Geomorphological and historical records of the surge-type behaviour of Hansbreen (Svalbard)

Abstract

This paper presents geomorphological and historical records of the surge-type behaviour of Hansbreen, one of the most studied tidewater glaciers in Svalbard. The surge-type behaviour of the glacier has not been considered before due to the lack of evidence of this phenomenon. We integrate geomorphological mapping of the terrestrial and submarine forefields with historical data from the 19th and 20th centuries to reconstruct the glacier dynamics and identify the possible timing of surging. Landform assemblages are representative of the surging glacier landsystem, including crevasse-squeeze ridges (CSRs) and submarine streamlined glacial lineations. Abundant CSRs in the outer part of the terrestrial forefield were also documented in the 1980s, but most have been obliterated since then. We suggest the identified surge landsystem was produced during a surge of Hansbreen detected from photographs taken during the Austro-Hungarian expedition in 1872. Historical photogrammetric photos from the Norwegian expedition in 1918 revealed surge-diagnostic features in the glacier surface, including a folded medial moraine and a dense, complex network of crevasses. A potential next surge remains questionable in the following decades due to the low-lying accumulation area of the main stream hindering the mass build-up, but potential surges of the tributary glaciers should not be excluded.

Introduction

Surge-type glaciers exhibit multiyear, quasi-cyclical velocity fluctuations between a prolonged quiescent phase of slow flow and a shorter active phase of fast flow, during which ice velocities can increase by 10–1000 times (Meier and Post, 1969; Raymond, 1987; Sund and others, 2009). Most surge-type glaciers concentrate in spatial clusters within an optimal climatic envelope, such as Svalbard (74–81°N) (Sevestre and Benn, 2015). Glaciers cover 57% of the land of this archipelago and more than 60% of the glaciated area is represented by tidewater glaciers (Błaszczyk and others, 2009; Nuth and others, 2013). There is general agreement that surging is a common phenomenon in Svalbard (e.g. Liestøl, 1969; Jania, 1988; Lefauconnier and Hagen, 1991; Hagen and others, 1993; Sevestre and Benn, 2015; Farnsworth and others, 2016; Benn and others, 2023), although the estimated percentage of surge-type glaciers varies from 13% (Jiskoot and others, 1998) to 90% (Hagen and Liestøl, 1990). The duration of the quiescent phase typically ranges from 30 to 150 years, but can be as long as 500 years (Dowdeswell and others, 1991; Hagen and others, 1993; Murray and others, 2003). Therefore, the length of surge cycles often exceeds the observation period, which poses a risk of misclassification of glaciers as non-surging if the active phase has not been observed and associated geomorphological and glaciological records of surging have been obliterated since then (Benn and Evans, 2010).

The frequency of surges is influenced by climatic conditions through changes in the mass balance and modification of the thermal structure of glaciers (Dowdeswell and others, 1995; Sevestre and others, 2015). In recent centuries, Svalbard glaciers reached their maximum extent during the Little Ice Age (LIA) and many of these advances were surge-related (e.g. Liestøl, 1969; Lefauconnier and Hagen, 1991; Christoffersen and others, 2005; Sevestre and others, 2015; Farnsworth and others, 2016; Zagórski and others, 2023). The main sources of data about the timing of the late LIA surges are historical observations of glaciers that have recently advanced (e.g. the surge of Kongsvegen dated to c. 1800; Liestøl, 1988) or evidence of glaciers experiencing the active phase of a surge as recorded in historical drawings and maps (e.g. Recherchebreen in 1838; Liestøl, 1993, Kongsvegen in 1869; Liestøl, 1988, Nathorstbreen and Paulabreen mapped in 1898; Ottesen and others, 2008). Although the phenomenon of glacier surges was unknown at that time, this unique photographic and cartographic documentation can be reinterpreted to identify the evidence for past surge events (e.g. Zagórski and others, 2023). However, in regions with less archival data on glacier fluctuations during the LIA, such as Hornsund, most surges have only been recorded in the second half of the 20th and in the 21st century (Błaszczyk and others, 2013). The first maps presenting glacier termini in this area were prepared during the Austro-Hungarian North Pole Expedition 1872–1874 (Peterman, 1874) and the Swedish-Russian Arc-of-Meridian Expedition to Spitsbergen 1899–1902 (Wassiliew, 1925) when the glaciers were already extensive or even at their LIA maxima. The scales of these maps and the simplified cartographic representation of glaciers on the older map do not allow the identification of evidence of ongoing surges, such as extensive crevassing or folded medial moraines (Copland and others, 2003). Therefore, some surges may have occurred in this area during the LIA that are not documented in existing inventories.

In the case of sparse archival data, the application of surging glacier landsystem models based on characteristic landform assemblages in the terrestrial (Evans and Rea, 1999, 2003) and submarine forefields (Ottesen and Dowdeswell, 2006; Ottesen and others, 2008, 2017) allows the detection of surge-like behaviour of glaciers with no documented surging history (Christoffersen and others, 2005; Farnsworth and others, 2016, 2017; Flink and Noormets, 2018). The landforms are a record of a fully active surge with glacier advance (stage 3 according to Sund and others, 2009) marked with pro- and supraglacial moraines and characteristic meltwater drainage patterns (Lønne, 2016). One of the key features of the surge landsystems are crevasse-squeeze ridges (CSRs), which indicate basal crevassing and saturated basal sediments during the active surge phase, while the preservation of the ridges implies abrupt surge termination and ice stagnation during the quiescent phase (Evans and Rea, 1999; Ben-Yehoshua and others, 2023). Fast ice flow during the surge can be inferred from streamlined glacial lineations (Ottesen and Dowdeswell, 2006). Glaciotectonic moraines, often associated with debris flow lobes in the marine environment, reflect an advance onto and deformation of proglacial sediments (Lovell and others, 2018). Ice-cored hummocky moraines record entrainment and transport of high volumes of sediments during the advance, and subsequent melt-out on stagnant ice (Evans and Rea, 1999). Small submarine retreat moraines reflect minor winter advances of quiescent tidewater glaciers (Flink and others, 2015).

Hansbreen (Hornsund, south Spitsbergen) is one of the most studied glaciers in Svalbard and its possible surge behaviour has long been debated (e.g. Jania, 1988; Jania and Głowacki, 1996; Rachlewicz and Szczuciński, 2000; Ćwiąkała and others, 2018). The glacier and surrounding area have been the subject of numerous investigations (e.g. Jania, 1988; Glazovskiy and others, 1992; Vieli and others, 2002; Pälli and others, 2003; Grabiec and others, 2012; De Andrés and other, 2018; Laska and others, 2022) and its contemporary dynamics are well understood (Błaszczyk and others, 2021, 2024). According to the Randolph Glacier Inventory (RGI) V7.0, there is ‘no evidence’ that Hansbreen is a surge-type glacier (Maussion and others, 2023; RGI 7.0 Consortium, 2023). In contrast to most medium-sized or large tidewater glaciers in Hornsund, Hansbreen has not surged since the regular monitoring of glacier dynamics initiated in 1982 and there is no unequivocal evidence for surging earlier in the 20th century (Błaszczyk and others, 2013; Szafraniec, 2020). Over the last 40 years, three tidewater glaciers in Hornsund underwent a surge: Paierlbreen (advance of 380 m in 1992–1995), Mendeleevbreen (advance of 1.5 km in 1997–2002) and Svalisbreen (advance of c. 400 m in 2016–2020; Błaszczyk and others, 2023). Evidence of an ongoing surge was also reported in the case of Mühlbacherbreen (in 1961; Sund and others, 2009) and Körberbreen (in 1938; Liestøl, 1969). Marine-terminating Storbreen, Hornbreen and Samarinbreen are also surge-type glaciers (Fig. 1b; Jiskoot and others, 2000). Minor advances (c. 100–300 m) of Hansbreen were observed in 1957–1959 (Kosiba, 1960), 1973–1977 and 1993–1995, and an episode of increased surface velocity occurred in the early 1990s, but these events were interpreted as ‘mini-surges’ rather than fully developed surges (Jania, 1988, 1998). Landforms diagnostic of surging were not detected in the submarine forefield of Hansbreen when it was mapped with multibeam echosounder in 2008 (Tegowski and others, 2016; Ćwiąkała and others, 2018). Farnsworth and others (2016) classified Hansbreen as a surge-type glacier based on the presence of CSRs in the terrestrial forefield. The geomorphology of the recently deglaciated submarine and terrestrial marginal zones presents an opportunity to investigate newly exposed geomorphological evidence for possible surging. Several studies have also demonstrated the importance of integrating terrestrial and submarine geomorphological records with historical datasets and observations in order to evaluate evidence for past surging of tidewater glaciers (e.g. Ottesen and others, 2008; Lovell and others, 2018; Aradóttir and others, 2019; Zagórski and others, 2023).

Figure 1. (a) Location of Hansbreen in the Svalbard archipelago and (b, c) Hornsund. (b) Glaciers: Ho, Hornbreen; K, Körberbreen; Me, Mendeleevbreen; Mu, Mühlbacherbreen; P, Paierlbreen; Sa, Samarinbreen; St, Storbreen; Sv, Svalisbreen. (c) Glaciers (black letters): D, Deileggbreen; F, Fuglebreen; K, Kvitungisen; S, Staszelisen; T, Tuvbreen; V, Vrangpeisbreen. Mountain ridges; peaks and saddles (yellow letters): Fa, Fannytoppen; Fb, Fugleberget; Fp, Flatpasset; Fr, Flatryggen; VT, Vesletuva. Peninsulas and islands (orange letters): B, Baranowski Peninsula (Baranowskiodden); Hh, Hansholmane; O, Oseanograftangen. Bays: Hb, Hansbukta; I, Isbjørnhamna. PPS, Polish Polar Station Hornsund. Figure map: © Norwegian Polar Institute (Norwegian Polar Institute, 2014). Sentinel-2B from 27 July 2023.

Here, we present the geomorphology of Hansbreen's terrestrial and submarine forefields and combine this with historical data to evaluate its surge-type behaviour.

Study site

Hansbreen is a polythermal tidewater glacier that terminates in Hansbukta on the northern side of Hornsund in south-west Spitsbergen (Fig. 1; Jania and others, 1996; Błaszczyk and others, 2019). The system consists of the main trunk flowing from north to south, four main tributary glaciers on the west (Fuglebreen, Tuvbreen, Deileggbreen and Staszelisen) and several minor tributaries on the east. The glacier is approximately 14 km long with a c. 2 km-long grounded calving front, 51.3 km² in area (as of 2015), a total ice volume of 9.6 km³, a mean surface slope of 1.8° along the centreline and up to 5.6° along the centrelines of the tributaries (Grabiec and others, 2012; Błaszczyk and others, 2019, 2024). There is a distinct ice divide between Hansbreen and Vrangpeisbreen at the highest part of the main trunk at c. 490 m a.s.l., whereas the tributary glaciers reach up to 565 m a.s.l. (Błaszczyk and others, 2019). The delineation of the eastern border is not straightforward due to the transfluence of ice from the accumulation field of Hansbreen to Kvitungisen, a tributary of Paierlbreen. Around 2010, the mean and the maximum thickness of the glacier were 171 and 386 m, respectively, and the mean surface elevation was 306 m a.s.l. (Grabiec and others, 2012). The surface mass balance has been surveyed since 1989 and the mean annual value in 2000–2019 was −0.26 m w.e. (Schuler and others, 2020). In 2007–2015, the mean annual velocity measured c. 3.5 km from the calving front varied from 49.3 to 88.2 m a⁻¹ and the central part of the front was 4–5 times faster. The average retreat rate in 1992–2015 was 38 m a⁻¹ with a maximum of 311 m a⁻¹ (Błaszczyk and others, 2021). The bedrock of the main trunk is overdeepened below sea level for 10 km from the calving front, and this overdeepening corresponds to a quarter of the total bedrock area beneath the glacier system (Grabiec and others, 2012; Otero and others, 2017). The geological structure is complex and diverse with several stratigraphic units of N-S orientation. The main valley has developed within Middle and Late Proterozoic phyllites, dolomites and conglomerates as well as Cambrian dolomites and quartzites. The west tributary glaciers have eroded Middle Proterozoic quartzites, marbles, phyllites and schists (Birkenmajer, 1990; Ohta and Dallmann, 1999). In 1979–2018, mean annual air temperature and precipitation measured at the nearby Polish Polar Station were −3.7°C and 235 mm with increase trends of +1.14°C and +61.6 mm per decade (Wawrzyniak and Osuch, 2019, 2020).

Similar to other glaciers in Hornsund, the first maps of the Hansbreen frontal position were based on surveys by the Austro-Hungarian (1872–1874) and Swedish-Russian expeditions (1899–1902) (Peterman, 1874; Wassiliew, 1925). At this time, Hansbreen occupied Hansbukta, the Baranowski Peninsula (Baranowskiodden), part of Isbjørnhamna, Oseanograftangen and raised marine terraces at the foot of Fugleberget and Fannytoppen. The geomorphology of the forefield was investigated by Birkenmajer (1960), Karczewski and others (1984), Pękala (1989), Lindner and others (1992), Rachlewicz and Szczuciński (2000). Since that time, there have been no new studies of the entire terrestrial forefield. The submarine geomorphology of Hansbukta was outlined by Jania (1988, 1998) and Tegowski and others (2016), and described in detail by Ćwiąkała and others (2018).

Materials and methods

We mapped the terrestrial forefield by combining field investigations with remote sensing data, as recommended by Chandler and others (2018). Field observations were carried out in 2021–2023 with detailed mapping in August 2022. Our map is based on a very high-resolution orthophotomosaic and DEM generated and published by Błaszczyk and others (2022) from aerial photos acquired on 22 June 2020 using a Dornier DO228 aircraft. The resolution of the orthophotomosaic and the DEM are 0.0843 and 0.169 m, with a maximal vertical DEM error of 0.54 m. Mapping of the submarine forefield was based on two datasets of bathymetric data. We mapped most of the forefield using a high-resolution DEM (1 m) based on multibeam bathymetric data from the Norwegian Hydrographic Service collected in 2008 (The Norwegian Mapping Authority, Kartverket). This dataset was used in the previous geomorphological studies of Hansbukta (Tegowski and others, 2016; Ćwiąkała and others, 2018). In addition, we used a DEM generated by Błaszczyk and others (2021) from multiple data sources to map the areas not surveyed in 2008. The DEM was based on measurements from a small boat using a Norbit Wideband Multibeam Sonar with a positioning system (2014, 2016, 2017) and data collected from a small boat with a single beam echosounder Lowrance HDS5 with positioning (2015). The assessed accuracy of these bathymetric data was ±5 m (Błaszczyk and others, 2021). We mapped the landforms in QGIS using the WGS84/UTM33N spatial reference system. Some small areas of the forefields were not covered by the bathymetry data or the orthophotomap and a small part of the seafloor with a surveyed topography remained unclassified in the geomorphological map (Figs 2 and 3). Changes in the frontal position of Hansbreen were taken from supplementary data to Błaszczyk and others (2013, 2021) and vectorized from Sentinel-2 images in natural colour compositions for the period 2016–2018.

Figure 2. Hansbreen and its forefield. Orthophotomap from June 2020 after Błaszczyk and others (2022) and bathymetric map based on the Norwegian Hydrographic Service multibeam data (2008) combined with bathymetric data (2014–2017) after Błaszczyk and others (2021) were used to produce the geomorphological map (Fig. 3). Changes in the frontal position of Hansbreen after Błaszczyk and others (2013, 2021), Sentinel-2A from 2 August 2016 and 18 September 2018, and Jania (1988). The dashed line marks the glacier extent in 1938 based on terraphotogrammetric measurements conducted by Pillewizer (1939).

Figure 3. Geomorphological map of the forefield of Hansbreen. Glacier extent on 22 June 2020.

To characterize the changing glacier and its forefield over time, we used historical maps and photographs from 1872 (Austro-Hungarian North Pole expedition, Peterman, 1874), a map from 1899 (Swedish-Russian Arc-of-Meridian Expedition, Wassiliew, 1925), a map and archival photographs from the Norwegian expedition to Spitsbergen in 1918 (Hoel, 1929), oblique aerial photographs from 1936 by the Norwegian Polar Institute and the orthophotomap from these data (Geyman and others, 2022), terrophotogrammetric documentation of the Hansbreen forefield in 1982–1996 from the archives of the University of Silesia (Jania and Kolondra, 1982) and the orthophotomap based on vertical aerial photographs from 2011 by the Norwegian Polar Institute, available in the TopoSvalbard online archive (https://toposvalbard.npolar.no/). The historical photographs from 1872 were acquired from the Austrian National Library Digital Image Archive (https://onb.digital/) and from 1918 from the digital image archive of the Norwegian Polar Institute (https://bildearkiv.npolar.no/fotoweb/). The purpose of the historical data was to help interpret the formation of landforms and to assess past glacier frontal positions and any geomorphological or glaciological evidence for surging. As the surge criteria, we considered (1) looped moraines, (2) deformed ice structures, (3) heavy surface crevassing, (4) a surge bulge and (5) shear margins on the glacier surface (Lefauconnier and Hagen, 1991; Copland and others, 2003; Grant and others, 2009).

Hansbreen surging glacier landsystem

Glacial and glacifluvial landforms are classified according to their origin and morphological features and presented in the context of their associations (subglacial, supraglacial, ice-marginal, proglacial, glacimarine). We describe the terrestrial and submarine marginal zones separately to highlight the similarities and contrasts between landform assemblages in different environments.

Glacier ice and debris covered glacier ice

The main tongue of Hansbreen is densely crevassed at the active calving front, and similarly crevassed at the confluence zone between Tuvbreen and the main trunk (Figs 1 and 2). The dominant crevasse pattern is transverse with additional sigmoidal crevasses near the confluence zone. A dense network of crevasses with different orientations and crevasse traces (healed crevasses) can be observed in the eastern lateral zone of Hansbreen, including the land-terminating part, which transitions into the ice-cored hummocky moraine. The extensively crevassed central and eastern parts of Hansbreen contrast with the almost uncrevassed surfaces of Tuvbreen and Fuglebreen, which are dominated by crevasse traces and exhibit a few supraglacial streams and moulins.

Two medial moraines can be traced on the western part of the glacier (Figs 1 and 2). The medial moraine between the main trunk of Hansbreen and Tuvbreen takes the form of a wide belt of highly dispersed debris a few centimetres thick, with fine clasts of metamorphic rocks from the nunatak Tuva. Snow patches partially mask a scattered medial moraine between Tuvbreen and Fuglebreen composed of fine-grained, homogeneous debris sourced from the metamorphic rocks of the Vesletuva mountain. We interpret the former as an ice-stream interaction medial moraine and the latter as an ablation-dominant moraine.

Terrestrial forefield

Ice-moulded bedrock: Ice-moulded bedrock is exposed on abundant glacially eroded skerries and capes in Hansbukta and Isbjørnhamna, which often take the form of roches moutonnées (Fig. 3). The most prominent of these features is the Baranowski Peninsula, which is covered by a thin, discontinuous layer of fluted till with numerous bedrock outcrops of amphibolites. The surface of the peninsula is streamlined with minor roches moutonnées and rock drumlins (Fig. 4a). Similarly, streamlined bedrock with a fluted till surface occurs at the foot of Fannytoppen, across Hansbukta from the Baranowski Peninsula (Fig. 4b). A large, 9.5 m high roche moutonné, partially draped with a thin layer of fluted till is located in the north-western forefield near the glacier terminus (Figs 5a and c). In the south-eastern forefield, ice-moulded bedrock is also represented by glacio-isostatically uplifted marine abrasive terraces and sea stacks covered with till.

Figure 4. Streamlined topography of (a) the Baranowski Peninsula and (b) at the foot of Fannytoppen with parallel-sided flutings on a thin layer of till. Note CSRs transverse and oblique to flutes on the Baranowski Peninsula. Shaded relief maps after Błaszczyk and others (2022).

Figure 5. Fluted till surface with CSRs at the foot of Fugleberget in the north-western part of the forefield. (a) Overview of the flutings and CSRs networks on the orthophotomap from 2020 (Błaszczyk and others, 2022). (b) Flutings and CSRs emerging from downwasting glacier ice in August 2020. Hummocky moraine and trimline on the Fugleberget slopes in the background. (c) The same area in September 2023. Note two drumlins (dr) and a roche moutonné (rm) overprinted with flutings and well-developed CSRs. (d) A prominent CSR c. 0.8 m high with the extent marked with the dashed line. (b–d) Photos A. Osika.

Fluted till surface: Ice-moulded bedrock is strewn with a thin layer of coarse-grained, matrix-supported diamicton interpreted as subglacial traction till and a discontinuous layer of supraglacial diamicton of subglacial origin with similar properties (Skolasińska and others, 2016). The clasts are sub-angular to sub-rounded, ranging from gravels to boulders up to 2 m in diameter. The total thickness of diamictons is usually less than 0.5–1 m (Rachlewicz and Szczuciński, 2000).

The till surface is densely fluted, except in areas reworked by meltwater streams, where only the most prominent flutes can be recognized. Flutes are low, flow-parallel ridges composed of till similar to the till plain, but with significantly smaller clasts. Their formation is connected with infilling of till into a subglacial cavity in the lee-side of a boulder (e.g. Boulton, 1976; Benn, 1994; Ives and Iverson, 2019). Flutings concentrate in four spatial clusters: (1) the streamlined topography of the Baranowski Peninsula (Fig. 4), (2) the recently exposed north-western part of the forefield, at the foot of Fugleberget (Fig. 5), (3) raised marine terraces at the foot of Fannytoppen (Fig. 4) and (4) remnants of till plain on Oseanograftangen (Fig. 6). Most of the flutings are parallel-sided and up to 84 m long with nearly constant width between 0.3 and 2.4 m and height below 0.3 m, commonly with a boulder lodged on the stoss-side of the flute. There is a second group of short tapering flutes up to 14 m long, with boulders embedded at their heads and decreasing width and height in a downstream direction. The assemblages of long flutings associated with CSRs, which can be observed in the Hansbreen forefield, are vital components of the surging glacier landsystems (e.g. Evans and Rea, 1999, 2003; Christoffersen and others, 2005; Benn and Evans, 2010; Aradóttir and others, 2019; Ives and Iverson, 2019).

Figure 6. (a, b) Roches moutonnées on Oseanograftangen overprinted with fluted till and CSRs. Note the extensive beaches and storm ridges around the hills and remaining of the plain south of the lake, indicating the flutes and CSRs reworking by coastal processes. (c) Levelled flutes (parallel; blue lines) and CSRs (oblique and transverse; red lines). The latter take the form of low ridges (0.3–0.5 m) or diamicton trails a few centimetres high.

Crevasse-squeeze ridges (CSRs): Low, diamicton ridges oriented transverse or oblique to ice-flow direction and superimposed on flutes are interpreted as CSRs. CSRs are composed of the same till as flutings and their formation involves till infilling of basal crevasses that have opened during a surge and subsequent emergence of ridges from stagnating ice during the quiescent phase (e.g. Boulton and others, 1996; Rea and Evans, 2011; Farnsworth and others, 2016; Ben-Yehoshua and others, 2023). The landforms are a key characteristic of the geomorphological record produced by surging glaciers and have been reported in the forefields of numerous glaciers in Svalbard with both confirmed and undocumented surge histories (e.g. Heintz, 1953; Boulton and others, 1996; Glasser and others, 1998; Ottesen and Dowdeswell, 2006; Farnsworth and others, 2016; Ottesen and others, 2017; Flink and Noormets, 2018; Aradóttir and others, 2019). The morphology of CSRs in the forefield of Hansbreen ranges from distinct, intersecting ridges c. 0.5 m high to remnants of ridges or mounds and sediment trails reflecting the characteristic criss-cross pattern of CSRs. A network of well-defined cross-cutting ridges c. 0.5 m high, up to 3.5 m wide and up to 50 m long that overlay dense flutes can be seen emerging from the downwasting glacier snout in the northwestern part of the forefield (Fig. 5). The most prominent ridges approach the height of 0.8 m (Fig. 5d). In contrast, CSRs on the Baranowski Peninsula appear as indistinct ridges in the field and are best identified on the orthophotomaps. However, in the 1980s they were visible as abundant, distinct diamicton ridges transverse or oblique to underlying flutes (Fig. 7). In the eastern forefield, CSRs occur less frequently and are mostly individual diamicton ridges and mounds 6–40 m long and 0.3–0.5 m high overprinted on highly degraded flutes (Figs 6 and 8). The ridges are located among numerous cross-cutting diamicton trails a few centimetres high, which mimic the networks of CSRs and flutings. Similar features interpreted as the remains of CSRs were described in western Svalbard by Ben-Yehoshua and others (2023). In the southernmost forefield, transverse and oblique CSRs are superimposed on fluted till draped on two roches moutonnées (Fig. 6). Their morphology ranges from distinct, short ridges and mounds approaching 0.3–0.5 m in height, to degraded, intersecting ridges a few centimetres high taking the form of diamicton trails.

Figure 7. (a, b) Terrestrial photogrammetric photograph of the Baranowski Peninsula from 1982. Note long, parallel-sided flutings and low CSRs (white arrows) oblique and transverse to the flutes. Today, CSRs in this site are sparse and have an indistinct appearance in the field.

Figure 8. (a) Low, obliterated CSRs c. 0.3–0.5 m high among sediment trails reflecting the cross-cutting pattern of CSRs and flutings at the foot of Fannytoppen. (b) Ice-cored hummocky moraine in this area. In the background, hummocky moraine and trimline on the slopes of Fugleberget. (c) Hummocky moraine (h) and trimline (t) on the Fannytoppen slopes. The arrow marks the lateral terminal moraine and the rectangle illustrates the location of the roches moutonnées with CSRs on Oseanograftangen. (a–c) Photos A. Osika.

Eskers: Low, locally eroded sinuous eskers are visible in the eastern forefield on the till plain with fluvial overprinting. The ridges are 70–90 m long, 1–3 m wide, up to 0.5 m high and dissected by meltwater channels. They can be recognized by their surface of compact, sub-angular to rounded clasts ranging from fine gravels to 30 cm boulders with absent finer fractions. Based on their sinuous morphology, we interpret them as landforms developed in stagnant ice during the quiescent phase.

Hummocky moraine: Hummocky moraine occurs on the ice-proximal side of the lateral terminal moraine (Fig. 3). In the western part of the forefield, it takes the form of low, ice-cored, aligned hummocks, depressions and ponds, transitioning to the north into regular, ice-cored, low-amplitude linear ridges, interpreted as controlled moraine (Evans, 2009). The formation of hummocky and controlled moraines involves entrainment, transport and melt-out of debris on a stagnant glacier snout, leading to the accumulation of supraglacial debris cover (Boulton, 1968; Rachlewicz and Szczuciński, 2000; Schomacker and Kjær, 2008; Evans, 2009; Ewertowski and others, 2011). In controlled moraine, the supraglacial debris retains distinct linearity reflecting the inherited pattern of englacial debris concentrations (Evans, 2009). In the eastern forefield, hummocky moraine occurs as a complex of irregular, partly ice-cored, low-amplitude mounds and depressions at the proximal zone of the lateral terminal moraine (Fig. 8). The second assemblage is a complex of low, recently exposed ice-cored hummocky moraines adjacent to the slopes of Fannytoppen and Flatryggen. The surficial sediments are clast-rich, matrix-supported diamictons with large clasts of metamorphic and sedimentary rocks (Skolasińska and others, 2016). Abundant fragments of marine mollusc shells can be found up to the Flatpasset saddle between Fannytoppen and Flatryggen at c. 150 m a.s.l. (Fig. 2). The development of hummocky moraines requires entrainment and transport of large volumes of debris, which is particularly effective and widespread during the active phase of surging (Evans and Rea, 1999; Lovell and others, 2015a). Therefore, they are an important component of the surging glacier landsystems, although they are not independently diagnostic of surge-type behaviour (Evans and Rea, 2003).

Lateral terminal moraine: The outer, most prominent moraine belts are interpreted as lateral terminal moraines that extend underwater as a frontal terminal moraine (Fig. 3). In the western forefield, the lateral terminal moraine occurs as a conspicuous controlled moraine belt with sub-parallel units reflecting the former englacial structures in the glacier snout (cf. Evans, 2009). The outer moraine belt is 1130 m long, up to 200 m wide, 30 m high and ice-cored as evidenced by electrical resistivity tomography (Glazer and others, 2020), but the internal structure is unknown. The transverse profile is asymmetric with a steeper distal slope. In the south-eastern forefield, the lateral terminal moraine takes the form of two arcuate ridges 570 and 580 m long, 30–50 m wide and up to 8 m high, dissected by proglacial channels. The ridges are probably ice-cored based on the evidence of ongoing de-icing processes. The relief of their southern margins is subdued and resembles the adjacent hummocky moraine. In turn, the lateral terminal moraine on Flatpasset (Fig. 8c) occurs as a distinct, arcuate ridge 420 m long, up to 30 m wide, 6 m high and probably ice-cored. The northern part merges with ice-cored hummocky moraine, whereas the south-western part is associated with a trimline on the adjacent Fannytoppen slope. The lateral terminal moraines are composed of clast-rich diamicton with sub-angular to sub-rounded, and occasionally angular large clasts of metamorphic and sedimentary rocks. We interpret them as the outer and oldest ridges of controlled moraines, formed by debris entrainment during glacier advance and supraglacial melt-out of sediments from stagnating ice, preserving the ice core from degradation (cf. Evans, 2009).

Trimline: Trimlines can be traced on the eastern and south-eastern slopes of Fugleberget at 174–70 m a.s.l. and the western slopes of Fannytoppen at 173–32 m a.s.l. (Figs 3 and 8). We interpret them as the maximum thickness of Hansbreen during the surge (cf. Yde and others, 2019). The imprint is discontinuous due to paraglacial reworking and formation of talus cones. A gradual elevation decline corresponds to widening and lateral spread-out of the glacier tongue. Well-developed trimlines from the LIA are typical for Svalbard glaciers and their formation and distribution are primarily controlled by bedrock erodibility (Rootes and Clark, 2022). According to their study, erosional trimlines are slightly more common among surging and tidewater glaciers in Svalbard compared to land-terminating glaciers or glaciers with no known surge history.

Submarine forefield

Drumlinized seafloor and streamlined glacial lineations: The drumlinized seafloor comprises c. 750 m wide zone in the proximal part of the submarine forefield, taking the form of streamlined topography with drumlinoids, low-amplitude streamlined ridges and grooves (Figs 2 and 3). In the central zone, an oval, symmetrical ridge 260 m long and 90 m wide, interpreted as a drumlin, can be observed between a similar ridge with an indistinct appearance and a prominent drumlinoid 190 m wide. In the western part, an assemblage of distinct, flow-parallel, low-amplitude ridges and grooves up to 430 m long can be observed on the slopes of an overdeepening that extends to a depth of 94 m. The ridges are up to 5 m high, 30 m wide and the distance between their crests is usually 40–60 m. We interpret these as streamlined glacial lineations, which indicate high velocities during glacier advance and are a key characteristic of the submarine forefields of tidewater surge-type glaciers in Svalbard (e.g. Flink and others, 2015, 2018; Dowdeswell and Ottesen, 2016; Ottesen and others, 2017; Streuff and others, 2018). Based on their morphology and elongation ratios, the ridges represent a continuum between elongated drumlins and mega-scale glacial lineation, indicating subglacial streamlining (MSGL; Stokes and Clark, 2002; Spagnolo and others, 2014).

Esker: A single, sinusoidal ridge in the inner part of the forefield, which can be traced to an R channel in the nearby land-terminating part of the glacier is interpreted as an esker, resulting from the infilling of the subglacial conduit with sediments during decreasing meltwater discharge (Ottesen and others, 2008). The ridge is 390 m long, up to 10 m wide, 2 m high and faces east for the first 185 m before turning south. The distinct morphology of the first segments contrasts with blurred distal part c. 90 m long ending at the recessional moraine. Eskers have been documented in the forefields of many tidewater glaciers in Svalbard and, in the case of surge-type glaciers, inferred to have developed after surge termination (Ottesen and others, 2008; Dowdeswell and Ottesen, 2016; Flink and others, 2018; Noormets and others, 2021).

Small moraine ridges – CSRs and / or De Geer moraines: Small ridges oriented subparallel to the calving front and overprinting the flow-parallel ridges of streamlined bedforms can be traced across the inner part of Hansbukta. The transverse, locally branching and cross-cutting ridges approach 10–20 m in width and 3 m in height with spacing of 15–50 m. Their spatial pattern and abundance compared with the number of winter glacier advances (Błaszczyk and others, 2021) and the proximity of CSRs in the terrestrial forefield (Fig. 3) imply that the ridges are probably submarine CSRs formed during a surge (e.g. Ottesen and others, 2008; Dowdeswell and Ottesen, 2016). Based on a similar appearance, some ridges could be De Geer moraines, also known as small annual retreat moraines or push moraines (e.g. Streuff and others, 2015; Flink and Noormets, 2018). The formation of De Geer moraines involves the pushing of sediments during seasonal winter advances (Boulton, 1986; Ottesen and Dowdeswell, 2006). The landforms are widespread in the proximal but not visible in the central or outer parts of the forefield, possibly because they have been buried by sediments from meltwater plumes.

Recessional moraines: Several distinct transverse ridges, larger than those described above, can be observed primarily in the central and outer parts of the forefield. The ridges are oriented sub-parallel to the calving front, up to 10–15 m high and partly buried in glacimarine sediments (Ćwiąkała and others, 2018). In the proximal forefield, a chain of crescent ridges c. 5 m high corresponds to several winter advances to a similar position in 2013–2015. The most conspicuous ridges correspond with glacier advances in 1957–1959, 1973–1977 and during the 1990s (Jania, 1998; Ćwiąkała and others, 2018; Błaszczyk and others, 2021). We interpret the ridges as recessional moraines associated with minor readvances interrupting the overall glacier retreat (Robinson and Dowdeswell, 2011; Flink and others, 2015; Streuff and others, 2018).

Terminal moraines: Two terminal moraines can be observed in the submarine forefield of Hansbreen, each of them associated with a debris flow lobe. The outermost terminal moraine takes the form of a large, crescent, transverse ridge beyond Hansbukta, emerging onshore as the ice-cored lateral terminal moraines and marking the LIA maximum of Hansbreen (Figs 2 and 3). The total length of the LIA terminal moraine approaches 5100 m with a submarine part of c. 3300 m. The arcuate form indicates the radial splaying of the glacier beyond the constraining mountain ridges. The indentation in the distal flank is related to the abrupt break in the slope into a deeper fjord basin, which has prevented the full development of the moraine (cf. Dowdeswell and others, 2020). The outermost terminal moraine of Hansbreen was described in detail by Ćwiąkała and others (2018). The submarine ridge is plain-topped, c. 13 m high, with clear margins in the central, but extensively blurred in the lateral parts and located in a water depth of 10–25 m. Rock outcrops, fields of megaripples, iceberg pits and ploughmarks can be identified on the moraine surface (Ćwiąkała and others, 2018). On the proximal flank, four oval, flat-floored depressions 28–55 m in diameter and 2 m deep are interpreted as kettle holes. Near Oseanograftangen, the moraine surface is eroded by a channel-like feature interpreted as a glacifluvial incision, 2–7 m deep, 160–250 m wide, and with blurred margins reworked by sediment flows (Figs 2 and 3). In Isbjørnhamna, the outermost terminal moraine is subdued and overprinted with a debris flow lobe associated with an inner terminal moraine.

The inner terminal moraine occurs as transverse ridge segments up to 3 m high, semi-parallel to the outer terminal moraine and corresponding to the glacier extent in 1938 (Figs 2 and 3). We imply that terminal moraines have been formed by surges based on their resemblance to surge terminal moraines of Svalbard tidewater glaciers, in particular the association with well-developed debris flow lobes (e.g. Flink and others, 2015; Burton and others, 2016; Noormets and others, 2021). The morphology of the inner terminal moraine of Hansbreen is similar to the moraines of Tunabreen or Blomstrandbreen, corresponding to surges in the 20th century (Flink and others, 2015; Burton and others, 2016). In each site, the terminal moraines of subsequent surges are significantly smaller than the LIA terminal moraine and often discontinuous across the fjord. The formation of submarine terminal moraines involves thrusting and pushing of sediments during glacier advances as well as depositional processes during glacier stillstand, which may be long-lasting on a sill (e.g. Benn and Evans, 2010; Ottesen and others, 2017; Lovell and others, 2018).

Debris flow lobes: We identified two glacigenic debris flow lobes in the forefield of Hansbreen. The outermost lobe overprints the distal flank of the large terminal moraine, the adjacent flat-floored area of Isbjørnhamna and a steep northern side of the Hornsund Fjord beyond the outermost terminal moraine (Figs 2 and 3). In Isbjørnhamna, the lobe is flat-surfaced with clear outer margins, several rock outcrops and small, circular depressions, which are probably iceberg pits. Towards the break in slope into a deeper part of Hornsund, the lobe surface becomes hummocky with abundant iceberg pits, ploughmarks and minor sediment slides and flows. Similar topography can be observed beyond the central and eastern parts of the outermost terminal moraine with several prominent downslope-oriented flow structures. The ice-proximal margins of the moraine are clear, apart from the area adjacent to the moraine incision affected by mass-wasting processes. The second debris flow lobe is associated with the inner terminal moraine and superimposed on the top of the outermost, large terminal moraine in Isbjørnhamna. The landform has distinct margins and flow direction to the south-west (Figs 2 and 3). According to published backscattering data, the sediment flow lobes are composed of fine-grained or mixed material contrasting with coarse-grained surficial sediments of the terminal moraines (Ćwiąkała and others, 2018). Glacigenic debris flow lobes result from the bulldozing or pushing of seafloor sediments in front of the advancing terminus during surging and subsequent slope failure on the distal side of the glaciotectonic moraine (Dowdeswell and Ottesen, 2016; Lovell and others, 2018). The steep topography beyond the outermost terminal moraine has presumably controlled the development of the associated debris flow lobe, as they usually take the form of extensive gently sloped lobes in flat-floored areas (e.g. Dowdeswell and others, 2016; Flink and others, 2018; Streuff and others, 2018).

Smooth glacigenic seafloor: Smooth seafloor with flat-floored depressions separated by transverse moraine ridges comprises a major part of the submarine forefield (Figs 2 and 3). The maximum depth of depressions increases towards the current glacier front. We interpret these as glacially overdeepened basins filled with sediments from meltwater plumes, previously described by Tegowski and others (2016) and Ćwiąkała and others (2018). The basin-fill sediments partly bury the moraine ridges from minor readvances of Hansbreen. The sediment thickness ranges from 10 m in the central to 25 m in the southernmost part of the submarine forefield and the estimated mean sediment accumulation rate in particular basins ranged between 0.25 and 0.4 m a⁻¹ (Ćwiąkała and others, 2018). In the topography of the smooth seafloor, Ćwiąkała and others (2018) documented fields of megaripples, pockmarks, iceberg pits and ploughmarks.

Pockmarks: Pockmarks occur in the southernmost part of the forefield as small, circular depressions with diameters and depths approaching 15 and 1 m, respectively. The distribution of pockmarks in the basin adjacent to the frontal terminal moraine was documented by Ćwiąkała and others (2018), who suggested their possible origin as gas migration along faults or seepage of pore water due to very high sedimentation rate. Widespread craters and mounds related to methane release develop in the Arctic seabed after deglaciation (Andreassen and others, 2017). Apart from geological factors (e.g. Weniger and others, 2019), pockmarks in Svalbard fjords may form as a result of water escape from rapidly deposited glacigenic debris flow lobes related to surging (Ottesen and others, 2008; Forwick and others, 2009). Pockmarks in Hansbukta may be of similar origin considering the high sediment accumulation rate in the basins (cf. smooth glacigenic seafloor; Ćwiąkała and others, 2018). In addition, we infer that their formation may be associated with the burying of organic-rich sediments during glacier advance.

Hansbreen on archival maps and photographs (1872–1936)

Historical maps

The first published sketch map of the Hansbreen snout (scale 1: 200 000) was prepared in 1872 during the Austro-Hungarian North Pole Expedition (Peterman, 1874). The glacier was also presented on a 1: 200 000 map from the Russian–Swedish expedition in 1899–1901 (Wassiliew, 1925). The frontal position can be traced on a 1: 100 000 map based on reconnaissance photogrammetric surveys of the Hornsund area in 1918 during the Norwegian expeditions in 1917–1921 (Fig. 9a; Hoel, 1929). In these maps, Hansbreen occupies Hansbukta and the north-eastern part of Isbjørnhamna, covering the Baranowski Peninsula and Oseanograftangen. However, the maps are not detailed enough to accurately assess the extent and possible fluctuations of the glacier. Hence, our study is primarily based on archival photographs documenting the extent and surface morphology of Hansbreen.

Figure 9. (a) Hansbreen on a 1: 100 000 map ‘Spitsbergens Kyst, Hornsund til Dunder Bay’ based on the Norwegian Spitsbergen expeditions in 1917–1921 with photogrammetric measurements in Hornsund in 1918 (public domain, https://data.npolar.no). (b–g) Terrestrial photogrammetric documentation of Hansbreen from 1918. (b) The radially expanded glacier snout after surge advance. Note a folded medial moraine in the eastern part of the glacier. (c–f) A complex network of narrow, cross-cutting crevasses and crevasse traces documented from Fannytoppen. (g) The glacier front in Isbjørnhamna. Note a waterfall from an active supraglacial channel in the background. (b) © Olaf Holtedahl, Norsk Polarinstitutt (modified), (c–f) © Jørgen Gløersen, Norsk Polarinstitutt (modified), (g) © Adolf Hoel, Norsk Polarinstitutt (cropped and modified).

Archival photographs

Unique photographic documentation of Hansbreen from the late 19th and early 20th centuries was collected by the Austro-Hungarian expedition in 1872 and the Norwegian expedition in 1918. In 1936, Hansbreen was documented on oblique aerial photographs during the Norwegian mapping campaign of Svalbard (Luncke, 1936). The photographs provide data on the glacier extent, surface morphology, ice structure in the ice cliff and geomorphological features, which shed light on the glacier dynamics. In addition, they are a vital source of data on geomorphological processes.

1872: Photographs from the Austro-Hungarian expedition in 1872 documented the terrestrial margins of Hansbreen and part of the calving front. In the western terrestrial zone, the glacial terminus had already reached its LIA maximum position (Fig. 10a). The marine-terminating zone was densely and chaotically crevassed with a conspicuous surface bulge (Fig. 10b). At the foot of Fannytoppen, the terrestrial margin of Hansbreen terminated as a steep ice cliff, exposing inclined and folded debris layers. The photograph also documented top-down and bottom-up crevasses and a convex surface profile (Figs 10c and d). In this part of the glacier, the terminus had not yet reached the LIA maximum, today marked by the lateral terminal moraine adjacent to rocky raised marine terraces. The proglacial outwash plains mapped outside the terminal moraine were also absent in this photograph (Figs 3 and 10c).

Figure 10. Photographs of Hansbreen from the Austro-Hungarian expedition in 1872. (a) The western terrestrial margin of the glacier. (b) Crevassed marine-terminating zone with a conspicuous surface bulge. (c, d) The eastern terrestrial margin, where the glacier had not yet reached the maximum extent. Note folded and inclined debris layers, top-down and bottom-up crevasses exposed on the ice cliff and a convex surface profile of the glacier. © ÖNB Vienna: Image ID 00299043, 00480587 and 00444839 (cropped and modified).

1918: The Norwegian Spitsbergen expedition in 1918, led by Adolf Hoel and Sverre Røvig, conducted reconnaissance photogrammetric measurements in Hornsund. The comprehensive documentation, including photogrammetric photos, overlays the entire glacier. At that time, the glacier front was anchored on the small Hansholmane islands and Oseanograftangen (Figs 9a and b). The extent of the terrestrial margins was similar to that of 1872, but the calving front positions are difficult to compare due to insufficient data from the previous photographs. Glacier-free unnamed islands within the terminal moraine belt suggest a slight retreat from the maximum LIA position. A photograph from the southern side of Hornsund revealed a folded median moraine between Hansbreen and one of the eastern tributaries (Fig. 9b). The two black features on the calving front can be observed in other photographs and could be interpreted as the mouths of subglacial channels. The photos from Fannytoppen documented a dense, complex network of intersecting crevasses and crevasse traces extending high up on the main trunk of the glacier (Figs 9c–f). Most crevasses were narrow, which allowed the expedition to cross the glacier (Hoel, 1929). The photographs also documented a steep ice cliff in Isbjørnhamna with supra- and englacial sediments and a waterfall at the ice front (Fig. 9g). Compared to 1872, this part of the glacier had a smooth topography in 1918.

1936: By 1936, the glacier had retreated beyond Hansholmane, but the terminus was still anchored on Oseanograftangen and the Baranowski Peninsula (Fig. 11). The glacier front was relatively straight and slightly expanded in the central zone. The downwasting of the stagnant terrestrial margins had led to the development of the ice-cored moraine complexes. The lateral moraines of the eastern tributary glaciers had merged into a wide belt along the glacier margin. As a result, the folded moraine of the 1918 photo was no longer visible as it was probably squeezed to the eastern lateral moraine. The surface morphology was dominated by crevasse traces and wide transverse crevasses in the frontal zone. Some crevasse traces reflected the intersecting pattern of crevasses documented in 1918.

Figure 11. Hansbreen in the oblique aerial photograph from 1936. © Norwegian Polar Institute.

Discussion

Geomorphological records of surging

The forefield of Hansbreen with CSRs, fluted till plains, hummocky moraines, high trimlines, drumlinized seafloor, glacial lineations, De Geer moraines, terminal moraines with debris flow lobes and other components (Fig. 3) indicates at least one surge event. Such landform assemblages are consistent with surging glacier landsystem models in terrestrial and submarine environments (Evans and Rea, 1999, 2003; Ottesen and Dowdeswell, 2006; Ottesen and others, 2008, 2017; Flink and others, 2015; Aradóttir and others, 2019).

The key features diagnostic for the surges of Hansbreen are CSRs, glacial lineations and two glacitectonic terminal moraines with debris flow lobes. Terrestrial CSRs and flutings occur up to the southernmost part of field (Figs 2–8). In Svalbard, CSRs networks emerging from stagnating ice may be related to surge episodes from years to over a century ago (Lovell and others, 2015a; Farnsworth and others, 2016). CSRs and flutings in the terrestrial forefield of Hansbreen correspond with the submarine features. The rapid ice flow during the surge is evidenced by the streamlined glacial lineations (Figs 2 and 3; Ottesen and others, 2008, 2017; Streuff and others, 2018), although their absence would not preclude a surge event (Aradóttir and others, 2019). The submarine CSRs and glacial lineations appear only in the inner part of Hansbukta, but they might have been more widespread in the past and buried later with glacimarine sediments. A restricted extent or absence of CSRs on the fjord floor, but occurrence on land, were evidenced in the forefield of surge-type Blomstrandbreen (Burton and others, 2016; Farnsworth and others, 2016) and the glacial system of Trygghamna (Aradóttir and others, 2019). Hansbreen has not developed a terrestrial thrust-block moraine diagnostic of surging, but such landforms are not widespread in the lateral forefields of surge-type tidewater glaciers in Svalbard (Lønne, 2016; Lovell and Boston, 2017). The slightly curvilinear ridges of the lateral terminal moraines and the submarine large terminal moraine indicate the radial expansion of the glacier snout when entering the Hornsund Fjord (Figs 2, 3, 9). The occurrence of terrestrial CSRs up to the southernmost part of the forefield (Figs 3 and 6), the dimensions of the submarine terminal moraine and the occurrence of a debris flow lobe on its distal flank imply that the maximum LIA extent of Hansbreen was related to a surge. The lateral terminal moraine on the Flatpasset saddle (Figs 2 and 3) marks the glacier's attempt to pass the mountain ridge during this surge advance when the glacier entrained marine sediments with mollusc shells and deposited with diamicton up to 150 m a.s.l.

The inner terminal moraine with a debris flow lobe is the main evidence of a possible second surge. The lobe has overprinted the large terminal moraine and the flow direction of sediment was to the south-west, perpendicular to the inner moraine (Figs 2 and 3). The inner moraine corresponds to the frontal position of Hansbreen in 1938, which was similar or, in some parts, slightly more extensive than in 1936 (Fig. 2). The dimensions of the moraine and the sediment lobe are notably smaller than those from the LIA maximum. However, in Svalbard, surge terminal moraines from the 20th century tend to be less prominent than surge moraines from the LIA and may have similar morphology to recessional moraines (Burton and others, 2016). Surge end moraines of Tunabreen from 1930, 1971 and 2004 are much lower than the outermost moraine belt from the 1870s surge and only two surge moraines have adjacent debris flow lobes (Flink and others, 2015). The younger debris flow lobe of Hansbreen has similar dimensions to debris flow lobes on the outermost surge moraines of Blomstrandbreen (Streuff and others, 2015, 2022). This glacier has not developed debris flow lobes at all during the subsequent surges in the 20th century (Burton and others, 2016). The formation of the inner surge moraine of Hansbreen may have been limited by the scarce supply of sediment from resistant metamorphic rocks in the glacier basin that had been redeposited during the older surge. Low surge end moraines with small or without debris flow lobes may also reflect a short-active surge phase, which limits the sediment supply to the glacier terminus (Flink and others, 2015). In addition, most of the moraine ridges in the central and southern Hansbukta have been partly buried with glacimarine sediments, which mask their primary size. If the inner terminal moraine had a continuation across Hansbukta, it also might have been buried and subdued.

Historical records of surging

1872 – ongoing surge: The photographs of Hansbreen from the Austro-Hungarian expedition in 1872 reveal characteristic features of an active surge phase (Fig. 10). The glacier surface was densely crevassed with a conspicuous bulge, which we interpret as a surge front propagating downglacier (e.g. Sund and others, 2009; Lovell and Fleming, 2023). A surge bulge reflects the mass displacement from the reservoir area and builds up in a high friction zone near the glacier front (Haga and others, 2020). The ice thickness at the south-western margin corresponded to the trimline on edge between the eastern and southern slopes of Fugleberget (Fig. 2). Most of the southern slope is overprinted with talus cones and it was not possible to identify a clear trimline over there (Figs 2, 3, 10a). The vertical ice extent in 1872 was not documented on the other mountain slopes. The south-eastern margin of Hansbreen had a convex surface profile and a steep ice cliff with debris-rich structures (Figs 10c, d). Similar features are typical for the termini of surging glaciers (e.g. Lovell and others, 2015a; Ingólfsson and others, 2016; Sobota and others, 2016; Lovell and Fleming, 2023) and we suggest they fed the subsequent controlled moraines in this area (Fig. 3). The glacier extent in this area was smaller than marked by the LIA lateral terminal moraine. Hence, we imply that the surge was underway and the glacier was advancing to the maximum LIA position.

1918 – quiescent phase: Photographic documentation from the Norwegian Spitsbergen expedition in 1918 provides evidence for a glacier in the quiescent phase. The radially expanded snout was anchored on capes and small islands, and the glacier extent corresponded to the outermost submarine terminal moraine and the lateral terminal moraines. The ice thickness was similar to the trimlines mapped on the Fugleberget slopes and the southern slope of Fannytoppen (Figs 3 and 9). The frontal position did not change significantly between 1872 and 1918 (Koryakin, 1974; Jania, 1988). We suggest it was linked to anchoring the terminus on the terminal moraine and skerries with the prevailing thinning of the glacier tongue by the surface ablation rather than limited frontal ablation in shallow water. The water depth on the terminal moraine ranges from 2 to 15 m and does not exceed 25 m on the ice-proximal side of the belt (Fig. 2). The southernmost extent of the calving front was on the Hansholmane islands, which were probably one of the pinning points within the moraine belt (Figs 1, 2, 3, 9a and b). The mouths of subglacial channels could be observed on the calving front (Figs 9a and b) and the location of one of them seemed to correspond with an incision in the submarine terminal moraine (Figs 2 and 3), which suggests possible glacifluvial origin of this feature. The surface morphology of Hansbreen revealed compelling evidence for a surge before 1918, such as a folded medial moraine and a dense, complex network of intersecting crevasses and crevasse traces extending high up on the main trunk of the glacier (Fig. 9). We infer that abundant healed crevasses and their cross-cutting pattern were associated with rapid ice flow while surging in 1872. In 1918, the glacier terminus in Isbjørnhamna had a smooth topography and an active supraglacial channel with a waterfall at the ice cliff, which suggests closed crevasses in this part of the glacier during the quiescent phase (Fig. 9g). Similar spectacular features can be observed on the front of Bråsvellbreen (https://toposvalbard.npolar.no/) after decades of the quiescent phase in the aftermath of its surge in 1938 when the glacier tongue was badly crevassed (Schytt, 1969).

Late 1930s – limited evidence for a subsequent surge: In the western part of the forefield, the inner terminal moraine with a debris flow lobe suggests a younger surge of Hansbreen, which might have occurred in the first half of the 20th century (Fig. 2). Oblique aerial photographs from 1936 do not indicate a fully active surge (stage 3 according to Sund and others, 2009) as the surface morphology was dominated by crevasse traces, wide transverse crevasses were confined to the frontal zone, the terrestrial margins were stagnant, and the calving front was relatively straight and slightly expanded only in the central zone (Fig. 11). These data alone do not allow a qualitative assessment of the upper part of the glacier for evidence for the first two stages of the active surge phase expressed by changes in the glacier thickness (Sund and others, 2009). In August 1938, Pillewizer (1939) measured the unusually high surface velocity of Hansbreen using the terrestrial photogrammetric method of Finsterwalder (1931), which could suggest an active surge phase, and his map indicates an advance in some parts of the glacier front compared to 1936 (Fig. 2). If such event occurred, it may have been related to the inner terminal moraine with a debris flow lobe. However, additional evidence is necessary to support this interpretation in its current state.

Reconciling glacial geomorphology and past dynamics of Hansbreen

The possible surge-type behaviour of Hansbreen has long been debated for the lack of evidence of an ongoing surge in the observation period (e.g. Jania, 1988; Jania and Głowacki, 1996; Rachlewicz and Szczuciński, 2000; Ćwiąkała and others, 2018). We present the records of at least one surge event. Photographic documentation of the glacier extent, surface morphology and ice structure in the ice cliff revealed characteristic features for the active surge phase in 1872 (Fig. 10) and quiescence in 1918 (Fig. 9). This advance was marked by the outermost terminal moraine, trimlines and the lateral terminal moraines (Figs 2 and 3). Sparse historical data may suggest a minor surge episode in the 1930s, but further investigation is necessary to confirm this event. The imprints of the possible younger surge are much weaker than the LIA surge and the only compelling geomorphological evidence is the inner terminal moraine with a debris flow lobe (Fig. 3).

Our interpretation differs from previous investigations, which did not conclude that surges had occurred. Ćwiąkała and others (2018) analysed the submarine forefield and excluded the possibility of surging in the last c. 120 years based on the absence of streamlined glacial lineations, overridden recessional moraines, CSRs or a surge moraine with a debris flow lobe. However, drumlinized seafloor, streamlined glacial lineations and CSRs are not visible in the multibeam data collected in 2008, which they used for geomorphological mapping. Such landforms occur in the inner zone of Hansbukta surveyed in 2014–2017 (Fig. 3). CSRs were also not identified in the terrestrial forefield of Hansbreen (Skolasińska and others, 2016 and references therein), but part of the debris-rich structures in the passive ice cliff described by Rachlewicz and Szczuciński (2000) resembled CSRs at the base of surge-type glaciers (e.g. Evans and Rea, 2003; Lovell and others, 2015a). Apart from the north-western forefield, most CSRs are subdued and take the form of low ridges or sediment trails (Figs 6 and 8), similar to those described in Trygghamna (Aradóttir and others, 2019; Ben-Yehoshua and others, 2023). CSRs in the Hansbreen forefield are less abundant, lower and worse preserved than in many surge-type glaciers in Svalbard, e.g. Nathorstbreen (Lovell and others, 2018), Sefströmbreen (Boulton and others, 1996), Tunabreen, Von Postbreen (Lovell and others, 2015a) or the Hornbreen-Hambergbreen glacier system (Osika and others, 2022). Therefore, previous investigations could have omitted them without high-resolution remote sensing data.

Our study shows the importance of compiling terrestrial and submarine geomorphology with historical records for reconstructing past glacier dynamics. However, if the archival data are sparse, the identification of surge-type glaciers is impeded by the low preservation potential of surge-diagnostic landforms or topographic or geological conditions meaning that they do not form in the first place (e.g. Hansen, 2003; Farnsworth and others, 2016; Ingólfsson and others, 2016; Gądek and others, 2022). Therefore, we infer that the number of Svalbard surge-type glaciers or decaying glaciers that surged in the past can be underestimated if the geomorphological fingerprints have been obliterated and historical records are lacking or unrecognized. One example may be a cold-based valley Tellbreen, which was thought not to have ever surged (Bælum and Benn, 2011). However, the basal sequence and glaciological structures revealed that it had experienced warm-based conditions and dynamic ice flow, most probably during the LIA (Lovell and others, 2015b).

Interestingly, the 1870s surge of Hansbreen occurred in a period of numerous surge events in Svalbard between 1860s–1900s (e.g. Fridtjovbreen in 1861; Liestøl, 1993, Kronebreen, Kongsvegen and Kongsbreen in 1869; Liestøl, 1988, Von Postbreen in 1870; Liestøl, 1993, Nathorstbreen in the 1870s or 1880s; Liestøl, 1977; Ottesen and others, 2008, Sefströmbreen in 1882–1886; De Geer, 1910, Recherchebreen and Renardbreen at the turn of the 1870s and 1880s; Zagórski and others, 2023, Scottbreen in 1880s; Zagórski and others, 2023, Hambergbreen c. 1900; Wassiliew, 1925 or Paulabreen before 1898; Ottesen and others, 2008). The 1930s was another decade of apparent surge clustering (e.g. Bråsvellbreen in 1936–1938; Schytt, 1969, Negribreen in 1935–1936; Liestøl, 1969, Körberbreen in 1938; Liestøl, 1969, Markhambreen in 1930–1936; Hagen and others, 1993, Arnesenbreen in 1925–1935; Hagen and others, 1993 or Etonbreen, Rijpbreen and Clasebreen in 1938; Hagen and others, 1993), although, despite some evidence, a surge of Hansbreen in this period remains questionable. Such synchronicity of surge advances might suggest regional factors affecting glacier dynamics. The reconstructed winter surface air temperatures in Longyearbyen were higher in the 1860s and early 1870s than in the 1850s or late 1870s (Divine and others, 2011). The 1920s and 1930s corresponded to the early 20th-century climate warming in Svalbard (Nordli and others, 2020) and we speculate that these surges may have been related to increased water content at the glacier beds from enhanced summer melt and produced by frictional heating (cf. enthalpy balance theory by Benn and others, 2019). However, this observation requires further studies on the LIA climate conditions and investigation into triggering factors of glacier surges (e.g. Benn and others, 2023).

Repeated surge events of many Svalbard glaciers with different durations of the quiescent phase (e.g. Tunabreen; Flink and others, 2015, Blomstrandbreen; Burton and others, 2016, Negribreen; Haga and others, 2020) raise the question about a potential future surge of Hansbreen. In our opinion, the surface topography of Hansbreen, i.e. the low-lying accumulation area of the main trunk and significant ice supply by the western tributary glaciers, as well as the subglacial topography with overdeepenings along the centreline, will hinder the build-up of mass and triggering the next surge. A dramatic increase in solid precipitation will be critical to fill up the reservoir area(s) and allow the rapid transfer of large amounts of ice to the receiving zone. We can speculate that such conditions will not be fulfilled in the following decades under projected climatic conditions in Svalbard (Hanssen-Bauer and others, 2019). However, one should not exclude possible dynamic advances of the western tributary glaciers after retreat of the main trunk. The loose of the back-stress from the Hansbreen main tongue may lead to re-organization of the surface profiles of the tributary glaciers through re-advances (Farnsworth and others, 2017).

Conclusions

We present a reconstruction of the past dynamics of Hansbreen by combining geomorphological mapping of the terrestrial and submarine forefields with historical data from Spitsbergen expeditions in the 19th and 20th centuries. Landform assemblages, including CSRs, flutings, glacial lineations and terminal moraines with debris flow lobes are consistent with the published surging glacier landsystem models and indicate at least one surge of Hansbreen.

Photographic documentation from the Austro-Hungarian expedition in 1872 revealed evidence of an ongoing active surge phase, such as a surge bulge, extensive surface crevassing, top-down and bottom-up crevasses and a convex surface profile of the terrestrial margin. The terminus had not yet reached the surge maximum. This surge corresponded to the maximum LIA extent of Hansbreen, marked by the outermost submarine terminal moraine with a debris flow lobe, terrestrial lateral terminal moraines and trimlines.

In 1918, the glacier was in the quiescent phase and the calving front was anchored on small islands and capes. The photogrammetric photos from the Norwegian expedition revealed further evidence of the surge before 1918, including a folded medial moraine and a dense, complex network of intersecting healed crevasses extending high up on the main trunk of Hansbreen. There may have been a subsequent surge in the 1930s based on a minor submarine terminal moraine with a debris flow lobe but this potential event requires further investigation.

Under projected climatic conditions, the low-lying accumulation area of the main trunk will hinder the mass build-up, and a potential next surge will probably remain questionable in the following decades. However, surges of the western tributary glaciers should not be excluded, in particular as a dynamic response when they lose the back-stress from the receding main tongue.