maximum southerly extent between 19–17 ka. The collapse of all BIIS marine sectors was thought to have occurred by 17 ka (Clark et al., 2012). However, this chronology was largely based on three ages; an early thermoluminescence age from beneath weathered Skipsea till at Eppleworth of 17.5 1.5 ka (Wintle and Catt, 1985), a radiocarbon date from silts beneath till from the UK LGM typesite at Dimlington of 21.7 cal ka (Catt and Penny, 1966), which possibly has a hard water effect problem, and a luminescence age of 17 ka dating Lake Humber (Bateman et al., 2008), which had to have been impounded by the NorthSealobe. Bateman et al. (2011) reﬁned this with new luminescence ages, indicating ice advances at the Dimlington typesite within the periods 21.7–16.2 ka and 16.2–15.5 ka. Thus Livingstone et al. (2012) when producing a six-stage model of the central sector of the BIIS based on geomorphic mapping and stratigraphy showed ice persisting through to 16 ka, with oscillations occurring from their ‘‘stage III’’ (20 ka) onwards. The ice source for the NorthSealobe, as proposed by Busﬁeld et al. (2015) based on erratic lithologies, was the Midland Valley in Scotland. Ice ﬂowing eastward from the upland dispersal centres of northern Britain appears to have been deﬂected southward and this may have been due to the presence of ice in the central and northern NorthSea. Part of the BIIS has been shown to have advanced southeast in the Witch Ground Basin around 17.5 cal ka BP (Fladen 1; Sejrup et al., 2015). However the extent of this advance remains largely unknown and it would have had to have extended much further south to affect ice from the Midland Valley. Merritt et al. (2003) postulated a deﬂecting dome in the northern NorthSea. One of the scenarios proposed by Clark et al. (2012) was that an ice dome in the central NorthSea caused the NorthSeaicelobe to be deﬂected southwards, but evidence of this has yet to be found. Both Bateman et al. (2011) and Busﬁeld et al. (2015) observed that the ﬂow occurred in a topographical low where soft sediment and the Jurassic mudstone substrate probably enhanced basal sliding (presumably also deforma- tion) between the north-south striking outcrop of Cretaceous Chalk along the western NorthSea margins and the southwest-northeast trending topographic high of Dogger Bank. The topography of the latter may have been enhanced by the FSIS forebulge which uplifted the centre of the NorthSea basin by 10–20 m (Busschers et al., 2007). Whilst this would explain the southerly ﬂow direction of the icelobe it does little to explain the mechanism for the westerly ice oscillations reported on the County Durham and Yorkshire coastlines (e.g. Davies et al., 2009; Evans and Thomson, 2010).
Beyond the MIS 2 limit in both the lowland area of The Wash/ Fenland and the North Norfolk coast, the existence of MIS 10, 8 or 6 (i.e.i.e., “ Wolstonian ” ) glacial limits is a subject of considerable and protracted debate. A chronostratigraphic framework in the Fenland area has facilitated Gibbard et als. (2018) proposal for an MIS 6 surging icelobe during the Tottenhill Glaciation(Fig.1a). Although an inset series of ice-marginal glacitectonic landforms and ice-dammed lakes appear to record oscillations of this NorthSeaLobe over the Fenland/Wash basin, the apparent close juxtaposition of Middle Pleistocene (possible post MIS 12) glacial features in North Norfolk (cf. Straw,1979; Hart,1990; Lee et al., 2013, 2017; Pawley et al., 2005, 2008) makes it dif ﬁ cult to decipher whether or not any similarly substantial MIS 6 glaciation signatures exist in that area. Rose (2009) has entertained the notion of MIS 6 glacial deposits in North Norfolk and tentatively equated the Cromer Ridge glacitectonic moraine and its associated outwash fans and esker with the Saalian Glaciation in mainland Europe, but acknowledges the lack of chronological constraints on this correlation. Oursampling and datingof sediment- landform assemblages located south of the River Stiffkey, previously classi ﬁ ed as “ Wolstonian ” by Straw (1979) and entertained as potentially of MIS 10, 8 or 6 age by Hamblin et al. (2005) and Rose (2009), indicate that they are indeed likely of MIS 6 age and record the modi ﬁ cation of local drainage by advancing glacier ice during the Tottenhill Glaciation. The juxtaposition of the MIS 2 and MIS 6 ice limits along this coast, in contrast to their greater separation over the Fenland/Wash basin gives some credence to Gibbard et als. (2018) notion of topographically-induced lobe surging in the region. The continuation of the MIS 6 limit northwards into Lincolnshire remains unknown but could be demarcated by the arcuate assemblage of glaci ﬂ uvial deposits and till mapped by the BGS between Mareham- le-Fen and Wildmore Fen, to the west of the Stickney Moraine. Importantly in this respect, the Kirkby Moor Sands are dated to MIS 6 and reach altitudes of ca. 30 m, the highest level of proglacial lakes proposed for the Fenland area during both MIS 6 and 2 due to the spillway control altitude of the River Waveney valley.
The dominance of angular to sub-angular clast forms within the gravels at East Heslerton indicates the mechanical breakdown (frost shattering) of cold climate conditions but also of short travel distances, the former being compatible with the development of intraformational ice wedges in LF2. They also reflect low-energy fluvial conditions and/or short travel distances and hence are not likely to have been delivered to the site after significant transport through and then along the margin of a glacier snout. This further supports the notion that the depocentre represented at East Heslerton is a scarp base fan fed by runoff from the Wolds. A fan interpretation, however, does not explain the Sherburn Sands ‘shelf’ located just below the altitude of the lower Glacial Lake Pickering shoreline (45 m OD), unless a series of channels through the escarpment were used by runoff to prograde sediments into the vale in a series of fans that coalesced over time and, at the last stages of sediment production, aggraded to a base level below the 45-m shoreline. If deposited at the margin of Glacial Lake Picker- ing, the Sherburn Sands were probably deposited in a scarp foot fan delta. The occurrence of a small outcrop of rhythmites in LF1a is probably the stratigraphic equivalent of the lake sediments previously reported from the former lake floor and hence the lake deposits appear to continue under the Sherburn Sands. Additionally, the more gravel-rich sedi- ments of LF2 interdigitate with sands in a distal-fining architecture at East Heslerton (Foster, 1987a). The paucity of lake deposits in areas covered by the Sherburn Sands can be explained by their location within the limits of the former Vale of Pickering icelobe, so that fan progradation only started once the ice margin began its recession eastwards. Moreover, the subaerial nature of LF2, as indicated by the fluvial bedforms and intraformational ice wedge development at East Heslerton, indicates that lake water levels had fallen to below 34 m OD by the time it was deposited. This must have taken place sometime after the 45-m shoreline phase of the Cayton–Speeton Stage (Penny and Rawson, 1969) and a lower lake stand is recorded by the Pickering delta at 30 m OD (Kendall, 1902), potentially recording a later incision level at the intake of the Kirkham Priory spillway. However, the 45-m lake stand must have impacted upon the Sherburn Sands because they extend up the Wolds to at least 60 m OD. Lake waters probably trimmed fans west of the Flamborough Moraine, East Heslerton being an example, rejuvenating them with new base levels. This enabled these fans to continue
To perform the scaling analysis of seaice deformation, we implemented a coarse graining method similar to the one pro- posed by Marsan et al. (2004) and applied it to the unfiltered and filtered versions of our RGPS Image Pair data set. Seaice shear and absolute divergence rates are computed at different spatial scales ranging from 7 to 700 km. For the lowest scale, which is also the scale of the triangular cells, all the cells are taken into account. For the other scales, the coarse graining procedure covers the domain with boxes of different sizes (14, 28, 56, 112, 224, 448 and 896 km). The boxes actually overlap since a distance equal to half the box size separates their respective centers. For each box, we select the cells that have their center in the box. When the sum of the area of those cells is greater than half the box area, the deformation over the box is defined by averaging the spatial derivatives of the displacement weighted by the surface of each cell. The spatial scale for this new estimate of the deformation is the square root of the sum of the cell areas. The shear and abso- lute divergence for each box are then reported as a function of the spatial scale on a log–log plot (see Fig. 10 for the abso- lute divergence rate). The mean value h ˙ L i (where ˙ L is either
effects on other experiments (e.g., Else et al., 2015). Ice sam- ples were collected using ceramic knives or a Kovacs Mark II coring system depending on the ice thickness. Sampling was performed from a movable bridge to avoid walking on the ice surface and to ensure only undisturbed sites were sampled. Ice cores were collected from one end of the pool (half me- ter away from the edge of the pool) and at least 20 cm away from previous cored sites. Ice cores were packed in clean plastic bags and kept frozen during the 20 min transport to a cold laboratory and processed within a few hours. Seawater was sampled for total alkalinity (TA) and total dissolved in- organic carbon (TCO 2 ) with a peristaltic pump (Cole Parmer
3.1.2 S2: response to strong regional runoff variations The runoff distribution used in S2 introduces regionally var- ied coastal surface freshwater fluxes. The responses of the seaice properties can therefore be expected to be strongly region-dependent. For the seaice concentration (Fig. 2g), we find changes of high statistical confidence in the coastal area. Increases in ice concentration and thickness (Fig. 2h) occur in the eastern Weddell Sea, in the western Ross Sea, close to the coast of East Antarctica, and east of the tip of the Antarc- tic Peninsula. Areas of strongly reduced seaice are located adjacent to the coast of the Amundsen and eastern Ross seas and in the southern Weddell Sea. Since in S2 the freshwater input is varied regionally along the coastline, the ice drift ve- locities are also altered depending on the location (Fig. 2i). Compared to CTR, the westward ice drift is faster along the coast of the Amundsen and Ross seas. From the Prydz Bay to the southern Weddell Sea it is slower than in CTR. From here, seaice speeds up compared to CTR, moving northward along the Antarctic Peninsula, to slow down again on the western side of the peninsula toward Bellingshausen Sea. In S2, the changes in seaice velocity cause most of the local changes in seaice concentration and thickness.
5 Uni Research Climate, Bjerknes Centre for Climate Research, Bergen, Norway
(Manuscript received 10 August 2017; in final form 22 January 2018)
A B S T R A C T
A data assimilation method capable of constraining the seaice of an Earth system model in a dynamically consistent manner has the potential to enhance the accuracy of climate reconstructions and predictions. Finding such a method is challenging because the seaice dynamics is highly non-linear, and seaice variables are strongly non-Gaussian distributed and tightly coupled to the rest of the Earth system – particularly thermodynamically with the ocean. We investigate key practical implementations for assimilating seaice concentration – the predominant source of observations in polar regions – with the Norwegian Climate Prediction Model that combines the Norwegian Earth System Model with the Ensemble Kalman Filter. The performances of the different configurations are investigated by conducting 10-year reanalyses in a perfect model framework. First, we find that with a flow-dependent assimilation method, strongly coupled ocean–seaice assimilation outperforms weakly coupled (seaice only) assimilation. An attempt to prescribe the covariance between the ocean temperature and the seaice concentration performed poorly. Extending the ocean updates below the mixed layer is slightly beneficial for the Arctic hydrography. Second, we find that solving the analysis for the multicategory instead of the aggregated ice state variables greatly reduces the errors in the ice state. Updating the ice volumes induces a weak drift in the bias for the thick ice category that relates to the postprocessing of unphysical thicknesses. Preserving the ice thicknesses for each category during the assimilation mitigates the drift without degrading the performance. The robustness and reliability of the optimal setting is demonstrated for a 20-year reanalysis. The error of seaice concentration reduces by 50% (65%), seaice thickness by 25% (35%), sea surface temperature by 33% (23%) and sea surface salinity by 11% (25%) in the Arctic (Antarctic) compared to a reference run without assimilation.
Abstract. A realistic representation of sea-ice deformation in models is important for accurate simulation of the sea-ice mass balance. Simulated sea-ice deformation from numerical simulations with 4.5, 9, and 18 km horizontal grid spacing and a viscous–plastic (VP) sea-ice rheology are compared with synthetic aperture radar (SAR) satellite observations (RGPS, RADARSAT Geophysical Processor System) for the time period 1996–2008. All three simulations can reproduce the large-scale ice deformation patterns, but small-scale sea- ice deformations and linear kinematic features (LKFs) are not adequately reproduced. The mean sea-ice total deforma- tion rate is about 40 % lower in all model solutions than in the satellite observations, especially in the seasonal sea-ice zone. A decrease in model grid spacing, however, produces a higher density and more localized ice deformation fea- tures. The 4.5 km simulation produces some linear kinematic features, but not with the right frequency. The dependence on length scale and probability density functions (PDFs) of absolute divergence and shear for all three model solutions show a power-law scaling behavior similar to RGPS obser- vations, contrary to what was found in some previous studies. Overall, the 4.5 km simulation produces the most realistic divergence, vorticity, and shear when compared with RGPS data. This study provides an evaluation of high and coarse- resolution viscous–plastic sea-ice simulations based on spa- tial distribution, time series, and power-law scaling metrics.
Although snow thickness and distribution are variable and primarily result from wind-induced redistribution dur- ing storms (Weeks, 2010), the impact of snow cover on the thermal evolution of seaice can be significant (Massom et al., 2001). Snow, which has a low thermal conductivity com- pared to seaice (Massom et al., 2001), provides thermal in- sulation between the cold air and the ice. The presence of a thick snow cover also affects the isostatic balance, poten- tially resulting in negative freeboard (i.e., the snow–ice in- terface is submerged below the seawater level). If the seaice is permeable throughout the entire ice column, negative free- board causes vertical flooding to at least the snow–ice in- terface through open brine channels. The percolation thresh- old above which columnar seaice is considered permeable to fluid transport corresponds to a brine volume (which is controlled by temperature and salinity) of 5 % (Golden et al., 1998, 2007).
Simulation, version 00.2 (CATS00.2) that we use here is an updated version of the CATS model used by Padman and Kottmeier  and Rignot et al. . The governing equations (depth-integrated momentum equations plus con- tinuity) are presented by Robertson et al. . The CATS00.2 domain includes the entire ocean south of 56S. For the seaice model region outside the CATS domain, between 52S and 56S, tides were set to zero. (As only a very small fraction of this area is ice covered in winter, the influence on the results is completely negligible.) The grid spacing is 1/4 1/12, which gives 10 km resolution near the Antarctic coast. For the ocean cavities under the ice shelves, water depth is replaced by ‘‘water column thickness,’’ i.e., the vertical distance between the base of the glacial ice and the seabed. The water column thickness values for the Filchner-Ronne Ice Shelf region are obtained from Johnson and Smith . For all other ice shelves in the Weddell Sea the water depth is taken unmodified from ETOPO-5 since very few glacier thickness data are available. The model incorporates depth data for the southwestern Weddell Sea that were acquired during the 1998 Ronne Polynya Experiment (ROPEX-98) [Nicholls et al., 2003].
184.108.40.206 Since 1991 a dedicated satellite image interpretation system has been in use operationally, and additional data from microwave equipped satellites (both active and passive) are being used. The ERS-1/2 SAR and RADARSAT have proven very efficient in the ice mapping, and hopefully the air observations may become obsolete. However, in the waters around Cape Farewell (southernmost Greenland), which are the most important from a navigational point of view, the mapping of seaice has proved extremely difficult. The sea is frequently rough due to stormy weather, and the ice consisting mainly of small floes in medium concentrations cannot generally be directly detected on the images due to the heavy backscatter from the surrounding water. As the thickness of the ice is several meters, the ice floes (and bergy bits) form a severe menace to navigation. In recent years, much emphasis has been placed on first-order statistical filtering of the images, whereby even moderate concentrations of small and medium ice floes may be detected. A number of simultaneous underflights has been carried out to compare the satellite-based SAR data with visual observations, video recordings and photographs from the reconnaissance aircraft.
Underway measurements of f CO sw 2 and sea surface temper- ature (SST) were obtained aboard two containerships MS Trans Carrier (operated by Seatrans AS, Norway www. seatrans.no) and MV Nuka Arctica (Royal Arctic Lines of Denmark). The route of MS Trans Carrier has changed over time and today the ship crosses the NorthSea along a tran- sect at roughly 5 ◦ E (Fig. 1). At the start of the project, the ship track had a triangular shape as the ship also called on the port of Immingham (UK) in addition to Bergen (Norway) and Amsterdam (The Netherlands). For the present analyses, we use data exclusively from the line connecting Norway and The Netherlands, since it has been the most persistent track. Note that even within this transect the ship track can change slightly, for example, due to weather conditions. Moreover, north of 58.5 ◦ N the ship frequently stopped at several small ports before Bergen. For consistency, data acquired within the geographic rectangle 53.2 ◦ N–58.5 ◦ N and 4.4 ◦ E–5.5 ◦ E
Linking movement data from satellite-tagged marine mammals with biological information on viral shedding illustrates that exposed animals have the potential to carry PDV long distances. Movement of PDV seropositive bearded seals and northern fur seals occurred within their species-specific predicted viral transmission distances and in close proximity to known locations of PCR positive individuals from all species tested. These data demon- strate the potential for animals exposed to PDV to carry the virus to areas with conspecifics and sympatric species. Data from satellite-tagged Steller sea lions suggest infectious animals could move over 100 km in one week, reach- ing nearby and distant rookeries (Fig. 4a). Bearded seals and northern fur seals have the potential to move over 200 km and 500 km in one week, respectively. Spotted seals and ribbon seals may bridge the gap between northern ice-associated seals and Steller sea lions and northern fur seals living in the southern Bering Sea (Fig. 4a). The ability to move long distances and timing of movements associated with life history cycles likely influence transmission patterns in the North Pacific Ocean and the potential to transmit PDV to species living in southern habitats.
In this section we show results from ECHAM5/MPI-OM when we disable ocean heat transport by reducing the ocean model to one level of 50 m thickness and enforcing zero ocean velocities. All simulations use the low bare sea-ice albedo of 0.45. Ocean velocities not only transport heat but also exert drag on seaice (see Sect. 3). All simulation there- fore use disabled sea-ice dynamics such that we truly mea- sure the effect of ocean heat transport alone. In essence, this variant of ECHAM5/MPI-OM is an atmospheric gen- eral circulation model coupled to a mixed-layer ocean with zero ocean heat transport but with exactly the same sea-ice thermodynamics as the full ECHAM5/MPI-OM model. The latter implies that it allows fractionally sea-ice covered grid boxes and tracks snow on seaice, which is not the case for the stand-alone ECHAM5 mixed-layer ocean setup used in Abbot et al. (2011).
Historically, NEMO-LIM3 seaice dynamic parameters were tuned in a heuristic, ‘‘trial-and-error’’ process, with the goal to best match the ice areal export through Fram Strait (M. Vancoppenolle, personal communication, 2013). While seaice extent and thickness are realistically simulated in the Arctic in the default conﬁguration [Vancoppenolle et al., 2009], the Arctic seaice drift simulated at the daily time scale exhibits clear biases. Figure 1a is a snapshot of observed seaice motion observed during a 2 day period in April 2012. Because NEMO- LIM3 is forced by atmospheric reanalyzes, it is expected to reproduce this motion. The direction of the ﬂow is simulated overall correctly, but its intensity clearly is not (Figure 1b). In addition, seaice is nearly motionless in a large portion of Beaufort Sea and the gradients of seaice velocity are clearly too smooth compared to obser- vations. At the model’s resolution, oceanic eddies are not explicitly represented and their effect is instead par- ameterized in terms of global state variables [Madec, 2008]. The fact that the modeled drift is globally too slow may be in part attributed to this issue, but this is probably not the only reason.
error estimates of the different reference data sets, described above. It would be interesting to compare the SAR derived data to ship observations. However, this turns out to be impractical as each SAR scene would result in only one match against a ship observation and present-day SAR coverage will give less than one match per day. Generally, ice charts, as produced by ice operators, are widely held as the best available data set of the ice edge, and they are regularly validated by mariners, who are critically depen- dent on a correct classification. It should be noted that the ice charting community have the most stringent require- ments to detail and accuracy and have commonly left airborne reconnaissance in favor of satellite SAR observa- tions and that passive microwave imagery is used only for strategic navigation requirements [Bertoia et al., 2004]. Still, some indication of the uncertainty in the SAR derived ice concentrations was obtained through classifying the same scene by two independent analysts [Bøvith and Andersen, 2005]. It was found that the classifications into seaice and water agreed within 2.1% between the two classifications. It is unclear whether this is a representative estimate of the error in the SAR classification method. In some scenes over high-concentration seaice, the ice con- centration variability is clearly less than the above 2.1% and the accuracy of the SAR classifications is therefore bound to be better. Interestingly, this means that we can put an upper limit to the errors in the SAR concentrations, by defining subsets with more or less stringent quality control.
While extensive work still needs to be performed on this new proxy, we can circumvent most of the aforementioned issues by combining HBI records with diatom assemblages. Whilst B. adeliensis is rarely preserved in marine sediment archives due to their thin silica shells and inherent vulnerability to water column and sediment dissolution, other diatom species, associated with seaice (e.g., Fragilariopsis curta and Fragilariopsis cylindrus) or seasonally open water/ice free conditions (e.g., Thalassiosira antarctica, Chaetoceros resting spore or Fragilariopsis kerguelensis), can be used to infer seaice conditions [88,111,118,121]. In Antarctica, the IPSO25 and IPSO25/triene ratio has been utilized in eight marine cores located in the western and eastern Antarctic Peninsula [111,118,121], Prydz Bay  and the Adélie Basin [88,100,104], East Antarctica. Recently, a modified IPSO 25 /triene and IPSO 25 /brassicasterol ratio (termed PIPSO 25 index in reference to the Arctic PIP 25 index ) was evaluated regarding its potential use as semi-quantitative seaice proxy . Most of the existing HBI records span the Holocene (i.e, the last 11kyrs BP), of these, four cover most of the last 2000 years at decadal or sub-centennial resolution [88,111,121]. Four records from the Vega Drift and Andvord Drift in the Antarctic Peninsula, the Dumont d’Urville Trough and Prydz Bay on the East Antarctic Margin cover the last centuries [100,118] and decades .
In most situations, the seaice covers only a part of the total water surface and is a mixture of ice types differing in structure and properties – level, rafted and ridged ice, possibly with cracks and leads (Lepp¨aranta and Myrberg, 2009). The numerical sea-ice models operating on the scale of tens up to thousands of kilometers reduce this informa- tion to a few parameters, typically to the concentration and mean thickness of the specified ice (and snow) classes in a given grid cell (for studies concerning the Baltic Sea see, e.g., Haapala and Lepp¨aranta, 1996; Meier et al., 1999; Haa- pala, 2000; Lehmann and Hinrichsen, 2000a; Zhang, 2000). As described below, the model used in this study belongs to that class of models. Modelling studies of the Baltic Seaice thermodynamics and dynamics can be broadly divided into two categories. One concentrates on problems of climate and climate change, seasonal and interannual sea-ice variability, and the influence of large-scale atmospheric circulation pat- terns on the sea-ice processes in the Baltic Sea (e.g., Haapala and Lepp¨aranta, 1996; Omstedt and Nyberg, 1996; Lehmann and Hinrichsen, 2000a,b; Schrum et al., 2003). These studies are based on medium- and long-term simulations and mostly involve variables such as the maximum annual ice extent or the length of the ice season, i.e., parameters which can be understood as proxies of winter severity and which are there- fore good indicators of a climate change. The other category consists of studies which tackle smaller spatial and tempo- ral scales with the aim of analyzing effects of synoptic-scale weather patterns on the thermodynamics and dynamics of the seaice. Good examples are provided by Uotila (2001), Br¨ummer et al. (2002), Rudolph and Lehmann (2006), Wang et al. (2006) or Bj¨ork et al. (2008). The need for more ex- tensive observational and numerical research is widely rec- ognized, particularly with respect to short- and medium-term ice dynamics in the Baltic Sea.
Abstract. This paper presents a first implementation of a new rheological model for seaice on geophysical scales. This continuum model, called Maxwell elasto-brittle (Maxwell- EB), is based on a Maxwell constitutive law, a progressive damage mechanism that is coupled to both the elastic mod- ulus and apparent viscosity of the ice cover and a Mohr– Coulomb damage criterion that allows for pure (uniaxial and biaxial) tensile strength. The model is tested on the basis of its capability to reproduce the complex mechanical and dy- namical behaviour of seaice drifting through a narrow pas- sage. Idealized as well as realistic simulations of the flow of ice through Nares Strait are presented. These demonstrate that the model reproduces the formation of stable ice bridges as well as the stoppage of the flow, a phenomenon occurring within numerous channels of the Arctic. In agreement with observations, the model captures the propagation of damage along narrow arch-like kinematic features, the discontinuities in the velocity field across these features dividing the ice cover into floes, the strong spatial localization of the thickest, ridged ice, the presence of landfast ice in bays and fjords and the opening of polynyas downstream of the strait. The model represents various dynamical behaviours linked to an overall weakening of the ice cover and to the shorter lifespan of ice bridges, with implications in terms of increased ice export through narrow outflow pathways of the Arctic.