Western University Scholarship@Western

Geography Publications Geography Department

6-1988 Effective Discharge for Suspended Sediment Transport in Streams of the Saskatchewan Basin Peter Ashmore University of Western Ontario

T J. Day Sediment Survey Section, Water Survey of

Follow this and additional works at: https://ir.lib.uwo.ca/geographypub Part of the Geography Commons

Citation of this paper: Ashmore, Peter and Day, T J., "Effective Discharge for Suspended Sediment Transport in Streams of the Saskatchewan River Basin" (1988). Geography Publications. 294. https://ir.lib.uwo.ca/geographypub/294 WATER RESOURCESRESEARCH, VOL. 24, NO. 6, PAGES864-870, JUNE 1988

EffectiveDischarge for SuspendedSediment Transport in Streams of the Saskatchewan River Basin

P. E. ASaMO•

DepartmentofGeography, University ofSaskatchewan, Saskatoon, Saskatchewan, Canada

T. J. D^Y

SedimentSurvey Section, Water Resources Branch, Inland Waters/Lands Directorate, Environment Canada, Ottawa, Ontario

Effectivedischarge for suspended sediment load was determined for 21 sites in theSaskatchewan River basinat whichsediment records range from 5 to 29years in length.The drainage areas for these streams rangesfrom 10 to over 300,000 km•. Thesediment discharge histograms havea varietyofforms ranging fromthe classic unimodal form in whichthe peak occurs at dischargeswith a durationof 1-3%to those in whichthe effective discharge isthe extreme event of recordand cases in whicha singleeffective dischargeis difficult to define.The percentage duration of the effective. discharge ranges from less than 0.I % toover 15%, a greaterrange than previously hasbeen reported. There is an obvious tendency for thepercentage duration of theeffective discharge to increase with drainage area and hence downstream throughthe drainagesystem.

INTRODUCTION menttype is likelyto be manifestedin theeffective discharge One of the tenetsof fluvial geomorphologyis that while for bed load being a higher magnitudeevent than that for sedimenttransport occurs over a widerange of discharges,the suspendedload, becausethe threshold of motion is much flowswhich, over the longterm, accomplish the most work are higherfor coarserbed load particles than for finersuspended thoseof moderatemagnitude, close to bank-full,typically re- sediment[Baker, 1977; Richards, 1982, p. 145]. curringonce or twice/year[Wolman and Miller, 1960].The Therehave been few attempts to confirmthe validity of the notionof "workdone" (sediment transported) by eventsof a effectivedischarge concept for fluvialsediment transport and givenmagnitude and frequency is distinctfrom, but related to, to establishthe magnitude of variation in its duration. Benson the effectivenessof these events in controllingchannel form. and Thomas[1966] showedthat histogramsof total load Wolmanand Miller [1960]suggested that in humidtemperate transportedby incrementsof the dischargerange couldbe environmentsthe effectivedischarge for sedimenttransport used to identify the effectivedischarge. Their data indicate wouldbe very similar in magnitudeto thedominant discharge that the average percentage duration of theeffective discharge controllingchannel morphology. for streamswith drainageareas ranging from 584 to 357,000 Wolmanand Miller [1960'1proposed that the amount of km2 is approximately!2%. Theseare morefrequent dis- sedimenttransported by flowsof a givenmagnitude depends chargesthan thosesuggested by Wolmanand Miller [1960]. on the form of the relationshipbetween discharge and sedi- The fact that similarquantities of sedimentwere transported mentload and on thefrequency distribution of thedischarge by a wide rangeof dischargesalso lead Bensonand Thomas events.The productof transportrate and frequency gives the to questionthe validity and applicabilityof the effectivedis- cumulativesediment load transportedby a givendischarge. charge concept. In particular,they pointed out that the ef- Given a lognormaldischarge frequency distribution and a fectivedischarge for suspendedsediment transport was a powerrelationship between discharge and sedimentload, the smallermagnitude event than the "dominant"discharge to curve relating dischargeto cumulative sedimentload has a which the channelform responds.Contrary to Bensonand singlemaximum at somefairly highmagnitude, but not ex- Thomas's[1966] results,more recentanalyses by Andrews treme,discharge referred to asthe effective discharge [Pickup [1980] for the total load, Pickupand Warner[1976'1 for bed andWarner, 1976; Andrews, 1980]. load and Webband Walling[1982] for suspendedsediment The durationof the effectivedischarge varies with the form haveshown that the effectivedischarge is a relativelyfrequent of the frequencydistribution of dischargeand thereforewith event exceededbetween 1 and 10 days/year(0.35-3% of the factorssuch as drainage area, drainage basin topography, and time).Nolan et al. [1987] reporteffective discharges with fre- geologyand the temporal pattern of precipitationinputs, and quenciestoward the lowerend of thisrange (2-4 days/year). is also influencedby the nature of the sedimentload. Thus The relationshipbetween the effectivedischarge for sedi- Wolmanand Miller [1960]demonstrated that the significance ment transport and the dominant dischargefor channelfor- of highmagnitude events was greater in smallerdrainage mation appearsto vary. Andrews'[1980] samplesof streams basinsand in ephemeralstreams, while Andrews [1980'1 in the YampaRiver basin have an effectivedischarge for total showedthat theeffective discharge for totalsediment load was sedimentload of almostidentical magnitude to the bank-full less frequent in smaller drainage basinswhich have more discharge.However, Pickup and Warner [1976] foundthat the skeweddaily dischargeduration curves. The influenceof sedi- effectivedischarge for bed load transporthad a lower return periodthan the bank-full discharge,and Bensonand Thomas Copyright1988 by the AmericanGeophysical Union. [1966] clearlydemonstrate that the effectivedischarge for sus- Paper number 7W5038. pended sediment load is well below bank-full. Recent data 0043-1397/88/007W-5038505.00 from northernCalifornia [Nolan et al., 1987] also showthat

864 AS•MORE^•n D^¾:StJs•,œ•r>•D S•m•,tEN'r TRANSPORT 865 theeffective discharge is considerablybelow bank-full. It is the year, althoughoccasionally a large summerrain storm nowapparent that the dominantdischarge for channelmor- may dominate. phologyis alsosubject to considerablevariation controlled in partby environmental conditions, the recovery time for large DATA floods[Wolman and Getson, 1978], and the threshold for sedi- The Water Surveyof Canada (WSC) began a program of menttransport [Baker, 1977]. Thus it appearslikely that not sediment data collection in the Saskatchewan River basin in onlyis the effectivedischarge for suspendedsediment trans- 1962.As of 1983(the last year usedin this analysis),there were portmore variable in relativemagnitude than Wolinartand 27 stationswith more than 4 yearsof data and 18 with records Miller[1960] suggested but alsothe effectivedischarge for 10 yearsor more in length,up to a maximumof 29 years.Of sedimenttransport may not alwaysbe similar in magnitudeto these,21 were chosenfor effectivedischarge analysis. These thedominant discharge for channelmorphology. stations are listed in Table 1, and their location is shown in Thepurpose of thispaper is to addto the existinginfor- Figure 1. mation on the magnitude and duration of the effectivedis- Data are collectedaccording to standardWater Surveyof chargefor suspendedsediment transport and in particularto Canada practice.A single depth-integratedsuspended sedi- presentevidence from a large drainagebasin in western ment sample is collected at a cross section which normally Canadathat theduration of theeffective discharge may vary coincideswith the dischargerated section.Usually, data are morewidely than has previouslybeen reported, as a resultof collectedonly during open water season(March or April to differencesin basin area, streamflow regime, and perhaps the October)but in somecases (particularly during the 1960sand natureof the sedimentload. The drainageareas of manyof early 1970s)sampling has been carried out year-round. The thesestreams are comparablewith thoseused by Bensonand suspendedsediment load carried during winter averagesless Thomas[1966] and considerablygreater than those used in than 5% of the annualload on theserivers lAshmore, 1986]. mostother previousinvestigations. Samplesare collectedevery few days on averagebut more frequentlyduring high discharges. The concentrationin the singlevertical is adjustedto give a SASKATCHEWAN RIVER BASIN crosssection average using an adjustmentfactor that is calcu- The Saskatchewan River upstream of The Pas, Manitoba lated from occasionalsampling at severalverticals in the cross drains a total area of 347,000 km2 east of the continental section.The measuredsuspended sediment concentrationis divide including much of the and Saskatchewan plotted alongsidethe continuousstage record and a continu- Plains(Figure 1). The drainage systemconsists of two main ous record of concentrationis interpolatedbetween the sam- branches, the North and which pies. Daily mean concentration is calculated from this curve join in central Saskatchewanabout 300 km upstreamof The and multipliedby the daily mean dischargeto give the daily Pas. The river empties into Cedar Lake and from there into suspendedsediment load. This method differs from the dura- . tion curve-sedimentrating curve techniqueon which previous Thereare few large natural lakesin the systemapart from analysesof effectivedischarge for suspendedsediment trans- the Cumberland Lake and Delta region in east-centralSaskat- port have been based.Consequently, the inherent bias in the chewan.However, there are severallarge reservoirs,notably rating curve method which has been revealedrecently [Fer- on the , Lake Die- guson, 1986], and which, unless corrected, often leads to un- fenbaker on the South Saskatchewan River, and Tobin Lake derestimationof the concentrationfor high dischargesis not on the Saskatchewan River. In addition, there a number of presentin thesedata. This may have a bearing on the results smallerpower dams in the headwatersof the and reportedhere comparedwith the previouslypublished analy- several onstream and offstream reservoirs associated with irri- ses discussed above. gation in the Bow, Oldman, and South Saskatchewan River basins.Flow regulation at these sites may influencethe ef- ANALYSIS fectivedischarge at somedownstream stations on theserivers; For each station the total suspendedsediment load trans- details of this are discussed below. ported by incrementsof the dischargerange was calculated By far the largestportion of the area of the basinlies within and plottedas a histogram.The modal value, the discharge the Alberta and Saskatchewan Plains where relief is low increment with the largest cumulative load, is taken to be the exceptwhere the major have inciseddeep valleys.The effectivedischarge. The duration of the upper and lower topographychanges dramatically on the westernmargin of bounds of this dischargeincrement were read from the flow the basinwhere the Plains give way to the duration record. and Foothills. It is this region of the basin, where annual precipitationis highest,which supplies much of the runoffto RESULTS the large Plains streams.Roughly 70% of the streamflowin the larger streamsis suppliedfrom the portion of the basin SedimentDischarge Histograms upstreamof Edmonton, Red Deer, Calgary, and . The form of the sedimentdischarge histograms varies con- Close to the Mountains the annual runoff regime showsa siderablyamong the stationsanalyzed, and Figure 2 contains singlepeak in May, June,or July(depending on theproximity illustrativeexamples. Contrary to the expectationof unimodal to the Mountains)derived from snowmelt(and in somecases histogramswith clear effectivedischarges, a variety of forms glaciermelt as well) and springand early summerrainfall. emerged.These can best be describedby groupingthem into This runoff is routed downstream but further east it is often particular types. preceededby a smallerpeak derivedfrom Plainssnowmelt The first of these are stations that fit the established model. andspring rain. In the caseof riverswith no Mountainsource Effectivedischarge at thesestations is relativelyfrequent and thisearlier snowmelt peak is usuallythe major runoff event of the histogramhas a well-definedsingle mode. This type is 866 ASHMOREAND DAY' SUSPENDEDSEDIMENT TRANSPORT ASHMOREAND DAY: SUSPENDEDSEDIMENT TRANSPORT 867

TABLE 1. DrainageArea, Mean Annual Discharge, and Effective Discharge for Stationsin theSaskatchewan River Basin Lengthof Sediment % Durationof Effective Drainage Record,years Discharge Station Area, WSC Station Number km2 Seasonal Annual Seasonal Annual

CrowsnestRiver at Frank 05AA008 402 5 I 1.93-3.52 01dmanRiver near Waldron's Corner 05AA023 1,440 8 2 0.11-0.45 OldmanRiver near Brocket 05AA024 4,400 18 12 0.64- !. 52 04). 02 WillowCreek near Claresholm 05AB021 1,100 11 10 0.04-0.17 0.03-0.11 WillowCreek above Chain Lakes 05AB028 162 13 11 0-0.12 0-0.08 OldmanRiver near Lethbridge 05AD007 17,000 12 6 1.87-4.44 SouthSaskatchewan River at Highway 41 05AK001 66,000 18 12 6.22-12.54 4.29-8.24 MarmotCreek main stem near Seebe 05BF016 9.4 21 15 0.05-0.21 0.05-0.21 ElbowRiver at BraggCreek 05B J004 792 7 0.75-1.60 HighwoodRiver nearthe mouth 05BL024 3,990 ! 1 7 3.26-5.68 at Red Deer 05CC002 11,600 13 3 7.56-14.01 Red Deer River at Drumheller 05CE001 24,800 9 2 8.73-15.62 Red Deer River near Bindless 05CK004 44,700 18 12 15.24-47.54 9.11-29.54 SouthSaskatchewan River near Lemsford 05HB001 119,000 I0 9 2.96-4.98 Swift Current Creek near the mouth 05HD037 3,910 9 0.64-0.99 SouthSaskatchewan River near Outlook 05HF001 136,000 12 8.17-13.55 SouthSaskatchewan River at Saskatoon 05HG001 141,000 6 6 6.80-22.45 5.43-22.45 North SaskatchewanRiver at Whirlpool Point 05DA009 1,920 10 6 11.82-16.97 North Saskatchewan River at Prince Albert 05GG001 131,000 22 15 8.76-20.61 6.01-13.38 Carrot River near Smoky Burn 05KC001 9,250 7 2.04-2.96 SaskatchewanRiver at The Pas 05KJ001 347,000 29 21 20.73-32.20 13.95-21.71

exemplifiedin Figure 2 by the North SaskatchewanRiver at Lakes(05AB028), and Marmot Creek Mainstem near Seebe PrinceAlbert (05GG001) and includesmost of the other sta- (05BF016)also exemplifythis type. Theseare all fairly small tionson the large Prairie rivers with Mountain sources(Red Foothillsstreams upstream of the stationsin the third group. Deer River at Drumheller (05CE001), Red Deer River at A fifth type of histogramcan be identifiedin whicha single Bindloss(05CK004), South SaskatchewanRiver at Highway modeoccurs at a dischargewith a duration similar to thosein 41 (05AK001),South SaskatchewanRiver near Lemsford the firstgroup but in whichthe histogramhas a broad,almost (05HB001),and South SaskatchewanRiver near Outlook flat, peak coveringa wide dischargerange. Both the North (05HF001)). SaskatchewanRiver at Whirlpool Point (05DA009) and the Thesecond group of stationsare thosein whichan effective SaskatchewanRiver at The Pas (05KJ001) have this form. In dischargecan be recognizedbut the histogramhas a very the case of the former the regime is dominated by summer erraticform and dischargesof widelydiffering durations trans- snowand glaciermelt from the Rocky Mountains,while the portsimilar sediment loads. In somecases the effectivedis- latter has a regimestrongly influenced by upstreamreservoir chargeis a relativelyextreme event. Swift Current Creek is an and lake storage. exampleof thistype illustrated in Figure2. Otherexamples are the CrowsnestRiver at Frank (05AA008)and the Carrot Duration of Ef[kctiveDischarge Rivernear SmokyBurn (05KC001).These are all relatively Despitethe erratic nature of someof the sedimentdischarge smalldrainage basins whose average annual flow regimes histogramsit is possibleto identifya singledischarge in- showa single,sharp snowmelt peak. crementwhich has a larger cumulativesediment load than any A third groupof stationsare thoseat whichthe effective other increment. The duration of the discharges at the upper dischargeis withinthe normalduration range but theextreme and lower ends of this effective increment are shown in Table eventsduring the periodof recordtransport sediment loads 1. In all cases the duration is based on the seasonal flow and almostcomparable to the effectivedischarge. The consequencesediment record {April to October). During winter, discharges isa histogramwith at leastone significant secondary peak at are consistentlylow and sedimentload is negligiblein all these theextreme upper end of the dischargerange. The curvefor streams,and therefore the sediment load histogramsare af- theRed Deer River at Red Deer (05CC002)shown in Figure2 fectedvery little by the use of the seasonalrecords. However, is an exampleof this type. Other examplesinclude the an importantconsequence of this is that the flowdurations of OldmanRiver at Lethbridge(05AD007), the ElbowRiver at the effectivedischarge increment are higher (by about one BraggCreek (05BJ004), and the HighwoodRiver near the third) than they wouldbe if annualdata wereused. Where the annual recordsare sufficientlylong, the durationsfor the ef- mouth (05BL024). Thestations in thefourth group have a formsimilar to that fectivedischarge increment for the annual data are also given ofthe third but differby beingdominated by theevents at the in Table 1 and can be comparedwith the seasonalresults for upperend of the dischargerange. Thus in somecases, the the same streams. largestdaily flow on recordis theeffective discharge. Cases Table 1 reveals a considerable range in duration of the suchas these have not previously been reported. The example effectivedischarge from less than 0.1% of the time to over in Figure2 is WillowCreek near Claresholm (05AB021). The 40%. It is difficult to identify a modal duration range for the OldmanRiver at Waldron'sCorner (05AA023), the Oldman effectivedischarge, although there is a predominanceof values in the range!-10%. This rangeof valuesis more extremethan Rivernear Brocket (05AA024), Willow Creek above Chain . 868 ASHMOREAND DAY: SUSPENDEDSEDIMENT TRANSPORT

uJ uJ

I

rr' 00• I [,mlff f JillIll! i Jml,t, , h,.,[,• I[.[[,,, It[,,,,,] J , •

[,,t[[] , • [,],][ [ [

(S•INN01) (SXYQ) avo• •N•lw[aas a3aN:•dsns

Iiiii I i i I Iiiii i I i i

I

,

u. iz (S=JNNOJ.) (SAYO)AON::IFIO::11:I=I rr ,,- avoa 1NaV•Ia•S a•ioN;iasns

! Jill! I ! [ i Jill ! i ! ! I z

rl-

f,r,,, , , , "'l,,,[, , , r

(S=INNO1) (SAVO)AON•]nogt:l::l (:]¾O'1J.N=I•10::IS I-

uJ

uJ I...... z

r!' illIll ] I I Jllll • [ ! I >_. •r , (S=•NNOJ.) (SX¾O)AON=JNO•]Idd 0¾09 IN:ql•ila•]s a•]ON•dSnS

0

(S•NN01) (sx¾a) XON•inoau:l avo3 •Nam•oas aaaNaasns ASI-IMOREAND DAY' SUSPENDED SEDIMENT TRANSPORT 869

100

10

0.1

0.01 lO lOO lOOO 1000o 1oooo0 DrainageAreaOcrn 2) Fig. 3. Drainage basin area versusthe percentageduration of the effectivedischarge increment. Stations are labe!ed by an abbreviated station number. any previouslyreported and doesnot supportany widelyap- This relationship between effective discharge duration and plicable generalization about the duration of the effective dis- drainage area is probably primarily a reflection of differences charge. These data include sites at which the effective dis- in dischargeregime and flow duration characteristics.Thus for charge is well in excess of the mean annual flood as well as example,the skewnessand range of the daily dischargestends severalat which the effectivedischarge is exceededmore than to decreasedownstream lAshmore, 1986] and therefore is in- 10% of the time. verselyrelated to drainage area. An additional factor account- A brief inspection of Table 1 reveals that the variation in ing for this relationship is that in the larger Prairie rivers duration of the effective discharge is related to the drainage suspendedsediment concentration tends to remain relatively area of the streams.This is illustrated in Figure 3 from which high even at low discharges,perhaps becauseof a large wash it is appa'?entthat the larger basinstend to have effective load, which may help explain high frequency of the effective dischargeswith higherpercentage exceedantes, the patternex- dischargein those rivers. pectedon the basis of previous work by Wolmanand Miller Another factor to be considered in explaining this down- [!960] and Andrews[1980]. The resultis a downstreamtrend stream trend in the duration of the effective discharge is the alongthe major riversfrom small Foothills streamsin which effect of flow regulation and consumptive use on the larger the effectivedischarge is a fairly extremeevent occurring1 streams.In the absenceof preregulation data it is impossible day/yearor less,to large Prairie riversin whichthe effective to establish the extent of this effect but two points can be dischargeis a common event exceeded30 or more days per made in this connection. First, the downstream increase in the year.This can be seenin Figure 3 by followingthe sequenceof percentageduration of the effective discharge occurs in the stationsdownstream along the Red Deer or Oldman-South Red Deer River which was unregulated during the period for SaskatchewanRivers (see Figure 1). For example,the down- which the analysiswas carried out. Second,the potential influ- streamsequence in percentageduration of the effectivedis- ence of flow regulation on the effective discharge is unlikely to charge at stations along the Oldman-South Saskatchewan be great except in two or three cases.Calculation of the natu- River is 0.1!-0.45, 0.64-1.52, 1.874.44, and 6.22-12.54 for sta- ral flow of these rivers by Prairie Provinces Water Board tions 05AA023 (Oldman River at Waldron's Corner), [1982] indicatesthat in the South Saskatchewansystem up- 05AA024(Oldman River near Brocket),05AD007 (Oldman streamof Lake Diefenbaker consumptiveuse has reducedna- River near Lethbridge),and 05AK001 (South Saskatchewan tural flows by about 15-25%. These reductions are fairly Riverat Highway41), respectively. evenly distributedthrough the open water seasonand there- 8'70 ,A•SHMOREAND DAY: SUSPENDED SEDIMENT TR^NSPO•ZT fore there has been no pronounced alteration of the annual of moderatemagnitude as previouslysuggested by Wo!man flow regime. andMillet' [1960] and confirmedby Andrews[1980]. How- In the North SaskatchewanRiver the annualflow regime ever,this is not true of all streams;in somecases extreme hasbeen altered by flow regulation.Average monthly flows in eventsmay be the effective discharge while in others,particu- July and April are now almost identical in contrast to the larly thosewith largedrainage areas (greater than 10,000 naturalregime in whichthe averageJuly dischargewas almost km-'),very frequent events transport the greatest proportion of twicethat of April. This broaderpeak to the annualdischarge the load. The relationshipbetween effective discharge and regime may have increasedthe percentageduration of the bank-fulldischarge is presumablycorrespondingly variable, effectivedischarge on the North SaskatchewanRiver at Prince makingquestionable any generalizations on thispoint. Albert (05GG001). The same is true of two other stations: Finally,in severalstreams similar quantities of sedimentare SouthSaskatchewan River at Saskatoon(05HG001) and Sas- transportedby flows of quite different magnitudeand fre- katchewan River at The Pas (05KJ001), both of which are quency.In thesecases the concept of an effectivedischarge for strongly influencedby upstream reservoirs.Note that the sta- suspendedsediment transport is inapplicable. tion on the South Saskatchewan River near Outlook Acknowledgments.The work describedin this paper wascarried (05HF001) was discontinuedprior to the impoundmentof out whilethe firstauthor held a NaturalSciences and Engineering Lake Diefenbaker and hence the sediment transport data at ResearchCouncil of Canada Visiting Fellowshipin CanadianGov- that site are not influencedby the reservoir. ernmentLaboratories. Thanks to JosephMcllhinney for computer programmingand to KeithBigelow and Zofia Wolinski for drafting. CONCLUSIONS REFERENCES The form of the effectivedischarge histograms and the dura- Andrews,E. D., Effectiveand bankfull dischargesof streamsin the tion of the effectivedischarge for suspendedsediment trans- Yampa River basin, Colorado and Wyoming, J. HydroI., 46, 311- port in streams of the Saskatchewan River basin are highly 330, 1980. variable. In a sample of streams with drainage areas ranging Ashmore, P. E., Suspendedsediment transport in the Saskatchewan from 10 to over I00,000 km 2 and with sediment records be- River Basin, Rep. IWD-HQ-WRB-SS-86-9, Sed. Surv. Sect.,Water Surv.of Can., Water Resour.Branch, Inland WatersDirectorate, tween 5 and 29 years in length, the duration of the effective Environment Canada, Ottawa, Ont., 1986. discharge is less than 0.1% in some cases and over 15% in Baker, V. R., Stream-channelresponse to floods,with examplesfrom others, with the majority of stations having values between 1 central Texas, Geol. Soc. Am. Bull., 88, 1057-1071, 1977. and I0%. In extreme casesthe effectivedischarge is the largest Benson, M. A., and D. M. Thomas, A definition of dominant dis- flow to occur during the period of sediment record. These charge,Bull. lnt. Assoc.Sci. HydroI., 11, 76-80, 1966. Ferguson,R. I., River loads underestimatedby rating curves,Water results are largely based on seasonal records and must be Resour. Res., 22, 74-76, 1986. comparedcautiously with previousanalyses of annual data. Nolan, K. M., T. E. Lisle, and H. M. Kelsey,Bankfull dischargeand In many casesthe effective dischargehistograms are not the sediment transport in northwestern California, Erosion and Sedi- simple unimodel distributions envisaged by Wolman and mentation in the Pacific Rim, IAHS Publ., 165, 439-449, 1987. Miller [1960] and shown, for example, by Andrews [1980], Pickup,G., and R. F. Warner, Effectsof hydrologicregime on mag- nitude and frequencyof dominant discharge,d. HydroI., 29, 51-75, but rather they have a complex form sometimeshaving peaks 1976. of similar magnitude at two or more dischargeswith quite Prairie ProvincesWater Board, Water demand study,Historical and different durations. current water uses in the Saskatchewan-Nelson Basin, Prairie The duration of the effective discharge is related to the Provinces Water Board, 1982. Richards,K., Rit,ers: Form and Processin Alluvial Channels,358 pp., drainage area of the streams, perhaps reflecting differencesin Methuen, London, 1982. the form of the discharge duration curves. The smaller head- Webb, B. W., and D. E. Walling, The magnitude and frequency waters basins in the Rocky Mountains and Foothills tend to characteristicsof fluvial transport in a Devon drainage basin and be dominated by the most extreme events, while the large some geomorphological implications, Catena, 9, 9-23, 1982. Wolman, M. G., and R. Gerson, Relative scales of time and ef- Prairie rivers with Mountain sources are dominated by fre- fectivenessof climate in watershedgeomorphology, Earth Surf.Pro- quent eventswith durations often greater than 10%. The high cessesLandforms, 3, 189-208, 1978. wash load of the large Prairie rivers may also be partially Wolman, M. G., and J.P. Miller, Magnitude and frequencyof forces responsiblefor the frequencyof the effectivedischarge in these in geomorphic processes,d. Geol.,68, 54-74, 1960. rivers. The effective discharge at some stations on the down- P. E. Ashmore, Department of Geography, University of Saskat- stream sections of the main stem North Saskatchewan, South chewan, Saskatoon, Saskatchewan, Canada S7N 0W0. Saskatchewan,and SaskatchewanRivers may be influencedby T. J. Day, Sediment Survey Section, Water ResourcesBranch, In- upstream flow regulation. The tendency in thesecases would lands Waters/Lands Directorate, Environment Canada, Ottawa, On- be to increase the percentage duration of the effective dis- tario, Canada K1A 0E7. charge. (ReceivedJuly 20, 1987; It is apparent from thesedata that the effectivedischarge for revisedJanuary 29, 1988; suspendedsediment transport may be in many casesan event acceptedJanuary 29, 1988.)