High Arctic coasts at risk—the case study of coastal zone development and degradation associated with climate changes and multidirectional human impacts in Longyearbyen (Adventfjorden, Svalbard)

1 INTRODUCTION

Arctic coastal communities are special in their relationship with the polar environment. Having faced one of the largest climatic transformations on the planet, from the glacial period dominated by ice masses to the present period dominated by glacial retreat and permafrost thaw, Arctic communities have been constantly forced to adapt to new environmental conditions and develop new survival strategies. Over the last century the natural changes were superimposed with a cultural shift transforming their tribal and hunting legacies by urbanization and race for Arctic resources and pivotal military locations. The seminal report on the state of Arctic coasts (Forbes, 2011) emphasized the rapid change of circumpolar coastal environments that led to irreversible changes in coastal communities over the last century. The recent appeal by Fritz, Vonk, and Lantuit (2017) accentuated an urgent need for transdisciplinary research effort to investigate the physical and human impacts resulting from collapsing Arctic coastlines.

The observed acceleration of Arctic coastal change is mostly associated with a decreasing sea ice extent and duration that increases shoreline exposure to storm wave energy and erosion (e.g., Barnhart, Overeem, & Anderson, 2014). Loss of sea ice is only one of the processes that transform Arctic coastal zone. The functioning of the present-day Arctic coastal system is also influenced by permafrost degradation (e.g., Wobus et al., 2011), storm-surge floodings (e.g., Pisaric et al., 2011), increased sediment supply from glacial catchments (e.g., Mercier & Laffly, 2005), and even such extreme events as tsunamis (e.g., Buchwał, Szczuciński, Strzelecki, & Long, 2015; Dahl-Jensen et al., 2004; Long, Szczuciński, & Lawrence, 2015). Majority of those changes have a strong impact on circum-polar Arctic coastal communities and their historical (including archaeological) and modern infrastructure (e.g., Ford, Bell, & St-Hilaire-Gravel, 2010; Gorokhovich, Leiserowitz, & Dugan, 2014; Hatcher & Forbes, 2015; Mason, Jordan, Lestak, & Manley, 2012; Radosavljevic et al., 2016).

In the European sector of the Arctic—Svalbard serves as a barometer of environmental, climatic, and economic change (Figure 1). It is noteworthy, that although Svalbard serves as a reference site for High Arctic environmental change research and significantly contributed to improving our understanding of glacial systems; state of permafrost or sensitivity of fjord environments to climate shifts; the coastal zone changes and adjustments to climate warming remained one of the under-explored topics (Overduin et al., 2014). The study of circum-arctic coastal erosion rates carried out by Lantuit et al. (2012) suggests the relative stability of Svalbard shorelines (0 m yr⁻¹ change). More recently, Sessford, Bæverford, and Hormes (2015) modified this view by a detailed analysis of multiyear erosional processes operating on unlithified cliffs developed along inner fjords of Spitsbergen (main island of Svalbard) and observed that erosion rates range from 0.34 (for ice-poor cliffs) to 0.47 m yr⁻¹ (for cliffs with higher ice content).

In general, over the last decade, Svalbard coastal studies have been concentrated on the coastal zone response to shifts in sediment supply associated with changes in local ice masses and paraglaciation (e.g., Bourriquen et al., 2016; Mercier & Laffly, 2005; Sessford, Strzelecki, & Hormes, 2015; Strzelecki et al., 2018; Zagórski, 2011); ephemeric pulses of sediments from snow-fed streams (Lønne & Nemec, 2004; Strzelecki, Long, & Lloyd, 2017); or the controls of coastal permafrost development (Kasprzak et al., 2017). Recent years had also brought some advances in local rocky coast systems that dominate over coastal landscape in numerous Svalbard fjords (e.g., Kasprzak et al., 2017; Strzelecki, 2011; Strzelecki, 2017; Strzelecki et al., 2017; Świrad, Migoń, & Strzelecki, 2017). Most recently, Jaskólski, Pawłowski, and Strzelecki (2017) reported on the impact of geohazards including coastal changes on the state of abandoned Arctic town—Pyramiden in Northern Billefjorden.

Coastal zone in the surroundings for Longyearbyen has been a subject of important sedimentological investigations focusing on development of High Arctic deltas (e.g., Lønne & Nemec, 2004; Prior, Wiseman, & Bryant, 1981) and modern sedimentation rates (e.g., Zajączkowski, 2008). The richness of marine environment and complexity of ecological processes operating in Adventfjorden have been thoroughly described in a state-of-the-art review by Węsławski et al. (2011). More recently, Guégan and Christiansen (2016) presented the detailed study of seasonal changes along bedrock cliffs overlain by unconsolidated sediments at the entrance to Adventfjorden. In recent years, the functioning of central Spitsbergen coastal zone has been modified by increased duration of open-water conditions and destabilization of coastal slopes (e.g., Larsson, 1982; Lønne & Nemec, 2004; Muckenhuber, Nilsen, Korosov, & Sandven, 2016).

Strangely, although the main community infrastructure is located within narrow coastal zone including the airport, main road, ports, and several housing and storage facilities, no coastal zone change monitoring and environmental risk assessments have ever been carried out in Longyearbyen. In this paper, we identify the major coastal changes and potential hazards to Longyearbyen community infrastructure in this strategic region for Arctic tourism, governance, and scientific activity.

2 REGIONAL SETTING

Longyearbyen (78°22′N and 15°64′E) is the most populated settlement and administrative centre of Svalbard. Over the last century, the town established in 1906 has evolved from important coal mining centre into a popular tourist destination, science hub for polar research, for example, University Centre in Svalbard, Norwegian Polar Institute, and Svalbard Science Forum, and strategic infrastructure such as Svalbard Satellite Station or The Svalbard Global Seed Vault. The town has recently experienced a rapid growth both in terms of population from 1,100 inhabitants in 1992 to 2,150 inhabitants in 2014 (Statistics Norway, 2014). The increase of population associated with growth of tourism and scientific activities led also to development of community infrastructure and construction of new buildings (Figure 2). The surroundings of Longyearbyen are mountainous and over the Quaternary has been transformed by glacial and periglacial processes. Due to the sedimentary geology (Cretaceous to Tertiary), flat plateaus, reaching around 500 m a.s.l., dominate over the landscape. Two glaciers are situated at the end of the valley and supply braided river system of Longyearelva. Channelization and constant river management of Longyearelva, to some extent, reduced the risk of channel shifts and flooding of buildings already located on the floodplain. Longyeardalen is typical glacially eroded U-shaped valley with active slope processes. Numerous snow avalanches are triggered on a yearly basis in the study area, and most of local slopes are reshaped by avalanche activity (e.g., Eckerstorfer & Christiansen, 2012; Eckerstorfer, Farnsworth, & Birkeland, 2014). One of the most recent avalanches that hit the town (December 19, 2015) killed one person, injured nine more persons, and destroyed several buildings (http://www.bbc.com/news/world-europe-35144236). Apart from avalanches, the major natural hazard affecting the community infrastructure are debris flows (e.g., Bernhardt, Reiss, Hiesinger, Hauber, & Johnsson, 2017; Jahn, 1976; Larsson, 1982). Study area is underlain with continuous permafrost. According to Humlum, Instanes, and Sollid (2003), permafrost at sea level and in the valley bottom is most probably of the late Holocene origin.

The southern shore of Adventfjorden, analysed in our study, is characterized by a mixture of natural and anthropogenic landforms. The entrance to the fjord is dominated by low bedrock cliffs and gravel beaches. Coastal zone between airport and ports is strongly transformed by human action (shore protection constructions and installations). The most diverse coastal landscape is found between Longyearelva and Adventelva river deltas where mosaic of mixed sand–gravel barriers and spits enter muddy tidal flat (Figure 1b).

3 MATERIALS AND METHODS

In order to assess the impact of coastal hazards on Longyearbyen infrastructure, as well as the influence of town development on the coastal environment, we studied an approximately 10 km-long section of Longyearbyen coast from Vestpynten to the tidal flat at the mouth of Adventdalen (Figure 1b). We used the term ‘hazard’ to describe any process or situation that can cause harm, damage, or loss in the coastal environment. (e.g., erosion, landslides, extreme weather events, and waste dumping). We defined the environmental risk as the likelihood or probability of occurrence of hazard. In this study, risk refers to risk to the coastal environment including community infrastructure built along the coast.

The CVI presented in our study is based on CVI developed by Gornitz, Daniels, White, and Birdwell (1994). The CVI has been calculated as the square root of the product of the scored variables divided by the total number of variables as

where

• a = geomorphology;
• b = coastal slope;
• c = rate of relative sea-level rise;
• d = shoreline erosion/accretion rate;
• e = mean tide range; and
• f = mean wave height.

The index allows the six environmental variables to be related in a quantifiable manner. Classification of coastal geomorphology in each 500-m-long section was supported by field-based mapping and interpretation of aerial images from 1990 to 2009, provided by Norwegian Polar Institute. Coastal slope was calculated from Numerical Terrain Model with 20-m resolution provided by Norwegian Polar Institute. Due to the lack of a reliable tide gauge record in Longyearbyen, we used the mean relative sea-level change data for Barentsburg (ca. 40 km from Longyearbyen) obtained from Permanent Service for Mean Sea Level based in Liverpool, England at the National Oceanography Centre (−1.76 mm yr⁻¹ of relative sea-level fall). After digitizing shorelines from 1990 to 2009 in QGIS software, recent shoreline displacement was calculated using the DSAS extension for ESRI ArcGIS (Thieler, Himmelstoss, Zichichi, & Ergul, 2009). The program was imposed to create 300-m-long transects every 10 m, and the Net Shoreline Movement (NSM) method was used in the calculation. NSM calculates the distance between the oldest and the youngest coastline for each designated transect. Measurement values are expressed in negative (erosion) and positive (accretion) values. The standard error was 2 m as it was not possible to distinguish if the visible shoreline comprises ephemeral gravel berms or storm ridges, which are currently separated by about 2 m. The confidence interval in DSAS was at 95%. Based on the field observations and information obtained from the Longyearbyen Port Office, we used value of 1.5 m as a local tidal range and 0.8 m as a mean wave height. For each of 20 coastal sections delimited for CVI and LM analyses, we have used the mean shoreline change value from 50 DSAS transects.

The LM was developed by Leopold, Clarke, Hanshaw, and Balsley (1971), in response to the National Environmental Policy Act of 1969. It is a two-dimensional matrix (a × b) that consists of columns (factors of impact/kind of impact) and rows (recipients exposed to the influence/recipient of impact). The LM provides a system for the analysis and numerical weighting of probable impacts. When assessing the impact of the factor on a particular element of the environment, we take into account the ‘weight’ of the factor (qualitative assessment) and its ‘strength’ (quantitative assessment). Weight of the factor determines whether and to what extent the phenomenon affects a particular element. Strength defines the degree of intensity and the nature of impact denoted by + and − (positive or negative impact). In our analysis, the positive impact is an effect of a process that does not endanger the coastal environment and does not pose a risk to town development. An example of a positive impact in our study area is the protection of shoreline by fast ice (ice anchored to the shore). Negative impact was given to any process that threatened the coastal environment and damaged the community infrastructure. An example of a negative impact is the coastal erosion that damages the shore protection infrastructure and activates landsliding.

Based on the field observations and a literature review of geomorphological and other environmental processes operating in the surroundings of Longyearbyen, we selected and assessed their impact on the state of Longyearbyen coast by the following 17 factors (of both natural and anthropogenic origin): waste dumping, sewage dumping, coastal erosion, coastal accumulation, mass movements (rockfalls, landslides), solifluction (periglacial processes), fluvial accumulation, fluvial erosion, stranded ice, fast ice, sea ice, wind action, precipitation, snow cover, storms, permafrost, geology.

We divided the studied coast into 20 even sections (500 m alongshore, 100 m offshore, and 900 m onshore). In each section, we documented the presence, type, and state of infrastructure, as well as the record of past and present-day environmental hazards (Data S1). Our results enabled an assessment of the state and safety along the Longyearbyen coast in terms of the potential impacts on: housing, service buildings, industrial buildings, recreational buildings & areas, roads, airport, ports, coastal protection, technical equipment, and utilities.

In order to reduce the potential subjectivity of LM method, we analysed also the results of CVI and DSAS commonly used in global coastal protection strategies (e.g., the National Coastal Vulnerability Assessment to the Atlantic Coast and European Climate Adaptation Platform).

Fieldwork was organized in summer seasons 2014 and 2015 to ground-truth remote-sensing analyses and map small-scale coastal landforms and effects of human impacts on coastal stability (waste dumping, construction works, shoreline protection, etc.). Short revisit in 2016 was carried out to check if observed changes and degradation of coastal zone continues.

4 RESULTS AND DISCUSSION

4.1 Coastal change and vulnerability in Longyearbyen

The coastal change rates for Longyearbyen for the period 1990–2009 range from −0.5 to 4.5 m yr⁻¹ (Table 1). The recent evolution of Longyearbyen coastal zone is an interesting example of an interaction of natural and anthropogenic processes in rapidly changing Arctic environment. The natural course of coastal development has been transformed by expansion of human infrastructure. For instance, the major change observed along the Longyearbyen coast was a build-up of new coastal zone in the Longyearbyen port section, where construction of road and reinforced coastal embankment led to over 80 m advance of a shoreline (Sections 15 and 16—Figure 3a and Table 1). In comparison, the natural progradation rates of the delta system (Sections 18–20—Figure 3a and Table 1) and gravel-dominated barriers (Section 5–7, 9, and 10) were almost 10-times slower. Similarly, the coastal retreat rates observed along anthropogenically transformed sections of the coast such as Section 20 (Figure 3a and Table 1), where mixture of construction waste, garbage, and soil was used to heap up a cliff, are several-times faster than along natural coastal sections (e.g., Sections 2, 4, and 8).

Table 1. Coastal change rates based on Digital Shoreline Analysis System method

Section no.	1	2	3	4	5	6	7	8	9	10	11	12	13	14	15	16	17	18	19	20
Total shoreline change (m)	0.4	−5.8	2.8	−3.9	0.3	7.2	0.5	−1.1	6.8	0.6	−0.1	18.6	22.7	20.0	80.1	87.2	4.2	7.5	4.9	−10.3
Mean annual shoreline change (m yr−1)	0.0	−0.3	0.1	−0.2	0.0	0.3	0.0	0.0	0.3	0.0	0.0	0.9	1.2	1.0	4.2	4.5	0.2	0.4	0.2	−0.5

The CVI values calculated for 20 delimited sections of Longyearbyen coast ranged from 4.9 (very low) to 9.49 (high). The high CVI value of Longyearelva delta coast (Sections 18 and 19; Figure 3a) is linked with low coastal slope of delta system and potential increase of erosion due to limited sediment supply from channelized river system. Other sections, Sections 1–17 and 20, were diagnosed with relatively low to moderate coastal vulnerability (4.9–8.49). However, the results of CVI are adulterated by shoreline changes associated with the development of port and road infrastructure. In most cases, the build-up of artificial coasts during the construction of new port and roads changed the geomorphology (in terms of accretion and mechanical composition) and slope of coastal zone (steepened).

4.2 Environmental risk assessment analysis

The data on the environmental state of Longyearbyen coastal zone and the inventory of community infrastructure analysed in modified LM indicated that coastal and river erosion and permafrost-related processes (thawing and solifluction) are factors that have the strongest negative impact on Longyearbyen infrastructure (Figure 4a). Based on recent dramatic events observed in Longyearbyen, associated with avalanches and landslides (2015–2017), we expect that in coming years, the influence of snow cover changes, and mass movements on the state of coastal zone may increase. By the time of this study, the types of community infrastructure most affected by coastal zone changes were ports, quays, and coastal protection installations (Figure 4b). The moderate risk was characterized for Longyearbyen road system, however, the results of analysis are based by the high risk of landslides and coastal erosion along major road linking town with the airport, as well as side roads threatened by snow avalanches and debris flows. In Figure 3b, we present the summary of environmental assessment analysis carried out for each of 20 sections. Most of the analysed sections (Sections 1–8, 11, and 12—Figure 3b) were found by moderate negative impact of coastal zone changes on community infrastructure. Lack of housing infrastructure, limited pollution, and coastal accumulation (both natural and man-made) resulted in low negative impact of processes operating in coastal zone on community infrastructure in Section 9, 10, 15, 16, and 20. The strongest negative impact on Longyearbyen infrastructure was calculated for Sections 13–14 (LM value up to −1.86) and 17–19 (LM value up to −1.96). In the first case, the main road linking the town with the airport crosses the narrow coastal strip affected by both coastal erosion and landslides, whereas in the second case, the effects of natural hazards (coastal and fluvial erosion) are enhanced by poor coastal protection infrastructure and uncoordinated waste management operations (including unorganized waste dumping and incorporation of industrial waste in coastal and fluvial reinforcement structures).

4.3 Impact of natural and human hazards on Longyearbyen coastal zone

We have mapped the state of Longyearbyen coastal zone over the years 2014–2016 focusing on the effects of human actions and natural hazards including coastal and periglacial processes. Along the gravel-dominated barrier system located in front of Longyearbyen Airport (Site a—Figure 5), the total erosion of low-lying, vegetated coastal bluffs was approximately 4 m (in period 1990–2009). We agree with Guégan and Christiansen (2016) that observed erosion rates threaten the stability of road linking Longyearbyen and Bjørndalen. Across the whole barrier (from Vestpynten to Port 3), beach surface was polluted with fragments of coal, construction wood, glass, asphalt, and plastic trash posing a threat for the coastal ecosystem, especially for birds occupying the ponds along the shore zone.

The pollution linked with a presence of mine dumps (coal, corroded and abandoned constructions, and machines) is concentrated between Ports 2 and 3 (Site b—Figure 5). One of the most critical sections of the Longyearbyen coast is a narrow gravel-dominated barrier (Site c—Figure 5) stretched between Port 1 (yacht port) and Port 2. Here, the barrier is occupied by a main road linking town with the airport, and over observation period was a subject to several hazards including coastal erosion, sea-ice pile-up, rockfalls, and debris flows from relict rock cliff and unstable mountain slopes. However, the highest number of hazardous processes affecting the state of Longyearbyen coast was mapped between Port 1 and tidal flat developing at the mouth of Adventdalen (Sites d–i—Figure 5). This section of the coast is the most transformed and polluted part of Longyearbyen coastal zone. Two major storage and industrial districts of the town are located in this area, and all buildings situated at the seafront are at risk of coastal erosion. For instance, the recent collapse of the cliff, composed mainly by industrial waste mixed with marine gravels, led to over 5-m shoreline retreat and threatens the stability of a rubb-hall (Site f—Figure 5). Even more critical is the state of a waste-dump cliff located on the eastern side of Longyearelva delta (Site i—Figure 5), which is seriously destabilized by slumping and numerous debris and mudflows. This area is in particularly poor state due to the dumping of industrial and municipal waste (including remains of dead seals). In years 2014–2016, we have mapped over dozen fresh slumps and 20 debris flow lobes composed of mixed waste and slope deposits. The erosion of the cliff base that occurs when storm waves enter Adventfjorden at high tide leads to further instability of the cliff and transport of eroded waste to the fjord system.

Another critical issue observed along Longyearbyen seafront is the state of coastal protection infrastructure. Most sea walls and breakwater structures in Longyearbyen are rather basic, constructed of concrete waste, and rusty scrap metal (see Sites d, e, and g—Figure 5). In summer 2016, most of these primitive structures were heavily damaged and vulnerable to coastal erosion and destructive sea-ice action. The surroundings of the Longyearelva delta (Site h—Figure 5) are example of one of the most heavily anthropogenically transformed natural system in Svalbard. The rapid growth of town, observed over the last decade, led to limitation of investment area (area predisposed or/and prepared for development). To obtain new construction lands, the floodplain of Longyearelva was channelized limiting the migration of braiding channels. The delta system was also subject to river engineering that focused on construction of artificial channel banks and maintaining single river channel. Regardless of the significance of fragile barrier and delta environments for local wildlife, the artificial banks were constructed from mixed fluvial deposits and industrial waste (mainly plastic, concrete, and scrap metal).

5 CONCLUSIONS

AUTHOR CONTRIBUTIONS

M. W. J. conceived the study. L. P. assisted with fieldwork and carried out GIS analyses. M. C. S. guided the intellectual direction of the research. M. W. J. wrote the manuscript with input from all co-authors.