Saline groundwater extraction from the fjord delta aquifer , Sunndalsera , Mere og Romsdal , Norway

Saline groundwater extraction from the fjord delta aquifer , Sunndalsera , Mere og Romsdal , Norway

ODDMUN DSOLDAL,NORALF RYE,EIRIK MAURING&OLAM.SA: THER

Soldal,0.,Rye,N.,Mauring,E.& Seether,O.M. 1992: Saline grou ndwater extrac tion from the fjord delta aquifer, Sunnda lser a,Mere og Romsdal,Norwa y.Nor. geol.unoers.422.37-46.

Thefjorddelta atSunndatse ra hasbeen investigatedby geop hysic almeasurements,drilling and pumping inorder to examine the possibilities for abstrac tio nofsalinegroundwater foruse infish farming.A geological model ofthe delta'sIithostratigraphyispresented.The hydraulicconductivi- ty showsa cyclicitywith the highestvalues on the landwardsidein each cycle.In the outerpartof the deltathe salinegroundwateris located at shallower depths than along the river and on the landward side.The freshwaterlayerin the fjord andin the aquifer,the hydraulicgradientand the distribution ofhydraulic conductivity in the delta all contribute to a reduction in salinityduring pumpingof the groundwa ter.Itis estimatedthat lateral,freshwater-inducedrecharge from the river to the wells along coarse-grained layers contributes about 18% of the total yield.By means of two-levelpumping in the outer area ofthe aquifer,the salinity of the abstracted water may be increased to at least 20per mille.

Oddmundsotoet,^NoralfRye,Department of Geology. SectionB,UniversityofBergen,Allegt.41, N-500 7Berge n.Norway.

Eirik Mauring,OlaM.Seettier,Geological SurveyofNorway.P.O.Box 3006-Lade.N-7002 Trond- tieim,Norway.

Introduction

In many fish farms infectious diseases have caused great problems.These aretransmitted through the sale of infectedfry or by the esca- pe of infected fish from other farms. Such escapees may also infect natural fish stocks. In order to reduce these problems , ster ile, saline groundwater may be an important re- sou rce for on-shore fish farms.

Fjord deltas can be important aquifers for both saline and fres h groundwat er.Lowering of piezomet ric heads in coastal areas due to pumpinghasleadto salinizat ion of many aqui- fers throughou t the world (e.g. Inouchi et al.

1990). It is not known to what extent this ef- fect operates in afjord delta, mainly because the variation in hydraulic conductivity and the grou ndw ater flow regime is poor ly under- stood. Fjor d deltas are usually of Gilberttype (Gilbert 1885), characterized by coarse top- sets, steep sandy fore-sets and fine grained bott om- set layers (e.g. Corner et al. 1990).

Fjorddeltas havepreviously beeninvestigated by Prior et al. (1981), Boge n (1983), Syvitski and Farrow (1983), Kostaschuk and McCann (1987), Prior and Bornhold (1986,1988) and Corner et al. (1990).

It is to be expected that deltas located in sheltered areas with high relief will have a large supply of sediments and that the wave- energy will be relatively low (e.g. Colella et al. 1987).A progress ive lowering of sea-level in postglacialtimes in Norway has typically led to erosion ofolderdelta deposit s andredeposi- tion of coarser mater ial on the delta front.

Such erosion and redeposition leads to an unconformity between the top-sets and fore- sets, and to the coarsest sediments usually being found on the part of the delta clos est to the fjord (Corn er et al. 1990).Alluvial fans which are con stricted laterally have a large potential for build-ou t. They have a relatively small facies gradient downwards along the depo sit (Nemec and Steel 1988, Rachocki 1981). It is likely that the same is true for a fjo rd delta. These sedimentolog ical factors govern the grain-s ize distributio n and hydr au- lic conductivity and are probablyimporta nt for the hydrogeology of the delta.

Fjo rd deltas constitute a special type of coas tal aquifer because of the stratification of the fjord water. For example, at the delta front at Sunnda lsora,Mere og Romsdalcoun- ty,the upper brack ish layer ofthe fjord water

38 Oddmund So/da/,NoralfRye,EirikMauring&O/aM. Seetner

LEGEND: • Drilling no.

0

NGU -BULL.422.1992

Fig.1.Locationof the investigated area(inset).Location sitesofbore- holes,seismic prof ilesandpump ing wells .

I d

is up to 2-4 m in thick ness, while at 40 m depth the salinity is 31-32 per mille (A.Kiltel- sen,pers. comm.).The thickness of the brac- kish layervaries withthe freshw ater discharge from incoming rivers. The salinity of botto m water in Norwegian fjord s is close to 35 per mille (Scelen 1976).

The investigated area

The delta at Sunndalsera is situated between the mouths of two rivers, the Driva and the Litledalen rivers, and isbuiltoutinto theSunn- dalsfjo rd which is up to 200 m deep in its inner basin (Fig.1). Geological investigations on the delta have been carried out by Ander- sen (1984), Follestad (1984, 1987), Hillestad (1984), Kummeneje (1985, 1989), Henninq

(1985),Storre (1986), Nielsen (1988)and Sol- dalet al. (1990).

Bothinthe Sunndalenvalley (con taining the river Driva) and in the Litledalen valley there exist ice-margin deltas of Younger Dryasage (10,000-11,000 yrs BP).These sand and gra- vel depo sits were built up to c.140 m above present sea-leveland indicate theupper mari- ne limit in the area (Follestad 1987).

The surface sediments in the valley floor consist of coarse, fluvial deposits. Centrally in thedeltathereisatleast500mofQuate rna-

ry deposits (Hillestad 1984), probably domi-

nantlyfine-grainedmarinesedim ents(Follestad 1987).

The western portion of the fjorddelta,which is located at the mouth of Litledalen (Fig.1), has been investigated. The investigated area

NGU•BUll.422.1992 Salinegroundwaterextraction 39

Results

Fig.2.Simplifieddrilling logs(seeFig.1 for location).The marked boundaries indicatedifferencesin sediment type. although not necessarilyinobservedgrain-size.

processing included static corrections and band pass filtering. Velocity analysis of multi- channelrecords wasused to convert two-way travel time. Processed records are shown in Fig.3.

Sedimentology

Inspectionof aerial photographstaken indiff e- rent years reveals rapid changes in the river outlet at the delta front. This is in agreement with Follestad (1987),who states thatthe del- ta is river dominated. The Litledal River has probab ly only contributed to the construction of the western part of the delta. The delta associatedwiththeLitledalRiverisconstricted by the mountain cliffs on the west and the delta of the River Driva to the east (Fig.l).

The postglacial uplift of this area was most rapid between 10,000 and 9,000 years BP, with an estimated rate of 50 mm per year.

Between9,000yrs and thepresent,the avera-

7 8 9 10

.

~:~ :.:.

/!i ^:

~!;

? .

Hj ^;:.

,

~ :::

.

: ^{", ..}

:~~.

i: ~{-:

.;

~

, ^, .;;.

..

c ..

.

::

LEGENO

G ...

Ef]

~^Sand

D

D

Drilling no,

5 6

:

.

:::

:~:

tl .

~~ :~~ ,.

-::.- f· ^~

~::^r

~ { f..

.'

...:

Et

~

"

:

.::

~

:~:: :::

..::

W ^{;~ :::'}

.} I;:

t

" ^".'-':-

:~.; j,

~':,

z:

t

3

.

_.

:~~.:~:

10

12 ':''':-

14

16 ^"

K

_~

J: 20

'.

...

w ₂₂ 0

"

24 26 28 30 32 34

36 38 40

lies within the tidal zone.The investigation is part of a project to evaluate the possibilities for abstraction of saline groundwater for use at Akvaforsk fish farm.

Two water-works supplyingfresh groundwa- ter are situated at Sunndalsera (Fig.l). One is for municipalsupplyand one supplies Akva- forsk,a research station for aquaculture,whe- re salmon fry and halibut are reared. The nearest water-works is located 400 m to the south of the investigated area.

Methods

Ten borehol eswere drilledand sampled.Test- pump ing was carried out in two pumping wells over a three month s period. 600 m of seismi c reflection pro files were measured (Fig.l).

The wells were drilled using a Borro s ham- mer-rotation mobilerig emp loying water asthe drilling fluid.Penetration rates,water pressu- resand sound-responsewere recordedduring boring. These data were used to inter pret sediment type. After borinq, 5/4" diameter cast-iron well-points, with,elongated slots at the tip,weretemporarily installed to the desi- red depth. Samples of sediment and water werecollected and tne water temp eraturewas measured. The borehole logs are shown in Fig.2. The water samples were anaiyze d for cations by atomic emiss ion spectroscopy and anions by ion-chrornatoqraphy. Alkalinity was measured by Gran titration,pH with a glass- elect rode, and electrical conductivity (E.G.)by conductivimeter. A field ref ractometer was used for the majority of the measur ements of salinity , a subse t of which were controlled by measuring osmotic press ure with a 'Semi- mikro osmometer,type M'. Salinity is directly related to disso lved solids (1per mille = 1000 pprn)and electricalconductivity(salinity=0.68 x E.G.(inmS/cm)).The results of these analy- ses are presented in Soldal et al. (1990).

The seismic reflection profiling was carried out using the «optimum offset»method (Hun- ter et al. 1988, Hunter and Pullan 1989) and recorded with a digitalseismograph (Scintrex

S-2 'Echo' with 24 channels). Six profiles

were measured. The distance betw een the energy source(in-hole shotgun with 12gauge shells)and the first geophone was 18m.The geophoneswereplacedwiththreemetreinter- vals for profiles number 1 and 2, and two- metre intervals for profiles 3 to 6. Seismic

40 Oddmund So/da/,NoralfRye,Eirik Mauring&otsM.Ssettie r NGU•BULL.422.1992

upperboundaryof these sedimentsis astrong reflecto r which can be recognized in most of the seismic pro files .The fine-grained depo sits are near horizont aland can beinterpreted as dista l bottom sedim ents.

Seismic pro files 1 and 2 (Fig A) show incli- ned reflec tors which flatten out to the north.

In the area with more steeply-dipping layers, the sediments encounte red in the boreholes are at their coarsest (Fig.2). The preferred interpretation is, therefor e, that the coarse, grave llysands weredepo sited asfor eset beds.

The coarsest sediments are thought to have had their origin in submarine channels at the river mouth (e.g. Prior and Bornhold 1988).

The finer,sandy sediments were probably eit- her deposited somewhat more distally from the river mouth or laterally,at the sides ofthe river mouth channels. In front of the river mouth,where the sediment supply wasgrea- test,the delta advanced rapidly withtransport along subaquatic channels.

Fig.4reveals the cyclicity of sedimentation, which we interpret as being due to erosion of older depo sits follow ed by deposition of new foresets, due to meandering of the two rivers. This resulted in step-wise change s in grain-size .Withineachcyclethe coarsest sedi- ments are located nearest land.

Profile no.6 Profileno.4

Prolileno.5 Profile no.3

Fig.3.Processed seismicreflectio nprofiles, see Fig.1 for locatio ns.

ge emergence rate was about 4 mm per yr (Svendsen & Mangerud 1987). The current rate of uplift is 3-4 mm per yr (Serensen et al. 1987).Theemergence,erosion and subse- quent sediment supply from the ice-margin delta has been of great importance in gover- ning the postglacial construction of the fjord delta.

The seismic profiles (Fig.3) reveal several reflector sbeneat h the fine-gr ained sediments.

These are,however,of little relevance to this investigation .Inter pretationof shallow erstruc- tures (0-15m)from the profiles alone is diff i- cult,butin combination withobservatio nsfrom the boreholes,a mode lcanbededuced(FigA).

In the investigated portion of the delta the

coarse topset (cobbles, gravel and sand) is underlainbya c.20 m thick sequenceofsand which wedges out towards the ice-margin delta(E.Danielsen,pers.comm.).Under these layers there is a sequence of fine sand and silt/clay with a low hydraulicconductivity.The

Hydrogeology

The uppermost part of the delta is an uncon- fined aquifer,limited by the fjord and the Lit- ledal River. Test pumping has been carried out in the two pumping wells (Fig.1) over a three month period. The transmissivity was calculated as 1.6 x 10-' m'/s (1382 m'/d) and the storage coefficient as appro ximately 0.3 (Soldal et al. 1990).

Thesalinity of thegroundwaterin theinvesti- gated area varies betwee n1and32 permille.

The measurements reveal that the freshest groundwaterislocatedalongtheLitledalRiver and with in the southern part of the delta.

The salinity increaseswithdepth and toward s the fjord.The investigated part of the delta is floodedby water with asalinity of up to 5 per mille during high tide. The average difference betwee n low and high tide in the fjord is 1.5 m. Gro undwate r levels inthe delta change as afunctionof thetidesinthefjord.Consequent- ly thereis also a changeinhyor auticgradient in theouter part of the aquifer.Thenet effect

NGU-BULL.422.1992 Saline groundwater extraction 41

240m 4

Fig.4. Interpr etation of ^{11. . .}4f--- - - - Profileno.1----_~^{1...4- - - - - Profileno.2 - - - --}

the seismicprofilesand WELLS

borehole data. Hydro- S 6

~7~~~[~I~~~3J

geologicalobs erva tio ns ?

are alsoused in the in- terpre tation s. Some of the borehole s show no changes in sediment type acrossthe seismic reflectors. The maxi- mum depth to the groundwater level is1.5 m below the surface.

1 "'-

3 _1

1 _

- -I 1 ...

LEGEND

~Gravellysand with

~pebble s

Sandandgravelly sand

o

B

1

I

of the tidal fluctuations (Serfes 1991)is a hy- draulic gradient from the river towards the aquifer,asillustratedby the salinewaterdistri- bution.

Many studies have show n that sedimentary facies isadeterminingfactor for the hydraulic conductivity(permeability)ofclastic rocks (e.g.

Dreyer et al. 1990),beca use there is a known corr elation between grain size and hydraulic conductivity in a porous aquifer (e.g. She- pherd 1989).If the presentedsedimentological modelis corr ect,theremust be stepwise chan- ges in the hydrau lic conductiv ity of the delta, within each step the highest values being on the landwar d side.

The zone of mixing between fresh and salt water is wide and the overall distribution is independent of the sedimentology (Soldal et al. 1990). Boreholes 2 and 4 (Fig.5) provide

an exception to the overall picture of a gene- ralincrease in groundwater salinity downwards in the aquifer.Decreases in salinity (E.C.)are detected between 17 and 19 m depth in boreho- le 2 and between 11 and 15 m depth in boreho - le 4. In borehole 3, groundwater becomes gradually more saline downwards in the aqui- fer, but the measurements of Br- and alkalini- ty show a considerable vertical variationinthe profile (Fig.6).The large increase in alkalinit y at 21 m depth in borehole 3 is probably due to analyticalerror because electricalconductivi- ty, pH and Ca-content do not exhibit a similar variance(Soldalet al. 1990). Bromide ions are good indicato rs of sea water intrusion (Morell et al. 1986). The local decreases in salinity, alkalinity &bromide at some levels in bore ho- les no. 2,3 and 4 are interpreted as due to inflowof fresher water along coarse foresets.

42 OddmundSo/da/,NoralfRye,EirikMauring & O/a M.Seettier NGU· BULL. 422.1992

I 8

i

~¹⁰

)

"-

~¹² Na+

" ^{- Cl'}

16 EC Temp.

18

mon th s.In thescreened level of thewells,the average salinity of thewater was 27 permille before the pumping started .Afte r two weeks of pumping the salinity of abstracte d water had stabilized at 10-14 per mille in well Iand at 7-8 per mille in well 11. There is a large difference in the salinity of the water from the two wells in spite of their being located adja- cent to each othe r.Nothinginthe local geolo- gy can exp lain this difference. One poss ible explanat ion is that a natur al grave l filter has been developed outside the well screen and casinginwell 11.Thismightlead totheintroduc- tion of fresher groundwate r from theaquifer's coarse top layer dow n to the slotted screen.

Inborehole 7,however,whichislocated betwe- en the river and thewells, influx of waterwith 2 per mille salinity was detected at 11-12 m depthduring thetest pumping.Inaddition,the sedimentological modelindicates thatthereis a highe r hydraulic conducti vity in the part of the aquifer nearer the river than in that part nearerthe fjord.Consequently,it appearsthat the inflow of water from the Litledal River causes thelow salinity in well11. Well 11,which is located closest to the river,receivesa hig- her prop ortion of this fresh water than well I,which presumably derives much ofits water from the fjord-dominated part of the aquifer.

It is assumed that the fresh/brackish water that reache s the wells has a salinity of 2 per mille, as measured in borehole 7 during the pump ing period.The fjord water whichcovers the area at high tide also hasa similar avera- ge salinity.The salt water component can be assumed to have an average salinity of 27 per mille, as measured in bor ehole 7 prior to pumping. A simple relation ship between salt and fresh water inflow to the wells is:

Temp.

Temp.

7 8 9

Temp·I'CI EClm S/cml 7 8 9 TempI'CI

EC(m S/cm]

Sa1nty1'"1 EC lm S/cml 2 3 •

1 2 3 4 5 6 7 8 9 Temp I'CI 10 20 30 40 50

o 05 1.0 15 2.0'''')

0'

I 8

I

~10

W

I

)

Cl12

"

II

16

18

20 Na+ Cl' EC

DRILLINGNO. 3

0 0.5 1.0 1.5 20I"') ^{10 20 30} '0 50

0

2 3

•

\ \ \

I 8

\ \

1

I

\

i ^I

ti:¹⁰

W

\

Cl12

"

16

I

18

I ^j

20

Na+ Cl' SalinIty EC

DRILLING NO.4

( A x F ) + ( B x S ) = (A + B) W

Fig.5.Changes in differentparameters in three boreholes located at the margins of the investigated area. Plotted pointsarelocatedin the centreof the 1 m samplinginterval.

Test pumping

Two abstraction wells were placed 20 m NE of borehole 7 (Fig.1). The distance between them was 2.0m and theywere screened fro m 8 to 18 m below the surface.The sediments below 18mdepthconsisted of fine sand.Well 11 is located nearest to Litledal River. The to- tal capacity of the wells was 0.02 m'/s (1200 I/min) and the test-p umping period was three

where A= Volume of fresh/brackishwater,B=

Volume of salt water, F= Salinity of fresh/

brackish water, S=Salinity of salt water, W=

Salinity of the water in the well. A + B = 1.

F = 2 per mille,S = 27 per mille.

Afte r one month's test-pumping,the average salinityinwellI was 12.0per mille,andinwell 11 7.5 per mille.This gives:

Well I: A = 0.6 (60%) and B = 0.4 (40%)

Well 11: A = 0.78 (78%) and B = 0.22(22%)

NGU-BULL422,1992 Saline groundwater extraction 43

...

···1

...

.... . ...

Br' [ppm)

Alkalinity [mmol/I)

I

'" _'"

'" " .

'" '" '"

Alkalinity [m mol/I)

2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17

30 40 50 60 70 80 90 Br-[ppm]

I· ...

'" '"

....

LEGEND

'iij

>.

<3ca

1 2 3 4 5

6 b":':'-!''f"I---LoT:T---~""O;;::..__---____j 7

8 E

1 ~

li

13

~: ^iHi\W{

:: i ~; ,~

~ ^W;?N

~ ~!tt11iJ

Fig.5.Lithostratigraphy andhydrochemical profilesfrom borenolenO.3.Note the decrease inBr"-content and thechange in alkalinity in thecoarse layers.The extreme'alkalinity at21mis probably due to analytical error.

The saline water constitutes 40% of the total yield in well I,and 22% in well 11. If it is assu- med that any inflow from the top part of the aquifer is similar in the two wells, the 18%

differe nce in salinity is due to the lateral flow of fresh water.

The observat ions in boreholes 2, 3 and 4 reveal that fresh water mixes with saline wa- ter along coarse foresets. Small variations in grain-size lead to great differences in the hy- draulic conductivity and thus to an increased interfin gerin g when the gro und water velocity increa ses(Gillham & Cherry 1982). The hydrau- lic conduct ivity , which is highest in the area betweenthewells and the river,the interfinge- ring effec t and the hydraulic gradient from the river to the aquifer cause this lateral flow of fresh water.

After two months of pumping, a test with two-level pumping was performed(Fig.7). The screen level was adjusted to 5-10mbelow the surface in wellI andto 12-18 m in well 11.The total pumping rate remained 0.02 mvs, This

change caused thesalinity inwellI to stabilize at0-5 per mille, whereas inwell11 itincreased to 15-16 per mille. Using the same relation s- hip as above, we estimate that 54% of the yield in well 11 is saline water. In well I the percentage of saline water is around 2% . It is estimated that 18% of the fresh water in well 11 was transmitted laterally through coarse layers from the river-dominated area of the aquifer and 28% was due to vertica l flow of fresh groundwater from the top of the aqui- fer. Figure 7 shows a qualitative model of gro undwater inflow to the wells.

No drawdown effect was observed in the freshwaterwells at the Akvaforsk waterworks during the test-pumping period. The radius of influence around wellsI & 11 is thus small or is limited in the direct ion of the waterworks.

Thesalinity of thewater obtained from wells I& 11 is too low to be used by the fish farm. Preliminary investigations indicate that the general stratigrap hy of the whole delta front is similarto that inthe investigated area. They

44 Oddmundsotoet,Noralf Rye.Eirik Mauring & OlaM.Scether NGU-BULL. 422.1992

WELL II WELL I

Salinity:7·8 %0 (Brackish -salt) Temp. :35-50'C

Salinity:10-14%o(Brackish·salt) Tem p. :4.9-6.2'C

®

WELLI

Sahlllt y:15-16%0[Brac k rsn-salt) Temp .56-60'C

Salinity:0·5%,(Fresh tobrackish) Temp. :7.0 -8.5'C

2m 2m

I

I 10

l-a.

~ 12 14 16 18

-

-

I

I 10

l-a.

w 12

o

14

16

18 LEGEND:i

EZJ ~~~~:It~i~~rp~~~~:nd l22JJ Fore sets layers,sand

B Well screen

-

^-

Fig.? A qualita tivemodelfor inflow tothewells.A:pumping from one level,B:two- level pumping.Design ationofthe water types is in accordancewith Stuyfzand.1986.

also show that the aquifer's coarse top layer carries relatively fresh water throughout the whole delta.Thus.the potentialinflow of fresh water from the aquifer'stop layer to groundwa- ter wells will probably be large everywhere on the delta. New pumping tests. 200 m to the north of borehole 3,reveal that the salini- ty of the abstracted water eventually decrea- ses to below 15 per mille with a pumping rate of around 0.04m'/s.Preliminary investigations do, however, indicate that a reduction in the pumping rate to c.0.02rnvs may give a stab- le situation with a salinity of 25-30 per mille (G. Storm,pers. comm. 1991).

The maximum groundwater salinity will be att ained in areas with no lateral flow from the fresh water aquifer.Consequently,wells desig- ned to abstract saline water must be placed far out on the delta. The lateral fresh water flow that constitutes c. 18% of the well yield in the investigated area can then be avoided. If the wells were placed far out on the delta, at locations where the net hydraulic gradient is very small and where coarse fore sets with connection to the top of the aquifer are lac- king. 72% of the well yield would be saline water. Using the same salinities of fresh and salinewaters as above,and the same relations- hips, we estimate that the salinity would be

at least 20 per mille in the water abstracted from the deepest well usingtwo level pumping.

Conclusions

Investigations at the extreme northern edge of the delta (Kumm eneje 1985,1989;E.Daniel- sen pers.comm.1991) didnotreveal any coar- ser-grained materialthan wasobserv edin the investigated area.The absence of atrend to- wards coarser -grained material outward s in the investig ated part of the delta is prob ably due to the fact that this part of the delta was formed after 9000 yrs BPduring aperiod with little uplift. Uplift may however be respon sible for the wedge-shaped form of the coarser- grained sediments between the fjord- and the ice margin deltas.

The wide zone of mixing between fresh and salt groundwater in the aquifer is probably the result of the influence of tidal changes in the fjord on piezometric heads in the delta aqui- fer. In a dynamic system such as the one described here,the Ghyben-Herzbe rg principle (Badon Ghyben 1888, Herzberg 1901) is not applicab le. Glover (1964) has developed an approximate equation for the shape of the freshwater-saltwater interface in a situation

NGU-BULL.422.1992

where groundwater dynamics are involved.

Theinternal dist ributionof hydraulicconduct ivi- ty,thehydraulic gradient andthetidal fluctuati- ons are probably the most important factors influencing the flow of grou ndwater in the aquifer. The coarse-grained top layer carries large amo unts of fresh water and to a large extentnegatestheflowof deepersaltground- water to the investigated wells.

If ahigh salinity is to beachieved,two-level pumping is the simp lest solution.Other possi- bilitiesto obtaingro undwater withahigh salini- ty are to reduc e the pumping rate of the wells,to infiltrate saline water from the fjord into the shallow aquifer,or to search for dee- per aquifer horizons.In order to reduce fresh water inflow, it is important to avoid localities with hydrau lic gradients from rivers and high- permeability for esets.

Saline gro undwater canbe pumped out from wells I & 11 at Sunndalser a in large amounts without causing conflict with the utilization of freshgroundwaterintheexisting water-works.

Acknowledgements

WethankAkvaforsk forpermission to publish these data.

The work was supported financially by Akvatorsk ,NTNF and NVE-Vas sdragsdire ktoratel. Three anonymousref ere- es.D. Banks and B.Rob ins corrected the language and suggestedimpro vements.J.Elling sen helpedwith the figu- res.Tothese persons and institutionswe profferour since - re thank s.

Reference s

Andersen.A.B.1984:Grunnvannsundersokelserpa Sje land - sera i Sunndalkornmune,Mere og Rom sdal. Nor. ge- 01.unaers.report84.149.

BadonGhyben.w.1888:Notainverbandmet de voorge no- men putboring nabij Amsterdam . Tijds chr ift Konink lljk tns tituutvanIngenieurs.The Hague.Netherland s.8-22 . Bogen.J.1983:Morphology and sedimentology of deltas

in fjord and fjord valley lakes. Sedimentary Geology 36.245-267.

Corner.G.D..Nordahl. K..Munch-Ellings en. K.&Robe rt- sen,K.R.1990:Morpholog yand sedimentology of an emergent fJord- head Gilbert-type delta: Alta delta.

orway.SpecialPublicat ion s Int.Assoc.Seoimentoto- gists to.155-168.

Couela.A..deBoer.P.L.&Nio.S.D.1987:Sedimentolog y ofa marine inlermontanePleisloceneGilbert- type fan- delta comp lex in the Crati Basin.Calabria. southern Italy.Sedimentotoqy34.721-736.

Salinegroundwaterextraction 45

Copper Jr..H.H..Glover. R.E.. Henry. H.R.&Kohout.FA 1964:Sea water in costal aquifers .US.Geol.Survey Wat.Supp.Paper 1613-C.84 pp

Dreyer.T..scneie.A.& Walderhaug.O.1990:Minipermea- meter- based study of permeability trend s in channel sand bodies. Amer. Assoc. Petroteum Geot. Buttettn 74.359-374.

Fouest ad,BA 1984:SunndalsoraBOP109110. Quate rn ary geologicalmap.Nor.geol.unders.

Follestad.BA 1987:Sunndalsora 1420 Ill. Beskrivelse til

~.va rt<erg eo l ogi s k kart - M 1:50000.Nor.geol.unoers..

Skrofter 79. 1-32.

Giloert.G.K.1885:The topoqrapnic features of lake sho- res.US.Geol.Survey.5thAnnualReport. 69-123.

Gillham.RW.&Cherry.J.A.1982:Contaminant migration In saturatedunconsonoatecgeologicoeposits.In: ara- simnan T. . (ed.). Recent Trends in Hydrogeology.

Geol.Soc.America.Spec.Publ. 189.31-62.

Glover. R.E. 1964: The pattern of fresh -water flo w in a coastalaquifer.In:Sea Water in Coastal Aqu ifer s.US.

Geot.Survey Wat. Supp.Paper. 16 13-C.32-35. Herzberg.B.1901 :Die Wasserversorungeiniger Nord seeba-

cer. Journal Gasbeleuchtung und Was serversorgung.

44.815·8 19.842-844.

Hutestao.G.1984:Seismiske malinge r.Sunndals era.Nor .

geol.unders.report 84.097.

Hunter. J.A.. Pullen. SE. Burn s. RA . Gaqne. R.M. &

Good.R.R.L. 1988:Applications of a shallow seismic reflection method to groundwaterand engineeringstu- dies.In:Ministry of NorthernDevelop ment and Mines.

Proc.01 Explorations 87.Spec.Vol.3.704-71 5.

Hunter.JA &Pullan. S.E.1989:The optimum offset shal- low reflection technique.In:Proceedings ofthesymposi- um on the application of geophysicstoengineering and environmentet prootems. March 13-16. 1989.Colorado School of Mines.Golden.Colorado.

Inouchi,K.. Kishi,Y. & Kakinuma.T.1990:The motion of coastal groundwater in responseto the tide.J.Hydro lo- gy. 115.165-191 .

Kostascnuk, R.A. and McCann. S.B. 1987: Subaqueous morphology and slope processesin a fjorddelta.Bella Coota,Brotish Columb ia.Canad.J.Earth SCiences.52.

52-59 .

Kurnrneneje 1985:Sunndalkommune.havne- og industri- plan Hasoran.Orienterendegrur.nundersokelserav sta- brntets -og tundamentennqstorbold.O.KummenejeAIS.

report no.I.0.5260.Trondheim.

Kummeneje 1989: Sunndal kommune.havneplan - Haso- ran. Suppterende qrurmunderseketser og stabilitets- vurdering av ytre oeuaornra ce.O.Kummeneje AIS .re- portno.3.0.5260.Tro noneirn .

Morell. I..Med ina.J..Pulido .A.& Fernandez-Hub io,R.1986:

The use of bromide andstro ntium ions as indica tors of marine intrusionin the aquifer of Oropesa-Tor re blan- ca (Castellon ,Spain) Proceedings oftne9th Salt Wa- ter Intrusion meeting.

oe«

Nemec.W.&Steel.R.J.1988:What is a fandelta and how do we recognize it? In:Nemec.W.&Steel.R.J.(ed s.) Fan Deltas: Sediment ology and Tectonic Settings.3- 13.Blackie and Son.London.

Nielsen.J.T.1988:GrunnvannsforsyningtilAkvatorsk,Sunn- catsora.More og Romsdal. Nor.geol.unders.report 89.001.

Prior.D.B.&Bornh old,B.O.1986:Sed iment transport on subaqueous fan -delta slopes.Britannia Beach .British Columbia.Geo-Marine Letters.5.217-224.

Prior.D.B.&Born hold.B.D.. 1988:Sub marine morphology and processes of fjord fan deltas and related high-

46 OddmundSo/da/,Noralf Rye,EirikMauring&oteM.Seetner NGU-BULL. 422.1992

gradient systems:modern examplesfromBritish Colum- bia. In: Nemec. W. & Steel. R.J. (eds.) Fan Deltas: Sedimentologyand TectonicSet tings.125-143.Blackie and Son.London.

Prior,D.B.Wiseman.W.J.&Bryant.W.R.1981:Submarine chutes on the slopes of fjord deltas. Nature. 290.

326-328.

Rachocki,A.198 1:AlluvialFans.An attemptatan empiri- cal approach .John Wiley&Sons.Ltd.16 1pp.

Ronning.J.S.1985:vernkate etektriskesonderinger ogSP- malinger pa Sjolandsora.Sunndal.Mor eog Rom sdaJ.

Nor.geol.unaers.repo rt 85.012.

Santing.G.1986:Circu lat ingflowin a confinedsaltground- water body. Proceedings of the 9th salt waterintrusion meeting.Dettt, 12 - 16May 1986.73-84.

Serfes. M.E. 1991:Determining the mean hydr aulic gradi- ent of ground water affected by tidal fluctuations.

Grou ndwater.29.549-555.

Sheparc,R.G.1989:Corre lationsof permeabilityand grain size.Ground water.27.663-638.

SoldaJ. 0..Rye.N.& Mauring. E.1990:Hydrogeolo giske undersokingarved Litledalselva.Sunnd alsora.Sunnda l, More ogRomsdal. Nor.geol.unders.report 90.067.

Storm.G.1986:Hydroqe otoqlskeundersokelserpasjerano- sera, Sunndal kommune. sluttrapport. Nor.geol. un- ders.repo rt 86.176.

Stuyfzand. P.J. 1986: A new hydrochem icalclassification ofwatertypes.Principlesandapplicationto thecoastal dunesaquifersystem of the etherlands.Proc.9thSalt WaterIntrusion Meeting.Dettt. 12-16May 1986.641-655.

Svendsen.J.J.& Mangerud.J.1987:Late weicnsenanand Holocenesea-leve lhistoryfor across- sectionof wes- tern orway.J.Ouart. Sci..2.113-132.

Syvitski,J.P.M.&Farr ow.G.E.1983:Structuresand proces - ses in bayhead deltas: Knight and Buteinlet. British Columbia.SedimentaryGeology.36.217-244. Saslen.O.H.1976:Generalhydrogr aphy offjords.In:Skres-

let. E.. t.embe. R.. Matthews.J.B.L. & Sakshaug.E.

(eds.) Freshwater on the Sea.Association Norwegian Oceanographe rs.Oslo.43-49.

Sorensen.R..Bakkelid.S.& Torp.B.1987:Nasjonalatlas for Norge.map sheet 2.3.3. landhevning (land uplift).

Statens Kartver k.Honefoss.

Manuscript received November 1991; revised typescrip t acceptedFebr uary 1992.