport is dominated by clay and silt sized
particles. Particle size analyses by the USGS in these rivers indicate that
over 82% of the suspended solids by weight is clay (<0.004 mm) and \*>% is silt
(0.004 - 0.062 mm) (Antilla & Tobin, 1978). Consequently, the use of depth
integrated samples is not necessary for accurate suspended solids measurements
in these rivers. - -
FLOW DATA
All flow data are obtained from the USGS." Copies Of the "'Primary
Computation of Gage Heights and Discharge" are obtained for each station soon
after the gage house tapes are processed. These prin-touts include the stage
measurements at hourly intervals for each day. From these tables the gage
height at the time of sample collection is Obtained and entered into the data
set for that sample.
For each station expanded rating tables are obtained from the
These rating tables are stored in the computer and a computer program then
generates the flow values at times of sample collection using the gage height
and the rating table.
DATA STORAGE
To illustrate the information which is stored for each sample, a printout
of our archive files is shown in Table 12. The printout includes information
for samples collected at the Melmore gaging station between 1900 hours April
7, 1979 and 0100^ hours April 17, 1979- The year, month, day and time" of
sample collection are shown in the first two columns. All data are stored in
a time sequence for a given station. The day of the water year is shown in
the next column. It is used as a time base' for several of the plot programs.
The stage data, in feet, are shown in the next column. The flow in CFS is
generated by a program that refers to a rating table supplied by the nqas.
The flow table lists flows for each tenth of a foot increment in stage. Our
29
-------
Table 11. Paired T-test comparisons of pumped and grab samples.
Lo
o
Variable Number
of Pairs
Grab SRP
238
Pumped SRP
Grab TP
229
Pumped TP
Grab NO23
242
Pumped NO23
Grab NH3
239
Pumped NH3
Grab SS
239
Pumped SS
Mean
0.0815
0.0820
0.2074
0.2112
2.7586
2.7724
0.1190
0.1332
57.4695
63.3544
Standard
Deviation
0.096
0.096
0.162
0.170
3.261
3.342
0.190
0.219
127.197
146.268
T- TEST
Mean Standard Corr. 2-Tail T 2-Tail
Difference Deviation (r) Prob. Value Prob.
-0.0005 | 0.041 0.938 <0.001 -0.19 0.952
I
-0.0038 0.074 0.902 <0.001 -0.77 0.439
-0.0138 0.358 0.994 <0.001 -0.60 0.549
-0.0142 0.209 0.486 <0.001 -1.05 0.295
-5.8869 66.395 0.891 <0.001 -1.37 0.172
-------
Table 12. Sample'Archive Printout for Transport Stations.
1 'V.L \ 26-FE&-81
t'VhMlili TlhF' DAY OF YR
7-KI407 1900
79P40B 1900
79040? 1300
79&409 1900
790410 700
7V0410 1900
7^041 1 700
/V0411 1900
7V0412 /'OO
790412 1900
7VO413 700
790413 1900
770414 l&O
7V0414 700
7V0414 1000
790414 1300
7V0414 1600
790414 1900
790415 100
790415 400
790415 700
7V0415 1000
,'90415 1300
7*0415 1600
7V0415 1900
790415 2200
7V0416 100
790416 400
7V0416 700
790416 1000
790416 1360
790416 1600
790416 1VOO
790416 2200
790417 100
IBB. 792
1P9.792
190.542
190.792
191 .292
191 .792
192.292
192. 792
193.292
193.792
1V4.292
194. 792
195.042
195,292
195.417
195.542
195.667
195.792
196.042
196.167
196. 292
196.417
196.542
196.667
196.792
196.917
197.042
1 97. 167
197.292
197.417
197.542
197.667
197. 792
197. 91 7
178.042
09144:52
STAGE
3.96
3.63
5.94
6.11
6.17
3.94
5.28
4.6
4,21
4,03
3.89
4.72
8.62
9.02
10.28
10.37
10.26
10.01
9.47
9.26
9.06
8.84
8.6
8.36
8. 13
7.89
7.67
7.44
7.18
6.94
6.69
6.46
3.?6
4.06
5-. 09
FLOW
304.68
233.97
930.64
1002.8
1029,6
930.64
P79.26
462.3
362.23
320.41
289.67
496.04
2405.8
3302.4
3702.8
3786.1
3684.6
3458.9
3024.9
2865
2716.8
2559
2392
2232
2086,6
1940. 1
1810.9
1679.2
1532
1402.2
1274. 1
1162.2
1070
9&1 . 16
910.2
MLM79
MT OF'
24 .033
21 .031
12 .075
9 ,069
12 .06
12 .062
12 .059
12 .052
12 .042
.12 .04
12 .052
9 .069
6 .067
4. .07
3 .057
3 .056
3 .044
4. ,043
4. .041
3 ,036
3 .043
3 .023
3 .036
3 .03?
3 .018
3 .015
3 .03
3 .028
3 .023
3 .045
3 .034
3 .029
3 .019
3 .029
3 .03
SAMF'LE
TF' SS
.207 67, 1
.158 46
.304 204.
,32 137.
.306 131.
.246 87.3
.186 70.3
. 1 44 50,6
.142 41
.106 39.1
.09 39.5
.538 323.
1.31 982.
1.46 1061
1.56 676
1.54 1098
1.58 780,
1 .49 944.
1.56 961
1.52 1008
1.42 884.
1.38 760.
1.24 645.
1.14 496.
1.07 446.
1.01 403
.976 358
.912 329.
.904 300.
.828 273.
.792 257.
.784 247.
.768 247.
.73 235.
.71 330.
ARCHIVE F'RINT OUT
N023
6.57
7. 43
b..?5
5.94
6.65
7,45
7.68
7.38
6.9
6.64
6,58
3.13
2.7
2.72
2.59
2.5
2.39
2.58
2.52
2.57
2.68
2.78
2. 79
3.41
3.57
3.52
3.59
3.60
3.74
3. 71
3.86
3.93
4.03
4
4.03
NH3
.108
.197
.224
1 . 10
.206
. 121
.091
. 109
.081
.304
.115
.161
. 164
.109
. 149
.19
.066
.188
.101
.156
.32
.899
.048
.215
.114
.078
.072
'.09
,123
.219
. 1
.098
.058
.159
.067
COND TEMF' F'H SI02
451
491
346
334
333
345
380
416
442
463
484
400
200
188
1B6
17B
171
174
173
169
173
181
191
208
212
193
321
216
224
233
239
"246
252
258
263
-1 7. 79 7.11
-1 7.83 4.62
-1 7.71 6.06
-1 7.33 8.13
'-1 7.49 6,62
•1 7.58 8.4
-1 7.65 8.42
-1 7.75 9.01
-1 7.52 8.6
-1 7.91 9.73
-1 8 7.87
- 1 7,73 6.03
-1 7.54 S.59
-1 7.43 4.74
-1 7.34 4.38
--1 7.34 4.36
-1 7.31 5.78
-1 7.27 4.67
-1 7.3 4.69
-1 7,32 5.04
-1 7.36 5.03
-1 7.36 6.53
-1 7.42 6.46
-1 7.71 7.59
-1 7.55 9.82
-1 6.94 9. 19
-• 7.33 9.?7
- 7.26 7.5
- 7.38 7.2
- 7.44 7.27
- 7.47 7.9
- 7.51 /.98
7,5 9.86
- 7.53 11.1
7.54 9.05
CL TNN TOC CA
-1 1 .
-1
-1 1 .
-1 1 .
-1 1 .
-1 1 .
-1 1 .
-1 .
-1 1
-1
— 1
-1 1.
-1 4.
-1 5.
-1 6.
-1 5
-1 5.
-1 6.
-1 5.
-1 6.
-1 5.
-1 5.
-14.
-1 3
-1 3.
-1 3.
-1 3.
-1 2.
-1 3.
-1 3.
-1 8.
-1 2.
-1 3.
-1 3.
-1 2.
4U - 1
97 - -1
87 - -1
69 - -1
75 - -1
77 -1 -1
26 -1 -1
96 -1 -1
.3 -1 1
88 -1 -1
85
93
58
35
66
O _
98
09
61
15
57
07
75
.7
33
18
47
93
-1
-1
-1
-1
-1
-1
- 1
-1
-1
-1
-1
-1
-1
-1
_ 1
_ t
-1
-1
14 -1 -1
08 1 -1
23 -1 -1
86 -1 -1
29 -1 -1
15 -1 -1
66 -1 -1
MG
• 1
-1
1
-1
-1
-1
-1
-1
-1
-1
-1
-1
-1
-1
-1
-1
-1
-1
-1
-
-
-
-
-1
- 1
-1
-1
-1
-1
-1
-1
-1
-1
-1
K
-1
1
••1
-1
-1
1
-1
- 1
-1
- 1
-I
-1
-1
-•1
-1
-1
-1
-1
-1
-1
-1
-1
-1
- 1
-1
-1
1
-1
-1
-1
-1
-1
-1
-1
-1
NA
-I
-1
- 1
-1
-1
-1
-1
-1
-1
-1
-1
-1
-1
-1
-1
- 1
-1
-1
-1
-1
-1
-1
-1
-1
-1
- j
-1
- I
-1
-1
_ 4
-1
-1
-1
-1
FE
--1
- 1
-1
-1
-1
-1
-1
-1
-1
-1
-1
-1
-1
-1
-1
-1
-1
-1
-1
-1
-1
-1
-1
- 1
-1
-1
-I
- 1
-1
-1
-1
-1
-1
-1
-1
CLi
1
1
-1
-I
-1
-1
- 1
-1
- 1
1
- 1
-1
-1
-1
-1
- 1
-1
-1
-1
-1
-1
-1
-1
-1
-1
-1
-1
- 1
-1
-1
-I
-1
-1
-1
-1
MN
-1
- 1
• 1
-1
- 1
- 1
1
-1
- 1
-1
-1
-1
-1
-1
- 1
- 1
-1
-1
-1
- 1
— 1
-1
-1
-1
-1
-1
-1
-1
-1
-1
•-1
-1
-1
-1
-1
I Nil OF hl_M79
- FF.fr-81
SAMF'LE ARCHIVE f-'RINT OUT
-------
computer program does a linear interpolation to generate the flow for the
hundredth place in the stage data.
The column labeled MT contains a time multiplier in hours. The time
multiplier is obtained by a program which, for a given sample, calculates the
time interval between the preceeding sample and the following sample. This
time is divided by 2, to give the duration of time for which that sample is
used to characterize the stream transport. If the calculated multiplier
exceeds 24 hours the multiplier is set equal to 24 hours so that no single
sample is used to characterize the river for more than a 24 hour period. The
sample archive printout illustrates that during periods of high flow, more
samples are analyzed, and consequently the time multipliers are smaller. The
time multipliers are used in the programs for calculating material flux, total
flow, and time exceedency tables.
The concentrations of dissolved reactive phosphorus (OP), total
phosphorus (TP), suspended solids (SS), nitrate + nitrite nitrogen (N02?>)
ammonia nitrogen (NH3), conductivity in umhos at 25 degrees C (cond.), pH,
Si02, chloride, Total Kjeldahl Nitrogen (TKN) and metals are listed in the
remaining columns. All concentrations are in the units of mg/1 for the
element (ie. as P or N) with the exception of silica which is reported as
Si02. A minus one (-1) is used to indicate that the sample was not analyzed
for that parameter.
Following data entry into the laboratory's PDF 11 computer system, a
preliminary archive printout is obtained for use in data verification. The
values are compared with the laboratory bench sheets. Corrections are made as
necessary and the archives files are transferred to the college's main
computer system, a PRIME 550.
All Of the daba collected at the transport stations have been transferred
to the STORET system.
COMPUTATIONAL METHODS
A variety of Fortran programs have been written for conducting routine
calculations on the data sets. These include programs to calculate weighted
mean concentrations, fluxes, exceedency tables and annual loading, using flow
duration tables. Examples of the Outputs of these programs are presented
below, along with explanations of the way in which the calculations are done.
Flux Calculations
Table 13 is an example of the summary table from a flux calculation. The
program allows the selection of the river station, the beginning and ending
dates and times, and the parameter- The total load in metric tons and short
tons over the time interval are then listed. The total load is the sum of the
loads of the individual samples collected during the time interval. This is
illustrated in Table 14 which is an Optional printout for the flux summaries.
For each sample the load is calculated as:
-------
Table 13. Sample Printout of Program for Flux Calculations.
FLUX SUMMARY FOR THE MELMORE.RIVER ,
IETUFEN THE DATES 790*07 AND 790417
CDR THE: PARAMETER TP
TDTAL LOAD: METRIC TONS 25,88192
TOTAL VOLUME: CUBIC DETERS 2.6856644E
HEAN FLOW .' 31.67804
MONITORED TIME (HRS)'- 235.5000
1EAN CONCENTRATION (MG/L) 0.8446275
^LUX ytlGHTED MEAN CONC. : 0.9637064
INSTANTANEOUS FLOW UT. MEAN CONC. (H6/LK 1
TIME WEIGHTED AVERAGE CONC. (MG/L): 0.5357952
YJMBER OF SAMPLES 35
SHORT TONS 28.52187
"7 CFS-DAYS 10989.74
1119.365
,164046
Table 14. Sample printout of flux summary option which includes
data for individual samples and cumulative totals.
<=M
1
2
3
5
6
7
a
9
10
11
12
13
L*
: 5
1 6
i.7
ia
1 9
J3
3 1
i ->
_ c
;> 4
"•5
76
?7
'•-.a
50
u
i2
13
i5
. DATE
790*07
790*03
790*09
793*09
790*10
790*10
790*11
790*11
790*12
790*12
790*13
790*13
790*1*
790*1*
790*1*
790*1*
790* 1*
790*1*
790*15
790*15
790*15
790* 15
790*15
790*15
790*15
790*15
790*16
790*16
790*14
790*16
790*16
790*16
79fl*lS
790*16
790*17
TINE
1900
1900
1300
1900
0700
1900
0700
1900
0700
1900
0700
1900
0130
0700
1000
1300
1630
1900
0130
0*00
0700
1000
1300
1630
1900
2200
0100
0*00
0700
1000
1300
1600
1900
2200
0100
TP
IH5/L)
2.0700E-01
1.5800E-01
3.8*30E-01
3.2000E-01
3.0600E-01
2.*600E-01
1.86DOE-01
l.**OOE-01
l.*200E-01
1.0600E-01
9.0000E-02
5.3900E-01
1.31*OE 00
1«*663E 00
1.56*OE 00
1.5*33E 00
1.5820E 00
l.*900E 00
1.5623E 00
1.5200E 00
l.*2*OE 03
1.38*OE 03
1.2*80E 03
1.1*80E 00
1.0760E 00
1.0100E 00
9.7600E-31
9J1200E-01
9.0*OOE-01
8.2800E-01
7^9200E-01
7.8*OOE-01
7.6800E-01
7.3000E-01
7.1000E-01
FLOy
(CFS)
30*.9
23*.0
930.6
1003.
1030.
930.6
679.3
*62.3
362.2
320.*
289.7
*96.3
?*06.
i302.
3703.
3 786.
3685.
3*59.
3025.
2665.
2717.
2559.
2392.
2232.
2087.
19*0.
1811.
1679.
1532.
1*02.
127*.
1162.
1070.
981.2
910.2
NT -
(MRS)
2*.0
21.0
12.0
9.0
12.0
12.0
12.0
12.0
12.0
12.0
12.0
9.0
6.3
*.5
3.0
3.0
3.3
*.S
*.5
3.0
3.0
3. 0
3.0
3.0
3.0
3.0
3.0
3.0
3. 0
3. 0
3.0
3.0
3.0
3. 0
3.0
TOTAL LOA3
' t«T^
l.5*31E-01
2.33*OE-
01
6>.7030E-ai
9.6*56E-31
1.3*97E
1.6296E
1.78*1E
1.8655E
1.928*E
1.9699E
2.&018E
2.2*6*E
*.1788E
6.398*E
8. 168*E
9.9539E
1.1736E
1 .*098E
1 .b26*E
1.7595E
1.8778E
1 .9860E
2.0773E
2.1556E
?.22*2E
2.28*1E
2.3381E
2.38*9E
2.*273E
2.*627E
2.*936E
2.521*£
2.5*65E
2.56&*E
2.5682E
30
00
30
00
30
30
00
00
30
30
00
30
M
01
31
31
31
01
31
31
01
01
01
01
31
01
31
31
01
31
31
TOTAL FLOy
7.*547E
1.2*60C
2.3836E
3.3033E
4.5621E
5.6999E
6.5303E
7.0955E
7.538*E
7.9301E
8.28*2E
8.7390E
1.0210E
1.172*E
1.2855E
l.*013E
1.5139E
1.672*E
1.8111E
1.8987E
1.9817E
2.0599E
2.1330E
2.2013E
2.265CE
2.32*3E
2.3797E
2.*310E
2.*776E
2.5207E
2.5596E
2.5952E
2.6279E
2.6578E
2.6857E
05
06
06
06
06
06
36
06
06
06
06
36
07
07
07
0 7
07
07
07
07
07
07
37
37
0 7
3 7
07
0 7
0 7
07
07
0 7
0 7
07
07
TOTAL TIME
(HAS)
2.*OOOE
*.5000E
5.7300E
6.6300E
7.8000E
9.3000C
1.0200E
1.1*OOE
1.2600E
1.3800E
1.5000E
1.5900E
1.6500E
1.6950E
1.7250E
1.7550E
1.7850E
1.830QE
1.8753E
1.9050E
1.9350E
1.9650E
1.9950E
2.0250E
2.0550E
2.0850E
2.1150E
2. 1*50E
2.1750E
2.2050E
2.2350E
2.26*OE
2.2953E
2.3250E
2.3550E
01
01
01
01
01
01
02
02
02
02
32
02
02
02
02
32
02
32
32
32
02
02
02
02
02
02
02
32
02
02
02
02
02
02
02
33
-------
mg/1 X CFS X hours X 1.018810 X 10-4 - metric tons
Also the flow volume is calculated as:
CFS X hours X 101.88 - cubic meters
In the optional printout (Table 14) the cumulative total load, cumulative
flow and cumulative time are reported for each successive sample collected
during the time interval. The cumulative values for the final sample are
reported in the summary tables. This method of flux calculation is equivalent
to the midinterval method which the USGS uses to calculate sedmiment loads at
daily sediment stations (Porterfield, 1972). The various time multipliers are
the equivalent of subdivided days in the USGS calculations.
The flux summary tables also list the total volume of water monitored
during the time interval, both in cubic meters and CFS-days. The mean flow
during the time interval is calculated by dividing the total volume of water
by the monitored time. The mean flow is listed both in cubic meters per
second and cubic feet per second. The total monitored time in hours is also
listed in the summary. By listing the monitored time, the extent of gaps in
the monitoring record over the time interval can be determined.
The flux summary includes 4 calculations of concentration during the time
interval. These are calculated as follows:
m
Mean Concentration - i 0)
Flux Weighted Mean - £ cf^ Total Flux (2)
Concentration ' * concentration of ith sample
ti" * time multiplier of the ith savple
q^ • instantaneous flow for ith sample
N » number of samples
The interpretation of these various types of average concentrations is
described in the next section. The flux sumnary also lists the total number
of samples collected during the time interval.
Our flux calculation programs include a number of useful options. Flux
calculations can be obtained for any of the chemical parameters. In
specifying the dates, the flux for individual storms, months, water years or
the entire period of record can be selected. One Option selects all of the
samples for a specified month within a specified period of years. Table 15 is
a summary printout for all of the samples collected during the month of April
for the period of record.
34
-------
Another Option useful when calculating fluxes over a long period of time
includes, along with the summary table, a listing of the 50 largest flow
volumes (discharge X time multiplier) for individual samples. The dates of
the samples, concentrations and flows along with percent of the total flux
accounted for by each sample is listed. Similar information is also listed
for the 50 largest fluxes. These lists are useful for screening and verifying
those samples which are major contributors to the calculations of weighted
average concentrations. Examples of these outputs as applied to a flux
calculation of total phosphorus at the Melmore station during the 1979 water
year are shown in Tables 16, 17, 18. This type of screening has been done for
all of the flux weighted mean concentraions reported in the next section.
Exceedency Calculations
A second set of program deals with calculations of the percent of time
given concentrations, flows or fluxes are exceeded. In addition the program
calculates the percent of the total load accounted for by concentrations,
flows or fluxes up through each sample in the exceedency ranking. Tables 19,
20 and 21 provide examples of concentration, flow and flux exceedency tables
for the same sample data set from the Melmore station as shown in the sample
archive printout (Table 12).
The programs first allow selection of: the type of exceedency table (ie
concentration, flow" or flux); the stations; the initial and final dates and
times; and the parameter. In the case of a concentration exceedency table,
(see Table 19} all of the samples are sorted into a sequence Of increasing
concentrations. The printout includes the date, the time multiplier, the
instantaneous flow in CFS and the instantaneous flux in kg/hr for each sample.
The time multipliers are sequentially added giving a cumulative sum- of-time.
For each sample the current sum-of-time is-expressed, as a percentage of the
total time. The total time is the sum-of-time for the sample with the highest
concentration. Although the percentage is labeled "Percent Exceedency" it is
actually the percent of time that the concentration is less than the sample
concentration in the ranking. The actual "Percent Exceedency" is 100 minus
the listed Percent Exceedency.
The relationship between concentration exceedency and total loading is
shown in the final two columns of the printout. For each sample the total
load (ie the product of the time multiplier and the instantaneous flux) is
calculated and sequentially added, producing the column labeled sum-of-flux.
For each sample, the sum-of-flux is expressed as a percentage of the total
flux. Although these percentages are also labeled "Percent Exceedency", they
actually show the percent of the total flux accounted for by that sample plus
all of the samples with lower concentrations.
In the case of flow exceedency tables (Table 20"), samples are ranked in
order of increasing flow. The column labeled "Percent Exceedency" and
adjacent to the sum-of-time column shows the percentage of time that flows are
less than the sample flow. The listed "Percent Exceedency" should be
subtracted from 100 to obtain the actual percent of time the flows exceed the
listed flow for that sample. The relationships between flow exceedency and
total transport are shown in the final 2 columns of the flow exceedency
printout. The sum-of-flux and the "Percent Exceedency" are calculated in the
-------
Table 15. Sample printout of flux summary option for selected months.
MONTHLY FLUX SUMMARY FOR THE MELMORE.RIVER
IN THE APRILS OF 76 THROUGH 79
FOR THE PARAMETER TP
TOTAL LOAD METRIC TONS 42.38480 SHORT TONS 46.70305
TOTAL VOLUME CUBIC METERS 7.7128112E 07 CFS-DAYS 31560.82
MEAN FLOW (M**3/SEC) 8.252888 291.6215
10NITORED TIME (HRS) 2595.998
MEAN CONCENTRATION (MG/L) 0.3640593
FLUX WEIGHTED MEAN CONC. (M6/L) 0.5495377
INSTANTANEOUS FLOW UT. MEAN CONC. (MG/L) 0.7623307
TIME UEIGHTED AVERAGE CONC. (MG/L) 0.2073411
NJMBEK OF SAMPLES 223
Table .16. Sample printout of the flux summary for the 1979 water year
JLoading of total phosphorus at the Melmore gaging station.
FLUX SUMMARY FOR THE MELMORE.RIVER
BETWEEN THE DATES 781001 AND 790923
FOR THE PARAMETER TP
TOTAL LOAD METRIC TONS 82.99326 SHORT TONS 91.45856
TOTAL VOLUME CUBIC METERS 1.5828461E 08 CFS-DAYS 64773.05
MEAN FLOW (M**3/S£C) 6.263243 ' 221.3160
MONITORED TIME (HRS) 7019.997
MEAN CONCENTRATION (MG/L) 0.3826081
FLUX UEIGHTED MEAN CONC. CMG/D 0.5243292
INSTANTANEOUS FLOW UT. MEAN CONC. (MG/L) 0.6503026
TIME WEIGHTED AVERAGE CONC. (MG/L) 0.2638122
NUMBER OF SAMPLES 570
36
-------
Table 17. Sample printout showing the 50 largest flow volumes contributing to
the phosphorus loading at Melmore during the 1979 water year.
DATE
790306
790307
790414
790414
790414
790415
790224
790526
790406
790310
-790224
790305
790410
790308
790224
790406
790414
790410
790409
790915
790414
790414
790306
790305
781209
790224
790622
790622
790309
790223
790223
790223
790306
790621
790409
790623
790305
790304
790415
790305
790622
790304
790622
790304
790305
790225
790311
790411
790415
790829
790304
790306
790304
790305
790223
TIME
1900
1900
1900
700
100
100
700
1000
1900
1900
100
1900
700
1900
1300
100
1300
1900
1300
100
1000
1600
100
1300
1900
1900
1300
1900
1900
700
1300
100
700
1900
1900
100
100
2200
400
400
700
1900
100
1600
700
100
1900
700
700
1300
1300
1300
1000
1000
1900
MT
15.00
24.00
4.50
4.50
6.00
4.50
6.00
13.50
21 .00
24.00
6.00
6.00
12.00
24.00
6.00
12.00
3.00
12.00
12.00
24.00
3.00
3.00
6.00
4.50
24.00
6.00
6.00
6.00
24.00
6.00
6.00
6.00
6.00
6.00
9.00
6.00
3.00
3.00
3.00
3.00
6.00
3.00
6.00
3.00
3.00
6.00
21.00
12.00
3.00
24.00
3.00
6.00
3.00
3.00
6.00
FLOW
30.92
18.92
97.89
93.46
68.08
85.60
63.90
27 .77
17.84
15.52
61 . 17
59. 58
29. 14
13.87
55.41
27. 16
107. 15
26. 34
26. 34
13. 13
104.79
104. 27
51.57
68.28
12.40
47.85
47/36
46.70
11 . 51
45.28
45.28
45.28
43.83
43.51
28. 38
41 .96
82.36
•81 .29
81 .08
80.87
40.43
79. 76
39.53
77 .92
77.92
38.94
11 .08
19.22
76.89
9. 53
75.44
36.63
73.21
73.01
36. 48
TP
0. 21200
0. 18800
1.4900
1.4660
1.3140
1.5620
0.56700
0.46300
0.32700
0. 17000
0.58600
0.29900
0. 30600
0.17000
0. 56100
0.42400
1.5430
0. 24600
0. 38400
0.48300
1.5640
1.5820
0.27000
0.34200
0. 32600
0.51700
1.3260
1.1020
0. 16100
0.46000
0.74400
0.40800
0.24700
3.4920
0.32000
0.85600
0.39800
0. 38900
1.5200
0.38000
1.6360
0.37400
1.8040
0. 37700
0.35700
0.43500
0. 15600
0.18600
1.4240
0.35100
0. 39000
0.22800
0. 39700
0. 34400
0. 59000
FLOW VOL
463.81
454.18
440.49
420.56
408. 50
385.22
383.41
374.85
374.72
372.38
367.02
357.46
349.66
332.81
332.47
325.88
321.44
316.05
316.05
315. 12
314.37
312.82
309.43
307.26
297.55
287.10
284.14
280. 22
276. 28
271 .68
271 .68
271 .68
262.95
261 .07
255.43
251.75
247.08
243.88
243.24
242.60
242. 59
239.29
237. 19
233.76
233.76
233.63
232.76
230.68
230.66
228.73
226.33
219.77
219.64
219.04
218.91
% VOL
1.055
1.033
1.002
0.957
0.929
0.876
0.872
0.853
0.852
0.847
0.835
0.813
0.795
0.757
0.756
0.741
0.731
0.719
0.719
0.717
0.715
•0.711
0.704
0.699
0.677
0.653
0.646
0.637
0. 628
0.618
0.618
0.618
0.598
0. 594
0.581
0. 573
0.562
0.555
0. 553
0. 552
0.552
0. 544
0.539
0.532
0.532
0.531
0.529
0. 525
0.525
0.520
0.515
0. 500
0.500
0.498
0.498
FLUX
0.35398
0.30739
2.3628
2. 2196
1.9324
2.1662
0.78261
0.62480
0.44112
0.22790
0.77427
0.38477
0. 38518
0.20368
0.67145
0.49743
1.7855
0.27989
0.43690
0.54794
1.7700
1.7816
0.30076
0.37830
0.34921
0.53435
1.3564
1.1117
0. 16013
0.44990
0.72767
0. 39904
0. 23382
3. 2819
0. 29425
0.77578
0.35401
0.34152
1.3310
0. 33188
1.4288
0.32218
1.5404
0.31726
0.30043
0. 36586
0. 13072
0. 15446
1.1824
0.28903
0.31776
0. 18039
0.31390
0.27126
0.46496
% FLUX
0.427
0. 370
2.847
2.674
2. 328
2.610
0.943
0.753
0. 532
0.275
0.933
0.464
0.464
0.245
0.809
0.599
2.151
0. 337
0.526
0.660
2.133
2.147
0.362
0.456
0.421
0.644
1.634
1.339
0.193
0.542
0.877
0.481
0.282
3.954
0.355
0.935
0.427
0.412
1.604
0.400
1.722
0.388
1.856
0. 382
0.362
0.441
0.158
0.186
1.425
0.348
0.383
0.217
0. 378
0. 327
0.560
37
-------
Table 18. Sample printout showing the 50 largest phosphorus fluxes contributing
to the phosphorus loading at Melmore during the 1979 water year.
DATE
790621
790414
790414
790415
790414
790414
790414
790414
790622
790622
790622
790415
790415
790622
790415
790415
790709
790621
790415
790224
790623
790224
790417
790223
790415
790224
790526
790415
790915
790623
790416
790224
790406
790416
790223
790405
790223
790406
790511
790409
790416
790223
790623
790410
790305
790305
790225
790405
790416
790305
790306
781209
790606
790304
790405
TIME
1900
1900
700
100
100
1300
1600
1000
100
700
1300
400
700
1900
1000
1300
1900
1300
1600
700
100
100
1300
1300
1900
1300
1000
2200
100
700
100
1900
100
400
1900
100
700
1900
100
1300
700
100
1300
700
1900
1300
100
700
1000
100
1900
1900
100
2200
1300
MT
6.00
4.50
4. 50
4. 50
6.00
3.00
3.00
3.00
6.00
6.00
6.00
3.00
3.00
6.00
3.00
3.00
9.00
6.00
3.00
6.00
6.00
6.00
3.00
6.00
3.00
6.00
13. 50
3.00
24.00
6.00
3.00
6.00
12.00
3.00
6.00
6.00
6.00
21 .00
6.00
12.00
3.00
6.00
6.00
12.00
6.00
4. 50
6.00
6.00
3.00
3.00
15.00
24.00
6.00
3.00
6.00
FLOW
43.51
97.89
93.46
85.60
68.08
107. 15
104.27
104.79
39.53
40.43
47. 36
81 .08
76.89
46.70
72.42
67.69
13.24
17.36
63. 17
63.90
41 .96
61 . 17
19.42
45.28
59.05
55.41
27.77
54.90
13.13
36.20
51.25
47.85
27.16
47.52
36.48
29.01
45.28
17.84
14.80
26.34
43.36
45. 28
31 . 18
29.14
59.58
68.28
38.94
29.14
39.68
82.36
30.92
12.40
4.57
81.29
28.25
TP
3.4920
1.4900
1.4660
1.5620
1.3140
1.5430
1.5820
1.5640
1.8040
1.6360
1. 3260
1.5200
1.4240
1.1020
1.3840
1.2480
1.9580
2. 1440
1.1480
0. 56700
0.85600
0. 58600
3.6020
0.74400
1.0760
0.56100
0.46300
1.0100
0.48300
0.70000
0.97600
0. 51700
0. 42400
0.91200
0. 59000
0.72500
0. 46000
0. 32700
1 . 3670
0. 38400
0.90400
0. 40800
0. 59000
0. 30600
0. 29900
0. 34200
0.43500
0. 58000
0.82800
0. 39800
0.21200
0. 32600
3.4760
0. 38900
0. 55600
FLOW VOL
261 .07
440.49
420.56
385. 22
408.50
321 .44
312.82
314.37
237.19
242.59
284 . 14
243.24
230.66
280.22
217.26
203.08
119.16
104.17
189. 50
383.41
251.75
367.02
58. 27-
271 .68
177.15
332.47
374 .85
164 .71
315.12
217.17
153.75
287.10
325.88
142.56
218.91
174.07
271.68
374 .72
88.83
316. 05
130.07
271.68
187.09
349.66
357. 46
307.26
233.63
174.83
119.05
247.08
463.81
297.55
27.45
243.88
169.53
% VOL
0.594
1.002
0.957
0.876
0.929
0.731
0.711
0.715
0.539
0.552
0.646
0.553
0.525
0.637
0.494
0.462
0.271
0.237
0.431
0.872
0. 573
0.835
0. 133
0.618
0.403
0.756
0.853
0.375
0.717
0.494
0.350
0.653
0.741
0.324
0.498
0.396
0.618
0.852
0.202
0.719
0.296
0.618
0.426
0.795
0.813
0.699
0. 531
0.398
0. 271
0.562
1.055
0.677
0.062
0.555
0.386
FLUX
3.2819
2.3628
2.2196
2. 1662
1.9324
1.7855
1.7816
1.7700
1.5404
1.4288
1.3564
1 . 3310
1.1824
1.1117
1.0825
0.91240
0.83993
0.80402
0.78315
0.78261
0.77578
0.77427
0.75556
0.72767
0.68622
0.67145
0.62480
0.59890
0.54794
0.54728
0. 54020
0.53435
0.49743
0.46807
0.46496
0.45433
0.44990
0.44112
0.43715
0.43690
0.42329
0. 39904
0.39737
0. 38518
0. 38477
0. 37830
0. 36586
0. 36504
0.35485
0.35401
0.35398
0.34921
0. 34350
0.34152
0.33933
% FLUX
3.954
2.847
2.674
2.610
2.328
2.151
2.147
2.133
1.856
1.722
1.634
1.604
1.425
1.339
1.304
1 .099
1.012
0.969
0.944
0.943
0.935
0.933
0.910
0.877
0.827
0.809
0.753
0.722
0.660
0.659
0.651
0.644
0.599
0.564
0.560
0.547
0.542
0.532
0.527
0.526
0.510
0.481
0.479
0.464
0.464
0.456
0.441
0.440
0.428
0.427
0.427
0. 421
0.414
0.412
0.409
38
-------
Table 19.
\
Sample printout of a concentration exceedency table for a sample data set from
the Melmore station.
DATETIME
7904130700
7904121900
7904120700
7904111900
7904081900
7904110700
7904071900
7904101900
7904100700
7904091900
7904091300
7904131900
7904170100
7904162200
7904161900
7904161600
7904161300
7904161000
7904160700
7904160400
7904160100
7904152200
7904151900
7904151600
7904151300
7904140100
7904151000
7904150700
7904140700
7904141900
7904150400
7904141300
7904150100
7904141000
7904141600
MT
12.
12.
12.
12.
21.
12.
24.
12.
12.
9.
12.
9.
3.
3.
3.
3.
3.
3.
3.
3.
3.
3.
3.
3.
3.
6.
3.
3.
5.
5.
3.
3.
5.
3.
3.
PARAMETER
TP
0.900000E-01
0. 106000
0. 142000
0. 144000
0. 158000
0. 186000
0.207000
0. 246000
0.306000
0.320000
0.384000
0.538000
0.710000
0.730000
0.768000
0.784000
0.792000
0.828000
0.904000
0.912000
0.976000
1.01000
1.07600
1.14800
1.24800
1.31400
1.38400
1.42400
1.46600
1.49000
1. 52000
1.54300
1.56200
1.56400
1.58200
FLOW
289.67
320.41
362.23
462. 30
233.97
679.26
304.88
930.64
1029.62
1002.86
930.64
496.00
910. 20
981.16
1070.00
1162.20
1274 . 10
1402. 20
1532.00
1679.20
1810.90
1940. 10
2086.60
2232.00
2392.00
2405.80
2559.00
2716.80
3302.40
3458.90
2865.00
3786. 10
3024.90
3702.80
3684 .60
FLUX
2.65762
3.46225
5. 24343
6.78630
3.76846
12.8794
6.43348
23. 3380
32. 1177
32.7142
36. 4300
27.2026
65.8782
73.0145
83.7706
92.8846
102.867
118.355
141.180
156. 115
180. 173
199.752
228.875
261 .206
304.314
322.256
361 .038
394.379
493. 526
525.377
443.930
595.531
481.658
590.355
594.215
SUM OF
TIME
12.0000
24.0000
36. 0000
48.0000
69. 0000
81.0000
105.000
117.000
129.000
138.000
150.000
159.000
162.000
165.000
168 . 000
171 .000
174.000
177.000
180.000
183.000
186.000
189.000
192.000
195.000
198.000
204.000
207. 000
210.000
214.500
219.000
222.000
225.000
229. 500
232.500
235.500
PERCENT
EXCEEDENCY
5.096
10. 191
15. 287
20. 382
29.299
34 . 395
44.586
49.682
54 .777
58.599
63.694
67.516
68.790
70.064
71.338
72.611
73.885
75. 159
76.433
77.707
78.981
80.255
81.529
32.803
84. 076
86.624
87.898
89. 172
91.083
92.994
94.268
95.541
97.452
98.726
100.000
SUM OF
FLUX
31.89141
73.43842
136. 3601
217.7957
296.9333
451 .4860
605.8894
885.9452
1271.358
1565.786
2002.946
2247.769
2445. 403
2664. 447
2915.758
3194.412
3503.012
3858. 077
4281 .617
4749.961
5290. 480
5889.737
6576.361
7359.978
8272.920
10206. 46
11289.57
12472.71
14693.57
17057.77
18389.55
20176.15
22343.61
24114 .67
25R97.31
PERCENT
EXCEEDENCY
0. 123
0.284
0.527
0.841
1. 147
1.743
2.340
3.421
4.909
6.046
7.734
8.680
9.443
10.289
11.259
12.335
13.527
14.898
16.533
18.342
20.429
22.743
25. 394
28.420
31.945
39.411
43. 594
48.162
56.738
65.867
71.010
77.908
86.278
93.117
100.000
-------
Table 20.
Sample printout of a flow exceedency table for a sample data set from the
Melmore station.
DATETIME
7904081900
7904130700
7904071900
7904121900
7904120700
7904111900
7904131900
7904110700
7904170100
7904101900
7904091300
7904162200
7904091900
7904100700
7904161900
7904161600
7904161300
7904161000
7904160700
7904160400
7904160100
7904152200
7904151900
7904151600
7904151300
7904140100
7904151000
7904150700
7904150400
7904150100
7904140700
7904141900
7904141600
7904141000
7904141300
MT
21.
12.
24.
12.
12.
12.
9.
12.
3.
12.
12.
3.
9.
12.
3.
3.
3.
3.
3.
3.
3.
3.
3.
3.
3.
6.
3.
3.
3.
5.
5.
5.
3.
3.
3.
PARAMETER
TP
0. 158000
0.900000E-01
0.207000
0. 106000
0. 142000
0. 144000
0.538000
0. 186000
0.710000
0.246000
0. 384000
0.730000
0. 320000
0. 306000
0.768000
0.784000
0.792000
0.828000
0.904000
0.912000
0.976000
1.01000
1.07600
1.14800
1.24800
1. 31400
1.38400
1.42400
1.52000
1.56200
1.46600
1.49000
1.58200
1.56400
1.54300
FLOW
233.97
289.67
304.88
320. 41
362.23
462.30
496.00
679.26
910.20
930.64
930.64
981. 16
1002.86
1029.62
1070.00
1162.20
1274.10
1402.20
1532.00
1679.20
1810.90
1940. 10
2086.60
2232.00
2392.00
2405.80
2559. 00
2716.80
2865.00
3024.90
3302.40
3458.90
3684.60
3702.80
3786.10
FLUX
3.76846
2.65762
6. 43348
3.46225
5.24348
6.78630
27.2026
12.8794
65.8782
23.3380
36.4300
73.0145
32.7142
32.1177
83.7706
92.8846
102.867
118.355
141 .180
156.115
180. 173
199.752
228.875
261.206
304.314
322. 256
361 .038
394.379
443.930
481.658
493.526
525.377
594.215
590. 355
595.531
SUM OF
TIME
21.0000
33.0000
57.0000
69.0000
31.0000
93.0000
102.000
114.000
117.000
129.000
141.000
144.000
153.000
165.000
168.000
171 .000
174.000
177.000
180.000
183.000
186. 000
189.000
192.000
195.000
198.000
204.000
207.000
210.000
213.000
217. 500
222.000
226.500
229.500
232. 500
235.500
PERCENT
EXCEEDENCY
8.917
14 .013
24.204
29. 299
34.395
39.490
43.312
48.408
49.682
54.777
59.873
61. 146
64.968
70. 064
71.338
72.611
73.885
75.159
76.433
77.707
78.981
80.255
81.529
82.803
84.076
86.624
87.898
89.172
90.446
92.357
94.268
96.178
97.452
98.726
100.000
SUM OF
FLUX
79.13763
111.0290
265.4325
306.9795
369.9012
451.3367
696. 1602
850.7129
1048. 347
1328.403
1765.563
1984.607
2279.035
2664.447
2915.759
3194.412
3503.012
3858.077
4281 .617
4749.961
5290.480
5889.737
6576.361
7359.978
.8272.920
10206.46
11289.57
12472.71
13804.50
15971.96
18192.82
20557.01
22339.65
24110.71
25897.31
PERCENT
EXCEEDENCY
0.306
0.429
1.025
1.185
1.428
1.743
2.688
3.285
4.048
5.130
6.818
7.663
8.800
10.289
11.259
12.335
13.527
14.898
16.533
18.342
20.429
22.743
25.394
28.420
3) .945
39.411
43.594
48.162
53.305
61.674
70.250
79.379
86.262
93.101
100.000
-------
Table 21. Sample printout of flux exceedency table for a sample data set from the
Melmore station.
Di,TETIMF
7904130700
7904121900
7904081900
7904120700
79040^1900
790411 19QO
7904110700
7904ini°00
7904131900
7904100700
7904091900
7904091300
7904170100
7904162200
7904161900
7904161600
7904161300
7904 161000
790416070C?
7904160400
7904160100
7904152200
7904151900
79041 51600
7904151300
7904140100
79P4151000
7904150700
7904150400
79041 50100
7904140700
7Q041 41 900
7904 1 4 1000
7Q041 41 600
79^4141300
MT
12.
12.
21.
12.
24.
12.
12.
12.
9.
12.
9.
12.
3.
3.
3.
3.
3.
3.
3.
3.
3.
3.
3.
3.
3.
6.
3.
3.
3.
5.
5.
5.
3.
3.
3.
PARAMFTFR
TP
0. 900000R-01
0. 106000
0. 158000
0. 142000
0. 207000
0. 141000
0. 186000
0. 246000
0. 538000
0. 306000
0. 320000
0. 384000
0. 710000
0. "HOOOO
0. 768000
0. 784000
0.792000
0.828000
0. 904000
Q. 912000
0. 976COO
1. 01000
1.07600
1 . 14800
1. 24800
1.3MOO
1. 38400
1. 424QQ
1. 52000
1.56200
1. 46600
1.49000
1. 56400
1. 58200
1.54300
FLOW
289.67
320. 41
233.91
362. 23
304. 88
462. 30
679. ?6
930.64
496.00
1029.62
1002.86
930.64
910. 20
981. 16
1070. 00
1162. 20
1274. 10
1402. 20
1532.00
1679. 20
1S1 0. 90
1 940. 10
2086.. 50
2232. 00
2392. 00
7405. 80
2559. 00
2716.80
2865.00
3024.90
3302. 40
3-158.90
3702. BO
3684.60
378G. 10
FLUX
2.65762
3. 4622S
3. 76846
5. 24348
6. 43348
6. 78630
12. 8794
23. 3380
27. 2026
32. 1177
32.7142
36. 4300
65. 8782
73.0145
83.7706
92.8846
102. 867
118. 355
141. 180
156. 115
180. 173
199.752
2?B. 875
261. 206
304.314
322. 256
361.038
394. 379
44 3. 930
4R1. 658
493. 526
525. 377
590. 355
594. 215
595. 531
SUM OF
TIME
12. 0000
24. 0000
45. 0000
57. 0000
Rl.OOOO
93. 0000
105.000
1 17.000
126. 000
138. 000
1 47. 000
159.000
162.000
165.000
158.000
171. 000
174. OCO
1 77. 000
180. POO
1 83. CQO
196. 000
1 89.000
1 9?. 000
1 95.000
1 98. 000
204. 000
207. COO
210.000
?1 3.000
217. 500
222.000
226.500
229.500
232. 500
235.500
PERCENT
EXCFEDENCY
5. 096
10. 191
19. 108
24. 204
34. 395
39. 490
44. 586
49.682
53. 503
58. 599
62.420
67.516
68. "790
70.064
71. 338
72.611
73.835
75. 159
76. 433
77. 707
78. 981
80. 255
81. 529
82. 803
84. 076
86.6?4
87. 898
89. 172
90. 446
92. 357
94. 258
96. 178
97.452
98.726
100. 000
SUM OF
FLUX
31. R9141
73.43842
152.5760
215.4977
369.9012
451.3367
605.8894
885. 9452
1130.769
1516. 181
1810. 603
2247. 770
2445. 404
2664. 447
2915.759
3194.412
3503.012
3858.07?
4281.617
4749. 961
5290. 480
58P9.737
6576. 361
7359.978
8272. 920
10206.45
11289.57
12472.71
13804.50
15971.96
18192. 82
20557. 01
2£328.07
24110. 71
25897. 31
PERCENT
EXCEEDENCY
0. 123
0. 284
0.589
0.832
1.428
1.743
2. 340
3.421
4.366
5.855
6.991
8. 680
9. 443
10. 289
11. 259
12.335
13. 527
14. 898
16. 533
18. 342
20. 429
22. 743
25. 394
28. 420
31.945
39.411
43. 594
48. 162
53. 305
61.671
70. 250
79. 379
86.218
93. 101
100.000
-------
same fashion as in the concentration exceedency tables.
For the flux exceedency tables the samples are ranked by instantaneous
flux (Table 21 ). The percent of time fluxes are less than the indicated value
is shown in the "Percent Exceedency" column. The relationship between ranked
instantaneous fluxes and the total flux or load are shown in the last two
columns.
One modification of the flow exceedency Output is to substitute a dummy
variable equal to one (1 ) into the concentration column. This converts the
sum of the flux into the sum of the flow and the relationship between flow
exceedency and total flow can be examined.
Normally the exceedency calculations described above are done on an
entire data set. Some potential uses Of this type of exceedency data in water
quality management are described in subsequent sections of this report.
Flow Duration Table Calculations
The USGS provides daily flow duration tables for its continuous stream
gaging stations. Table 22 is an example of such a table for the Fremont
stream gaging station. At each station the mean daily flows are divided into
35 classes (0-34). The flow classes represent approximately equal logarithmic
increments in flow. The duration tables consist of two parts. In the first
part, the number of days in each water year that the mean daily flows fell
into each flow class is listed. In the second part, for each flow class there
is listed the flow value, the total number of days that the flow was in that
class, the accumulated number of days with flows greater than that flow class,
and the percent of days with flows greater than -that class. Flow duration
tables from the USGS are stored in our computer for each station.
'
One set of programs uses the flow duration tables to calculate mean
annual loading. Table 23 is an example of this method as it is applied to the
calculation of mean annual loading of total phosphorus at the "Fremont Station
(Tindall Bridge). The program sorts the entire Fremont total phosphorus data
set into the flow intervals between the listed flow class values from the flow
duration table. Figure 7 shows alternate flow class values superimposed on a
graph shoeing the relationship of phosphorus concentration to stream flow,
expressed on a log scale. The flux weighted mean concentration of total
phosphorus for all of the samples with flows in each interval is calculated.
The number of samples in that flow interval is listed as N in the table. The
concentration for the interval is multiplied by the arithmetic mean flow for
the interval. The resulting instantaneous flux is multiplied by the
proportion of a year when the flows fall in that flow interval. This is
determined by the change in percent flow exceedency between the two flow
classes forming the flow interval. This percent of a year is shown as delta
percent in the table. Appropriate conversion factors are introduced so that
the products are in kg/yr for each flow interval. These are totaled to
provide an estimate of the mean annual loading.
One variation of the program results in the calculation of the
concentration rather than the flux weighted mean concentration for each flow
interval. In this case, the standard deviation and the standard error of the
42
-------
Table 22. U. S. Geological Survey Flow Duration Table for the Sanctusky River near Fremont.
STATION NUMBER 04198000
DISCHARGE, IN CUBIC FEET PER SECOND
MEAN
SANDUSKY RIVER NEAR FREMONT, OHIO
CLASS
YEAR
1966
1967
1968
1969
1970
1971
1972
1973
1974
1975
1976
1977
1978
0 1
8
11
DURATION TABLE OF DAILY VALUES FOR YEAR ENDING SEPTEMBER 30
9 10
11 12 13 14 15 16 ' 17 18 19 20 21 22
NUMBER OF DAYS IN CLASS
23 24 25 26 27 28 29 30 31 32 33 34
4 13 19 27 35 35 40 28 29 26 20 19 6 15 8
6 44 25 16 8 17 22 18 18 30 25 14 15 17 18
6 17 19 17 16 19 33 42 49 34 19 12 9 14 8
1 18 21 8 20 6 19 22 34 33 29 24 24 .14 24 17
7 10 21 22 20 31 21 18 21 21 23 23 20 19 17
27 28 22 37 20 21 27 26 22 25 12 17 10 17 12
3 35 25 23 13 7 19 37 23 18 30 29 18 12 14 10 10
1 20 11 6 7 5 13 15 24 27 22 25 33 28 23 36 16
1 3 13 17 27 29 16 24 20 13 26 33 17 24 16 17 9
2 3 10 21 16 17 11 10 13 19 32 30 39 25 18 20 16 15 15
1 1 14 14 10 25 42 36 30 31 25 30 32 12 16 7
55 68 42 31 16 16 11 16 9 10 15 8 18 12
7 13 24 16 4 14 22 22 20 30 26 27 24 21 9 20 9
7
11
13
4
12
10
10
23
8
15
7
7
6
6
14
5
10
8
5
5
36
13
15
7
10
5
5
11
9
5
9
7
10
16
8
7
6
5
8
6
9
8
7
12
9
10
20
12
7
7
7
4
1
9
8
9
7
3
7
13
5
7
5
5
5
4
7
2
4
7
1
4
11
5
7
4
6
2
5
3
7
3
2
7
2
2
2
6
2
5
2
3
2
2
2
2
4
1
1
1
3
1
2
2
3
.1
3
3
1,
1
6
*Years 1924-1965 not shown in the above table.
-------
Table 22 continued.
CLASS VALUE TOTAL ACCUM PERCT
CLASS VALUE TOTAL ACCUM PERCT
0
I
2
3
4
5
6
7
8
9
10
11
VALUE
V95
V90
V75
V70
V50
V25
V10
0.00
5.00
6.50
8.50
11.00
15.00
19.00
25.00
32.00
42.00
55.00
72.00
0
6
14
19
169
146
526
612
1090
1211
1241
1093
EXCEEDED 'P'
26
36
68
86
230
790
= 2600
.00
.00
.00
.00
.00
.00
.00
19054
19054
19048
19034
19015
18846
18700
18174
17562
16472
15261
14020
PERCENT
100.0
100.0
100.0
99.9
99.8
98.2
98.1
95.4
92.2
86.4
80.1
73.6
OF TIME
12
13
14
15
16
17
18
19
20
21
22
23
94.0
120.0
160.0
210.0
270.0
360.0
470.0
610.0
800.0
1000.0
1400.0
1800.0
877
1101
1073
962
1176
1132
1009
891
633
922
570
489
12927
12050
10949
9876
8914
7738
6607
5598
4707
4074
3152
2582
67.8
63.2
57.5
51.8
46.8
40.6
34.7
29.4
24.7
21.4
16.5
13.6
24
25
26
27
28
29
30
31
32
33
34
2300
3000
3900
5200
6700
8800
11000
15000
20000
26000
33000
484
404
380
283
238
135
98
59
9
1
2
2093
1609
1205
825
542
304
169
71
12
3
2
10.9
8.4
6.3
4.3
2.8
1.5
.8
.3
-------
Table 23. Sample printout of calculation of mean annual loading using flow duration intervals and
flux weighted mean concentrations for each interval.
YIELD BY FLOW
STATION NAME:
Flow Class
34-33
33-32
32-31
31-30
30-29
29-28
28-27
27-26
26-25
25-24
24-23
23-22
22-21
21-20
20-19
19-18
18-17
17-16
16-15
15-14
14-13
13-12
12-11
11-10
10-9
9-8
8-7
7-6
6-5
5-4
CLASS
TINDALL RIVER R3
Interval
0.0- 0.0
0.0- 0.0
0.0- 0.3
0.3- 0.8
0.8- 1.5
1.5- 2.8
2.8- 4.3
4.3- 6.2
6.2- 8.4
8.4- 10.9
10.9- 13.5
13.5- 16.5
16.5- 21.3
21.3- 24.6
24.6- 29.3
29.3- 34.6
34.6- 40.5
40.5- 46.6
46.6- 51.7
51.7- 57.3
57.3- 63.1
63.1- 67.7
67.7- 73.5
73.5- 80.1
80.1- 86.5
86.5- 92.2
92.2- 95.4
95.4- 98.2
98.2- 98.9
90 - 9— 99 - a
Delta %
0.0
0.0 '
0.3
0.5
0.7
1.3
1.5
1.9
2.2
2.5
2.6
3.0
4.8
3.3
4.7
5.3
5.9
6.1
5.1
5.6
5.8
4.6
5.8
6.6
6.4
5.7
3.2
2.8
0.7
0.9
N
0.
8.
55.
42.
44.
76.
109.
94.
89.
75.
66.
90.
124.
76.
38.
196.
115.
137.
130.
132.
138.
120.
71.
73.
78.
48.
12.
3.
2.
3 -
'' Mid Flow
Ft**3/Sec
29500.00
32000.00
17500.00
13000.00
9900.00
7750.00
5950.00
4550.00
3450.00
2650.00
2050.00
1600.00
1200.00
900.00
705.00
540.00
415.00
315.00
240.00
185.00
140.00
107.00
83.00
63.50
48.50
37.00
28.50
22.00
17.00
13 - OO
PARAMETER :
Mean Cone.
KG/L
0.0000000
0.5618126
0.5922673
0.7196640
0.4936485
0.5589405
0.4530399
0.3730553
0.4116686
0.3168694
0.2882024
0.3150172
0.2418206
0.2358314
0.2061968
0.2658683
0.2009223
0.1580711
0.1863328
0.1745594
0.1922386
0.1404692
0.1509163
0.1190366
0.1383464
0.1461058
0.1445343
0.1496249
O.O738666
O - O77S666
Yield
KG/YR
0.0000000
0.0000000
27767.12
41773.12
30549.56
50288.00
36107.58
28799.98
27902.51
18740.49
13717.62
13502.97
12438.54
6254.776
6101.307
6795.008
4393.212
2712.355
2036.688
1614.941
1393.961
617.4139
648.7782
445.5038
383.4811
278.9345
117.1176
82. 30742
7.828341
0-3. 14676
-------
Sandusky River At TindaI I Bridge
A I I Data 1969 Through 1979
3.00 3.50
4.80 4.50
LOG C Flow
5.00 5 50
) MT3/S
Figure 7. Flow duration flow class intervals for the Fremont gaging
station superimposed on a graph of total phosphorus
concentration versus log of stream flow.
mean are also calculated for each interval. The standard error of the mean
concentration for each flow interval is also multiplied by the mean flow and
percent of the year to obtain an error estimate for each interval. These
error estimates are totaled for all of the flow intervals to obtain an error
estimate for the mean annual load as shown in Table 24.
A second variation sorts the data into summer and winter months. mhe
average concentrations and standard errors are calculated for each flow
interval. These seasonal concentration data are used to compare winter and
summer concentrations and to calculate fluxes in cases" where chemical data are
missing. Summer and winter concentration data are stored for each parameter
for each flow interval for each station. After entering the station, month
and mean daily flow, the program generates an estimated daily flux for each
parameter. The program totals the fluxes for each parameter for each mean
daily flow entered in a given month. Fluxes for missing sample days are then
added to measured fluxes for each month in order to estimate total monthly and
annual loads.
46
-------
Table 24. Sample printout of calculation of mean annual loading of total
phosphorus using flow duration tables and average concentrations,
standard deviations and standard errors for each interval.
Yield by flow class
Station name: Tindall.river,R3
Parameter: TP
Flow
class
34-33
33-32
32-31
31-30
30-29
29-28
28-27
27-26
26-25
25-24
24-23
23-22
22-21
21-20
20-19
19-18
18-17
17-16
16-15
15-14
14-13
13-12
12-11
11-10
10-9
9-8
8-7
7-6
6-5
5-4
4-3
3-2
2-1
% Interval
0.0- 0.0
0.0- 0.0
0.0- 0.3
0.3- 0.8
0.8- 1.5
1.5- 2.8
2.8- 4.3
4.3- 6.2
6.2- 8.4
8.4- 10.5
10.5- 13.5
13.5- 16.5
16.5- 21.3
21.3- 24.6
24.6- 29.3
29.3- 34.6
34.6- 40.5
40.5- 46.6
46.6- 51.7
51.7- 57.3
57.3- 63.1
63.1- 67.7
67.7- 73.5
73.5- 80.1
80.1- 86.5
86.5- 92.2
92.2- 95.4
95.4- 98.2
98.2- 98.9
98.9- 99.8
99.8- 99.9
99.9-100.0
100.0-100.0
Delta %
0.0
0.0
0.3
0.5
0.7
1.3
1.5
1.9
2.2
2.5
2.6
3.0
4.8
3.3
4.7
5.3
5.9
6.1
5.1
5.6
5.8
4.6
5-8
6.6
6.4
5.7
3.2
2.8
0.7
0.9
0.1
0.1
0.0
N
0.
8.
55.
42.
44.
76.
109.
94.
89.
75.
66.
90.
124.
76.
38.
196.
115.
137.
130.
132.
138.
120.
71.
73.
78.
48.
12.
3.
2.
3.
0.
0.
0.
2244.
Mid Clow
ft /sec
29500.00
23000.00
17500.00
13000.00
9900.00
7750.00
5950.00
4550.00
3450.00
2650.00
2050.00
1600.00
1200.00
900.00
705.00
540.00
415.00
315.00
240.00
185.00
140.00
107.00
83.00
63. 50
48.50
37.00
28.50
22.00
17.00
13.00
9.75
7.50
5.75
Total 1 Mean cone.
flow
0.00
0.00
52.50
65.00
69.30
100.75
89.25
86.45
75.90
66.25
53.30
48.00
57.60
29.70
33.14
28.62
24.49
19.21
12.24
10.36
8.12
4.92 .
4.81
4.19
3.10
2.11
0.91
0.62
0.12
0.12
0.01
0.01
0.00
951.1
mg/1
.0000
.5702
.6051
.6002
.5833
.5191
.4654
.3819
.4227
.3452
.3181
.3249
.2531
.2831
.2413
.3137
.2625
.1928
.2473
.1798
.1907
.1378
.1496
.1237
.1410
.1433
.1397
.1450
Yield
kg/yr
.0000
.0000
.2837E
.3484E
.3610E
.4671E
.3709E
.2949E
.2865E
.2042E
.1514E
.1393E
.1302E
7510.
7140.
8019.
5740.
3309.
2703.
1664.
1383. -JS
605.7
643.0
463.1
390.8
270.0
113.8
79.77
05
05
05
05
05
05
05
05
05
05
05
Sdev
mean
.0000
.2230
.2565
.4029
.2612
.2231
.1888
.1514
.3748
.1531
.1468
.2563
.1304
.2435
.1355
.3223
.224
.1551
.2670
.7270E-01
Serr
mean
0000
7886E-01
3458E-01
6218E-01
3938E-01
2559E-01
1808E-01
1561E-01
3973E-01
1768E-01
1807E-01
2701E-01
1171E-01
2793E-01
, 2198E-01
. 2302E-01
, 2093E-01
.1325E-01
.2341E-01
.6328E-02
.8070E-01.6869E-02
.7400E-017.865
.7767E-018.116
.0000
.0000
.0000
.0000
.0000
. 0000
.3438E
06
.5257E-01
.7412E-01
.5091E-01
.5318E-OL
.6654E-01
.7856E-01
.6974E-01
.1414E-02
.1154E-02
.0000
.0000
.0000
.4799E-02
.8796E-02
. 5959E-02
. 6021E-02
.9604E-02
.2268E-01
.4026E-01
.1000E-02
.6663E-03
.0000
.0000
. 0000
Serr
yield
.0000
.0000
1622.
3609.
2437.
2303.
1441.
1205.
2693.
1046.
860.2
1158.
602.6
740.8
650.4
588.5
457.7
227.3
255.9
58.55
49.82
21.10
37.82
22.30
16.69
18.09
18.47
22.15
.1063
.6962E-01
.0000
.0000
.0000
.2216E 05
-------
SECTION 6
CONCENTRATIONS OP NUTRIENTS AND SEDIMENTS
AT THE TRANSPORT STATIONS
Studies of both ambient water quality and material transport in streams
require measurements of the concentrations of various substances. For
nutrients and sediments in the streams of northwestern Ohio, a major feature
of the concentration data is the large amount of variability present at a
given station. This variability is reflected in the graphs showing the
relationship between concentrations of various materials and stream flow
(Figure 8). Examination of the concentration data reveals that, within this
variability, there exist several patterns of concentration in relation to
flow, location and time. In this section of the report these patterns of
concentration are presented along with a discussion of some of the factors
which contribute to the variability.
DESCRIPTION OF DATA SETS
The extent of the data available at each of the transport stations is
summarized in Table 25- The summary includes the number of nutrient and
sediment samples analyzed during each water year for which automatic samplers
were in place at- the sampling stations. The total discharge for each water
year as monitored by the USGS is listed. The percent of the total discharge
and time monitored at each station for each water year was calculated from the
total flow and total time data listed in the flux summaries for suspended
solids. Only data collected through the 1979 water year have been included in
the data analyses presented in this report. The- sampling and analyses program
did continue at eight of the stations during .the 1980 and 1981 water years.
In addition to the parameters listed in Table 2^, conductivity, pH and ammonia
data are also available for approximately the same numbers Of samples.
Chloride, sulfate and total Kjeldahl nitrogen data are available for
approximately 25% of the samples.
TYPES OF WEIGHTED AVERAGE SEDIMENT AND NUTRIENT CONCENTRATIONS
As noted in the description of the flux summary programs, four types of
average concentrations are calculated for each data set. The equations for
calculating these average concentrations are repeated below.
1 . Average Concentration (unweighted)
E c.
. _
L. —
n
2. Time Weighted Average Concentration
E c.t.
E t.
i
48
-------
Table 25. Numbers of Samples Analyzed and Percent of Flow and Time
Ohio Sampling Stations.
Monitored by Water Year for Northwest
ern
Maumee
Portage
Bucyrus
Nevada
Upper
Sandusky
Tymochtee
Mexico
Melmore
Water
Year
1975
1976
1977
1978
1975
1976
1977
1978
1975
1976
1977
1978
1979
1976
1977
1978
1979
1975
1976
1977
1978
1979
1975
1976
1977
1978
1979
1975
1976
1977
1978
1979
1976
1977
1978
1979
Discharge
m /yr .
4.76309
5.03509
3.05309
6.16609
2.89908
3.26208
2.41308
4.43808
1.02708
8.21607
5.72207
9.55607
1.20308
4.83?7^
5.21907
1.22508
9.91707
2.75308
2.20508
1.41108
3.0-6808
2.75708
1.77108
7.64107
6.60207
2.07308
1.44208
6.66608
4.78308
3.60708
7.51308
6.37508
6.91607
7.29207
1.47808
1.50208
Percent
Monitored
% Flow % Time
73%
86
70
96
69
81
78
98
35
78
97
99
58
86
79
65
90
92
91
113
72
81
89
67
99
73
76
6
121
112
109
105
59%
90
69
78
57
82
71
82
31
66
84
89
54
60
75
73
64
72
79
88
69
76
58
64
82
80
61
8
50
76
63
89
96
80
Suspended
Solids
386
601
406
418 '
465
565
381
445
119
529
575
500
336
458
512
429
456
621
647
558
384
486
555
428
484
408
323
105
413
315
436
477
656
554
569
-
Numbers of Samples Analyzed
Soluble
Total Reactive Nitrate
Phosphorus Phosphorus Nitrogen
396
615
407
418
427
565
380
444
118
519
575
501
330
464
513
430
455
572
635
535
385
488
535
454
482
413
323
132
412
332
437
477
668
555
570
413
585
395
405
487
566
361
436
121
537
513
455
330
440
479
423
455
631
622
552
380
488
576
412
466
390
323
129
384
315
436
470
658
539
570
397
604
382
414
502
563
360
445
122
524
560
478
336
473
487
431
454
631
630
550
386
483
580
433
486
419
324
132
332
436
477
667
555
569
====--.
Conductivity
368
617
398
418
488
575
370
445
118
516
578
503
335
490
517
431
456
621
640
559
378
489
567
457
491
420
323
128
331
437
483
660
555
570
49
-------
Table 25 continued
Wolf West
Wolf East
Fremont
Huron
Water
Year
1976
1977
1978
1979
1976
1977
1978
1979
1975
1976
1977
1978
1979
1975
1976
1977
1978
1979
Percent
Monitored
Discharge % Flow % Time
3
2
6
4
4
1
1
6
1
7
6
1
1
3
2
2
4
2
.016°7
.855°7
•743°7
.597°7
.405°7
.iso08
121°8
.279°7
.030°9
.716°8
.290°8
.391°9
.088°9
.058°8
.671°8
.535°8
.048°8
.955°8
80%
86
91
105
99
99
46
67
67
54
93
83
— . 83
80
83
86
87
27
55%
80
72
64
57
74
73
68
51
68
84
83
62
60
84
62
75
31
Suspended
Solids
382
467
437
413
372
486
411
385
421
433
452
456
420
568
560
346
377
176
Numbers of Samples Analyzed
Soluble
Total Reactive Nitrate
Phosphorus Phosphorus Nitrogen
388
446
435
415
358
487
412
388
397
426
456
455
424
552
518
339
379
175
378
462
428
411
345
481
404
388
426
416
444
446
422
586
521
317
374
171
366
466
438
414
371
466
407
388
403
429
455
454
423
546
538
328
379
176
Conductivity
388
468
436
415
363
489
412
389
406
423
456
456
423
546
560
345
377
168
50
-------
3. Flow Weighted Average Concentration
C =
L c.q.
11
Zq.
4. Flux Weighted Average Concentration
where Ci = concentration of ith sample; qi = instantaneous flow for ith
sample; and ti = time multiplier for ith sample; and n = number of samples.
The necessity of distinguishing among four different methods of
calculating average concentrations is a consequence of the nature Of the
sampling program used in this study. In order to accurately measure stream
transport, the frequency of collections during high stream flows was greater
than the frequency during low flows. Consequently, the individual samples
have varying time durations associated with them. If the samples had been
collected on a uniform time schedule (as for example daily or weekly) then
equations 1 and 2 would have been identical as would equations 3 and 4.
In Table 26 the four methods are illustrated using the data from the
Melmore gaging T station. The average and time weighted average concentrations
provide information of relevance for ambient water quality. If, for example,
water were to be withdrawn at a constant rate from the river, the time
weighted average concentrations in the stream- would reflect the average
concentrations of the pumped water. In contrast, the flow weighted and flux
weighted average concentrations provide informa'tion "T>f relevance to stream
transport. If all of the water discharged by the stream were collected for a
long period of time, the flux weighted concentration in the stream would be
the average concentration of the collected water, at least for conservative
parameters.
Comparison of the flux weighted and time weighted average concentrations
also indicates whether the concentrations of a given parameter tend to
increase or decrease with increasing stream flows. Where the flux weighted
average concentration exceeds the time weighted concentration, as for
suspended solids, total phosphorus and nitrates, the concentration of the
parameter tends to increase with increasing stream flow. Conductivity and
chlorides (not shown) tend to decrease in concentration as stream flow
increases. For these the flux weighted concentration is less than the time
weighted concentration. The time and flux weighted concentrations of soluble
reactive phosphorus are almost equal, indicating that at the Melmore gaging
station the soluble reactive phosphorus is, on the average, independent of
stream flow.
In Table 27 the time and flux weighted average concentrations of
sediments and nutrients are listed for each of the twelve sampling stations.
In each case the concentration is calculated from the entire data set
available for the station through bhe 1979 water year. This includes both the
51
-------
Table 26. Comparison of time weighted average, flow weighted
average and flux weighted average concentrations of
sediment and nutrients at the Honey Creek, Melmore
sampling station for the period between 76-01-28
and 79-09-30.*
Parameter
Suspended Solids
Total Phosphorus
Souble Reactive P.
Nitrate/Nitrite N
Conductivity (umhos)
Ammonia-N
N
2256
2270
2237
2268
2268
2216
Average
mg/1
99.6
.271
.0861
4.30
518
.208
Time Wt.
Average
rag/1
56.3
.193
.0781
3.68
585
.187
Flow Wt.
Average
mg/1
203.7
.466
.0851
4.56
320
.293
Flux Wt.
Average
mg/1
164.17
.386
.0789
4.57
379
.241
*The average concentrations reported, in this table were calculated after the
addition of missing flow data based on U. S. Geological Survey estimates of
mean daily flows. Consequently, these average values are slightly different
than those presented in other tables. See section 7 for further description
of corrections for missing data.
full year records described in Table 25 and partial year records from earlier
water years. The number of samples included in the calculations for suspended
solids is listed for each station. Similar numbers o-f samples were used for
the nutrient calculations. Those stations located downstream from significant
municipal sewage outfalls are evident as locations where the time weighted
mean concentrations of soluble reactive phosphorus are larger than the flux
weighted mean concentrations. This Occurs at the Portage, Bucyrus, Upper
Sandusky and Huron stations.
PATTERNS OF SEDIMENT AND NUTRIENT CONCENTRATIONS IN RELATION TO STREAM FLOW
Typical relationships between sediment and nutrient concentrations and
stream flow for northwestern Ohio rivers are shown in the scattergrams of
Figure 8. The patterns for strictly agricultural watersheds are illustrated
by the East Branch of Wolf Creek (upper graphs) while the patterns for
agricultural watersheds containing significant point sources immediately
upstream from the sampling station are illustrated by the West Branch of Wolf
Creek (lower graphs)- The Wolf-West sampling station is immediately
downstream from the town of Bettsville. In this figure a log (base 10) scale
is used for stream flow so that the points will be spread out across the
graphs rather than concentrated on the left (low flow) portion of the graph.
Because of the log scale for flow, care must be exercised in judging the
presence or absence of trends in the data. The graphs include all of the data
52
-------
Table 27. Comparison of Flux Weighted and Time Weighted Mean Concentrations of Suspended Solids and Nutrients
in Northwestern Ohio River Basins.
Gaging Station
Maumee , Waterville
Portage, Woodville
Huron, Milan
Sandusky,
Sandusky,
Sandusky,
Sandusky ,
Tymochtee
Honey Cr.
Fremont
Mexico
Upper S.
Bucyrus
, Crawford
, Melmore
Broken Sword, Nevada
Wolf Cr. ,
Wolf Cr. ,
East Br.
West Br.
Flux
Time
Flux
Time
Flux
Time
Flux
Time
Flux
Time
Flux
Time
Flux
Time
Flux
Time
Flux
Time
Flux
Time
Flux
Time
Flux
Time
Wt.
Wt.
Wt.
Wt.
Wt.
Wt.
Wt.
Wt.
Wt.
Wt.
Wt.
Wt.
Wt.
Wt.
Wt.
Wt.
Wt.
Wt.
Wt.
Wt.
Wt.
Wt.
Wt.
Wt.
Mean
Mean
Mean
Mean
Mean
Mean
Mean
Mean
Mean
Mean
Mean
Mean
Mean
Mean
Mean
Mean
Mean
Mean
Mean
Mean
Mean
Mean
Mean
Mean
N
1769
1842
2027
2237
1578
2729
2299
2631
2115
1766
1654
1699
Suspended
Sediment
mg/1
242
106
164
62
220
69.6
217
83.1
239
85.8
235
105
173
49.6
205
68.7
180
57.8
244
78.3
181
42.4
183
40.8
Total
Phosphorus
mg/1
.516
.340
.402
.360
.362
.343
.453
.244
.428
.250
.518
.482
.573
1.13
.419
.181
.403
.195
.401
.157
.416
.161
.394
.232
Soluble
Reactive P
mg/1
.116
.116
.119
.191
.104
.201
.093
.073
.070
.069
'.134
.234
.219
.837
.069
.040
.101
.102
.061
.042
.118
.063
.100
.133
N03-N02
Nitrogen
mg/1
4
3
5
3
3
2
4
3
3
3
3
2
3
3
5
3
4
3
4
3
4
2
6
3
.92
.91
.89
.80
.61
.50
.61
.09
.50
.49
.90
.60
.42
.11
.12
.48
.85
.74
.87
.11
.71
.92
.14
.45
Conduc-
tivity
umhos
517
652
554
854
540
685
487
716
624
750
478 !
709
460
702
397
751
397
587
428
637
578
764
464
747 ;
I
for the Wolf-East and West stations summarized in T'able 25- Note that in
several instances the scaling factors (Scientific notation with E format) vary
for both concentrations and flows.
The pattern for suspended solids
The suspended solid concentrations
flows increase. The concentrations
particularly at the higher flow
relationship between rainfall induced
transport by the storm runoff water
resuspension as stream flows increase
contribute to the observed pattern.
(Figure 8) is similar for both streams.
are low at low flows and increase as the
are very variable at a given flow,
values. This pattern reflects the
erosion on the landscape and sediment
in the streams. The effects of sediment
and of stream bank erosion may also
For the East Branch of Wolf Creek the pattern of total phosphorus
concentrations (Figure 8b) in relation to stream flow is very similar to the
sediment pattern. Much of the total phosphorus is particulate phosphorus that
is adsorbed onto sediment particles. For the West Branch, the total
phosphorus concentration is lowest in the midrange of flows and increases as
the flows both increase and decrease. The high total phosphorus
concentrations at low flows are due to point source inputs of phosphorus. As
the stream flows decrease, there is less stream water present to dilute the
incoming point sources. The high concentrations of total phosphorus at high
flows are related to the high sediment concentrations and are similar to those
found in the East Branch.
53
-------
Uolf Creek Eosl Bronch
'*
-e 00 0 se i 00 i be 2 00 2 se j GO 3 se
-e.se -e.ee e.se i.00 i.se 2.ee 2.se 3. ee 3.se
Figure 8. Relationship of concentrations (mg/1) and log flow in CFS at the Wolf, East and West Stations.
-------
Wolf Cr.ak Eo«l Brano1-
All Do to 1975 Through Pr«..nl
£ CO
a. .
• to
<***t
28 5 34. S
Uolf Cr*«k U»«t Braocfi
All Data 1975 Throuoh Pr.««nt
se i 0e i .be 2 ee 2 se 3.88 3.50
Wolf Cr.«k Coil Branch
All Dala 1975 Throuoh Pr.
D
«nt
-1.5 45 105 16.5 22.5 28 5 34.5
I (iR t Flow ) C-l
Wolf U«.t U««t Brooch
All Data 1975 Through Pr.««nt
-B se -e
.SB 1.00 1 .SB 2 00 2.50 3.00 3.50
-8.50 -0.
-------
UI
CTv
Wolf Cr,.k Ea«l Branch
All Data 1975 Through Pr«s.nt
-I S
<.S IP 5 16.S
LOG C t- lo- 3 F-1
22.5 28.5 34.5
Uolf Cr..k U..I Branch
All Dato 1975 Throueh Pr«««nl
i N
ui -
t t
- 0 56 -8 09 0 50 I 00 1.50 2 00 ? 50 3 00 3 50
LOG (. Flow )
Uolf Cro.k Eoel Branch
All Dolo 1975 Through Pretanl
0)
S>
-------
Uolf Cr.ok Eo.l Branch
All Data 1975 Through Pr.s.nt
I **W*
* *
IT*T*T I I fc -tf* * A* * * i. *
U 1 Tftt lt4*WP t \L F" rf ^
* inOTXt%«t^%f*n*k ******
p- •^M^^v^piite^
I-.. '.I'v^.^M^S^
-I S
Wolf Cr.«k U.«t Branch
.All Doko I97S
•45 185 185 22.5 28 S
LOG C Flow 3 E-l
-------
The concentrations of soluble reactive phosphorus in the East Branch do
increase slightly as stream flows increase (Figure 8c). This increase is not
nearly as large as the increase in total phosphorus concentrations with
increasing stream flow. Not all of the agricultural watersheds show
increasing soluble reactive phosphorus concentrations with increasing stream
flow, but in all of them the soluble reactive phosphorus concentrations are
significant during periods of high stream flow. For the West Branch of Wolf
Creek the dominant feature of the soluble reactive phosphorus concentration
versus flow graph is the large increase at low stream flows. During the low
flows below point source inputs a large proportion of the total phosphorus is
in the form of soluble reactive phosphorus. Point source sewage effluents are
characterized by a high proportion of soluble reactive phosphorus present
within the total phosphorus.
The patterns of nitrate + nitrite nitrogen concentrations are similar for
the two streams (Figure 8d) . In both cases the nitrates tend to increase as
stream flows increase but the concentrations are extremely variable at all
flows. The point source inputs at Bettsville do not cause increases in
nitrate concentrations under low flow conditions on the West Branch of Wolf
Creek.
The patterns for ammonia nitrogen ("Figure 8e) are similar to those for
total phosphorus except that they are less pronounced. For the East Branch
the highest ammonia concentrations did occur under conditions of high flow.
These would reflect Occasional high ammonia concentrations in runoff water
from agricultural lands. However, many low values for ammonia were also
observed at high flows. High ammonia concentrations are also characteristic
Of sewage effluents and consequently, in the West Branch of Wolf Creek, high
ammonia concentrations were present at both high- affd low flows.
The concentration-flow pattern for To'ta'l K-jeldahl Nitrogen
paralleled the sediment patterns in the East Branch (Figure 8f). For the West
Branch, some high TKN values were also observed under low flows. (Note that
different TKN concentration scales were used for the two stations).
Significant nitrogen export does occur in the TKN fraction associated with
suspended solids.
For both of the stations conductivity decreased as stream flows increased
(Figure 8g) . Again considerable variability was present in the conductivity
at a given flow. A number of factors probably account for the conductivity
patterns. The conductivities of surface runoff water would be less than tile
effluent which, in turn, would be less than ground water. Various mixtures of
water from the above three sources would produce varying conductivities.
Under low flow conditions evaporation in the summer and ice formation in the
winter can increase the conductivity of the stream water.
The patterns of concentration versus flow illustrated for the two
branches Of Wolf Creek are typical of the patterns observed at the other
transport stations in northwestern Ohio. The major variations in the patterns
are associated with the extent of "pdint source impact under low flow
conditions. As the volume of the point source discharge increases in relation
to stream flow and as the concentrations of the point source effluent increase
in relationship to background stream concentrations, the effects of the point
58
-------
source effluents on the concentration versus flow patterns become larger.
However, the further downstream the sampling station is located relative to
the point source Outfall, the less will be the_impact On the low flow stream
concentrations. This is especially so for nonconservative parameters, such as
phosphorus and nitrogen, which are processed in various ways by the stream
system.
Although various types of concentration-flow patterns are evident in the
scattergrams described above, the large amounts of variability in sediment and
nutrient concentrations at a given station and flow rate stand Out as
important features of the data sets. It is evident that basing conclusions
concerning stream chemistry on the results of sampling programs involving
small numbers of samples could give misleading results. In the following
sections, sources of this variability in concentration-flow relationships will
be discussed.
HYDROGRAPHS, SEDIMENTGRAPHS AND CHEMOGRAPHS
Part of the variability in chemical concentrations at a given station and
flow is due to the nature of the runoff events which move through stream
systems. As a storm event moves past a given sampling station, the stream
flow and concentrations of sediment and various chemicals change in
characteristic ways. A given flow less than the peak flow will occur twice,
once during rising flows and once during falling flows. The chemical
concentrations during these two periods are often quite different. Plots
illustrating the changes in stream flow, sediment concentration and
concentrations of chemicals as a function of time are referred to as
hydrograplis, sedimentgraphs and chemographs. Where the chemicals are
considered pollutants the term "pollutograph" is-'also used. Analysis of these
graphs can often yield information concerning the source and transport of
materials. ' ~~
Figure 9 shows a typical runoff event which occurred between May 2 and
May 11, 1977 on the East Branch Of Wolf Creek. In these figures, the sediment
graphs and various chemographs are shown in relationship to the hydrographs.
The hydrograph can be divided into the ascending limb, the peak discharge and
the descending limb. In a runoff from a single rainstorm the rise in flow on
the ascending limb is usually steeper than the decrease in flow on the
descending limb and consequently the ascending limb transports less total
volume of water and lasts a shorter time than the descending limb.
Accordingly, the midpoint in the mass of water moving past a sampling station
during a storm event occurs after the peak discharge has passed.
The concentration of suspended sediment increases very rapidly during the
ascending limb of the hydrograph (Figure 9a). The peak sediment concentration
occurs prior to the peak discharge and the sediment is said to have an
"advanced" peak. After the peak sediment concentration is reached the
sediment concentration decreases more slowly. The chemographs for total
phosphorus (Figure 9b) and Total Kjeldahl Nitrogen (Figure 9c) closely
parallel the sediment graph. Much of the total phosphorus is associated with
the transport Of particulates.
59
-------
213.0 214.8 215.0 216.0 217.0 218.0 219.0
Day of the Waler Year
213.0 214.0 215.0 216.0 217.0 218.0 219.0
Day of the Voter Year
CTi
O
213.0 214.0 215.0 216.0 217.0 218.0 219.0
Day of the Water Year
D
213.0 214.0 215.0 216.0 217.0 218.0 219.0
Day of the Water Year
Figure 9. Hydrographs (+), sediment graphs (Q), and anemographs ($) for a runoff event at the Wolf
East Station between May 2 and May 11, 1977. Concentrations in mg/1 or ymhos and flow
in CFS.
-------
0
213.8 214.0 215.0 216.0 217.0 218.0 219.0
Day of Ihe Water Year
(0
213.0 214.0 215.0 216.0 217.0 218.0 219.0
Day of the Water Year
0
0
00
OS
0 i
-(0
H
213.0 214.0 215.0 216.0 217.0 218.0 219.0
Day of the Water Year
213.0 214.0 215.0 216.0 217.0 218.0 219.0
Day of the Wcfer Year
-------
In contrast to total phosphorus, the chemographs for chloride (Figure 9e)
and for conductivity (Figure 9f) show decreasing concentrations as the flow
increases. Often the shape of these chemographs is the inverse of the
hydrographs. The chemograph for nitrate plus nitrite nitrogen (Figure 9g^
shows an increase in concentration in association with the hydrograph. The
peak nitrate concentration trails the peak sediment and total phosphorus
concentrations and occurs during the descending limb of the hydrograph.
Soluble reactive phosphorus (Figure 9d) also increases in concentration
as the flow increases. It should be noted that the actual concentrations of
soluble reactive phosphorus are much lower than the total phosphorus
concentrations. After its peak, the concentration of the soluble reactive
phosphorus decreases more rapidly than the nitrate concentration. The timing
of the elevated soluble reactive phosphorus concentrations is more like the
timing of the elevated concentrations of suspended sediments and total
phosphorus than like nitrate concentrations.
The changes in concentrations which occurred during the runoff event
illustrated in Figure 9 represents the net effect of several different
processes. Prior to the rainstorm the stream channel upstream from the
sampling station contained water with the chemical composition characteristic
of low or base flow for the watershed. This water is characterized by high
concentrations of chloride, high conductivity and low suspended sediment and
nutrient concentrations. During the runoff event, this base flow water in the
stream channel would have been pushed ahead by the storm runoff water.
Consequently, during the early portion of the hyrdograph, the discharge of
base-flow water stored in the channel will increase above prestorm discharges
of that water.
As the rainfall input to the watershed exceeds" the infiltration capacity
of the soils, water accumulating on the surface begins to drain into the
stream network. This surface runoff water con'tains-'- high concentrations of
sediment and sediment-bound chemicals such as phosphorus. Since this water
has limited internal contact with the soil prior to reaching the drainage
network it does not contain high concentrations of dissolved substances, ^he
decreasing chloride and conductivity at the sampling station reflects the
effects of this surface runoff water, as well as rainwater falling directly on
the surface waters. Surface runoff originating in close proximity to the
g'aging station will arrive early in the hydrograph. The timing of the arrival
of the surface runoff from more distant parts of the watershed depends on the
routing of water through the drainage network.
Some of the rainwater which infiltrated into the soil will move laterally
to the stream systems. This pathway, which is often referred to as interflow,
can be very important in agricultural lands with tile drainage systems.
Chemically, this water contains much higher concentrations of certain soluble
nutrients, such as nitrates, than does the surface runoff. Also, its content
of total dissolved ions, as reflected in its conductivity, will be higher than
the surface runoff. Generally this water will have a very low concentration
of suspended sediments. The timing of the arrival of the interflow water at
gaging stations will vary with the distance and pathway of water within the
drainage network. This movement of water will be delayed relative to the
surface water movement to the gaging station. As the hydrograph progresses at
the gaging station the proportion of water derived from surface runoff
62
-------
O
00
LU
Q
,
en'
o
a
o
q
d"
in
03
o
q
tog
—
u_o
in
in
o
o
PORTOE 1977
1.82 1.84 1.85 1.87 1.88
DAT OF WflTER TERR E 2
a
q
d
a
m
UJ
Q
-------
o
q
CD"
o
o
o
o
..
Luo
o
^o
i°.
a
in
CM
o
o
o
o
LO
X
Oo
510 1979
a.
_j
a:
t—r\j
a
o
o
o
o
03
O
O.
O
o
o
CD
«§§
a
q
CD
o
o
CM
o
o
Flow
& Total Phosphorus
X Suspended Solids
3.10 3.14 3.18 3.22 3.26 3.30 3.34
DRY OF THE WflTER TEflRE 2
o
q.
cd
o
q
ca"
cr.
-o
q
OJ
o
o
o
q
o"
o
03
o
.
S
510 1979
Jo
q
o"
o
CM
a
a
Figure 11,
Flow
O Conductivity
X Nitrate
3,10 3.14 3.18 3.22 3.26
DRY QF THE WRTER YERR
3.30
E 2
3.34
Compound hydrographs, sedimentgraphs and chemographs for August,
1979 runoff events at the Fremont gaging station.
64
-------
relative to interflow will change, having a high proportion of surface runoff
during the early portion of the hydrograph and a low proportion of surface
runoff during the later stages of the runoff. This is illustrated in Figure 9
where the peak concentration of nitrates trails the peak sediment
concentration and minimum conductivity.
The pattern of changes in sediment and chemical concentration at the
sampling station during a hydrograph is also affected by sediment resuspension
from the stream bottom, as well as "by deposition Of sediment both Onto the
stream bottom and, during larger runoff events, onto flood plains. The
resuspension of sediment is related to the movement of the flood front through
the drainage network. The flood front moves downstream faster than the
velocity of the water. As the flow increases, the sediment carrying capacity
of the water increases and sediments are picked up from the stream bottom.
These resuspended sediments, along with sediments derived from surface runoff,
determine the pattern of changes in suspended sediment concentration which
occur at the sampling station.
Although the general sequence of peak and minimum concentrations shown in
Figure 9 is common, the positions of the peaks and minimums can be shifted
relative to the peak of the hydrograph. Figure 10 illustrates storms in which
the peak sediment concentration is simultaneous to (10 a) Or trailing (10 b, c
& d) the hydrograph peak. The position of the peak concentration relative to
the hydrograph is determined by the interaction of the various processes
mentioned above. "Trailing sediment peaks can most easily be explained by
localized storms on a small portion of the watershed. The storm runoff water
containing high^ suspended sediments enters the drainage network and initiates
the propagation of a storm wave through the stream system. The sediment-laden
storm water moves more slowly through the stream system and arrives at the
sampling station after the peak of the storm wave"has passed. In the case of
advanced sediment peaks, it is probable that both resuspension of sediment and
the routing of nearby surface runoff water contribute- to the high sediment
concentrations present early in the hydrograph. For large storms much Of the
water passing the stream gage during the late portions of the hydrograph may
have been in storage over flood plains during which a portion of its sediment
load was deposited.
Very often rainstorms closely follow one another. The associated
hydrographs become superimposed giving rise to complex sediment graphs and
chemographs. Figure 11 illustrates some storms with complex hydrographs and
their associated complex sediment graphs and chemographs.
VARIATIONS AMONG RUNOFF EVENTS AT A SINGLE GAGING STATION
Part of the scatter present in the concentration versus flow graphs of
Figure 8 is a result of the large variations in concentrations that accompany
runoff events with approximately equal peak flows. This is true for storms
with high, medium and low peak flows.
This variability is apparent within a set of 52 storm events for which
good sampling records were available at the Upper Sandusky .gaging station. In
Table 28, there is listed the beginning and ending dates, the number of
samples collected, the storm duration, the peak flow, the time weighted
65
-------
Table 28. Summary of Storm Data at Upper Sandusky Gaging Station
Storm
no.
1
2
3
A
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
Starting date
Ending date
741207
741212
741212
741230
741223
741230
741231
750106
750107
750116
750125
750128
750128
750204
750216
750222
750222
750303
750304
750311
750329
750404
750830
750904
751018
751025-
760208
760214
760215
760221
760221
760229
760303
760311
760319
760327
760618
760623
760707
760715
760723
760728
Mean flow
Peak flow
(m3/sec)
20.93
50.01
27.03
63.53
24.12
66.34
15.58
39.45
37.29
95.31
16.09
38.37
35.90
99.79
29.84
73.66
56.65
182.45
15.30
38.80
14.30
41.88
14.12
38.15
4.28
7.90
19.81
29.38
83.41
165.36
21.27
62.91
27.29
101.31
12.89
47.43
5.49
15.55
9.17
42.08
11.55
34.41
Samples
Duration
N, hrs
20
120.0
35
246.0
23
162.0
25
146.0
30
219.0
12
92.5
26
157.8
23
139.5
36
221.0
20
138.0
18
170.5
19
122.4
19
174.7
10
150.0
20
136.0
17
213.0
24
222.0
21
216.0
18
135.0
26
189.0
31
115.5
Susp. solids
Wt. mean
Position
rns/1
92.7
70.5%
79.3
56.5%
113.40
67.7%
71.20
72.9%
223.3
67.2%
131.3
50.0%
342.8
76.4%
110.9
61.9%
538.5
74.8%
136.9
59.7%
247.3
85.5%
444.4
72.4%
32.8
66.5%
131.3
47.2%
716.7
70.9%
235.5
80.3%
420.1
68.7%
328.5
73.3%
368.7
61.9%
726.8
69 . 9%
1123.8
79.47,
Total P
Wt. mean
Position
me/1
0.248
55.2%
0.210
49.6%
0.297
65.1%
0.160
57.2%
0.445
59.4%
0.671
70.0%
0.495
62.8%
0.408
62.7%
0.422
76.97.
0.994
67 ._5%
0.370
66.0%
0.358
34.4%
0.957
66.4%
0.370
68.2%
0.720
66.3%
0.594
68.1%
0.707
59.8%
1.070
67.8%
0.962
62.7%
Nitrate
Wt. mean
Position
ms/1
6.29
46.0%
6.92
51.7%
6.50
52.3%
5.03
50.1%
4.89
49.3%
3.57
51.7%
3.45
44.4%
5.54
46.1%
3.42
52.4%
4.31
47.2%
4.23
30.5%
2.80
44.3%
2.50
36.2%
2.45
32.0%
4.12
47.3%
3.81
50.5%
3.25
50.4%
2.82
48.8%
12.41
32.5%
4.74
40.4%
2.49
44.1%
Cond.
Wt. mean
Position
umhos
477.
54 .-2%
443.
51 . 0%
421.
50.3%
490.
49.2%
362.
46.0%
483.
58 . 5%
336.
A3. 6%
417.
51.4%
288.
44.8%
496.
54 . 2%
479.
43.8;:
365.
46.7%
711.
51.4%
400.
39-8%
311.
48.2%
431.
43.0%
386.
46.7%
493.
49.6%
637.
51.2%
440.
50.3%
394.
50.7%
66
-------
Table 28. Continued
Storm
no.
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
Starting date
Ending date
760807
760812
770317
770302
770317
770322
770322
770401
770402
770410
770503
770509
770630
770704
770704
770708
770709
770713
770722
770728
770805
770810
771130
771205
771208
771224
780313
780318
780518
780521
780524
780528
780612
780615
781202
781206
781208
781217
781231
790116
790222
790301
Mean flow
Peak flow
(m3/sec)
3.96
10.06
26.88
90.16
27.87
63.53
16.0
43.75
31.12
100.04
15.71
40.55
8.98
29.55
6.18
14.60
2.89
6.97
5.68
21.01
2.56
5.31
18.05
44.71
56.55
237.72
96.41
197.17
15.03
30.88
15.27
41.88
2.50
5.78
5.70
18.17
4.62
26.65
28.69
71.43
42.72
115.12
Samples
Duration
N, hrs
21
126.0
26
228.0
14
117.0
20
273.0
18
213.0
19
155.0
16
114.0
16
102.0
16
108.0
25
162.0
20
141.0
13
117.0
37
360.0
21
141.0
14
84.0
15
117.0
10
93.0
15
120.0
14
234.0
28
403.0
21
114.0
Susp. solids
• Wt. mean
Position
mg/1
199.8
60.5%
220.5
58.1%
227.1
58.4%
215.5
69.8%
196.3
59.0%
1447.8
63.7%
563.0
37.1%
326.7
50.3%
621.7
40.9%
436.6
59.3%
176.4
67.6%
181.1
58.8%
36.9
51.9%
363.9
74.0%
370.5
63.4%
554.7
82.3%
267.6
67.5%
184.7
82.1%
85.0
90.5%
193.2
66.0%
Total P
Wt . mean
Position
rag/1
0.621
62.0%
1.083
76.3%
0.506
52.1%
0.258
52.1%
0.508
63.5%
0.423
56.9%
1.480
56.3%
0.875
40.6%
0.599
48.3%
1.120
47.3%
0.792
50.7%
0.555
63.0%
0.390
59.3%
0.405
53.9%
0.666
65.0%
0.556
52.3%
0.830
62.1%
0.962
66.8%
0.575
68.1%
0.367
63 . 9%
0.624
69.0%
Nitrate
Wt . mean
Position
mg/1
2.04
52.9%
2.86
35.7%
6.38
42.8%
6.48
56.4%
5.22
44.0%
6.49
47.4%
7.30
22.3%
9.35
53.2%
7.06
61.2%
2.34
21 . 5%
1.42
35.8%
7.02
40.3%
4.68
46.1%
2.20
47.6%
7.20
49.2%
2.63
62.3%
6.18
40.4%
7.79
48.6%
8.72
55.0%
3.03
44.4%
Cond.
Wt . mean
Position
pmhos
531.
48.8%
447.
56.1%
485.
54.4%
534.
47.6%
437.
46.7%
539.
49.8%
499.
54.2%
535.
57.7%
546.
51.8%
482.
58.5%
665.
51.3%
592.
46.7%
423.
54.9%
292.
56.9%
513.
48.2%
509.
49.7%
666.
43.8%
689.
55.0%
612.
47.6%
632.
41.1%
338.
56.6%
67
-------
Table
Storm
No.
43
44
45
46
47
48
49
50
51
52
28. Continued
Starting
date Mean flow ,
Ending date Peak flow
(m /sec)
790301
790312
790402
790407
790408
790412
790412
790422
790510
790522
790523
790529
790610
790616
790709
790713
790714
790723
790913
790918
42.
172.
17.
51.
26.
61.
36.
190.
5.
23.
15.
49.
6.
37.
4.
10.
4.
12.
14.
56.
60
57
39
22
38
16
37
6
18
25
38
41
15
53
76
73
84
64
19
83
Samples
Duration
N, hrs
32
276
14
144
10
90
29
264
15
162
162
17
157
13
186
11
126
9
72
19
138
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
Susp.
Wt.
solids Total P
mean Wt. mean
Position
mg/1
293.
71.
230.
60.
310.
60.
902.
73.
529.
83.
71.
70.
331.
81.
364.
75.
318.
45.
295.
61.
3
8%
0
4%
9
8%
7
8%
5
87.
5
5%
5
0%
5
3%
3
4%
8
8%
Position
mg/1
0.
63.
0.
58.
0.
57.
1.
65.
0.
72.
0.
56.
0.
56.
0.
45.
0.
39.
0.
59.
517
9%
495
5%
466
9%
216
7%
906
6%
337
2%
452
1%
622
6%
722
9%
592
5%
Nitrate
Wt . mean
Position
ms/1
3
37
7
44
7
48
4
48
5
37
24
41
9
39
5
50
3
55
2
45
.83
.5%
.45
.5%
.35
.5%
.21
.5%
.54
.42
.48
.9%
.19
.4%
.18
.8%
.56
.5%
.11
.3%
Cond.
Wt. mean
Position
jjmhos
318.
42.6%
523.
52.5%
472.
51.4%
324.
45.6%
554.
-46.0%
625.
50.7%
569.
48.9%
533.
48.7%
489.
56.2%
409.
52.4%
average flow and the flux weighted concentrations of suspended solids, total
phosphorus, nitrate + nitrite nitrogen and conductivity. The table also
includes a calculation of the -percent of the total storm flux of suspended
solids, total phosphorus, nitrate + nitrite nitrogen and conductivity which
accompanied the first half of the water mass of each storm. This percentage
is labeled "position" in the table.
In Figure 12, the complete hydrographs and sediment graphs for eight of
the 52 storms have been plotted. To facilitate comparison of the hydrOgraph
and sediment graphs, all of the storms have been plotted using the same scales
for the concentration, flow and time axes. Storms A., B and C were large
storms in terms of peak flows. The concentrations of suspended sediments in
storm A were very high, in B were moderate and in C were very low. Storms I),
E and F had moderate peak flows and again were accompanied respectively by
Tiigh (P), medium (E) and low (F) concentrations of suspended solids. Storms G
and H had very low peak flows, one of which (G) was associated with moderate
suspended solids concentrations while the other (H) was associated with very
low suspended solids concentrations.
The storm to storm variability present for sediment concentrations is
also present for nutrients. In Figure 13 the flux weighted average
concentrations of suspended solids, total phosphorus, and nitrate-nitrite N
68
-------
A 790412-790422
B 750222-750303
p-4
\
tjl
CO
T)
•H
rH
o
CO
T)
(1)
t)
^
Q)
a
w
5
o
a
d
CD
0
"-)
a
q_
d
o
o
<\j
o
o
d
a
0
*"""*
o
q_
a
q_
d"
a
—a
CO CD
[LI
Ho
0
^d"
DO
-ia
u ^
a
o
d"
o
0
(\i
CD
q_
CTl
194 197 200 203
Day of water year
M Q,
13 C?
•H,d
•^ o
13
? a
OJ CD
3
CO
L44 1.47 L50
Day of water year
1.53
E 2
C 780313-780318
D760723-760728
^^
rH
^^
tjl
tn
-a
•H
rH
O
CO
T~1
VJ
(D
T3
C
Q)
a
to
?
CO
o
q_
d
a
a
en
a
q^
d^
Q
CD
r\l
a
q_
d
o
a
o
q_
o
o
d "
o
^°
C0[£)
fia
o
5?°
DO
-10
U- •^
o
a _
d "
0
o
f\j
o
q __
1.G3 1.G6 1-K3
Day of water year
O
O
5s
en a
g ~\J
CO
•a
o
o
2.96 2^99
Day of water year
. . '»•»»_»«»
l«I9U«tlk--?TC«B ffa3ia5,-u£
3.02 3.05
-------
750830-750904:
o
'- a
!§
r
a
•o _
o a
13 O
Is
w a
3,33
3.36
Day of water year
o
o
c °"
C1 a
gi °
~" 0
tn o
"3 d
"^ °
o °
LO ^
TJ q
(1) CD
•a ci
% °
0) o
w rt
co °
q _
o
q
d"
a
a
CO
o
a
d
a
ga
So
o
£ d"
o a
" O
h oj
o
q_
tt
308
3.16
Day of water year
Figure 12 continued.
F 741231-75010S
CD
—> ^
">• o
g. °
M 3
•a q.
•n a
•a °.
£ a
oj a
9.00 9.30 g.'RO
Day of water year
9.90
E 1
H75-1018-751025
^
s.
tr>
,§
to
T)
>r^
o
•a
O
T)
0)
a
to
CO
o
a
d"
o
o
en
0
o
d"
o
o
C\J
0
q_
d
o
o
— >
o
q.
0
a_
a
a
0
ID
o
a
a
^^o
r/i CD
tj ^*
CJ
q_
oo
_JO
U_CM
o
q_
--m
ti^-^—
1 70 2.00 2.30
Day of water year
-*-p-»..
2.60
E 1
-------
Sr
WLO
~o —
~o Q
if) in
CM
-o -
0)
-£> '
c
01 °
00 •
c ^v
§ ^
s:
"Si
-LO
O>
-
0) (3
^ LfJ~
X ™
h
0
* i LTT
t o-j
L
5
t ^. L-O"
C
0
t t « .
* TJ ^
- -S "
X
* + + .?
^ 4. i + H UJ /
+ t + + .+ + -1
* * '
+ +"^J. T i ,3
. t4
+ t t+++ 0
X
T +
+ ++ iV+ C
+ + + T + +
•f -t 4, 4- "^
+ + T + ++ + v
+ •" ' 4 4
• 1 1 i 1" 1 1 1 (
5 1.8 1.5 2.0 2.5
, * , | | | | | * , 1 OGC Penk Flnu 5 MT3/S
5 ' 1.0 l' 5 ' 2 '. 8 ' 2^5
LOGC Peok Flow 3 MT3/S
m
L ~~
0
KOI
O
x
a. ~~
I3®
0
7 °
0 jo
ECO.
jj
~°>®
e ^
''S
X
rH
>^\-
_j ^
r^"* ' -i-
> T
'4 H Q + +
"f 3^1 4- ^~
+ + O '-0 I -
4- * ^~ "*"
•i + -4- 0 i
^i T
f + ' , L/V" + ^ ^ + N
+ ^ 0 LO t t +
"^ + + 2; ^+'^"T'I' +
^ ^ A t "^ — i v— i-J~ j_
/+^ ; * ^ + t^ "
* + -C t
+ + + + + + f ~ .? \ /
i _ (*o i ~
— T lj ^—^4-
f ^ ^
X J- f
3
i i 1 1 1 -f- — t • r-. iii i ...
-1 T , T . • o PI ' *-— '^ -I ' ' ' ' ' ' — ~ ' '
5.5 10 1 .b 'w ^ fa 05 |0 s 23 25
uOG^ Peok f 1 cw ^ r'.T3/ ^ LOGC Peak - ov. ) MT3/S
Figure 13. Variability in flux weighted mean concentrations of suspended solids, total phosphorus.
nitrates and conductivity in relation to peak flows for individual storms.
-------
along with the flux weighted conductivity are plotted as of function of the
peak flow (log scale) for the set of 52 individual storms at Upper Sandusky-
Again the wide variation in flux weighted concentrations for storms of similar
sizes is very evident. Furthermore the scattergrams reveal no obvious trend
between peak flows and flux weighted concentrations except for conductivity
where average conductivity decreases as peak flows increase. Linear
regressions of concentrations on peak flows, together with correlation
coefficients are shown in Table 29. The correlation for conductivity is much
greater than for the other parameters.
Table 29. Linear Regressions of Flux Weighted Mean Concentrations on
Peak Flows for the Upper Sandusky Storm Data Set.
Parameter
Suspended Solids
Total Phosphorus
Nitrate-N
Conductivity
Slope
-0.0522
-0.0004
-0.0100
-1.37
Intercept
339
0.642
6.01
566
R2
0.0%
0.5%
2.4%
50.4%
DF
49
48
49
50
Probability
» 0.1
» 0.1
» 0.1
« 0.001
The large variability in sediment and nutrient concentrations associated
with runoff events Of approximately equal size is present at all of the stream
gaging stations in northwestern Ohio. The extent of this variability leads to
the following conclusions:
1. Most of the time the sediment transport in the streams is less than
the sediment carrying capacity of the streams at a given flow. The
sediment carrying capacity would be equal to or greater than the peak
sediment concentrations observed at that flow. There apparently is
not a ready source of erodible sediment from stream banks or bottom
that replace sediment "deficits" in surface runoff water entering the
stream network. Thus programs that reduce the sediment concentration
in surface runoff water should reduce the sediment yield from large
river basins.
2. The interaction between watershed factors and precipitation
intensity, duration, and distribution have major impacts on the
concentration of sediments and nutrients present in runoff water.
During the time period of the study at Upper Sandusky there were no
significant changes in farming practices that would have accounted
for variations in stream chemistry. The seasonal changes of ground
72
-------
cover associated with the prevalent rotations and tillage practices
coupled with variations in precipitation patterns are sufficient to
cause the observed variations.
3. The extent of the variability illustrates the complexity of
calibrating models where Output includes sediment graphs and
chemographs for individual storms from large watersheds. The inputs
for such models must include the major variables which cause the
variation in concentrations within storms with equal peak flows.
4. Studies of the sources of the variability in concentrations
associated with runoff events of approximately equal size may be
quite helpful both in planning sediment reduction programs and in
assessing the effectiveness of programs which are implemented. Given
the extent of the variability present within these 52 storms, further
study of these may reveal some of the major factors affecting the
concentration of materials during storm events.
VARIATIONS ASSOCIATED WITH SEASON AND RAINFALL INTENSITY
Individual Storms
One important factor influencing the concentrations of materials present
during runoff events--is the season of the year. In particular, those runoff
events which are produced as a result of snow melt often have much lower
sediment concentrations than events associated with rainfall. This is readily
apparent in the data set of Upper Sandusky storm events. In 'Figure 14, the
month in which each storm event Occurred has been added to the graph of
weighted mean sediment concentrations in relation fo peak flow. It is evident
that the flux weighted average concentrations Of suspended solids are much
smaller in December (12). January (1), February'(2) and March (3) than for
Other months. This is particularly so for storms with medium and small peak
flows.
Storm #4-8 (Table 28) provides an interesting exception to the above
pattern. This storm, which occurred May 23-29, 1979, had a very low sediment
concentration in relation to the spring storms of similar peak flows. This
runoff event followed a series of prolonged gentle rains. In Figure 15
tracings from a recording raingage located in the watershed are shown from the
above storm as well as for a storm which Occurred on April 13, 1979. The
latter storm had much higher sediment concentrations (Table 28, Storm -46).
Storm #48 was also accompanied by extremely high nitrate concentrations
(flux weighted average = 24.48 mg/l). Apparently a high proportion of the
rainfall infiltrated into the soil. The resulting tile effluents had high
nitrate and low suspended solids concentrations. There apparently was little
surface runoff water to dilute the tile effluent.
73
-------
Upper Sandusky Storms
Q
0)
T!
col
o)
2:
LD
.c
CD
0.5
Figure 14.
-4-
i
1
.0 1.5
LOGC Peak Flow
2.9
MT3/S
2.5
Flux weighted mean concentrations of suspended
solids in relation to peak flow with month of
occurrence marked for individual storms.
(Jan. + 1, Feb. + 2, etc.)
Seasonal Sediment and Nutrient Hating Curves
Since the sediment concentrations during snow melt events are lower than
during rainfall induced events, separate sediment rating curves are often
produced for winter and summer conditions (strand, 1975). These curves
involve a separate plotting of sediment concentrations in relation to stream
discharge for winter periods and for summer periods, "^igure 16 shows such a
plot for the Melmore gaging station.
In Figure 16a the sediment concentrations are plotted in relation to
stream discharge while in Figure I6b sedimenb flux is shown in relation to
stream flow for the same data. The period from December through March is
taken to include the period when snow melt generated events are likely to
occur. The April through November period is taken as the non-snow melt
period. In this region, midwinter thaws are common and snow seldom
74
-------
Ul
April 14
TIME (hrs.)
24
'I'""
24
May 24
May 25
14 17
May 26
24
Figure 15.
Raingage tracings from storms on April 13 and May 24-26, 1979. The intense
rainfall of April 13 was associated with runoff #46 at Upper Sandusky while
rainfall on May 24-26 generated runoff #48.
-------
3 HFLMORF SJMMFR WINTER
C3
O
CD
C3
C.I
LLJ
CIO
cn
o
a
* i
*
* A
»*
",
2.00 2^27 2.53 2.80
- — LOG OF FLOW CFS
MELMQRF SUMMER WINTER
3.07
3.33
3.60
B
i. *"'
Ji .00
Figure 16.
2.00
-Ujm:
250
FI nu rp.s
3.50
Seasonal sediment rating curves at the Melmore
gaging station: 16A Summer (A) and winter
(X) sediment concentrations; 16B Summer (£.)
and winter (O) instantaneous sediment fluxes.
76
-------
accumulates over the entire winter. More northerly regions are characterized
by single, large, snow melt events in the spring.
A convenient way of examining the seasonal aspects of concentration
versus flow relationship in streams is to use the flow intervals provided
within the USGS flow duration tables, as described in the methods section.
All of the winter samples were sorted into the flow intervals between adjacent
flow classes and within each interval the means and the standard deviations
for sediment and nutrient concentrations were determined. Similar
calculations were made for the samples collected during summer conditions.
For each flow interval the significance of the differences between the winter
and summer mean values were tested using a student t-test.
Tables 30 to 33 show the resulting means, standard deviations, and
probabilities for the Melmore stream gage. In Figure 17 the summer and winter
mean values are plotted according to the arithmetic midpoint of each flow
interval. In the cases of suspended solids (Table 30 and Figure 17a), summer
values are significantly higher than winter values over the entire range Of
flows. For total phosphorus summer concentrations are significantly higher
than winter concentrations (Table 31 and Figure 17b) in the high and medium
flow ranges. At lower flow ranges there is no significant difference in
summer and winter values except at the lowest flows where winter values are
statistically significantly higher than summer values.
For nitrates (Table 32 and Figure 17c ) there is no statistically
significant difference between winter and summer for most of the flow
intervals across the entire range of flow values. For a few flow ranges,
summer values were higher than winter values. During both summer and winter
the highest concentrations tended to Occur just below the peak flow intervals.
In the case of conductivity (Table 33 and Figure 17d) at high flows there
was either no significant differences or else the summer values had higher
conductivity than winter values. Over the medium and low flow intervals, the
winter values for conductivity are significantly higher than the summer
values. This is probably due to ice formation which concentrates the solutes
in the remaining liquid water. The use Of road salt may also have some effect
on the elevated conductivities in the winter.
The summer and winter mean concentrations for each flow interval were
calculated for sediments and nutrients for all Of the transport stations. T'he
resulting summer and winter rating tables were used to calculate loading when
concentration data was lacking (see Section 7). The seasonal data for the
other stations were similar to the examples for the Melmore station shown
above.
ANNUAL VARIATIONS IN SEDIMENT AND NUTRIENT CONCENTRATIONS
The storm to storm and seasonal variability in sediment and nutrient
concentrations described above results in significant variations in flux
weighted mean concentrations of sediments and nutrients as measured on an
annual basis. In Table 34, the annual flux weighted mean concentrations are
shown for the twelve gaging stations. The total annual discharge according to
final USGS figures is shown for each station along with the percent of the
77
-------
Table 30. Comparison of sunnier and winter suspended solid* concentrations by
flow intervals at the Melmore sampling station.
Table 31. Comparison of sunnier and winter total phosphorus concentrations by
flow interval* at the Melmore sampling station.
Flow Class
33-32
32-31
31-30
30-29
29-28
28-27
27-26
26-25
25-24
24-23
23-22
22-21
-j
00 "-20
20-19
19-18
18-17
17-16
16-15
15-14
14-13
13-12
12-11
11-10
10-9
N
4
9
38
30
37
49
41
69
52
46
72
61
91
97
103
102
100
142
129
89
78
42
15
4
Sumer
Mean
ng/1
426.5
957.0
218.3
137.7
225.4
193.4
264.3
175.7
235.1
233
287.9
190.6
130.3
93.00
90.25
51.35
31.98
21.25
17.86
19.46
15.77
17.22
31.61
24.12
S.D.
59.08
1160
279.2
97.64
219.3
387.6
428.8
185.7
228.1
222.1
569.6
345.7
140.4
112.0
196.9
54.7
38.6
19.1
14.54
15.66
8.3
13.25
13.66
10.11
N
14
10
46
58
38
44
42
38
22
22
25
. 39
47
41
18
20
13
5
16
19
11
11
Winter
Mean
216.3
131.9
130.0
88.66
84.13
77.58
111.5
66.03
39.53
47.69
38.87
28.10
30.23
15.33
10.98
7.30
12.39
4.38
7.531
5.154
8.236
5.855
S.D.
193.0
88.45
93.94
61.98
82.83
53.92
119.9
64.82
42.66
50.50
38.18
27.44
39.53
14.79
9.044
10.62
21.18
2.066
4.839
3.685
4.850
3.155
D.F.
16
17
82
86
73
91
81
105
72
66
95
98
136
136
119
120
111
145
143
126
87
51
T P1
2.10851 x
2.2493 xx
2.0135 xx
2.8730 xxx
3.709 xxxx
1.96404 x
2.2223 xx
3.51B05 xxxx
3.97695 xxxx
3.8519 xxxx
2.17688 xx
2.92431 xxx
4.78179 xxxx
4.41959 xxxx
1.70186 x
3.57676- xxxx
1.79036 x
1.96813 x
2.814 xxx
5.62551 xxx
2.93149 xxx
2.80509 xxx
Flow Class
' 33-32
32-31
i
31-30
30-29 /
29-28
28-27
27-26
26-25
25-24
24-23
23-22
22-21
21-20
20-19
19-18
18-17
17-16
16-15
15-14
14-13
13-12
12-11
11-10
10-9
N
4
9
38
30
36
49
41
69
52
46
72
61
92
97
104
103
106
142
126
91
79
42
15
4
Sumar
Mean
ng/1
1.052
1.286
.5317
.4657
.5787
.434
.489
.407
.439
.407
.478
.360
.317
.251
.217
.189
.154
.132
.122
.137
.143
.139
.165
.225
S.D.
.076
.8868
.2579
.1263
.5589
.3369
.3551
.2170
.2253
.2018
.5106
.2652
.1713
.1490
.1967
.1325
.1243
.0721
.0761
.1147
.0792
.0751
.0388
.2059
N
14
10
46
58
38
44
42
38
22
21
25
39
47
41
18
20
13
5
17
43
13
11
Winter
He an
.5203
.359
.383
.324
.315
.305
.355
.272
.216
.193
.172
.147
.137
.120
.131
.139
.204
.176
.149
.094
.237
.199
S.D.
.268
.142
.131
.087
.157
.122
.208
.115
.112
.106
.088
.085
.134
.065
.073
.076
.132
.097
.124
.092
.090
.089
D.f.
16
17
83
86
72
91
81
105
72
65
95
98
137
136
120
121
117
145
141
132
90
51
T
3.846
3.27
3.416
6.179
2.80
2.40
2.10
3.56
4.41
4.57
2.97
4.85
6.28
5.41
1.82
1.63
-1.36
-1.32
-1.26
2.15
-3.89
-2.28
P1
xxx
xxx
xxx
xxxx
xxx
XX
XX
xxxx
xxxx
xxxx
XXX
xxxx
xxxx
xxxx
X
HS
HS
HS
HS
XX
xxxx
XX
1. Probabilitiesi M.S.
- <0.001.
not significant, x - <0.10, xx - <0.05, xxx - <0.01.
1. Probabilities: N.S. - not significant, x - < 0.10, xx
- < 0.001.
< 0.05, xxx - < 0.01,
-------
Table 32. Comparison of lunmer and winter NOrN concentrations by flow class
interval* at the Helraore (anpling station.
Table 33 • Comparison of stunner and winter conductivities concentrations by flow
intervals at the Melmore sampling station.
Flow Cla»«
33-32
32-31
31-30
30-29
29-28
28-27
27-26
:t ::
25-24
24-23
23-22
22-21
21-20
20-19
19-18
18-17
17-16
16-15
15-14
14-13
13-12
12-11
11-10
10-9
N
4
9
38
30
37
49
41
69
52
46
71
61
91
97
105
101
10*
142
127
89
79
42
14
4
Susmer
Mean
3.522
3.813
7.590
6.869
6.646
5.817
4.791
5.784
7.231
5.801
S.MS
5.689
5.610
S.697
4.721
5.110
1.441
2.940
2.062
2.195
2.017
2.220
1.511
2.617
S.D.
.0806
.2602
6.279
4.863
3.088
3.110
2.787
3.171
4.607
4.287
4.670
4.092
4.086
4.519
3.716
4.541
1.260
2.858
1.951
1.699
.493*
.63*8
.6215
1.096
N
14
10
46
58
18
44
42
38
22
22
25
19
47
41
18
20
11
5
17
41
11
11
Winter
Mean
•9/1
3.326
3.416
5.011
4.791
5.916
5.795
5.541
4.901
4.465
5.628
5.221
4.896
4.284
3.999
1.618
4.261
4.100
4. MO
1.915
1.828
2.268
2.156
S.D.
1.590
1.704
1.898
2.124
2.494
2.696
2.700
2.181
1.828
2.169
1.989
2.144
1.606
1.415
2.057
1.26S
.9144
.7158
.4816
.4618
.2948
.1416
D.r.
16
17
82
86
73
91
81
105
72
66
94
98
116
116
"1
121
117
145
142
110
90
51
T
.241144
.68978
2.64616
2.71836
1.1274
.0162
-1.2452
1.475
2.71815
.178191
.167716
1.112
2.11787
2.15196
1.218
1.02016
- .722892
-1.1062,8
.108525
1.18917
-1.7769
.127948
P1
NS
NS
XX
XXX
NS
NS
NS
NS
XXX
MS
NS
HE
XX
XX
MS
KB
MS
MS
NS
MS
X
NS
Flow Class
33-32
32-31
31-30
30-29
29-28
28-27
27-26
26-25
25-24
24-23
23-22
22-21
21-20
20-19
19-18
18-17
17-16
16-15
15-14
14-13
13-12
12-11
11-10
10-9
N
4
9
38
30
37
49
41
69
52
45
72
61
93
98
105
103
106
143
129
91
79
42
15
4
Summer
Mean
umhos
233.5
205.6
349.2
354.7
345.6
375.4
376.4
397.9
420.4
428.8
410.5
469.9
510.3
554.7
575.5
611.5
623.8
639.0
630.6
631.5
617.6
612.2
601.9
642.3
S.D.
58.9
69.13
84.87
64.73
80.88
62.32
67.58
59.78
71.11
85.77
92.99
111.2
100.1
106.9
99.42
95.84
79.75
81.06
64.64
68.35
35.20
41.75
42.85
57.31
N
14
10
46
58
38
44
42
38
22
21
25
39
46
41
17
19
9
4
16
43
13
11
Winter
Mean
umhos
230.1
276.6
310.7
311.7
358.0
387.6
409.7
408.1
452.9
520.3
541.1
581.2
621.8
675.2
717.1
750.5
749.9
818.5
867.7
852
955
909
S.D.
51.6
129.0
46.53
47.16
46.46
76.88
64.48
84.94
55.41
90.06
78.68
76.69
71.92
80.97
65.61
47.80
38.25
61.19
124.5
40.45
78.12
89.31
D.F.
16
17
82
86
73
96
81
105
72
64
95
98
137
137
120
120
113
145
143
132
90
51
T
.1130
-1.4694
2.6355
3.5586
- .816848
- .861824
-2.29714
- .724526
-1.90974
-3.973S7
-6.27957
-5.4695
-6.73B01
-6.47694
-5.66527
-6.1666
-4.68362
-4.38772
-12.2118
-19.57
-25.947
-15.0078
P1
NS
NS
XX
xxxx
NS
NS
XX
NS
X
xxxx
xxxx
xxxx
xxxx
xxxx
xxxx
xxxx
xxxx
xxxx
xxxx
xxxx
xxxx
1. Probabilitiesi M.S. - not lignifleant, x • < 0.10,
xxxx • < 0.001.
- < 0.05,
- < 0.01,
1. Probabilities: N.S. " not significant, x
xxxx - < 0.001.
< 0.10, xx - < 0.05, xxx - < 0.01,
-------
ao
o
Honey Creek Near Me I more
1000 „.
50 2
LOGC
00 3 50 4 00 4 50 5 00
Midpoint Flow > Ft.T3/S
1B0 _
60..
0
§20
•
2 50 3 00 3 50 4 00 4 50 5 00
LOGC Midpoint Flow ) F-tT3/S
1000 _
2
50 3
LOGC
Figure 17.
4 50 5
FtT3/S
00
50 3 00 3 50 4 00 4 50 5
LOGC Midpoint Flow ) Ft.T3/S
00
00 3 50 4 00
M i dpoin t FIow
Su^er and Winter average concentrations plotted in relation to the midpoint of the flow duration
class intervals for Honey Creek at Melmore.
-------
Table 34. Annual variation* in flux waightad Man concentrations of a*di*anta and nutrlanta at northMitam Ohio
gaging etntlona.
Mauma
Portage
•ucyxue
Nevada
Upper
Sanduaky
Tymochtee
Mexico
Helmore
Uolf Neat
Nolf Eaat
Hater
Year
1975
1976
1977
1978
1975
1976
1977
1978
1975
1976
1977
1978
1979
1976
1977
1978
1979
1975
1976
1977
1978
1979
1*75
1976
1977
1978
1979
1975
1*76
1977
1978
1979
1976
1977
1978
1979
1976
1977
1978
1979
1976
1977
1478
1979
Diechaxge
io7.3
476.3
503.5
305.3
616.6
28.99
32.62
24.13
44.38
10.27
8.21
5.72
9.55
12.03
4.83
5.21
12.25
9.91
J7.53
22.05
14.11
• 3O.68
27.37
17.71
7.64
6.60
20.73
14.42
66.6*
47.83
36.07
75.13
63.75
6.91
7.29
14.73
15.02
3.01
2.85
6.74
4.59
4.40
11.50
11.21
6.27
Percent
Honitored
73
86
70
96
69
81
78
98
35
78
97
99
58
e«
79
65
90
92
»1
113
72
81
89
67
W
73
76
•
6
121
112
109
105
80
86
91
105
99
99
46
67
Suapended
Sol Ida
-9/1
279
315
404
138
275
161
128
132
96
219
ISO
110
256
316
198
81
310
226
338
212
13*
Ml
310
205
121
61
1*3
125
ai
352
134
226
87
67
371
205
77
ISO
269
346
77
194
272
Total
Phoephorua
•9/1
.577
.554
.739
.396
.496
.400
.389
.359
.441
.596
.673
.460
.604
.451
.276
.226
.522
.402
.570
.646
.431
.613
.570
.362
.325
.223
.413
.335
.112
.657
.310
.4*4
.284
.252
.524
.431
.231
.355
.498
.558
.323
.373
.580
Soluble
Heactive Nitrate
Phoaphoru* Nitrogen
•9/1 »9/l
.114
.107
.098
.111
.133
.111
.140
.100
.197
.190
.284
.195
.161
.054
.051
.074
.062
.0*1
.105
.18*
.175
.123
.06*3
- -9*«
.078
.072
.070
.069
.070
.119
.073
.053
.0*3
.083
.081
.136
.083
.091
.101
.088
.135
.095
.132
6.72
3.58
6.82
4.40
8.14
3.82
8.86
4.84
2.56
3.15
4.26
3.00
4.60
3.6*
5.13
3.72
5.86
4.42
3.71
4.23
3.65
5.88
5.7*
5.W
5.84
3.43
6.17
7.78
t.DO
4.81
3.01
3.71
5.83
3.78
5.17
3.S8
8.16
5.26
7.46
5.14
4.17
5.42
4.87
Conductivity TP/S8
UBhoa 9/*g
522
503
554
448
582
556
645
489
484
44O
501
432
428
424
485
4 JO
374
438
46O
591
449
483
370
317
495
445
422
4*1
7O3
553
513
366
45)
364
364
411
567
396
525
497
661
552
464
2.07
1.76
1.83
2.87
1.80
2.48
3.04
2.72
4.59
2.72
3.74
4.18
2.36
1.43
1.39
2.7*
1.68
1.79
1.69
3.05
3.10
2.04
1.84
1.77
2.6*
3.66
2.14
2.68
5.33
1.87
2.31
2.14
3.26
3.76
1.93
2.10
3.0O
2.37
1.85
1.61
4.19
1.92
2.13
pp/ss1
9/X9
1.66
1.42
1.59
2.06
1.32
1.80
1.95
1.96
2.54
1.85
2.16
2.41
1.73
1.26
1.14
i.ea
1.48
1.38
1.38
2.16
1.84
1.63
1.64
1.45
2.04
2.48
1.78
2.13
2.00
1.53
1.78
1.90
2.21
2.52
1.63
1.44
1.92
1.76
1.47
1.36
2.44
1.43
1.65
81
-------
Table 34. continued
Ma tar
Tear
Fremont 1975
1976
1977
197B
1979
Huron 1973
1976
1977
1978
1974
Discharq*
io7.3
103.
77.
62.
139.
1O8.
3O.
26.
25.
40.
0
16
90
1
8
M
71
35
48
29.55
Pvrcant
Kmitorwi
67
54
93
83
81
ao
83
B6
87
27
Suapendad
Solids
•9/1
294
198
160
1*»
272
2S1
232
279
119
209
Total
Phosphorua
-9/1
.513
.401
.416
.357
.531
.403
.293
.436
.291
.490
Soluble
RaactivB
Phosphorus
•9/1
.067
.072
.106
.075
.104
.080
.141
.088
.108
.125
Nltrata
Nitrogen
•9/1
4
3
4
4
4
3
2
5
2
5
.99
.82
.96
.12
.87
.66
.31
.30
.99
.26
Conductivity TT/SS
ua*tos 9 /Kg
439
533
577
431
414
465
578
515
556
581
1.74
2.03
2.60
2.41
1.95
1.43
1.26
1.56
2.44
1.70
PP/SS
9/Kg
1.52
1.66
1.94
1.91
1.57
1.15
.66
1.25
1.54
1.26
Pmrtlculata phoaphoraa (TV) la aatlawead by subtracting SVf trom TV.
82
-------
annual discharge which was monitored through the sampling program. The latter
value was based on the instantaneous flows and time multipliers associated
with sediment samples.
Values differ from 100$ for a variety of reasons including Our use of
interim flow data for hourly stages and missing data due to sampler or stage
recorder malfunction. The number of samples analyzed for each parameter each
year at each station has been shown in Table 25-
In the case of sediments, the ratio of the highest to the lowest annual
mean concentration exceeds 2 at all of the stations. Total phosphorus and
nitrate concentrations also show large annual variations at most of the
stations while smaller variations are present for soluble reactive phosphorus
and conductivity.
Comparison Of the 1978 and 1979 water years is interesting in that for
both years the total discharge was large and rather similar. The sediment,
total-phosphorus and nitrate flux weighted concentrations were much higher in
1979 than in 1978. The monthly distribution of runoff for the Fremont and the
Melmore stream gages is shown in Table 25. In 1978 the December through March
period accounted for 76.1$ and 73.1$ of the total annual discharge at the
Fremont and Melmore gages. During the 1979 water year these months accounted
for 45-5$ and 43.7$ of the annual discharge. The seasonal distribution of
runoff undoubtedly accounted for the major differences in the mean annual
concentrations of sediments and nutrients between 1978 and 1979.
The extent., of the variations in concentrations from year to year, even
when the sampling programs are based on many samples collected throughout the
year, illustrates the need for long term studies to document loading of
nutrients and sediments from nonpoint sources*. Seasonal analysis of runoff
may account for major portions of the variability but defining the seasonal
concentration-flow relations may require extended studies.
VARIATIONS ASSOCIATED WITH LOCATIONS RELATIVE TO POINT SOURCES
Another factor affecting the patterns of concentrations observed at
sampling stations is the position of the station relative to point source
inputs. Where sampling stations are located immediately downstream from point
source inputs of phosphorus, the concentration versus flow plots show
increasing concentration of phosphorus as flows decrease in the low flow
range. This reflects decreasing dilution of the point source inputs by the
stream water. The actual concentrations observed depend on the concentrations
and flow of the point source effluent in relation to the stream flow and
.concentration. As the sampling location is moved downstream, the phosphorus
concentrations under low flow conditions decrease due to uptake by benthic
algae, chemical precipitation, or adsorption onto bottom sediments.
During the summer of 1974, from 32 to 52 samples were collected during
nonstorm conditions at each of twenty six stations along the mainstream of the
Sandusky River. In Figure 18 the mean concentrations Of total and soluble
reactive phosphorus are shown as a function of distance from the river mouth.
The effects of inputs from the "Rucyrus, Upper Sandusky and Tiffin sewage
treatment plants are evident. None of the plants had phosphorus removal
83
-------
Table 35.
Comparison of Monthly Distribution of Runoff between 1978 and 1979
Water Year.
Fremont Gage
1978 1979
Discharge
106m3
October
November
December
January
February
March
April
May
June
July
August
September
32.06
14.08
305.97
90.56
36.19
626.64
209.29
49.64
22.53
9.23
12.53
2.48
Percent
0.
1.
22.
6.
2.
45.
15.
3.
1.
0.
1.
0.
9%
0%
0%
5%
6%
0%
0%
5%
6%
7%
0%
2%
Discharge Percent
106m3
4
5
35
55
96
308
272
85
52
31
96
43
32
.58
.10
.71
.53
.12
.77
.34
.18
.51
.69
.75
0
0
3
5
8
28
25
.5%
.6%
.2%
.1%
.9*
.3%
.0%
7.8%
4
2
8
4
.8%
.9*
.9%
.0%
Melmore Gage
1978 1979
Discharge Percent
106m3
.977
1.63
39.27
12.21
4.51
58.05
22.54
4.53
1.49
.337
2.13
.779
0.8%
1.3%
28.0%
1.0%
3.6%
41.0%
16.3%
3.2%
1.1%
0.3%
1.5%
0.7%
Discharge Percent
lo6™3
.23
.32
6.86
9.53
15.08
34.04
39.64
12.13
16.83
3.78
6.86
4.89
0.
0.
4.
6.
1%
2%
6%
3%
10.2%
22.
26.
a.
11.
2.
4.
3.
6%
4%
1%
2%
5%
5%
3%
Total
1391.2
1087.6
148.5
150.2
programs in operation at that time. The decreases in phosphorus concentration
below each town reflect deposition of phosphorus rather than dilution effects
(Baker and Kramer, 1976).
Another effect of point source inputs which may he present deals with
diurnal fluctuations in the loading rates from the plants. Figure 1Qa
illustrates the patterns of phosphorus concentrations observed at a series of
stations downstream from Bucyrus, Ohio. Three grab samples were collected
each day at each station. Although the phosphorus concentrations generally
decreased in a downstream direction, the pattern at the upstream stations
"sometimes showed highest concentrations at Kestetter Road and sometimes at
Denzer Road. If samples had been collected only on 760826 at 1600 or 2100
hours one might have concluded that a source of phosphorus existed between
Kestetter and Denzer Roads. Collections at other times could also have lead
to similar conclusions.
Simultaneously with the grab sampling program used to obtain the above
data, automatic samplers were used at Kestetter, Denzer and Caldwell Roads.
The samplers were set to collect samples at two hour intervals. The results
84
-------
c
o
•^
JJ
"3
^
-P
0)
U
Ǥ
to
o
•a
to
o
Bucyrus
STP
1.50
1.00
0.50
Total Phosphorus
Soluble Phosphorus
200
150
100
50
Kilometers from river mouth
^ Direction of Flow
Figure 18. Profiles of mean phosphorus concentrations along
the Sandusky River during June - September 1974.
of this program are shown in Figure 19b. Large diurnal variations in
phosphorus concentrations were present at Kestetter Road, which is located
about 0.9 km. downstream from the Bucyrus sewage treatment plant. The rather
confusing pattern (Figure 19a) apparent from the grab sampling program was
composed of isolated parts of clearly defined diurnal inputs from the sewage
treatment plant. The occurrence Of higher values at Denzer Road were a
consequence of the travel time between the two stations. It is apparent from
a comparison of the Kestetter and Denzer Road concentrations that longitudinal
mixing of the water during its passage between the two stations dampened the
concentration variations present at Kestetter Road. At Denzer the low
concentrations were higher than the low concentrations at Kestetter Road.
The data presented in Figure 19 clearly illustrate the need for
examination of possible diurnal effects at sampling stations located
immediately downstream from treatment plants. Where the diurnal loading from
85
-------
CD
cn
Figure 19a
Yr/Mo/Day Hour o b
M
O
ui
ob
o
'o
in
b
o
o
760825
2100
760826 0900
1600
2100
760827
760828
760829
760830
1000
1500
2100
1000
1500
2100
1500
2000
ISOO-i
2100
Kestetter Rd.
i i i i
Denzer Rd.
i i i i
Caldwell Rd.
1 L _L J
Direction of Flow
Figure 19b
ui
o
O
o
ui p
ooo
2400 _
1200
2400
1200 -
I
Kestetter Rd. -
Denzer Rd.
i i i i
Caldwell Rd.
Direction of Flow
Figure 19. Comparison of phosphorus concentration profiles obtained from 3 samples per day and 12 samples
per day at three bridges on the Sandusky River downstream from Bucyrus, Ohio, for the period
between August 25 and August 30, 1976.
-------
point sources is associated with significant diurnal variations in discharge
relative to stream flow, calculating material fluxes at sampling stations
downstream from a stream gage are extremely difficult. The peak discharge
rates at the point source input will generate wave fronts which move
downstream faster than the water is flowing. For example, if the peak
concentrations and peak flows occurred simultaneously at Kestetter Road the
peak flow would have proceeded the peak concentrations at Denzer HOad, since
the wave front moves downstream faster than the velocity Of the water. T'hese
effects may be important when attempting to quantify processing rates.
Verhoff and Baker (1980) have used the diurnal variation data to calibrate a
phosphorus deposition model for the Sandusky River below Bucyrus.
CONCENTRATION EXCEEDENCY RELATIONSHIPS
An important aspect of water quality is the percentage Of time that the
concentrations of some parameter fall vfithin particular ranges. Although
these studies were primarily directed toward measuring tributary loading,
which is dominated by high flow periods, the automatic samplers were Operated
continuously during both event and non-event periods. During nOn-event
periods a single sample per day was analyzed. As described in the methods
section, different time multipliers were attributed to the samples, depending
On the frequency of collection Of the analyzed samples. The inclusion of
daily samples during low flows provides a data base which allows calculation
of the time exceedency distribution of various concentrations. The
calculational procedures used for producing exceedency tables are described in
the methods section.
Table 36 contains concentration exceedency data for suspended solids at
selected stations. For the Maumee station the 'data indicates that 80^ of the
time the suspended solids concentration exceeded 23-9 mg/1, and 1 $ of the time
the concentration exceeded 778 mg/1. In comparing the concentration
exceedency data for the various stations, it is noteworthy that the Tlaumee
River, which is the largest of the watersheds, had much higher sediment
concentrations than the other stations for much Of the time. particularly for
the low range of concentrations (ie. high exceedency percentages) the
concentrations exceeded fixed percentages of the time were higher for the
larger rivers and tended to decrease as the watershed area decreased. At the
high concentration range (ie low exceedency percentages), concentrations were
highest for the watershed with the highest gross erosion rate (Nevada) and
lowest for that with the lowest gross erosion (¥olf, West). It should "be
recalled that high suspended solid concentrations Occur at high flows.
Table 37 contains concentration exceedency data for total phosphorus.
High total phosphorus occurs at times of high suspended solid concentrations
and also at low flows where the stations are affected by point source inputs.
The concentrations of total phosphorus which were exceeded large percentages
of the time were also higher in the Maumee than in the Other watersheds. mhe
total phosphorus concentrations exceeded 0.5% of the time were higher at the
Portage and Nevada stations than at the Maumee. For the Maumee and Nevad?.
stations the high total phosphorus concentrations occurred during high flows
while for the Portage most of the high concentrations Occurred at low flows.
87
-------
Table 36 . Percentage of time the indicated concentrations of suspended solids
(mg/1) were exceeded at representative gaging stations.
Percent
Exceedency
99%
98%
95%
90%
80%
70%
60%
50%
40%
30%
20%
10%
5%
2%
1%
0.5%
Maumee
3.5
4.0
5.4
10.0
23.9
42.0
54.8
67.5
83.8
103.0
138.0
221.0
319.0
604.0
778.0
1059.0
Portage Tindall Melmore Wolf West
mg/1 2.1 mg/1
2.9
4.2
5.9
8.2
12.2
18.6
27.0
39.3
51.8
74.9
134.0
220.5
429.0
570.0
938.0
2.1 mg/1
2.9
4.4
6.1
11.5
17.5
25.4
34'. 7
47.9
66.9
100.0
176.0
285.0
483.0
738.0
874.0
1.4 mg/1
2.3
3.3
4.4
6.7
9.4
13.2
19.0
26.7
38.8
61.5
123.2
191.0
381.0
676.0
961.0
1.7 mg/1
2.3
3.4
4.8
6.9
10.2
14.2
18.5
22.9
31.6
44.3
78.4
134.0
271.0
463.0
667.0
Nevada
1.3 me
1 .9
3.1
4.5
7.5
14.3
21.6
29.6
40.8
61.4
85.2
142.0
230.2
436.0
841.0
1436.7
Watershed
area Km 16,395 1,109 3,240 386 171.5 271
Gross Erosion
m.t./ha/yr. 6.84 5.00 8.25 6.86 4.19 9.31
Table 37. Percentage of time the indicated concentrations of total phosphorus
(mg/1) were exceeded at representative gaging stations.
Percent
Exceedency
99%
98%
95%
90%
80%
70%
60%
50%
40%
30%
20%
10%
5%
2%
1%
0.5%
Maumee
.142
.158
.173
.190
.221
.246
.273
.296
.321
.353
.401
.504
.685
.943
1.128
1.480
Portage
mg/1 -059 mg/1
.073
.106
.133
.166
.197
.230
.273
.325
.401
.494
.673
.943
1.29
1.49
1.72
Tindall
.051
.058
.073
.088
-111
.127
.142
.164
.184
.220
.287
.406
.571
.848
1.080
1.279
Melmore
mg/1 -028
.035
.050
.067
.089
.101
.119
.140
.163
.199
.257
.366
.483
.725
.978
1.24
Wolf West
mg/1 -022 mg/1
.043
.054
.067
.092
.118
.143
.176
.208
.250
.313
.464
.601
.928
1.160
1.290
Nevada
.013 mg/1
.018
.024
.034
.047
.062
.078
.101
.128
.154
.197
.312
.416
.705
1.23
1.58
88
-------
Table 38. Percentage of time the indicated concentrations of nitrate-
nitrogen (mg/1) were exceeded at representative gaging stations.
Percent
Exceedency
99%
98%
95%
90%
80%
70%
60%
50%
40%
30%
20%
10%
5%
2%
1%
0.5%
Table 39.
Percent
Exceedency
99%
98%
95%
90%
80%
70%
60%
50%
40%
30%
20%
10%
5%
2%
1%
0.5%
Maumee Portage Tindall Melmore
.01 raq/1 -02 mg/1 .02 mg/1 .480 mg/1
.03 .06 .03 .770
.13 .21 .09 .970
.44 .47 .240 1.260
1.45 1.07 .510 1.670
2.01 1.70 1.190 2.020
2.57 2.21 1.800 2.340
3.20 2.93 2.560 2.820
4.30 3.79 3.140 3.330
5.32 4.80 4.100 3.990
6.53 6.43 5.410 4.990
8.02 8.69 7.200 7.380
9.10 10.00 B.800 9.360
10.60 11.89 10.900 13.300
12.30 13.10 13.820 16.110
13.00 14.10 16.460 18.520
Wolf West
.040 mg/1
.070
.110
.240
.500
.900
1.55
2.66
3.61
4.54
6.16
8.21
9.70
11.30
14.60
16.39
Nevada
.059 mg/1
.070
.130
.240
.500
.800
1.620
2.230
2.870
3.800
5.000
7.200
9.440
13.00
16.00
18.80
Percentage of time the indicated concentrations of total dissolved
solids as respresented by conductivities (urahos) were exceeded
at representative gaging stations.
Maumee Portage Tindall Melmor-e " Wolf West
319 umhos 293 urahos 267 umhos 202 umhos.,-
348 385 307 236
404 486 389 310
464 585 467 369
520 684 553 465
558 746 623 533
590 792 667 580
627 834 704 607
6f>5 883 743 632
707 930 779 659
765 1017 865 689
891 1146 974 736
981 1384 1077 822
1112 1460 1217 876
1239 1510 1305 919
1245 1554 1360 953
261 urahos
320
435
525
614
665
691
717
736
779
887
1037
1136
1331
1419
1462
Nevada
' 225 umhos
279
375
460
546
590
617
647
674
702
731
801
859
887
904
926
89
-------
The drinking water standard for nitrate-nitrogen in Ohio is 10 mg/1. At
all Of the stations this concentration is exceeded more than 2% of the time
(Table 38). At the Portage station 10 mg/1 is exceeded 5$ of the time.
Concentrations of 7 mg/1 are exceeded more than 10$ of the time at all of the
stations. Since shifts to no-till agriculture in the basin may increase tile
flow in proportion to surface runoff, the proportion of time nitrate
concentration standards are exceeded may increase. Most of the high nitrate
concentration occur in Nay and June.
In Table 39 concentration exceedency data for conductivity is presented.
The Ohio drinking water standards for dissolved solids, when expressed as
specific conductance, set a limit of 1200 umhos and a monthly average of 800
umhos. Both of the above conditions are exceeded within the major streams of
northwestern Ohio. In the case of the Portage River, conductivities of 1200
are exceeded 8.0$ of the time. At each station the highest conductivities
occur during the winter low flow periods (see Figure 17). The relative
importance of ice formation and salt use for road deicing is not known at this
time.
Concentration exceedency data Of the type discussed above are useful for
assessing in stream water quality and the impact of proposed control programs
on stream water quality. This format of presenting the data also lends itself
to risk assessment considerations. In comparing the time exceedency data
among the various watersheds it is also apparent that some systematic
differences occur in"relationship to drainage basin size. The concentrations
of sediments and total phosphorus which are exceeded high percentages of the
time tend to increase as drainage basin size increases. The probability of
thunderstorms occurring within a basin increases as the basin size increases.
Longer travel times are required for storm generated water to move through
large basins. Even at low flows, the linear velocities of water would tend to
increase with increasing stream order and. this increased velocity could
increase suspended solids transport. This is consistent with the concept of
increasing transport of fine particulate organic matter in high order streams
(Cummins, 1975).
SEDIMENT-PHOSPHORUS RELATIONSHIPS
A large portion of the total phosphorus transported in river systems is
associated with suspended solids. Often phosphorus loading is estimated
through first estimating sediment yields and then multiplying by a
phosphorus-sediment ratio. The latter can be obtained from empirical
determinations or the use of phosphorus enrichment ratios and soil phosphorus
values.
In Table 40 the nutrient-sediment ratios as measured for the northwestern
Ohio river basins using the entire data sets for each station are listed. Tn
this case the estimates are based on the mean annual fluxes of nutrients and
sediments (see Section 7) rather than on the flux weighted mean
concentrations. The highest ratios of phosphorus to sediment were observed at
the Bucyrus station. This station is located a short distance downstream from
the Bucyrus sewage treatment plant and the high ratios observed at that
station undoubtedly reflect the effect of phosphorus derived ^rom coint
sources.
90
-------
Table 40. Nutrient-Sediment Ratios for Agricultural Watersheds
of Northwestern Ohio, 1974-1979
Gaging Station
Maumee, Waterville
Portage, Woodville
Huron, Milan
Sandusky , Fremont
Sandusky, Mexico
Sandusky , Upper S .
Sandusky, Bucyrus
Tymochtee , Crawford
Honey Cr. , Melmore
Broken Sword , Nevada
Wolf Cr. , East Br.
Wolf Cr. , West Br.
y
Mean
St. Dev.
TP/SS
g/kg
2.13
2.45
1.64
2.09
1.79
2.20
3.31
2.04
2.24
1.64
2.30
2.15
2.17
.44
SRP/SS
g/kg
.48
.73
.47
.43
.29
.57
1.26
.34
.56
.25
.65
.55
.55
.26
PP/SS
g/kg
1.65
1.72
1.17
1.66
1.50
1.63
2.05
1.71
1.68
1.39
1.65
1.61
1.62
- .21
NO N/SS
g/kg
20.3
35.9
16.4
21.2
14.6
16.6
19.8
25.0
26.9
20.0
26.0
33.6
23.0
6.7
The most constant of the phosphorus-sediment ratios is that for the
particulate phosphorus fraction. This is estimated by subtracting the
soluble reactive phosphorus from the total phosphorus values. The lowest
value for this ratio is found at the Nevada station, which has the highest
gross erosion rates (Table 7). Nitrate ratios were also calculated although
the transport of nitrates is through soluble rather than particulate forms. "
_ One important aspect of the data on phosphorus-sediment ratios is that
the values show considerable variability from year to year. This variability
is shown in the data of Table ^4. At the Melmore station in 1978, the -P/ss
and PP/SS ratios were 3.76 and 2.52 while in 1979 the ratios were 1.9^ Pnd
T.fcO. Both years had similar total discharges Of water. The higher
phosphorus sediment ratios of 1978 were caused by much lower sediment
concentration in 1978 than in 1979. As described earlier the differencps in
sediment concentration in 1978 and 1979 are attributable to the season*!
variations in runoff for the two years. The annual variability m
nutrient-sediment ratios points out that such ratios cannot be determined on
91
-------
the basis of detailed One year studies. Much care should be exercised in
determining these ratios for watersheds.
For individual samples there is a great amount of variability in both
TP/SS and PP/SS ratios at a given station. This variability is illustrated
for the Melmore station in Figure 20. The TP/SS ratios for individual samples
are plotted as a function of stream flow in Figure 20a. In Figure 20b the
same data is presented with the ratios plotted as a function of suspended
solids concentration. In this form it is evident that higher suspended solids
concentrations are associated with lower phosphorus-sediment ratios. The same
is true for particulate phosphorus ratios (Figure 20c). One possible
explanation for the above would be that higher sediment concentrations may
tend to have larger average particle size distributions. Since larger
particles have relatively smaller surface areas, they may have less phosphorus
per unit weight than finer particles.
In Figure 20 it is also evident that even at particular suspended solids
concentrations, there is considerable variation in phosphorus-sediment ratios.
In examining this ratio at particular sediment concentrations, it appears that
the higher phosphorus-sediment ratios are associated with higher stream flows.
This also could be explained in terms of particle sizes, if, at a given
sediment concentration, higher stream flows are associated with lower average
particule sizes. Since particle size data has not been collected in this
study, the above speculations for the variations in phosphorus sediment ratios
cannot be evaluated. ~ ~-
The variability in nutrient sediment ratios for individual samples points
Out the difficulties that may be encountered when attempting to measure such
ratios based on the collection of a small number of samples. Since control
programs aimed at reducing phosphorus loading from agricultural sources are
based on erosion control programs and such programs may have different degrees
of effectiveness for different particle siz'es, "§ better understanding of
nutrient-sediment ratios in relation to particle size distribution is needed.
Also the relationship between the phosphorus/sediment ratios and the
bioavailability of the phosphorus needs investigation.
92
-------
IP/SS*10GO VS FLOWx.OOl FOR MLM
a
o
"
o
o
.00 .50
TP/SS*1000 VS SS FOR MLM
LOO L50 2,00
FLOW*.Q01 CFS
2.50 3.00 3.50
o
o
o
coo
O
o
o
q
to"
o
o
,00 300.00 600.00 900.00 1200.00 > 1500.00 1800.00 2100.00
O «J
PP/SS»1000 \IS SS FOR HIM
o
o
a
CL
Q-
o
a
* + *
300.00 600.00 900.00 1200^00 1500.00 1800^00 2100^00
SS (MG/L)
Figure 20. Phosphorus/sediment ratios (x 1000) for individual samples
plotted in relation to stream flow (A) and suspended
sediment concentrations (B & C) for Honey Creek at Melmore,
93
-------
SECTION 7
NUTRIENT AND SEDIMENT LOADING AT TRANSPORT STATIONS
In the preceeding section, various patterns of sediment and nutrient
concentrations present in the rivers of northwestern Ohio were described. In
this section material loadings as products of concentrations and their
associated stream flows will be considered.
MEAN ANNUAL LOADS OF NUTRIENTS AND SEDIMENTS
For water quality management planning, information on mean annual
transport of materials at various locations in a watershed is often useful.
Three different methods have been used to calculate mean annual loading using
the data sets available at the transport stations. These methods use either
mean annual flows Or flow duration tables as described below.
1 • Flux Weighted Mean Concentration and Mean Annual
In this procedure, the flux weighted mean concentration of each
parameter at a particular station is multiplied by the mean annual
discharge observed at that station for the entire period of hydrological
record. This, together with appropriate conversion factors, provides
mean annual loading values. A major assumption in the method is that the
flux weighted mean concentration, as based on the sampling period covered
by the study, is representative of the long term flux weighted
concentrations characteristic of that station.
2. Flow Duration Tables and Flux Weighted Mean Concentrations For Each
Flow Interval.
One way to calculate loading using flow duration tables is to
calculate the flux weighted mean concentration for each flow interval, as
described in Section 6. This concentration is multiplied by the mean
flow of the adjacent flow classes that form the flow interval. The
percent of time the flows fell into that interval is then used as the
percent of a year in which flows of that size would occur. For each
interval multiplying the product of the mean flow and the flux weighted
mean concentration by the expected duration of these flows gives a flux
for that interval. The fluxes for each flow interval can then be summed
to provide a mean annual flux. The mean annual flow can also be
calculated and should correspond to the published long term mean flow for
~the station. Table 23 provides an example of the flow duration method of
calculating mean annual fluxes. A. principle advantage of using the flow
duration method is that it adjusts the data to the long term distribution
of flows in the stream. Here the assumption is that the
concentration-flow relationships within the data set are representative
of long term concentration-flow relationships for the station.
94
-------
3- Flow Duration Tables and Average Concentrations for Each Flow Interval
The third method we have used in calculating mean annual loading is
a variation of the method described above. For each flow interval
instead of calculating a flux weighted mean concentration the arithmetic
mean is calculated along with the standard deviation and the standard
error of the mean. Both the mean and the standard error are multiplied
by the mid flow and duration so that a "loading error" estimate can be
determined for each flow interval. The above loading errors for each
interval were then summed to give an error estimate associated with the
mean annual load. Table 24 provides an example of the calculations of
mean annual loading using this procedure.
The mean annual loadings of sediments and nutrients as calculated by the
three methods described above are shown in Table 41. Flux weighted mean
concentrations are shown in Table 27 and mean annual flows in Table 1. Since
the Nevada, Melmore, Wolf West and Wolf East stations have relatively short
periods of hydrological records, mean annual discharges had to be estimated
for these stations. The estimation method included comparing the discharges
observed at nearby long term stations during the period of operation Of the
new gaging stations with the average annual discharge observed at the long
term stations. The mean annual discharge at the new stations was calculated
using the measured discharge at the new station and the ratio of the observed
flow during the same period to the long term flow at the nearby long term
stations. Flow duration tables were not used for the new stations because of
the short duration of hydrological records.
In general the three methods of estimating mean annual loads gave very
similar results. Also the standard error of the estimate was relatively small
considering the annual variability in sediment-and nutrient loadings. The
standard errors as a percentage of the mean were much less for the nutrients
than they were for sediments. The average" leadings Of the three calculated
values were used in subsequent calculations of unit area loads.
UNIT AREA LOADS
In Table 42, the mean annual loads of sediments, phosphorus forms and
nitrate-nitrite nitrogen are expressed On a load per unit area basis. In all
cases the mean annual loads were divided by the total drainage area for the
station to obtain the unit area load. Although the dominant land use in each
watershed is cropland, significant percentages of other land uses do Occur in
each watershed (see Table 2). The phosphorus loadings were not corrected for
upstream point source inputs. Such corrections are described in Section 8.
The unit area loads for phosphorus and nitrogen observed in the study
watersheds represent rather high values in comparison with other areas. As
part of the PLUARG study, data on unit area phosphorus loads were compiled
from the PLUARG watersheds as well as from Other studies in the Great Lakes
region and elsewhere (Johnson, et al., 1978). A summary from that review
relating unit area phosphorus yields to land use and land forms is reproduced
in Table 43- The unit area yields at Our river transport stations are similar
to those Of plowed fields On fine textured soils.
95
-------
Table 4L. Mean Annual Sediment and Nutrient Loading at Transport Stations
Station Calculation
Method*
Maumee
Portaqo
Tindall
Huron
Ducyrus
Upf>cr Sandusky-
Tymochtee
Mexico
Nevada
Melmore
Wolf West
Wolf East
1
2
3
Avc>.
I
2
3
Av« .
1
2
3
Avp .
1
2
3
Ave .
1
2
3
Ave.
1
2
3
Ave.
1
2
3
Ave .
1
2
3
Avo .
1
1
1
1
Suspended
Solids
10 m. tons/yr.
1,043
981
1,087 + 10.3%
1,037
45.6
41.3
47.3 + 18.5%
44.7
185
156
164 + 11.9%
168
58.9
53.4
57.9 + 18%
56.7
11.1
13.0
13.3 + 13.4%
12.5
50.8
45.2
50.7 + 11.8%
48.9
31.3
29.8
31.9 + 13.2%
31.0
123.
124
141 + 18.8%
129
18.9
18.9
8.18
12.9
Total
Phosphorus
m. tons/yr.
2,224
2,131
2,237 + 6.1%
2,200
111.7
103.4
108.4 + 7.2%
108
386.4
335.5
343.8 + 6.4%
355
97.0
97.8
100.5 + 10.3%
98.4
36.9
45.2
45.6 + 7.5%
42.6
111.9
109.5
113.5 + 6.3%
112
64.0
61.6.
63.2 + "7.5%
62.9
221.
232
250 + 10.6%
234
31.1
42.2
17.6
29.8
Dissolved
Ortho Phosphorus
m. tons/yr.
500.1
491.0
479.0 + 5.1%
490
33.1
33.3
34.6 + 6.2%
33.7
79.3
75.1
75.1 + 6.1%
76.5
27.9
2R.1
27.2 + 8.5%
27.7
14.1
18.6
18.4 + 7.1%
17.0
29.0
30.6
30.5 + 5.5%
30.0
10.3
10.4
-f 10.5 + 5.5%
10.5
36.1
47.1
43.9 + 7.4%
42.4
4.71
10.56
4.46
8.42
Nitrate-
Nitrite-N
m. tons/yr.
21,200
21,500
22,000 +
21,600
1 , f,40
3 ,690
1,710 +
1,680
3,933
4,164
4,050
967
971
1,010 +
983
259
261
277 +
266
842
883
914 +
880
782
795
814 +
797
1,804
2,688
2,853 +
2,450
376
507
274
336
5.3%
6.2%
8.f>».
8.3%
4.5%
6.0%
7.3%
•Calculation Methods: 1. Flux weighted mean concentration times mean annual flow; 2. Flow duration
table and flux weighted mean concentration per flow interval; 3. Flow duration
table and average concentration per interval plus or minus standard error of
the estimate.
96
-------
agricultural waters)
Haginq
Station
Maumec
Portage
Huron
Frr'mont
Mexico
Upper Sandusky
Bucyrus
Crawford
Honey
Nevada
Wolf, Ea;;t
Wolf, West
Suspended
Area Sediment
Km Tons/ha/yr
16,395
1 ,109
061
3,240
2,005
772
230
593
386
217
213
171.5
.63
.40
.59
.52
.65
.63
.54
.52
.49
.87
.61
.48
Total
Phosph
kg/ha/yr
1
1
1
1
1
1
1
1
1
1
1
.34
.97
.02
.10
.17
.45
.85
.06
.09
.43
.40
.02
Soluble
Reactive
Phosph
kg/ha/yr
.30
.30
.29
.24
.21
.39
.74
.18
.27
.22
.39
.26
'•>
Nitrogen
kq/ha/yr
13.
15.
10.
12.
12.
11.
11.
13.
13.
• 17.
15.
16.
1
1
2
5
2
4
5
4
1
3
a
0
(kg/ha/yr)
Table 43. Total Phosphorus Unit Load/by Land Use and Land Form (in U.S.A.)*
Form
Use
Plowed fields
Grassland
Dairy (pasture)
Brush
Orchard/
truck crops
Forest
Fine Textured
Level Sloping
1.06
.23
.40
.23
1.25
.10
1.25
.23
.63
.23
1.25
.10
Medium
Level
.87
.10
.23
.23
1.25.
.10
Textured
Sloping
.07
.10
.23
.23
1,25
r
.10
Coarse
Level
.23
.10
.10
.23
1.25
.10
Textured
Sloping
.63
.10
.10
.23
1.25
.10
*From J^lmson, M. G. , et al.. Management Information Base and Overview Modelling,
1978.
97
-------
Uttormark, et al. (1974) reviewed 19 studies on nutrient transport from
agricultural lands by streamflow and found the following ranges and means:
Total-N Total-P
kg/ha/yr kg/ha/yr
max 13.0 2.3
min 1.2 0.03
ave 5.1 .38
For seven of the northwestern Ohio watersheds (Table 42) the export of nitrate
nitrogen exceeded the maximum total nitrogen export listed in the studies
reviewed by Uttormark. Also the unit area phosphorus yields in northwestern
Ohio are three times higher than the mean loads from the studies reviewed by
Uttermark.
In a nationwide survey of the relationships between nonpoint sources and
stream nutrient levels, Omernik (1977) noted the following mean values for
phosphorus and nitrogen export:
Ortho- Total Inorganic Total
phosphorus Phosphorus Nitrogen Nitrogen
kg/ha/yr kg/ha/yr kg/ha/yr kg/ha/yr
>_ 75^ Agriculture 0.094 0.255 3.26 5.54
> 90$ Agriculture 0.118 0.266 7.81 9-54
The total phosphorus export rates for th,e northwestern Ohio watersheds
(Table 42) are about four times higher than the mean" values in the nationwide
survey. Likewise, the orthophosphorus and inorganic nitrogen export rates are
two - three times higher than the mean value in the nationwide survey.
ANNUAL VARIATIONS IN NUTRIENT AND SEDIMENT LOADING
One of the objectives of this study was to determine the extent of annual
variations in nutrient and sediment export from the study watersheds. A
preliminary estimate of the annual loading for each water year can be obtained
by using the flux summary programs described in Section 5 and -selecting for
the water year dates. However, in any year there is generally a significant
amount of missing data as noted in Table 25.
Erroneous or missing stage data and missing chemical analysis data
generally result in the need to correct the preliminary loading calculations.
These corrections are made as follows:
98
-------
1) Erroneous Stage Data
The formation of ice jams Occasionally results in high stage readings
which are not associated with high flows. The hourly stage data from the
USGS preliminary reports are not corrected for ice jam effects, and thus
when these stages are entered into the chemical data sets, the associated
flows are too large. The final flow data as presented in the Water
Resources Data for Ohio for each water year presents corrected mean daily
flows and both monthly and yealy total flows.
A computer program is used to compare the total flows, summed from
the chemical data sets for each month, with the U.S. Geological Survey's
reported monthly flows. Table 44 shows a sample printout from the
program. The ratio of the USGS flow values to the calculated flow from
the chemical data sets is listed along with the number of samples
collected, the total hours monitored and the weighted mean concentration.
For any month when the ratio is less than one, the flow data from the
chemical data set is compared with the mean daily flows from the USGS
records and corrections are made in our data archives to bring the flows
into agreement with the USGS values.
2) Missing Stage Data
Malfunctions in the stage measuring and recording equipment
occasionally Occur and result in gaps in the stage data. The automatic
samplers continue Operation, the samples are analyzed and the results
entered into the computer. Thus concentration data is available but flow
data at the specific times of sample collection is not available. In such
cases the monthly summary tables show higTTer USGS flows than the
calculated flows.
•* -*—
The USGS estimates daily flows during these gaps by comparison with
nearby gaging stations. These estimates are published in the Water
Resources Data and are indistinguishable from the mean daily flows
calculated from hourly stage data. In Order to calculate fluxes for these
time periods, the mean daily flows from the Water Resources Data are
entered into Our archive data set and used in conjunction with the
measured chemical concentrations.
3) Missing Chemical Data
Gaps in the chemical records are generaly caused by pump failures at
the gaging stations or by problems with the automatic samplers. In these
cases, the archive sets contain neither stage (and flow) data nor chemical
data. If the period of missing data includes significant flows, then the
monthly summaries again show higher USGS flows than the calculated flows
associated with the chemical data set. In this case corrections are made
using the mean daily flow from the USGS records and seasonal
concentration/flow relationship for each station and parameter as
described in Section 6.
99
-------
Table 44. Sample printout of monthly flux and flow summaries used for identifying
and correcting erroneous or missing flow and concentration data.
Thdi JUL ?4 1960
STATICN:UPPEF. .SANDUSKY.MVER
PARAPETERITP
BEGINNING DATETI HE:751Ot1063P
ENCII.G DATETI^E :760930190C
»ATEK DUALITY LAfi
HEIDELBERG COLLEGE
MONTH
OCT.
NOV.
DEC.
O JAN.
O
FEB.
MAR.
APR.
HAY
JUNE
JULY
AUG.
SEP.
N
55.00
12.00
23.00
38.00
59. OC
61.00
33. CO
28.00
89.00
127.0
95.00
33.00
653.0
WT MEAN
MG/L
.3106
.3215
.3199
.3232
.7022
.5929
.1641
.5873
.7330
.9145
,55b5
.4264
.5695
CUM TIME DBS FLCU
MRS W**3
732.2
246.1
375.5
310.2
634.0
750.0
696.0
585.0
723.0
741.0
7?3.0
573.0
7119.
.5236E
.8848E
.6039E
.3970E
.7057E
.3965E
.8057E
.3982E,
.1115E
.1350E
.4722E
.1091E
.29ffcE
07
06
07
08
08
08
C7
07
08
OB
07
07
09
uses FLCU
.4784E
.2914E
.2145E
.4572E
.6747E
.3676E
.8122E
.4720E
.1031E
.1245E
.4600E
.1171E
.?2C':E
07
07
08
08
08
08
07
07
08
08
07
07
ON-
FLOW RATIO CBS FLUX CALC FLUX
METRIC TONS METRIC TGMS
.9137
3.294
2.b68
1.152
.9271
1.008
1.185
.9252
.9224
.9743
1.074
1.626
.2845
2.572
12.8?
49.55
23.51
1.322
2.339
8.170
12.35
2.b2J
.4652
117.6
1
.
t,
.466
9370
.862
14. 7P
47.38
2
i
2
7
1
i
1
1 .60
.33J
.772
.559
1.39
.556
4995
19.3
-------
Table 45. Annual variability in aadlatant and nutrlant export few Mlactad northmatarn Ohio rinra.
Ha tar
Station tear
Fortaga 1975
1976
1977
1978
Preannt 1975
1976
1977
1978
1979
TymochtM 1975
1976
1977
1978
197*
MalJkon 1976*
1977
1978
1979'
•Tabruary through 8»
uses
Dlachar^a
H3
,.7
28.
32.
24.
44.
99
62
13
38
103.0
77.
62.
139.
108.
17.
7.
6.
20.
14.
6.
7.
14.
15.
pfaamji
16
90
1
a
71
64
60
73
42
9i _..
29
78
02
>r only.
8u*p« ivied
Solid*
Cono. Load
•g/1 tonnaa
213
155
161
131
293
161
146
119
261
2*2
148
126
64
1*2
274
105
74
207
61,740
50,550
38,930
58,340
302,200
124,000
91,560
193,500
285,500
50,010
11,120
a, 142
13,180
27,720
19,090
7,677
ll.OOO
41,115
Total
Phoaphorua
Cone. Load
•g/1 tonnaa
.191
.381
.410
.367
.407
.516
.409
.313
.518
.526
.278
.134
.207
.411
.5*8
.335
.276
.5M
111.1
124.1
98.8
162.8
418.9
198.9
257.5
461.4
561.4
91.1
21.1
22.1
42.*
59.1
4O.49
24.41
4O.81
U.OO
Soluble
Reactive
Vhoaphorua
Cone. Load
•9/1 tonivaa
.128
.114
.137
.107
.068
.064
.104
.084
.103
.0*1
.079
.061
.072
.080
.115
.090
.093
37.29
37.12
33.01
47.51
70.42
49.67
65.76
116.9
112.2
10.87
5.20
12.96
10.14
5.51
8.33
11.15
11.96
Nltrata-
Nitrita
Nitrogon
Cone. Load
•7/1 tonnaa
7.45
4.20
7.94
4.96
4.57
1.19
4.M
1.56
5.16
5.49
4.17
6.16
3.10
5.88
4.56
6.78
4.14
5.93
2i5a
1371
1916
2200
4709
2621
3115
4958
5615
972.6
134.1
406.8
641.7
848.5
115.1
494.8
611.1
8*0.5
TJ>/8»
9/V9.
1.81
2.46
2.54
2.79
1.39
3.22
2.81
2.19
1.97
1.86
i.ea
2.65
3.25
2.11
2.11
3.18
1.71
2.04
« fh B;*hod8/;r correcting annual loading estimates have been
o J i Ji T 8tatl0na sho™ in Table 45. The Table includes a listing
the ^ ^ ^ USGS "^ year' and the estimated lead and
.„,«? £.-£££ ?:r:a rr^r'*1-"in °°dim°"t -nd -utrient
1 .
""
6
•
ty larg. annual variations
'-
101
-------
Table 46. Ratio of high annual yield to low annual yield.
Station Years SS TP SRP Nitrate-N
Portage
Fremont
Tymochtee
Melmore
4
5
5
3
1.
3.
5.
5.
58
30
99
613
1.
2.
4.
3.
65
19
38
61
1.
2.
2.
1.
42
35
49
68
1
2
2
1
.60
.14
.91
.80
2. Annual sediment yields are not directly proportional to annual discharge.
For example at Fremont the 1978 discharge was 23% higher than the 19^9
discharge but the sediment export in 197B was 72% lower than the 1979
sediment export. The flux weighted mean concentration of sediment in 1978
was 139 rng/1 while in 1979 it was 263 mg/1. Large variations in annual
flux weighted mean concentrations occur at each station and the values do
not correlate with flow. At Tymochtee the highest annual discharge (1978)
had the lowest average sediment concentration ("64 mg/l) while the second
highest annual discharge (1975) had _the highest average sediment
concentration (282 mg/l). As discussed in 'the previous section the major
factor affecting this variability is probably the proportion of the annual
runoff that occurs during winter months.
3. Annual yields of total and particulate phosphorus are not directly
proportional to annual yields of suspended solids. This is reflected in
the rather wide variations in the ratios of total and particulate
phosphorus yields to suspended solids yields (Table 45). As discussed in
the previous section, the phosphorus to sediment ratios decrease with
increasing sediment concentration. This is illustrated in Figure 21 where
the TP/SS ratios of the annual yields are plotted as a function of annual
flux weighted mean concentrations of suspended solids.
FLUX EXCEEDENCY RELATIONSHIPS
The range in fluxes of substances at a particular station greatly exceeds
the range in concentrations. One way to characterize the range in fluxes is
to calculate the percentage of time a given flux is exceeded. The computer
programs for calculating flux exceedency are described in Section 5.
102
-------
3.8
3.6
3. H
3.2
3.0
2.8
I2'6
"a. 2.4
^00
"3 2.2
Q<
H 2.0
1.8
1.6
1.4
1.2
1.0
0
LEGEND
O A FREMONT
D MELMORE
* PORTAGE
O TYMOCIITEE
'
-
* *
0
*
*
A
o a
D
A
° * °
-
A
-
1 1
100 200
SUSPENDED SOLIDS («a/l)
300
Figure 21.
Annual variations in the ratio' of total phosphorus
export to suspended sedimen,t export in relation
to annual flux weighted suspended solids concentra-
tion.
Table 47 provides an example of a flux exceedency table for the 'Fremont
Station on the Sandusky River. The time exceedencies are based on
approximately 30,000 hours of sampling data at the station during which time
2150 discreet samples were collected and analyzed. The wide range of fluxes
is evident in the data.
The importance of the high flux and high flow values in terms of total
transport at the sampling stations is shown in Table 4-8. The information in
Table 48 was summarized from both flux exceedency and flow exceedency tables
in which the cumulative percent of the total flux was also listed (see Table
20 and 21). The data in Table 48 indicate, for example, that for the ""aumee
River at ¥aterville, the flows exceeded 1 *, of the time accounted for 11.7* of
the total discharge, 13.6$ of the suspended solids flux, 12.9* of the TP flux,
10,0 of the SRP flux and 7.6£ of the dissolved solid flux. The suspended
103
-------
Table 47. Flux exceedency values for suspended solids, total
phosphorus and soluble reactive phosphorus at the
Fremont gaging station.
Percent
Exceedency
99%
98
95
90
80
70
60
50
40
30
20
10
5
2
1
0.5
0.1
Suspended
Solids
kg/hr
20.9
35.2
49.6
78.4
134.9
235.5
410.5
814.8
1,596
3,650
10,220
~~ 42,970
125,600
320,700
496,700
914,400
1,341,000
Total
Phosphorus
kg/hr
.285
.355
.467
.709
1.257
1.973
2.844
3.991
6.119
12.78
30.99
93.61
276.9
653.1
991.4 -
1515,
2140
Soluble
Reactive P
kg/hr
.008
.022
.082
.158
.337
.675
1.109
1.764
2.818
4.714
9.939
28.65
53.68
111.2
148.6
244.0
389.0
solids fluxes exceeded 1^ of the time accounted for 30.8$ of the suspended
solids flux, the TP fluxes exceeded \% of the time accounted for 20.^ of the
TP flux, the SEP fluxes exceeded *\% of the time accounted for 13.1^ of the S?^>
flux and the "conductivity" fluxes exceeded \% Of the time accounted for 8.7^
of the total dissolved solids flux.
The differences between the percentages of the total fluxes accounted for
by the flows and fluxes exceeded 1$ of the time are greatest for suspended
solids and least for total dissolved solids. It is evident at all of the
stations that, for suspended solids, peak flows cannot account for the peak
fluxes.
104
-------
Table 48. Percent of total flux accounted for by fluxes associated with flows and fluxes exceeded fixed
percentages of time.
o
en
Percent
Exceedency
(Time) Discharge
1%
2%
5%
10%
20%
30% '
L
1% '
2%
5%
10%
20%
30%
1%
2%
5%
10%
20%
30%
11.
20.
39.
55.
72.
81.
15.
27.
47.
65.
80.
87.
16.
26.
44.
61.
77.
85.
Maumee
7%
6%
2%
6%
0%
7%
Tindall
4%
3%
7%
5%
7%
7%
Melmore
2%
3%
4%
3%
9%
6%
Percent of total monitored flux during 1974-79 water years
SS TP SRP TDS*
Flow Flux Flow Flux Flow Flux Flow Flux
13.6%
27.8%
61.5%
80.6%
90.6%
94.5%
18.9%
40.8%
66.3%
84.5%
93.2%
96.2%
35.2%
52.8%
66.4%
81.2%
92.3%
96.8%
30
46
70
83
92
95
35
50
73
87
95
98
51
62
78
88
96
98
.8%
.1%
.0%
.1%
.2%
.7%
.9%
.9%
.8%
.4%.
.8%-
.2%
i
.4%
.8%
.5%
.9%
.1%
.4%
12.9%
24.8%
52.6%
71.8%
84.3%
90.0%
21.0%
38.8%
63.6%
80.2%
90.8%
94.9%
29.5%
44.1%
61.0%
77.2%
89.6%
94.3%
20.9%
33.3%
56.9%
72.8%
85.0%
90.7%
28.3%
43.2%
66.4%
81.9%
92.0%
95.8%
36.0%
46.9%
65.1%
79.8%
91.1%
95.2%
10.0%
17.8%
37.8%
54.4%
71.5%
81.4%
20.8%
31.3%
52.1%
68.1%
84.1%
90.8%
15.7%
27.3%
45.9%
62.9%
79.7%
86.8%
13.1%
21.9%
39.4%
55.5%
72.9%
82.7%
22.8%
34.5%
54.5%
70.9%
85.9%
92.0%
19.9%
30.6%
49.9%
66.4%
81.9%
88.5%
7.6
14.3
29.9
45.1
63.5
75.3
9.6
17.2
34.0
51.8
70.3
80.1
8.6
14.6
29.7
46.6
65.9
76.5
8.7%
15.4%
30.3%
45.9%
64.1%
76.0%
10.0%
17.9%
34.9%
52.3%
70.8%
80.6%
9.4%
15.9%
30.4%
47.1%
66.3%
76.8%
*TDS (total dissolved solids) flux is based on conductivity measurements.
-------
As the percent exceedency increases, the differences in percentage of
total transport associated with flow and flux exceedencies decrease. At 30$
exceedencies, the percentage of the total fluxes for both flow and flux
exceedency are about the same.
The three stations summarized in Table 48 have greatly differing drainage
areas. The Maumee, Fremont and Melmore station have watersheds of 16,395,
3,240 and 386 sq km respectively. As the size of the drainage area decreases
the flux and flow exceedency patterns appear to change in a systematic way.
With respect to sediment transport, flows exceeded \% of the time accounted
for 13.6$ of the total transport at the Maumee station, 18.9"? at "Fremont and
35.2$ at Melmore. Suspended solids fluxes exceeded 1$ of the time accounted
for 30.8$ of the sediment transport at the Mauraee station, 35-9$ at Fremont
and 51-4% at Melmore. Parallel patterns were present for total phosphorus
transport. The transport of total dissolved solids did not show this
characteristic. These data do emphasize that as watersheds become smaller,
greater proportions of the transport of sediment and sediment related
pollutants Occur in small percentages of time.
It should be noted that the time exceedency data presented above do not
represent contiguous time intervals. Rather the data reflect the cumulative
role of peak flows or peak fluxes taken from a number of large runoff events.
The data nevertheless underscore the role of peak transport events in total
material transport. In all three stations the highest 10$ of the time with
respect to either flows or fluxes accounted for more than 80^ of the total
sediment transport, more than 70% of the total phosphorus transport and more
than 54$ of the^ soluble reactive phosphorus transport.
106
-------
SECTION 8
WATER QUALITY MANAGEMENT IMPLICATIONS
The river transport data described in the proceeding sections are useful
in addressing a number of problems in the area of water quality management
planning. These problems include: 1) the assessment of the relative costs
and environmental effectiveness of point and nonpoint phosphorus control
programs; and 2) the identification of critical areas for nonpoint control
programs.
POINT AND NONPOINT SOURCE COMPONENTS OF STREAM PHOSPHORUS TRANSPORT
The mean annual nutrient loads calculated for the transport stations and
summarized in Table 41 include the effects of both point source and nonpoint
source inputs. A standard procedure for calculating the nonpoint source
components of the transport is to subtract upstream point source inputs from
the total stream transport (Baker and Kramer, 1973; SonzOgni; et al. , 1978;
COE, 1975a). Usually it is assumed that all Of the point source inputs are
transported through the stream system, although large portions of these inputs
may be stored in temporary sinks On the stream bottom. By assuming 100$
transmission Of point source inputs and subtracting this value from the total
stream transport, the resulting value for nonpoint components represents a
minimum value. If the transmission of point source inputs is less than 100$,
then smaller point source components would be subtracted from the total load
and consequently the nonpoint components would be larger.
Point source loading estimates for the study watersheds were taken from
sewage treatment plant records where these were available and estimated in
other cases. Plants with flows both greater than and less than 3770 m3/day d
million gallons per day) were included • i-n . the point source summaries.
Phosphorus removal requirements apply primarily to^plants with flows greater
than 3770 m3/day. Point source phosphorus loadings have been summarized in a
number of recent publications (COE, 1975b; DePinto, et al., 1979; UC, 19^9;
IJC, 1980). Point sources of phosphorus in the Sandusky Basin are listed in
Table 49.
Calculations of nonpoint phosphorus loading for the study watersheds are
shown in Table 50. Table 50 also includes data which suggests that point
source inputs do not have 100$ transmission through the stream system. When
the calculated nonpoint yields are divided by the total watershed area to
obtain unit area nonpoint yields, the watersheds with the highest percentage
of point source (Bucyrus, Portage and Huron) have the lowest unit area
nonpoint phosphorus yields. These same watersheds, however, have unit area
sediment yields comparable to adjacent watersheds lacking point source inputs.
Consequently the nonpoint phosphorus to sediment ratios are much lower for the
watersheds with higher percentages Of point sources than for adjacent
watersheds with similar unit area sediment loads and lacking point source
inputs. There is no reason to expect nonpoint phosphorus to sediment ratios
to be lower in watersheds containing point sources than in adjacent watersheds
lacking point sources. The discrepancies could be resolved if the
transmission of the point source phosphorus inputs were less than 100$. Even
if point source inputs have 100$ transmission through the stream systems they
would account for Only 16.3$ percent of the total loads observed at the river
107
-------
Table 49. Indirect municipal point source phosphorus
discharges in the Sandusky River Basin 1978.
Location
10
Crestline
Bucyrus
Upper Sandusky
Gary
Attica
Bloomville
Tiffin
Total
Loading
, , Flow Total Phos. Rate
J MJ/day
2.1
7.7
6.4
2.3
.87
1.1
13.
33
(MGD)
(-55)
(1.9)
(1.7)
(.6)
(.23)
(.28)
(3.5)
(8.8)
mg/1
6*
8.0
2.5
6*
6*
4*
0.9
Kg/hr.
.52
2.6
.67
.57
.22
.18
.49
5.3
Annual Load
M Ton/yr.
4.6
22
5.8
5.0
1.9
1.6
4.3
45
*Estimated.
Table 50. Minimum nonpoint source phosphorus yields for the study watersheds.
Total
Phosphorus
Transport
Watershed. kg/ha/yr.
Maumee
Portage
Huron
Sandusky Basin
Fremont
Mexico
Upper
Sandusky
Bucyrus
Nevada
Tymochtee
Melmore
Wolf, East
Wolf, West
2200
108
, 98.4
355
234
112
42.6
31.1
62.9
42.2
29.8
17.6
Point
Source
Input
kg/ha/yr.
321
40
44
45
37
32
27
3.5
Minimum*
Non-Pt.
Yield
kg/ha/yr.
1879
68
54.4
310
197
80
15.6
31.1
39.0
29.8
17.6
Unit Area
Non-Pt.
Yield
kg/ha/yr.
1.14
.61
.57
.96
.98
1.04
.68
1.43
1.06
1.01
1.40
1.03
Suspended
Solids
Yield
MT/ha/yr.
.63
.40
.59
.52 ~
.65
.63
.54
.87
.54
.49
.61
.48
Non-Pt.
TP/SS
Ratio
gAg
1.81
1.52
.97
1.85
1.51
<•
1.65
1.25
1.64
1.96
2.06
2.29
2.13
Point Source
% of Total
Phosphorus
Yield*
14.6%
37.0*
44.7*
12.7%
15.8*
28.5*
63.3%
0*
0%
8.3*
0*
0*
Watershed
Population
Density
#/km
31
26
46
96
16
8
18
29
29
mouth stations (Maumee, Portage, Huron and Sandusky at Fremont.)
Septic tank systems are another potential source Of phosphorus in the
stream system. Within the Sandusky Basin upstream from Fremont the total
population is about 99,600 (preliminary 1980 census data). Approximately <50^
Of the population is served by centralized sewage collection systems and
treatment plants while the other 50$ is served by septic tanks. Although some
septic tank effluents containing phosphorus enter the stream systems, studies
conducted below the town of Tyro in the Honey Creek watershed indicated chat
stream loading associated with septic tank inputs was quite small (Kneger, et
al., 1980).
108
-------
The dominance of agricultural land use within the study watersheds
suggests that the bulk of this nonpoint phosphorus loading is derived from
rural, rather than urban land uses. The unit area nonpoint phosphorus loading
rates are actualy higher in watersheds which lack urban areas (Nevada,
Tymochtee, Melmore, ¥olf East and ¥olf West) than for watersheds containing
urban areas. For the Lake Erie Basin as a whole the PLUARG studies indicated
that 61% of the tributary phosphorus loads were derived from cropland, 5% from
pasture, 21$ from urban and 13$ from other land uses (PLUARG, 1978). The
ratios Of urban land to cropland wOuld be much lower in the study watersheds
than the average ratio for Lake Erie.
STREAM PROCESSING OF POINT SOURCE PHOSPHORUS INPUTS
The phosphorus concentration profiles presented in Figures 18 and 19
illustrate that phosphorus entering stream systems rapidly disappears from the
flowing water. Biological uptake by the periphytic and benthic communities,
adsorption onto sediments and chemical precipitation could all be involved in
the removal of total phosphorus from the flowing water. This processing Of
phosphorus by stream systems has been noted by numerous authors (Keup, 1968;
Thomann, 1972; Verhoff, et al., 1978; Verhoff and Baker, 1980).
The extent of phosphorus deposition below point sources in the Sandusky
Basin has been analyzed using both concentration data and flux data (Baker,
1980). The combined point source loading rate in the Sandusky Basin upstream
from Fremont is about 5.3 kg/hr (see Table 49). The phosphorus flux
exceedency tables at Fremont (see summary in Table 48) indicate that 55% of
the time the flux is less than 5.3 kg/hr and the cumulative flux during that
55% of the time accounted for only 2.14$ of the total phosphorus flux observed
at the Fremont station.
The computer printouts for flux exceedency include both the cumulative
hours and the cumulative flux (see Table 21 ). In this case, 55% of the time
amounted to 16,806.4 hours and during this time the stream transported
30,043.7 kg of P. Thus the average phosphorus flux during the 55$ Of the time
with the lowest flux was 1 .79 kg/hr. This flux includes both phosphorus from
point sources and "background" phosphorus following the hydrolOgical pathways
Of water under these flow conditions.
At Fremont, 55$ of the time the flows were less than 326 CFS and flows
during this 55$ of the time accounted for 2.41$ of the total observed
phosphorus flux. Thus at these exceedency levels, flux and flow rankings are
very similar in terms Of their accompanying percent Of total transport. The
average flow during the 55$ of the time with the lowest flows was 145 CFS.
Since the combined flow from all the sewage treatment plants is only about 14
CFS it is clear that the bulk of the flow (131 CFS) during this portion of the
time is derived from hydrolOgical pathways such as ground water, tile effluent
and surface water rather than point sources.
In the watersheds lacking point sources, the total phosphorus
concentrations averaged about 0.12 mg/1 during the 55$ of the time with the
lowest flows. Assuming that this concentration characterized the nonpoint or
"background" concentration of the stream flow, the average nonpoint loading
109
-------
rate would be about 1.6 kg/hr (131 CFS X 0.12 mg/1 X 0.10188). Since at
Fremont, the average total loading during this time period is 1 .8 kg/hr, it is
clear that most of this phosphorus can be attributed to nonpoint rather than
point source inputs. With point source phosphorus input rates of 5.5 kg/hr,
and with only about 0.2 kg/hr of the average Output for 55% of the time
allocable to point sources, it is also clear that these streams are capable of
removing most of the point source phosphorus inputs from the flowing water.
The 55% time period in the above calculations was selected because in the
flux exceedency tables 55% of the time the fluxes at Fremont were below the
point source input rate of 5-5 kg/hr. Since much of that phosphorus loading
would be due to nonpoint sources, deposition of point source phosphorus must
also be occurring at higher flows and fluxes.
Municipal point sources which load phosphorus into tributaries upstream
from gaging stations nearest Lake Erie are considered indirect point sources
with respect phosphorus loading (IJC, 1979). Direct point sources are those
whose- Outfalls enter streams below the gaging station nearest the lake, or
enter estuaries or the nearshore zone of the lake. The effectiveness of
phosphorus control programs at indirect point sources in reducing
eutrophication in Lake Erie depends On the extent of resuspension and eventual
delivery of this phosphorus to the lake and on its bioavailability upon
reaching the lake. The significance of information on transmission and
bioavailability in phosphorus management is illustrated in the Watershed Model
developed by the Great Lakes Basin Commission (SonzOgni, et al., 1980;
Montieth, et al., 1980).
Analyses of hydrograph patterns at Sandusky gaging stations, using wave
and water routing techniques, indicate that sediment resuspension is an
important component of sediment transport in this basin (see Section 9). It
is usually assumed that phosphorus derived.from point sources becomes part of
the particulate phosphorus which is resuspended'and iffbved downstream by the
passage of wave fronts associated with storm flows (Verhoff, et al., 1978).
The question of the transmission Of point source phosphorus to the lake
becomes a question of the instream transmission or delivery ratio of
particulates to the lake. Reservoir sedimentation Or flood plain deposition
could transfer point-source-derived phosphorus into long term sinks and
consequently reduce the transmission of this phosphorus below 100 .•>. Tt is
also possible that phosphorus could be removed from stream systems through
food chains Or other biological means.
As noted earlier, indirect evidence associated with the calculation of
nonpoint phosphorus loads and phosphorus-sediment ratios suggests that the
transmissions of point source phosphorus through stream systems is less than
_J_00$. Precise and direct measurements of point source phosphorus transmission
factors will, however, be very difficult to obtain because of the large annual
variations in phosphorus transport and phosphorus/sediment ratios. It is also
probable that the "transit time" of phosphorus movement through the stream
will be variable. The best opportunities for measuring transmission factors
would be in locations where point source phosphorus comprises a large portion
of the phosphorus output and where long term yield measurements are conducted
both prior to and following a substantial reduction in point source inputs.
In the Sandusky Basin the best opportunity for such studies would be at the
Upper Sandusky gaging station following implementation of a phosphorus removal
110
-------
program at Bucyrus, Ohio.
¥ith respect to bioavailability, it is probable that the pOint-sOurce-
derived phosphorus that does reach the lake is largely incorporated into the
particulate fraction. ¥ithin the phosphorus transport model developed by the
U.S. Army Corps of Engineers in the LF,¥MS study, there is no evidence Of a
release of soluble phosphorus from sediments during the passage of storm
events (Yaksich & Adams, 1980). Since point-source-derived phosphorus is
largely bioavailable upon entry into a stream, it is also considered largely
bioavailable upon subsequent delivery to receiving waters (Monteith, et al.,
1980). Data from the Sandusky do not support this conclusion.
Considerable attention is now being directed to the bioavailability of
particulate phosphorus (Sonzogni et al., 1981; Logan 197Sb; Armstrong, et
al., 1979; Lee et al., 1980). These studies suggest that TTaOH-extractible
phosphorus provides a good measure of bioavailable particulate phosphorus. In
Table 51, data from Logan (l978b) on the NaOH-extractible -P from the study
watersheds is presented. The NaOH-extractible fraction comprises, On the
average, 36.4$ of the persulfate-digestable particulate -phosphorus.
Persulfate rather than percloric acid digestible particulate phosphorus was
selected from Logan's data, since the total phosphorus procedure used in Our
lab involves persulfate digestion. The data in Table 51 indicate that there
is considerable variability in the proportion Of bioavailable particulate
phosphorus at a given station. The average value of 36.4$ does, however,
agree very closely with the 35$ bioavailable portion reported for the Maumee
River by Armstrong, et al., (1979).
The average percent availability is slightly higher for strictly
agricultural watersheds such as ¥olf, ¥est and Broken SwOrd than it is for the
Other stations which include much larger point source inputs in their
watersheds (see Table 50). This would not be expected if point-source-derived
phosphorus became largely incorporated into ' the ^bioavailable particulate
fraction. At the Fremont station, the mean annual particulate phosphorus
loading is 278.5 metric tons per year. Assuming 35$ of this is bioavailable,
the bioavailable particulate phosphorus loading would be 97.5 metric tons per
year. If all of the point source phosphorus inputs upstream from Fremont (45
metric tons per year) were exported as bioavailable particulate phosphorus,
the point source inputs would make up 46$ of the total export of bioavailable
phosphorus from the basin. It would then be expected that the percent
bioavailability would be much greater along the mainstream of the river below
point sources than from watersheds lacking Or with very small point source
inputs. Data presented in Table 51 do not support this. Instead the data
suggest that, upon delivery to the lake, either the bioavailability of
point-source -derived phosphorus is no greater than the bioavailability of
nOnpoint source particulate phosphorus Or that the transmission of
point-source-derived phosphorus is much less than 1i
It should also be noted that most of the soluble reactive phosphorus
delivered to the lake from tributaries is derived from nonpOint sources. The
flux exceedency tables for the Fremont gage indicate that 70$ of the soluble
phosphorus flux is delivered in 10$ of the time. On an annual basis this
corresponds to 53.5 metric tons of soluble reactive phosphorus during 10$ of
the year. Since point source inputs are relative constant year round, they
could account for only 4.5 metric tons during the time when the total export
111
-------
Table 5]_. Analyses of NaOH-Extractable P from suspended solids collected
during runoff events of study watersheds. (after Logan, 1978b).
Station
Waterville
Portage
Fremont
Mexico
Upper Sand.
Bucyrus
Honey Creek
Broken Sword
Wolf, West
Wolf, East
Huron
Suspended
Date Solids
mg/1
3/10/77
4/26/77
3/28/77
7/5/77
3/28/77
7/2/77
7/3/77
3/28/77
3/28/77
3/28/77
3/21/77
3/28/77
4/5/77
7/1/77
7/1/77
7/2/77
7/1/77
3/28/77
7/5/77
143
248
160
318
376
278
730
340
288
224
120
160
114
1184
1204
194
370
198
1008
Persulfate-Total NaOH-Ext.
Phosphorus Phosphorus
ug/g ug/g
1168.2
1197.9
1267.2
1049.4
1019.7
1297.0
970.2
1029.6
900.9
702.9
940.5
663.3
1194.0
861.3
910.8
861.3
1148.4
871.2
792
426
309.9
442.2
276.4
480.8
395.2
290.6
359.5
317.3
261.5
434.9
423.9
388.3
354.1
332.3
259.2
311.5
283.6
267.8
Percent
NaOH-Ext.
Phosphorus
36.4%
25.8%
34.9%
34.9%
47.1%
30.1%
30.0%
34.9%
35.2%
37.2%
46.2%
64.9%
32.5%
41.1%
36.4%
30.1%
27.1%
32.5%
33.8%
was 53-5 metric tons. During periods of high flow, if soluble phosphorus from
point sources moved directly through the stream system without processing, it
could account for Only a small portion of the soluble phosphorus loading to
the lake.
In considering both soluble reactive phosphorus and bioavailable
particulate phosphorus the evidence cited above indicates that
nonpoint-derived phosphorus has a higher bioavailability upon delivery to the
lake than indirect point-source-derived phosphorus. This is primarily because
of the large amount of soluble reactive phosphorus derived from nonpoint
sources that moves into the lake during storm events.
COMPARISON OF PHOSPHORUS INPUTS FROM TRIBUTARIES AND DIRECT POINT SOURCES
Since the early 1970's, phosphorus loading to the Great Lakes from point
sources has been decreasing due to municipal phosphorus removal programs and
detergent phosphorus controls. In the mid 1970's point source loading became
smaller in magnitude than nonpoint-derived phosphorus loading (OQfi, 1975;
112
-------
PLUARG, 1978). Although there has teen much progress in reducing phosphorus
loading to Lake Erie through the implementation of phosphorus removal programs
at municipal sewage treatment plants, still further reductions in phosphorus
loading are required to meet the target loads adopted in the 1978 Great Lakes
Water Quality "Agreement (iJC, 1980; IJC, 1981 ). Two programs are attempting
to identify the most cost effective strategies for achieving the target loads
(PLUARG, 1978; COE, 1979).
A primary issue in the development of these strategies concerns the
relative costs and effectiveness of point source and nonpoint source controls
in reducing the eutrophication of Lake Erie. Considerable data is available
on the costs of varying levels of phosphorus removal at municipal wastewater
treatment plants (Drynan, 1978). Much less information is available on the
costs of nonpoint controls. In the Lake Erie basins these controls are aimed
primarily at reducing cropland erosion. Both the PLUARG and LEWS studies
have included demonstration projects in which estimates of the costs of
implementing nonpoint controls can be based. A variety of conservation
tillage programs appear to be very economical to farmers in this region (COE,
1979; Forster, 1978). It is significant that the soils most suitable for
no-till crop production are also the soils with the highest gross erosion
rates (Table 7). However, there is no correlation between gross erosion rates
and nonpoint phosphorus loading rates (Figure 23). Unfortunately, the
effectiveness of erosion control in reducing phosphorus export from large
river basins has yet to be demonstrated through large scale implementation and
long term monitoring programs. At present, estimates of phosphorus load
reductions are largely based On assumptions about sediment delivery ratios and
enrichment ratios. Consequently information on the costs per unit Of
phosphorus load reduction is much less certain for nonpoint source controls
than for point source controls.
The relative effectiveness of point and nonpoint phosphorus controls in
reducing the eutrophication of Lake Erie i's alstJ subject to question. As
noted above, upon entering the stream Or lake system, point-source-derived
phosphorus is largely bioavailable while nonpoint-derived phosphorus is Only
about 50% available. The bioavailable phosphorus from nonpoint sources is
composed Of approximately equal portions Of soluble reactive phosphorus and
bioavailable particulate phosphorus. It has already been suggested by
numerous authors that consideration of bioavailable phosrihorus should be
incorporated into future analyses of cost effectiveness (see review by
Sonzogni, et al., 1981). These considerations tend to increase the cost
effectiveness of point source controls relative to nonpoint controls.
Research in northwestern Ohio rivers suggests that temporal, spatial and
hydrodynamic aspects of phosphorus loading could also be very important in
^comparing the effects of point and nonpoint inputs. In stream systems, point
source inputs are processed very rapidly, resulting in deposition and probable
reduction in subsequent bioavailability. Direct point source phosphorus
inputs into the lower section of rivers, estuaries, bays, and even the near
shore zone of the lake, may also be subject to rapid processing, resulting in
deposition and/or conversion to less bioavailable forms. The susceptibility
of point source inputs to efficient processing may be associated with the
constant rather than pulsed nature Of these inputs. The annual point source
inputs are delivered to aquatic systems at approximately constant daily rates.
The processing of this phosphorus may be associated with significant localized
113
-------
water quality problems.
In contrast the nonpoint-derived phosphorus is delivered in association
with runoff events. Large portions of the annual loads are delivered in a
small percentage of the time. During periods of nonpoint loading the
retention time of water in the lower sections of rivers, estuaries, bays and
even the nearshore zone of the lake would be much shorter. Consequently
soluble nonpoint-derived phosphorus may be much less susceptible to processing
within these zones than point-source-derived phosphorus. Since much of the
particulate phosphorus of nonpoint Origin is associated with clay-sized
particles, this material may also be delivered rather efficiently through the
estuaries, bays and the nearshore zone to the open lake.
It should also be noted that the concentrations of both soluble and total
phosphorus in rivers during periods of high flows are much higher than the
phosphorus concentrations in lake water. Also, point source phosphorus from
the Detroit area enters the western basin of Lake Erie at much lower
concentrations than nonpoint-derived phosphorus, due to the large volume of
water with low phosphorus concentration entering the Detroit River from the
Upper Lakes.
The above discussion suggests that temporal, spatial and hydrodynamic
aspects of phosphorus loading and accompanying processing need to be included,
along with bioavailability, in refining cost-effectiveness analyses for
phosphorus management of Lake Erie.
SEDIMENT DELIVERY RATIOS AND CRITICAL AREA IDENTIFICATION
The data on mean annual sediment yields (Table 41) coupled with the data
on gross erosion rates (Table 7) allow calculation of sediment delivery ratios
for each watershed. These are summarized in ^ffble 52. The calculated
sediment delivery ratios ranged from 6.2 to 11.9$.
There have been very few measurements of sediment delivery ratios for
watersheds in this size range. In Figure 22, the delivery ratios measured in
these watersheds are superimposed On a plot of delivery ratios in relation to
drainage area. The original data is presented in Soil Conservation Service
publications (SCS, 1971) and is Often used in modeling studies (T.TcElrOy et
al., 1976). Although the study watersheds do occur in roughly the appropriate
position in the plot, they do not follow the trend Of decreasing delivery with
increasing drainage area exhibited by the Other watersheds.
Linnear regressions between sediment delivery ratios and other watershed
parameters listed in Table 52 are shown in Table 5?. Of these, the only
significant correlation was an inverse correlation between gross erosion rate
and sediment delivery ratio. In addition, there was no significant
correlation between gross erosion and sediment yield or gross erosion and
nonpoint phosphorus yields for these watersheds. Figure 23 illustrates a plot
of gross erosion and nonpoint phosphorus losses.
The above data indicate that neither sediment yields nor nonpoint
phosphorus yields are predicted very well by average gross erosion rates in
watersheds of the sizes included in this study. The lack of correlation
114
-------
Table 52. Sediment delivery ratios for Northwestern Ohio agricultural river
basins
Gaqinq
Station
Maumc-e
Portaqe
Huron
Sandusky
Mexico
Upper Sandusky
Bucyrus
Tymochtee
Honey
Nevada
Wolf, East
Wolf, West
Watershed
Are,a
Km
16,395
1,109
961
3,240
2,005
772
230
593
386
217
213
171.5
Mean
Annual
Flow
cm/yr.
26.3
25.1
27.9
26.3
25.7
28.0
32.9
25.8
27.1
35.6
33.5
26.0
Gross
Erosion
Tons/
ha/yr.
6.84
5.00
7.51
8.25
9.37
9.35
7.85
8.41
6.86
9.39
5.11
4.19
Sediment
Yield
Tons/
ha/hr.
.63
.40
.59
.52
.65
.63
.54
.52
.49
.87
.61
.48
Delivery
Ratio
%
9.2%
8.0%
7.9%
6.31!,
6.9%
6.8%
6.9%
6.2%
7.1%
9.2%
11.9%
11.5%
Table 53. Linear regressions relating to delivery, ratios, sediment yields, and
nonpoint phosphorus yields. ""
Parameters
X
Watershed Area
Watershed Area
Runoff
cm
Gross Erosion
rate
Gross Erosion
rate
Gross Erosion
rate
Y
delivery ratio
delivery ratio
delivery ratio
delivery ratio
sediment yield
unit area
nonpt. phos.**
Slope
0.0000
-0.3639
0.184
-0.737
0.0298
-0.0126
Intercept
8.09
9.58
2.95
13.6
0.3517
1.222
2
0.5%
5.15%
11.6%
47.1»*
22.5%
2.0%
DF
10
10
10
10
10
7
•significant at the p - 0.05 level.
"excludes Bucyrus, Huron and Portage basins.
115
-------
CTt
« » • T • ti
Iowa Tech. Sote Eng. 16 3/31/73^
SEOIMCNT DELIVIWT H»TIO
VS.
DRAINAGE ARE*
u » DtPAjmoxr Of AGUJCULTUM
SMLCOHSEHVATICW SERVKI
Figure 22. Relationship between sediment delivery ratio and drainage area.
-------
1.6
1.4
1.2
en
x
•a
o
.c
a.
W
o
•§.
1
1.0
.6
.4
45678 9
Gros3 Erosion Rate MT/ha/yr
Figure 23. Relationship between nonpoint phosphorua export and
gross erosion rates in the study watersheds.
10
between gross erosion and yields of sediment anfr phosphorus raises significant
questions about the effectiveness Of programs which propose to reduce sediment
and phosphorus yields by reducing gross erosion in the watersheds. It should
be noted that the differences in gross erosion between the various watersheds
listed in Tables v and 52 reflect erosional difference under conventional
tillage where ground cover is very low. The rather uniform sediment and
phosphorus yields from these watersheds may reflect approximately equal unit
area clay export from these soils. Those areas with higher gross erosion
rates may simply have lower delivery ratios due to deposition of larger
particles. Conservation tillage practices which increase ground cover could
reduce the entrainment of clay particles, thereby reducing sediment and
particulate phosphorus losses as well as gross erosion. This effect could be
independent of gross erosion rates. Thus with respect to reducing sediment
and particulate phosphorus yields to Lake Erie, "critical areas" may not exist
and, instead, efforts should concentrate On implementing various conservation
tillage practices wherever the combination of suitable soils and technology
are present. There may be more Opportunities for implementing such programs
in areas with lower gross erosion rates than in areas with higher gross
erosion rates. The benefits to the lake could be proportional to the areal
extent of conservation tillage implementation rather than to the reduction in
gross erosion per se. It is quite possible that the proportional reduction in
sediment and phosphorus yields could be greater than the reduction in gross
erosion for a large watershed.
117
-------
The interpretation presented above differs in some significant ways from
interpretations upon which many current nonpoint control planning programs are
based. Most current programs involve critical area identification through
using some combination of information On gross erosion rates, sediment
delivery ratios, or hydrologically active areas. Projected decreases in
particulate phosphorus loading are presumed to be proportional to reductions
in gross erosion, multiplied by a factor less than one (1) which takes into
account shifts in particle sizes and accompanying phosphorus export (COE,
1979). In small watersheds the proportional reduction in phosphorus export is
less than the reduction in gross erosion. In large watersheds the same does
not necessarily hold since delivery ratios may differ in various parts of the
watershed and these ratios are inversely correlated with gross erosion rates.
If control programs concentrated on areas of high delivery and low gross
erosion, the prOpOrtionalreduction in sediment and total phosphorus yields
could be much greater than the reduction in gross erosion for the watershed as
a whole.
118
-------
SECTION 9
TRANSPORT MODELING
The network of transport stations in the Sandusky Basin has been used for
the study of material transport along the mainstream Of river systems. Of
particular interest has been the development of an explanation for the
advanced sediment and phosphorus concentration peaks relative to the
hydrograph peaks that persist from station fao station as storms move down the
mains tern of the river. Since the storm wave moves downstream faster than the
water flows downstream, the water containing peak sediment and phosphorus
concentrations during the early stages of a hydrograph at an upstream station
cannot be the same water that contains the peak concentrations of sediment and
phosphorus during the early phase of a hydrograph at a downstream station.
Two explanations for the persistence of advanced peaks have been
proposed. One involves routing of materials from various areas in the
watershed to the stream gages. The second involves the incorporation of
deposition/resuspension phenomena applicable to all components of the
suspended sediment transport, including the clays.
A model was developed for material transport in the river which included
both the wave and water velocities. Only when deposition and resuspension of
sediments and phosphorus was incorporated into the model, could the patterns
of sediment and phosphorus concentrations at downstream stations be
duplicated. The routing models also allow calculations of deposition and
resuspension of materials as shown in Figure 24.
These models have been developed as part of the Lake Erie ¥astewater
Management Study and are described in detail in the Technical Report Series
associated with that study (Verhoff et al., 1978;" Melfi & Verhoff, 1979;
Yaksich & Adams, 1980; COE, 1979).
' ~r
As part of this study, a generalized river transport model based on the
above work was developed at the University of West Virginia under subcontract
from Heidelberg College. A paper presenting the generalized model is included
in the appendix of this report.
119
-------
200
NJ
o
1400
- 1200
DEPOSITION
H-ULU RESUSPENSION
I 20
1400
0.00
DEPOSITION
Sj RESUSPENSION
0.00
9 10 II 12 13 14 15 789 10 II 12 13 14 15
JULY JULY
Figure 24 - Deposition and Resuspension of (a) Total Phosphorus and (b) Orthophosphate in the Sandusky River
near Upper Sandusky - Storm beginning 7 July 1976. Source: Melfi & Verhoff, 1979.
-------
REFERENCES
Antilla, Peter ¥. and R. L. Tobin. 1978. Fluvial Sediment in Ohio.
Geological Survey Water Supply Paper 2045. United States Government
Printing Office, Washington, D.C. 20402. 58 pp.
Armstrong, D. E., J. J. Perry and D. E. Flatness. 1979. Availability of
Pollutants Associated with Suspended or Settled River Sediments Which Gain
Access to the Great Lakes. EPA-905/4-79-028, U.S. Environmental
Protection Agency, Chicago, Illinois. 90 pp.
ASTM Manual on Presentation of Data and Control Chart Analysis. ASTM Special
Technical Publication 15D, 1976. Guidelines for Control of Analytical
Procedures in an Intralaboratory Quality Control Program. 11 pp.
Bachmann, R. ¥., 1980. The Role of Agricultural Sediments and Chemicals in
Eutrophication. Journal Water Pollution Control, Vol. 52, No. 10. pp.
2425-2431.
Baker, David B. and Jack W. Kramer. 1973. Phosphorus Sources and Transport
in an Agricultural River Basin of Lake E-rie. Proc. 16th Conf. Great
Lakes Res., Internat. Assoc. Great Lakes Res. 858-871.
Baker, David B. and Jack W. Kramer. 1975a. Distribution of Nonpoint Sources
of Phosphorus and Sediment in the Sandusky River Basin. The Sandusky
River Basin Symposium, Proceedings. Int. Ref. Group on Great Lakes
Pollution from Land Use Activities, pp. 61-88.
Baker, David B. and Jack W. Kramer. 1975b. - Effects of Advanced Waste
Treatment and Flow Augmentation on Water Quality During Low Stream Flows.
The Sandusky River Basin Symposium, Proceedings. Int. Ref. Group on
Great Lakes Pollution from Land Use Activities. pp. 123-142.
Baker, David B. 1980. Upstream Point Source Phosphorus Inputs and Effects.
Seminar on Water Quality Management Trade-Offs: Point Source vs. Diffuse
Source Pollution (Conference) September 16-17, 1980, Chicago, Illinois.
pp. 227-240.
Bliss, Norman B., et al. 1975. Land Resource Measurement for Water Quality
Analysis: The Land Resource Information System. Tri-County Conservancy
of the Brandywine, Inc., Chadds Ford, Pa. 20 pp.
Cahill, Thomas H. and Robert W. Pierson. 1979. Honey Creek Watershed Report.
Lake Erie Wastewater Management Study. U.S. Army Engineer District,
~ Buffalo, New York. 79 pp.
Cahill, Thomas H., R. W. Pierson, Jr., and B. R. Cohen. 1979. Nonnoint
Source Model Calibration in Honey Creek Watershed. U.S. EP/V. Atlanta,
Georgia. EPA-600/3-79-054. 133 pp.
COE, 1975a. Lake Erie Wastewater Management Study. Volume 1, Main Report.
Corps of Engineers, Department of the Army, Buffalo, New York. 172 pp.
121
-------
COE, 1975b. Lake Erie Wastewater Management Study. Volume II, Preliminary
Feasibility Report, Water Quality Inventory. Corps of Engineers,
Department of the Army, Buffalo, New York. 202 pp.
COE, 1978. Lake Erie Wastewater Management Study. Water Quality Data for
Sandusky River Material Transport. U.S. Army Corps of Engineers, Buffalo,
New York. 262 pp.
COE, 1979- Lake Erie Wastewater Management Study Methodology Report. Corps
of Engineers, Department of the Army, Buffalo, New York. 146 pp.
Cummins, Kenneth W., 1975. The Ecology of Running Waters, Theory and
Practice. The Sandusky River Basin Symposium, Proceedings. Int. Ref.
Group on Great Lakes Pollution from Land Use Activities, pp. 277-294.
DePinto, Jospeh V-, et al. 1980. Phosphorus Removal in Lower Great Lakes
Municipal Treatment Plants. U.S. Environmental Protection Agency,
Cincinnati, Ohio. 147 pp.
Drynan, W. R. 1978. Relative Costs of Achieving Various Levels of Phosphorus
Control at Municipal Wastewater Treatment Plants in the Great Lakes Basin.
IJC, Windsor, Ontario, Canada. 57 pp.
Environmental Protection Agency. 1979. Methods for Chemical Analysis of
Water and Wastes. EPA-600/4-79-020, U.S. Environmental Protection Agency,
Cincinnati, Ohio.
Forster, D. Lynn. 1978. Economic Impacts of Changing Tillage Practices in
the Lake Erie Basin. Lake Erie Wastewater Management Study. U.S. Army
Engineer District, Buffalo, New York. 59 pp--
Forsyth, Jane L. 1975. The Geological Setting of-the Sandusky River Basin,
Ohio. The Sandusky River Basin Symposium, Proceedings. Inbernational
Reference Group on Great Lakes Pollution from Land Use Activities, pp.
13-60.
Honey Creek Joint Board of Supervisors. 1980. Honey Creek Watershed ^reject.
Lake Erie Wastewater Management Studies. U.S. Army Corps of Engineers,
Buffalo, NY. 61 pp.
IJC, 1979. Inventory of Major Municipal and Industrial ^oint Source
Discharges in the Great Lakes Basin. Great Lakes Water Quality Board,
Remedial Programs Subcommittee. Windsor, Ontario, Canada.
_IJC, 1980a. Pollution in the Great Lakes Basin from Land Use Activities.
Windsor, Ontario, Canada. 141 pp.
IJC, 1980b. Report on Great Lakes Water Quality Appendix. Great Lakes Water
Quality Board. Windsor, Ontario, Canada. 82 pp.
IJC, 1981. Supplemental Report under the Reference on Pollution in the Great
Lakes System From Land Use Activities on Phosphorus Management Strategies.
Great Lakes Water Quality Board. Windsor, Ontario, Canada. 24 pp.
122
-------
Johnson, Murray G., et al. 1978. Management Information Base and Overview
Modelling. International Joint Commission. International Reference Group
on Great Lakes Pollution from Land Use Activities. 90 pp.
Keup, L. E. 1968. Phosphorus in Flowing Waters. Water Res. 2:37?-^86.
Krieger, Kenneth A., et al. 1980. Environmental Quality of Upper Honey
Creek: A Preliminary Assessment. U.S. Army Engineer District, Buffalo,
New York. 74 pp.
Lee, G. F. , R. A. Jones, and W. Rast. 1980. Availability of Phosphorus to
Phytoplankton and its Implication for Phosphorus Management Stragegies.
IN Phosphorus Management Strategies for Lakes. Ann Arbor Science pub.,
Ann Arbor, Michigan, pp. 259-310.
Logan, Terry J. 1978a. Maumee River Basin Pilot Watershed Study.
International Joint Commission. International Reference Group on Great
Lakes Pollution from Land Use Activities. 96 pp.
Logan, Terry J. 1978b. Chemical Extraction as an Index of Bioavailability of
Phosphate in Lake Erie Basin Suspended Sediments. Lake Erie Wastewater
Management Study, U.S. Army Corps of Engineers Buffalo District, Buffalo,
New York. 49 pp.
McElroy, A. D., S. Y. Chiu, J. W. Nebgen, A. Aleti, and F. W. Bennett. 1976.
Loading Functions for Assessment of Water Pollution from Nonpoint Sources.
U.S. Environmental Protection Agency, Washington D. C. EPA-600/2-76-151.
445 pp.
Melfi, David and Frank Verhoff. 1979- Material.-Transport in River Systems
During Storm Events by Water Routing. Lake Erie Wastewater Management
Study, U.S. Army Corps of Engineers Buffalo'District, Buffalo, New York.
70 pp.
Monteith, Timothy. 1980. A Management Technique for Choosing Among Point and
Nonpoint Control Strategies, Part 2 - A River Basin Case Study. Seminar
on Water Quality Management Trade-Offs: Point Source vs. Diffuse Source
Pollution (Conference) September 16-17, 1980, Chicago, Illinois, pp. 129
- 162.
Omernik, James M. 1977. Nonpoint Source—Stream Nutrient Level
Relationships: A Nationwide Study. U.S. Environmental Protection Agency,
Corvallis, Oregon. 151 pp.
JPLUARG, 1974. The Effect of Residential and Commercial-Industrial Land Usage
on Water Quality. Windsor, Ontario, Canada. 38 pp.
PLUARG, 1978. Environmental Management Strategy for the Great Lakes Joint
Commission. Environmental Management Strategy for the Great Lakes System.
Windsor, Ontario, Canada. 115 pp.
RMA, 1979- Land Resource Summary, Vol. I, Major River Basins, Vol. II,
Sandusky River Basin, Vol. III. TT.S. Army Engineer District, Buffalo,
New York. 407 pp.
123
-------
SCS, 1971. Sedimentation, section 3• National Engineering Handbook. Soil
Conservation Service, U.S. Department of Agriculture, Washington B.C.
Sonzogni, William C., Timothy J. Monteith, William N. Bach, and V- Gregory
Hughes. 1978. United States Great Lakes Tributary Loadings.
U.S. Environmental Protection Agency. 187 pp.
Sonzogni, William. 1980. A Management Technique for Choosing Among Point and
Nonpoint Control Strategies, Part 1-Theory and Process "Framework. Seminar
on Water Quality Management Trade-Offs: Point Source vs. Diffuse Source
Pollution (Conference) September 16-17, 1980, Chicago, Illinois.
pp. 87-124.
Sonzogni, William, and Steven C. Chapra. 1981. Bioavailability of Phosphorus
Inputs to Lakes: Significance to Management. Great Lakes Basin
Commission, Ann Arbor, Michigan.
Strand, Robert I. 1975. Bureau of Reclamation Procedures for "Predicting
Sediment Yield. Proceedings of the Sediment-Yield Workshop, USDA
Sedimentation Laboratory, Oxford, Mississippi, November 28-30, 1972.
U.S. Department of Agriculture, pp. 10-15.
Thomann, Robert V. Systems Analysis & Water Quality Management.
Environmental Research and Applications, Inc., New York, 1972. 286 pp.
Urban, Donald R., Terry J. Logan and John R. Adams. 1978. Application of
the Universal Soil Loss Equation in the Lake Erie Drainage Basin. Lake
Erie Wastewater Management Study. U.S. Army Corps of Engineers, Buffalo
District, Buffalo, New York. 45 pp.
Urban, Donald R., Terry J. Logan and John R. Ad'ams. ---1978. Application of the
Universal Soil Loss Equation in the Lake Erie Drainage Basin. Appendix I.
U.S. Army Corps of Engineers, Buffalo District, Buffalo, New York. 521
pp.
Uttormark, Paul D., John D. Chapin and Kenneth M. Green. 1974. Estimating
Nutrient Loadings of Lakes from Non-Point Sources. U.S. Environmental
Protection Agency, Washington, D.C. EPA-660/3/-74-020. 113 pp.
Verhoff, Frank H., David A. Melfi and David B. Baker. 1978. Phosphorus
Transport in Rivers. Lake Erie Wastewater Management Study, U.S. Army
Corps of Engineers, Buffalo District, Buffalo, New York. 88 pp.
Jerhoff, Prank H. and David B. Baker. 1981. Moment Methods for Analyzing
~~~- River Models With Application to Point Source Phosphorus. Water Research,
15:493-501 .
Wischmeier, Walter H. and Dwight D. Smith. 1978. Predicting Rainfall Erosion
Losses—A Guide to Conservation Planning. Agriculture Handbook No. 537.
U.S. Department of Agriculture. 58 pp.
124
-------
Yaksich, Stephen and John Adams. 1980. Sediment and Phosphorus Transport.
Seminar on Water Quality Management Trade-Offs: Point Source vs. Diffuse
Source Pollution (Conference) September 16-17, 1980, Chicago, Illinois.
pp. 241-278.
Zison, Stanley W. 1980. Sediment-Pollutant Relationships in Runoff From
Selected Agricultural, Suburban, and Urban Watersheds: A Statistical
Correlation Study. U.S. Environmental Protection Agency, Athens, Georgia.
136 pp.
125
-------
APPENDIX 1
ANALYTICAL METHODS
126
-------
Soluble Reactive Phosphorus (Modified colorimetric ascorbic)
Storet No. 00671
Scope and Application - Methods for phosphorus are based on reactions
specific for the orthophosphate ion. The methods cover the determination of
specific forms of phosphorus present in surface waters. The sample
pretreatment determines the form of phosphorus to be measured.
For dissolved orthophosphorus the sample is filtered through a prewashed
follows using a modified Method 365.1 from Methods for Analysis of Water and
Wastes. U.S. Environmental Protection Agency, National Environmental Research
Center, Cincinnati, Ohio, 1979. EPA-600/4-79-020, pages 365.1-1 - 365-1-9-
The method has been modified from the single reagent method in several
ways. First, the ascorbic acid is a separate reagent and is added to the
analytical stream in place of the distilled water, and in effect we make our
mixed reagent within the analytical system. By doing this our mixed reagent
has a shelf life of several months, not the four hours as in the EPA method.
We also put a much larger volume (sample & reagent) through our analytical
system. This results in a more reliable system and improves flow cell clean
out.
Nitrate and Nitrite Nitrogen (automated cadmium reduction)
Storet No. 00631
Prior to analysis by cadmium reduction from Method 353-2 from Chemical
Analysis of Water and Wastes, U.S. Environmental Protection Agency, National
Environmental Research Center, Cincinnati, Ohio,_ 1979- EPA-600/ 4-79-020,
pages 353.2,2-7. The samples are filtered " through a .45 micron pore size
membrane filter.
• -,-
Ammonia (colorimetric automated phenate) Storet No. 00608
The sample is filtered through a .45 micron pore size membrane filter and
analyzed using Method 350.1 from Chemical Analysis of Water and Wastes,
U.S. Environmental Protection Agency, National Environmental Research Center,
Cincinnati, Ohio, 1979. EPA-600/4-79-020, pages 350.1-1 - 350.1-6.
Chloride (colorimetric automated Ferricyanate) Storet No. 00940
Following filtration through a membrane filter with a pore size of from
Jthe Chemical Analysis of Water and Wastes, U.S. Environmental Protection
Agency, National Environmental Research Center, Cincinnati, Ohio. 1979.
EPA-600/4-79-020, pages 325.2-1 - 325.2-3.
Sulfate (colorimetric, automated, methylthynol Blue) Storet No. 00945
The samples are filtered through a .45 micron pore size membrane filter
prior to analysis by Method 375.2 from Chemical Analysis of Water and Wastes,
U.S. Environmental Protection Agency, National Environmental Research Center,
127
-------
Cincinnati. Ohio. 1979. EPA-600/4-79-020. pages 375-2-1 - 375-2-3-
Silica (colorimetric molybdate reactive) Storet No. 00955
Samples are filtered throuth a .45 micron membrane filter prior to
analysis using Technicon Industrial Method 105-71W from Technicon Industrial
Systems, Tarrytown, New York, 10591, February 1973-
Total Phosphorus Storet No. 00665
A 50ml aliquot of a well mixed sample is placed in a 125ml Erlenmeyer
flask. One milliter (automate pipette) of strong acid and four-tenths of a
gram (pre measured scoop) of ammonium persulfate are then added to the flask.
Two blanks and four standards are prepared with each batch of samples. After
autoclaving (45 minutes) at 15 Ibs. pressure and 121 degrees C. the samples
are filtered through a glass fiber filter (pre filter pore size) to remove
turbidity and poured out into tubes for analysis on the total phosphorus
autoanalyzer. This method is a modification of Method 365.1 (see SRP) from
the Chemical Analysis of Water and Wastes, U.S. Environmental Protection
Agency, National Environmental Research Center, Cincinnati, Ohio. 1979-
EPA-600/4-79-020, pages 365-1-1 - 365-1-9-
When the samples are removed from the autoclave, those samples which had
high amounts of suspended matter are checked to confirm that the residue is
white to light grey in color- If color is present the sample(s) are set up
again and digested with larger amounts of persulfate.
Total Kjeldahl Nitrogen (colorimetric, automated phenate) Storet No. 00625
A 10ml* aliquot of a well mixed sample is placed in a 75ml digestion
tube. Four milliliters (by repipet) and 2-3 Teflon boiling chips are added to
each tube, and the sample-acid solution is vortexed before the sample set is
put on the block. The block is set for 2 temperatures and 2 times. The first
temperature is 180 degrees C for 1 hour, the second temperature is 380 degrees
C for 3 hours. When the digestion is completed, the tubes are taken off the
block and allowed to cool down. Ten milliliters of distilled water is added
to each tube to get back to the original volume. Before pouring the sample
out into test tubes, the samples are vortexed again to get a thorough mixture
of distilled and digested acid. At this point any residue from suspended
solids should be light grey or off white in color, if color is present,
consult with the lab manager. Analysis follows using Method 351-1 from
Chemical Analysis of Water and Chemical Analysis of Water and Wastes,
U.S. Environmental Protection Agency, National Agency, National Environmental
Research Center, Cincinnati, Ohio, 1979- EPA-600/4-79-020, pages 351-1 -
351.1-4-
*Volume varies with sample type.
128
-------
Conductivity
Specific conductance is measured using a Water Quality Laboratory
conductivity meter (product of WQL) and a Yellow Springs Instrument (YSl)
conductivity cell. Samples are run at 25 degrees. This is in accordance with
Method 120.1 from Methods for Analysis of Water and Wastes, U.S. Environmental
Protection Agency, National Environmental Research Center, Cincinnati, Ohio,
1979. EPA-600/4-79-020, page 120.1.1.
pH (potentiomentric) Storet No. 00400
pH measurements are determined using a glass combination electrode with
an Orion 701 digital pH meter. The samples are run at 25 degrees C following
the guidelines for Method 150.1 in Methods for Analysis of Water and Wastes,
U.S. Environmental Protection Agency, National Environmental Research Center,
Cincinnati, Ohio, 1979. EPA-600/4-79-020, page 150.1-1.
Specific Conductance
Specific conductance is measured using a Water Quality Laboratory
conductivity meter (product of WQL) and a Yellow Springs Instrument (YSl)
conductivity cell. Samples are run at 25 degrees. This is in accordance with
Method 120.1 from Methods for Analysis of Water and Wastes, U.S. Environmental
Protection Agency, National Environmental Research Center, Cincinnati, Ohio,
1979. EPA-600/4-79-020, page 120.1.1.
Suspended Solids (Residue, Non-Filterable)
A 100 ml aliquot of a well mixed sample is vacuum filtered through a
pre-weighed glass fiber filter. The residue retained on the filter is dried
to a constant weight at 103 degrees - 105 degrees as described in Method 160.2
in Methods for Chemical Analysis of Water and Wastes, U.S. Environmental
Protection Agency, National Environmental Research Center, Cincinnati, Ohio.
1979. EPA-600/4-79-020, pages 160.2-1 - 160.2-3.
129
-------
APPENDIX 2
Integral Methods for Approximate Water
and Pollutant Transport In Rivers
W. C. Peterson
Department of Mathematics
and
F. H. Verhoff
Department of Chemical Engineering
West Virginia University
Morgantown, WV 26506
130
-------
ABSTRACT
Integral Methods for Approximate Water
and Pollutant Transport In Rivers
W.C. Peterson, Department of Mathematics
and
F.H. Verhoff, Department of Chemical Engineering
West Virginia University
Morgantown, WV 26506
Mass balances in an integral form are used for both water and
transported material in solution.
The' integral form for the water leads to a general approximation
for water routing of which the Musklngum method is a special case.
The integral form of the mass balance for the transport of material
leads to a general approximation for material transport which can be
coupled with the method above for water routing to obtain both flow
rate and concentration at a downstream location from known upstream data
based primarily on knowledge of the flow vs area curve at upstream and
downstream locations. Parameters in the approximation are obtained
from knowledge of the channel and material in solution. The parameters
are obtained to be functions of the wave and/or water velocity.
The approximation was applied to data from the Sandusky River In
Ohio.
131
-------
INTRODUCTION
The input of substances from point and non-point sources causes
concentration changes within the river itself and within the receiving
water body. The basic concept used to mathematically model these
chemical changes Is the mass balance. This balance can be applied to
both the water Itself and/or to the chemical species of interest.
Also, It can be used for the prediction of In-river water quality or
It can be used to predict the amount of substance transported through
the river. Streeter and Phelps [1925] probably were the first to use
the mass balance on a chemical substance, in this instance BOD. They
were interested in steady flow and hence the water mass balance took
the simple form of constant water velocity. Their mass balance on
BOD then predicted the BOD and oxygen concentrations as a function of
distance in the river, i.e. in river quality.
This paper focuses on the transport of substances through river
> ^
systems particular My during storms rather than on the in-river water
quality. The transport of substances during storms is important
because much of the water and associated chemicals reach the receiving
water body, e.g., lake, during storms. For example, total phosphorus,
a principal nutrient In lake eutrophlcation, is primarily transported
during storms. Choride also is transported during winter storms from
roadways of northern areas.
The modeling of chemical transport during storms requires a mass
balance for both the water and the chemical substance of interest. The
132
-------
solution of the water mass balance during storms has been Investigated
for many years and is often called water routing. The main goal of
this prior research has been prediction of the flood peak as 5t moves
downstream. Much less work has been done on the solution of both the
transient water mass balance and the transient chemical mass balance as
is applicable during storm transport.
Herein, an integral method for solving the water and substance mass
balances between two different stations on the river Is developed. This
is in contrast to the finite difference approximate solutions which have
been the main methods previously employed. Essentially, the mass
balances are Integrated between the two stations of the river, including
the input from tributaries and unmeasured areas. Thus if the concen-
trations and flows are known at the upstream station and at the tributary
mouths, and the flow and concentration of the unmeasured inflow can be
estimated, the concentration of the pollutant at the downstream station
on the river can be calculated. This integral method could be used to
predict the concentration or aid in the understanding of the transport
of various chemical species in rivers.
Much research has been done related to the transport of substances
In rivers. Basically, the research work can be classified as storm vs
steady flow, and correlational vs deterministic models. Various
.correlations have been found between the concentration of different chemical
species during steady flow (see Enviro Control [1972]). Also, the
relationship between various chemical concentrations and between concen-
tration and flow has been the subject of numerous authors, e.g. Foster
[1978] and Feller and Kimmins [1979]- They have found that some
133
-------
substances as total phosphorus and suspended sediment concentrations
increase with flow and that there is a correlation between these two
concentrations. On the other hand species such as chloride usually
decrease in concentration with increasing flow. The variability of
the concentration is dependent upon the particular locality and upon
the season of the year. These studies are not deterministic and do not
provide a quantitative understanding of the processes involved in the
transport, however, they do help discriminate between various plausable
mechanisms for chemical transport.
Deterministic models for the transport of materials in rivers has
been attempted since the original work of Streeter and Phelps [19251
which was applicable to rivers under steady flow. These river models
have been made more complex to Incorporate more of the nutrient and
chemical cycling mechanisms (see Chen [1970] and Sandavol , et al. [1976])
who have used these models for predicting the expected concentrations
of various substances as a function of stream distance and time.
Further, steady flow models have been used with transient concentration
variations to discriminate between models and to determine parameters
within models, e.g. Verhoff and Baker [1981 ].
However, the number of researchers who determlnistlcally modelled
the transport of substances during unsteady flow Is rather limited,
except for the transport of suspended sediment. These suspended
sediment models usually involve mechanisms for the resuspensIon and
deposition of sediment particles of different sizes and they are primarily
used to predict changes in stream bed morophology ie, points in the stream
134
-------
where scour and deposition can occur e.g. U.S. Army Engineers HEC
[19771. Yalin [1972] has summarized various of these deterministic
models used to describe this suspended sediment transport. In a
recent paper, O'Conner [1976] considers the interaction of ground-
water and surface water in the transport of chemical substances.
He primarily was interested in steady flow but he does present examples
of unsteady flow. Since unsteady hydrologlcal models have been
available for some time, there have been several attemps to couple the
transient chemical dynamics to these transient water models. These
storm water models are primarily finite difference solutions of the
governing equations coupled with various correlations. Schultz and
Wilmarth [1978] have applied the Hydrocomp Simulation Programming model
to water quality in Southwestern Illinois. This model starts with the
precipitation and predicts the hydrograph in the Hver followed by the
concentration of various substances as a function of-time at different
points in the river. For the one storm event present in the article
the predicted hydrograph did not correspond closely with the measured
one. They also showed reasonable comparlsions between predicted and
measured concentrations under relatively stable flow conditions.
135
-------
INTEGRAL REPRESENTATIONS
The technique to be described uses the assumption that water is
moving as a kinematic wave. The data calculated wi I 1 he derived from
information given at one point on the river reach, thus the assumption
that a given reach of the river is simular to the measurement point has
been made. In the following discussion parameters will be determined
dependent on knowledge of the wave velocity and of the water velocity at
an upstream and downstream measuring station over a short river reach.
The water and wave velocity at a point on the river reach can be deter-
mined from the flow vs channel cross-section area curve. This relation-
ship, such as given in figure 1 ,can be represented by the equation
1 = f(A) (I)
where Q. is" the river flow rate and A is the channel cross-sect i on area,
f may be a multi-valued function relationship depending as to whether the
hydrograph at a point in the river is rising or" falling with respect to
time. From kinematic theory of hydrologic events the wave velocity is
determined by the slope f'(A) of the curve given by Q. =• f(A) and
the water velocity is given by Q/A .
For kinematic waves, the continuity equation for open channel flow
has the form
3A . 30 .. _ ,„,
in which t = time , x = distance measured along the channel reach,
A = A(x,t) is the channel cross -sect i on area, Q. = Q.(x,t) is the flow
rate and q = q(x,t) is the lateral inflow per unit length. Writting
136
-------
equation (2) in che form
and integrating each side with respect to s from x to x+Ax over
a channel reach Ax , it follows that
x+Ax x+Ax x+Ax
jm(s,t)ds = - / 3jl(s,t)ds + / q(s,t)ds
x 3t x 3x x
x+Ax
= d(x,t) - Q(x+Ax,t) + / q(s,t)ds (3)
x
Now, assuming in equation (3) that integration and differentation can be
interchanged, which should be the case for these physically continuous
functions, at least not near a strong shock, let
) _^
x+Ax
dS.(t) _ 1/x Ms.Ods
"dt 3t
where S(t) is the volume of water in the river reach from x to x+Ax
at time t , so it follows that
.r/ \ X+AX
^y ' = Q(x,t) - l(x+Ax,t) + / q(s,t)ds
Equation (k) is the continuity equation (2) described as a maroscopic
mass balance. Equation CO is often used as the starting place to
137
-------
develop a finite difference type approximation for water routing as in
the Muskingum method as in Weinmann and Laurenson [19791 •
Integrating each side of equation CO from time t to tine t
o
and re-arranging the terms, the volume of water in the channel reach Ax
at time t , S(t) , can be written in the form
x+Ax
S(t) - / A(s,t)ds
x+Ax t t
A(s,t )ds + / Q.(x,w)dw - / Q.(x+Ax,w)dw
r\ O U L
O O
t X+AX
/ / q(s,w)dsdw (5)
L X
The values, of the integrals in equation (5) will be approx ima ted later
by a choice of parameters. This then is the-integral form of the water
mass balance to be used herein.
The same procedure with the conditions about the river form can be
used for material transport. Relative to the kinematic wave for water,
the mass balance for a dissolved material such as orthopospate or suspended
material such as total phosphorus in open channel is given by
where C - C(x,t) is the concentration at x and time t on the
channel reach and C. = C.(x,t) is the lateral inflow concentration of
138
-------
the lateral inflow q per unit length.
Equation (6) is of a form similar to that of equation (2). To
solve equation (6) for the concentration, C. and q must be supplied as
input functions, Q and A must be obtained from the water mass balance or
be known and boundary conditions must be given. Again an integral form
of this equation will be obtained, following the procedure used to obtain
the integral form of the water mass balance in equation CO . From the
material mass balance given by equation (6) an expression for the total
mass, S(t) , at time t over a channel reach fron x to x+Ax can
be obtained where
x+Ax
S(t) = / A(s,t)C(s,t)ds
and
x+Ax
-------
Equation (7) is then an integral Corn of the n.itc^LiI noss balanc" If
'•v i 1 1 be use'1 in the following approximation schrnr', .
Eqiu t i ons ('<•) and (?) will he t ho rm i n pq'in f i ons "'• O'l in rlpwloninn,
predictive water quality node Is for storm «vonts. '/c ..'ill look at methods
to approximate these equations, at procedure:; tn r^l/ito 1 to A .1^.1 to
estimate q and r. . . The- follo-.'in.j sort ions '-.'ill r.-v-c ;.',>r these nrohlcms.
METHOD OF APPROXIMATION
The goal of this section is to develop approximations of equations
(5) and (7) with second order accuracy to the flow Q or concentration C at
the downstream location using information from the upstream hydrograph,
unmeasured lateral inflow and the Q. vs A curve. The interals given by
equations (5) and (7) will be approximated to third order accuracy by use
of parameters and intervals of length Ax and At =* t - t where by the
o
parameters can be estimated from the available data.
In a previous paper [1981] the authors consider methods of approximation
') .„
of equation (5) for water routing. The method of approximation is based
on the following, applied here to a function h(x) of one variable with
at least second order derivatives on an interval from x to x, using
o 1
Taylors theorem. The function h(x) can be written in the form
h(x) - ah(x) + (1 - a)h(x)
- a[h(xo) + h'(xo)(x - XQ) + h"(xQ)(x - XQ)2/2]
+ (1 - aJChUj) + h'Cx^Cx - Xj) + h-'UjHx - Xl)2/2]
140
-------
- o[h(x ) + h'(x )(x - x )] * (1 - a)[h(x.) + h'(x,)(x - x.)l
O O O III
+ [oh"(x )(x - x )2 + (1 - o)h"(xj(x - x,)2]/2 (8)
O O 11
where x x, are between x and x and x and x, , respectively,
o 1 o 1 ' '
depending on x .
Now integrate each side of equation (8) from x to x. , simplify,
use the mean-value theorem for integrals to simplify the remainder term,
Xl
/ h(x)dx - [oh(x ) + (1 - a)h(x,)]Ax + [ah'(x ) - (1 - a)h'(x,)](Ax)2/2
X O 1 O 1
o
+ 0((Ax)2) (9)
where Ax » x - x . The details for the approximation given by
equation (9) are given by the authors in U981]. Equation (9) simplifies
to the trapezodial rule for integration for the special case of a - 1/2 ,
The above form of approximation will be used to express each of the integrals
In equations (5) and (7) in terms of parameters a,6,and \i> similar to what
has been done above in equation (9). a,B, and \l> will then be chosen so
that the terms Involving the derivatives will be of order (Ax) . In
equation (9) taking a - 1/2 forces the term Involving the derivative to
be of order (Ax) , other choices for a are possible if additional properties
of h(x) are known.
Equation (5), for the water volume in the channel reach Ax, can then
be wrltten in the form
141
-------
S(t) - [aA(x,t) + (1 - a)A(x-(-Ax.t)]Ax
CaA(x,t ) + (1 - a)A(x+Ax,t ) ]Ax
[BQ(x,to) + (1 - B)Q.(x+Ax,t ) + (1 - 40Q(x+Ax,t) ]At
t x+Ax -
+ / /x q(s,w)dsdw + 0((Ax)^) (10)
o
where Ax - kAt for some constant k and .a/B.and fy are arbitrary.
a,6, and ty are determined so as to increase the order of the terms
involving the partial derivatives to order (Ax) , such choice is not
unique and will be chosen to depend on the knowledge of the Q vs A curve
at the upstream location.
'n [1981 ] tne authors write A(x+Ax,t) and A(x+Ax,t ) In terms of
Q(x,t), Q(x,t ), Q(x+Ax,t) and Q(x+Ax,t ) substitute into equation (10)
and simplify where At - K - Ax/f'(A(x,t )) yielding an expression
similar in form to that of the Muskingum method for water routing.
The choice for K, the time parameter is the same choice as is often used
in the Muskingum method for water routing as has been noted by Cunge [1965]
and Nash C 1953 J.
142
-------
Equation (10) with the above change can then be written in the form,
after re-arranging the terms
KCad(x,t) + 0 - a)Q(x+Ax,t) ] - K[aQ.(x,t ) + (1 - a) Q_(x+Ax, t) 1
. a)(|A(x+Ax,t) _ |A(x+Ax,to))](Ax)2/2
3x 3x
+ [o(3A(x.t) . lA
3x 3x
- KCBQ(x,t
- S)Q(x,t)]
at
t x+Ax
/ q(s,w)dsdw
o X
(11)
Solving for Q(x+Ax,t) , the flow rate at the downstream location at
time t , equation (11) can be re-written in the form
2 - a -
|Q(x,to)
1 - a -
2 - a -
0.(x+Ax,tQ)
3x
3x
. o)(|A(x+Ax.t)
ox
t x+Ax
/. / q(s,w)dsdw/K(2 - a
t X
-f 0((Ax) )
(12)
143
-------
In equation (11) or (12) the parameters a, 8, and ^ are to be chosen
so that the order of the terms involving the partial derivative terms
is increased, this can be done with any one of several possible choices
for a, 6, and \l> . To aviod computing with Q(x+Ax,t ), data dependent
on knowledge of the new downstream computed data, one choice is to take
a - 1 - i|» . In [ 1981 ] the authors show that a - 0 - 1 - 41 Is such a
choice where a - f'(A(x,t ))/2f'(A(x,t)). This choice for a can be
computed from the upstream data and knowledge of the Q vs A curve at
the upstream location, so here we have a as a function of the wave
velocity at the upstream location. Thus, In equation (12), the downstream
flow rate, Q(x+Ax,t), can be predicted from knowledge of or previously
computed upstream flow rates Q(x,t) and Q(x,t ) and knowledge of the
Q vs A curve at the upstream location. With the above choice for a, 6,
and 4» equation (12) has the form
t x+Ax -
Q(x+Ax,t) - (1 - 2a)d(x,t) + 2aQ(x,t ) + / / q(s,w) dsdw. •»• 0((Ax)-) (13)
o '*•
In a manner similar to the procedure used to obtain an approximation
for the downstream flow rate from knowledge of the upstream flow rate and
the upstream Q. vs A curve an approximation for the downstream concentration
C at some time t can be obtained from knowledge of the upstream flow
rate, cross-section area of the channel and the Q vs A curve. From
equation (9), the material mass balance given by the Integral form of
equation (7) and in terms of the (new) parameters a, 8, and i|» , to be
chosen, can be written as
144
-------
S(t) - Ax[aA(x,t)C(x,t) + (1 - a)A (x+Ax ,t) C (x+Ax
+ [ai(_A(x,t)C(x,t)) _ (J _ a)9J_A(x+Ax,t)
oX "X
- Ax[aA(x,t )C(x,t ) + (1 - a)A(x+Ax,t )C(x+Ax,t )]
+ [al(A(x,to)C(x,to)) . (1 _ a)|lA(x+Ax,to)C(x+Ax,to))](Ax)2/2
3x 3x
At[Bd(x,to)C(x,to) + (1 - e)Q(x,t)C(x,tn
3 1 9 1
,t )C(x-»-Ax,t ) •»• (1 - ^)Q(x+Ax ,t)C(x+Ax,t)
3 1 3 t
t x+Ax .
+ / / q(s.w)C.(s.w)dsdw + 0((Axr) (H)
t X I
o
As previously noted for water routing, chose Ax - f'(A(x,t ))At so that
time and distance are related to the wave velocity. Now to obtain a
form of approximation similar to that of equation (12) from equation (1*0
solve equation (1*0 for C(x+Ax,t), the downstream concentration. From
the form for C(x+Ax,t) the routing parameters a, 3, and *l> will be
determined. a will be determined so as to be a function of the wave
velocity of the flow relative to the Q vs A curve and the upstream and
downstream hydrograph for water.
145
-------
For convience, let f (A(x,t )) » W and
(1 - a)WA(x+Ax,t) + (1 -i|>)Q(x+Ax,t) =• D. solve equation OM for
C(x+Ax,t), giving the form
C(x+Ax,t) -[-aWA(x.t) + (1 - B)d(x ,t) ]C (x ,t)/D
CaWA(x,t ) + BQ(x,t )J C(x,t )
o o o
[-4-Q(x+Ax ,t ) + (1 - a)WA(x+Ax,t )] C(x+Ax,t
. . ,3(A(x+Ax,t)C(x+Ax ,c
U - aM—
f }]
3x
at 3 t
t x+Ax „
+ /t /x q(s,w)C.(s,w)dsdw/DAt + 0((Ax)Z) (15)
o
In equation (15) force the coefficient of C(x+Ax,t ) to be zero or
2
to contribute an error of order (Ax) so as to avoid computing with
downstream concentration data. Here the choice is made so that
-4/Q(x+A*,t ) + (1 - a)WA(x+Ax,t ) - 0 . (16)
In equation (16) the choice for a and ty is not unique. Here let
a •• 4/ as one possible choice and solve equation (16) for a giving
146
-------
WA(x+Ax,t )
WA(x+Ax,t ) + Q.(x+Ax,t )
W (17)
W + v
o
where v = Q(x+Ax,t )/A(x+Ax,t ) is the water velocity of the
o o o
hydrograph at the downstream location and time (x+Ax,t ), where
also the flow, Q.(x+Ax,t ). is previously calculated by a water
o
routing method such as that given by equation (13). Also the channel
cross-section area, A(x+Ax,t ), can be calculated from knowledge of the
0, vs A curve at x+Ax on the channel reach.
The choice of the parameters a, 8, and , equation (1 5l ,for-the downstream
concentration, has the form
C(x+Ax,t) - ((1 - o)Q(x,t) - WA(x,t))C(x,t)/D
+ o(WA(x,tQ) + Q(x,t ))C(xft )/D
t x+Ax «
•*•/,/ q(s,w)C.(s,w)dsdw/DAt +0((Ax)') (18)
t X I
o
A second choice for the routing parameters, a • '(* *• 1 - B and
the same choice for a, as given by equation (17), so that equation (15)
147
-------
can be written in the form
C(x+Ax,t) - o(d(x,t) - WA(x,t))C(x,t)/D
* (aWA(x,t ) + (1 - ct)Q(x,t ))C(x,t
o o o
t x+Ax
+ / / q(s,w)C.(s,w)dsdw/DAt + 0((Ax) ) (1?)
t A I
o
In the next section a discussion of how the integrals representing
the lateral inflow and lateral inflow concentration might be approximated,
This information is necessary in equations (13), 08), and (19).
148
-------
UNMEASURED INFLOW
In equation (13), to predict the downstream flow rate Q(x+Ax,t),
or equation (19), to predict the downstream material concentration
C(x+Ax,t), knowledge of the amount of unmeasured lateral inflow q and
the amount of unmeasured lateral inflow concentration C. is required,
both relative to the location on the stream reach x and the time t.
The amount of unmeasured lateral water inflow or the material concentration
in the lateral inflow over a given channel reach for a storm event can be
estimated with some knowledge of the river basin. This knowledge
includes water flow and material concentration for storm events occuring
at a simular time of year and with a simular intensity and duration. In
the following discussion a method will be developed which can give an
indication of when, relative to time and the knowledge of the upstream
hydrograph, the relative amount of available unmeasured inflow should be
added to the computed downstream hydrograph. It will be assumed that
* mf.
the unmeasured material concentration C. is carried with the unmeasured
water inflow q . In the following discussion it will be noted that the
largest contribution of the unmeasured inflow q to changes in the down-
stream hydrograph at a position x on the channel reach when compared to
the upstream hydrograph at position x+Ax on the channel reach will occur
in time after the peak in the hydrograph at x . Knowledge of q then
allows an approximation for the integrals in equations (13), 08) or (19).
That this is reasionable follows by re-writting the form of the
momentum equation for open channel flow and comparing q with the slope
of the free water surface under some steady state conditions.
149
-------
The momentum equation for open channel flow, Stoker [19571, can be
wri tten in the form
3v 9v fLv 1 / \
a t 9x A Ax
where v is the water velocity, s is the slope of the free water
surface (positive downward for positive downstream channel distance), f
the friction force, L the wetted perimeter of the channel bed and u
the component of the velocity of the lateral water inflow q in the
downstream direction.
With trie assumption that -rr + v-— is small or zero, that is over
a t oX
a short channel reach x to x+Ax and small change in time, the change
in the velocity of the water is negligible (velocity constant), equation (20)
can then be solved for the lateral Inflow q » (gsA - fLv )/(v - u ) .
/%
For unmeasured lateral inflow q, assume that the component
u ». 0 for the component of the velocity of the lateral inflow q In
the downstream direction, so that
2
(21)
Now a relationship between the lateral inflow q and the slope of the
free wate"- surface s wi 1 1 be obtained with the assumption that there
is a steady state condition for v near the peak of the wave at time t
fixed over the channel reach (peak in the wave in the channel reach x
direction). Also g and f are assumed to be constant and q> 0 .
Near the peak of the wave there are points in the x direction on
either side of the peak where the cross-section areas of the channel bed
150
-------
and also the wetted perimeters are the same. Suppose that at these
two respective points the lateral inflows are q and q , such that
qi.7 V
?\
Let s • s - -p- where s is the slope of the channel bed and
o 3x o
y is the river stage. Let y. and y be the river stage corresponding
to q and q_ respectively. From equation (21)
9(=0 - £')A - Rv2
3Y1 3Y2
Implies that -r— > -r — and since, relative to the positive channel
K 3x 3x
reach x direction, for -^- > 0 the water wave is increasing over the
dX
channel reach as occurs behind the peak over the channel reach and for
T*- <0 the water wave stage is decreasing over the channel reach as
dX
occurs before the peak over the channel reach. Thus it follows that
near the peak, with the above assumptions on v,A,L7 and f» q- > q,
3 1 32
Implies that T— > — , so q_ occurs before' the peak in the wave in
dX dX 2.
the x direction and q occurs after the peak of the wave in the x
direction along the channel reach. With the same assumptions on v,A,L
3 1 32
and f , It also follows that — "> — implies q»> q, . Thus
dx dx 2 1
larger lateral inflow occurs before the peak in the wave in the down-
stream x direction for the same cross-section area A (or Q) occuring
before and after the peak with the steady state assumption on v .
3Y1 3V2 3Q1 3Q2
Generally, for rivers, — > T— implies that — > T— for
oX oX dX oX
corresponding flow rates Q, and Q
Now from the continuity equation (2), where — = -^V^r and with
3 t 31 dA
151
-------
assumption as above on the lateral inflow q it follows that, from
.
3t 3x
dA/
dQ
where — is negat i ve , -rr2 is positive and q is positive, then
Q dx art L
32
— is positive, which occurs after the peak in the positive t
direction in the hydrograph at position x on the channel reach. In
a similar manner to the above for Q. ,
& _L &
q1 3x " dQ 3t
dV
3Q1
where - — is positive, q is positive and for rivers that q. is
10
31 31
smaller than — it follows that — is negative, which occurs before
OX a t
the peak In the hydrograph at position x on the channel reach.
The above discussion indicates that relative to the peak of a
hydrograph at a point x on the channel reach ,the relative amount of
unmeasured inflow q is higher after the peak at x in time t than
before the peak at x in time t .
152
-------
IMPLEMENTATION
In this section a discussion of how the approximations given
by equations (13), 08), or (19) can be implemented so as to compute
Q(x+Ax,t) and C(x+Ax,t), the predicted flow rate and material concen-
tration at a downstream location with channel reach Ax . To compute
both the predicted flow and material concentration knowledge of the
upstream hydrograph from either a given curve or from actual flow and
concentration measurement data is needed. Also knowledge of the
Q vs A curve or measurement data is required both at the upstream
and downstream measurement stations (here assumed to be single valued).
To compute Q(x+Ax,t) it is necessary to obtain an earlier time t .
o
Suppose the upstream measuring station is at x » 0 on the channel
reach. At x,» 0 the initial flow, Q(0,t.) and concentration
C(0,t.) is known at times t1.t2't3 fk (tlmes not necessary to
be equally spaced). The downstream hydrograph at _x - L will be
obtained by first obtaining the predicted hydrograph at point on the
channel reach between 0 and I so that the relative length of Ax
and the channel length L , Ax/L , is less than 1 (the channel reach
length L can be scaled to length 1 if desired). The channel reach
of length L is partitioned into N subintervals (not necessary to be
equally spaced), here of length L/M , so that Ax » L/N.
--- The partition of the channel reach can be chosen so as to locate a
point on the reach from 0 to L pertaining to the mouth of a
tributary at which measured inflow data is known. Relative to the
computed hydrograph from the main branch of the river and knowledge of
153
-------
the hydrograph of the tributary there is a juncture problem at this
location on the channel reach. For a tributary with flow rate not
differing greatly from that of the main stream of interest the
tributary can be treated as a point source.
The Q vs A curves are computed at points x and x+Ax between
0 and L on the channel reach, with the assumption that the river
channel characteristics change slowly from 0 to L so that linear
Interpolation can be used to obtain an approximation to the Q vs A
curves at x and x+Ax (—^- also is approximated from the Q vs A
QA
curves). The needed values of -rj- ™ f'(A), Q, and A for equations
(13), (18), and (19) can then be approximated from the constructed
Q. vs A curves at x and x+Ax on the channel reach.
Now with the channel reach from 0 to L partitioned into N
sublntervals and taking x. - jL/N as a point on the channel reach
where the hydrograph is known, has been computed. Equation (13) then
Is used to compute the hydrograph at x.+1 - (J+1)L/N ; the following
will outline the proceedure. Suppose at x. the predicted hydrograph
has been computed at times t.,t ,t , ....,t, so that Q(x.,t.) ,
Q(x.,t ), ---- , Q(x t ) and C(x t ), C(x t ) ...... C(x t )
J •£ J K J ' J *• J *
are known. To use equation (13) to compute Q(x.+1.t.) where
Q(x .tj), Q.(x.+1,t2), ..., QU..t) have been previously calculated,
it is necessary to first obtain the time t. , f'(A(x.,t. )), and
to j 10
f (A(x..t.)). Time t. can be obtained by iteration and since from
Q(x.,t. ,) to Q(x.,t.) linear interpolation is to be used, it is
154
-------
sufficient to compute A(x.,t.) and At = Ax/f ' (A(x . ,t.)) from
j
the Q, vs A curve at x. , take an approximation for t. as
t. - t. - At then compute A(x.,t. ), At - Ax/f ' (A(x. , t. ) and
10 i J 10 j to
take t, - t. - (At + At )/2 . Note that at t , the first recorded
io i i
measurement time on the hydrograph at x. , to compute t. where
t. < t. , use linear interpolation from Q(x.,t ) to O.(x.,t ) and
extend this straight line approximation to values less than t. so that
t. can be computed as previously.
Once t. is computed then a =• f'(A(x.,t. ) )/2f ' (A(x ,t.)) can
be computed for equation (13) so that Q(x ,t.) can be computed
with a correction to be given for the unmeasured lateral inflow q . To
approximate the term for the lateral inflow, as noted in the previous
section, relative to the hydrograph at x. , the inflow for water is
accumulated when the hydrograph is increasing in Q. for increasing time,
with a small percentage of the acculated inflow' added to the computed Q
and the inflow is treated as in equation (13) when the hydrograph is
decreasing in Q for increasing time. For example, if the hydrograph
at x. begins increasing at time t and is also increasing at time
'mil then the accumulation (storage) Ad at t is Ad • q(x t )Ax
m-M m m m j m
so that //qdsdw/K - rAd and take a new Ad =• (1 - r)Ad . Since
m mm
the hydrograph is also increasing at Q(x.,t ) take
j m+1
*dm+\ " ^dm + q(xj>t:m+1^Ax and rcPeat what was done previously for Ad
In this way much of the lateral inflow, when the hydrograph at x. is
increasing, is stored until the hydrograph changes from increasing to
decreasing in flow rate Q as the time t increases. In the examples
155
-------
of the next section the parameter r was chosen to be .05. Note
that when the hydrograph is decreasing Ad n //qdsdw/K.
Once Q(x.+j.t.) is determined from equation (13) there is
sufficient information to compute C(x. t.). The knowledge of the
Q. vs A curve at the downstream location permits calculation of
t.), A(xj + 1,t,), and a = f ' (A(Xj , t .Q) )/ (f ' (A(Xj ,t .j ) + VQ)
where v - 0_(x t.)/A(x t. ) needed for equation (18) or equation
o J ' ' o J ' ' o
(19). C(x. .,t.) is then corrected for the lateral inflow concentration
using the information with regard to the lateral water inflow. As for
the water above a new C(x.,t ) Is computed from C(x.,t ) as
j m j m
new C(x.,t ) - (Q(x.,t )C(x.,t ) + CA(x.,t )Ad )/(Q(x.,t ) + Ad ) where
J m jmjm jmm J m m
CA(x.,t ) os the concentration of the accumulated lateral inflow.
J m
Then take new CA(x.,t ) a new C(x.,t ) so that the routed water and the
j m j m
i
lateral inflow accumulation is mixed before routing to x+Ax down-
stream. Note that the integral
t x+Ax
/ / C. (s,w)q(s ,w)dsdw
o * '
will be approximated from the upstream data only, and will be taken to
be C. (x ,t)q(x ,t)AxAt in this discussion. For the portion of the
accumulated flow added when the hydrograph is decreasing in Q for
Increasing time t in equation (18) or (19) use //C.qdsdw/D •
C (x. ,t.)q(x. ,t.)AxAt/D . Here q(x t^Ax Is Ad. to be compatable
with the previous notation.
A further discussion of some of the above will be continued in the
next section when the computed results of a flow and concentration are
156
-------
given For a particular example of a reach of the Upper Sandusky River
in Ohio.
STORM EVENTS
In this section the numerical approximations of the mass balance
equations in the integral form (equations (13) and (19)) and the
method of treating the unmeasured inflow was implemented using data
available from the USACOE [1978] on the Sandusky River (Ohio).
The 33 mile river reach from the measurement station near Bucyrus
to the measurement station near Upper Sandusky has a tributary, Broken
Sword Creek, with a measurement station at Nevada. The mouth of the
tributary was about 16.2 miles from the upstream measurement station
at Upper Sandusky. The total river basin area above the Upper Sandusky
measurement station was 298 square miles with 88.8 square miles accounted
for by the Sandusky River basin above the Bucyrus measurement station
and 83.8 square miles accounted for above the Groken Sword Creek
measurement station on the tributary. This leaves 125.^ square miles '
of drainage area with unmeasured inflow. The unmeasured lateral inflow
area along the 33 mile river reach was approximated by a trapazodial area
roughly matching the geometry of the river basin. This total area
then was partitioned for each Ax on the stream reach based upon the
percentage of the trapazodial area which would drain into this Ax
of river. The percentage of available area apportioned to stream reach
Ax was then multiplied by the estimated amount of total available
unmeasured inflow for the unmeasured drainage area to obtain the
157
-------
unmeasured inflow available for the channel reach from x to x+Ax.
This percentage was then used to approximate the value of the integrals
t x+Ax
/ / q(s ,w)dsdw
L J\
o
and
t x+Ax
/ / C.(s,w)q(s,w)dsdw
t X 1
o
needed for equations (13) and (19) .
From the USACOE [1978] data two storm events were considered. For
one storm event, data from 3/1/76 1300 hours to 3/10/76 1900 hours
scaled so the initial time corresponds to time t - 0 and the final
time corresponds to t =» 216 (hour). During the time at the upstream
measurement station (Bucyrus) when the hydrograph was changing most
rapidly; the measurements in flow and concentration (specific conductance)
were available in approximately 6 hour in-tervals. Simular data was
available near the mouth of the tributary (Broken Sword Creek). The
initial upstream data was interpolated linearly to obtain approximate
measurements at three hour intervals from t * 0 to t * 216 . Using
the procedure suggested in (Gray [1973], section seven), the upstream
hydrograph (Bucyrus) was used to estimate the duration of the storm
event (111 hours), beginning at t = 25.5 and ending at t - 136.5 .
For the time period, t = 25.5 to t - 136.5, the intensity of the
rainfall was assumed to be uniform over the unmeasured river basin area.
Since the downstream hydrograph was known a mass balance for water
and and material (conductance) could be used In conjuction with data at
158
-------
the upstream location (Bucyrus) , the tributary, and the downstream
measurement station (Upper Sandusky) to calculate the water and material
inflow from the unmeasured area over the time t • 0 to t - 216.
If the water flow and concentration at the downstream location are
known, an estimate for the unmeasured lateral inflow might be obtained
from the knowledge of previous storm events on the river basin for a
corresponding time of year and storm duration. For this storm, from
t - 25.5 to t - 136.5 the unmeasured Inflow was estimated to be
1»28.2 ft /sec over the entire 125.*» square mile river basin , so
that the integral
x+Ax
q(s,t)ds
X
can be estimated to be ^28.2 times the percentage of area available
from x to x+Ax . Simular, by a mass balance on the specific
conductance C. was estimated to be 389-9 uHHO . To account
for the groundwater inflow during the periods of time 0 to 25.5
and 136.5 to 216 before and after the storm events the uniform
Inflow was estimated to be 70 ft /sec and C. estimated to be
700 yMHO based on available data before and after the storm
event at the downstream location. To account for storage of unmeasured
inflow when the hydrograph was rising, as discussed in the section on
unmeasured inflow, the parameter r was chosen to be .05 (for r in
the range .03 to .08 only small variations were noted in the
location of the peak of the hydrograph at Upper Sandusky). The 33 mile
channel reach was partitioned into ISO subreach distances of length
968 ft, as the choice for Ax
159
-------
The predicted hydrograph is given In Figure 2 for both the
flow and specific conductance as well as the values of the measured
data over the time at the downstream location. The accuracy of the
numerical method was checked by integrating the mass balance over the
whole storm for water predicted at the cownstream measurement station
and finding it to be 0.48 percent higher than the integrated Inflow
data indicated. Using the same procedure , the mass balance for specific
conductance was 1.21 percent higher at Upper Sandusky than the actual
inflow data indicated.
A second storm event over the same river reach has been considered.
This storm event covered the time interval from 2/21/77 1900 hours
to 3/3/77 1300 hours, a total time period of 234 hours. In this
storm the storm duration has been estimated to be 111 hours, from
t - 31 .5 to t - 142.5 .
For the period from t » 31-5 to t - 142.5 the unmeasured Inflow
rate, over the channel reach, was estimated to be 225.2 ft /sec and
the inflow concentration for specific conductance was estimated to be
814.2 viMHQ. During the period from t ** 0 to t - 31.5 the water
inflow was estimated to be 130 ft /sec over the channel reach and
the specific conductance 900 uMHO . Also for the period t - 142.5
to t - 234, the inflow was estimated to be 40 ft /sec for water and
the specific conductance 900 uMHO over the channel reach. The
predicted hydrograph is given in Figure 3 for both the flow and specific
conductance as well as the values of the measured data
160
-------
over the time at the downstream location. The integrated mass balance
for water at the downstream measurement station at 'Ipper Sandusky was
calculated to be 0.7 percent higher than the actual integrated flow
indicated and 2.1 percent higher for specific conductance.
Remark: Both of the above examples were inplimented using FORTRAN
Level G. Also the CPP time for each example was about 10 seconds
on a Amdahl *»70 Model V/7A computer.
SUMMARY AND CONCLUSIONS
An approximation technique based on an integrated form of the mass
balance for water flow and dissolved substances is developed. The
parameters in this method of approximation are determined from knowledge
of the 0_ vs A curve and are functions of the v/ave velocity and/or the
water velocity. In terms of the flow Q. or concentration C this
approximation yields errors of order two; the order of error of the
integral approximation is three.
An input element in the use of these mass balanc'es is the estimation
°f the unmeasured inflow. At a point on the stream reach a relatively
large amount of the unmeasured inflow is added to the hydrogrnph shortly
after the peak in the water flow. This corresponds qualitatively to
the discription of lateral inflow as obtained from the momentum equation
under psuedo steady state conditions.
An algorithm was formulated based in these approximations and applied
to water flow and conductance data from the Sandusky River in Ohio. The
predicted water flow and condictivity at the downstream stations correspond
161
-------
well with the measured values. Over various stream reaches the predicted
total mass of water in a specified period of time was within one percent
of the total mass of the water flov/ing into the strcan reach during the
same period of time. The theorical mass balance on the conductivity
was within two percent.
ACKNOWLEDGEMENT
V/e gratefully acknowledge the CPA Laboratory at Athens, Georgia for
partial financial support for this work. Also v.re appreciate Pr. ^avid
Baker of Heidelberg College, Tiffin, Ohio for supplying the data from the
Sandusky River.
162
-------
REFERENCES
Chen, C.W., Concepts and utilities of ecolog i c model s , J. San. Eng.
Vi.v. ASCE,96, p 1083, 1970.
Cunge, J.A., On the subject of a flood propagation computation method
(Muskingum method), J. Hydnautic. Re4 . , 7(2), p 205, 1969.
Enviro Control Inc., National assessment of trends in water quality,
Report PB-210669, Nat. Techno). Inform. Serv., Springfield, VA,
1972.
Feller, M.C. and J.P. Kimmins, Chemical characteristics of small
streams near Haney in Southwestern British Columbia, (&£eA ReaouA.
Re*., 15, p 2i»7, 1979
Foster, I.O.L., Multlvariate model of storm-period soluble behavior,
J. Hydwtogy, 39, p 339-353, 1978.
Gray, O.M., Handbook ofi -Che P-tx.nctp£e4 o& Hydiology, Water Information
Center, Inc., Port Washington, N.Y., p 7.1-7-2'*, 1973-
Lighthill, M.J. and G.B. Witham, On kinematic waves, Proceedings 0&
thz. RoyaL Science Academy, 229, p 281-316, 1955.
Nash, J.E., A note on the Muskingum flood-routing method, J. o{\ Geo-
Re4., 64 , p 1053-1056, 1359.
O'Conner, D.J., The concentration of dissolved solids and river flow,
ReAoo/i. Re4.,/2 (2), p 279-29V, 1976.,
Peterson, W.C. and F.H. Verhoff, Muski ngum-L I ke approximations for water
routing, submitted 1[)8l
Sandaval, M., Verhoff, F.H. and T.H. Cahill, Mathematical modeling of
nutrient cycling in rivers, Mocfe£oig Bcocne>ncca£ PA.oce44e6 -en
Aquatic Eco4t/A-£em6 , R.P. Canale ed , Ann Arbor Science Publications,
Ann Aroor, Ml, p 205, 1976.
Schultz, N.N. and A. WMmarth, Water quality simulation and Public Law
92-500. case study: Southwestern Illinois, lionet. Re4eoA.cn
, 1978.
Stoker, J.J., IvtteA Ciavei , Interscience Publishers, Inc., New York,
p 452-'456, 1957.
163
-------
Streeter, N.N. and E.P> Phelps, A study of the pollution and the
natural purification of the Ohio River III Factors concerned
in the phenomena of oxidation and reaeration, U.S. Public
RiO-vtin 146, p 75-89, 1925-
U.S. Army Corps of Engineers, Hydrologic Engineering Center,
, HEC-IHD-1200, 1977.
U.S. Army Corps of Engineers, Lake Erie waste water management study,
QuiotUj./ Dcuta., Buffalo District, Buffalo, M.Y., Aug 157^ .
Verhoff, F.H. and O.B. Baker, Moment methods for analyzing river models
with applications to point source phosphorus, accepted for
pup li cat ion is llkteA R&>ouAC.U R
-------
LIST OF FIGURES
Figure 1. - Discharge Versus Area Curves for Stations in Sandusky P.ivcr
Basin at U.S. Geological Survey Stations (l = Samiusky p.iver near Pucyrus;
2 = Sandusky P.iver near Upper Sandusky; 3 = Tymochtee Creek at Crawford;
*4 = Sandusky River near Mexico; 5 = Sandusky Piver near Frcnont)
Figure 2. - Hydrograph and Chemograph at ".S. Geological Survey Caging
Station on Sandusky River near Upper Sandusky, Ohio, for Storm Event
3/1/76 to 3/10/7r>
—' — Predicted Hydrograph near Upper Sandusky
x x x x x Measured Hydrograph near Upper Sandusky
- - - - - Predicted Chenograph near Upper Sandusky
A A A A A 'Measured Chenograph near Upper Snndusky
Figure 3. - Hydrograph and Chenograph at U.S. Ge'ological Survey Gaging
Station on Sandusky River near Upper Sandusky, Ohio, for Storm Event
2/21/77 to 3/3/77
Predicted Hydrograph near Upper Sandusky
x x x x x Measured Hydrograph near Upper Sandusky
----- Predicted Chenograph near Upper Sandusky
A A A A A Measured Chenograph near Upper Sandusky
165
-------
Figure I
12OOO
10OOO -
5OO 1OOO 15OO 2OOO
DISCHARGE AREA IN SQUARE FEET
2500
166
-------
STORM 1/3/76
(M
0)
-750
3000-
O
X
o
z
o
z
LU
o
z
o
u
-500
-250
0
LL
U
I
30
60
90 120
TIME(HRS)
150
1HO
210
-------
ro
OJ
M
3
30
60
90 120
I
TIME (MRS)
150
180
210
oo
------- |