Late Glacial and Holocene deglaciation and sedimentation in Lågendalen, southeastern

PER JØRGENSEN & ROLF SØRE NS EN

Jørgensen, P. & SØrensen, R.: Late Glacial and Holocene deglaciation and sedimentation in Lågendalen, southeastem Norway. Norsk Geologisk Tidsskrift, Vol. 59, pp. 337-343. 1979. ISSN 0029,196X.

At the end of the Ra period, approximately 10,600 years ago, the icefront retreated mostly by calving with intermittent stops through the long and narrow Numedalfjorden. Average recession rates Iie between 90-100 m per year. Detailed mapping of Glacial and Postglacial sediments, combined with radiocarbon and geophysical data, aUowed reconstruction of Late Glacial and Holocene development, dating of glaciofluvial and river valley terraces, and calculation of average sedimentation rate during Preboreal time in a deep ice-eroded basin.

P. Jørgensen, Institu/1 for geologi, Universitet i Oslo, Blindern, Oslo 3, Norway. R. Sørensen, Institul/for geologi, 1432 Ås-NLH, Norway.

which can be followed eastward to ·the Mona Deglaciation ridge at Mysen (Holtedahl1974, Andersen 1975). The )argest of the Younger Dryas moraines, the The age of the large glaciofluvial deposits in the Ra, crosses the lower Lågendalen at Bom­ region will be discussed below. mestad, northeast of Larvik (Fig. 1). A radiocarbon dating indicates an early advance to the Ra position around 11 ,000 B .P. (T-26 1, Marine limits Andersen 1968). How far behind the Ra the glacier subsequently retreated during the AllerØd The upper limit for wave-washing at Larvik is period, when marine clays were deposited, has approximately 155 m a.s.l. Farther up Lågenda­ not yet been established. len the upper marine limit is at Svarstad 173 m The ice readvanced to the Ra around 10,700 a.s.l. (KorbØl1972), 184 m a.s.l. at Goverud, and B.P. Whole shells in a glaciomarine clay lying 175-177 ni a.s.l. at Skollenborg (Marthinussen, below disturbed bedded sediments and till near in Holtedahl 1953). Field work in connection Sandefjord, have been dated to 10,650 ± 150 with the Project (Rosenqvist 1969) has B.P. (T-426, Andersen 1968). verified the Skollenborg area value, while the A glaciomarine clay with paired, whole stated marine limit at Goverud is probably 4- 5 m Macoma shells, overlain by disturbed marine too high. sediments and glaciofluvial gravel near Other marine limits used in this paper are the TØnsberg, was dated to 10,850 ± 11 0 B.P. (Lu- ones developed in the Ski moraine at DrØbak, 71 6, Sørensen 1979). Based on this and other 195 m a.s.l. (SØrensen 1979), and the Mona evidence presented below, it is assumed that the locality at Mysen, 208 m a.s.l. (Holtedahl1974). final retreat from the Ra took place around 10,600 years ago. Several lateral and frontal deposits, mainly of Equidistant shoreline diagram glaciofluvial character, indicate a stepwise re­ treat of the glacier front up the valley from A large number of glaciofluvial and fluvial (river Larvik to Kongsberg (Fig. 1). The glaciofluvial valley) terraces are found in the valley from deposits at Passebekk-Goverud have been tenta­ Larvik to Kongsberg. For estimating the tively correlated with the Ski event, which has approximate time of formation of these terraces been dated to 10,000-10,200 B.P. (SØrensen an equidistant shoreline diagram was con­ 1979). Mapping of the moraine ridges between structed (Fig. 2). The diagram is parallel to line Sande and Goverud has been done by Nygaard A- A on Fig. l. (1 952). The Sande frontal deposits are most In addition to the datings and marine limit likely contemporary with the Ski moraines, determinations already mentioned, the diagram

22 - GeologiskTidsskr. 4/79 338 P. Jørgensen & R. Sørensen NORSK GEOLOGISK TIDSSKRIFT 4 (1979)

A Deglaciation deduced from the shoreline diagram l The first large glaciofluvial deposits in Lågenda­ len north of the Ra are found at Styrvoll and GRAVENFOSS Svarstad (Fig. 2) These are thought to represent short stops in the recession of the glacier. The approximate age of the Svarstad glaciofluvial deltas is 10,300 B.P., and the Styrvoll deposits are slightly older. An approximate recession rate for this Ra-Svarstad phase is calculated as 100 m per year. The Svarstad deposits are tentatively correlated with the Ås event (SØrensen 1979). From Svarstad to Passebekk the average reces­

10 I

m.a.s.l 180 m.a sl 160 160 .to.l0600�100 BP

It,() 140

120 120

100 100

80 80

60 60

40 40

20 - 20 - -- -3000

o o STYRVOLL BOMMES TAD KJERRAFOSSEN LARVIK ÅSRUM STEIN SHOLT SVARSTAD HOLM FOSS t ICE� 10600 10300 � 10100 9800 MARGIN 11000 1 (RA) p 2pkm

GLACIOFLUVIAL TERRACE ---- FLUVIAL TER RACE M.L. MARINE LIMIT (RIVER VALLEY TERRACE ) .to. DATED LOCALITY � BEDROCK TRESHOL D

Fig. 2. Profile along the river Lågen from Larvik to Gravenfoss showing bedrock thresholds and elevation of glaciotluvial and tluvial terraces. Equidistant shoreline diagram for the lasi l 0,600 years is superimposed. The profile is parallel to line A-A on Fig. l.

the valley may therefore have little meaning. relative sea level lowering of at least 35 m (Fig. From Skollenborg to Pikerfoss (Fig. 5) there is 2). Consequently, by 9500 years ago, sea lev el considerable evidence for dead-ice wasting, and was approximately at the level of the rock it is difficult to define an actual glacier front threshold at Kongsberg (Fig. 5). About 300 years retreat. later the river had cut down through the large glaciofluvial terraces between Kongsberg and Skollenborg, and the erosion base was then controlled by the rock threshold at Labrufoss. Holocene sea lev el changes in Between 100-105 m a.s.l. remnants of a few Lågendalen terraces are found outside the delta front at The sea followed the retreating ice-front, and the Skollenborg. These are all older than 9000 B .P. long narrow Numedalsfjorden reached from From 9000 to 8500 B .P. a number of terraces Kvelde to Skollenborg (Fig. l) in late Preboreal­ were lifted from sea level and above, since the early Boreal time. Material carried by meltwater sea dropped approximately 18 m during this first was deposited as glaciofluvial deltas in the inner half of the Boreal chronozone. The highest part of this fjord (Heistadmoen, Saggrenda and number of terraces in Lågendalen were lifted Skollenborg, Figs. 3 and 5). About 9800 years above sea level during the latter half of Boreal ago, when the ice had receded to Kongsberg, the time. The drop in sea level amounted to 14-16 m, sea level was 176 m a.s.l. at the ice terminus. and the sea regressed relatively quickly from With a shoreline gradient of 0. 75 m/km, the Gravenfoss to Hvittingfoss (Fig. 2). corresponding sea level at Larvik would have An important consequence of these sea level been 125 m a.s.l. changes was the fact that all the sediments used The isostatic rebound attained its maximum in building up the terraces above the '8000 years during the next 300 years with a corresponding B.P. line' (Fig. 2) must have been transported to 340 P. Jørgensen & R. Sørensen NORSK GEOLOGISK TIDSSKRIFT 4 (1979)

Ed FLUVIAL SAND � MARINE SILT AND CLAY

GLACIOFLUVIAL � GRAVEL AND SAND CJ THIN COVER WITH TILL

TERRACE IN MARINE SEDIM ENTS

GLACIOFLUVIAL TERRACE

Contour interval 100m

Fig. 3. Sediments deposited in Numedalljorden. Marine limit in the Kongsberg area is 176 m a.s.J. Line A-B-<: is a seismic proftle line.

Lågendalen during Preboreal and Boreal time. races in Lågendalen, Odbergmoen, corresponds This is reasonable since farther inland Hardan­ with the sea level of 7000 B .P. Except for gervidda was ice free at the end of the early Odbergmoen, relatively few terraces were lifted Boreal subchronozone (Sandmoe 1960) and rem­ above sea level during the wet Atlantic nants of the ice sheet were found only in the chronozone. At the end of Subboreal time the valleys north of Kongsberg. Considerable sea bad withdrawn to outside Bommestad, amounts of sediments were still being supplied to where the Ra still functions as a local erosion the sedimentation basins, but at the end of the level. As in Boreal time, during a dryer climate, Boreal chronozone the meltwater discharge was hetter defined terraces were lifted above sea probably insignificant and the generally dry eli­ level also in Subboreal time. mate must have resulted in a marked drop in It appears as if sediment transport was at a annual river sediment discharge. maximum during wet periods, and the deposits The surface of one of the largest fluvial ter- formed during these periods were lifted above NORSK GEOLOGISK TIDSSKRIFT 4 (1979) Deglaciation and sedimentation in Lågendalen 341

o l km GRAVEN FOSS B A ! l l 100

50-'

Vl

sea level during the dry periods, to become bound. Since this is not the case (Fig. 5), it is terraces. The lowest terrace surface south of natural to conclude that the glaciofluvial de­ Bommestad corresponds to the sea levet at the posits in the 'dead-ice zone' were deposited Subboreal/Subatlantic chronozone transition. during a gradual retreat of the ice. Consequently, when the terrace material at Pikerfoss was being deposited, the river had already eroded about 20 Sedimentation in the Kongsberg m into the Kongsgårdmoen-Skollenborg ter­ races. It is important to remember that only 600 area years were available for dead-ice wasting, Between Skollenborg and Pikerfoss several glaciofluvial deposition, and subsequent fluvial glaciofluvial terraces are found (Figs. 3 and 5). erosion down to the erosion base levet of The bulk of the sediments is sand and grav el, and Labrufoss. frequently a very coarse upper cobble layer can Silt and clay were laid down in deeper parts of be observed. Particle size and sedimentary the fjord simultaneously with the glaciofluvial structures reveal that these sediments were de­ deposition. As already pointed out, a large over­ posited in a high velocity stream regime. deepened sedimentation basin exists just south Distinct dead-ice phenomena, such as kettle of Skollenborg (Fig. 3). U sing sei smie methods, hoies, bear witness to the ice-contact nature of the depth to bedrock has been determined the deposits. If all the se terraces had been depo­ (Geoteam A/S 1977, internat report). The profile sited at the same time, the upper accumulation line is shown on Fig. 3 and the results are given surface would increase more than 25 m from in Fig. 4. The maximum depth to bedrock be­ Heistadmoen to Pikerfoss, due to isostatic re- tween A and B is 160 m and the seismic vel-

-----DEAOICE ZONE ------

200 ... KONGSBERG HEISTADMOEN 'I'TT"'!"'" ..- KONGSGÅRDMOEN - -- -"".....,_ , 0'·>�SKOLLENBORG "'1111 --;;wo... � TERRACE ... � [!!!) BOULDUS AND STONES GRAVEl ... [ll]) ... IITTI) SAND � SILT AND CLA'f GRAVEN TI EltPOSED BEDROCK FOSS

Fig. 5. Proflie along the river Lågen from Pikerfoss to Gravenfoss showing elevation of the river bed and of the glaciofluvial terraces. 342 P. Jørgensen & R. Sørensen NORSK GEOLOGISK TIDSSKRIFT 4 (1979)

[Id SAND TERRACE I EDGE

WEATHERED Atter berg WJ CLAV li mils 9 � QUICK � CLAV BORE HOLE NR. 9 BEDROCK SURFACE

Undrained Ill shear strength W(%} t/m Depth 20 40 O 4 8 Sens. m r-L-��-L-+--�-0

R • • • 17 l • • • • 27 • • 50 80 • • • 65 • 5 • • 160 ·= 180

• fl' 540 • • • 590 • • 280 • • 180 • 10 •• 180 • 410 Il 9. 100

c D 4. 3. 5. 6. 80 ["60 60 50

40

Fig. 6. Section through sediments at Hvittingfoss. Geotechnical data are from an intemal report of Norges geotekniske institutt. NORSK GEOLOGISK TIDSSKRIFT 4 (1979) Deglaciation and sedimentation in Lågendalen 343 ocities indicate that this deeply ice-eroded basin lifted above sea leve! and infiltration of precipi­ is filled with fine-grained deposits. The upper 30 tation built up the hydraulic gradients necessary m of these sediments are exposed in an erosion for this flow. channel which probably formed during the first Due to postglacial weathering, podzol proflles half of the Boreal chronozone. The sediments have developed in the sandy deposits con­ are composed of 'varved' layers of alternating temporarily with weathering of the clay just clayey silt and sandy silt. Since erosion in these below the clay/sand boundary. sediments started about 9000 B.P. (Fig. 2), the The Hvittingfoss area demonstrates that clay maximum time available for deposition was 1000 slides have been an important factor in the years (from when ice-front retreated, about forming of Lågendalen in Holocene time, to­ 10,000 B.P.). Consequently, the average deposi­ gether with fluvial processes such as river bank tion rate for the deepest part of this basin was erosion and extensive ravination by small around 17 cm per year. streams.

Acknowledgements. - The support of The Norwegian Re­ search Council for Science and the Humanities (NAVF) is Sedimentation and postglacial gratefully acknowledged. The authors also express their appreciation to Professor I van Th. Rosenqvist for all help processes in the Hvittingfoss area received in connection with the Numedal Project. We wish to thank A/S Geoteam and Norges geotekniske institutt for Fig. 6 shows a section through the sediments permission to use data from their intemal reports. deposited just south of the Hvittingfoss threshold (Fig. 2). The height of the sandy ter­ race, 97 m a.s.l., corresponds with the sea level of 9000 B.P. It seems reasonable to assume that References the valley, from Gravenfoss to the very narrow Andersen, B. G. 1968: Glacial geology of Western Troms, cross-section at Steinsholt (Fig. 2), was ftlled to North Norway. Nor. Geo/. Unders. 256. this level with sediments. A major part of the Andersen, B. G. 1975: Glacial geology of Northem Nordland, glaciomarine sediments was probably deposited North Norway. Nor. Geo/. Unders. Bull. 33, Nr. 320. 1977: 4742.01. during the ice recession, while the sandy top­ Geoteam A/S Rapport Fellesutbyggingen Labro/Graven/Tofstad/Landefoss. Seismiske målinger for sediments were deposited during the frrst 800 Elektrisitetsverk og Vestfold Kraftselskap. years of the regressive period. Holtedahl, O. 1953: Norges Geologi. Nor. Geo/. Unders. 164. Further lowering of sea-level caused erosion Holtedahl, O. 1974: Noen glasifluviale isrand-avsetninger i den into these sediments and the present river eleva­ sydlige del av Glomma-vassdragets (nuværende) drenerings­ område. Nor. Geo/. Unders. Skr. Il, Nr. 306. tion was established about 7500 years ago. KorbØI, B. 1972: En kvartærgeologiskog sedimentpetrografisk Geotechnical studies have shown that the undersøkelse i området omkring Svarstad. Unpubl. cand. clays have high sensitivities (ratio between shear real. thesis, Universitetet i Oslo. 1959: strengths of undisturbed and remoulded sam­ Nydal, R. Trondheim Natura! Radiocarbon Measure­ ments l. Radiocarbon Supplement I, 76-80. ples). The scattering of the observed values for Nygaard, Ø. 1952: Kvartærgeologiske undersøkelser i Botne undisturbed shear strength (Fig. 6) must be attri­ og Hof i Vestfold. Unpubl. cand. real. thesis, Universitetet i buted to the high sensitivities and accompanying Oslo. l. 1955: problems for obtaining so-called undisturbed Rosenqvist, Th. Investigations in the clay-electrolyte­ water system. Norwegian Geotechn. Inst. Publ. 9. samples. High sensitivity is generally attributed Rosenqvist, l. Th. 1969: Numedalsprosjektet - en presenta­ to the removal of salt (Rosenqvist 1955), which sjon. (The Numedal Project - a presentation.) Nor. Geo/. has caused a lowering of liquid limit below the Tidsskr. 49, 330-332. 1960: natura! water content (Fig. 6). Diffusion alone Sandmo, K. Problemer omkring furuskogens innvand­ ring og dens senare tilbakegang. Tidsskr. for Skogbruk 68, cannot explain the salt removal and conse­ 204-207. quently it must be assumed that hydraulic flow is SØrensen, R. 1979: Late Weichselian deglaciation in the Oslo­ responsible for the salt removal. This process fjord area, south Norway. Boreas 8, 241-246. started when the bedrock along the valley was