EPISODIC VARIATION IN STREAMWATER CHEMISTRY AT BIRKENES, SOUTHERNMOST
NORWAY: EVIDENCE FOR THE IMPORTANCE OF WATER FLOW PATHS
T.J. Sullivan, N. Christophersen, R.P. Hooper, H.M. Seip,
I.P. Muniz, P.D. Sullivan, R.D. Vogt
ASS TRACT
We have investigated changes io stream and soil water chemistry and
water ttble variations in tn• Birkan•' catchment, soutnern~o•t Norway,
during autumn rainfall and sprin~ snowmelt episodes . Labile ~nomeric
aluminium lAl . I concentrations in streamwater exhibited variable
response to ~episodic increases in discharge. During events precede~
b~ extended periods of low flow,
episodic discharge coincided with
large increases in {Al.]. Wet conditions were characterised by
high but generally decrealing [Al . ) sometimes with drops in concentration c lose to discharge p~a ks. Streamwater was saturated with
respect to synthetic gibbsite at baseflow, but consistently undersaturated durin~ wet periods . Water table measurements in piezometers
situated in a subcatcnment at 8irkenes suggested generally higher
interception of soil layers as the autumn rain season progressed. In
some piezometers . later storms (when streamwater [Al . ] decreased)
resulted in interception of organic horizons which ~ere not intercepted duriog earlier storms {when t~l . l increased!. These data support the hypothesis that streamw~ter [Al.J is controlled pri1
marily bY changes in water flow path during nigh discharge episodes.
INTROOUCTIOH
~cid deposition is thought to cause a decrease in alkalinity and pH of
surface waters in acid sensitive areas and also an increase in
aluminium mobiliza tion from watershed soils. Resulting elevated H+
and aluminium concentrations in lakes and streams have been linked
with fish kills and toxicity to other aquatic biota lleivestad et al.
1976, Saker and Schofield 1982, Henriksen et al. 1984). Aqueous aluminium and hydrogen ions are particularly elevated during periods of
Piezp~~ttr
Rtsults
Data from four representative piezometers are presenttd in figure 4.
Despite considerable variability in responst, there was a consistent
trend of rise in water table for 6 out of 1 pitzometers with sustaintd
water table betw.tn tvents. Of tht rtmaining 1 pitzo~eters. situated
in upland hillsidt artas, 2 never gave water and 5 dried out complettly between tvtnts (e.g. I 8, Figure 4). In pitzomtttr I 14 , situated
in a deep boggy area adjacent to the stream, the water table intercepted organic deposits throughout the study period and rost steadily
during that ti... but showed no episodic response. This sugge.s ts that
water moving through this are• contributes lesa to episodic discharge
than to baseflow. The water table rise caused interception of the organic horizon at four additional sitts during the laat
event
(piezometer I 3, 7, 13, 15, Figure 41. During the first evant none of
these 4 piezoeeters respondtd by intercepting the organic horizon but
two did so in each of the second and third events.
Hydrological lmplicatipnl
These data allow us to draw some conclusions regarding the influence
of hydrology on episodic discharge ch.-istry. First, s~• areas of
the hillslope probably contribute disproportionately to episodic runoff. Stcondly , later events tended to cause interception of higher,
organic toil layers.
Thia was .ast pronounced for the last event,
which was both preceded by the wetteet conditions and was also the
largest event of those studied . Soil solution (lysi. .ter) studies at
8irkanes have . indicated a generel trend of increeaing Al.
1
concentration with increasing depth during both autumn and spring,
although variebility is high !Sullivan et al. 1986a, 198lal. These
data therefore support the hypothesis of streamweter [Al . l being
1
determined largely by flowpath of soil water through verious soil horizons. The hypothesis is coneiatent with the observed decreasing
trend of [Al J in streamwater during the wet period of the autuan
1
1986 and with the episodic drop in (Ali] at the last event
l'igure 2) .
we
earlier 1uggested (Sullivan et al. 1986a, 198la) thlt aluminium
che~istry in strea..ater wat likely regulated
by reaidence time of
water in contact with soil, changes in flow paths during episodes,
and/or r•disaolution of alu.inium which accumulated in soil pores when
water evaporated during dry periods. Soil water at 8irkenes showed
little temporal variation in concentrations of H+ and aluminiu• species despite sa~ling under widely varying hydrological conditions in
both autu.n and spring (Sullivan et al . t9t7a) ~ It is therefore un-
likelv that &ither residenc~ time or redissolution of aluminium in
soil pores is the major controlling factor.
In a recent rtport. Henriksen et al. {19171 have found rapid release
of aluMinium from stream su~1trate 1nd
mots
with
artificial
ac:_idification of a tributary to the rivet Vikedll in wutern Norway.
They attributed this release to dis s olution of Al-hydroxides which had
precipitated from aluMinium-rich ~roundwater loosing co in the stream
2
resulting in increased pH . This pheno.enon is probably of less i~or­
tance in streems such as Birkenes which are chronically acidic . The
Vikedal tributary generallY shows ~H v1riationt of 5 . 0 - 8 . 0 (Henriksen et al. t98Tl, while Birkenes atream values 1re typically 4.25.2. Further~re , Menrixsen et al . acidified the tributary to as low
as pH z 3 . 6 at 5 m below tha point of acid addition, representing Z41
~eqL- 1 of H• above the atream's nor••l •aximuM. It is therefore difficult to equate the resulting Al pulse with snowmelt/rainfall episodic processes . If strea.water Al dissolution were i•portant at 8irkenes we would ••peet to ••• low flow and rising li.O hyd~ograph Al concentrations in equilibrium wit~ an aluMiniUM hydroxide phase su~h •• emorphous or ~icro-crystalline gibbsite. Figures 2 and 3 indicate lowflow saturati on with respect to synthetic gibbsite, but atre~.water
remains well undersaturated with respect to other aluMin i um hydroxide
mineral pha-ses.
In sum.ary , the piezometer d•ta presented in this paper support the
hypothesis of wtter flow path as the major deter minant of . strea~t•r
(Al l at B i r~enes. Other processes which coul d be important in
1
regulating episodic streamwater Al chemistry appear to be of less
significance at this catchment .
ACKNOWLEO&EMEMTS
The authors gratefully acknowledge support ~rom NATO Research Grant
Ho. R615/0672, the Surface Water Acidification _Progra-.e and from the
Al~inium
Biogeoche.istry Study (contract AP 2165 fro. Electric Power
Research lnst . • Palo Alto . -Ca) . . Tt'e Norwegian Water hsources and
Electricity Board INVEI provided unpub1ished discharge data .
Z11
REFERENCES
Baker, J.C. ~nd
acidic wate~s.
C.L. Schofield. 191!. Alumintum toxicity to fish in
~ater Air Soil Pollut. tl pp.289-J09.
Barnes, R.I. 1975. The deterainat1on of specific forms of alu•iniu. in
natural water. CheM. Geol. 15 pp. t71-19t.
C.W., S.A. Gherini, N.E. Peters, P.S. Murdoch, R.M. Newton and
Goldstein. 1914. Hydrologic analyses of acidic and .llkal:ine
l.aktl. Mater Resour. Res. zo pp. 1115-1812.
Chen,
fi.A.
Cbristophersen, N. and R.F. Wright. '911. Sul,ate flu• and a Model for
sulfate concentratio-n in strealft water at lirlllenes, a 1111111 forested
catc~ent ift soutern.ost Norway.
Water Resour. Res. 17 pp. l77-J89.
Christophersen,
stre•• wat•r
pp. 977-996.
H.H. Seip and I.F. Wright. 1912. A IMdel f'or
cheMistry at lirkenes, Norway. Water tesour. lea. 11
N.,
Driscoll, C.T. 1914. A procedure ~or the fractionation o~ aqueout
alu•iniwn in dilute acidic waters. Inter.J.Environ.Anal.Ch... l i
pp. 261-283.
Driscoll, C.T., J.P. Iaker, J.J. lisogni aRd C.L. Schofie1d. 1910 .
Effects of alu•iniu• speciation on ~ish in dilute acidified waters.
Nature 214 pp. 1&1-1&4.
HenrikSen,
A.,
O.K.
Henriks•n.
A.,
A.
Skogheim
1914. Episodic
ctl~nges
in pH and aluminiuRI-Speciation kill filh in a Norwegian
salMon river. Vatten 40 pp. ZSS- 2&0.
Fjalltl.ia.
and
6.6.
1.0.
Rosseland.
RaddUIII, 1.0. ·ttoueland and O.K ._
Skoghei• 1917.
The release of alu.iniw. from moss during acid
episodes in streaMs. Natur•. In review.
Leivestad, H., G. Hendre~. J.P. Huniz and E. Snekvik. 1976. Effectt of
acid precipi ution on fruhwater organis1111, pp. u~ 1 t 1. In: Br•kke,
F.H. (Ed.). Impact of acid ~recipih.Uon on f'Onst and freshwater
ecosyst. .• .in Norway, SNSf -Project, Fl 1/T&, Norwegian tnst. for
water Res., Oslo.
212
Neal , c., t.J . Smith, J. Walls and c. s. Dunn . 19&6 . Ma j or , minor and
trace element mobility in the acidic upland forested catchment of
the upper River Severn, Hid Wales . J . Geol.Soc.London 143, pp. 635&48.
Rosenqvist, l.lh . 1978. Acid precipitation and other po ssible sources
for acidification of rivers and lakes . Sci.Total Environ. 10 pp.J9.U.
Schecher , W. and C.T. Driscoll . 1987 . Evaluation of uncertainties
asaociated with alu~inium equilibrium calculation, . Water Resour.
Re-!L In pres.s .
Seip, H.M . 1980. Acidification of freshwater - sources and mechanisms.
pp.J58-365 . In: D. Orlbles and A. Tollan (Eds.), Ec ological Impact
of Acid Precipita tion . SHSF- Project. Norwegian Inst. Water R•s .,
Oslo .
Sullivan , T. J .• N. Christophersen. I.P. Muniz, H. H. Seip and P . O.
Sullivan. 1936a . Aqueous aluminium chemistry response to episodic
increases in discharge. Nature 323 pp . 324-327.
Sullivan, T.J .. I.P. Muniz and H. M. Seip . 19B6b. A comparison of freQuently usad methods for the determination of aqueous aluminium.
lnter . J.EnvirO·n.Anal..CheM . 26 pp . 69 - 75 .
Sulliv• n. T .J .• H. H. Seip and R.O . Vogt . 1987a. Episodic changes in
aluminium concentrations and ot her i mportant paramet ers i n stream
and soil water in the Birkenes catchment, southernmost Norway.
Can.J . Fish Aquat . Sci. In review .
Sulliv•n. T. J ., K. Christophersen and r.P. Muniz. t937b . Soil- and
stream water chemistry during snowmelt at the 8irkenes catchment,
southernmos t Norway . Can.J . Fish Aquat .Sci. In review .
high runoff such as autu•n rain and sprin~ snoWMelt IDriscoll et al.
19110, Christophersen et al.
1982, Neal et al. 19116). Earl~t life
stages of fish are often present during these high discharge periods
and appear to be especially sensitive to Al toxicity (Baker and
Schofield 1912l.
several researchers have suggested that hydrological factors such as
flow path and contact tiMe with mineral soil may be iMportant in regulating surface water cheMistr~t (Rosenqvist 19711, Seip 1910, Chen et
a 1. 1!U4l . Thus, the role of hydrology in influencing surface water
cheMistry has been recognized, but the quantitative iaportance of
hydrologic processes in episodic acidification is largely unknown.
We have investigated changes in streaM and soil water concentrations
of H•, Al-species and other important parameters in response to high
discharge episodes at the Birkenes catchment since autuMn 1984
(Sullivan et al. t986a, 1987a, 1987b). The purpose of this paper is to
summarize our findings to date regarding episodic changes in stream
and soil water cheMistry, and to report preliminary results of our
investigations on groundwater table fluctuations during autuMn rainfall events. Based on earlier studies, the working nypothesis was
that observed .declines in inorganic monomeric aluminiuM concentrations
in streamwater during later autumn events occured because proportionately more water flowed through upper, organic horizons which contained less inorganic alu.inium.
SITE DESCRIPTION
2
Sirkenes is a small 10.41 km 1 forested catchment with shallow pod ·
zolic soils on granitic bedrock. It is situated 30 km north of Kristianund ·in southernmost Norway and receives an atmospheric deposition
loading of nearly 1 g so4 2- m-2 yr ·1 (Christophersen and Wright 1981).
Streamwater is chronically acidic with a volume-weighted mean pH in
the main brook of 4.5.
METHODS
Stream and lysimeter samples were collected during high discharge autumn (t98,-86l and spring (1985-86) seasons. Meltwater and snow core
samples were also collected in spring. Analyses of pH . conductivity
and aluminium were accomplished immediately at an on-site field
laboratory. Aluminium fraction concentrations were measured using the
Barnes/Driscoll method (Barnes 1975, Driscoll 1984) at ambient tempe·
rature. Speciation of tne labile (mainly inorganic) monomeric aluminium fraction was done with the 4LCHEHI computer program (Schacher and
274
Driscoll 191tl . Additional analvses included Najor · anion• · •ajor
cations , total organic carbon , total fluoride and silica. Details on
s~pling
anf analytical . .thodologie• are presented elsewhere !Sullivan et al. 1986b, 198la. 198lb).
, ... ---~~~-----.
,'
\
:
.
I
, ... ----'"---~
\
,'
/
~
:
"-"
•
,-'
. . ....,
•~
I
I
I
,'
I
I
/
'
I
I
I
I
\
•'
.''.
I
I
.
I
I
'
/
,-'
I
I
I
I
I
~
•
l
''
'
I
I
I
,.
I
,
I
•
I
,
I
\
.•,........
'\
.
''
.\·· - ... --.......,
..' ,
I
~
I
I
'•
, _____,I''
'
' 1'··:
"'•
..
--
'-~
!lUt1
I
,.J'
,
1.--
1. Hap of the lirkenes catc~nt showi ng •ain br ook aa-.ling
station ( 8lE01 1 and piezometer locations .
Fig~re
A subcatchment in the lirkenes watershed was instrumented with 14 piezometers (Figure 11 to record changes in the groundwater table level
during 1 tor• events in the autu111n of 19 81. lila wished to determine i f
a different floWS~ath, as evidenced by different soil ·horizons being
intercepted by the saturated zone ,. could account for the ~ifferenees
i n streamwater cheMi stry observed between initial and later events .
The piezoMeters were ·construct ed of PVC tubing which was slotted up to
15 em over the botton~ . Conical tips were ce•ented on the lower end,
and the tops were capped with" a reeov able PYC cover ; a smlll hole was
drilled into the tube below the cap to allow equilibration wi th atmospheric prusure . The piezOMeters were hand augered into .place and the
holes were ba ckfilled with extr·a c:ted Material and ta111ped · at the
surface to prevent water front runn i ng down the sides of ·the tube . nie
levels of the different soil horizons were measured before the piezoMeter s were instal led.
Zl5
Four piezo.eters were pl~ced along • transect edjacent to the stream
just ~bove the valley floor Ia 10- 131 . The others were installed at
suitable placet along three approxi. . te tre~sect• up the hillslopes.
,iezo. .ter I 14 wet placed in a 2 ~ deep bOIIY area close to the
str••• lfiture 1} .- Sitea wert selected to llle u nearly planar aa possible.
Tht interior o~ each piez~eter wat equipped with a clean polystyrene
tube containinQ a styrofoa• float ~ltich would rise to the surface and
adhere to the walls ~f the tube if the water level dropped. These
••exi.u. rise indicators· also prov1ded· a back-up to our .. nual .anitorint prograM. Frequency of Measur~nt we• approxi•ttely every two
hours during the risint li.O of the hydrograph aftd
frequently ts
it declined. Precision of water depth Muur. . .nts wu approd•ately
!. 1 Cll.
1•••
RESULTS ANQ DlSCUSSIQH
Cbnical.a tr•Df•
During both autuan rain and spring sn~lt episodes the patterns of
strea• response were ·st•ilar for ttcut ch. .ieal eo..,onents for all the
five field seasons studied. Variations in discharge and concentrations
of K• an6 labile MOnomeric (eainiy inorganie) alu.iniua (Al Jfor
1
the Main brook IIIE01) for eutu.n end sprint of 1111 art presented in
Figures 2 and 3: results for pr~vioua rears are . given elsewhere
ISulli~an
et al. 1981a, 111Ta. 1tllb) . Hydrogen ion, nonlabile
mono. .ric (Mainly organically co.plexedl alu.iniu•. and total organic
carbon increased with episodic dit~harge While basic cations and total
fluoride decreased. Cation and anion su•s decrtased during most of the
sn~lt. Patttrs for A1 . however, were More variable . The first
1
episodes following extended baseflow periods resulted in large pulses
of . Ali during both aut~n and sprinl eventa. As autUMn rain and
spring snowmelt
seasons progreued, (Al1. J.. though re•aining
high,
.
.
followed 1 generally decreasing trend, so.eti••s with drop• in concen~
tration close to discl'larve peaks . The thne Ma)or tributariell to th•
Mtin brook followed si~lilar patter~•· although ab•olute concentra tions varied soaewhat aaong the four sites
'I16
07-0ct
11-0ct
15-0ct
19-0ct
23-0ct
21-0ct
31-0ct
0
-1 L----_.....~-"1
-2
-3
60
50
30
20
10
0
12
8
'
200
150
100
50
OtSCHARGE (I Is)
0+-~~~-+r--T~~~~~~~~~*+~~~~
07-0ct
11-0ct
15-0d
31-0ct
Figure 2. Autumn 198& stream discharge, concentrations of H• and
Al. and saturation indices fSII for synthetic gibbsite tSSGI and
1
microcystallina gibbsite ISMG) at 8IE01. SaMpling paints for H+ and
Al. are indicated on time axis. SI's have only been computed for
1
.
3+
+ -l .
samples w1th complete anal~ses. ISl = log ((Al )[H l /Ksol where Kso
is the gibbsite solubility product at the appropriate temperature).
The largest peak in the hydrograph on 28-29 Oct. was not sampled.
27-~
19-Mar
0
_,
-2
?0
10
so
40
Ali(~M)
DISCHARGE (tis)
28-Ap
Figure l . Sp~ing 1115 str•a• discharg•, H+ and Ali conc•ntrations
and saturation indices . Cfr . Fig. 2 .
Z78
11-0ct
15-0ct
19-0ct
23-0ct
2'1-0ct
31-0ct
___A
100
E .
v90
80
70
60
50
'4l
30
20
10
0
90
E .
uao
Piezometer 14
80
50
~
30
250
20
10
200
Ptezometer 7
0
150
100
100
50
so
DISCHARGE ( 1/s)
11-0ct
IS-OCt
l9-oct
23-0Ct
27-0ct
Figure 4. Str••• discharge and water table height at four seleeted
piezometers during the autumn 198&. organic horizons are indicated as
hatched areas to the right.