Temporal variations in the through the Knabeana-Kvina Fedafjord, Norway

Temporal variations in the through the Knabeana-Kvina Fedafjord, Norway

MARIANNELANGEDAL

transport of river system, .

mine tailings and into the

Langedal,M.1996:Temporal variations in the transport of mine tailings through the Knabefma-Kvinariver system, andintothe Fedafj ord,Norway.Nor.geol. unders. Bull.430,95-101.

Mine tailingsfrom the Knaben Molybden umMines were discharged into twolakes in the headwaterarea of the Knabeana-Kvina drainage basin inthe period 1918-1973.Adamwasbuiltdownstreamofthetwo lakesin 1976to prevent spreadi ngofthe waste.However, large amou nt s oftailingswere previously washed outof thelakes,and approximately420,000tonsarenow depositedalong the river. Chemicalanalysisof8 samp lesofsuspendedsedi- ment,collected in1993-1994,wasperfo rmed to examinewhethe rthe tailingscouldbe recognisedinthesuspen- dedsedimentload.The medianacid-solubleconte ntsofCu andMo were120 and 48 ppm, respecti vely.The con- centrat ionsare similartothoseinthespoil heap andoneorderof magnitude high erthan inthelocal nat uralsedi- ment .Thesuspendedsedimenttransportrateintheperiod March-December 1993 wasmonitoredby freque nt sampli ngof suspendedsediment(1-4timesaday) andcont inuou s measurementsofwaterdischarge.Inthemoni- toredperiod,the tota lsuspended sedimentyieldwas 600 ton s.Of these, 90%were transportedduring7%of the time.Thesediment transportoccurred in pulses correspondingto period s wit h high wat er discharge.Rapidvariati- onsrequiremonitoring by frequentsampling in order to assesstheamount sof tailingsthat are transported towards the fjord.The relationshipbetween sedi ment transportandwater dischargeseemsto be governedby major flood s that open upsediment sourceswhich remainvulnerablefor erosion in thecoming years. After largefloodsthe transportrate of tailing smay thereforeincreasesignificantly in thesup ply-limited Knabeana-Kvinariver.This impli- catesthat future internationalnegotiations and abatementst rategies concerningfluvial input of heavy metalsto marine areas should assessthe risk of input of particle-bound histor icalpollution,bothin normal flow sit uati ons and in extremeflood events.

Marianne Langedal,Depart m entofGeolog y and Min eralResourcesEngineering,Norw egian UniversityofScienceand Technology,N-7034 Trondheim,Norwa y.

Introduction

The objective of this paper is to demonstratethe tempo- ral variations in the particulate heavy metal load carried by rivers,and to pointout implications for the input of heavy metalsto marine areas.

The North Sea Declaration of 1987 aims at 50-70 % reductions in the riverine input of certainheavy metalsby 1995, with 1985 as a reference year.In Norway,monito- ring of the fluvial inputisbased on monthly watersam- pling of the mainriversenteringthe North Sea(Holtan et al. 1991).Little attentionhasbeen paidto particle-bound heavymetals,alt hough several studieshave shown that the majo rpart oftheheavy metalstransported by rivers exists as particles or is adhering to particu late matt er (F6rstn er&Kerstens 1988,Leenaers1989).The transport rate of particle-boun d heavy metals is prob ably more dependenton sediment transportratethan on theheavy metal concent rat ion in the particul atematt er since the formerexhibits larger temporalvariati ons (Walli nget al.

1992a,Langedal et al. 1996).

Heavymetals in the particulate strea m loadcanorigi- nate from bot h natural and ant hropoge nic sources.

Region algeochemi calmappinghas shown that the natu-

ral contents of metalsin soilsa nd sed iment

Ottesen&Belviken 1987).Superimposedon thisdistribu - tion pattern are metalsfrom anthropogenicdischarges in present and earlier times. An example is the 17 billion tons of heavy metal containing waste left by histor ical mining in Europe(RIVM 1991).The spoil heaps are com- monly situ ated nearrivers,where the waste is subjected to natural erosion, transport and deposition cycles.

Chemical analyses of overbank sediment on floodplains have shown that mine waste hasbeen scatt ered down- stream from spoil heaps in approximately30%of all drai- nage basins in Europe(Demet riades et al. 1990,Bolviken et al. 1993).Inthisway,large amounts of heavy metals are temporarilystored indrainagesediment that may be entrained inthe river load and brought int o themarine enviro nmen t.

Thispaperdiscussesthe temporal variati ons of fluvial transport of histori cal mine waste from the Knaben Molybdenum Mines through the Knabeana-Kvina river syst em and int o the Fedafj ord, Norway (Fig. 1).The dis- cussions are based on chemical analysisof suspend ed sedim ent load and moni t oring of suspended sedi ment transport.Theprobable influence offut uremajorfloods on theinput rateof historic al pollut ion intothe Fedafjo rd isdiscussed, basedon sediment erosion and transport data from rivers that have recentl y been subjected to maj or floo ds. Practical impli cations for moni t orin g pro-

96 Marianne Langedal NGU-BULL 430, 1996

Fig. 1.Map overthe unregula tedpartoftheKnabeAna-Kvinadrainage basin.

withsampling locationof suspended sedimentmarkedby S.

Water discharge measurem ent,sampling and sample treat- ment

Water discharge was measured continuouslyin the peri- od March-Dec ember 1993, using a Ste vens Lim nig raf.

Samplesofwat erwith suspend ed sediment were pum- ped up from a turbulent section of the Knabeana river (Fig. 1) by an ISCO automatic pumping sampler,accor- ding to a proceduredescribed byBogen(1988, 1992).A 50-litre sam p le wascollected eve ry 10 daysfor ch e mical and grain-size analysis,while 1-litre sam p leswere collec- ted 1-4 times a day forsediment transport monitoring. The samples were filtered on 0.45urn milliporefilte rsto separate suspended particles from wat er.Since few 50- litre samplescontainedenoughparticu latesfor the selec- ted chemica lanalysis(> 1g particulat e material),7sam- ples from 1993were combinedto4samples.Inaddit ion, 4 were collecte din 1994.

barren rock whi le therem ainin g drainageareahas rather thin glacial and glaciofluvial deposits wit h ast rong vege- tati oncover. Fluvial depositsare restrictedto the im medi- atevicinity oftheriver.The larger portionof the river bed isarmoured by gravel and peb bl es,indicating that the sediment transpo rt capacity exceeds the sedi ment sup- plyBat hurst 1987).Such rivers were called'sup ply lim it ed' by Pitlick& Thorne(1987), andinthefollo w inq thisterm willbeusedaboutthe Knabeana-Kvina river.

The Knaben Molybdenuminesare situatedsome 50 km upstr eam from the outlet of the Kvin a rive r int o the Fedafjo rd. Mini ng and milling took place from 1918 to 1973.In this periodatleast 8 milliontonsof wastemat eri- al were prod uced and deposited in two nearby lakes.

Thesetailingscont ain approxim ately200 ppm acid -solu- ble Cu and 40 ppm acid-soluble Mo (median values, Langedal 1995).To redu ce the spr eading of tailing s by fluvial processes, a dam was built down stream of the lakesin 1976.However,conside rable amountsof tailings had previously been washed downst ream by the Knabeana-Kvina river.Today,an est imated 420,000 tons of thismaterialare visible aslarge sandbarswithi n the rive r channel and asoverbank deposit s on the f1ood- plains(Lang edal 1995).The medianconcentrationsof Cu and Mo in these deposits are oneorder ofmagnit ude hig- her than in the pre-m in ing sediment from the bottom partof overbanksedim entprofiles(Fig 2,Langeda l1995, Langedal & Ottesen 1995).The construction of the tai- lings damhelp edout tocut off the sedimentsup plyfro m the majorwasteheap.Todaythe most erodible material in the drainagebasin is the tailing sdeposite dwithi n the chan nel(Langedal&Ott esen1995).Minor sediment sour- cesare tailing s onfloodplains and natural till,glaciofluvial and fluvial deposits. Open natural sedi m ent sources, however, are few, due to the lim it ed amo unt of Quat ernarydepositsand the strongveget ati on cove r.

Methods

o

. . _ . . . . I

CONFL UE CEWITH THELIT LEA. ARIVER

-"'

I

I

I I ~l

/

~)

\

\

,...-

/

__

,,-,

/ / " , /

. / /

('

(

REGULATED / ) )

/ / ...1

. / )

.v -.

i {

J

. s '. . I

: '.' ./

. /J

.: : .) - \!

I

THEK ABE

(\MOLYBDENUM MINE

;;

I

( (

\ r -- -- - - - - --i )

\

(

/ I'

\

\

I.r

\

,

I

<,

"

) /

.... "

I

l.-

\

»:.... /

\

~..--,

( - _ . . -

<, ) / I / f I .) (

l' , -

\

I

\ ' \

.J 1 ,- (

\..,.-'"\

I

l

\

KVINESDAL THE0Y E SMELTE R

grammes, risk analysis, and new treaties concerning inpu t of heavy met als to coastal areas arealso poin te d out.

Study area

The Knabeana-Kvinadrainage basin belongstoa geoche- mical Mo-province underlai n by Precambrian granites and gneisses (Sigmondet al. 1984,Ottesen et al. 1989).

About50%ofthedrainage basin consistsof outcropping,

NGU-BULL 430,1996 MarianneLangedal 97

ned by weighing the filter residue and the remaining water sample. The suspended sediment concentrations between the times of sampling were estimated by linear interpolation. The suspended sediment transport rate could then be obtained by multiplying sedimentconcen- trationsby the continuouslymeasured water discharge.

To test whether suspended sedimentsampling once a month would give a similar result as the frequent sam- pling describedabove,the first samples collected on the 3rd of every month were selected.A regressionanalysis of logtransformed data for suspended sediment concen- tration versuswater dischargewas performed.For every day in the period March to December,the suspended sediment concentrations were calculated by useof the regressionequati on and theobserved meandailywater discharge. Sincethe sedimentrating curveswere obtai- ned from logtransformed data,transformation bias was corrected for by multiplying the calculated suspended sediment concentrati ons by the factor e(2,651 • 52),where 52is the standard error of the estimated rating-curve in log 10 unit s(Ferguson 1987).Estimated daily sediment transportrates could thenbe obtainedbymult iplyingthe sed iment concentration wit h the observed mean daily water discharge. The same procedure was repeated for samples collected on the 13th,the 18th and the 24rdof every month.

- • - Sandbarsedime nt - . . - F:iord sediment

20 40 60

Downstream distance Crom the tailings dam (km)

Overbank sediment, upper section Overbank sediment. bottom section

o

1000

-=

-e

..

e

a

... ^--

::i 10 o

0 20 40 60

Downstream dista nce Crom tailings dam [km)

1000

-=

al 100 _{_-e}

z ^a

.:

0 ::0

Fig.2.Median acid-solubleCuandMo concentra tionsin drainage sediment

versus thedistance fro m the spoil heaps(Langedal& Ottesen1995).

Re sults and discussion

Chemical analysis

Calculationofsuspended sedimenttransport

Suspended sediment concent ration(mg/I) was determi-

Table 1:Acid- soluble Moand Cuconcentrations in suspendedsediment samplesfrom the Knabeanariver

Chemical analyses of the suspended sediment show that tailings constitute the larger part of the suspended load (Table 1).The acid-soluble contents of Cu and Mo are similar to those in thespoil heap and one order of magni- tude higherthan inthenatura lsediment. Sinceonly sam- plescollected at high dischargeswere largeenough for chemicalanalysis,the sediment load in low flow situati- ons may have a different composition. However, the major part of the monitoredsed imenttransport occurred in periods with high water discharge (Fig. 3) and the obtained concent rat ions are therefore thought to be representativefor the suspended load.

From March to December,1993, the suspended sedi- ment yield ofthe Knabeanariverwas approximately600 to ns.Of these,more than 150to ns were transported on the 19th of December,another 200tonsin a three-day period in April,and another 70 tons in a five-day periodin March.In fact, 90%of thesedi ment transportoccurred in

Mo(ppm) Cu(ppm)

Samples of suspended sediment were analysedfor Cu, Moand30 other aquaregia solubleelementsby ICP-AES at the Geolog ical Survey of Finland (Eden & Bj6rklun d 1994).The same procedure waspreviously used for che- mical analysisof materialfrom the spoil heaps and sedi- ment withinthis drainagebasin (LangedaI 1995).

Date(s)

26March1993 3+10 May1993 17May 1993 24 + 31May

+7 June1993 25April 1994 2May 1994 8May1994 11 May1994

142 255 65 97.5

54.4 37.4 179

19.5 50.7 33.5 44.3 82.9

Table2:Estim atedsediment yieldMarch-December1993 using data from thesame date every month.

Date Sedimentyield %ofmeasured

(tons) yield

3rd 270 44

13th 130 21

18th 500 82

24rd 215 35

98 ~arianneLangedal NGU-BULL430,1996

200 £ ⁴⁰

-

ca

.$ 1993 _§~ ^IQ

:((. ^{q;; "#§} -

\ ^~

en

-

= ¹⁵⁰

³⁰ e

+J0 Waterdisc harge

-

-

_...

0

£ ...

~

°1 ^ca

en 100

20 .=

= ^§~

^en

ca _"#§

.l:2

^'0 ...

+J _Q)

5 ^{50 10}

ca

.8 ₌₌

'0

r.n 0 0

i

^Is.

⁴

^f ;; ^f

^!fJ

^if

^....

^~

^at

~ ~ ~ ""'f

~ ^{~ ~}

~ ~

Fig ,3.Sedimenttransport (Gs)andwa ter discharge(0)in the Knabeana river in1993,based on1-4samples of suspendedsedimenta day.90%of the trans- port occurs in7%of thetime(Langedal &Ottesen1995).

100

" "

• •

•

• ISt. residual l < 2

"St .residual > 2 - St.residual <-2 - Fitted line

10

,

•

Q (nf Is)

•

1

• _{• •}

• •• •• •

••• .>.

• •• J'

I

". ..

... ^"'. _• ..

. ^{. ...}

... .. .

...

.. _ ^-

1

Cs

5.6

10 Q R = 0.7, p = 0.000 0.1

0.00001 0 .1

-

^{0 .01}

"

- ^en

U ^0.001

I t ^•

0.000 1

Fig.4.Regressionanalysisof sediment concentration(Cs)versus water discharge(0)in theKnabeanariver in 1993.St.residual= standardisedresidua l,the vertical distancebetween an observationand the fitted line,divided bythestandard deviationof the residuals(Noru! is 1990).

NGU-BULL430,1996

Fig.5.Erosionscarbythe 200-year flood inthe 0sterdalenvalley,Norway.

Theprot eetionalstructu resalong theriverhave causedturbulence on the floodplain and alargetrench was carved outby theflood(PhotobyJ.

Bogen,1995).

7% of the time, closely corresponding to periods with high water discharge.Several other st udi es also show that the transport rate of suspended sediment andasso- ciated heavy metals are largest in flood sit uati ons (Bradley 1988, Leenaers 1989, Walling et al. 1992a, Langedal et al. 1996).

Suspended sediment yields estimated from monthly sampling of suspended loadin the Knabeana riverrange from 21 to 82%of the measured yield (1-4samples a day, Table 2). The large deviations between the estimates show that more frequent sampling of suspended sedi- ment is necessary to detect the rapid fluctuations in the sediment transport rate (Fig.3),and thus to determine the amount of particle-bound heavy metals that are broug httothe sea.Similarresultswere obtained for the river Ex,UK,wherethe suspended sediment yield was cle- arly underestimated when calculated from rating curves based on weekly sampling (Walling et al. 1992a).

Application of parametric and non-parametriccorrection factors suggested by Ferguson(1986)and Koch&Sm illie

,

I

,

» I

, ,

0- ^f ,

I

,

c..

,

'"

.,

E _I^I

'"

.13

r

'1;l ,

'"

,

en

U ^,

f---- ^I

Time

Maj orflood Major floodfuture?

Fig.6.Modelforvaria tionsinsedimentsupply accordingtothe occurrence oflarge-magnitude floodsinsupplylimitedrivers.Large-magni tude floods willopensedimentsources whichsupplysediment for rivertransport,until theyaredepletedorclosed byre-vegetatio norhum an interference.

Marianne Langedal 99

(1986) reduced the underestimation.However,the preci- sion obtained by 50 replicate data sets was not satisfacto- ry,indicatingthat even weekly sampling is too infrequent to estimatethe sediment yield.

In the Knabeana river,there is a loglinear correlation between suspended sediment concentration and water discharge(Fig. 4).If observations that fall more than two standard deviations from the fitted line are treated as outliers,there are tw elve outliers above the regression line. Of these, eight are connected to situations with heavy rain on snow and frozen ground, giving rapidly increased water discharge.It is,however,uncertain whet- her the rapid discharge-increaseit self is responsible for the high sediment transport rate, or if the within-channel depositsare morevulnerable for erosion during the pre- vailing thawing cond it ions. Of nine outliers below the regression line eight are connected to falling water dis- charges. In other supply limited rivers,a reduced trans- port-discharge relationsh ip on waning stages has been recorded on severaloccasions(e.g. Walling&Webb 1987, Batalla & Sala 1994, Bogen et al. 1994).Thisfeat ure is generallyattributed to the previou soutwash of the avai- lable sedime ntat the risingorpeak stage.

Both the daily and the annua l sedi ment transport rat es depend on natural facto rs,suchas runoff, frequ ency of water level fluctuation,temperature,previousexhaust ion of sediment supply, form of precipitation , and ground wetness (Bogen et al. 1994).A comb inat ion of these fac- tors may cause largevariations betweenthe annualyields of tailings in the Knabeana-Kvina drainage basin. For instance, did the annual sediment yield in the Leira river, Norway,double from 1989 to 1990 due to the occurrence of winter floods at the time when the topsoil was vulnera- ble for erosion (Bogen et al. 1994).The flood didnot have an extraordinarymagnitude, but the timing of theevent was unusual.

Building of protecti onal structures along the river banks may prevent entrainment of historical pollutionin the suspended sediment load at normal discharges.

However,the situation during a large-magnitude flood would be entirely different. During the 200-year flood in the 0sterdalen valley, Norway, in June 1995,the protecti- onal struct ures alo ng riverbanks created strong turbulen- ce on the adj acent floodplain s, and theflood carved out large trenches behi nd the protectional walls (Fig. 5,J.

Bogen pers. comm .). Unprotected floodplains were, in fact, lessinfl uenced by this large-mag nit ude flood.The erosional impact on unprotectedflood plains seems to be stronglydepen dent on the hydraulic relat ionship betwe- entheriver and the flood plain (Brakenridge 1988, Hickin

&Sichingabula 1988,Nanson&Croke 1992).Extreme flo-

ods may even strip downcomplete floodplains in head- wat erareas(Nanson&Croke 1992,Bourke 1994).Indrai- nage basins with historical pollution, large amounts of

con ta m i na ted sed i m e n t will the n be brought to the

downstreamarea and finallyinto the sea.

Major floods may also play an important rolefor the long-term sediment supplyin riversas shown by the fol-

100 Mariann eLangedal

low ing exam ples:

1) In the river Leira, Norw ay, a 1aa-year flood in 1987 low ered the erosion base in several tributary gullies (Bogen et al.1994). Channel-scourundercut adjacent slo- pes andsub sequentlyseveral slideswere triggered.This greatly increased thesediment supply in the follow ing years.Bogen et al.(1994)pred ictedthatthe rate of lands- lideswill persistuntiltheadjacentslop es regain stability.

Increasedsediment erosionand transportrates werealso detected asincreased deposition rateson flood plains in theperiod1986- 1990com pared with 1954-1985 (Walling etal.1992b).

2) In theHowg illFellsarea, UK, large sedim ent sources were opened bya1aa-yearfloodin 1982(Harvey 1987).

During the flood the meandering Langdale and Bowderdalechannels crossed a threshold intothebrai- ded regim e. Threeyearsafterthe large flood the sedi- ment supplywasstill richenoughto maintainanessent i- ally braidedpattern,indicatingthat thesediment supply to therivers was st ill governedby themajorflood(Harvey 1987).

3)Inthe Fall river,Colorado,USA,a damburst in1982cau- sed a flood with a magnitude 2-30times the estimat ed 500-years floo d (Pitlick & Thorne 1987). The sedi ment transpo rtin the two follow ingyearswasst rong ly depen- dent on sedimentsupply from sediment sources created duringtheextremeevent.

These examples indi cate that extreme flood events create sediment sourcesbychannel avulsions,gullying, channel low ering and und ercutting of adjacent slop es.

Until such sou rces are depleted orclosedby re-vegetat i- on orthe const ructi on ofprotectionalstr uct ures,they are vulnerable for erosion during future normal floods.

Sediment supply(and thus sediment transportin supply limited systems) may therefore be governed by the occurrence of large magnitude floods. The model in Figu re 6 is tho ught to be valid for temperate regions where re-vegetation of erosion scars is rath er slow . Accordingto the model,amajor flood in the Knabeana- Kvinadrain agebasinwo u ld no t onlycause asin g le episo- de of high input of tailing sto the Fedafjord;itwould also cause higher input s during subsequent normal floods due to the increased availability of tailings for erosion.

Thus,the occurrence of majorfloods may controlthe spe- ed at which the estimated 420,000 tonsof tailings,depo- sited dow nstreamof thetailingsdam,will be broughtto the sea.

Pract ic al implications

Results from chemical analysesof the suspended sed i- ment load and sediment transport monitoring in the Knabeana river show that hist orical mine tailings are remobilised and brought downstream by natural river processes. Since particle-bo und heavy metals constitu te a large part of the total metal load in rivers (F6rstner &

Kerstens1988,Leenaers 1989),monitoringof the riverine

NGU-BULL430,1996

input of heavy metals to coastal areasshould include monitoring of sedi ment yield and composition. Fluvial transport of particulate heavy metalsoccurs in pulses, and freque ntsam pli ng istherefo renecessaryto estimate the amo unt sofparticle-boundheavymetalsbroughtto the sea eachyear.

As the indu strialeffl uents and discharge of municipal wastewater arebeing reduced bygovernmenta lregula- tions(Longva et al.1994),the relative im port ance of his- toricalheavymetalpollut ion willincrease.The occurren- ce oflarge-magnitude floods may governthe availab ilit y of histo ricalpollu tion forfluvialtransport. In this way the total riverineinput of heavymetalsto coastalareasmay, in fact, increase as the discharges from present-day humanactivit ies decrease. In futureinternati onal negoti- ation s concerning riverineinput ofheavy metals tomari- neareas,riskanalysis andabatement strategiesdirected towar ds historicalpollut ion shouldtherefore beincluded.

Ackno wledgements

Thisproject hasbeen finan ced by the Commissioner of Mines,The Research Cou ncil of Norway,The StatePo llutio n Cont rol Aut horit y,and the NorwegianWaterandEnergy Admin istration.The author is thank ful to Odd Kvinlaug whoruns thesam plingst ation,andto the No rweg ian Wate rand Ene rgyAdministrationthatwas respon sibleforthesedim en t monit o ringprogra m m e.Thanks are alsoexp ressedtoRolf T.Ottesen, Bjo rnBo lviken and AIIan Krillforcommenting on draftsof the manus- cript.

Refe rence s

Batalla,R.J.& Sala,M.1994:Temporalvariab ili tyof suspend edsedi m ent transportina Medit erranean sandygravel-bed river.InL.J.Olive, R.J.

Lough ran&J.A.Kesby(ed s.):Variabilityinstream erosion and sedi- ment transport.IAHs Pub!.no.224,307-31 S.

Bathu rst,J.e. 1897:Differencesbetwee ngravel and sand bed rivers.

Discussio n.In e.R. Thorne, J.e.Bathu rst & RD.Hay(eds.): Sediment TransportinGravelbedRivers. Jo hnWiley& Sons,Chiche ste r,12-13.

Bogen, J. 1988: A monitoring programme of sedi m ent transpo rt in Norwegian rivers.In: Sediment Budgets(Proceedings ofthe Porto Alegre Sympo sium,Decem ber1988).IAHSPub !.no. 174,149-159.

Bogen,J.1992:Monitoringgrainsizeof suspended sed imentsinriver s.

In J.Bo gen,D.E.Walli ng& T.Day(ed s.):Erosionandsedime nt trans- port monitoringprogrammes in riverbasins. IAHs Publ,no. 210,183- 190.

Bogen,J.Berg,H.&sandersen,F. 1994:The contributionof gullyerosion to the sedim en t bud get of the River Leir a. In: LJ. Olive, RJ. Loughran& J.A.Kesby(ed s.):Variabilityin streamerosion and sedi- ment transport.IAHs Pub!.no.224,307-315.

Bo ur ke,M.e.1994:Cyclicalconstructionand destruct ionofflood domi - natedfloodplains in semiaridAustralia .InL.J.Olive,RJ. Lo ug hran&

J.A.Kesby(eds.):Variabili tyinstreamerosion andsediment transport.

IAHs Pub!.no.224,113-123.

Bradley,s.B.1988:Sediment-wate rint eract ions:The physica ltransport ofheavymet alsin the fluvial syst em s.In:Metalsand metalloidsinthe hydra sphere; impact through min ing and industry and prevention technology.Proceed ing s of an IHP-w o rksho p,Bochum,Germa ny, UNESCO,Paris,111 -134.

Brakenr idg e,G.R.1988:Riverfloo dreg im eand floodplain st rati g rap hy.

In V.R.Baker,R.e.Kochel& P.CPatt o n(ed s.):Flood geomorp hology.

JohnWiley&Sons Chichester,139-156.

Bo lviken,B.,Bergstrom,J., Bjorklun d,A., Konti,M.,Leh muspelto, P., Lindholm,T., Magnusson,J.,Ott esen,R.T.,Steenfelt,A.& Volden,T.

1986: Geoche m ical Atlas of North ern Fennoscandia scale

NGU-BULL430,1996

1:4,000,000.GeologicalSurveysof Finland,Norway andSweden, 20 pp,150 maps.

Belviken,B.,Bogen,J.,De Vos,W.,Ebbing,J.,Hindel, R.,Ottesen,R.T., salminen,R.,schermann,O.&Swennen,R.1993:Finalreportofthe Working Group on Regional Geochemical Mapping 1986-1993.

Forum ofEurop ean Geologicalsurveys(FORGES).Nor.geol.unders.

Report 93.092,18pp,6 appendices.

Demetriades, A., Locut ura, J. & Ottesen, R.T. 1990: Geochemical MappingofWestern Europetoward sthe year 2000,Pilot project Report.Nor.geol.unders.Report 90.705,9pp,10 appendices.

EdenP.&Bj orklund,A. 1994:Ultra-Iow density samplingof overbank sedimentin Fennoscandia.J.Geochem. Explor.57,265-289.

Ferguson,R.I.1986:Riverloadsunderest imatedby rat ingcurves.Wat.

Resour.Res.22,74-76.

Ferguson,R.I.1987:Accuracy and precisionof met hodsforestim ating river loads.Earth.Surf.Process,12,9S-104.

Forstner,U.&Kerstens,M.1988:Sediment-waterint eract ions;chemical mobilizat ion.In:Metal sand metalloidsin thehydrosphere;impact throughminingandindustryandpreventi ontechnology.Proceedin gs of an IHP-wor kshop,Bochum,Germany,UNESCO,Paris,111 -134.

Harvey A.M.1987:Sediment supplyto upland streams:Influence on channeladjustme nt. In:CR.Thorne,J.CBat hurst &RD.Hey,(eds.):

Sediment Transport in Gravel-bed Rivers. John Wiley & Sons, Chicheste r,121-146.

Hickin,E.J.&Sichinqabu la,H.M.1988:Thegeom orphicimpactof the catast rophicOctober1984floodon theplanformofSquamishriver, southwestern British Colombia .Can.J.EarthSci.24,1078-1087.

Holtan, G.,Berge,D.,Holtan,H.&Hopen,T.1991:Paris Convent ion.

Annual reportondirect and riverineinputsto Norwegian costal watersduring the year 1990.Principles,resultsand discussions.

NIVA-report452A/91.

Koch,RW.&Srnillie,G.M.1986:Commenton'Riverloadsunderestima- ted byratingcurves'byR.I.Ferguson.Wat.Resour.Res.22,2121- 2122.

Langedal, M. 1995: Spredning av tungmetaller fra Knaben Gruber:

Forurens- ninqsstatus.Rep. 30, Inst. for geologi og bergteknikk, Trondheim, 1995,69pp.

Langed al, M.&Ottesen,R.T.1995:Particlebound MoandCu inthe Kvinariver basin.In RD.Wilken,U.Forst ner&A.Knochel(eds.):

Hea vy metals in the Env iro nment, Internat ional Conference Hamburg,Septembe r 1995.Vol 1,332-335.

MarianneLan gedal 101

Langedal,M.,Ottesen,R.T.&Bogen,J. 1996:Time variatio n in the chemi- calcompositionofsediments from tree Nor wegi an rivers-Natural andant hropogen ic causes.Arch.Hydrobiol.47,29-39.

Leenaers,H.1989:Thetransportofheavymetalsduringflood eventsin the pollut ed river Geul(TheNetherlands).Hydrol ogical processes 3, 325- 338.

Longva, 5., Schei, P.J. & Rensvik, H., 1994: Naturmiljeet i tall 7994.

Universitetsforlage t,Oslo,431pp.

Nanson,CG.&Cro ke,J.C1992:Ageneticclassificationof floodplains.

Geom orphology4,459-486.

Norusls.M.J.1990:SPSS/PC+4.0BaseManualfor theIBM Pc/XT/AT and PS/2.SPSS Ine.Chicago,511pp.

Ottesen R.T.&Belviken,B.1987:Use ofgeochemical mapsin regional planning.In:Wolff,F.C (ed):Geology forEnvironme nta l Plannin g.

Nor.geol.unders.Special Publication2,46-51.

Ottesen,R.T.,Bogen,J.,Be lviken,B.&Volden,T.1989:Overbanksedi- ment:A representative samplemedium forregionalgeochemical mapping.J.Geochem.Expl or.32,257-277.

Pitlick,J.C&Thorne,CR.1987:Sedime ntsupply,movemen tandstora- gein anunstablegravel-bedRiver.In:CR.Thorne,J.C Bath urst&

RD.Hey(eds.):Sediment TransportinGravel-bed Rivers.JohnWiley&

Sons,Chichest er,151-178.

RIVM(Nat ional Inst it ute ofPub licHealth andEnviron ment alProt ection ) 1991:Susta inab leuse of groundwater. Problemsandthreatsinthe European Communities.RIVM- repo rt no. 600025007, Novem ber 1991.

Sigmond,E.M.O.,Gustavson,M.& Rob erts D. 1984:Bedrock map of Norway1:1million Nor.geol.unders.

Walling,D.E.&Webb,BW.1987:Suspendedloadin gravel-bedrivers:

UK experience.In: CR.Thorne, J.C Bathurst & RD. Hey (eds.):

Sediment Transpor t in Gravel-bed Rivers, Joh n Wiley & Sons, Chichester,691-732.

Walling,D.E.,Webb,BW.&Woodward,J.C1992a:Somesamplingcon- siderations in the designofeffective strategies formonitoringsedi- ment associatedtransport. In:J.Bogen,D.E.Walling&T.Day(eds.):

Erosion and sediment transport monitoring programmes in river basins.IAHSPubl. no.210,279-288.

Walling,D.E.,Quine,TA&He,Q.1992b:Investigating conte mporary ratesof floodplainsediments.In:P.A.Carling&G.E.Petts(eds.):

Lo wland Floodpla in Rivers Geomo rpholog ical Perspectives. John Wiley & Sons,Chichester, 165-184.

Manuscriptreceived June/995;revisedversionaccepted November/995.