depositional model for the lower cretaceous ... · modern depositional systems with braid-plains or...

41

IntroductionThe Early Cretaceous succession on Spitsbergen consists of the muddy Rurikfjellet Formation (Hauterivian), the sand-rich Helvetiafjellet Formation (Barremian) and a mixed muddy-sandy, marine Carolinefjellet Formation (Aptian-Albian) (Parker 1967; Nagy 1970) (Figs. 1 and 2). There is a general consensus on the sharp, erosive and unconform able character (in mid-Spitsbergen) of the boundary between the marine Rurikfjellet and the fluvial-paralic Helvetia-fjellet formations (Midtkandal et al. 2008), though the dura-tion of the time hiatus is not documented. It was pointed out by Steel and Worsley (1984) and by Gjelberg and Steel (1995) that this erosive hiatus likely increased to the north-west on Spitsbergen because the eroding Helvetia fjellet riv-ers flowed from uplands in this direction and because the unconformity likely originated from slight southeastward tilting of pre-Helvetiafjellet strata. Conversely the hia-tus likely decreased to the south and southeast where this boundary appears to be less erosive (Edwards 1976; Steel and Worsley 1984). This inferred Barremian tectonic uplift

of northern Spitsbergen was a precursor of the well-doc-umented, Late Cretaceous tilting of the region, when the Aptian-Albian Caroline fjellet Formation was eroded pro-gressively deeper to the north (Nagy 1970). The bound-ary between the Helvetia fjellet and Carolinefjellet forma-tions appears to be an inter fingering one, as there are some alternations of brackish-water and shallow-marine facies in the uppermost parts of the Helvetia fjellet Formation (Steel and Worsley 1984; Gjelberg and Steel 1995). The discussion below concerns the debate on the nature, i.e. the gradient, of the base Helvetiafjellet unconformity, and the regressive vs. transgressive character of the immediately overlying flu-vial-paralic succession.

The regional depositional system of the Helvetiafjellet Formation on Spitsbergen has recently been discussed in several papers by Midtkandal and coworkers (Midt kandal et al. 2007, Midtkandal et al. 2008 & 2007, Midtkandal & Nystuen 2009, Onderdonk and Midtkandal 2010). One important aspect of their study has been to emphasize that the basal Helvetiafjellet depositional system developed on

Gjelberg, J. & Steel, R.: Depositional model for the Lower Cretaceous Helvetiafjellet Formation on Svalbard - diachronous vs. layer-cake models. Norwegian Journal of Geology, Vol 92, pp. 41-54. Trondheim 2012, ISSN 029-196X.

The Barremian Helvetiafjellet Formation on Svalbard has been assigned two different stratigraphic/architectural models: (1) one in which an allu-vial-paralic-marine succession overlies and transgressively onlaps an uplift-created unconformity, and (2) a layer-cake model with regressive, very low-gradient braided-stream deposits on the basal unconformity, passing upwards into coastal-plain, shoreline and shelf deposits without diachron-eity of the constituent facies belts. The transgressive model has been criticized because its basal fluvial deposits necessarily (gradient considera-tions) become slightly younger sourcewards (northwest and northwards) and because the model appears to transgress landwards too steeply. The layer-cake model also has some flaws, as the synchronous basal braided-stream unit has a lateral extension landwards of some 150 km. The latter is problematic because a gradient of at least 1m/km would be expected in such a fluvial system, as seen from modern data, and the updip extent of such braided sheet sands is usually limited to 20-50 km. The layer-cake model also ignores a published, more extensive database for the Helvetia-fjellet Formation, showing that there is, despite the overall vertical change from fluvial to shoreline deposits, a significant facies variability from place to place, with fluvial channels sometimes occurring high in the succession and coastal deposits near the base. The transgressive model meets this problem by having a diachronous, upward-deepening succession, as would be expected during transgression, but with significant autogenic environmental variability in time and space. The fluvial sandstone tongues in the transgressive model have less than 40 km of downdip extent and their overall diachronism predicts that timelines pass from fluvial through shoreline to shelf lithosomes. The layercake model therefore has some difficulty to explain the geometry, lateral facies relationships, gradient, grain-size variation and geomorphology of the system.

John Gjelberg, North Energy ASA, Nesttunveien 98, 5020 Bergen Ron Steel, Jackson School of Geosciences, University of Texas (email: [email protected])✝ Deceased

John Gjelberg✝ & Ron Steel

NORWEGIAN JOURNAL OF GEOLOGY Depositional model for the Lower Cretaceous Helvetiafjellet Formation on Svalbard

Depositional model for the Lower Cretaceous Helvetiafjellet Formation on Svalbard – diachronous vs. layer-cake models

42 J. Gjelberg & R. Steel NORWEGIAN JOURNAL OF GEOLOGY

is therefore a paradox that it is this very system that should be characterized as having an extremely low-angle, low-energy landscape. On the contrary, we argue that grain size and facies suggest that this was one of the highest energy depositional systems in the post-Carboniferous succes-sion on Spitsbergen and that the braided-stream landscape passed to shoreline and open shelf environments, usually within 30-50 km. This alternative, diachronous and trans-gressive depositional model evolved from a series of earlier works (Steel et al. 1978, Steel & Worsley 1984, Nemec et al. 1988, Nemec 1992) and was elaborated upon by Gjelberg & Steel (1995). It may look controversial and unrealistically

a very low-gradient shelf or platform, in the northwestern Barents Shelf epicontinental seaway. A second key inter-pretation was that the basal Helvetiafjellet Formation stra-tigraphy was everywhere fluvial and regressive, and sub-sequently shows rapid synchronous changes of the depo-sitional system over long (>150 km) downdip distances, as the relative sea level rose across the low-gradient shelf. We do not disagree that the Helvetiafjellet landscape dur-ing deposition of the Helvetiafjellet Formation had a rela-tively low gradient. However, the grain size of the basal part of the formation is one of the coarsest found in the entire Carboniferous through Cenozoic succession in Svalbard . It

Fluvial facies associations

Destructive sand sheets

Open shelf facies associations

Tidal shannel facies associations

Interdistr. buy facies associations

Fluvial dominatedMouth bar

Wave dominated

Lateral development of the Helvetiafjellet Formation «Layercake» model

70

60

50

40

30

20

10

0

M

20

10

010 km

BohemansflyaHanaskogsdalen

Helvetiafjellet

Skolten

Drønnbreen

Breikampen

Glitrefjellet

StorlengjaLundkvistfjellet

Richterfjellet

Forkastnings- fjellet

BillefjordenFault Zone

??

?

Carolinefjellet Fm

Helvetiafjellet Fm

Rurikfjellet Fm

150 km

200 m

Glitrefjellet Mbr

Festningen Mbr

B

BB’

B’

x x’

A

A

A’

A’

A A’

Rurikfjellet Fm

Figure 1. Above: cor-relation along depo-sitional dip (A-A’) in a “layer-cake” model as initially presented by Gjelberg & Steel (1995). The lateral facies relationships in this correlation are problematic, as dis-cussed in text. Light blue colour represents shallow-marine depo-sits and bold black, sub-horizontal lines are coals. Below: deve-lopment of the Helve-tiafjellet Formation in a ”layercake” model (B-B’) as proposed by Midtkandal and Nys-tuen 2009. The gradi-ent of the braidplain in this system is very low

43NORWEGIAN JOURNAL OF GEOLOGY Depositional model for the Lower Cretaceous Helvetiafjellet Formation on Svalbard

of the general depositional environment of the forma-tion, the main purpose of his study was a stratigraphic overview without any sequence stratigraphic or dynamic stratigraphic approach. Midtkandal & Nystuen (2009) have claimed that the for-mation has an extensive, aggrading layer-cake architecture, without interfingering of fluvial and marine facies. How-ever, convincing arguments were lacking, and they appar-ently did not consider any of the problems related to this model, though these problems (as discussed below, and seen in Fig. 1) had been specifically raised by Gjelberg & Steel (1995). For this reason we find it necessary to further explain the route towards a diachronous model.

Why a diachronous model?The initial layer-cake model of Gjelberg and Steel (1995) was worked out as a serious attempt to correlate the Helvetia fjellet Formation, with its multiple sedimentary facies, between a large number of locations across Spitsber-gen, where the top of the Festningen Member was chosen as a datum (Fig. 1). This correlation panel is oriented more or less parallel to the depositional dip direction , as verified by numerous paleocurrent measurements. However, the problems with this correlation were:

steep at first glance (Fig. 2), but if the scale of the figure is more closely considered, this is not the case. The trans-gressive model resulted also from a series of regional stud-ies, driven partly by the Statoil expeditions in late 1970s and early 1980s, where close to 50 vertical profiles were mea-sured throughout Spitsbergen, and calibrated to hundreds of paleocurrent measurements.

Gjelberg & Steel (1995), in their first attempts to synthe-size the Helvetiafjellet depositional system, did consider and illustrate a layer-cake model similar to that recently proposed by Midtkandal and others (2008), and this is shown in Figure 1 . However, there arose many prob-lems in the attempt to integrate the database of measure-ments across Spitsbergen, as discussed below, and the layer-cake model (Fig. 1, lower) was eventually deemed unlikely. The diachronous model (Fig. 2) arose from the need to have a dynamic and reliable understanding for the abrupt occurrence of this coarse-grained sediment wedge, where Walter´s Law (Walter 1894) was honored, as well as the generally upward-deepening character of the succession.

The name Helvetiafjellet Formation with its subdivi-sion into the Festningen Member and the overlying Gli-trefjellet Member was introduced by Parker (1967) who also recognized the disconformity at the base of the for-mation. Even though Parker had a good understanding

Lateral development of theHelvetiafjellet FormationDiachronous model

Bohemansflya

Forkastningsfjellet

Helvetiafjellet

Hanaskogsdalen

Skoltenn

Drønnbreen

Fleksurfjellet

Breikampen

Glitrefjellet

Storlengja

Lundkvistfjellet

Richterfjellet

Bolt-odden

2

3

4

5

6

7

8

9

10

11

12

13

1

100 m

50 km

Carolinefjellet Fm

Rurikfjellet Fm

BillefjordeF.Z.

Helvetiafjellet FmOutcrops andPaleocurrents

Open marine shelf

Prodelta, distal delta front

Paralic, delta plain, bay, lagoon

Delta front, shallow marine sand complexes

Fluvial sandstone complexes,braided and meandering channels

Incised valley

Figure 2. Correlation along the same transect as in Figure 1, A-A’, but with a diachronous trans-gressive onlap of the basal unconformity. The model shows down-current facies trans itions from braid ed stream, through meandering channel lower delta plain with bays and, in the most distal part, shallow shelf deposits. The model has a basin ward gradient consistent with modern braided fluvial-coastal land scapes. Sub-horizontal, bold black lines are coals.


the lower delta-plain reaches of large river systems such as Niger, Ganges/Brahmaputra or Yukon (data calculated from Google Earth) . Such a low gradient is not compara-ble to any modern braided stream (braidplain) depositional system and is difficult to explain by known hydrodynamic processes. We believe that the “layer-cake” model is severely flawed.

Modern depositional systems with braid-plains or braid deltas.

All published research on the Helvetiafjellet Formation agrees that the basal sandstone unit (the Festningen Mem-ber) is mainly fluvial, consisting dominantly of braided stream deposits that coalesced into extensive braidplains (Nemec 1992, Gjelberg & Steel 1995, Midtkandal & Nyst-uen 2007, 2009). Modern braidplains all have relatively high gradients. Examples of such modern braidplains from pro-glacial gravelly outwash braidplains/fans in North Amer-ica show the following size and gradients: The Scott fan (Boothroyd & Ashley 1975), 24 km long: 1.5 m/km in dis-tal part to 17.6 m/km in the most proximal part. The Yana fan (Boothroyd & Ashley 1975), 8 km long: 3 m/km in dis-tal part to 7.6 m/km in the proximal part. The Malaspina outwash braidplain (Gustavson 1974): 2.4 km long, 3.3 m/km in distal part to 9.5 m/km in its proximal part; up to 50 m/km only in the most proximal, uppermost 400 m. The Peyto outwash braidplain (McDonald & Banerjee 1971): 1.1 km long, gradient 16.1 m/km in distal part to 21.8 m/km in proximal part. The Bow outwash braidplain (McDonald & Banerjee 1971): 0.8 km long, gradient 17.8 m/km in its dis-tal part to 20.8 m/km in its proximal part.

Common for all these braidplains is that they are coarse-grained (sandy and gravelly) and relatively small. Usually the gradient decreases with increasing size of the braid-plain, and decreasing grain size. The braidplain deposits of the Helvetiafjellet Formation were dominantly sandy but with gravelly units at the base, and they were not connected to glacial processes. Braidplain deposits like those of the lower part of the Helvetiafjellet Formation are mainly well channelized, sand-dominated systems with relatively deep channels. It is reasonable to assume that the gradient of this system was much lower than the ones listed above, but not as low as 0.15 m/km over a distance of 150 km as is the case for the model by Midtkandal & Nystuen (2009).

Most of the modern examples of river systems on a rotated uplifted slope (as discussed above) are located within incised valleys and occur as meandering, braided and anas-tomosing channels, depending on gradient. The average gradient of the rivers within incised valleys may be very low over large distances (0.03 m/km – 1m/km), but are still able to by-pass large amounts of relatively coarse-grained sed-iments, since they have limited cross sectional area How-ever, such river systems cannot be compared with the dep-ositional system of the Helvetiafjellet Formation. The huge

1) Marine-influenced facies close to the basal unconfor-mity in the more proximal region (e.g., at type location Helvetiafjellet, Fig. 1) would appear to correlate with braided stream deposits in distal locations on the Helve-tiafjellet landscape.

2) There is an irregular but overall grain size increase in the Festningen Member from proximal to distal localities. If the Festningen Member was deposited in a continuous braidplain that covered most of Spitsbergen as suggested by Midtkandal and Nystuen (2009), it should be possi-ble to see some grain size reduction from the most proxi-mal to the most distal localities across the 150 km wide braidplain.

3) The layer-cake correlation forces deltaic mouth bars of the Helvetiafjellet Formation at various locations to be interstratified with or completely wrapped within braided stream deposits, also in a down-current direc-tion (Fig. 1).

4) We suggest that the “layercake” model would imply an average gradient of the braidplain (over 150 km) close to 0.15 m/km (our calculation) which is virtually flat. This contrasts strongly with modern braidplains which com-monly show gradients of up to 20 m/km (but may be as low as 1.5 m/km at their distal segments).

The problems above have not been addressed by Midtkan-dal and coauthors, who claimed significant time changes (rather than lateral changes) of facies across much of south-ern Spitsbergen, during vertical aggradation of the facies belts (e.g., fluvial, delta plain, etc). They also claimed that there is a regional marine transgression surface on top of the Festningen Member (at about level 44m in Fig. 1, upper). Such a unique transgressive surface at this strati-graphic level has not been observed by us within our large database of locations, but we have seen multiple transgres-sive surfaces at individual locations. The most character-istic feature within the Helvetiafjellet Formation, on the other hand, is gradual and interfingering facies-transi-tions. In order to deal with the four points of concern listed above, we looked at alternative ways to correlate, honoring the overall vertical data trend of fluvial in the lower part to marine-influenced delta plain and to marine shorelines in the upper part. We thus proposed the diachronous model (Fig. 2), with overall transgression punctuated by high-fre-quency regressions, as the optimal stratigraphic solution.

Braidplains or braided delta-plains are well known to develop where there is a decrease in drainage gradient, often located where a steeper drainage gradient with sedi-ment by-pass reaches a lower gradient area. Modern exam-ples of deltaic braid plains, such as the sandurs in south-eastern Iceland, show gradients as low as 3m/km in the lowermost 5 km of the plain (data calculated from Google Earth). Common for all such low-gradient deltaic braid-plains is that they have a limited down-current extension, usually not much more than 10 km. The braidplain intro-duced in the depositional model of Midtkandal & Nyst-uen (2009) implies a gradient (at the basal unconformity) of some 10-20 cm/km, which is less than the gradient on


braided rivers) with some estuarine components, develop only at the very distal reaches, close to the delta. Otherwise the system consists of incised valleys with meandering and anastomosing rivers, mainly with sediment by-pass.

The Ebro River in Spain is mainly an incised valley with rela tively straight or meandering channels that by-pass sediments. Only the lowermost 25-30 km of the river is a depositional system proper, represented by the Ebro Delta. There is no braided stream zone between the by-pass region and the Ebro Delta, probably because the river gradi ent is mature and close to equilibrium at the transition to the delta itself. The gradient of this river system is very close to the gradient that has been proposed for the Helvetia fjellet system by Gjelberg & Steel (1995). The main difference is the lack of braided stream (braidplain) deposits because of a relatively mature river profile with deep incision that terminates towards the present delta plain (Fig.5).

Other drainage systems that could be compared with the Helvetiafjellet Formation are the drainage systems in North America entering into the Gulf of Mexico. The Nueces River of S. Texas is an incised meandering river of very low gradient , and no braided streams have developed as part of the river system. The overall gradient over a distance of some 200 km is approximately 0.4 m/km which is still

river systems of the world are old and well established, and are associated with incised valleys close to an equili-bri um profile, whereas the Helvetiafjellet system is a broad, immature sheet-like system probably developed on numer-ous small ravines or incised valleys. No one has suggested that the Helvetiafjellet system was deposited within a large incised valley.

The drainage profile of the Helvetiafjellet depositional model, with the gradient presented by Gjelberg and Steel (1995), can be compared with some modern depositional systems that show some similarities with respect to facies and depositional environment, such as the Tagliamento River (Italy), the Palana River west coast of Kamtsjatka, the Ebro River (Spain) and the Sandur systems of Iceland (Figs. 3 and 4). These examples represent some low-gradient modern river systems that terminate in delta-plain depos-its, but they are still of much higher gradient than that pre-sented in the diachronous model of Gjelberg & Steel (1995).

The Palana river system on Kamtsjakta has a gradient which is comparable to the one proposed for the Helvetia-fjellet system by Gjelberg & Steel (1995), and it may repre-sent a partial analogue. Still, a coalescing, widely distri-buted braidplain is not a major part of the Palana river system. Instead, anastomosing rivers (and transition to

0

50

100

150

200

250

300

020406080100120140160180200

Ele

vatio

n a

bove

sea-leve

l (m

)Palana R

iver Kam

tsjatka

The diachronous modelHelvetiafjellet Formation(Gjelberg and Steel 1995)

Sandur Island

Taglia

mento

Rive

r

North

ern

Italy

Distance from the shore (km)

Nueces River Texas

Brazos River Texas

The «layercake»model of Midtkandal and Nystuen (2009)

10m/km

1m/km

2m/km

0,5m/km

0,1m/km

Gradients m/km

The lower reches of the Mississippi River

Ebro River Spain

Figure 3. Profiles of some modern river systems compared to the diachronous model of Gjelberg and Steel (1995), and the “layercake” model of Midtkandal and Nystuen (2009). It is difficult to find a relevant, low-gradient analogue to the model of Midtkandal and co-workers, where braided streams are an important part of the system. The gradients of the Ebro River, Spain or the Palana River, Kamtsjatka Peninsula correspond fairly well with the diachronous model. The Texas rivers draining into the Gulf of Mexico are all meandering rivers in incised valleys, with very low gradients (between 0,05 and 0,4 m/km) but still much higher gradients than the “layercake” model. The lowermost 200 km of the Mississippi River shows gradi-ents as low as 0,05 m/km which is significantly less than the “layercake” model of Midtkandal and Nystuen (2009).


the “layer-cake” model of Midtkandal and coworkers. How-ever, it is very difficult to apply the muddy Mississippi River as an analogue for the Helvetiafjellet depositional system.

From a facies development point of view a better analogue would be the Tagliamento River in northern Italy. This river emerges from its catchment area, and is dominated

much higher than the gradients implicit in the “layercake” model of Midtkandal and coworkers (Fig. 3). The Brazos River is gentler, and has an average gradient over its lower-most 200 km as low as 0. 22 m/km, also much steeper than implied in the “layer-cake” model. The lowermost 200 km of the Mississippi River shows a gradient that is not more than 0.05 m/km (Fig. 3), which is significantly less than in

100 km

Island sandur

TagliamentoRiver, NorthernItaly

Palana RiverKamtsjatka

Reindalenbraidplain

Typical extent ofmodern braidplain(orange)

Ebro

Helvetiafjellet FmOutcrops andPaleocurrents

Figure 4. Some modern analogues that may be helpful to explain the depositional setting of the Helvetiafjellet Formation. None of the analogues are good because braidplain facies are lacking (except for the sandur on Iceland, but this is not a good analogue because of the poorly channelized braid-plain).


The two models thus agree on the presence of a widespread unconformity, but differ in the interpretation of the depos-its lying immediately above. Gjelberg & Steel (1995) believe the unconformity was backfilled during long-term punctu-ated transgression, with the regressive-transgressive turn-around not far (a few km) to the southeast of Kvalvågen, whereas Midtkandal et al. (2009) believe the basal fluvial deposits are regressive, the turnaround is higher in the suc-cession and far out into the Barents Sea area.

Onlap geometries observed directly from field and seismic data.

One argument used against the diachronous model by Midtkandal and co-workers is that short-distance, lateral interfingering of facies belts (e.g., fluvial, bay, mouth bar) is not observed in the field. Instead, it was proposed that laterally extensive aggradational architectural elements of paralic facies associations with limited interfingering domi-nate the upper part of the Helvetiafjellet Formation (Midt-kandal et al. 2007, Midtkandal & Nystuen 2009). This is incorrect, as significant fluvial units occur at high levels in the upper parts of the formation, as a glance at the type sec-tion of the formation shows (Location 3, Fig. 2; Gjelberg & Steel, 1995), and from this a likely interfingering of the flu-vial architectures can be reasonably inferred. The model presented by Gjelberg & Steel (1995) would require at least 5 km of continuous exposure in order to be able to see such interfingering directly. Such exposures are not pres-ent on Spitsbergen. If we look at a segment of the diachro-nous model of Gjelberg & Steel (1995) in the same verti-cal and lateral scales (Fig. 6) it becomes clear how difficult it would be to see interfingering lateral facies-belt relation-ships directly in the field, and how difficult it would be to distinguish them from a purely aggradational setting with limited interfingering. In addition, it has to be mentioned that the segment shown in Figure 6 has been chosen delib-erately in an area of most pronounced facies transitions.

by sediment by-pass into a braided river belt that further develops into a meandering rivers system before it ends onto the delta plain with distributary channels, bays and mouth bars. In a hypothetical transgressive setting, this lat-eral systems tract could represent a vertical transition from braided stream, through meandering stream to lower delta plain environments. However, the model for the Helve-tiafjellet Fm presented by Gjelberg and Steel (1995) has a much lower gradient than found in the Tagliamento River. The Ebro River may be a useful analogue with respect to gradient, but it lacks a braidplain between the by-pass zone and the delta plain, suggesting that the gradient change is not big enough to accommodate braided river deposits (Fig. 5).

The Basal UnconformityThe diachronous transgressive model has been rejected by Midtkandal and coworkers partly because of the observa-tion of an erosion surface with associated local conglom-erate at the base of the Helvetiafjellet Formation both in several northwestern areas and also in the Kvalvågen area on the east coast, suggesting an unconformity of regional extent. They additionally concluded that the overlying braided stream deposits were regressive and that the turn-around to transgression was located far out into the Bar-ents Sea to the south and southeast. The observation of the basal erosion surface in the Kvalvågen area is not new, and was also made by Nemec et al. (1987) and Gjelberg & Steel (1995). In fact the presence of an unconformity across much of the region is not controversial, and is included in the diachronous model of Figure 2. The latter authors inter-preted the formation of the unconformity in terms of uplift of the northwestern corner of the Barents shelf, as discussed above, but without much preservation of sediment (except occasional conglomeratic pockets) during the concurrent forced regression and bypassing of shorelines down this southeastward sloping surface.

Braided streams

Meandering streams

Distributary channels

Sediment by-pass

Ebro River

A

Figure 5. Imma-ture river profile with deposition of braided, meande-ring and straight distributary chan-nels respectively in the accommo-dation zone along the profile. The Ebro rive profile is enclosed for comparison. The width of Ebro delta plain is some 20 km.


and Richardbreen), thick units of braided stream depos-its are located immediately above the unconformity (Figs. 1 and 2). This is difficult to explain unless a diachronous model is applied.

Unfortunately, we do not have access to offshore seismic data with good enough resolution to see eventual onlap or erosional truncation along the unconformity. However, it is important to note that the model presented by Gjel-berg & Steel (1995) represents very low angle relationships that would display only modest onlap geometries on com-mercial seismic data. On a regional seismic line where an

The only outcrops that are continuous enough to reveal a diachronous onlapping relationship are the Annaber-get-Ullaberget localities at the northwestern side of Van Keulen fjorden. At this locality Midtkandal and coworkers have demonstrated very clearly the complexity of the basal unconformity and the onlapping character of the overlying sediments. It is puzzling why this was not used as an argu-ment for a strongly diachronous onlapping setting, where the marginal marine Louiseberget Bed represents a termi-nating onlapping segment. We find it important that in the down-current direction from the shallow marine deposits of the Louiseberget bed (e.g. at Kvalvågen, Thomsonbreen

Figure 6. A random 2 km long section of the diachronous model of Gjelberg and Steel (1995) in the same vertical and lateral scale. Even if there are con-tinuous exposures along the 2 km long section the diachronous onlapping relationship would be impossible to see in field outcrops. It is therefore unfair to use the lack of such observations as an argument against the model since there are no exposures of such quality along the depositional dip direction.

1 km

5 km

Bohemansflya

Forkastningsfjellet

Helvetiafjellet

Hanaskogsdalen

Skoltenn

Drønnbreen

Fleksurfjellet

Breikampen

Glitrefjellet

Storlengja

Lundkvistfjellet

Richterfjellet

Bolt-odden

2

3

4

5

6

7

8

9

10

11

12

13

1

Carolinefjellet Fm

Agardhfjellet Fm


Barremian time over a large part of the northwesterly Arc-tic area, and it is tentatively suggested that this was a result of tectonic movements probably related to intraplate stress (Gjelberg & Steel 1995).

Stratigraphic ArchitectureThe subdivision of the Helvetiafjellet Formation into the Festningen and the Glitrefjellet members as introduced by (Parker 1967) was questioned by Gjelberg & Steel (1995), particularly because of the diachronous and interfingering nature of the fluvial channel deposits (Festningen Member) and the lower delta plain facies (Glitrefjellet Member). Both with respect to facies, lateral connectedness and age, the flu-vial depositional system in the lower part of the formation is complex and is developed into thick units of braided and meandering stream deposits which are not always directly connected along the unconformity. These types of deposits also occur in the upper part of the formation, particularly as meandering channel deposits. The main problem with this subdivision emerges when trying to define the boundary between the two members (Fig. 8). Except in cases of a very limited dataset (e.g., in single wells), it turns out to be almost impossible to define this boundary even within a small area (few kms), and the problem increases with increasing data density. The reason for this is that the complexity of the sys-tem is not recognized with limited data, and there are sev-eral examples where it seems to be much easier to correlate sections that are located several tens of kilometers away from each other compared to sections only a few kilometers apart.

The stratigraphic revision of the Mesozoic succession by Mørk et al. (1999) removed the subdivision of the Festnin-gen and Glitrefjellet Members because of the interfingering character of this lower fluvial unit. However, it was reintro-duced by Midtkandal et al. (2008) who claimed that there is a regional transgressive surface on top of the Festnin-gen Member (Fig. 8B). Such a transgression surface is pres-ent locally, but not regionally, as is seen in the larger data-base of measured sections (Gjelberg & Steel 1995). Conse-quently the boundary between the two members cannot be defined on the basis of this assumed transgressive surface. Instead a complex zigzag facies transition line will define

onlapping relationship similar to that outlined by Gjel-berg & Steel (1995), covering a distance of 160 km and with velocities somewhat similar to the Upper Cretaceous and lower Cenozoic succession on the Norwegian shelf (and with no tectonic disturbances), the onlapping geom-etry would scalewise be approximately as shown in Figure 7. Such vertical/lateral display is common for regional seis-mic lines, and onlapping geometries like the one outlined in Figure 7 is commonly seen along many unconformities, both on the Norwegian Shelf and numerous other places around the world. This suggests that the onlapping geome-try presented in the diachronous transgressive model is not unusual with respect to angular relationships compared to many other unconformities.

Regional implicationsIt is not well established what caused the sudden change in basin configuration in Barremian time, leading from the generally more muddy offshore sediments below to the dis-persal of the coarser grained Helvetiafjellet Formation, but it is highly likely that this very prominent basinward facies shift was a forced regression associated with a significant fall in relative sea-level. The reason for this fall was proba-bly a (rotational) tectonic uplift (strongest in the north) that continued in the later Cretaceous, as documented by the progressively stronger northward erosion of the overlying Carolinefjellet Formation (Nagy 1970). It is also well known that similar sandstone intervals are present in many other locations in the Arctic, such as the Isachsen Formation in the Sverdrup Basin. The de Geer Line between Spitsber-gen and Greenland, which coincides with the Wandel Hav Strike-Slip Mobile Belt was reactivated several times, with strike- slip activity at the transition between Early and Late Cretaceous successions (Stemmerik et al. 1998, Håkansson & Pedersen 2001), resulting in a local hiatus. The sandstone dominated Galadriel Fjeld Formation at Kilen and the upper part of the Ladegårdsåsen Formation on East Pearcy Land corresponds broadly to the Helvetiafjellet and Caro-linefjellet Formations (Dypvik et al. 2002). However, the marked unconformity as seen at the Barremian/Hauteri-van boundary on Svalbard is not developed in these areas. It is suggested that a relative sea-level fall took place in

00 50km 100km 150km

1

2

3

4

TW

T S

Carolinefjellet Fm

Rurikfjellet Fm

Agardhfjellet Fm

Helvetiafjellet Fm

Figure 7. Immature river profile with deposition of braided, meandering and straight distributary channels respectively in the accommodation zone along the profile. The Ebro rive profile is enclosed for comparison. The width of Ebro delta plain is some 20 km.


documented. Such ravines would probably not develop in an extremely low gradient system.

In summary, Figure 9 illustrates the phases of development proposed for the forced regressive-to- transgressive archi-tectural model for Helvetiafjellet formation.

Collapse development in Boltodden areaThere is also disagreement about a series of collapse fea-tures on the east coast of Spitsbergen. These were inter-preted as delta-front collapse (Nemec et al. 1988, Gjelberg and Steel 1995, Prestholm and Walderhaug 2000) and later as shelf-edge collapse and slope-gulley development (Steel et al 2001), not least because of a succession of turbidites and debris flows below the level of collapse. This was dis-puted by Midtkandal (2007) who suggested that the rotated blocks were part of an elongate, braided fault system, and later by Onderdonk & Midtkandal (2010) who further sug-gested that the slide blocks actually slid westwards from a tectonic uplifted area to the east. The debate on these fea-tures is not further pursued here as it deals with issues not related to the main stratigraphic model discussed above. Lately it has been documented that delta-plain slide blocks have been encountered in a fully cored CO2 injection research well in Adventdalen (Mork-Jansson et al. 2010).

the boundary (Fig. 8A). However, we agree that there is a clear dominance of fluvial channel facies in the lower part of the formation, because of the overall transgressive char-acter of the succession.

Similar experience with correlation has also been found within the time equivalent Isachsen Formation in the Sver-drup Basin where it also turned out to be very difficult to correlate between the complex deltaic facies distribution, even over short distances (e.g. Tullius et al. 2011).

The incised ravine or valley exposed in the middle part of the formation at Thomsonbreen on the east coast (see Gjel-berg & Steel 1995, their Fig. 12) represents a channel that was close to 4-500 meters wide and about 20-30 m deep, and filled with mouth bar or bayhead delta sediments. This is a relatively small incised drainage system, but according to the diachronous depositional model it would extend 10 -20 km landward before reaching fluvial channel facies, and another 50 km before reaching the sediment by-pass zone (Fig. 2). In the layercake model of Midtkandal and Nystuen (2009) such an incision would extend throughout the entire Spitsbergen and way out into the Barents Sea due to the very low gradient. Likewise, the ravines represented by the infill of the Louiseberget bed on the east coast (Midtkan-dal et al. 2009) may represent a somewhat similar setting, but here the direct contact to the basal unconformity is well

Figure 8. Two simplified correlation alternatives, to illustrate the difference between the diachronous depositional model (A) of Gjelberg and Steel (1995) and the “layercake” model (B) of Midtkandal and Nystuen (2009). The white lines represent timelines. SB = sequence boundary

Ca 10 km

Maxi

mun

exp

ose

d w

indow

Festningen Member

Glitrefjellet Member

Carolinefjellet Formation

Ca 5

0 m

Ca 10 km

Ca 5

0 m

Festningen Member


A B

Transgressive surface

SB

SB

Braided stream deposits

Meandering stream deposits

Shallow marine sandy deposits

Lower delta plain heterolithic deposits

Shallow shelf deposits


Janusfjellet Subgroup, and were probably also deposited within collapse gullies of the transgressing coastline.

These slide blocks probably represent delta-plain sedi-ments, possibly time equivalent to the prodelta shale of the

1) Deposition of the «Agardhfjellet Delta» complex.

2) Rotational uplift, relative sea level fall and erosion of the «Agardhfjellet Delta» complex. Forced regression system tract wedges developed gradually further eastwards.

3) Forced regressive systems tract developed close to a turnaround point to the east.Relative sea-level rise started to dominate.

4) An overall relative sea-level rise resulted in the deposition of the Helvetiafjellet Formation. Fourth order sequences developed as aconsequence of short periods of relative sea-level fall.

5) The final development of the Helvetiafjellet Formation, with the transition to the overlying Carolinefjellet Formation. The transitio is characterized by semi-continious transgression surfaces.

Agardhfjellet Fm, Janusfjellet Subgroup

Carolinefjellet Formation


Festningen Member

50 km100 m

Agardhfjellet delt

Figure 9. A simplified model to explain the development of the Helvetiafjellet Formation according to the diachronous model of Gjelberg and Steel (1995).


References

Boothroyd, J.C. & Ashley, G.M. 1975. Process, bar morphology and sedimentary structures on braided outwash fan, Northeastern Gulf of Alaska. In: A.V.Jopling & B.C. McDonald (eds): Glaciofluvial and Glaciolacustrine Sedimentation. Spec. Publ. Soc. Econ. Paleont. Miner. 23, Tulsa. 193-222.

Dypvik, H., Håkansson, E. and Heinberg, C. 2002. Jurassic and Cretaceous palaeogeography and stratigraphic comparisions in the North Greenland-Svalbard region. Polar Research 21, 1, 91-108

Gjelberg, J. & Steel, R. 1995. Helvetiafjellet Formation (Barremian-Aptian), Spitsbergen: characteristics of a transgressive succession. In: Steel, R., Felt, V.L., Johannesen, E.P. and Mathieu, C. (eds.): Sequence Stratigraphy on the Northwest European Margin. Norwegian Petro-leum Society (NPF). Special publication 5. 571-573

Gustavson, T.C. 1974. Sedimentation on Gravel Outwash Fans. Journal of Sedimentary Research, 44 (2), 374-389.

Haq, B.U., Hardenbol, J. & Vail, P.R. 1988. Mesozoic and Ceno-zoic chronostratigraphy ad cycles of sea-level change. In: Sea Level Changes: An Integrated Approach (Ed. by C.K.Wilgus, B.S. Hastings, C.G.St.C. Kendall, H.W. Posamentier, C.A. Ross & J.C.VanWagoner), SEPMSpec. Publ., 42, 71-108.

Håkansson, E. & Pedersen, S.A.S. 2001. The Wandel Hav Strike-Slip Mobile Belt – A Mesozoic plate boundary in North Greenland. Bulletin of the Geological Society of Denmark, 48, 149-158.

McDonald, B.C. & Banerjee, I. 1971. Sediments and Bed Forms on a Braided Outwash Plain. Canadian Journal of Earth Sciences, 8 :(10) 1282-1301.

Midtkandal, I., Nystuen, J.P. & Nagy, J., 2007, Paralic sedimentation on an epicontinental ramp shelf during a full cycle of relative sea-level fluctuation; the Helvetiafjellet Formation in Nordenskiöld land, Spitsbergen. Norwegian Journal of Geology, vol. 87, pp. 343-359

Midtkandal, I., Nystuen, J.P. Nagy, J. & Mørk, A., 2008. Lower Cretaceous lithostratigraphy across a regional subaerial unconformity in Spitsbergen: the Rurikfjellet and Helvetiafjellet formations. Norwegian Journal of Geology, Vol. 88, pp. 287-304

Midtkandal, I. & Nystuen, J.P. 2007. Sequence stratigraphy of a low-gradient, ramp shelf; the Lower Cretaceous on Svalbard. In: Brekke, H., Henriksen, H. and Haukdal, G. (eds): SEST - The First confe-rence on Shelf Edge and Shoreline Trajectories, a Dynamic Approach to Stratigraphic Analysis. Geological Society of Norway, Tromsø, Norway, 330.

Midtkandal, I. & Nystuen, J.P. 2009. Depositional Architecture of a Low-gradient Ramp Shelf in an Epicontinental Sea: the Lower Cretaceous on Svalbard, Basin Research, 21, 655 - 675.

Mørk, A., Dallmann, W.K., Dypvik, H., Johanssesen, E. P., Larsen, G.B., Nagy, J., Nøttvedt, A., Olaussen, S., Pchelina, T.M., and Wor-sley, D., 1999. Mesozoic lithostratigraphy, in: W.K. Dallmann (Ed.) Lithostratigraphic lexicon of Svalbard, Review and recommendations for nomenclature use. Upper Palaeozoic to Quaternary bedrock. Norsk Polarinstitutt, Tromsø, pp.127-214

Mork-Janson , W.A., Mørk, A., Olaussen , S. and Helland -Hansen , W. 2010 .Lower Cretaceous gravity and sandy shelf deposits in Adventdalen, Spitsbergen- a response of local tectonics? In: Brekke et al Arctic Days 2010 NGF Abstracts and Proceedings of the Geo-logical Society of Norway, p. 36

Nagy, J. 1970. Ammonite faunas and stratigraphy of Lower Cretaceous (Albian) rocks in southern Spitsbergen. Norsk Polarinst.Skr. 152, 58pp.

Nemec, W.,1992. Depositional controls on plant growth and peat accu-mulation in a braidplain delta environment: Helvetiafjellet Forma-tion (Barremian – Aptian), Svalbard, In: McCabe, P.J., and Parrish, J.T., (Eds.), Controls on the Distribution and Quality of Cretaceous Coals. Geological Society of America Special Paper 267, Boulder, Colorado 209-226.

Nemec, W. Steel, R.J., Gjelberg, J., Collinson, J.D., Prestholm, E. & Øxnevad, I.E. 1988. Anatomy of Collapsed and Re-established

SummaryThe discussion around the depositional and stratigraphic model for the Helvetiafjellet Formation can be summarized as follows:1. The transgressive, diachronous depositional model for

the Helvetiafjellet Formation involves braidplain gradi-ents that are comparable with modern braidplains, even though the gradient in the model is still low compared with many recent analogues. Modern braidplains com-monly show gradients up to 20 m/km, but may be less than 1.5 m/km at the downdip termination of the sys-tem.

2. The layer-cake model implies a gradient for the braid-plain between less than 20 cm/km. Such low gradients are problematic and difficult to explain in the context of modern braidplains.

3. The diachronous model explains how it is possible to have marine-influenced facies close to the basal uncon-formity at proximal sites (around Festningen and Helve-tiafjellet) whereas braided stream deposits occur close to the unconformity more than 100 km downdip.

4. In a layer-cake model we should expect to see a grain size reduction down the floodplain from a proximal position in the northwest to a distal position in the southeast, over a distance of 150 km. Such a grain size reduction is not seen. On the contrary, there is an overall grain size increase in this direction. In a punctuated transgressive model this is not difficult to explain.

5. The angle of transgressive onlap or the gradient of the basal unconformity is difficult to estimate, and there are several possible solutions as to how “flat” the diachron-ous model should be. However, if the Festningen Mem-ber represents a braidplain, a gradient of less than 1 m/km would be unlikely. On the other hand, if the Festnin-gen Member was deposited by meandering or anasto-mosing rivers in a broad incised valley system, a much lower gradient (<0.1 m/km) is possible, but in this case it would be difficult to explain the lateral facies transitions discussed above.

6. There is no evidence of a regional transgressive sur-face on top of the Festningen Member. The boundary between the Festningen and the Glitrefjellet members is interfingering, producing multiple, stacked transgressive surfaces.

7. An incised valley or ravine (4-500m wide) filled with mouth bar sediments that occurs in the middle part of the formation on the east coast of Spitsbergen (Gjelberg & Steel 1995, their Fig. 12), represents one of the charac-teristic, forced-regressive pulses within the Glitrefjellet member. This valley is southeast- northwest orientated and probably reflects a relative sea level fall of a few 10s of meters. It is likely connected to a fluvial drainage system updip. It is easier to explain this in a dynamic stratigraphic model rather than in an aggrading layercake model.

Acknowledgements – We thank William Helland-Hansen and Snorre Olaussen for helpful comments during revision of the manuscript.

53

Delta Front in Lower Cretaceous of Eastern Spitsbergen: Gravita-tional Sliding and Sedimentation Processes. American Association of Petroleum Geologists Bulletin 72, 454 - 476.

Onderdonk, N. & Midtkandal, I. 2010. Mechanisms of collapse of the Cretaceous Helvetiafjellet Formation at Kvalvågen, Eeastern Spits-bergen. Marine and Petroleum Geology, 27, 2118-2140.

Parker, J.R.,1967. The Jurassic and Cretaceous sequence in Spitsber-gen. Geological Magazine 105 (5), 487-505.

Prestholm, E. & Walderhaug, O., 2000. Synsedimentary faulting in a Mesozoic deltaic sequence, Svalbard, Arctic Norway- Fault geo-metries, faulting mechanisms and sealing properties. Am. Assoc. Petrol. Geol. Bull., 84, 505-522.

Steel, R. J., Crabaugh, J., Schellpeper, M., Mellere, D., Plink-Bjor-klund, P., Deibert, J., and Loeseth, T,. 2001. Deltas vs. Rivers on the shelf edge: their relative contributions to the growth of shelf-margins and basin-floor fans (Barremian and Eocene, Spitsber-gen). In: Weimer , P. et al. (eds) Deep-Water Reservoirs of the World, Proceedings of the GCSSEPM Foundation 20th Annual Research Conference , 981-1009.

Steel, R. J. & Worsley, D. 1984. Svalbard’s post-Caledonian strata – an atlas of sedimentational patterns and palaeogeographic evolution . In: Spencer, A.M. et al. (eds.): Habitat of hydrocarbons on the Norwegian continental margin, 109-135. Graham & Trotman, London .

Stemmerik, L., Dalhoff, F. Larsen, B.D., Lyck, J., Mathiesen, A. & Nilsson , I. 1998. Wandel Sea Basin, eastern North Greenland. Geology of Greenland Survey Bulletin 180, 55-62.

Tullius, D., Anfinson, O., Leier, A., Pedersen, P. & Guest, B., 2011. Outcrop Stratigraphy and Sedimentology of the Lower Cretaceous Isachsen Formation of the Sverdrup Basin, Ellef Ringnes Island, Arctic Canada (Abstract). Recovery – 2011 CSPG CSEG CWLS Convention.

Walter, J. 1894. Lithogenesis der Gegenwart Beobachtungen über die Bildung der Gesteine an der heutigen Erdoberfläche. Dritten Teil einer Einleitung in die Geologie als historische Wissenschaft: 535-1055. Jena: Verlag Gustav Fischer.

NORWEGIAN JOURNAL OF GEOLOGY Depositional model for the Lower Cretaceous Helvetiafjellet Formation on Svalbard

depositional model for the lower cretaceous ... · modern depositional systems with braid-plains or...

Documents