The Holocene 8,6 (1998) pp. 719–728

‘Little Ice Age’ nivation activity in northeast Greenland Hanne H. Christiansen* (School of Geography and Geosciences, University of St Andrews, St Andrews, Fife KY16 9ST, Scotland, UK) Received 9 January 1998; revised manuscript accepted 5 May 1998

Abstract: During the ‘Little Ice Age’ (LIA), the periglacial landscape in northeast Greenland responded relatively quickly to the climatic deterioration with increased nivation activity. Resulting nivation landforms and associated sediment basins are used as palaeoclimatic archives, documenting the significant geomorphological effect of nivation during the LIA period in the Zackenberg area. A nival basin studied in detail showed the onset and dominance of niveo-fluvial sedimentation in the period ad 1250–1420, followed by increasing niveoaeolian sedimentation between ad 1420–1500/1580. Finally, very rapid niveo-aeolian sedimentation prevailed during the period ad 1500/1580–1690. In the same period, some large nival fans accumulated in the Zackenberg lowland as a result of increased niveo-fluvial activity, and niveo-fluvial sediments were also deposited in preexisting fluvial valleys, demonstrating the large areal extension of the nivation activity. The chronology is based on nine AMS 14C determinations. The intense geomorphological terrestrial activity in northeast Greenland during the LIA is deduced to be primarily due to increased winter wind speeds enabling larger snowpatches to accumulate as a result of increased snowdrifting. Key words: ‘Little Ice Age’, niveo-fluvial sedimentation, niveo-aeolian sedimentation, nivation, northeast Greenland, Zackenberg, 14C AMS dating, snowdrift.

Introduction In periglacial, high Arctic northeast Greenland, seasonal and perennial snowpatches control the geomorphological development in large parts of the landscapes consisting of unconsolidated sediments (Christiansen, 1998). The Zackenberg area is typical of such landscape development, for which a nivation process-formsediment model was established (Christiansen, 1998). This shows that downslope of nivation hollows nival basins and fans are sedimentary sinks of sediment deposited as a result of nivation activity. In this paper nival basin and fan sediments are used for reconstructing dominant palaeo-wind directions and snowdrift activity in the Zackenberg area in northeast Greenland. Here no detailed record of the physical terrestrial conditions during the ‘Little Ice Age’ (LIA) exists. The study area is located at the edge of the North Atlantic basin (Figure 1), very close to the climatically important area where deep-water formation takes place in the North Atlantic Sea, which makes further palaeo-climatic information from this area important. When investigating former small-scale climatic variations, such as the LIA, investigations of nival sinks can be more useful than traditional glacier variation studies, especially as cold-based highlatitude glaciers are characterized by having relatively long *Permanent address: Institute of Geography, University of Copenhagen, Øster Voldgade 10, 1350 København K, Denmark.

 Arnold 1998

Figure 1 Location of places mentioned in the text in Greenland and surrounding area. Century represents the ice core in northwest Greenland, while GISP2 and Crete show the location of these two ice cores on the central part of the Greenlandic Ice Sheet. EGC is the East Greenland sea current.

response times to climatic changes (Porter, 1986). In contrast, snowpatches, nivation forms, processes and sediments are more directly controlled by the prevailing climatic conditions, which make them respond quicker to changes particularly in winter wind 0959-6836(98)HL282RP

720 The Holocene 8 (1998)

direction and activity, and/or the amount of snow precipitation (Christiansen, 1998).

The ‘Little Ice Age’ (LIA) Timing and climatic changes The LIA represents a high-frequency climatic variation of relatively low magnitude, but with widespread environmental changes (Grove, 1988; Porter, 1986). It is thought to be caused by injection of sulphate aerosols into the atmosphere during major volcanic eruptions affecting the radiative energy balance on a decadal scale (Porter, 1986). Reduced sunspot activity in the Maunder Minimum broadly coinciding with the coldest part of the LIA has been noted (Grove, 1988). This cooling in combination with its triggering of changes in the unstable thermohaline circulation in the North Atlantic has been suggested to explain the full range of the LIA cooling (Stuiver and Braziunas, 1993). The LIA climate of the Northern Hemisphere has also been related to the N–S movement of the main westerly wind stream (Lamb, 1979), so that a southward movement of the cyclone tracks associated with strong westerlies and advances of polar air masses caused the cooling. Traditionally the LIA is associated with considerable glacier expansion in many world regions, within a few centuries between the Middle Ages and the warm period of the first half of the twentieth century (Grove, 1988). The timing and duration of cold periods in the LIA varied considerably between regions (Grove, 1988), and certainly several cooling periods constituted the LIA, as demonstrated, for example, in Iceland (Ogilvie, 1992). There is great uncertainty about when this period of relatively recent climatic deterioration began (Jones and Bradley, 1992). The warming that took place in many regions during the middle part of the present century, is seen as the termination of the preceding cold climate period, even though there is variation as to when it did terminate in different parts of the world (Jones and Bradley, 1992). Due to the focus on glaciation, some restrict the LIA to the period when glaciers responded most dramatically in the late part of the period. According to a review on LIA delimitation by Jones and Bradley (1992), the period has been defined variably as at one extreme ad 1250–1920 (Porter, 1986), and at the other extreme as ad 1550–1850, with the main phase from ad 1550– 1700 (Lamb, 1977). During the culmination of the LIA the reduction of the mean annual air temperature was about 0.5–1.2°C for the northern middle latitudes (Porter, 1986). The maximum reduction during the coldest parts of the LIA (ad 1550–1750) was about 1.5–2°C in Scotland, 0.9–1.5°C in England, 1.5°C in Denmark, 1.5°C in Zurich and 1.3°C in central Europe (Lamb, 1995). In northwestern Europe, particularly around the North Sea, the LIA period was marked by increased incidence and severity of storms and sea floods from the thirteenth century (Lamb, 1995). The increased wind activity was presumably caused by a cooling of about 5°C of the surface waters in parts of the North Atlantic, which produced a strengthened thermal gradient between about 50° to 65°N (Lamb, 1995). There was also a marked increase in wetness of northwestern Europe during the LIA period, beginning soon after ad 1300 (Lamb, 1995). Physical conditions in Greenland The varying isotopic composition (␦18O) in the Crete ice core (71°N, 37°W; Figure 1) from Central Greenland demonstrates that the coldest part of the LIA was from ad 1150 to 1400, which is about 300 years earlier than the coldest LIA period in Europe and North America (Dansgaard et al., 1975). This difference in timing was explained as being due to an increase in the average wavelength of the quasi-stationary Rossby waves (Dansgaard et al., 1975). The Crete ice-core data, combined with some other

detailed, directly dated climatic records from other areas, enabled the establishment of a Northern Hemisphere temperature index (Hammer et al., 1980). According to this the LIA lasted from ad 1350 to 1700 or 1900. The coldest period was from ad 1350 to 1700, interrupted by a minor ‘interstadial’ between ad 1500–1550 (Hammer et al., 1980). After ad 1700 the temperature was generally higher than in the early part of the LIA period. The isotopic deuterium signal from the GISP2 ice core (73°N, 39°W; Figure 1) indicates that the fourteenth century was the period with the lowest temperatures in central Greenland during the last 700 years (Barlow et al., 1997). The ␦18O variation from an ice core in the NW part of the Greenland Ice Sheet (77°N, 56°W; Figure 1) indicated that here the LIA cooling probably started already at about ad 1100–1150 and lasted to about ad 1750–1850 (Lamb, 1995: Figure 36). Information on the LIA in the peripheral ice-free land areas of Greenland is less continuous, though some sporadic data exist. In Illulissat, central west Greenland (Figure 1), a mean annual air temperature of about 2°C below modern values between ad 1873 and 1925 was demonstrated (Humlum, 1996). In central west Greenland on Disko Island, the dominating upper-air wind direction was from the southeast during the LIA (Humlum, 1987). Greenlandic glaciers are thought to have advanced from about ad 1750; the onset of this advance is, however, not known, as detailed information on glacier fluctuations are available only from about ad 1850 (Grove, 1988). A detailed compilation of information, mainly from Ogilvie (1984), on the occurrence of sea ice around Iceland in the LIA shows the East Greenland sea current to have caused sea ice to spread primarily along the northern and eastern coasts of Iceland with various extensions during cold years from around ad 1200 to 1900 (Grove, 1988).

Study area The Zackenberg area is located in high Arctic northeast Greenland at 74°30⬘N, 20°30⬘W (Figure 1). Nivation forms and processes in the sedimentary landscapes of Aucellabjerg and in the Zackenberg lowland are widespread (Christiansen, 1998). The Aucellabjerg (965 m a.s.l.) consists of Cretaceous sandstone capped by Tertiary basalts, and is covered by a thin layer of glacial deposits, particularly on the lower 400 m. In the Zackenberg lowland glacial landforms likewise exist, primarily as moraine hills, ground moraine and meltwater plains. According to preliminary investigations the area was deglaciated at about 10 to 9 ka BP (Christiansen and Humlum, 1993). Continuous permafrost exists in the area, with a mean active layer thickness of about 60 cm (Christiansen, 1997). The annual precipitation is only about 200 mm, and the mean annual air temperature about −10°C (Humlum, 1997). Most of the precipitation falls as snow during winter, when surface winds are dominantly from the N (Humlum, 1997). The summers are dry, nearly without precipitation, and the climate is semi-arid. An almost continuous vegetation cover exists in the Zackenberg lowland, except at the wind-exposed tops of moraine ridges and fluvial bars that have very little vegetation.

Sampling and methods Nival sediments have been studied by excavating profiles in identified, selected nival landforms and in landforms affected by deposition of nival sediments. Stratigraphical and sedimentological studies include grain size analyses and the determination of the carbon content by loss-on-ignition. Sediment samples were taken from characteristic, different parts of stratigraphically separate layers. The sediment samples were analysed at the Laboratory at

Hanne H. Christiansen: ‘Little Ice Age’ nivation in northeast Greenland

721

the Institute of Geography, University of Copenhagen. Grain-size distribution was measured using dry sieving for particles larger than 63 ␮m and sedimentation for grains smaller than 63 ␮m in a SediGraph 5100 Particle Size Analyser. The chronology is based on nine 14C Accelerator Mass Spectrometry (AMS) dates, obtained from stratigraphically as wide spanning positions as possible in the nival sediment, primarily from the nival basin, but also from the other investigated nival sediments in the Zackenberg area. This method and the dates are discussed in detail in a separate section and in the sections on the various nival landforms and sediments.

Nival basin and fan Morphology of a nivation hollow and the associated deposits A study of palaeo-nival sedimentation was carried out at a filled nival basin located immediately below a nivation hollow. The basin was chosen due to its clear appearance as a former nival sink, and because fluvial erosion of a channel through the otherwise permanently frozen basin sediments had partly exposed the sediments (Figure 2). This enabled sediment studies to a much larger depth than usually possible in continuous permafrost areas in this part of Greenland. The investigated nival basin is located at about 70 m a.s.l. at the foot of Aucellabjerg (Figure 3). Here remnants of coastal cliffs and terraces probably indicate the highest Late Weichselian sea level in the Zackenberg area, developed at about 9.5–9 ka BP (Christiansen and Humlum, 1993). Nival sedimentation in the investigated basin must be younger than 9 ka BP, as the basin is situated at the former coastline. The backwall of the nivation hollow is about 10 m high (Figure 2). It is eroded into glacial sediments and faces south, like most other nivation hollows and snowpatches in the Zackenberg area (Christiansen, 1998). The nivation hollow is about 100 m long, and contains a semi-perennial snowpatch. There is no vegetation in the nivation hollow. The nival basin is roughly 100 m long and 50 m wide; the surface is completely covered primarily by grass, and without any recent sediment deposition. The nearly horizontal surface of the nival basin is dissected by a polygonal pattern, about 4–5 m in diameter of active ice-wedges only 0.1 m wide (Figure 2). Active

Figure 3 Location of the investigated localities in the Zackenberg area. N = the nival basin and associated nival fan; K = the Kamelkegle nival fan; R = the Rylekegle nival fan; Ke = the Kaerelv valley with nival sedimentation. At point N the light grey elongated areas running NW–SE are moraine ridges without vegetation. The area immediately southwest of the moraine ridges at the foot of Aucellabjerg is dominated by large pronival, alluvial fans, where Rylekegle and Kamelkegle are located. The large Zackenberg Delta dominates the SW part. North is towards the top. The photograph covers approximately 3 × 3 km. (Aerial photograph 17 July 1973, Danish National Survey and Cadastre.)

ice-wedges 1–6 m wide form large polygonal patterns in other parts of the Zackenberg lowland. Clearly the width of the icewedges in the Zackenberg lowland seems to be somehow correlated with the age of the sediments in which they formed, suggesting that the nival basin sediments are relatively young. Nival meltwater from the snowpatch in the nivation hollow as well as from other upslope snowpatches presently erodes the lowermost part of the nivation hollow and the innermost part of the nival basin. This is causing the development of a pronival stone pavement and erosion of an organic layer in the lower part of the nivation hollow and innermost nival basin (Figure 2). The meltwater also erodes the channel, about 1–2 m wide and from

Figure 2 The investigated nivation hollow and nival basin. A person is standing in the channel eroded through the inner part of the nival basin, where a pronival stone pavement is developing. C50 marks the position of a 14C AMS sample from a layer rich in organic material in the innermost nival basin. The channel in the front is where the basin sediment was studied in the section marked by S. Note the polygonal pattern in the basin sediments seen to the right of the person. (Photograph 16 July 1996.)

722 The Holocene 8 (1998)

0.5 to 1.3 m deep, crossing from the inner to the outer part of the nival basin (Figure 2). Through this channel the nival meltwater flows presently to the area downslope of the nival basin, where recent nival deposition takes place, forming a large pronival, alluvial fan (100 m long and maximally 100 m wide). Here nival sedimentation takes place along the sides of shallow meltwater channels as local, small levees or as thin sheets of sediment covering some few m2, while most parts of the fan are covered by vegetation. The sediment being deposited during summer on this fan seems to be primarily eroded from the nival basin along the channel, but also from the inner part of the nivation hollow, thus partly representing reworked nival sediments. Both the nival basin and the nival fan are dammed by moraine ridges running NW–SE (Figure 3). Nival basin sedimentology and stratigraphy In the central, deepest part of the nival basin, where the meltwater channel is about 1.3 m deep, a 115-cm thick section in the nival sediments has been investigated (Figure 4). The section is dominated by massive, structureless, light brown-coloured sediments with pockets of dark-coloured organic material. There is a weak overall horizontal stratigraphy, but with only a few distinct layers. The boundaries between the layers are often diffuse. Usually the

stratigraphy of nival basins and fans in Zackenberg is more distinctive than seen in this basin, with well-defined thin layers of intervening coarse- and fine-grained sediments (Christiansen, 1998). Grain-size analyses of the characteristic layers in the basin section are shown in Figure 5. At the bottom 115 to 110 cm of the section a grey-coloured diamict, with stones (5–25 cm) and pockets of sand and organic material, is exposed. The pockets of organic material are located on the surface of the diamict, in depressions between stones. Presumably this diamict represents a till beneath the nival sediments. The nival sediments comprise the upper 96 cm of the section. From the depth of 110 to 85 cm the nival sediments show some sorting (Figure 5). Up to 90% of the sediment in this part of the basin is within the sand fraction. The carbon content is 3.1% for sediment immediately overlying the bottom diamict and 0.2 and 1.0% in the upper part of this lower layer (Figure 6). Organic material is concentrated in horizontal layers in the otherwise massive, structureless sediment, resulting in a weak layering. Due to the sorting and layering of the sediment, these lower nival sediments were probably deposited by meltwater draining from a snowpatch in the nivation hollow. Between 14 and 85 cm depth the nival basin sediment is homogeneous. Generally the sorting is poor (Figure 5), and the mean grain size is about 0.03 mm. Organic material is concentrated as pockets within the sediment primarily from 85 and 65 cm depth, with some small pockets of gravel found next to the organic material. This distribution is reflected by the highest carbon content (3%), in the lower part of this layer (Figure 6). From 65 to 14 cm depth the sediment has no structures, but there are some spots of orange colour between 14 and 24 cm, probably showing a secondary precipitation of iron. As there are only few structures, such as the pockets of organic material in the sediment, and a mean grain size in the silt fraction, this upper layer presumably represents direct meltout and deposition of niveo-aeolian sediment and organic material from a melting snowpatch. Presently niveo-aeolian sediment is particularly seen to be deposited in the centre zone of often very shallow nivation hollows; normally when vegetation-free, deflated fluvial bars or moraine ridges, exist in the snowdrift source area (Christiansen, 1998). The concentration of organic material in the bottom of the layer must indicate that initially the snowdrift source area was covered with vegetation. As niveo-aeolian erosion continued the surface of the source area was eroded and only a scarce vegetation cover left. If this interpretation is correct, the former snowpatch must have been larger during the summer, when the upper part of the basin sediments was deposited, compared to its present size. Today the basin area is free of snow early in the summer, and only a thin winter snow cover exists in this area. The existence of almost 1 m of nivation sediment filling the basin below the nivation hollow indicates, that in the past, conditions for large nivation activity have been more favourable than at present. Below the uppermost 7 cm of moss and grass growing at the basin surface, from 7–14 cm depth, there is a 7–8-cm thick, darkcoloured and transformed organic rich layer, with the highest carbon content in the section (4.4%; Figure 6). This organic top layer represents a period when plants have been able to grow on the basin surface, when deposition of the nival sediments ceased. 14

Figure 4 The studied section in the central nival basin sediments. The section is 115 cm deep. Ruler for scale. The holes in the section are caused by sampling. G109, G98, G95, G80, G48 and G13.5 mark the position of the grain-size samples presented in Figure 5. C113, C80, C61, C14 and C7 indicate the position of 14C AMS samples shown in Table 1 and Figure 6. (Photograph 21 July 1996).

C AMS dating To determine the chronology of the nival basin sediments, five samples of macroscopic plant fragments (twigs or leaf stems) from organic pockets or layers in the basin section have been dated using the 14C AMS technique. By using this method, bulk samples were avoided, as rootlets penetrate the basin sediment. Samples for 14C AMS dating were taken from the surface of the bottom diamict below the nival sediments, at 113 cm depth, and

Hanne H. Christiansen: ‘Little Ice Age’ nivation in northeast Greenland

723

Figure 5 Grain-size distribution of sediment samples from the nival basin.

Figure 6 Reconstructed LIA deposition history of the investigated nival basin. The calibrated 14C AMS dates are marked by filled circles, while the horizontal lines indicate the uncertainty as ⫾1 standard deviation on each date. Average sedimentation rates of the basin based on the dates are presented. C% is the content of carbon determined as loss-on-ignition in the sediment samples.

724 The Holocene 8 (1998)

from the niveo-aeolian layers, at 80 cm and 61 cm depth. Samples were also taken from the top and bottom of the dark-coloured organic layer at the top of the basin sediments, below the living surface vegetation (Figure 4). The 14C AMS dating was carried out at the Institute of Physics and Astronomy, University of Aarhus, in Denmark. All 14C AMS ages mentioned in the text are in calibrated calendar years using the Seattle calibration programme, version 3.03 (Stuiver and Reimer, 1993; Table 1). Unfortunately, there is a rather long plateau in the 14C age calibration curve from about 100–200 14C yr BP and a minor one from 300–400 14C yr BP, within which age differentiation is difficult. These plateaux correspond in calibrated years approximately to the periods ad 1650– 1950, and ad 1450–1650. Ages are given as a time interval if there is more than one intercept of the measured 14C age with the calibration curve. 14C AMS dates, with the uncertainty of one standard deviation, are shown in Table 1 and Figure 6, including the supplementary samples from outside the nival basin discussed later. As organic material does not form separate continuous layers in the nival basin, except at the top of the basin (Figure 2), it is probably not autochthonous, but must have been deposited contemporaneously with the surrounding sediment. This is also indicated by the fact that the organic material consists of relatively well-preserved, large plant fragments only, and no complete plants. Presumably the organic material can then be only slightly older than the nival basin sediments. Erosion of organic material from the snowdrift source area probably was rather continuous when the organic rich individual layers or organic pockets were deposited. In all likelihood the difference between the primary age of the plant fragments, found by 14C dating, and the time of their deposition in the basin is well within the uncertainty of the dating results. Nival basin sedimentation Twigs deposited on the surface of the diamict, in the investigated section, are dated to ad 1250 (AAR-3093; Figure 6; Table 1), indicating that nival sedimentation started then. The 14C AMS

ages of nival organic material above are ad 1450 (AAR-3094) and ad 1510–1615 (AAR-3095; Figure 6; Table 1). The base of the organic layer at the top of the nival sediments has been dated to ad 1690–1955 (AAR-3096), while the top of this layer is modern, ad 1960–1990 (AAR-3097; Figure 6; Table 1). Average net sedimentation rates for the nival basin sediments were calculated based on the interpolated 14C AMS dates (Figure 6). Due to the difficulty with exact calibration of all the 14C ages, only intervals of sedimentation rates could be determined for the upper part of the nival basin. As the top of the lower, transformed organic rich layer, above the nival sediments, is dated to be modern (AAR-3097), and, as this organic rich layer is 7 cm thick, evidently this layer cannot entirely have accumulated during modern time. If the base of the organic layer should be from ad 1955 (AAR-3096), the accumulation rate within the organic layer would have been as high as 0.2 cm yr−1, which seems an unrealistically high value in this high Arctic environment. Most likely some modern carbon has contaminated the lower part of this layer, presumably as decomposition products of roots penetrating, or in percolation water, which are possible ways of downwards transport of carbon described in palaeosols (Matthews, 1981; 1992). Therefore basin sedimentation must have ceased closer to ad 1690 than to ad 1955. Accepting this, nival deposition presumably can be confined to the interval from ad 1250 to about ad 1690 (Figure 6) or not long after. In the niveo-fluvial lower part of the basin average net nival sedimentation was about 0.14 cm yr−1 from ad 1250 to 1420, assuming sedimentation to have been an approximately continuous process. In the upper part of the basin the average niveoaeolian sedimentation was 0.12–0.32 cm yr−1 from ad 1450 to 1510–1615, rising to 0.26–0.63 cm yr−1 from ad 1510–1615 to about 1690, assuming a constant accumulation rate. As the niveoaeolian sediment is homogenous without signs of any break in sedimentation, presumably sedimentation was continuous. It is not directly known if niveo-aeolian sedimentation was initiated immediately after niveo-fluvial deposition ceased, but the relatively short period between the two lower dates, from these two

Table 1 14C AMS dates from the Zackenberg area. All 14C results are reported in conventional radiocarbon years BP (before present 1950) and in calibrated calendar ages. The 14C AMS ages have been calibrated using the Seattle calibration programme version 3.03 (Stuiver and Reimer, 1993). 1955* denotes possible influence of bomb 14C, i.e., post-1955. Modern* denotes negative 14C ages, i.e, a 14C activity ⬎ 100 pmc (percent modern carbon = 0.95 × activity of the oxalic acid standard). The high activity values are normally due to anthropogenic 14C introduced by nuclear bomb test in the atmosphere. Based on the known variation of the atmospheric 14C content after ad 1955, a crude estimate of the calendar age of the sample is given in brackets Sample lab. no., no. from Figures 2, 4 and 7

14

Calibrated age, ± 1 standard deviation

␦13C (‰) PDB

7

AAR-3097, C7

(1960–1990)

–27

Base of organic layer, central nival basin

14

AAR-3096, C14

Modern* 117.3 ± 0.7 pmc 140 ± 40

–29.0

Niveo-aeolian organic material, central nival basin

61

AAR-3095, C61

360 ± 45

Niveo-aeolian organic material, central nival basin

80

AAR-3094, C80

435 ± 50

113

AAR-3093, C113

800 ± 50

50

AAR-3098, C50

435 ± 50

Niveo-fluvial organic material, Kamelkegle nival fan

23

AAR-3100, C23

320 ± 70

Niveo-fluvial organic material, Rylekegle nival fan

50

AAR-3101

370 ± 60

Niveo-fluvial organic material in valley bottom, Kærelv valley

70

AAR-3099

325 ± 60

ad ad ad ad ad ad ad ad ad ad ad ad ad ad ad ad

Material dated, site location

Top organic layer, central nival basin

Organic material below niveo-fluvial sediment, central nival basin Bottom of organic layer, inner nival basin

Depth (cm) of sample below terrain surface

C age BP

1690–1955, 1680–1955* 1510–1615, 1465–1635 1450, 1430–1480 1250, 1220–1280 1450, 1430–1480 1530–1640, 1480–1660 1490–1610, 1450–1640 1530–1630, 1480–1650

–29.4 –25.1 –28.5 –27.4 –27.6 –28.1 –26.4

Hanne H. Christiansen: ‘Little Ice Age’ nivation in northeast Greenland

layers, seems to suggest that no longer break in sedimentation occurred. Nival deposition took place during the ‘Little Ice Age’, and the increasing sedimentation rate upwards in the nival basin can be seen as a consequence of deteriorating climatic conditions during the coldest part of this period, when the snowpatch seems to have reached maximum size. The main concentration of organic material in the nival sediments between 85 and 65 cm depth was deposited from ad 1420 to 1500–1580 (Figure 6). A sample of leaves and stems (AAR3098) from the bottom of an organic enriched layer located in the innermost part of the nivation basin (Figure 2), next to the investigated basin part, have likewise been dated to ad 1450 (Table 1). This deposition of organic material is thus almost contemporaneous with the onset of niveo-aeolian sedimentation in the basin. The content of organic material in the basin sediments is reduced in the following period of maximal niveo-aeolian sedimentation. This distribution of organic material in the sediments could indicate that during the early part of the niveo-aeolian sedimentation period, a vegetation cover still existed in the snowdrift upland on Aucellabjerg. After about 80 to 160 years of sedimentation, at ad 1500–1580 (Figure 6), the vegetation cover presumably was much eroded, and barren deflation surfaces must have been dominant in the snowdrift upland, resulting in a lower content of organic material in the basin sediments. In the following period, from ad 1500–1580 to about 1690, niveo-aeolian sedimentation was further intensified and the basin filled, presumably partly due to better sediment availability in the then more deflated snowdrift upland area.

Nival fan below the nival basin The fan sediments below the nival basin were investigated in two shallow sections, in the apex and in the central part of the fan, 50 m from the apex. However, due to the location of the permafrost table at 53–63 cm depth, only the upper part of the sediments could be investigated. The apex fan section, 63 cm deep, was dominated by massive, homogeneous sediments, without sediment structures. The grain size was primarily in the silt and sand fraction, and the sorting was poor, as in the nival basin. As suggested for the upper part of the nival basin, the sedimentological and stratigraphical characteristics of these sediments, particularly the lack of layering, could mean that also this sediment had melted out directly from a snowpatch with a high content of niveo-aeolian sediment. In the 53 cm section in the central fan, horizontal, 0.2–14 cm thick, distinct layers were found. Thin layers primarily contained silt, while thicker layers were dominated by sand and gravel. Several layers were relatively well sorted. As an identical type of sediments accumulates due to modern niveo-fluvial sedimentation on this part of the fan, the characteristically layered stratigraphy in all likelihood represents niveo-fluvial sedimentation. When the snowpatch was larger, ad 1420–1690, causing niveoaeolian deposition in the nival basin, it might have reached as far as the upper parts of the fan area below, explaining the existence of the assumed niveo-aeolian sediments in this part of the fan. This would imply that the snowpatch was at least 50 m wider in the downwind direction during the LIA than today. When subsequently the channel was eroded through the basin sediments, the morphology of the fan evolved. The channel was initiated when more snow meltwater ran to this part of the basin, probably because the extension of the snowpatch was no longer so large as during the period of intense deposition of niveo-aeolian sediments in the nival basin. However, as the fan consists of partly resedimented material from the nival basin, in particular from the inner part where significant amounts of organic material have been recently eroded, no attempts have been made to date these sediments using 14C techniques.

725

Nival fans in the Zackenberg lowland Lining the foot of Aucellabjerg, close to the investigated nival basin and fan, several large alluvial fans exist (Figure 3). They are 100–200 m wide and more than 200–300 m long. Morphologically the fans are connected to large channels running down from 300–400 m asl. on Aucellabjerg (Figure 3). Meltwater from snowpatches above and along these channels has, since deglaciation, been draining through these channels, transporting sediment down to the Zackenberg lowland area (Christiansen, 1998). The channels are presently partly inactive and their slopes covered by vegetation, leaving only the central deepest parts still active. The associated large alluvial fans located at the downstream end of the channels are likewise covered by vegetation. Sedimentation here takes place only as small patchy sheets (10 × 10 m), deposited during the rapid early spring snowmelt, when niveo-fluvial discharge for a limited period, often no more than 1–3 days, is very high. Due to brimful conditions, or because of snow blocking the watercourses on the fans, water spills over parts of the fans, bringing sediment onto the fan surfaces. Two of the fans, Kamelkegle (40 m a.s.l.) and Rylekegle (45 m a.s.l.), were located about 600 m and 850 m WNW of the nival basin (Figure 3). The sediments in these fans were investigated to determine if the nival activity that caused these fans to accumulate was contemporaneous with nival sedimentation in the investigated basin. Shallow sections were excavated in the fans down to the frozen sediments, presumably close to the permafrost table, at respectively 55 and 50 cm below the fan surfaces. Both fans showed an identical stratigraphy, dominated by well-defined, horizontal layers of fine- or coarse-grained sediments with some organic material in some layers (Figure 7). The silt, clay and organic rich layers were generally not more than 2–6 cm thick,

Figure 7 Section in the Kamelkegle fan, showing a typical niveo-fluvial stratigraphy of relatively thin fine-grained layers as the C23 layer, and thicker coarse-grained layers as the one above C23. The section is 55 cm deep. C23 marks the location of the 14C AMS sample AAR 3100 (see Table 1) in the top of the layer rich in organic material in the middle of the section. Ruler for scale. (Photograph 29 July 1996.)

726 The Holocene 8 (1998)

while the sand and gravel layers typically were about 10 cm thick. Some layering was seen in the fine-grained layers, whereas the sand and gravel dominated layers were massive. As the large-scale morphology of these fans and their stratigraphy resembles the earlier described nival fan below the nival basin, it can be assumed that they have been deposited primarily as a result of large scale niveo-fluvial erosion on Aucellabjerg. Because nearly no precipitation falls as rain at present, it is assumed that, particularly in cold climate Holocene periods, rainfall must also have been rare, excluding the accumulation of the described large-scale fans as due to any significant rainwater activity. Obviously, the thick coarse-grained layers represent periods with intense, large scale niveo-fluvial deposition, while the fine-grained layers have been deposited during periods of only minor niveo-fluvial activity. A characteristic 4–6-cm thick organic rich layer exists in each fan section, consisting of silt and pieces of organic material. The organic content is primarily plant fragments, indicating that they must represent allochthonous plant material. As the fragments are relatively large, it seems that the transport distance was not long, stressing that they must have been derived from the nearby slopes of Aucellabjerg during fan accumulation, presumably in a period of major erosion of the vegetation cover on Aucellabjerg. Therefore is it assumed that the difference between the primary age of the organic material and the time of deposition is small. If the plant pieces should represents autochthonous plant material from an in-situ plant cover on the fan, as seen in the profile top layer, surely more organic material should remain in the layer, more complete plants should have been found, and the amount of sediment should have been much less. From each of the two fan sections a sample of large plant stems was collected for 14C AMS dating from the uppermost part of this well-defined organic rich layer, at 23 cm (AAR-3100, Kamelkegle, Figure 7) and at 50 cm depth (AAR-3101, Rylekegle). The samples were dated to ad 1530–1640 and ad 1490–1610 (Table 1), and evidently the two dated layers seems to have been deposited within the same period. This indicates probable contemporaneous deposition on the two fans, and that other similar fans located equally at the foot of Aucellabjerg were in all likelihood also deposited during this period. Moreover, the period when the organic layer was deposited in both fans, ad 1490–1610, is the same as when organic material was primarily deposited in the nival basin (Figure 6). This synchronization probably indicates a period of widespread nivation activity, most likely caused by high snowdrift activity, causing larger snowpatches to accumulate on the slopes of Aucellabjerg. This would have lead to increased niveo-aeolian and niveo-fluvial erosion, allowing areas that had not earlier been affected by nivation activity to become eroded, and lose their vegetation cover. Average net sedimentation rates calculated for the sediments above the dated organic layers are 0.05–0.06 cm yr−1 for Kamelkegle and 0.10–0.13 cm yr−1 for Rylekegle. These average values reflect lower rates of niveo-fluvial sedimentation since the termination of the LIA than during the early LIA period, when nivation activity still had not increased much, with niveo-fluvial sedimentation in the lower part of the investigated nival basin of about 0.14 cm yr−1 (Figure 6). The dating of the organic rich layer in Kamelkegle and Rylekegle furthermore indicates that, when the investigated nival basin was being filled quickly with niveoaeolian sediment including organic material, niveo-fluvial sedimentation dominated on the neighbouring nival fans.

Nival deposition in fluvial valleys in the Zackenberg lowland About 500 to 1000 m southwest of the foot of Aucellabjerg, some fluvial valleys originate in the Zackenberg lowland area. They are

typically 5–30 m deep, 50 to 200 m wide, and drain towards the southsoutheast (Figure 3). Since deglaciation these valleys have been modified by nivation activity (Christiansen, 1996), partly filling them with sediment, creating a flat valley bottom. Small watercourses, about 1 m wide, presently drain through these valleys, eroding the valley bottom sediments. Approximately 1 km south of the Kamelkegle fan, in the upper part of one of these valleys, Kaerelv (Figure 3), sediments were investigated in the sides of the recent watercourse. These sections, up to 80 cm thick, consist of distinctive, 1–5-cm thick, horizontal layers of silt, sand, gravel and organic material, overlying glaciofluvial sediment. This stratigraphy probably reflects primarily niveo-fluvial sedimentation, postdating the glacio-fluvial erosion of the valleys, with much the same sort of sediments and stratigraphy as found in the nival fans described earlier. The sediment and organic material were probably eroded by snow meltwater above and in the valley, and transported to the valley bottom, where deposition took place. In the bottom of the investigated section a sample of plant stems (AAR-3099) from an organic layer 2–3 cm above the glacio-fluvial sediment was dated by 14C AMS to ad 1530–1630 (Table 1). This is contemporaneous with the deposition of organic debris in the nival basin and in the Kamelkegle and Rylekegle nival fans, stressing that niveo-fluvial deposition then presumably was a widespread phenomenon in the Zackenberg lowland. This caused the onset of sedimentary filling of the glacio-fluvial valleys, contemporaneously with increased niveo-aeolian sedimentation started in the nival basin.

Discussion The presented results indicate that the sedimentary landscapes in the Zackenberg area are climatically sensitive, responding with increased nivation activity to the climatic deterioration in the LIA from ad 1250 to ad 1690 and that this geomorphological activity is contemporary with the coldest period recorded in the Greenlandic ice cores. Snowpatches were significantly larger especially during the period ad 1420–1690, when deposition of primarily niveo-aeolian sediment and organic material in the investigated nival basin took place, and when niveo-fluvial sedimentation was widespread in the Zackenberg lowland. Investigations of LIA physical terrestrial conditions in East Greenland have so far been scarce. Lithofacies and foraminiferal evidence from shallow sediment cores in Nansen Fjord in southeast Greenland (Figure 1) have shown variable climatic conditions with frequent intervals of severe cold to characterize the LIA from ad 1630 to 1900, when perennial sea-ice cover was frequent (Jennings and Weiner, 1996). Likewise an earlier cold interval lasted from ad 1270 to about ad 1370–1470, presumably attributed to the deteriorating climatic conditions at the onset of the LIA (Jennings and Weiner, 1996). There is evidence of enhanced geomorphic activity with increased snowdrift and aeolian activity caused by dominant northerly and northeasterly winds on Ammassalik Island, southeast Greenland, during ad 1100–1500 (Christiansen et al., 1998). The reconstructed nivation activity in Zackenberg thus seems to agree chronologically very well with these few other evidences of the LIA in East Greenland. West of Greenland, in the eastern Canadian Arctic, a build-up of permanent snowfields on Baffin Island (Figure 1) occurred during the LIA, as a widespread and rapid lowering of regional snow lines on upland plateaux took place (Barry et al., 1975). A large area in which lichens were killed on Baffin Island has been identified and is suggested to indicate the extent of a LIA extensive seasonal or permanent snow cover or glacial extension (Andrews et al., 1976; Grove, 1988). Cassiope tetragona plants at Ellesmere Island (Figure 1) in the Canadian high Arctic were killed by

Hanne H. Christiansen: ‘Little Ice Age’ nivation in northeast Greenland

increased snow accumulation, but not because of summer temperature lowering during the LIA from ad 1425 to 1665 (Havstro¨m et al., 1995). This evidence shows that in neighbouring high Arctic areas increased amounts of snow accumulated in the landscape during the LIA, presumably also primarily as an enlargement of snowpatches. Generally it has been suggested that, due to the LIA climatic impact, most of the vegetative cover in the Canadian high Arctic is of recent origin (Svoboda, 1982). This could easily also be the case for large parts of the Zackenberg area, particularly those affected by the LIA increased nivation activity. In Zackenberg, the reconstructed larger snowpatch extent in the investigated nival basin in parts of the LIA can only have accumulated if more snowdrifting took place. This was only possible if larger amounts of snow were available in the landscape, or if the winter wind speed was higher than at present. The prevailing winter wind direction must, as today, have been from the north during the LIA, because the enlarged snowpatches were located on the south side of Aucellabjerg as today. According to the GISP2 icecore data, snow accumulation in central Greenland was greatly reduced during cold periods in the Holocene, and the atmospheric circulation, rather than air temperature, seems to have been the primary control on snow accumulation over the past 18 000 years (Kapsner et al., 1995). Specifically, the LIA snow accumulation rate as reconstructed in the GISP2 ice core was generally lower than the preceding ‘Medieval Warm Period’ (Meese et al., 1994). Therefore, LIA precipitation in the Zackenberg area most likely was less than at the recent relatively warmer climate, and more intense LIA snowdrifting must primarily have been caused by augmented wind activity, particularly by enhanced winter wind speeds, enabling larger snowpatches to accumulate. This is also demonstrated by a recent 1995–97 interannual wind increase of 0.8 m/s, causing larger and longer lasting snowpatches in the Zackenberg area. Primarily the average annual winter wind speed increased, causing an early summer snowpatch downwind widening of 20–40 m, despite a contemporary interannual reduction in the total amount of snow precipitation of about 21%. The high LIA nival sedimentation rates of 0.14 to 0.63 cm yr−1 demonstrate nival sinks to be highly sensitive palaeoclimatic records, with a good resolution, suitable for detailed reconstructions. In a high-lying glacially eroded valley on the mountain Zackenberg, in the bedrock part of the Zackenberg area, an average Holocene nival sedimentation rate of only 0.03 mm yr−1 was determined for sediments younger than 9.0 ± 1 ka BP (Christiansen, 1994). This value is significantly lower than found in the investigated LIA nival sediments, stressing the lithological control on nival deposition rates. During the coldest part of the LIA, the Norse Greenlanders in southwest Greenland disappeared completely around ad 1350– 1500 after having been established at about ad 980 (Barlow et al., 1997; Lamb, 1995). Soil erosion took place in the area of the Norse society during the period ad 1000–1450 (Jacobsen, 1991), just as soil and aeolian erosion, sediment accumulation and slopewash in Iceland have been reported beginning after ad 900, with widespread change after ad 1510 (Gudmundsson, 1997). These geomorphological processes partly resemble the LIA niveo-fluvial and niveo-aeolian activity reconstructed from the Zackenberg area. However, the high Arctic Zackenberg area was far beyond the range of cultivation in the LIA period, and the documented nivation activity in this part of Greenland must be caused solely by changes in the climatic conditions. The contemporary geomorphic activity in southwest Greenland and Iceland was partly ascribed to the initiation of Norse settlement and cultivation. From about ad 1200 the connection to Iceland and Europe began to be impeded by increasing amounts of sea ice (Lamb, 1995). Regular contact ceased in ad 1369, and by ad 1500 the Norse Greenland settlement no longer existed. Norse burials took

727

place deep in ground that became permanently frozen not long after ad 1350 (Grove, 1988), clearly demonstrating the effect of the climatic deterioration. Contemporary with this initial LIA climatic deterioration in Greenland relatively high niveo-fluvial and niveo-aeolian sediment rates are found in Zackenberg. The erosion of organic material was concentrated in the early part of the period with highest nivation activity from about ad 1420 to 1500–1580. The significant and contemporaneous deposition of organic material in the nival basin, in the Kamelkegle and Rylekegle fans and in the Kaerelv valley, indicates that these landforms represent carbon sinks. After this period it appears that the source areas for nival erosion of organic material became partly depleted, indicating that the vegetation cover was then reduced. As high nival sedimentation continued for some time, probably until about ad 1690, the permafrost table was quickly raised in these landforms, whereby the accumulated organic material was preserved in the permafrost and thereby removed from the carbon cycle during the LIA. Only subsequent lateral erosion in these landforms releases small amounts of this LIA carbon. The 14C AMS dating of organic material incorporated in nival sediments show that it is possible to use this method in a semiarid high Arctic environment to date even relatively young material, despite the calibration plateau problems. Downwards transport of modern carbon in the investigated profiles has been restricted since deposition, as this can only take place in the active layer during the short and relatively dry summer period. In dry, continuous permafrost areas, it therefore seems possible to use 14 C AMS dating, in particular when sedimentation rates are high and redeposition of organic material rapid. The calibration plateau problem can be partly overcome by dating several samples in the same section, as demonstrated for the nival basin.

Conclusion Nivation forms and sediments have been demonstrated to represent important palaeoenvironmental archives in periglacial sedimentary landscapes, where the geographic extension of nivation is large. The LIA represents a minor climatic deterioration, lasting for 350–750 years in Greenland, during the period ad 1100–1900 according to the different palaeoclimatic evidences. The mean annual air temperature reduction was at least 2°C. One of the overall geomorphological effect of these climatic variations was significantly increased nivation activity, larger than at present in the form of rapid filling of nival basins, deposition of large nival fans and nival sedimentation modifying pre-existing landforms in the Zackenberg area. Here there was presumably less snow precipitation, but increased winter wind activity and consequently larger snowdrifting from the north from ad 1250 to 1690, but in particular from ad 1420 to 1690. Niveo-aeolian sedimentation rates were at a maximum in the period ad 1500–1580 to 1690, as a result of the previous intensified snowdrifting, which eroded the vegetation cover, primarily during the period ad 1420 to 1500–1580. This indicates that high nivation activity can be initiated in periglacial sedimentary environments by even small climatic deteriorations, especially when winter wind activity increases.

Acknowledgements This paper forms part of the results of the research project ‘The Arctic Landscape: Interactions and feedbacks among physical geographical and biological processes’, funded by the special programme for Polar Science (1995–1997) of the Danish Natural

728 The Holocene 8 (1998)

Science Research Council. Jan Heinemeyer, Institute of Physics and Astronomy, University of Aarhus is thanked for carrying out the 14C AMS dates. I thank C. Caseldine and an anonymous reviewer for improving the paper. Thanks also to the muskoxen in Zackenberg, normally grazing during the summer on the investigated nival basin and fan areas, for allowing me to investigate the sediments during their rather short visits to neighbouring areas.

References Andrews, J.T., Davis, P.T. and Wright, C. 1976: Little Ice Age permanent snowcover in the eastern Canadian Arctic: extent mapped from Landsat-1 imagery. Geografiska Annaler 58A, 71–81. Barlow, L.K., Sadler, J.P., Ogilvie, A.E.J., Buckland, P.C., Amorosi, T., Ingimundarson, J.H., Skidmore, P., Dugmore, A.J. and McGovern, T.H. 1997: Interdisciplinary investigations of the end of the Norse Western Settlement in Greenland. The Holocene 7, 489–99. Barry, R.G., Andrews, J.T. and Mahaffy, M.A. 1975: Continental ice sheets: conditions for growth. Science 190, 979–81. Christiansen, H.H. 1994: Thermoluminescence dating of nival sediments from Zackenberg, northeast Greenland. Quaternary Geochronology (Quaternary Science Reviews) 13, 491–96. —— 1996: Nivation forms, processes and sediments in recent and former periglacial areas. Geographica Hafniensia A4, PhD thesis, 185 pp. —— 1997: Periglacial and glacial geomorphological research. In Meltofte, H. and Thing, H., editors, Zackenberg Ecological Research Operations, 2nd Annual Report 1996, Danish Polar Center, Ministry of Research and Information Technology, 57–59. —— 1998: Nivation forms and processes in unconsolidated sediments, NE Greenland. Earth Surface Processes and Landforms 23, 751–60. Christiansen, H.H. and Humlum, O. 1993: Glacial history and periglacial landforms of the Zackenberg area, northeast Greenland: preliminary results. Geografisk Tidsskrift 93, 19–29. Christiansen, H.H., Murray, A.S., Mejdahl, V. and Humlum, O. 1998: Luminescence dating of Holocene geomorphic activity on Ammassalik Island, SE Greenland. Quaternary Science Reviews, in press. Dansgaard, W., Johnsen, S.J., Reeh, N., Gundestrup, N., Clausen, H.B. and Hammer, C.U. 1975: Climatic changes, Norsemen and modern man. Nature 255, 24–28. Grove, J.M. 1988: The Little Ice Age. London: Routledge, 498 pp. Gudmundsson, H.J. 1997: A review of the Holocene environmental history of Iceland. Quaternary Science Reviews 16, 81–92. Hammer, C.U., Clausen, H.B. and Dansgaard, W. 1980: Greenland ice sheet evidence of post-glacial volcanism and its climatic impact. Nature 288, 230–35. Havstro¨m, M., Callaghan, T.V., Jonasson, S. and Svoboda, J. 1995: Little Ice Age temperature estimated by growth and flowering differences between subfossil and extant shoots of Cassiope tetragona, an arctic heather. Functional Ecology 9, 650–54. Humlum, O. 1987: Glacier behaviour and the influence of upper-air con-

ditions during the Little Ice Age in Disko, central West Greenland. Geografisk Tidsskrift, 87, 1–12. —— 1996: Origin of rock glaciers: observations from Mellemfjord, Disko Island, central west Greenland. Permafrost and Periglacial Processes 7, 361–80. —— 1997: Meteorological station. In Meltofte, H. and Thing, H., editors, Zackenberg Ecological Research Operations, 2nd Annual Report 1996, Danish Polar Center, Ministry of Research and Information Technology, 11. Jacobsen, B.H. 1991: Soil resources and soil erosion in the Norse Settlement Area of Østerbygden in southern Greenland. Acta Borealia 1, 56–68. Jennings, A.E. and Weiner, N.J. 1996: Environmental change in eastern Greenland during the last 1300 years: evidence from foraminifera and lithofacies in Nansen Fjord, 68°N. The Holocene 6, 179–91. Jones, P.D. and Bradley, R.S. 1992: Climatic variations over the last 500 years. In Bradley, R.S. and Jones, P.D. editors, Climate since ad 1500, London: Routlege, 649–65. Kapsner, W.R., Alley, R.B., Shuman, C.A., Anandakrishnan, S. and Grootes, P.M. 1995: Dominant influence of atmospheric circulation on snow accumulation in Greenland over the past 18,000 years. Nature 373, 52–54. Koch, L. 1945: The East Greenland ice. Meddelelser om Grønland, Band. 130, no. 3. Lamb, H.H. 1977: Climate, present, past and future. London: Methuen, 835 pp. —— 1979: Climatic variation and changes in the wind and ocean circulation: the Little Ice Age in the Northeast Atlantic. Quaternary Research 11, 1–20. —— 1995: Climate history and the modern world (second edition). London: Routledge, 443 pp. Matthews, J.A. 1981: Natural 14C age/depth gradient in buried soil. Naturwissenschaften 67, 472–74. —— 1992: Radiocarbon dating of buried soils with particular reference to Holocene solifluction, in Frenzel, B., editor, Solifluction and climatic variation in the Holocene, European Science Foundation, Akademie der Wissenschaften und der Literatur, Mainz: Gustav Fischer Verlag, 309–24. Meese, D.A., Gow, A.J., Grootes, P., Mayewski, P.A., Ram, M., Stuiver, M., Taylor, K.C., Waddington, E.D. and Zielenski, G.A. 1994: The accumulation record from the GISP2 core as an indicator of climatic change throughout the Holocene. Science 266, 1680–82. Ogilvie, A.E.J. 1992: Documentary evidence for changes in the climate of Iceland, ad 1500 to 1800. In Bradley, R.S. and Jones, P.D., editors, Climate since ad 1500, London: Routledge, 92–117. —— 1984 The past climate and sea-ice record from Iceland, part 1: data to ad 1780. Climatic Change 6, 131–52. Porter, S.C. 1986: Pattern and forcing of Northern Hemisphere glacier variations during the last millennium. Quaternary Research 26, 27–48. Stuiver, M. and Braziunas, T.F. 1993: Sun, ocean, climate and atmospheric 14CO2: an evaluation of causal and spectral relationships. The Holocene 3, 289–305. Stuiver, M. and Reimer, P.J. 1993: Extended 14C database and revised CALIB 3.03 14C age calibration. Radiocarbon 35, 215–30. Svoboda, J. 1982: Due to the Little Ice Age climatic impact most of the vegetative cover in the Canadian High Arctic is of recent origin: a hypothesis (abstract). Proceedings of the 33rd Alaska Science Conference, 206.