MIDDLE–LATE HOLOCENE PALEOCLIMATE AND WEATHERING HISTORY IN THE NORRA STORFJÄLLET MOUNTAINS, SWEDEN: DEGLACIAL RECORD AND SOIL STRATIGRAPHY APPLIED TO NEOGLACIATION IN LOWER LATITUDES WILLIAM C. MAHANEY1 and VOLLI KALM2 1

Quaternary Surveys, Thornhill, Ontario, Canada and Department of Geography, York University, York, Ontario, Canada 2 Institute of Ecology and Earth Science, Tartu University, Tartu, Estonia

Mahaney, W.C. and Kalm, V., 2013. Middle–Late Holocene paleoclimate and weathering history in the Norra Storfjället mountains, Sweden: deglacial record and soil stratigraphy applied to Neoglaciation in lower latitudes. Geografiska Annaler: Series A, Physical Geography, 95, 145–158. doi:10.1111/geoa.12008 ABSTRACT. The retreat record of the Stabre Glacier into the Norra Storfjället mountains, after separation from the massive Tärnaån Glacier at some undetermined time in the Atlantic Chron, is documented by recessional moraines in the foreland. While poorly constrained by radiometric dating, the age of the middle group of moraines averages out to less than 4000 cal 14C yr BP, the older moraine group probably of Late Atlantic age, with the youngest deposits of Little Ice Age (LIA). Soils/paleosols range from Entisols (youngest) and Inceptisols (middle group) to mature Spodosols (outer group), existing either as single-story profiles or within pedostratigraphic columns, buried pedons either surfaced with weathered glacial or mass wasted deposits. Some profiles exhibit convoluted properties which could place them in the Cryosolic order. The physico-mineralchemical properties of soils/paleosols in recessional deposits across this sequence provide weathering indices over the mid to Late Holocene in the Swedish sub-Arctic climate. It is likely the middle group of deposits represents stillstand of the retreating glacier offset by climatic deterioration with the onset of Early Neoglacial climate which altered the glacial mass balance, at least until termination of the LIA. Correlation to other alpine areas in the middle and tropical latitudes with similar records is attempted and discussed. While the Stabre Glacier disappeared after the LIA, the nearby Tärna Glacier remains extant on the land surface, a presumed result of slight elevation differences between the two cirques which affects storm tracks and resultant variations in glacial mass balances. Key words: Holocene soils/paleosols, Neoglaciation, ice fluctuations, post-depositional weathering, Holocene paleoclimate reconstruction

Introduction The three main glaciers in the Norra Storfjället, Södra Syterglaciären, Norra Syterglaciären and Tärnaglaciären, all deposited sizable deposits of till and glaciofluvial sediment, draped to the east and northwest from the mountain massif that provided the vast bulk of mineral source material (Fig. 1 for topography; Fig. 2 for site location). With the breakup and segmentation of Atlantic Chronozone ice (>5 ka), these glaciers began to retreat away from the massive Tärnaån Glacier that occupied the Tärnasjön Valley. The Tärnaån Glacier, originally a broad trunk glacier fed by cirque glaciers straddling the Storfjället Massif south of the Syterbäcken Catchment (Fig. 1), eventually thinned into the main trunk glacier straddling the 600 m a.s.l. contour. As retreat accelerated north along the Tärnaån Valley, ice vacated the Stor-Laisan and Oltokjaure lakes during the Atlantic Chron retreating to the north. Feeder glaciers, with accumulation zones on the Storfjället, separated into cirque glaciers, Södra Syterglaciären, Tärna and Stabre, the latter two as one complete system in the early stage (Mahaney and Kalm 2012). Subsequent ice withdrawal at least by early Neoglacial time (Subboreal Chron) led to the separation of the StabreTärna Glacier into separate bodies. The weathering and soil morphogenesis record of the moraine/ glaciofluvial/mass wastage deposit sequence of the Stabre Glacier is the subject of this paper. Recession of the nearby Tärna Glacier, 3 km north of Stabre Cirque, left a small cirque glacier in place while the Stabre Glacier ablated out after the

© The authors 2013 Geografiska Annaler: Series A, Physical Geography © 2013 Swedish Society for Anthropology and Geography DOI:10.1111/geoa.12008

145

WILLIAM C. MAHANEY AND VOLLI KALM

Fig. 1. Topography of the field area.

Little Ice Age (LIA) leaving only well established firnpack (Fig. 3). The cirque floors at both sites are at nearly the same elevation and with similar aspects and average temperatures, whereas the cirque wall elevations on the massif are higher behind the Tärna Glacier. With similar cirque-floor microclimates and deglaciation between the two cirques, leading to ice retention in one (Tärna) and disappearance in the other (Stabre), the presence/ absence of ice is probably a function of either storm track differences or variable drifting and greater accumulation of snow across the broader higher massif to the north. Stratigraphic, clay mineral and chemical weathering research in Arctic-alpine Sweden is limited primarily to Allen et al. (2001), who correlated clay mineral distributions in a toposequence in the Kärkevagge area of northern Sweden. Their hypothesis that cold-based ice preserved preweathered clay minerals, lately distributed to lower sites 146

following deglaciation, correlates closely to findings discussed herein. Other work in the same area (Darmody et al. (2012) demonstrates that chemical weathering in Arctic terrain is an active process of denudation and that pyrite oxidation may prove to be an active early weathering process. Otherwise, soil stratigraphic studies in the Scandinavian Mountains are rather limited. Previous research on moraine recession elsewhere in European mountains, the Alps for example, has focused on radiocarbon dating (Matthews and Dresser 1983; Nicolussi and Patzelt 2000; Matthews et al. 2006), glacier variations (Ivy-Ochs et al. 2009), timberline fluctuations (Patzelt et al. 1997), soil geochemical discrimination of soils and paleosols (Mahaney et al. 1996), surface exposure dating (Kelly et al. 2004), and resultant post-depositional weathering and soil genesis. While multi-parameter lithologic/soil studies have been attempted in the Austrian Tyrol (Pindur and Heuberger 2010; Mahaney et al. 2011) to test relative-age dating methods, the database and interpretation presented here is the first to average soil profile development over ~4 ka in the Swedish mountains and assess the degree to which Neoglaciation slowed glacial recession, thus affecting glacial stillstand/readvance events. The approximate age of the outer group of deposits is based upon radiocarbon controls discussed previously by Mahaney and Kalm (2012). Moraines within the deglacial sequence are grouped in recessional order, representative profiles chosen out of a population of 12 pedons: outer group (STBY1, 4), middle group I (STBY3) and inner group II (STBY5 and 9) and LIA (STBY2). Regional geology and field area The regional geology is discussed in Mahaney and Kalm (2012). The only necessary addition of importance is the mean annual temperature (MAT) at Hemavan (473 m a.s.l.) which is –0.4°C. Mean annual precipitation at Hemavan is 681 mm (SMHI 1980), whereas climatic norms are not known with precision at 1000 m a.s.l., using the standard lapse rate of 1°C 100 m-1, the MAT must be to the order of around –6°C. The MAP would undoubtedly be greater but impossible to estimate. Paleoclimate at the time the outer group of soils developed into Spodosols was probably close to the present climatic norms at Hemavan (located just west of the Stor-Laisan on Fig. 1), or perhaps climatic norms were warmer and wetter (Mahaney et al. 1995b).

© The authors 2013 Geografiska Annaler: Series A, Physical Geography © 2013 Swedish Society for Anthropology and Geography

MIDDLE–LATE HOLOCENE PALEOCLIMATE AND WEATHERING HISTORY

Fig. 2. Satellite imagery and location of sites in the field area. Sites STBY6, 7 and 8 are indicated to show position within the outer group of moraines. Analysis of the pedostratigraphy in Stabre 6 and 7 and the polygonal Cryosol in Stabre 8 are in Mahaney and Kalm (2012).

Materials and methods Deposits were mapped in the field and sites placed on the satellite imagery in Fig. 2. Selected deposits were sampled and sites established on major groups of moraines. Sections were dug by hand and cut back to expose fresh sediment to depths of ~1 m. Sediment/soil samples were collected for laboratory analysis and where possible peats and organic clayey silts in sites STBY6 and 7 were selected for radiocarbon dating (Mahaney and Kalm 2012). Soil descriptions follow guidelines set out by the National Soil Survey Center (1995). The Cox horizon designation originates from Birkeland (1999) who uses ‘ox’ to refer to soil with colors stronger than 10YR 5/3-8, while that of the Cu (unweathered parent material) from Hodgson (1976). The ‘h’ designation in A horizons is applied where sufficient color strength indicates appreciable humus (Canada Soil Survey Committee 1977). Soil color assessments were made using Oyama and Takehara’s (1970) soil chips. Approximately 500 g

samples were collected at the sites to allow for particle size as well as clay and primary mineral analyses. Samples were air dried and treated with H2O2 to oxidize organic material. Particle grade sizes follow the Wentworth Scale with the exception of the clay/silt boundary (2 mm) which follows the US Department of Agriculture. The materials were then wet sieved, and the