SEDIMENT YRELQS mamas mf m " . soumsam PENWSHLAOFM10HIGAN: . - “ . Thesis for-the Degree of M; S; momma awe unwa‘asm —'JAMES N. WADE . » - 19?! ’ *0? 9173’?“ LIBRAR Y ' mmmmmnm rmmnnu “shim BIVCfSlty 31293 01103 4885 ' I . , ‘Kl: QQ-I'” - r' 3 1 «cl—— D. ”in-:93; 5.; 'I‘ '-:‘““' ~ 1" JUL 181999 ’ 9.2.9; 93117910 1 JULgr-g 79ml ABSTRACT SEDIMENT YIELDS OF RIVERS IN THE SOUTHERN PENINSULA OF MICHIGAN BY James N. Wade Sedimentation is a very important environmental and economic factor in the United States. There are many variables that influence the suspended sediment yield of a river or stream. Some of the more important variables are gross erosion of the soil, the slope or energy gradient of the river, the trap efficiency of lakes and ponds along the watercourse, and the particle size and density of the sediment. It has been shown by various research projects that it is possible to make sediment yield predictions for rivers within the same physiographic area. This research was undertaken to determine a meaningful relation- ship between the sediment yield of a river, and some other easily derived watershed parameter. The research for this thesis has been confined to the rivers of the southern peninsula of Michigan. James N. Wade Suspended sediment yields were computed using data from suspended sediment concentration records, and stream— flow statistics. The results show that there is a differ- ence in the rate of sediment yield between the summer and winter months in some Michigan rivers. A close correlation was established between annual sediment yield, and drainage area, for rivers with watersheds that are in multiple land use. This relationship can be used to predict the sediment yield of locations on rivers where sediment stations have not been established. SEDIMENT YIELDS OF RIVERS IN THE SOUTHERN PENINSULA OF MICHIGAN BY \ ..,_, 3 n. ” James NfrWade A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE Department of Resource Deve10pment 1971 ACKNOWLEDGMENT S To Professor C. R. Humphrys and the members of the staff of the Resource Development Department, I ex- press my appreciation for their encouragement and counsel during this research and throughout my graduate program. Special appreciation is expressed to the United States Soil Conservation Service and the members of the staff who offered to me their technical assistance. Thanks are also extended to my wife, Suzanne, for her patience and understanding that helped make this work possible. ii TABLE OF CONTENTS Page INTRODUCTION 0 O O O O O O O O O O O O l Erosion from Rainfall and Runoff. . . . . 2 Common Methods of Determining the Sediment Yield of Rivers. . . . . . . . . . . 6 Method of Intensive Sediment Sampling and Stream Gauging to Determine the Suspended Sediment Yield of a River . . . . . . . 6 Procedure Using the Universal Soil Loss Equation and Sediment Delivery Rates to Compute Sediment Yield . . . 8 Predicting Sediment Yield by Graphical Analysis of Daily Suspended Sediment Yields and Related Streamflow Parameters . . . . . . 14 Procedure for Sampling Suspended Sediment . . 15 Suspended Sediment Records . . . . . . . l8 Streamflow Records. . . . . . . . 18 Graphical Analyses of Suspended Sediment Concentration . . 20 Graphical Analysis of Sediment Yield Rates. . 23 Predicting Annual Suspended Sediment Yields in Ungauged Rivers. . . . . . . . . . . 35 Comparison of the Sediment Yield Curve to Results of Other Methods. . . . . . . . 38 Use of the Sediment Yield Predictions . . . . 42 REVIEW OF LITERATURE . . . . . . . . . . 44 POSSIBLE AREAS FOR ERROR TO OCCUR . . . . . . 46 SUGGESTIONS FOR FURTHER RESEARCH . . . . . . 48 SUMMRY O O O O O O O O O O O O O O O 5 0 BIBLIOGRAPHY O O O O O O O O I O O O O 5 1 APPENDIX . . . . . . . . . . . . . . 53 iii L I ST OF TABLES Table . Page 1. Daily Yield at Gauging Stations . . . . . 24 2. Sediment Yield Computation. . . . . . . 36 3. Sediment Yields . . . . . . . . . . 37 4. Sediment Yield Predictions. . . . . . . 41 iv Figure 10. 11. LIST OF FIGURES Sediment Delivery Rate Curve . . . . . Suspended Sediment Sampler Device . . . Water Quality Monitoring Stations . . . Streamflow versus Sediment Concentration Curves. . . . . . . . . . . . Streamflow versus Sediment Yield Curves: Watersheds Between 1070 and 6260 Square Miles I O O O O O O O O O O 0 Streamflow versus Sediment Yield Curves: Watersheds Between 1070 and 6260 Square Miles 0 O O O O O O O O I O O Streamflow versus Sediment Yield Curves: Watersheds Between 440 and 892 Square Miles 0 O O O O O O O O O O O Streamflow versus Sediment Yield Curves: Watersheds Between 440 and 892 Square Miles . . . . . . . . . . . . Streamflow versus Sediment Yield Curves: Watersheds Between 80 and 390 Square Miles. Streamflow versus Sediment Yield Curves: Watersheds Between 80 and 390 Square Miles. Drainage Area versus Annual Suspended Sedi- ment Yield Curve . . . . . . . . Appendix Figures--Original Plots of Stream— flow versus Suspended Sediment Yield STA. 1' RaiSin River. 0 I O O I O O Page 12 l6 19 22 26 27 28 29 3O 31 39 53 A-9. A-lO. A-ll. A-12. A-l3. A-l4. A-lS. A-l6. A-l7. A-18. A-19. A-20. A_21o STA. 2, Huron River . . . . . STA. 3, Rouge River . . . . . STA. 7, Black River at Port Huron STA. 4, Clinton River . . . . STA. 5, Belle River . . . . . STA. 9, Saginaw River . . . . STA. 13, N. Clinton River . . . STA. 12, Rifle River. . . . . STA. 15, Thunder Bay River. . . STA. 18, Boardman River. . . . STA. 24, Grand River. . . . . STA. 25, Black River at Holland . STA. 19, Manistee River. . . . STA. 26, Kalamazoo River . . . STA. 28, St. Joseph River . . . Raisin River, STA. 1 . . . . Middle River Rouge, STA. 7. . Upper Rouge River, STA. 8 . . Huron River, STA. 10. . . . L. Clinton River, STA. 12 . . vi Page 53 54 54 55 56 57 57 58 59 59 60 61 62 62 63 63 64 64 65 65 INTRODUCT ION Each year vast quantities of soil are eroded from the land by the forces of falling rain and running water. Some of this material reaches streams and rivers and is carried away as suspended sediment. Suspended sediment can greatly lower the water quality of a stream, making it less useful for fish and wildlife, recreational use, domestic use, agricultural use, and industrial use. Sus- pended sediment, when deposited, can clog rivers, lakes, and harbors. The removal of sediment is very costly, and in many cases the continued use of the watercourse or lake cannot justify the cost of the removal of the sediment. The sediment yield of streams and rivers has been studied extensively in many states. These studies have lead to methods of accurately predicting the sediment yield of rivers in the physiographic area studied. In Michigan there has been extensive stream gauging and water sampling, but a method of reasonably estimating the sus- pended sediment yield of ungauged rivers has not pre- viously been devised. This thesis discusses the process of erosion, transportation, and deposition of sediment. Methods of suSpended sediment sampling, sample analysis, and deter— mining suspended sediment concentration are outlined. The thesis refers to current sources of suspended sediment yield predictions and analyses that have been made pre- viously for Michigan rivers. Using the available data, analyses will be made to show the similarity and differ- ences between the suspended sediment transport charac- teristics of rivers in the lower peninsula of Michigan. A method for predicting the sediment yield of rivers in the lower peninsula of Michigan will be proposed and dis- cussed in detail. Erosion from Rainfall and Runoff There are two main types of erosion affecting the watersheds of Michigan; sheet erosion and stream scour. Sheet erosion begins when the force of falling raindrOps dislodges soil particles. Small soil particles less than 0.074 mm easily become suspended in rainwater. The water eventually begins to flow off of the land in micro- drainageways. The constant agitation of falling rain helps keep the small soil particles in suspension. Eventually the water will be concentrated into progres- sively larger watercourses, and some of the suspended soil will drop out of suspension and be left behind. Sheet erosion rates depend on many factors; soil type, slope steepness and length, vegetative cover, and rainfall intensity. In general, a fine textured soil will erode more easily than a coarse textured soil. Erosion will occur more rapidly on long steep slopes than on nearly level short slopes. Vegetative cover absorbs the energy of falling rain and helps hold the soil in place. Vegetative cover is easily controlled by man and can be an effective tool in preventing erosion. It has been shown that an acre of bare soil can produce 500 times as much eroded soil as an acre of forest land.1 Rainfall intensity is the initial driving force behind sheet erosion. The longer and more intense a rainfall, the higher the rate of sheet erosion. Sheet erosion is the predominant type of sediment producer in Michigan. Sheet erosion rates throughout Michigan are generally low when compared to many other physiographic regions. The soils in Michigan are pre- dominantly medium textured and not highly erOsive. The rainfall intensities are quite low.2 The percentage of the land that is covered by permanent vegetation is high, ranging from about 50 per cent in the southern counties to 90 per cent in the northern counties of the lower peninsula.3 1U.S., Department of Agriculture, Predicting Rain— fall and Erosion Losses from CrOpland East of the Rocky Mountains, Agricultural Handbook No. 282, Agricultural Research Service and Purdue Experiment Station (Washington, D.C.: Government Printing Office, 1965). 21bid. 30.8., Department of Agriculture, "Conservation Needs Inventory," (n.p., 1965). Stream bank and bed erosion is another source of sediment in Michigan streams. Flowing water exerts a force on soil particles, and if the force is great enough it will dislodge the particle. The force of water can be expressed simply as the product of the unit weight of water times the depth of water times the lepe or energy gradient of the stream.4 This relationship is known as the Tractive Force Equation, and is used by engineers to determine the force that will be generated by water flow— ing in canals and drainage channels before they are con- structed. Stream forces become quite complex, however, in natural streams where there are variations in flow, channel orientation, and channel shape. Stream bank and bed erosion is not a large pro- ducer of sediment in Michigan. The stream gradients in Michigan are generally low, therefore the energy gradient is low. It can be seen then in the tractive force relationship that if the energy gradient is low, the force producing erosion will be low. A large percentage of the soil that is moved by sheet erosion does not travel very far. It is deposited where slopes flatten out, in small depressions, in small channels, and is trapped by vegetation. Much of the 40.8., Department of Agriculture, Soil Conser— vation Service, Planning and Design of Open Channels, Technical Release No. 25 (Washington, D.C.: Government Printing Office, 1964). sediment that finally reaches a watercourse is deposited in reaches of slow moving water, or in lakes and ponds. If a particle of suspended sediment is not constantly agitated by turbulent streamflow, it will settle to the bottom and either remain there or continue to be trans- ported as bedload sediment. The physical ability of a lake or pond to retard the movement of sediment is called its trap efficiency. The trap efficiency of a body of water is related to the ratio of the quantity of water retained by the reservoir to the quantity of water flowing into it.5 It is also related to the amount of time a unit of sediment-laden water remains within the reservoir. Streams and rivers in Michigan's lower peninsula have characteristically low gradients and flow through numerous lakes, especially in the upper reaches of their watersheds. Much of the sediment that enters these streams is soon trapped, leaving only the finest sediment in suspension to continue on down the watercourse. 5U.S., Department of Agriculture, Soil Conser— vation Service, Procedure—Sediment Storage Requirements for Reservoirs, Technical Release No. 25 (Washington, D.C.: Government Printing Office, 1968). 6Sterling E. Powell, "Reservoir Sediment Accumu- lations in Southeast Michigan" (unpublished research paper for the Degree of M.S., Michigan State University, 1970). Common Methods of Determining the Sediment Yield of Rivers There are two common methods of determining the suspended sediment yield of rivers and streams. The first method involves daily sampling of water, and daily stream gauging. The second method involves the use of the Uni- versal Soil Loss Equation and Sediment Delivery Rate Curves. Both methods have been useful in determining the sediment storage requirements of reservoirs,7 and in river basin studies. Both methods are capable of making accurate determinations of sediment yields when they are properly used, and both methods have distinct limitations. Method of Intensive Sediment Sampling and Stream Gauging to Determine the Suspended Sediment Yield of a River The United States Geological Survey (USGS) main- tains numerous suspended sediment sampling stations throughout the United States; seven stations are main- tained in Michigan. In general, suspended-sediment samples are col- lected daily with U.S. depth-integrating samplers from a fixed sampling point at one vertical in the cross section. Depth-integrated samples are col- lected periodically at three or more verticals in the cross section to determine the ratio of the cross-sectional distribution of the concentration to the concentration at the daily sampling vertical. During periods of high or rapidly changing flow, samples are taken two or more times throughout the day. 7U.S., Department of Agriculture, Procedure- Sediment Storage Requirements for Reservoirs. During periods of inadequate sampling, daily loads of suspended sediment are estimated on the basis of water discharge, sediment concentrations observed immediately preceding and following the periods, and suspended-sediment loads for other periods of similar water discharge. The estimates are further guided by weather conditions prior to and during the questionable periods.8 Samples of sediment-laden water are filtered and the weight of the sediment suspended in a given quantity of water is determined. Sediment concentration is then expressed as parts per million. Parts per million or ppm is a unit for expressing the concentration of chemical constituents by weight, usually as grams of constituents per million grams of a solution. In the laboratory the results are ex- pressed as weights of solutes in a given volume of water. To express the results in parts per million, the data must be converted. For most waters this conversion is made by assuming that a liter of water weighs 1 kilogram; and thus, milligrams per liter is equivalent to parts per million. Parts per million, for suspended sediment, is computed as 1 million times the ratio of the weight of sediment to the weight of the mixture of water and sediment.9 The quantity of suspended sediment in tons per day that passed the station in the 24-hour sample period is computed next. This is done by multiplying the average streamflow in cubic feet per second times the sediment concentration in ppm times the factor 0.002697. The sedi- ment yield is then determined for the water year in question. 8U.S., Department of Interior, Geological Survey, Water Resources Data for Michigan. Part 2, Water Quality Records (Washington, D.C.: Government Printing Office, 1966), . 7. 9 Ibid. Water year in Geological Survey reports dealing with surface water supply is the 12-month period, October 1 through September 30. The water year is designated by the calendar year in which it ends and which in- cludes 9 of the 12 months. Thus, the year ending September 30, 1966 is called the "1966 water year." 10 The suspended sediment yield for the water year is the summation of the daily yields for that year. A high degree of confidence can be placed On the suspended sediment yield data obtained from this type of survey and analysis. These surveys, however, are very expensive because of the high cost of establishing the station and keeping it adequately staffed. Another problem is that sediment yields on most rivers vary quite a bit from year to year. In order to determine an aver- age annual sediment yield for a river there should be ten years on record for that river. In Michigan there are only seven stations of this type and most of these stations have been in operation less than six years. Procedure Using the Universal Soil Loss Equation and Sediment Delivery Rates to Compute Sediment Yield The sediment yield of a river can be estimated using the Universal Soil Loss Equation which predicts gross erosion, and sediment delivery curves which predict the percentage of the gross erosion that will be carried to a given location on the river. lOIbid. The Universal Soil Loss Equation is used to com- pute the quantity of sheet erosion that is lost annually from cropland. It is an imperical equation developed through twenty years of research by the Agricultural Research Service and the Soil Conservation Service. The main purpose of the equation is to compute sheet erosion losses from cropland. The Universal Soil Loss Equation is:11 A = KRLSCP where A, is the computed soil loss per unit area. Usually in tons per acre per year. R, the rainfall factor, is the number of erosion-index units in a normal year's rain. The erosion index is a measure of the erosive force of Specific rainfall. K, the soil-erodability factor, is the erosion rate per unit of erosion index for a specific soil in cultivated continuous fallow, on a 9 per cent lepe 72.6 feet long. L, the slope length factor, is the ratio of soil loss from the field lepe length to that from a 72.6-foot length on the same soil type and gradient. S, the slope-gradient factor, is the ratio of soil loss from the field gradient to that from a 9 per cent slope. C, the cropping-management factor, is the ratio of soil loss from a field with specified cropping and management to that from the fallow condition on which the factor K is evaluated. 11U.S., Department of Agriculture, Predicting Rainfall and Erosion Losses from Cropland East of the Rocknyountains. 10 P, the erosion-control practice factor, is the ratio of soil loss with contouring, strip- crOpping, or terracing to that with straight row farming, up and down slope. Conversion tables, maps, and charts are published 12 for the factors R, L, S, C, and P. Values in reference for the K factor are found in referencel3. Land cover percentages are determined from land use studies so that a weighted average value for the factor C can be computed for the watershed. Values for S and L can be determined from topographic maps. The K or soil erodability factor is derived by studying soil maps of the watershed to determine the percentage of each soil. The P factor is estimated from the known crop management practices in a given area. Once all of the factors have been determined they are simply multiplied together to compute the gross erosion in tons per acre per year for the watershed. The gross erosion in tons per year can then be computed by multiplying the gross erosion in T/A/YR times the total number of acres in the watershed. Sediment Delivery Rate Curves have been built for several regions of the United States. The delivery rate curve that has been used most widely is one developed by 121bid. l3Powell, "Reservoir Sediment Accumulations in Southeast Michigan," Table VI. 11 John Roehl14 (see Figure l). A delivery rate curve simply relates watershed size to the percentage of total erosion in the watershed that is carried past a specified point as suspended sediment. Sediment Delivery Rate Curves are built by knowing two watershed parameters; the average annual gross erosion from the watershed, and the average annual sediment out- flow from the watershed. The average annual gross erosion in tons is computed by the Universal Soil Loss Equation. The average annual sediment outflow at a selected point in the watershed can be computed from the data obtained by measuring the quantity of sediment that has accumu- lated in a reservoir of known age and trap efficiency. The annual sediment outflow (O) at the selected point is computed using the following equation: where O = Annual sediment outflow in tons. Q = Quantity of sediment measured in the reservoir in tons. TE = Decimal equivalent of the reservoir trap efficiency. T = The age of the reservoir in years. 14John W. Roehl, Sediment Source Areas, Delivery Ratios,_and Influencing Morphological Factors, Publi- cation No. 59, Commission on Land Erosion, International Association of Scientific Hydrology, 1962, pp. 202-13. 12 .m>u=o mumn mum>flawp ucmfiflpmmll.a mHDmHm mmaus unu=Um a menu uuucwnpo coed 06H on A d.o dc.o dqqqqd H d qfiuqd- a r quuddd d a 4.14:4 4 1 qduqqq I 1 :I o I. n e o H m o o u w 11 b m 0 0 0 Q 8 G I 00 o a o 1 o . 1 O .6. o n. .u auKWG _MV 0 1 w .u o .w. mrxw ” 1 o to. an I L n I OH OCH 13 The delivery rate of a watershed is then computed by dividing the annual sediment outflow from the watershed by the annual gross erosion. The curve develOped by John Roehl incorporated delivery rates from five different physiographic areas: Red Hills Physiographic Area, Texas and Oklahoma; Missouri Basin Loess Hills, Iowa and Nebraska; Blackland Prairies, Texas; Sand-Clay Hills, Mississippi; Piedmont Physio- graphic Area, North Carolina, South Carolina, and Georgia. The data from these areas was plotted as the drainage area of the watershed versus the delivery rate percentage. As can be seen from Figure l, the delivery rate plots tra— verse nearly one log cycle. This would be expected be- cause of the several physiographic areas represented in the curve. In order to use the curve one must determine whether to use the high side of the curve, the low side, or the actual curve. Attempts have been made in Michigan to use the delivery rate curve and the Universal Soil Loss Equation to predict the sediment yield of a number of rivers. It seems that the application of this method in Michigan should be done so with care since Michigan represents an entirely different physiographic area than any area represented in the delivery rate curve. The delivery rate curve must also be extrapolated before it can be applied to drainage areas greater than 250 square 14 miles. Some error in delivery rate percentages could result from the extrapolated curve. Many of the rivers in the watersheds that are represented in the delivery rate curve carry much of their sediment load as bedload. Bedload yields as well as suspended sediment yields were thus incorporated in the sediment delivery rate curve. It would seem unwise to apply the delivery rate curve to rivers in the southern peninsula of Michigan that carry most of their total sediment load as suspended sediment. This method, however, has one distinct advantage over the method of daily sampling and stream gauging, it is easy and inexpensive to use. Predicting Sediment Yield by Graphical Analysis of Daily Suspended Sediment Yields and Related Streamflow Parameters It seems that there are sizeable gaps in our knowledge of suspended sediment yields from Michigan rivers. Our knowledge heretofore has been based on the results of imperical computations made from data col- lected in other physiographic areas; or on intensive sampling and stream gauging on a small number of rivers for a short period of time. The problem was to develop a method of using streamflow and sample records that are available to determine annual suspended sediment yields that are realistic. The first step in the solution was to find out what records and data were available. It was found that 15 periodic suspended sediment samples had been taken by the Water Resources Commission (WRC) of the Michigan Depart- ment of Natural Resources on twenty-eight different rivers.15 Records were available from the Department of Natural Resources for the years 1963 through 1968. The Soil Conservation Service had sampled seventeen rivers over a two-year period and these records were available in published form. Daily streamflow records were avail— able from the United States Geological Survey for locations at or near many of the sediment sampling stations. A statistical summary of streamflow data had been published for the gauging stations that had ten years of record or more.16 There seemed to be very adequate streamflow records but only sporadic records of sediment concen- trations. The problem was to find a meaningful relationship between sediment concentrations and streamflow. This proposal will be developed fully in the preceding pages. Procedure for Sampling Suspended SedIment Suspended sediment samples can be easily taken with any device similar to the one illustrated in Figure 2. 15Michigan Department of Natural Resources, Water Qualitnyecords (Lansing, Mich.: Water Resources Com- missian, 1963-1968). 16U.S., Department of the Interior, Geological Survey, Statistical Summaries of Michigan Streamflow Data (Washington, D.C.: Government Printing Office, 1968). 16 .oofl>wp umadfidw udeprm cmodmdmsmlu.m madman .Gw m.oummnah.m0 annuldnn mnszH .No caurhuo.vm m.oa ooma manm «ma ma «non mmla mo ma vmma halo mm m moon omna moma hemumm om Nm oooa aauv noma hamlmw moo Edd muo mcaamemm mum and mum mcaamfimm \ucoa mpaaOm no new» COaumum \msoa mpaaom A no Humx acauoum pamaw popcwmmsm omnonomao mama mama» popcwdmsm mmuonomao mama .mCOaumum msamsmm um mama» maamaun.a Handy 25 Streamflow in cfs was plotted against tons/day sediment yield using the data from each station. The data were plotted on three types of graph paper; full arithmetic paper, semi-logarithmic paper, and full logarithmic paper. It was not possible to obtain a linear relationship on the graph paper used. The re— lationships were curvilinear and seemed to follow certain trends when compared with one another (see Figures 5, 6, 7, 8, 9, 10). Because of the vast range of magnitude of the data, the curves are presented on 3x4 cycle full logarithmic paper. As mentioned previously, the sediment concen- tration for the Saginaw River, Belle River, Lower Huron, and St. Joseph River seemed to be lower during the period of time between November 1 and March 31. For this reason the sediment yield versus cfs for this period of time was plotted with different color pencil than the data from samples taken in the period April 1 to October 31. A distinct relationship was found to exist for each period of time for the four rivers (see Figures 5, 6, 7, 8, 9, 10). The plots for some of the other rivers showed a trend toward this dual relationship, but the total point spread was low enough to assume a single relationship. It can be said with some degree of confidence that the winter conditions of frozen ground and snow cover must reduce sheet erosion and, therefore, the suspended sedi- ment yield of some Michigan rivers. 26 10M 1 UTTF G-Iolmzoo 1 LL11 1 Ton 5- bay I TTFTT 1L111 1 T 1O 18-5t "1030'; 1-51 Joseph 1 VTYT 1L11 I 1 1 1 1 111111 1_14_L111 _1 1111111 1 11 10 100 Cf. 1000 10000 lento ~r11--—- April to Iav.— -— -— Hotel-sheds between 1070 8 6260 SqJ!‘ Your Figure 5.--Streamflow versus sediment yield curves: watersheds between 1070 and 6260 square miles. 27 woo p ,- L- —4 r- —-4 )— -< p— —1 .. .4 r- —1 100 v- -1 Ton : "‘ oy _ _1 - -4 p a - -1 10 p- —. 9 3 L 1 p- 1" lS-Thunder‘ lay " P— —1 1 1 L111111 1 LL11111 1 1111111 1 1 10 100 cf: moo 10000 lav.» ~I‘ll"—- Apt-11 to Iov.— — ._ Duet-Me between 1070 i. 6260 $1.1“ Year Figure 6.--Streamflow versus sediment yield curves: watersheds between 1070 and 6260 square miles. 28 1000 b F .1 h H P A '- -I Z-Hurnn *- / Z-Hurun —4 / l I 100 / #’ /1-n.«.i.rin I -4 F - F — I)”, '- H h d ’- -4 ’- .4 111 he ' —‘ F. I -‘ I *- I I —-1 +- I h “-U.Hurnn .4 y I 1 1 111111 L 1 L11111 1 1 111111 1 L m “’0 mm .0000 c Nuv.1u April"--- 1'. April tn N"V.r——_— Year Figure 7.--Streamflow versus sediment yield curves: watersheds between 440 and 892 square miles. 29 1000 11- _ CHnton j I— -1 p _ b 0.1 P a L .4 3-Rougv *' 124.. c: Intnr. ‘1 1110 b --4 ~ ~ Tn“! ?- H I b IOK It Purl Hur‘nn 4 [’4y F a _ H) ’- -—4 P- -1 r / a L- —I L. .4 F — +- -1 F' 4 ' 1 11111 1 1 L11111 1 1 111111 1 1 1 "'0 mm m. Nuv.tn AIM-1]---- Pf! m H April to Nov.-—- ——— __. Veer Figure 8.--Streamflow versus sediment yield curves: watersheds between 440 and 892 square miles. 3O 10.0 E 1 ~ I b o-t L- a 'z r - lfl : ,5 ”11. -1 E $Lkn' r - I wanna. _ r a P / + IS-Iollc _ I I 4113*: .4 10 p- —4 .4 F2 4 L. .1 C -1 r— »- -1 1 1 1 4 LL 1 1 1 1 1 1 1 1 1 L 1 1 1000 Iov. to bran- 1.0:? to M.—-- — lam. bot-eon I) 8 3’0 «.111. Figure 9.--Streamflow versus sediment yield curves: watersheds between 80 and 390 square miles. 31 I. p- L- I- L "-th ’ muuum P fzr -4 15-le!“le 1” 13.11.13 inton r I F .1 *- 0.” 1. _ r -4 r -1 r- '1 _ J 10 l- A ~ -+ i- d d r D- —1 1. .1 fi- -4 I 1 11111 1 Llllll 1 1111111 11 no of. I. 1000 luv. to ---- with -—-"'" utmmnsmam. k” Figure lO.--Streamflow versus sediment yield curves: watersheds between 80 and 390 square miles. 32 Each graph is more or less the fingerprint of the rivers' rate of sediment transport at a given flow. Mean- ingful relationships between streamflow and sediment yield could not be determined for the following rivers: Che- boygan, Pere Marquette, Muskegon, and the AuSable. Several reasons for this lack of relationship can be assumed. The AuSable and the Muskegon rivers are regu- lated by retarding dams. Retention of floodwater runoff and its later release would distort the relationship be- tween streamflow and sediment yield. The two branches of the Cheboygan River, the Pigeon and the Black, each flow through a large lake before they coalesce to become the Cheboygan. The high sediment trapping efficiency of Black Lake and Mullet Lake would tend to distort the streamflow and sediment yield relationship. The soils in the watershed of the Pere Marquette River are composed primarily of sand. The Pere Marquette River transports most of its sediment load as bedload, which is predomi- nantly sand. This can be readily seen at the river's mouth where the Corps of Engineers annually dredges thousands of tons from the Ludington Harbor. Perhaps the relationship between streamflow and suspended sediment are not well defined for rivers that transport primarily bedload sediments. The relationship between streamflow and sediment yield becomes a qualitative relationship when streamflow 33 rate changes from day to day. A more meaningful expres- sion of sediment yield for a river would be the total mean yield over a given period of time; tons per month, tons per year, etc. In order to derive this expression there should be at least ten year's record of streamflow for the Specific location. A statistical analysis of stream- flow duration can be used in conjunction with the curve relating streamflow to sediment yield to determine monthly or yearly sediment yields. This method eliminates the tedious process of daily sediment sampling and analysis over a long period of time. Streamflow duration statistics are found in Sta- tistical Summaries of Michigan Streamflow Data, 1968.18 A duration table of daily discharge has been determined for each station. "The duration table of daily discharge shows the number of days in each water year during which flow for specified discharges were equaled or exceeded."19 Annual sediment yields for each station were computed using the following procedure: 1. Fifteen to twenty increments of flow, by per- centage of time, were chosen from the flow dur- ation table from 0 to 100 per cent. Example: .5%, 100 cfs; 2%, 90 cfs; 5%, 80 cfs; 10%, 70 cfs. 18U.S., Department of the Interior, Statistical Summaries of Michigan Streamflow Data. lgIbid. 2. 34 A suspended sediment yield for each flow was taken from the curve cfs versus tons/day (Figures 5, 6, 7, 8, 9, 10). Each flow duration percentage was converted to a decimal expression. Each sediment yield value was then subtracted from the next one greater, and that quantity was multiplied by the flow duration decimal equiva— lent for the streamflow represented by the larger sediment yield quantity. The result of this computation represents the average increase in sediment yield per day of a specific streamflow, over the yield at some lesser streamflow. When the values computed in step 4 are summed, the sum represents the mean daily suspended sedi- ment discharge for the station. If a summer and a winter streamflow versus sedi- ment yield curve have been developed for a station, the computation procedure is undertaken for both curves. The values for each season's daily mean sediment yield are then multiplied by a correction factor. These factors are simply the number of days in the year divided into the number of days in the season. The summer correction is 0.583 and the winter correction is 0.416. The summer and winter values are then summed to derive the 35 mean daily suspended sediment yield in tons for the station. 7. The mean daily suspended sediment yield in tons can be changed to the mean yearly yield by multiplying the daily yield by 365. An example of a sediment yield computation can be found in Table 2. A tabulation of the suspended sediment stations and the annual suspended sediment yield at these stations is found in Table 3. Predicting Annual Suspended Sediment Yields in Ungauged Rivers It is interesting and useful to know the yearly suspended sediment yield at a certain point along a river. The question is raised; is it necessary to set up a gaug- ing station and a sediment sampling station at each point where suspended sediment yield data is desired? As pre- viously mentioned, it is possible in some areas to pre— dict sediment yields using the Universal Soil Loss Equation and a Sediment Delivery Rate Curve. These pre— dictions are most valid when applied to the physiographic area from which the data was gathered to derive the delivery rate curve. It would seem possible then to make suspended sediment predictions using data from rivers in the southern peninsula of Michigan. The most desirable type of relationship in any method of predicting sediment yields is naturally the 36 o o o o o o o.H mo.~ m m v.n oom omo moo. omm. a v oa whoa ova mom. on.v m a mm oova ooaa mmo. o~.ma om am oh wham ooha moo. om.oH ma be «HA ooom ooam moo. mh.oa vm an ona ommm ooom moo. mo.ma mo mod mmN vauv oomm mom. om.m we oma oom onmm oomv wow. mv.o om oow oov omen oomm oNH. mm.m om omm o~o ovvo oooo goo. omm. we eon om> v~ooa oomo o~o. Hmmam vo.mm mmm. He mom ode. omo uuucaz vomaa o~.H oomm nuo. onIoN o o o o o o o.H o.- mm ma ow ooo omn moo. mm.m o mm on mnoa ovo mom. ao.om Hm om om mood ooHH mmm. no.av mm moa ooa onaw oohd moo. oo.m~ we mma oom moon ooaw moo. no.m~ co oHN ohm ommm ooom ovo. om.m~ vo oo~ oov v-v oomm mom. oo.o ov own omm oomm oowv ~om. oo.o no hon omo omoo oomm m~a. om.m co «me ooh mvvo oooo Hoo. m~.a vo mow omo vNooH oomo o~o. Hmmam onooo an.mma o~.a om mum mom. ovm umaasm vomHH o~.H oomm mac. nemnow um\a Huuoa mmmuo>¢ .ucH womuw>¢ cemwwm umm> uo x .ocH xmo\e .uou mwu .uou coflumuso Hmuoa mam» uo umm» cammmm .uso .uou cemmwm >mo\9 cammom .uou .<.o muo 30am :OMumum m no waxe smoxe .cofiumusmaoo pauwx acufiwpum11.~ mqm