NORWEGIAN JOURNAL OF GEOLOGY Postglacial changes from diatom analysis, Gardermoen 235

Postglacial changes in the groundwater table at Gardermoen, interpreted from diatom analysis �--�----��------�------�------�------·

Karin Bøe & Bjørg Stabell

Bøe, K. & Stabell, B. Postglacial changes in the groundwater table at Gardermoen, Norway interpreted from diatom analysis. Norsk Geologisk Tids­ skriftVol. 81, pp.235-244. Trondheim 2001. ISSN 0029-196X.

The postglacial environmental history of two kettle hole !akes, Bonntjern and Svenskestutjern, with very different water chemistry, has been inve­ stigated on the basis of the 6 000 year record of their diatom assemblage. The two !akes are situated only 300 m apart, one is acid and nutrient poor, and the other is neutral to alkaline and nutrient rich. Planktonic diatoms ( Cyclotella spp.) dominated in both !akes up to about 4 000 BP. In Svenskestutjern the Cyclotella flora was followed by a benthic/epiphytic flora up to about 2 500 BP, thereafter very few diatoms were recorded in the lake sediments. In Bonntjern the Cyclotella flora was followed by alternations of planktonic (mainly Cyclotella spp.) and benthic/epiphytic flora. The floralchange at about 4 000 BP is interpreted to reflect a lowering of the groundwater tab le. The floral development in the !akes seems to be closely related to regional climatic changes causing changes in the groundwater table.

KarinBøe and Bjørg Stabell, Department of Geology, Universityof , P. O. Box l 047 Blindern, 0316 Oslo, Norway. Present address Karin Bøe; Norsk Hydro Produksjon, Postboks 117, Stavanger, Norway

Correspondence address: Bjørg Stabell, Department of Geology, University of Oslo, P. O. Box 1047 Blindern, 0316 Oslo, Norway. E-mail: [email protected]

lntroduction Hongve (1977) and Hongve & Løvstad ( 1991) found that such lakes are isolated from the groundwater by the In order to study the lake development in the Garder­ deposition of low permeable amorphous organic mate­ moen area the post-glacial history of two neighbouring rial on the bottom, and the infiltration of groundwater lakes, Bonntjern and Svenskestutjern, with different will therefore occur very slowly. water chemistry has been investigated (Fig. 1). The Gar­ dermoen area is a coarse-grained ice-contact delta for­ med about 9 500 BP (Østmo 1976; Longva & Thoresen 1989). The area is one of the largest groundwater reser­ Site description voirs in Norway with a considerable number of kettle hole lakes. Limnological investigations of 28 of these The investigated lakes, Bonntjern and Svenskestutjern kettle hole lakes have been previously published (e.g. (Fig. 2), belong to separate lake types (Table 1). They are Hongve & Løvstad 1991; Brettum 1994). The water che­ only situated 300 m apart and illustrate the different mistry in these lakes is mainly determined by whether water chemistry of the kettle hole lakes. We studied the the lake is in contact with the groundwater or not. See­ diatoms in the post-glacial sediments to unravel the his­ page lakes, which are in contact with the groundwater, tory of these lakes and identify which processes might have water that is relatively rich in nutrients, the water­ have caused the differences we observe. leve! varying with the natural fluctuations in the The fact that the lakes are situated so closely together groundwater table. For lakes that are not in contact with eliminates factors such as climatic differences. The lakes the aquifer the water-level is generally much more stable, have previously been limnologically investigated (Halv­ and their water quality is poorer in nutrients these lakes orsen 1975; Hongve 1977; Hongve & Løvstad 1991; Bret­ are primarily fed by rainwater (Hongve 1977). tum 1994). Bonntjern, which is in contact with the Same lakes (e.g. Svenskestutjern) are deep with most groundwater is made up of two kettle hoies and has a of the lake basin below the groundwater table, but still maximum depth of 9 m. The basin is fairly steep in the not in contact with the groundwater. These lakes are sea­ western part, while the bay in the south-west is fairly led off from the surrounding aquifer (Kjensmo 1978). shallow. Bonntjern is fully circulating, naturally eutrop- 236 Postglacial changes from diatom analysis, Gardermoen NORWEGIAN jOURNAL OF GEOLOGY

...... ·.· . . . ' ' ·�······ · ......

Marine deposits ed rock 200 Groundwater contour . B ------EI] - line (m a.s.l.) 2m interv Eolian deposits c:::J Lake/bog Glaciolacustrine Fluvial deposits [2] Groundwater divide D deposits o Kettle hole Glaciofluvial � Till deposits deposits - lee front o Moreppen research (9500 '.C-yr BP) station

Fig. l. The Quaternary geology and hydrogeologyof the Gardermoen area (modified after Dagestad 1998) . NORWEGIAN JOURNAL OF GEOLOGY Postglacial changes from diatom analysis, Gardermoen 237 hic, with a fluctuating pH, neutral in the winter {dose to the pH of the groundwater) but up to around pH 8 in the summer, and with a moderate ionic content. Sven­ skestutjern, which consists of one single kettle hole, is small and deep, with a maximum depth of 17 m. It is iso­ lated from the aquifer, oligotrophic, and has a pH around 5 and a very low ionic con tent. The water-level in Bonntjern fluctuates with variati­ ons in the groundwater table. In the late 1970s the groundwater level was low {Erikstad et al. 1994) and the water-level in the lake was thereby lowered. Plants {induding birch) moved doser to the lake. As the water­ leve! started to rise in the 1980's, the root systems were drowned and the birch trees died. These dead birch trees can be observed around Bonntjern today. Small birch trees are presently beginning to grow among the dead trees, indicating a repeated lowering of the water-level.

Fig. 2. Aerial photo of the twoinvestigated !akes. (Photo: Øystein Søbye/Samfoto. In: Hongve, D., Skjeseth, S. & Nordli, K. 1994). Note the marsh in the southern end of Bonn tjern, which at the time the Methods photo was taken was not covered by vegetation. Birch is today gro­ wing on this marsh, see text. The lake sediments were cored from the ice in January and February 1995 with a Russian corer {Tolonen 1967) Table l: location1 morphometric and chemi- with diameter of 7.5 cm and 75 cm in length. During the cal values for the twoinvestigated lakes. winter of 1996 additional sampling was done with a • ••• Hiller corer {diameter 3 cm and length 50 cm), to pene­ Halvorsen 197 5 l •• Hongve 1977 l trate the deeper sediments. Altogether sediment cores Hongve & løvstad l 9911 • • • • Brettum l 994 258 cm in length were retrieved from Svenskestutjern Svenskestutjern Bonnt jern and 485 cm from Bonn tjern. The sediments in both lakes UTM-coordinates* PM:206727 PM:207731 consist of fairly homogeneous gyttja in the uppermost Elevation (m)** 198 196 part {upper 40 cm), while the remaining sediments have Area (ha)'* 2.2 4.6 alternating darker and lighter layers. Coarser sediments Max. depth (m)** 18 9 such as silt and fine sand were only found in the bottom Volume lO'm'* 1.57 1.58 5 cm of the Hiller cores. In Svenskestutjern these mine­ pH**'* 5.28/5.30 8.07/7.48 rogenic sediments were black and oily, and after expo­ Conductivity (mS/m)**** 1.70/1.16 5.62/5.50 sure to air their colour changed to red. XRF-analysis sho­ Trophy**** Oligotrophic Polyeutrophic wed increased Fe and Mn concentrations in these sam­ Hydrological type*** l 3 ples compared to samples above. Samples were prepared for diatom analysis every 5 cm in Bonntjern {32 samples analysed) and every 10 cm in Svenskestutjern {17 samples analysed). Approximately 0.5 cm3 of the sediment was boiled with H202 and there­ requirements) for the various diatom taxa were taken after washed and decanted. Thereafter 10 % HCl was from de Wolf (1982) and van Dam et al. {1994). added, the samples were then washed and centrifuged, Radiocarbon da ting was performed at the Laboratory and mounted in Naphrax. A Nikon labophot microscope for Radiological Dating in Tr ondheim (Table 2). The two with 15x ocular and lOOx objective was used for the ana­ cores were also analysed for pollen (Høeg 1997). The lysis. Approximately 500 diatom valves were identified in dating by pollen-analysis was based on a comparison of each sample. Identification was based on Krammer & other analytical pollen investigations in the area and 14C Lange-Bertalot {1986-1991); Germain {1981); Cleve­ dates of pollen events (Høeg 1997). There are, at times, Euler {1950-55); Hustedt {1930); and Patrick & Reimer large discrepancies between the 14C ages and the pollen­ { 1966). A few Cyclotella-species were stu di ed in SEM and analytically derived ages. The 14C ages from Bonntjern identified on the basis of Håkansson {1988, 1989). Eco­ and also the oldest one from Svenskestutjern appear to logical parameters (such as life form, pH and nutrient be too old. Høeg ( 1997) therefore suggests that the 238 Postglacial changes from dialom analysis, Gardermoen NORWEGIAN jOURNAL OF GEOLOG Y radiocarbon dates from groundwater fed lakes in the dominate, while subzone SIIb ( 110-90 cm) has a high per­ area tend to be too old. We have therefore used pollen­ centage of "Fragilaria" spp.. Pseudostaurosira (Fragilaria) analytically derived dates (uncalibrated) for Bonntjern- brevistriata reaches 33 o/o of the total assemblage at 90 cm.

Table 2: 14C ages obtained for this investigation and ages used based on pollen analysis , :!...- 1997), see discussion in text. . Locality Depth (cm) Referenceno. ••c age Agesused i

Svenskestutjern 87 T-12279A+B 2 475± IlO 2475 Svenskestutjern 170 T-12282A+B 4 415± 105 4415 Svenskestutjern 258 T-12280A+B 8 955± 120 6500 Bonntjern 120 T-12285A+B 1295± 90 1145 Bonn tjern 190 T-12286A 2405± 85 1700 Bonntjern 305 T-12283A+B 3650±125 2850 Bonn tjern 330 T-12287 A 3 620 ±125 3200 Bonntjern 370 T-12284A+B 3 975± 100 3975 Bonnt j ern 485 T-12281A+B 7 155± 95 6200

and those from the oldest sample in Svenskestutjern The diatom assemblage reflectsa change to a benthic (Table 2). and epiphytic community and to more acidophilous taxa from SI to Sil (Bøe 1997). The sediment above 90 cm is dose to barren of dia­ Results toms and therefore could not be analysed. For the sediments in Svenskestutjern the diatom analysis (Bøe 1997) resulted in a division into two zones, SI - Sil (Fig. 3) and four zones, BI-BIV (Fig. 4) for the sediments Bonntjern in Bonntjern. The zonations were mainly based on the In some samples the diatoms were severely fragmented, changes in the occurrence of planktonic and benthic and could not be analysed (hatched in Fig. 4). communities. Zone BI (485-385 cm) Svenskestutjern Cyclotella spp. do minate in this zone. Cyclotella kuetzing­ iana reaches a maximum of 76 o/o and C. comta is also Zone SI (250-150 cm) common. C. cyclopunctadominates in o ne single sample, reaching 60 o/o of the total assemblage at 450 cm. Cyclotella kuetzingiana dominates the assemblage from The diatom assemblage reflects a plankton commu­ 250 cm up to 170 cm, with a maximum of 84 o/o at 250 nity, and with circumneutral and alkaliphilous pH con­ cm. At 160 cm there is only a trace of C. kuetzingiana and ditions (Bøe 1997) C. stelligera dominates, making up 83 %. It should be noted that C. stelligera has been found only in two sam­ Zone BII (385-270 cm) ples in this lake, besides at 160 cm it comprises 4 o/o of the assemblage in the sample from 96 cm. Cyclotella Cyclotella comta and C. kuetzingiana decrease drastically comta, C. cyclopunctaand C. ocellata are also common in in this zone. Instead C. stelligera starts to appear (it this zone. At 200 cm Cymbella microcephala makes up dominates in same samples). Navicula radiosa and Syne­ 26 o/o of the diatom flora. dra acus are common and the number of Cymbella spp. The assemblage in Zone SI reflects a plankton com­ increases. Zone Bil contains a 40 cm interval of severely munity, and with circumneutral/alkaliphilous pH condi­ fragmenteddiatoms. tions (Bøe 1997). The diatom assemblage reflects a fluctuatingplankto­ nic and benthic/epiphytic community, and mainly cir­ Zone SII (150-90 cm) cumneutral species (Bøe 1997).

Species within the genera Navicula, Pinnularia and species Zone BIII (270-100 cm) earlier assigned to the genus Fragilaria (Pseudostaurosira brevistriata, Staurosira construens, and Staurosirella pin­ This zone is characterized by fluctuations in diatom nata) dominate this zone. Cyclotella spp. is scarce. Zone assemblages, especially between C. kuetzingiana and C. Sil is divided into two subzones. In subzone SIIa (150-110 stelligera. Species belonging to Achnanthes spp. and Navi­ cm) Pinnularia gibba, Navicula pupula and N. radiosa cula spp. are also common. NORWEGIAN jOURNAL OF GEOLOG Y Postglacial changes from diatom analysis, Gardermoen 239

Svenskestu1jem

,' l l t!/ IIft/ l , /f lll llt/1 l Il''l' !'l' 11/J II //J I �

Fig. 3. Diatom content in S�en�k�stutjern� s percentages plotted against depth. Note that only species with more than 5 % representatwn ts mcluded m the diagram.

- l ,---- - l> � l D� l l r l / FRA MNTs M ·Bi�ll l l ) l - 1-- 1- - - --+--- ,_- - - -� f1?:-�> '-l'> �·-- l co H l / �;;ol,-;;;; � 1 - -+- �

· Bl -t r -��r ,__ l > � Mr i'rl � m 1m- Mn r i'l'l lrrlh 1J'I m l'rl !mJ '�' imr"l"'1 1r1 r1'l mr-'r ___j -:;,��••� ��� �

Fig. 4. Diatom con tent in Bonntjernas percentages plotted against depth. Note that only species with more than 5 % representation is included in the diagram.

The dia tom assemblage reflectsa fluctuatingplankto­ Discussion nic and benthic/epiphytic community, and with mainly circumneutral pH condition (Bøe 1997). Zone Bill con­ In both lakes the chronology is based on pollen analysis tains two horizons of fragmenteddia to ms. and covers approximately 6 000 years. With the excep­ tlon. of the two youngest 14C dates in Svenskestutjern, the 14C dates (as shown in Fig. 5 and 6) appear to be too old Zone BIV (l 00-0 cm) This zone is characterized by large fluctuations between (Table 2, Høeg 1997). Up to about 4 000 BP (Zones SI and BI) the diatom analysis shows a striking similarity in C. stelligera (C.stelligera occurs in every second sample) the floral development in the two lakes, with a total and a diverse flora consisting of taxa such as "Fragilaria" dominance of plankton diatoms (Figs. 5 and 6). In Sven­ spp. and Navicula spp .. C. kuetzingiana reaches only 11 skestutjern there is a change to benthic/epiphytic and o/o in one sample. There is an increase in "Fragilaria"spp. even same aerophilous taxa, ending with a period from at the bottom of this zone. Synedra acus reaches 52 o/o at the top of this zone. about 2 500 BP to the present with no diatoms in the The diatom assemblage reflects an even more rapidly sediment. In Bonntjern, on the other hand, the flora fluctuating planktonic/benthic community, and with changes to alternations of planktonic and benthic/epip­ increased alkaliphilous pH conditions (Bøe 1997). hytic taxa. The fact that these two lakes, which are so dif­ ferent today, appear so similar between approximately 6 000 and 4 000 BP is intriguing. 240 Postglacial changes from diatom analysis, Gardermoen NORWEGIAN JOURNAL OF GEOLOG Y

Habitat Svenskestu�em

. "' PLANKTON IC BENTHIC EPIPHYTIC AEROPHILOUS UNKNOWN g' � o

o

2000

iL aJ (/) � c. w 4000 �

6000

o 20 40 60 80 100 o 20 40 60 80 100 o 20 40 60 80 100 o 20 40 60 80 100 o 20 40

Fig. 5. Percentage distribution of diatom-habitat preferences in Svenskestutjern against pollenanalytically derived time.

Habitat Bonn�em

PLANKTONIC BENTHIC EPIPHYTIC AEROPHILOUS UNKNOWN

o

1145

1700 2000

iL 2850 aJ 0:: 3200 iS c. � 4000 <

6000

o 20 40 60 80 100 o 20 40 60 80 1 00 o 20 40 60 80 100 o 20 40 60 80 100 o 20 40 60 80 100

Fig. 6. Percentagedistribution of diatom-habitat preferencesin Bonn tjern against pollenanalytically derived time. NORWEGIAN JOURNAL OF GEOLOGY Postglacial changes from diatom analysis, Gardermoen 241

Fig. 7. Distribution of --- Svenskestuljem Bonntjem- • • • • • Cyclotella spp. in o Svenskestutjernand Bonntjern against a pollenanalytically --- derived time scale. :..-;..-

------2000 --

- ­ _....._ --- a. m \ (/) \ \ � \ >- 4000

6000

o 20 0 50 020 40 60 80 100 o 20 40 60 80 100 c q q q Ch-, � . oteq�� ...",..",_ '· "��Qe�

An example of the floral similarity in the early phase fluctuations in dominance of either C. kuetzingiana, C. of the lakes is shown by the distribution of the Cyclotella stelligera or benthic/epiphytic diatoms, while Cyclotella group. The percentage curves for C. cyclopuncta, C. ocel­ species are scarce in Svenskestutjern. lata, C. comta, C. kuetzingiana, and C. stelligera (Fig. 7) To explain that the two lakes, which are so different in the two cores is very similar up to about 4 000 BP. Not today, had a similar development in the diatom flora up only does C. kuetzingiana dominate in both lakes up to to 4 000 BP it is tempting to assume that the lakes were approximately 4 000 BP, but the succession of the diffe­ previously connected. Measurements show that an in­ rent Cyclotella species also follows the same pattern in creased water-level of about 3.5 m would bring the lakes the two co res. There is a distinct peak of C. cyclopuncta at in contact. However, we have not been able to find any about 5 400-5 500 BP, followed by a peak in C. ocellata at ancient shorelines to support this hypothesis. Neither is about 4 900-5 200 and C. comta slightly later. C. cyclo­ there any indication that a brook connected the two puncta, C. ocellata, and C. comta disappear in both lakes lakes. The fact that the two lakes were so similar up to around 4 000 BP. At the same time C. kuetzingiana de­ about 4 000 years ago might, therefore, indicate that creases, disappearing completely above 3 700 BP in Svenskestutjern earlier was in contact with the ground­ Svenskestutjern, and does not increase substantially until water. In that case Svenskestutjern was isolated from the about l 900 BP in Bonntjern. C. stelligera occurs for the groundwater either at about 4 000 BP when a first time in Svenskestutjern at about 4 200 BP when it benthic/epiphytic diatom flora displaced the planktonic attains a peak of 83 o/o. C. stelligera then disappears and Cyclotella flora, or at about 2 500 BP when the diatoms there is a change from planktonic to benthic/epiphytic disappeared. diatoms. In Bonntjern C. stelligera occurs for the first In several papers treating the kettle hole lakes at time at about 3 000 BP (with 33o/o).However, no samples Gardermoen it has been discussed how and when the could be analysed between about 3 000 BP and 3 900 BP deep kettle hole lakes such as Svenskestutjern became due to fragmented diatoms, so C. stelligera might have isolated from the aquifer. It has been suggested that a arrived earlier. In both Svenskestutjern and Bonntjern dense layer or membrane may have caused a division the first appearance of C. stelligera occurs after the dis­ from the groundwater (Kjensmo 1978; Hongve & appearance of C. cyclopuncta, C. ocellata, and C. comta. Løvstad 1991; Erikstad et al. 1994). Hongve (1977) After about 4 000 BP Bonntjern is characterized by remarks that although groundwater holding deposits 2.42 Postglacial changes from diatom analysis, Gardermoen NORWEGIAN JOURNAL OF GEOLOGY surround the basins, same lakes do not have a ground­ pore water, therefore, silica organisms were dissolved water inflow. He suggests that this may be the result of a befare burial. However, the lake does not contain many lang-term infiltration, which has led to dogging of the siliceous organisms today (Brettum 1994), so lack of pares, either by fine particulate suspended matter or by nutrients is probably a more reasonable cause of the lack substances precipitated from the water. Kjensmo ( 1978) of diatoms. did chemical analyses of the lake sediments in the nearby Vilbergtjernet to determine how and when this lake was Bonntjern, which is continuously fed by ground­ separated from the groundwater. He concluded that water, was affected by fluctuations in the groundwater precipitation of calcium and iron compounds sealed off level. After the groundwater lowering at about 4 000 BP the lake water in Vilbergtjern. Kjensmo (1978) dated this this lake underwent repeated changes, reflected in the layer to l O 080 ± 200 BP and concluded that it has existed dia tom flora, as the water tab le increased (dominance of since the lake was formed (i. e. that the layer is primary). Cyclotella kuetzingiana or C. stelligera) and decreased Vilbergtjern, which is a small and deep lake within the (dominance of benthic/epiphytic species). The large same area as Svenskestutjern and Bonntjern, is today fluctuations in C. stelligera during the last l 000 years, nutrient poor and almost barren of diatoms, similar to shown here with peaks approximately every 250 years, Svenskestutjern. However Kjensmo (op.cit.) did report has also been seen in a detailed study of the upper 30 cm that the lower part of the sediments (approximately of sediment (Bakken 1997). 6 000-7 500 BP according to his da ting) is extremely rich Bonntjern also has layers where all diatoms are frag­ in centric diatoms. Kjensmo (1978) suggests that melt­ mented (Figs. 3 and 5), especially the intervals 370-330 water in the deglaciation phase was nutrient poor and cm (3 900-3 200 BP), 200-170 cm (l 800-1 550 BP) and thereby did not provide nutrients for a diatom flora, 150-110 cm (l 400-1 100 BP). These layers seem to be while the high amount of diatoms in early Atlantic time synchronous with homogeneous (not laminated) sedi­ (approximately 6 000-7 500 BP) suggests increased pro­ ment layers. The layers with fragmented diatoms may ductivity at that time. have been caused by disturbance by wind or bioturba­ We propose that changes in climate and thereby tion, but also by an extreme lowstand of water-level. fluctuations in the groundwater table may explain the Gaillard et al. ( 1991) discusses that fragmented diatoms lake development. While the bottom sediment in together with Fragilaria spp. can be explained by erosion Svenskestutjern showed high Fe-values, a similar layer at shore or littoral sediments at times of increasing was not found in Bonntjern. We therefore postulate that water-level. It is therefore possible that the layers with the isolation of Svenskestutjern from the groundwater is fragmented diatoms can be related to water-level related to the high Fe-content in this layer. In this con­ changes. It should be noted that the layers with frag­ text it is important to determine whether this layer is pri­ mented diatoms hardly seem to contain centric diatoms. mary or secondary. As discussed under Methods, the radiocarbon dates If it is primary, we must assume that the layer isola­ from groundwater fed lakes in the area tend to be too ting the lake water from the groundwater does not reach old. From pollen analysis, Høeg (1997) found that the up to the highest water-level possible in the lake (Hongve 14C dates from Svenskestutjern are reliable (except for & Løvstad 1991). When the groundwater table is high, the one from the bottom sediments, which is assumed to groundwater will reach the lake from above the isolating be far too old), while most dates from Bonntjern seem to layer, while at low groundwater levels the lake has no be too old. This is another indication that both lakes contact with the groundwater. were groundwater fed prior to 4 000 BP, and that Sven­ If secondary, the Fe-rich layer occurred either as a skestutjern thereafter ceased to be in contact with the consequence of a lowering of the groundwater level or groundwater. was gradually deposited at the interface between the Nesje & Kvamme (1991) and Harrison & Digerfeldt coarse deglaciation sediments and the fine lake sedi­ (1993) interpret the period around 4 000 BP as represen­ ments. The lake water would be sealed off from the ting a drier climate than at present, for example lakes in groundwater when we find the large change in the dia­ southern Sweden were at their lowest water-level bet­ tom flora (at about 4 000 BP). ween 4 500 and 3 400 years ago. It is highly possible that After the isolation from the groundwater the lake similar climatically induced water-level changes took chemistry in Svenskestutjern underwent a drastic place in the investigated area. change, which resulted in a change to the benthic/­ We interpret changes in diatom composition in the epiphytic flora in Zone Sil. Svenskestutjern gradually two lakes to reflect fluctuations of the groundwater table became poorer in nutrients as the groundwater infiltra­ (Fig. 8). A substantial, climatically induced, lowering of tion diminished and diatoms became rare. Surface sedi­ the groundwater table around 4 000 BP caused drastic ment studies (Stabell, in prep. ) show only a few benthic changes in the water chemistry in Svenskestutjern, resul­ diatoms. Sedimentation rates seem to be fairly constant, ting in the gradual disappearance of diatoms. In Bonn­ excluding a decrease in diatom abundance due to higher tjern the same lowering of the groundwater table resul­ sediment accumulation rates. It is possible to argue that ted in the lake becoming affected by smaller groundwa­ the change resulted in an undersaturation in silica in the ter fluctuations, reflected as fluctuations between plank- OF NORWEGIAN JOURNAL GEOLOGY Postglacial changes from diatom analysis, Gardermoen 243 tonic and benthic diatom assemblages. The Atlantic The similarity is seen in the succession of different Cyclo­ period is characterized as warm and humid. The domi­ tella species. This is interpreted to indicate that both nance of the planktonic Cyclotella species, and especially lakes were in contact with the groundwater at that time. C. kuetzingiana which is an indicator of a high and stable The dominance of planktonic diatoms points to high water-level (Wolin 1996), points to a high water-level in and stable water-levels up to 4 000 BP. both lakes during this period. This is also in agreement At about 4 000 BP, the diatom flora in both lakes with the data on water-level fluctuations in southern changed from a dominance of planktonic species to a Swedish lakes (Harrison & Digerfeldt 1993). The transi­ dominance of benthic/epiphytic species. This change is tion from the Atlantic to the Subboreal period is charac­ interpreted as indicating a lowering of the water tab le as terized by a change to a more arid climate. We would

Svenskestutjem Bonntjem Water leve!

Chronozones Climate o

SUB- COOLJ

ATLANTIC HUMID

2000

Il. CC C/) 0::: <( w SUB- WARMI >- BO REAL DRY 4000

WARMI ATLANTIC 6000 HUMID

Fig. 8. Comparison of the distribution of planktonic dia to ms in Svenskestutjernand Bonn tjern. Water level changes have been indicated based on interpretations fromfluctuations in planktonic diatoms. l =low water level, 2=medium water level, 3=high water level. Climate fo llows Hafsten (1956, 1960). expect a lowering of the groundwater table, as shown by a consequence of the transition to a drier climate. In a change to a more benthic/epiphytic flora in the two Svenskestutjern the lowering of the water-level resulted lakes. Pollen analysis (Høeg 1997) shows a decrease in in isolation from the groundwater. birch and an increase in pine, indicating drier and nutri­ At about 2 500 BP, diatoms disappear from the sedi­ & ent poorer conditions. In Sweden (Harrison Digerfeldt ments in Svenskestutjern, either because of dissolution 1993) 4 500 the water-level was at its minimum between of silica in the lake water or, most likely, because of low P 3 500 B and BP. In the Subatlantic period the climate nutrient -level. changed to colder and more humid conditions. From the In Bonntjern the diatom flora has alternated between diatom composition in Bonntjern we interpret this to be planktonic and benthic/epiphytic species during the last a period with large fluctuations in water-level. 4 000 years, suggesting fluctuations in the water-level. The development of the two lakes seems to be closely Conclusions related to regional climatic changes causing changes in the groundwater table. The diatom flora in the two lakes is strikingly similar between approximately 6 000 and 4 000 BP. 244 Postglociol chonges from diotom onolysis, Gardermoen NORWEGIAN JOURNAL OF GEOLOGY

Acknowledgements. - We extend our thanks to Lars Erikstad, Sylvi Hal­ Harrison, S. P. & Digerfeldt, G. 1993: European !akes as paleohydro­ dorsen, Dag Hongve, and Johannes Kjensmo for interesting discussi­ logical and paleoclimatic indicators. Quaternary Science Reviews ons and for reading through an earlier version of the manuscript. 12,233-248. Hilary Birks is thanked for helpful suggestions. We also express our Hongve, D. 1977: The ionic composition of !akes fed by groundwater thanks to the reviewers Atle Dagestad and Hans Schrader. Helge Irgens and precipitation in the upper Romerike district. Nordic Hydro­ 141-146. Høeg is thanked for all his help during fieldwork. The Research Coun­ logy8, & 1991: cil of Norway supported this study Hongve, D. Løvstad, Ø. Verneverdige innsjøer i Gardermo­ området. Limnoconsult. Unpublished report. 43 p. Hongve, D., Skjeseth, S. & Nordli, K. 1994: Grytehullsjøer på Hauer­ setertrinnet. Miljøvernutvalget i Kommune. Hustedt, F. 1930: Bacillariophyta (Diatomeae). In Pascher,A. (ed.). Die References Susswasser FloraMitteleuropas l O. Jena. 466 p. 1997: Bakken, T. B. 1997: Resent miljøundersøkelse i to nærliggende grytehull­ Høeg, H. I. Pollenanalytiske undersøkelser på øvre Romerike, sjøer på Romerike, belyst ved diatomeanalyse. Unpublished thesis, Ullensaker og , fylke (Gardemoprosjektet). University of Oslo, Norway, 124 p. Varia 46, Universitetets Oldsakssamling. 1988: Brettum, P. 1994: Referanseundersøkelse av grytehullsjøer i Håkansson,H. A study of dia tom species belonging to the Cyclo­ Gardermo-området 1993.- NNA-rapport 0-93150, 116 p. tella bodanicalcomta complex (Bacillariophyceae). Proceedings of !Jlh 1986. 329-354. Bøe, K. 1997: Postglasiale miljøforandringer og utvikling av to nær­ the International diatom Symposium, Bristol, liggende grytehullsjøer på Romerike, belyst ved diatomeanalyse. 1989: Unpublished thesis, University of Oslo, Norway, 116 p. Håkansson, H. Diatom successions during Middle and Late Cleve-Euler, A. 1950-55: Die Diatomeen von Schweden und Finland. Holocene time in Lake Krageholmssjon, Southern Sweden. Nova 1-5. Kungliga Svenska Vetenskapsakademi, Hand/ingar. Serie 4,2:1, Hedwigia48, 1-2,143-166. 1978: 4: l, 4:5, 5:4 and 3:3. Kjensmo, J. Postglacial sediments in Vilbergtjern, a small mero­ 25, 207-216. Dagestad, A. 1998: In situ luftinjeksjon i grunnvannsonen som opprens­ mictic kettle lake. Polskie Arectiwum Hydrobiologie & 1986-1991: ningstiltak i akviferen ved Gardermoen, Sørøst Norge. Dr. ing. the­ Krammer, K. Lange-Bertalot, H. Bacillariophyceae. In sis,Norwegian University of Science and Technology, Trondheim, Ett!, H., Gerloff,J., Heyneng, H. and Mollenhauer, D. (eds.) Suss­ Norway. wasserflora von Mitteleuropa 2 (1-4). Fischer,Stuttgart. & 1989: De Wolf, H. 1982: Method of coding of ecological data from diatoms Longva, O Thoresen, M. K. The age of the Hauerseter delta. 131-134. for computer utilisation. Mededelingen - Rijks Geologische Dienst, Norsk Geologisk Tidsskrift69, 36, (2),95-99. Nesje, A. & Kv amme, M. 1991: Holocene glacier and climate varia­ Erikstad, L., Halvorsen, G., Lenes, G. & Tuttle, K. J. 1994: Endringer i tions in : Evidence for early Holocene glacier vannbalansen Oslo Hovedflyplass-Gardermoen. Betydning for ver­ demise and multiple Neoglacial events. Geology 19,610-612. neverdiene i grytehullsjøene nord og øst for flyplassen. Patrick, R. & Reimer, C. 1966: The diatoms of the United States, exclu­ l. Unpublished NINA Notat 29.09.94 sive of Alaska and Hawaii. Vol. Fragilariaceae, Achnanthaceae, Gaillard, M. J., Dearing, J. A., El-Daoushy, F., Enell, M. & Håkansson, Naviculaceae. Academy of Natura! Science Philadelphia. Mono­ H. 1991: A late Holocene record of land-use history, soil erosion, graph, 13, 1-688. lake trophy and lake-leve! tluctuations at Bjæresjosjon (South Tolonen, K. 1967: Soiden kehityshistorian tutkismusmenetelmista. Il. Sweden). Journal of Paleolimnology 6, 51-81. Turvekairoista. (On methods used in studies on the peatland Germain, H. 1981: Flore des diatomees (diatomophycees) eaux douces et development. Il. On the peat samplers). Suo, 18, 88-92. saumatres du Massif Armoricain et des contrees voisines d'Europe Van Dam, H., Mertens, A. & Sinke!dam, J. 1994: A coded checklist and occidentale. 444 p. Boubee, Paris. ecological indicator values of freshwater diatoms from The Hafsten, U. 1956: Pollen analytic investigations on the Late Quaternary Netherlands. Netherlands Journal of Aqua tie Ecology 28, 117-133. development in the Inner Oslofjord area. University of . Wolin, J. 1996: Late Holocene lake-leve! development signals in Lower Årbok 1956, Naturvitenskapelig rekke 8, 163 p. Herring Lake, Michigan. Journal ofPaleolimnology 1.5, 19-45. Hafsten, U. 1960: Pollen-analytic investigations in South Norway. In Østmo, S. R. 1976: Kort beskrivelse av kvartærgeologisk kart, Garder­ Holtedahl, O. (ed.) Geology of Norway. Norges geologiske under­ moen 1:20 000, C QR 051 052-20 og Hydrologisk kart, Øvre Rome­ søkelse 208, 434-462. rike 1:20 000. Intern rapport. Norges geologiske undersøkelse, Halvorsen, K. 1975: En undersøkelse av to nærliggende, men limnologisk Trondheim. forskjellige grytehullsjøer. Unpublished thesis, University of Oslo, 186 p.