A sorting mechanism for a riffle-pool sequence

Geological Society of America Bulletin, Part 11, v. 90, p. 1142-1 137, 3 figs., Jirly 1979, Doc. no. M90703.

actions among coarse particles in motion INTRODUCTION in turb'ulent flow tend to concentrate

Tcansport of coarse, heterbgeneous them in groups with the qoarsest particles

debris MI a .natural stream' under a .wide at the surface (Langbein and Leopold, 1968).

range of flows usually results in a re; Incipient accumulations of coarse particles

markably stable, undulatory bed prof i.le', may be perpetu'ated 'by altering the flow

which manifests an in transit sqrting conditions which influence bed load trans-

process of the bed' material. In general, port.

finer material representative of the bulk' As flow increases, submergence of the

of the Dormal bed.logd resides in the riffle-pool bed topography modifies its

deep sections, or pools, below flood eff'ect on the*local hydraulic cbnditions.

stages. 'At highl'flows, pools may scour At low flow, mean volocity and water sur-

to immovable boulders or be'drock. face slope are greater, and mean depth

coarser material transportsd at more in- is less ata riffle than at a pool. Com-

qrequent flows forms the shallow sections, petence is greater at the riffle. As the

or riffles. Above a flow threshold, the stage increases, the water-surface profile

pool fill material.is often scou;ed and tends to even out as the hydraulic gra-

deposited i3 part over the riffles (Lane dient over a riffle decreases and that

and Borlan,d, 1954; .Andrews, 1977). Inter- over a pool increases (Leopold and others, 21 42

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/90/7_Part_II/1142/3429590/i0016-7606-90-7-1142.pdf by guest on 01 October 2021 1 43 c 1964). Corresponding vaiues of velocity those at a riffle. Andrews (197.7) shows

and depth tend to converge, although depth convergence of values of mean velocity

often less so (Richards, 1976; Lisle, with increasing discharge at sections

1976; Andrews, 1977). The convergence of comparable to pools and riffles until at

5. respective values of water-surface slope bankfull flow the velocity at some scour-

over a riffle and pool and the greater ing (pool) sections exceeds that at some

U depthpf the pool cause mean- shear stress, filling (riffle) sections. Under Keller's yRS (where y is the specific gravity hypothesis, coarse clasts entrained above T E

of water, R is the hydraulic radius .-or the reversal flow are more,likely to be

approximately the mean depth, and SE is deposited at a riffle than at a pool.

the energy gradient), to increase more Finer material transported below the re-

r'apidly' at the pool (Leopold an> others, versal flow tends to collect in the pool.

1964). As a result, competence as meas- Measurements of velocity near the bed

ured by velocity or bottom shear stress may show relative differences in compe-

should become more evenly distributed. tence between local bed areas. It is

over a riffle and pool (Kichards, 1976) difficult, howevef-, to translate point

or even reversed in hierarchy at high yelocity directly into a measure of com-

flow (Gilbert, 1914; Keller, 1971). petence, for which shear stress is the

Keller (1971) proposed the hypothesis most widely accepted parameter (Vanoni,

of "velocity revA-salll to explain-"tEe 1975). Several researchers have used

areal sorting of bed materials in a velocity profiles to calculate local shear

riffle-pool sequence by the reversal in stresses, but the technique is difficult

hiera'fchy of bottom velocity values. pis under some field cond'itions (Church, 1972; h data from Dry Creek, California, for four Smith and McLean, 1977). Measurement of

stages show the convergeace of mean values mean shear stress from the flow geometry

at the%ighest stage, above which bottom is a Rcactical method for quantitatively I velocities at the pool presumably exceed relating the sort3ng of grains to the -

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/90/7_Part_II/1142/3429590/i0016-7606-90-7-1142.pdf by guest on 01 October 2021 1 14

distribution of competence over reaches of several associated studies of river

as long as riffles or pools. hydraulics and sediment transpbrt (Bagnold,

This paper describes a sorting mechanism 1977; Andrews, 1977; Bennett and Nordin,

based on the reversal in hierarchy of 1977).

mean shear'stress at a pool and riffle. In this study reach, the channel is

Measurements of shear stress and grain formed in glacial outwash providing the

size distribution at a pool and riffle material for a basal gravel bed over

help substantiate this mechanism and. which coarse sand mostly derived from the

quantitatively demonstrate the operation early Tertiary Wasatch Formation is trans-

of the sorting process in one stream. ported as bed load. At low flow, the sand

Comparison with data from other streams tends to smooth the riffle-p?ol bed pro- 3 reveals some influences of sediment sup- file. At bankfull flow, about 23 m /s,

ply and the interactions of heterogeneous. the channel is about 18 m wide and 1.2 m

bed-load particles. deep; de velocity is about 1.1 m/s. The

average water surface slope over 1.5 km of PRESEARCH DESIGN channel length in this vicinity is *0.0007.

Data for this study were gathered in A 50-m-long riffle and a 90-m-long

the conth of a larger study (Lisle, ~ pool were chosen for study with the cri-

1976) of fr ctional resistance of a teria that they typify these channel forms,

riffle-pool-bend 'section of the East Fork are straight, and are near each other.

River, an upper tributary of the Green The intervening reach, 1.0 km long, is not

River'in western Wyoming. The site was joincd by any tributary. At low flow, the

chosen by Luna B. Leopold for direct bed slope of the,riffle is 0.0014 and of

measurement of total bed-load transport the pool is 0.0004.

by a bed-load trap. Data on that phase ' The data include daily measurements of

have been published (Leopold and Emmett, water-surface slopes, hydraulic radius it

1976, 1977), and the present paper is one each reach, and,percentage of sand-covered

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/90/7_Part_II/1142/3429590/i0016-7606-90-7-1142.pdf by guest on 01 October 2021 area of the pool bed. The water surface spring runoff.

slope, S, is the measure of the energy SCOUR AND FILL OF THE POOL AND RIFFLE gradient, assuming that the progressive ., variation of velocity with distance at The surface of the riffle bed is gravel

each reach is insignificant in deterrninihg .(&go = 46 mm, b-axis aiameter coarser than

The riffle channel is uniform relativc 90% of the bed; from a pebble count; after SE. -_ :.- , to other reaches; the pool channel con- Wolman, 1954) except at high flow when

stricts at about mid-section. Maximum. sand may overlie one-third of the channel.

error in the use of S for S because of Coarse sand repres.entative of the bed E

backw5ter effects at the riffle is esti- load (d5o = 0.5 to 1.5 mm) covers most

mated to be about 10%. Values of S and of the ba- gravel bed of the pool at

R were measured from cross sections, and low'flo$ to a thickriess under 1 m, but

arrays of surveyed bank pins were used to scours at high flow, progressively ex-

gauge local water-surface elevations. posing the gravel. From periodic sound-

The bed-surface composition of the pool' ings of cross sections within the study

was monitored by longitudinal sounding reach, Andrews (1977) show@ that most

transects run in*a boat equipped with a' scouring sections begin to scour at about 5 Model 1054 Ultrasonic Distance Meter and bankfull'stage, below which they fill.

chart recorder. The bottom profiles dis- The percentage of the pool bed covered

play the presence or absence of.sand in with sand began to decreqse'at discharges 3 bed forms and indicate changes in bed between 11 and 21 m /s (Fig. 1). Initial

elevation. Data were collected once scour, assumed to be marked by the in-

daily at gach reach every day except one creased exposure of the gravel bed, es- '- during the period from June 2 to June 26, sentially began no later than bankfull '3 1975. During this period three hydro- flow (Qbf = 23 m /s) and at progressively

graph peaks exceeded .bankfull stage. The lower flows during the rise to subsequent

first p,eak marked the beginning of major hydrograph peaks. A marked clockwise

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/90/7_Part_II/1142/3429590/i0016-7606-90-7-1142.pdf by guest on 01 October 2021 1146

I I I I -

X-Scour begins

I 1 I 1 I1 I 7 lo-. 15 20 30 D isch a rg e ( m3/ s)

Figure 1. Percent of the area of the pool bed covered

with sand ver'sus discharge. Qb, is bankfull discharge.

Lines connect consecutive flows. Three Line patterns

correspond to the rise and fall of three hydrograph'peaks.

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/90/7_Part_II/1142/3429590/i0016-7606-90-7-1142.pdf by guest on 01 October 2021 I L

hysteresis expresses the,lag in redeposi- for example, the remaining sand tends to

tion of-sand following each peak. be concentrateq in the downstream portiori

of the pool, relatively elevating the bed SHEAR STRESS REVERSAL there and maintaining a slightly greater

Figure 2 shows' the variation of hy- roughness (Lisle, 1976). This relative

draulic radius, water-surface silope, and elevation tends to reduce water-surface

mean shear stress with discharge at the slope, but this result is subordinate to

pool and riffle. The anomaly of the data the over-all increase of slope with dis-

points for the first four days (points charge at the pool. Different bed con-

marked- 'lx" and "0" in Fig. 2), partic-u- figurations at equal discharges (Fig. 1)

larly for slope values, resulted from the ' thus complicate the slope-discharge rela-

rapid adjustment of 'the channel during t ion.

the rise of the first hydrograph peak, Values of mean shear stress converge

when bed-load discharge was highest. over a range of discharges between 12 and 3 '2 The pool's values for hydraulic radi'us 16 m /s at T values between 5 aqd 7 N/m

are consistently greater than the riffle's (Fig. 2). This interval of shear stress.

but jncrease at a slightly lower rate values is designated T and the corre- r with discharge. sponding interyal of discharge values is

For each reach, water surface slope designated Q r . varies nonlinearly with discharge. The Significantly, Q falls within the r sets of values converge at less-than- range of the onset of scour at the pool.

bankfull discharge at a value roughly Keller's (1971) bottom velocities con-

equal to or greater than the slope of the verged at a'flow of, a 1.2-yr recurrence - over-all reach, S = 0.0007. In addition intdrval. Because Q represents less- r to measurement error,' the scatter of than-bankfull discharge, which often has

points results from variations caused by a recurrence interval of abdut 1.5 yr

scour and fill of the sand. At' high flow, (Leopold and others, 1964), the flow

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/90/7_Part_II/1142/3429590/i0016-7606-90-7-1142.pdf by guest on 01 October 2021 1148

2.0

1.5

1.0

.7

.5

.4

1s Figure 2. Valuss of hydraulic

radius, water surface slope, and mean E - .0010 Q P shear stress versus discharge for -0 v) .0007 the pool and riffle. The anomalous U 0 Y L' ' first four data points are designated 3 .0005 - L I I1 I1 Q "0". * x and S is the average slope 0 oo 00 -- I 3 over 1.5 Q is bankfull dis- km; bf .0003 charge; T and Q are, respectively, P r mean shear stress and discharge at - 10 n E reversal. \ Z - 7 v)

-2

Discharge (ml/s)

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/90/7_Part_II/1142/3429590/i0016-7606-90-7-1142.pdf by guest on 01 October 2021 1 49

frequencies at reversal for the two studies the riffle and pool is conjectural. Local

seem similar. shear stress may be appreciably higher

Above Q values of T in the pool ex- than the mean over irregular cross-sections. r' ceed those over the riffle. We can.expect Because of a more uniform distribution of

that the difference will become no greater velocity and depth, variation of time-

at still higher stages if values of S do averaged shear strecis over the riffle

not depart systematically from S and the should be less than over the pool. Shal-

R versus Q relations remain linear. lower depths in the riffle, on the other

hand, should praduce greater vertical AREAL SORTING velocity gradients and greater instantan-

Grain size distribution of the bed eous shear stresses than those in the

load changed little with flow. The per- ~001.~Church (1972) showed that imbrica-

centage of the coarsest size fraction of tion and close packing of particles in a

bed load sampled at a flow-slightly above stream bed inhibits entrainment, so that 3 bankfull (Q = 23.8 m 1s) exceeded that the Shield's (1936) criterion for incip- 3 for a moderate flow (Q = 10.6 m /s) by__ ient motion defines a lbwer limit. Possible only a few percent%t most (Fig, 3). Up differences in the bed structukes of the 3 to the highest sampled flow (Q = 45.0 m 1s) two reaches may thus cause differences /, I the size distribution was essentially un- in entrainment conditions. With only

changed. qualitative assessment of these factors,

Using threshold criteria, the maximum it seems valid to use the values of'T r' grain size theoretically entrained, at without adjustment for local hydraulic

L T~ = 7.N/m is 8 mm (Vanoni, 1975, p. 96, conditions, ,to infer the sorting mechanism.

after Shields, 1936). This size, desig- The value of d is fioer than 99% of r nated dr, is plotted (Fig. 3). How cT se the riffle gravel and coarser than 89% of _- s- ly the average entrainment conditions the bed load or the fill material covering

appkoximate the conditions on th.6 beds of the major part of the pool bed up to the

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/90/7_Part_II/1142/3429590/i0016-7606-90-7-1142.pdf by guest on 01 October 2021 1150

I I I ,y-----' /+7 f

Bedload ./

Entrained

0 .2 .4 .7 1 2 4 ,7 10 20 ' 40 70' 100 Sieve size (mm)

3 Figure 3. Grain size frequency for two bed-load samples at low flow (Q = 10.6 m /s) 3 and high flow (Q = 23.8 m /s) (Leopold and Emmett, 1976) 'and for'the surface of the

riffle bed from a pebble count (after Wolman, 1954). The bed-load sizes represent

the pool fill material. Grain size, dr, theoretically is entrained at the upper

value of T r' the mean shear stress at reversal.

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/90/7_Part_II/1142/3429590/i0016-7606-90-7-1142.pdf by guest on 01 October 2021 1

reversal stage. Keller (1971) stated, of tractive forces exists over a meander

"The size of the_ largest bed material be- bed at a given flow than in most straight

neath the fines in pools is dependent on reaches (Hooke, 1975) and thus provides

the revecsal velocity." More precisely, the hydraulic environment for the depo-

this grain size is

The coarser riffle gravel would be en- In the East Fork, gravel transport, as

trained at higher stages, and an excessive described by this inferred sorting mech-

tractive force would then exist in the anism, was inhibited by armoring of the

pool to convey the gravel to the next riffle bed, tfig distribution of transport

riffle or bar. As Keller (1971) argued, capacity, and the supply of sediment.

the riffle-pool form is maintained by the Consequently, gravel was not observed in

transient deposition of coarse clasts at quantity in the bed load up to more than

the riffles during flows above the rt- twice bankfull flow. Small pebble gravel

versa1 stage. comprised the uppermost size limit. of the

Small pebble gravel (4 to 8 mm), ap- bed load, in part because the coarser

proximating d is the deficient size riffle gravel was protected from transport r' fraction of the surface deposits of the by armoring and imbrication, which were

riffle and-pool, presumably because it is ineffective for the finer pool fill. Con-

a rare fraction of the introduced bed centration of coarse clasts tends to in-

load (Andrews, 1977), and because dis- hibit their transport (Langbein and Leopold,

persive stress and winnowing concentrate 1968).

coarser gravel fractions at the surface . Emmett (1976) demonstrated the effect of

of the riffle. As a surficial deposit armoring on gravel transport in the Snake

over the study reach, this size occurs and Clearwater Rivers. Until much of the

locally over point bars. A wider range cobble pavement of the bed was entrained,

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/90/7_Part_II/1142/3429590/i0016-7606-90-7-1142.pdf by guest on 01 October 2021 I 52

these rivers transported coarse sand, corresponding to about 0.75 bankfull depth

which was finer than the material which (Leopold and others, 1964). Similarly,

the streams are competent to carry. Only painted particles representative of the

when particles greater than 32 nun (the bed sizes of Dry Creek were eroded during

average median size of the bed material a flood with an instantaneous discharge

within 0.6 m of the svrfacej'were en- of a.l.5-yr recurrence interval (Keller,

trained, did the intermediate sizes appear 19?0) .

in quantity in the bed load. This occurred Twenty-four millimetre gravel, the

at a recurrence interval of from 1 to 2 median size of the gravel surface over

yr, or approximately at bankfull flow. the riffle, is theoretically entrained at 2 Below this, armoring limited the supply a shear stress of about 19 Nhn (after

of finer grains to the bed load and sig- Shields, 1936): This requires a dis-

nificantly reduced '@he transport efficiency charge which is over twice bankfull, our

Milhous (1975) found that the armor upper limit of observation, and much less

layer in Oak Creek. on the east edge of frequent than at the gravel stream cited

the Oregon Coast Range broke up when the above. Thus the gravel bed appears to be

d30 size was entrained. This occurred essentially static for flows up to those

at discharges exceeding 3.5% to 1.4% of of low frequency and high magnitude.

the time over 2 yr: These flows are more During high flow, the partial blanket

frequent than bankfull flows in English of sand over th& riffle further protects

rivers, which are exceeded on che average the gravel from t\rac"tive forces. The scour

0.6% of the time (Nixon, 1959). and fill of sand in the pool and riffle

Data from bther gravel streams also result from variation in distribution yf

show general entrainment of the bed sizes transport capacity (Andrews, 1977).

near bankfull flow. At Seneca Creek, The lack of supply of gravel from sources

Maryland, painGed pebbles of the median outside the stream was perhaps the predom-

size on a riffle were eroded at a stage inant inhibiting factor on gravel transport.

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/90/7_Part_II/1142/3429590/i0016-7606-90-7-1142.pdf by guest on 01 October 2021 I :2

Sand is the dominant Size supplied to the Although a stream's competence steadily

East Fork (Andrews, 1977). If gravel were increases with stage, the size,of bed load

introduced in abundance, the channel would material is Limited by supply. The protec-

adj us t C~Q&generate the necessarx transport tive assemblage of gravel deposits in

competence, and the riffle-pool configura-' stream beds inhibits entrainment of all of

t i6n with .the ' a t tend ant reversa 1 mechanism the material held within. Finer material,

would probably respond to more actively the largest size of which is related to

sort the coarser sizes. threshold conditions at the reversal flow,

L is avaifable for transport in pools at low SUMMARY AND CONCLUSION to moder'ate floks. At some stage above

Shear 'stress "reversal," equivalent to reversal, bed load grain size may incrkase

Keller's (1971) hypothesis of velocity abruptly when the coarser areas of the

reversal, 'offers a plausible mechanism bed, such as riffles, are entrained.

for the sortihg of bed material in the On the East Fork, there was a large

riffle-pool sequence of gravel streams at difference between the discharge at the

less than extreme flows. By either mech- reversal in competence and that required

anism (shear stress or velocity.reversal), to entrain the graeel in the bed. Pre-

the competence of flows in pools exceeds sumably the gravel framework of the bed

that over riffles above a certain range has become static without the infusion of

of flows and' thus tends to concentrate new gravel from sources other than the

coarse material in riffles. Data from1 bed. In more active gravel channels, this

the East Fork demonstrated the reverSal difference is smaller, as-host of the

in hierarchy of shear stress values at a gravel is set in motion near bankfull

riffle and pool. Deposition of gravel on stage. 'In either case, the switch in

riffles was not observed, but some influ- hierarchy of competence in a riffle-pool 5 ences of sediment supply on this sorting sequence near bankfull flow occurs in time

mechanism have been inferred. to favor deposition of the coarsest bed

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/90/7_Part_II/1142/3429590/i0016-7606-90-7-1142.pdf by guest on 01 October 2021 1 - I 14

7 load material on riffles and bars. The of .equal frequency in central Texas

reversal mechanism operates on the size locally Scoured pools and deposited coarse

of the supplied bed load. The bed topog- particles over bars, thus enhancing the

raphy. determines the local hydraulic con- form amplitude (Baker; 1977). Most likely

ditions for bed load transport and areal the causative difference between these

sorting of grain sizes, however, and may two examples is the greater relative vol-

be formed of co_a_rsermaterial of a dis- ume of sediment hontributed to the Cali-

tinctly different transport regime. fornia stre’ams from their highly e-fodible

,Of the elem&nts of the form-process re- baG^ins. The diminution of the riff le-pool

lation which govern ‘the dynamic equilib- . sequence in these streams possibly manifests

rium of riffle-pool morph,ology, the most a decrease in channel roughness in response

critical independent factor seems to be . to the higher- sediment load. A threshold

the quantity and grain size of the sediment magnitude of sediment input effectively

supply. In this respect, the East Fork perturbing the dynamic equilibrium of

represents a case of distinct transport gravel smeams, as suggested by these two

regimes for sand and gravel dependent on examples ,- warrants investigation.

flows of different magnitude and frequency. ACKNOTJLEDGMENTS The sediment.rey.mes of other streams may

also complicate or. perturb the simple The U.S. Geological Survey provided

sorting mechanism presented here. For field eqqipment, collected water an-d sedi-

instance, channel erosion and deposition ment- discharge data, \and sounded the riffle

of poorly sorted dkbris associated with cross section. Luna B. Leopold suggested

the flood of December, 1964, in northern the original idea. He and Robert Myrick

D California obliterated or diminished the gave helpful addice for data-collection

riffle-pool sequence of tome streams procedures. Victor Baker, William Emnett,

(Stewart and LaMarche, 1967; Kelsey, 1977, Roger Hooke, and Edward Keller critically

p. 283). In contrast, la-kge flood flows. reviewed the manuscript. Michael Church

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/90/7_Part_II/1142/3429590/i0016-7606-90-7-1142.pdf by guest on 01 October 2021 I 55

offered additional comments. Denver, Colorado, March 22-25, 1976,

p. 4-101 to 4-114.- REFERENCES CITED Gilbert, G. K., 1914, The transportation '

Andrews, E. D.; 1977, Hydraulic adjustment of debris *b,y running water: U.Si Geo-

of an alluvial channel. to the supply logical Survey Professional gaper 86,

. of sediment. [Ph.D. dissert.] : Berkeley, 363 p.

Califohia, University of California, Hooke, R. L., 1975, Distribution of sedi-

Berkeley, 152 p. ment transport and shear stress in a

Bagnold,, R. A., 1977, Bed load transport meander bend: Journal of Geology, v.

by natural rivers: Water Resources Re- 83, no? '5, p. 543-565.

search; v. 13, no. 2, p. 303.312. KellFr, E. A., 1970, Bedload movement ex-.

Baker, V. R., 1977, Stream-channel. response periments: Dry Creek, California:

to floods, with examples from Texas: Journal of SeMimentary Petrology, v.

Geological Society of America'Bulletin, 40, no. 4, p. 1339-1344.

V. 88, p. 1057-1071. 1971, Areal sorting of bed-load material:

Bennett, J. P., and Nordin, C. F., '1977, ,The hypothesis of .velocity reversal:

Simulation of sediment transport and Geological Society of America Bulletin,

armouring: Hydrological Science9 Bul; V. 82, p. ,753-756.

letin, v. 22, no. 4, p. 555-569. Kelsey, H. X., 1977; Landsliding, channel

Church, M., 19'72, Baffjn Island sandurs: changes, sedimekt. yield, and land use'

A study of Arctic fluvial processes:. in the Van Duzen River Basin, N. Coastal

'I Geological Survey of Canada: B,plletin California, 1941-75 [Ph.D. dissert. 1 : ,

216, 208 p. Santa Cruz, California, University of .

Emmett,, W. W.', $976, Bedload transport California, 370 p.

in two large, gravel-bed rivers, Idaho 'Lane, E. W., and Borland, W. Id., 1954;.

and Washington, in 1976 3rd Federal' , River-bed scour during flood?: American

.c Interagency Spd'ime'ntation-ConfSerence,

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/90/7_Part_II/1142/3429590/i0016-7606-90-7-1142.pdf by guest on 01 October 2021 1 56

V. 119, p. 1069-1080: State University, 232 p.

Langbein, W. B., and Leopold, L.-B., 1968, Nixon, M., 1959, A study of the bankfull

River channel bars and dune2 - 'Theory discharges of rivers in England .and

of kinematic waves: U.S. Geological (wales: Institute of Civil Engineering

Survey Professional Pager 522-L, 20 p. Proceedings, Paper no. 6322, p. 157-174.

Leopold, L. B., and Emmett, W. W., 1976, Richards, Ic. S., 1976, The morphology of .I .I Bedload measurements, East Fork River, ciffle-pool sequences: Earth Surface

Wyoming: Proceedings of the National Processes, v. 1, p. 71-88.

Academy of Science, U.S.A., vol. 73, Shields, A., 1936, Arwendung der Aenlich-

p. 1000-1004. keits-mechanik und der- Turbulenz-

Lep.pold, L. B., and Emmett, W. W., 1977, forschung auf die Geshienbebewegung:

1976 bedload measurements, East Fork Mitteilungen * de; Preussischen Versuch-

River, Wyoming: 'Proceedings of the sansta'lt fur Wasserbau and Schiffsbau,

National Academy df Sciences, v. 74: Berlin, Heft 26, 26 p.

nb. 7, p. 2644-2648. Smith, J.'D., ahd McLean, 5. R., 1977,

Leopold, L. B., Wolman, M. G., and Miller, Spatially averaged .flow over a wavy,

J. P., 1964, Fluvial processes in geo- surface: Journal of Geophysical Re-

morphology: San Francisco, W. H. search, v- 82,,no. 12, p. 1735-1745.

FreLrnan and Co., 552 p. Stewart, J. H., and Laarche, V. C., 1967,

Lisle, T. E., i.976, Components of flow Erosion and deposition produced by the

resistance in a natural channel 1Ph.D. flood of December 1964, on Coffee Creek,

dissert. 11: Berkeley, California, Trinity County, California: U.,S. 5eo-

University of California, Berkeley, logical Survey Professional Paper 422-K,

60 P! 22 p. Milhpus, R:T., 1975, Sediment transport Vanoni, V. d, ed., f975, Sedimenkation

in a gravel-bottomed stream [Ph.D. engineering: New York, herican Society

dissert-.]: Corvallis, Oregon, Oregon of-Civil Engineers, 745 p.

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/90/7_Part_II/1142/3429590/i0016-7606-90-7-1142.pdf by guest on 01 October 2021 Wolman, M:.G., 1954, A method of sampling

coarse river-bed material: American

Geophysical Union Transactions, v. 35,

no. 6, p. 951-956.

MANUSCRIPT RECEIVED BY THE SOCIETY

NOVEXBER 9, 1978

REVISED MANUSCRIPT RECEIVED MARCH 30, 1979

MANUSCRIPT ACCEPTED APRIL 2, 1'979

Printed in'U.'S.A.

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/90/7_Part_II/1142/3429590/i0016-7606-90-7-1142.pdf by guest on 01 October 2021