Groundwater contribution to a mountain stream channel , Hedmark , Norway.

Groundwater contribution to a mountain stream channel , Hedmark , Norway.

SYLVI HAL DORSEN,JENS-OLAF ENGLUND,PERJ0RGENSEN,LARS A.KIRKHUSMO& DAG HONGVE

Haldorsen,S..Englund,J-O.,Jorgensen,P.,Kirkhusmo,L.A. & Hongve,D. 1992:Groundwater contribut ionto a mountain stream channel,Hedmark.Norway.Nor.geol.unders . 422.3-14.

The Skvaldra stream in Hedmark,Southern Norway is characteri zed~y a stable base flow.This baseflow is controlled by a significant groundwater supply from Quaternary deposits.The area was chosen for an attempt at quantificat ionof the basenow'sthree groundwater components:(i) ground water from numerous springs on a sprinq-honzon at the boundary between permeable sediment-flowdeposits in the upper valley sideand compact basaltill depositedalong the lower part of the valley sides andthe valley bottom;(ii)groundwater from aquifers in compact basal till.

topographically lower thanthe spring horizon;(iii)deep groundw aterfrom till or underlyingbed- rock.Flood events in theSkvaldr a are due to surface run-off,usually confined to the valley bot- tom along the stream channel,or due to very.shallow subsurfaceinterflow,withvery little or no groundwater component. Even when the baseflowis very low«0.05rn's"),groundwater from the springs makes up more than70%of thetotal flow.Groundw aterfrom the compacttill constitutes about10%ofthebaseflow,while deep groundwatermakes up lessthan20% .Duringthe summe r the springcomponentis higher.During a year thebaseflowvariesbetween0.03and0.1rn's'",Most of thisbaseflo wisduetogroundwater,andthegroundwater componentis at least1x10'm'per year,corres ponding to10%ofthe annualprecipit ation.

Sylvirtetaorsen.PerJerqe nsen,Jens-OlafEnglund,tnst itutt forjord fag,seksjonfor geologi,Nor - ges landb rukshogskole,Postboks 28, 1432As,Norwa y.

Lars A. Kirkhusmo,Norges geologiske underseketse.Postboks3813 - Ullev;jlHageby, 0805-0510 ,Norw ay.

Dag Hong ve,Statens ins tituttfor toncenetse,Geitmyrsveien 75, 0462Oslo4.Norway.

Introduction

Terr ain covered by till or with a surface of strongly frost-weathered bedrock characteri- zes the mountainous regions ofsouth central Norway.In some mountaina: easthe ground- water contribution to streams and rivers is considerable. It gives a stab leand significant baseflowand has an important influence upon thechemicalcompos itionofthesewaters.The routing of water along surface and sub-s urfa- ce flow paths in such mountainou s areas is important because it influences the timing of stream flow and the chemical composition of water reaching the stream. In a catchment consisting of fractured bedrock overlain by glacial sediments,the rivers may receive wa- ter from deep regionalgroundwater flow paths as well as shallower water from the glacial overburden. Two general approaches have been used to better define the sources and actualflow paths ofwater incatchments (e.g.

Horton 1933, Pinder & Jones 1969, Kirkby 1978, Germann 1986, Kennedy et al. 1986.

Sklash et al. 1986); (i) physical observ ations from field studies, coupled withtheor etical flow modelling. and (ii) measurements of changes

inchemicalandisotopicratiosof flow compo- nents. Observations from limnigrammes give some inform ation about the hydrolo gical pro- perties of a catchment. but these cannot be anything but qualified guesses without field studies. Hewlett (1982) stated that «no graph i- calor mathematical operationperformed on a hydrograph will revealthe source or pathway of stormflow».Diffe rent combinations of sour- ces and pathways can lead to quite similar hydrographs.This study,therefore,usesche- micaldatain additionto the volume,rate,and timing of stream flow to clarify the sources of ground water andthe specific pathways along which groundwater moves through a catch- ment.

The Skvaldra catchment in Hedmark was selected for the study because the Skvaldra stream receives groundwater from thick Qua- ternary deposits along the valley sides and becausehuman influenceis small.Itis therefo- re well suited for an attempt to quantify the different groundwater components whichcon- tributetotheSkva ldra'sstreamflo w.Theinves- tigat ion ismainlybased on datafrom the peri- od 1987 to 1991.

4 By/viHa/dorsen,Jens-Ote!Eng /und,Per Jerqensen,LarsA.Kirkhus mo&DagHonqv e NGU· BULL.422.1992

Study area

Topography , climate and land use

The study was carried out in the upper part of the Skvaldra catchment,Godlidalen,which covers 13.5km'(Fig.1).Godlida lenis anasym- metric valley with a rather flat valley floo r, bounded by mountainous areas to the north and east, and low er hills between the Skvald- ra catch ment and the main Astdalen valley to the west. The altitude varies from 875 mabo- ve sea level (a.s.l.)in the valley bottom up to 1090 mat the highestmountain summit inthe east. The Skvaldra stream forms a dendritic pattern, with most of the tributary stream s

coming from the east. In the western part of thecatchment thereare twosmalllakes(Fig.1).

The annual precipitat ion is about 1100 mm (Fig. 2). Winters are normallycoldwithtemp era- tures wellbelow

a oe ,

and the snow fallsfrom the beginning of No vember. The main snow melt period is usually in the latt er part of May.Two of the study year s,1989 and 1990, were exceptiona lly mild,withlittle or no snow accumulationuntil lateDecember andwith the main snow melt in late April.

Peatland constitutes about 40 percent of

Godlidalen, and dominates completely below

the timbe r line. The Skvaldra catchment is used only for extensive grazing by sheep, huntingandrecreation.Othersources of pollu-

ion are insignificant.

40 50

D 5

o J F 60

o

15· 30

20 '

I "

10 .^{00 .: •} I °

0 ~ .., ^• f' ..

o~~!::-~^••~^••~~^{••-', ...•}~- ~:.~:-^,_:~~~~^{... •.:.}".: ;,::

i ;..;.

~_h_

..

~!,·:

J F M A M J J A S O N D

Discharge m3s-' 20-

Precip itat ion (mm) 70

Fig.2. Precipitation atSjusjeen 1990anddischarg e of the Skvaldra in1990.

...-/ Sk vald r aca tchm ent

o Gr ound w a t ertu bew el l

Y Precipitat ion gauge

""'V Limnigraph

' - - - ' _---''----'3km

Monit oredpa rtof

ca tc hment area River

Peat la nds Sprin ghorizon

Fig.1.The Skvaldra catchment withpositionofthemonito- redpart(Godlidalen).The profilea-a'is showninFig.4.

The western tube well is number 15006,the middle one 1500 7and theeastern one15008.Inset:0.d.=0sterdalen, G.d.=Gudbrandsdalen.

zo

C aJ Cr:

r

..

'"

!"

s

'"

ma.s.t 110 0

1050

100 0

950

900

850 a' 3000m 2500

Dommantgroundwa ter

oE- rec harge area ~

Thinsacrments andfrost Permeabl e weathere d bedroc k glacialseo tm ents

Iuva

l . -- -;r

I : ,;;:/

'0

--- j~

7

.r:

"

1500

springs 1000

saturatedflow unsaturatedflow Compactbas al (Ill and peatland

====

^{pe at}

- - - groundwaterlevel

·

^c.

· ·

-o -oc

·

~

~

500

Fig.4.(a)W5W-ENEgeologicalandhydrogeologicalpro file a-a' through thesouthern partof the Godlidalen based onthe seismic investigations of Hillest ad (1990).Position of pro fi le issho w nin Fig.1.(b)Main hydrologicalbudget of the Skva ldra.

1- 3·Base flow 4 Floo devents

o a

"I--''---LJ....:,I!:l""''''-''-'c=..,..J:==c-=~--_r.---__,_- --,_--N..J

20 00 Permeable

sediments 1km

Tate ru ngskampen

- -<..---

Fig.3.Dep o sitional model fortheQuatern ary sedime nts in the Skvaldra catchme nt.

Top:Relat ion tothe glacierduringglaciation.

Bott om:Present day,show ing positio ns of thesprings &

peatl and s.

Spri ng A- Springs

~= ~ ~=============

6 SylviHaldorsen,Jens-OlafEnglund.Per Jerqertsen,Lars A. Kirkhusmo &Dag Hongve NGU·BULL.422.1992

Bedro ck ge ology

The Skvaldra catchment is situated in the middle of a Late Precambrian sedimentary rock basin.The bedrock belongsto the upper part of the Brett urn formation and consists of alternating fine-gra ined conglomerates and arkosic sandstones (Siedleck a et al. 1987).

Seismic studies (Hillestad 1990) showed that the bedrock underlying the Quaternary sedi- ments can be regarded as only moderately fractured, Le. with seismic velocities betwe en 4500 and 5500 ms-I.

Quaternary geology

The valley bottom and the lowest part of the valley sides (about 6.5km')are covered by a compact silty basal till of thickness c. 10 m (Kehler 1985,Hillestad 1990)(Figs.3 & 4).Its petrog raphic and granulometric composition is very homogeneous because it was formed by comminution of the arkosic bedrock alone (Haldor sen 1982, 1983). Glaciofluvial material is found locally along the Skvaldra and this is probably underlainby till. Bedrock expos u- res are almost absent.

The upper part of the valley sides (about 7 km') is covered by a 5 - 10 m thick layer of more permeable and less compact material, much of it unsorted. Debris flow deposits dominate. The sediments as a whole are in- homogeneous and form a hummocky surface. Meltwa ter channels are frequent,andin many of these only coarse boulder lags remain.

Some of the channels extend deepinto the till.

The distinct boundary between compact till and more permeab le material canbefollowed at an altitude of 950 m in the south up to about 1000 m in the northern part of God- lidalen, where it continues into a prominent moraineridge .The boundary marks theupper limit of an active glacier during the Weichseli- an period (Fig.3).Compact tillwas thendepo- sited subglacially,whilethe material along the ice-free valley sides was exposed to water sorting, flow and frost activity.

Above an altitude of about 1050 m the co- ver of Quaternary sediments becomes thinner and thesummits are characterized byboulder fields, formed by intensive frost activity,pro- bably during a period of periglaci al climate.

General hydrology of the Skvaldra

The baseflow of the Skvaldra is defined as thelowestdischargewhichisobservedbetwe-

en clearly separ ated flood events, and it is normally between 0.03"and 0.1 m's-I(Fig. 2).

The annual baseflow is 1- 3x 10' m',corre- sponding to 110- 220 mm (Le. 10 - 20%)of the precipitation.During snowmelt floods,the discharge inthe Skvaldra canreach 20m's-I.

During the summer,stormflows of more than 2 rn's' have been measured.

Monitoring

The precipitation in summer was measured with a Plumatic precipitation gauge (Fig. 1) and the data wasprocessedby theNorwegian Meteoro logical Institute (DNMI Station 0676 Astdalen-Skvaldra).Duringthe winter,precipi- tation data from the meteorological stationat Sjusjeen, 30 km west of Godlidalen, were used (DNMI Station 1296 Sjusjeen - Stora- sen).The precipitation datafrom Sjusjce nare almos t the same as those from Godlidalen during the summerand fall,and itistherefor e assumed that theyarealso representative for the winter period.

The discharge of the Skvaldra was measu- red with an Ott limnigraph (Fig. 1). Glaamen og Laagen Brugseierforen ing was respons ible for theinstrumentation,and the limnigrammes were analysed by the Norwegian Water Resources and Energy Administration (NVE 2673-0).

For spring A (Fig. 3) the discharge was measured each time awater sample was col- lected,while for spring B(Fig.3)the dischar- ge was monitored closel y only during two shorter summer periods. The discharges of these springs were measured with a tipping bucket or a bucket and a stop watch.

Three groundwater monitoring tubewells of 5/4" diameter and with a filter tip of one me- trelengthwere placed inthe tillbytheGeologi- cal Survey of Norway (Fig. 1) (LGN Station 24/Astdalen15006,15007and15008).Gro und- water levels in these were measured with a manual measuring tape.

Groundwa ter t ypes and ground- water chemistry

Themaingroundwater recharge areasarethe area of frost weathered bedrock at the top of the mountains and the permeable sedi- ments along the valley sides (Fig. 4). Lowe r

NGU-BULL.422,1992 Groundwatercontributiontoamountainstream channel 7

A, Winter - Mar ch 1989

JJEqv/1 5001

Skvaidra

40 0

Spring C

150 0 8

30 0 Spring

8

Spring A

'00 Snow

I

150 0 7 lege nd

H' . HC0 3' Ca"

Mg" N0 3- K' SO;'

Na' Cl

Spri ngs Sk Tub ewens intill

8. Summer - Aug ust 1990

JJ Eqvll 40 0

Spring

C 150 0 6

JJEqv/l

300 1500

Skvalora

500 1000

OL

Tubeweusintill 15007 Spring

200 8

Spring A

10 0·

'Il ^{I I}

Rain

OL

~

~du,d

Springs

Fig. 5.Ion composition of themain water types in Astdaten. Precipitation(Pl,groundwater springsA,Band C,Skvaldra,and groundw ater tubewells 15008,15007and 15006.Bars on leftshow cations andbars on rightshow anions.Winte rdat aarefrom 1989.Summerdata are from 199 0.

down,the peatlands form an effective barrier against groundwater recharge, Three main groundwater types dominate (Fig. 4a,4b):

(i)Groundwater in theinhomogeneous per- meablesediments along theupper part of the valley sides. Data from Haldorsenet al. (1983)

suggest that the hydraulic conductivities of these sediments may be up to one hundred times higher than those of the compact basal till lower down. The groundwater in these upper sediments discharges via a marked spring horizon (Figs. 1,3 & 4) (Kehl er 1985).

8 Sy/viHetoorsen,Jens-O/afEng/und .Per Jerqensen,Lars A. Kirkhusmo&DagHongve ^{NGU -BULL.422,1992}

Basaltill

.,...,,'---

Hummoci< yarea

-

-::::-::::

- - - -

-::::-::::-::::-::::- /.

B.

A.

Hummoci<yarea

5m

Fig.7.(a) Po SItion of spring B in a humm oci<yarea.(b) ooel of spri ngsBand C.For geogr aphic alPOSitionofthe twosprings .seeFig.3.

Fig.6. Mod el of spring A(shallow qroundwater) and Its relationtothemeltwaterchannel. For pos itionof springA seeFig.3.

SUMMER

S \fe>'w a : ~ S

el'la.,,,,,, U

ECIIceree-a ne

Eo

-.

~ PO· ... • •

'"

~

""

~

..

So--0

..

S S

ye:...:.' Ye :"". :. '

ct1ann C" a" ~e'

c"

... '"

The springs are classified as contact springs. since they occur atthe bound ary betweenthe permable upper sediments and the compact basal till.The spring water follows numerous small streams from the spring horizon down to the Skvaldra.

(ii)Shallow groundwaterin thebasaltillalong the lowest part of the valley sides and in the valleybottom.Thiswaterdischarg es alongthe valleybottom.Some disch argemay occur from the till to the peatlands. but the main part probably flows from the till directly to the Skvaldra.

(iii) Deep groundwater in the lower parts of the tillor in the bedrock.with a long residen- ce time.which discharges in themiddleof the valley (see Englund 1986).

From a conceptualstandpoint .flowsystems in catchments can be considered to fallbetwe- en two end-member extremes;those that are dominated by near-surface flow paths and thosethataredominatedbydeepergroundwa- ter flow paths (e.g. Peters & Murdoch 1985.

Parsons et al. 1986. Winter 1986. Baron &

Bricker 1987).

The bedrock in Godlidalen is resistant to chemical weathering and mostof theground- water has a rather short residence time.The areaisdominated by near- surface flow paths. Asaresultthe gro undwater whichdischarges into the Asta (Fig.1)has alow ionic strength (Fig. 5) compared with other types of Nor- wegian groundwater (Englund 1983).

Calcium. magnesium and sodium are the majorcations.withpotassiumasaminorcons- tituent.Bicarbonateandsulphateare thedomi- nant anions.Calcium.magnesium.sodiumand bicar bonate constitute more than 90% of the total ion content. The difference in chemistry for the studied gro undwater types is mainly reflected in the different contents of cations and bicarbo nate.

Springs

The marked spring horizon along the eastern valley side is found at an altitude of930 m.a.

s.1. in the south.rising to about1000min the north (Figs. 1. 3 & 4). More than 80 distinct springs can be mapped over a distance of three kilometr es. with more diffuse seepage faces betwee n them.

Theseismicprofile indicatesthat theperme- able sediments are mainly unsatur ated (Fig. 4b).with seismic velocities betw een 800 and

1000ms-'.The satur ated zone mustberather thin withaflowparalleltothebedrock surface down towards the spring horizon.

The most typical springs are found in the following positions:

(i) Sing le springs downslope of meltwater channe ls which act as con fluence areas for the groundwate r (Spring A. Figs.3 & 6).

(ii) Groups of springs inside topograp hical depressionsdownslope ofgreater accumulati- ons of coarse sediments .The latter comm on -

NGU-BULL. 422.1992 Groundwatercontribution toamountain stream channel 9

Sum ofcations("Eqv/I)

500,---~

Fig.8.Cationcontent(ueqzl)of thespringsduring the years 1989-1991.Samplesfromall threeyearsare plotted in the same diagram to show thegeneralannualvariation.Springs A,Band C areshown bycurves. Smallblack squaresare samp lesfromotherspringsin the catchment.

tions in the springs are found at the end of winter(March-April) (Fig. 8) beforethe meltwa- ter reaches the water table. Variations in con- centration reflect degree of dilution and variati- ons in storage time for the groundwater.

The similarities in groundwater level fluctuati- on and chemical composition indicate that all the springs are fed bythe same type of aqui- fer, a shallow unconfined aquifer in coarse Quaternary sediments. Since the sediments are rather inhomogeneous they probably form many local, limited aquifer units rather than one single continuous aquifer along the valley side.

Three springs,A,B,and C, have been stu- died in more detail.

Spring A (Fig. 3) has a relatively low ion content (Figs. 5 & 8). Its upper, main outlet is active only when the groundwater level is high, i.e. from May to the middle of October (Figs. 6& 9).The rest of the year only a diffu- se seepage face is found some metres below the main spring A outlet. The highest ground- water level is established one to two months after the snow melt. After the startof an inten- sive rainy period during the summer it takes at least two weeks before the discharge from spring A starts to increase (Fig. 10).

Spring B has a medium ion concentration (Figs.5&8).Its discharge is more stable than that of spring A, but spring B also has its highest discharge in the middle of the summer.

Spring C has the highest ion content of all the springs. The lowest values, found in the middle of the summer, are higher than the highest winter values for springs A and B (Figs. 5 & 8).The discharge from spring C is significanteven in the late part of the winter.

C

V

^/

- .'

^{' . '}

.

^"

~ \ : ; . .~+- T

., ~ <, • ,~~ -c.~ .

I ~ ,-- --' __ •...,...

. ' " \ K!' ^{! -- -. 'A-~}

• Ii •

. -

100 200 300

Iy form a distinct hummocky topography (Spring B, Figs. 3 & 7).

(iii) Single spring outlets in topographical positions some metres below springs of types (i) and (ii) (Spring C, Figs.3 & 7b).

Between the marked spring outlets,a seepa- ge face extending along the whole valley side feeds the downslope peatland areas with water (Fig. 3).In dry summer periods no stre- ams are found above the spring horizon. The streams leading from the springs have a signi- ficant discharge even after long dry summer periods and also during the whole winter.

The ion concentration varies among the springs (Figs.5& 8),withthe highest concen- trations for the springs topographically lowest down in the valley side.The highest concentra-

SPRING A

Q) 4 jg2' 3

o

cS 2

J F M A J F M A M J J A SO N D

19 9 0

Fig.9.Discharge ofspringA (lis) in 1989 and 1990.

10 Sylvinetaorsen,Jens-OtetEnglund.PerJerqensen,Lars A. Kirkhusmo& DagHongve NGU· BULL.422.1992

5 I/s

20 Discharge

~cations

::

10

,

3 4

~ Eq vll

2001

100j

'---r--,---.---.---r--.~-

The change in chemistry along the stream flowin gfrom springAto theSkvaldra isinsigni- ficant during low disc harge situations in sum- mer,aswell as during winter.The water samp- led at the spring outletsis therefore representa- tive of the groundwater that drains into the Skvaldra via the small streams leading from the springs.

The water supp ly to the Skvaldra from the springs thus has an annualvariation with one main maximum in discharge during the sum- mer. This is the same annual variation which is observed for other aquifers in the inland moun tain areas of south Norway (Kirkhusmo

& Sensterud 1988).

Au g u st

Fig. 11. Fluctuations in groundwater level of tubswalts' 15006,15007 and 15008.

Fig.10.Variationindischa rgeandcationcontent ofspring A in Augusl 1988 in relation to prec ip itation inJuly and August 1988.

···· ···,5008

10 20

Au gus l

- - - -15007 Ground water levelsintill

- 15006

---

^'----

^-

- - . r - -

--

-

^{, ..}_'~:" r~·'··~~~~··

... ./ ...

^~~

.. I.. • •• • •.•..~ ."' . ''''1

, '., .,: ^-,

.•..~... ":'':':':'

I

I

"

^{, ,ae ,}'"

~

I

^I

.~!\tV\i'",,:,:,'.,J. ^/0";

I ^'~h

;:'i,

-a

"o 50

40

E

E c 30

rn

...

0 N

'0. ₂₀

'0

'"

c::

10

10 20

July

Gro undwater in basal till

About 80% of thecompactbasal till,stretch ing fromthe spring horizon down to the Skvaldra, is covered by peat (Fig. 1). The till forms lo- cal unconf ined aquifers along the valley side where it is not covered by peat. Along the valley bottom the peat forms a confining lay- er. Two groundwate r tubewells are located in an unconfined part of the compact till in the lower part of the valley side (the two tube- wells to the right in Fig. 1). The lower one

(NGU 15007) extends to a depth of 2.4 m

belowtheground surfaceand theupper(NGU 15008)to a depth of 5.1m.Thedistancebetwe- en them is about 100 m.

Falling head tests (method: Hvorslev 1951) gave hydraulic conductivities of 2 x 10-' ms-' (tubewell 15007) and 5 x 10-' ms-' (tubewell 15008), which are typical values for a com- pact basal till (Lind & Lundin 1990), The till thus has a limited ability to transport water dow n to thevalley bottom.Theannualground - water fluctuation in the two tubewe lls is about 1.5 m,with the highest values occurringduring the summer. This is a typical pattern for qrounc watc in the mountainous regions of southeaste rnI,,rw ay(Kirkhu smo&Sensterud 1988).

The contribution ofwaterfrom thetillto the Skvaldra has been calculated in two differe nt ways:

(i)Lowering ofthe groundwater table.During the winter months thereisno rechargeto the till, except for some water which may flow into the till from the perm eable sediments of the upper area(Fig.4b). There is no loss of water from the till by evapotranspiration.The

NGU·BULL. 422,1992 Groundwater contributiontoamountainstream channel 11 lowering of the ground water level (Fig. 11)

roughly corresponds,therefore,to the amount of water which has drained from the tillto the bottom ofthe valleyandlater contr ibutedto the discha rgeof the Skvaldra.From Dec.12^{t h}1989 to Feb. 25^th 1990 (75 days) the ground water level declined by 50 cm in both tube wells 15007 and 15008. Based on other studies of basal tills(e.g.Haldorsen etal.1983) specific yield isestimated to about10%and thelengt h ofthe till slope from the spring horizon down to the Skvaldra is about 1 km. If the whole till area was an unconfined aquifer ,the obser- ved lowering of the water table would corre- spond to a release of water from the till to the Skvaldra of 8 x 10-³m's-' per km'.Howe- ver,much of the till forms a confined aquifer, where the storativity is much less than the specific yield.Ageneral lowering of the piezo- metric surface by half a metre in the confined parts of the till aquifer would yield a much smaller amount of water to the Skvaldra than calculated for unconfined conditions . Based on typical values of storativities for confined aquifers ,the realflux from the whole till area to the Skvaldra is probably not more than a tenth of the calculated value.

(ii) Groundwater flow to Skvaldra . The groundw ater gradient betwee n the two tube- wellsis 1:10whichis equaltothe topograp hi- cal gradient. This is representative for the whole till slope from the tubewell 15007 up to thespring hori zon.The flow ofgroundwa- ter is thought to be mainly one-dimens iona l and directed down the valley slope.The hy- draulic conductivity around tubewell 15007 at 2 m depth was found to be 10-⁷ ms-'. The hydraulic conductivity nearer the tillsurface is obviously much higher due to the occurrence of fractures and root channels, while the valu- es deeper down probably are lower (as for tubewell 15008). If the hydraulic conductivity value from tubewell 15007 is used, and the average thickness of the till is estimated to be 10 m, Darcy's law implies a water trans- port of 10-' m's-' down to the Skva ldra per km widt h of the area.

The whole till slope area in Godlidalen is about 7 km' . If the calculation above isrepre- sentat ive,it indicates that the Skvaldra recei- ves a till water component of between 6 x 10-^l and 7 x 10-' rn's:' during the middle of winter.

Water samp les from the two tubewells in thetillhave ion concentrations which aresirni-

larto many of the springs (Fig.5).Water samp- les taken from tubewell 15008 have a higher ion concentration than those from tubewe ll 15007 because the former is placed deeper down in the till. Samples from the botto m of the peat between the Skvaldra and tubew ell 15007 give values in the same range. It is thusnot possible to dist inguishbetwee nwater from the till and water from the springs by means of chemistry alone;

The till water has been sampled only in the areabetween theSkvaldra and tubewell 15008 and only down to 5 metres depth. Till water from other parts of the area has not been studied. However, since the basal till is very uniform,the obtained values are probably re- presentative for other parts of the tilldown to a depth of 5 metres.

Groundwater with a long residence time

In the middle of the valley bottom,close to the river Skvaldra a groundwater tubewell in tillextends to a depth of 3.5 m below grou nd level (NGU 15006, the tubewe ll to the left in Fig. 1).The conditions are artesian thewhole year, with a press ure surface above ground level (Fig. 11).The ion concentration in water from this tubewell is about ten times higher than that of most of the other stud iedground- water types (Fig.5).Theion concentration has onlya smallvariation throughout the year.The artesian pressure could be caused by the unconfined aquifer in the upper part of the valley sides feeding a confined aquifer low er down.However,the highand rather stable ion concentration indicates that this is a deep groundwater,or a groundwater with a signifi- cantly long residence time, which discharges in the centre of the valley,as has earlier been proposed for the main Asta catchment (Fig.

1)(Englund & Haldorsen 1983, Englund 1986).

The small variations in pressure head and chemical composition indicate a nearly cons- tant annual flux of this groundwater type. It is not clear if the water originates from the deepest part of the till or from the bedrock below it.

Tubewell 15006 is the only locality where the deep groundwater has been sampled. Its representativeness for the catchment as a whole has not been verifi ed. However, the bedrockis very homogeneous,asisthe overly- ing till. The topogr aphical gradient of the val-

12 SylviHaldorsen,Jens-OlafEnglund,Per Jorgensen,LarsA.Kirkhusmo &DagHongve NGU·BULL. 422.1992

ley sides in the studied part of the catchment does not vary very much, and the gradient of the valley bottom is very small. There is, therefore,no reason to suppose thatthe deep groundwater compone nt along the valley sho uld varysignificantly.Streamwate rsamp les taken at several places along the Skvaldra in August 1987 and1988 showed verylittl e varia- tion in water chemistry (Bosch et al. 1988, Anema et al. 1989) indicating a rather uniform distribution of the different water components along the Skvaldra.

The influence of the deep groundwater is clearly seen in the chemistry of the Skvaldra.

In the winter the ion conce ntration in the Skvaldra'sstream wateris normallyhigherthan in spring C (Fig. 5).There are no other sour- ces for such water other than the deep ground- water which discharges in the middle of the valley.

Peatland hydrology

The 40% of the Skva ldra catchment which is covered by peatlands (Figs. 1 & 3) plays an important role in its hydrology.The hydraulic conductivityofthe peat wasmeasured at three localities betweentubewe lls 15006 and 15007 by falling or incr easing head methods(Hvors- lev 1951). For the upper part of the peat the value varies betwee n 5 x 10-' and 3 x 10-' rns" ,whilebelow50 cmdepththe valuedrops to 1- 2 x 10-<ms-I.In the deeper part of the peat the hydraulic conductivity is much lower than in thetill.The peatlands clear lyact,there- fore,as a confining layer overlying the till.The watertransport through the deeper part ofthe peat must be very slow,and therefo re of little import ance for the hydrological budge t of the Skva ldra.

Hydrological budget of the Skvaldra

Flood events

In 1989and 1990 the snowmeltfloodoccurred from April to May (Fig. 2). The conclusions above indicate that the ground wate r compo- nent of the snow melt flood is insignificant, since the discharge from the springs andthe

grou ndwater level inthe tillarelow during the whole snowmelt period. The snow melt flood is therefore completely dominated by direct supp ly of melt water, as surface run-off or shallow channelled interflow, down to the Skvaldra.

Mark edprecipitationevents in moistperiod s give a rapid increase in the discharge of the Skvaldra.Afte r a rainstormthedisch arge dec- reases very rapidly. The corresponding delay in the increase in discharge from spring A, which reacts quickly to rainstorms compared with the other springs,was found to beabout two week s. During this groundwater delay period,the discharge in theSkvaldra hastypi- cally nearly returned to baseflow conditions.

This means, in the same way as argued for thesnow melt flood,that thegroundwatersupp- ly is ofvery little importance for the disch arge of the Skvaldra even during the latt er parts of flood events.During a rainstorm,water di- scharges as surfacerun-off orfollows shallow channe ls in the peat or the soil dow n to the Skvaldra (Fig.4b).The peat is very important since itconstitutes mostof the area along the valley bottom.thisis very app arent after hea- vy rain;the Skvaldr athen becom esbrownish- coloured due to a high humic content in the water. The water soluble humic substances stored in the peat are partl ywashed out from it under such conditions.

It is therefore concluded that groundwater flow does not significantly influence flood di- schargein the Skvaldra during anypart of the year.

Baseflow and groundwater budget

Itis difficult to sayhow much of thebaseflow is realgroundwater and how much is surface water or shallowinterflowwater.In theSkvald- ra catchment, water from peatland s, as well as from the three small lakes, will cert ainly contribute to baseflow durin g dry periods. According to studi es by Yasuhara & Storm (1992)in Trondelaq(mid- orw ay)theflux from peatlands is important for the baseflow of ri- vers, even when the watersheds are rath er dry. Howe ver, from Octoberl November and during the winter months, the surface water influence decreases and the baseflow beco- mes more and more dominated bygroundwa- ter. It is thus thought thatthe lowest dischar- ge valuesrecordedduring winter approach the true groundwater basetlow value.

TheJanuarybaseflowintheSkvaldraduring

NGU-BULL.422, 1992 Groundwater contributiontoamountainstreamchannel 13

the winters of 1989 and 1990 was about 0.25 m's-'. By the latter part of February it had decreased to about 0.05 rn's'", In 1991 the minimum discharge was even lower and Feb- ruary discharge values of 0.03 rn's:' were measured.To make agroundwate r budget for these low baseflow situations, the following assumptions,whichare based on thediscussi- ons above,have been made:

(i) The total baseflow is due to the three groundwater components described earlier in the paper,whilethecontribution from surface water isinsignificant.

(ii)The deep groundwater sampled in tube- well 15006 is representative for the whole catchment.

(iii) The average groundwater component from the springs lies between the observed values for spring A and spring C,

(iv)The shallow tillgroundwater in the enti- re area has an average composition close to that of tubewells 15007 and 15008. The di- scharge from the till is not higher than 5 x 10-' m's-'.

(v)Thereexists.110groundwater witha com- position betweenthat of sprinqC and tube- well15006,i.e. thereis areal differencebetwe- en the deep groundwater component and the shallow groundwater, with no intermediate groundwater types.

Based on these assumptions one can make the following approximation of the groundwa- ter budget for late February 1991.The cation concentrations for spring C, tubewell 15006, tubewell 15008 and the Skvaldra are applied as input data.The relative water contribution from the springs (X) andfrom deep groundwa - fer (Y) are then calculated as follows:

X • cationconcentrationspringC + Y • cationconcentrationtubewell 15006 + Z • cationconcentrationtubewell 15008

= cation concentrationof the Skvaldra

zrs the relative water contribution from the till. The total discharge in Skvaldra in late February is 3.6 x 10-' rn's".Thus,

Z = 5 X 10-' / 3.6 x 10-' = 0.14 (i.e. 14%).

The second equation required to solve this problem is,

X+ Y + Z

=

¹ (total discharge is 100%).

Thecalculationyieldsasprin gwater compo- nent

of

76%,

a

deep groundwater compo nent of about 10% and a till water component of 14% of the total baseflow (Table 1).

Table 1.Calculat ed compo nents(%)ofdiffe rentgroundwa- ter types contibuting to Skvald ra during winter baseflow period s.Data from spring Agivesminimum.and datafro m springC gives maximum.spring water com po nents.

PERCENTAGES OF DIFFERENT GROUND- WATER BASEFLOW COMPONENTS

CALCULATEDCOMPONENTS(%) Based on datafrom Spring Deep Till

water groundwater water

Feb.1991 SpringA 66 20 14

SpringC 76 10 14

March1990 SpringB 82 13 5

March1989 SpringA 70 20 10

SpringB 72 18 10

SpringC 79 11 10

If the lower ionconcentration foundin spring Ais regarded as representative for the winter sprin g flow, the same calculation gives a spring water flux of 66%,a deep groundwater flux of 20%and a tillwater componentof 14%.

A similar calculation for early March 1990, using data from spring B,yields a spring flow component of about 82%,a deep groundwa- ter flux of 13% anda till water component of 5% (Table 1). For early March 1989 (Fig. 5) similar calculations give a spring component of 70,72 and 79% and a deep component of 20, 18 and 11%,using data from springs A, Band C respectively. The till water compo- nent is then calculated to be 10%.

It is concluded,therefore,that thecontributi- on from springs is the most important compo- nent of winter baseflow in the Skvaldra,com- prising considerab lymorethan 50%.The deep groundwater constitutes lessthan 20%ofthe totalbaseflow.The calculationofthefluxfrom the basal till is based on a seriesof assumpti- ons whichare not verifiedbyfield data.Howe- ver,thetill water chemistry is so close to the spring water chemistry that an error in the tillwater flux would mainly affect the calcula- ted spring water component and would have little effect on the calculated deep groundwa- ter flux. The calculationof the maximum deep groundwater flux is thusregarded as a rather good estimate.

It was diff icult to separate the baseflow part of the limnigrammes during the summer periods, because rain events were frequent during most of the observed summers. In addition the groundwater discharge responds very slow ly to rain events,with a delayof at

14 Sy/viHa/dorsen,Jens-otetEng/und,PerJerqensen,LarsA.Kirkhusmo&Dag Hongve GU -BULl.422.199 2

least two weeks. Compared with the winter, acalculationof the diffe rentgro undwate r com- ponents is complicated to carry out. Howe- ver, this study does support the follow ing deductions.

During the summer the totaldeepgroundwa- ter flux andtill water component are probably about the same as in the winter. The spring discharges are at a maximum in t:,e middle ofthe summer and earlyfall. Thetotalground- water flux is therefore expected to be higher in summer than in winter. The ioncontent of theSkvaldra baseflowislower duringsummer than during winter because all the springs have lower ion contents in summerthanwin- ter (Fig.5).

The mean annual groundwater flux to the Skvaldramust be more than0.03rn's",equiva- lent to at least 10% of the annual precipitati- on. Of this. most is recharged above the spring horizon. while precipitation along the lower part of the catchment mainlycontributes to the flood events.

Ackno wledge ments

Financ ialsupport was providedfrom the orweq ian auo- nalCommittee ofHydrology.Glaame n&LaagenBrug seier- fore ning and the NorwegianWater Resources and Energy Administration were responsible for the discharge station. The orweg ianMeteorolog icalInstitutehas been responsib- leforthe precipitatio n data.Facilities have been placed at our dispo sal by Plh lske Skogsameie.B.v.dWeerd.D.J.G.

ota andR.v.d .Berg from the Instituteof SoilScience and Geolog y. Univ,of Wageningen supervised two groups of Dutchstudentsin the field.Valuablecomments and cons- tructivecnncismhave beengivenbyD.Banks.H.Hueslat - ten. G. Storro and an anonymous referee.G. Bloc n. L.

Jakobse n,B.Sonstebyand R.Sonst erud assisted Inthe field.A.Borga nand B.Hopland draftedmo st of the figu- res.Tothese per son s andinstitutionswe renderour since - rethank s.

References

Anema.D..Jansse n.R.. [ssen,B.& Veenis.N.1989:A hydrogeologlcalreconnaissance ofthe Skvaldracatch- ment.Southeaste rn orway.Agricultu ra lUniversity of

orway -urnversuy ofWageninge n.113pp. Baron.J. & Bric ke r.O.P. 1987:Hydr ologic and chemical

flu x In LOCh Vale watershed.ROCk y Mountain an o n ar Park.In:Averett.R.C.& cKnight. D.M. (ed s.).Chemi- cal quality of water and the hydr ologic cycte. t.ewis Publishers.Chelsea. L

Bos cn,R.v.d..Eggln .H..Pelgrum.R.&Vermaas.J.1988:

Ahydrogeologlcalreconn aissanceof theskvstar« catch- ment. Southeast Norway. Agricultural Univers ity of

orway -Univers ity of Wagemngen.84 pp.

Englund. J-O. 1983: Chemistr y and flo w pattern in some groundwate rs of Southeastern orway. or.geol uruiers.380.221-234.

Englund.J-O.1986:Springcnarac erosICSand hydrolog ical rnocersof catchments. ora.Hyorot. 17.1-20.

Englund J-O.&Halc o rsen,S.1983:T e Astadalen catch- ment. sout heaste rn orw ay.geologyand general voro- log y.Rep.Dept.of Geol.Agric.UnlV.of orway18.42pp.

Germa nn.P.F.1986:Rapid drainagerespo nse 0orecipiai- on.Hydrol.Processe s 7.3-13.

Haldorsen. S. 1982: The gene sisof tills from Astadalen.

southeastern orw ay.Nor.Geot.ttassi«.62.17-38. Hatc ors en.S.1983:Miner alogy and geoc hemistry ofbasal

tilland theirrelat ions hipto \Ill-fo rmin g processes. or.

Geol.tiosskr.63.15·25.

Halcor sen. S.. Jenss en. P.D.. Kcnter.J.Chr. & yhr. E.

1983: Some hydraulic properties of sano v-suiv or- wegiannns.Acta Geol.Hisp. 78.191-198.

Hewlet t, J.D . 1982: Prmclp les of forest hydrology. The Univ,of GeorgiaPres s.Athens .Georg ia.U.S.A..183pp.

Hvorstev, .J. 1951: Time lag and soil permea biliy In ground waterobservauons,u.S.ArmyCorps of Engm.

Waterwa ys Exp enment.Stat.Bul/.36.50pp.

Hlllestad. G. 1990: Seisrmske rnaunqer i Astadalen. or.

qeot.unaers.Report 90.036.5pp.

Horton.R.E.1933:The roleofinfiltrationinthehydrologiC cycle.Ann.Geophys.Umon Trans. 14.446- 460.

Kennedy.V.C..Kendall.C..Zellweger.GW..Wye rma n.TA

& Avanz mo, R.J. 1986: DeterminaIon of he compo-

nentsof 5orm flow uSingwater chemistryand en "on- ment al ISOtOPeos. auore River Basin, Cahtornia. J.

Hydr.84.197- 140.

K" by. .J. 1978:Htllslope hydrology.John oIey& Sons.

ew York.386pp.

K" husmo.L.&Sonst erud.R. 988:Over va k runqav grunn- vann. or.qeot.uncers.Repo rt88.046.15pp.

Kohler.J.Chr. 1985:Kvaruergeologiskkart - Goauoeten 7:70000. LVF'sinst.forgeore ssurs - ogforu rensnings- Iorskninq (GEFO).As.Norw ay.

UndoB.B.&t.undm,L.1990:Saturate d hydr aulicconduc ti- vityof Sca ndina vian tills. ora.nyor.2t.107-118.

Parso ns.D.J..Stohlgren.T.J .&Graber.D. . 1986:Long term effectsof aciddepos ition on selected ecosystems of Sequia anonalPark.California.U.S.Dept.Interior.

at.ParkService.Ann.Rep..Dec.1986.ThreeRIvers.

California.

Peters. .E.&Murdocn,P.S.1985:Hydrologic comparison of an acidic- lake basin With neutr al-lake basinin the westcen tr al Aoironda ck Mount ains. ew York.Water.

Air& So" Pol/ut.26.387-402.

Pinder,G.F.&Jones.J.F.1969:Oeterrmna uo noftheground- water compon en tofpeak discharqefromthe Chemistry of totalrunoff.Wat.Res.Res.5.438-445.

Siedlecka.A.. ystu en.JP..Englund.J.O.&Hossack.J.

1987:Ulleha mmer -berggrunnskartM.1:250000.Nor.

qeot.unders.

Sktasn, .G..Stewa r. .K. &Pearce.A.J. 1986:Sorrn runo f generation In nurmo neacwatercatchm en5 2.A case studyofrutlsto peandloworder strea m respon se.

Wat.Res .Res.22.1273-1282.

'Vlnter. T.C. 198 6: Eff e c t of groundw3er recharge and

conIguraIon of he water able beneahsand dunes

and on seepage inlakesIn the Sand Hillsof eoras - a.USA.J.Hydr.86.221-237.

Yasuhara. .&Storro,G.1992:Baseflow procucion rorn paruatly peat-cov er edwater s ecs:Acasestudyin rmdd- le orw ay. or.geol.unaers.422. (thisvolume).

Manuscript received Oc tober 1991: revised typescrip t accepted February 1992.