Chapter 13 – sea ice hazards, risks and implications for disasters

Chapter 13 – sea ice hazards, risks, and implications for disasters Chapter 13 – sea ice hazards, risks, and implications for disasters Chapter 13 – sea ice hazards, risks, and implications for disasters Chapter 13 – sea ice hazards, risks, and implications for disasters Chapter 13 – sea ice hazards, risks, and implications for disasters Chapter 13 – sea ice hazards, risks, and implications for disasters Chapter 13 – sea ice hazards, risks, and implications for disasters

Chapter 13

Sea Ice: Hazards, Risks, and

Implications for Disasters

Hajo Eicken and Andrew R Mahoney

Geophysical Institute, University of Alaska Fairbanks, Fairbanks, AK, USA

ABSTRACT

The role of sea ice as a natural hazard is discussed with a focus on Arctic and Arctic regions where risks associated with human activities and ice processes are thegreatest Hazard assessment and emergency response need to consider a range ofcontrolling factors that can lead to events initiating an accident, failure, or full-scaledisaster These factors include environmental hazards, equipment, procedures andsettings, and people Quantifying risks associated with the presence of sea ice re-quires the joint consideration of the probability of specific hazards and the magni-tude of their impacts Both of these also depend on the type and level of humanactivity, such that disaster risks are substantially higher in the Arctic than in theAntarctic We identify three types of sea-ice hazards: (1) broad, long-term hazardsand associated risks associated with a rapid reduction in summer ice extent; (2) near-term hazards resulting from changes in sea-ice extent and dynamics such asincreased coastal erosion and threats to coastal infrastructure; and (3) immediaterisks and the potential for disasters derived from the combination of sea-ice hazardsand human activities such as shipping or offshore resource development A review ofkey properties and processes governing the role of sea ice as a hazard focuses onrecent rapid changes in ice extent and concentration in the Arctic and resultingthreats to coastal systems Other key factors include the distribution of old perennialice that has a greater thickness and higher mechanical strength than seasonal ice,patterns of ice movement that determine advection of ice hazards, and the degree ofice deformation that can generate thick, rough ice and represent a hazard in its ownright These factors are examined in the context of a case study for the Beaufort andChukchi Seas in the North American Arctic Linking specific environmental hazards

sub-to the geospatial distribution of human activities and vulnerable ecosystems allowsfor an integrated Arctic hazards assessment, currently still in its infancy The needfor coordinated environmental observations in informing hazard assessments andemergency response is discussed in the context of recent increases in maritimeactivities in the Arctic

Coastal and Marine Hazards, Risks, and Disasters http://dx.doi.org/10.1016/B978-0-12-396483-0.00013-3

Copyright © 2015 Elsevier Inc All rights reserved. 381

13.1 INTRODUCTION: ENVIRONMENTAL HAZARDS,

DISASTERS AND SEA ICE

The presence of sea ice in polar and subpolar oceans is a defining ronmental factor that governs weather and climate, ecology, and humanactivities in these regions Sea ice is also a significant natural hazard, boththrough direct interaction with assets and infrastructure and through theindirect impacts of variability and rapid changes in its distribution, inparticular in Arctic and sub-Arctic regions Although the former type ofthreat is clearly defined and falls well within the scope of classic hazardanalysis (e.g., Ogorodov et al., 2005; Zhang et al., 2013), the latter aspectrelates to the broader problem of threats to benefits or services provided bysea ice For example, delayed formation of coastal ice as a result of climatechange can expose shorelines to erosive action of fall storms and enhancethermal subsidence of coastal permafrost, over time greatly acceleratingrates of coastal retreat (Overeem et al., 2011) Nevertheless, both types ofthreats or hazards, rapid and slow onset, can be understood in the context ofsea-ice system services, which describe and potentially quantify the benefitsand threats that socialeenvironmental systems derive from sea ice over arange of scales (Eicken et al., 2009) This contribution considers both types

envi-of hazards and hence direct and indirect effects envi-of sea ice that can lead toharm or disasters

Sea ice is an integral part of socialeenvironmental systems, in particular inthe northern hemispheredhence, ice hazards and their potential role in di-sasters cannot be considered in isolation This broader perspective is sum-marized inFigure 13.1, which outlines the combination of factors initiating asequence of events that may culminate in failure of structures, loss of life andproperty, or escalate into a large-scale disaster Hazard mitigation and disasterprevention aim to curb or eliminate initiating events that can potentiallyescalate with negative consequences for human activities or ecosystems(Vinnem, 2007) It is important to recognize that in such a context, natural orenvironmental hazards, including sea ice, are part of a combination of factorsthat can lead to disasters Other factors include the design or hardware ofinfrastructure implicated in an event, human judgment, or error as well asprocedures in place to ensure safety and reliability of an organizationalstructure (Figure 13.1) The entire complex of factors, initiating events, pre-vention, escalation, and disasters can be understood in terms of risks associ-ated with specific consequences (Schneider et al., 2007) Typically, risk isevaluated in terms of a convolution of the probability of occurrence of aparticular event and the magnitude of its consequences

The concept of risk provides some guidance to the scope and thrust of thiscontribution By mapping the distribution of sea-ice processes and propertiesthat represent threats and hazards, we can help constrain the probability ofoccurrence of a disaster or loss event By evaluating harmful impacts through

direct (e.g., damage to offshore structures or vessels along shipping lanes) andindirect actions (e.g., enhanced coastal erosion or loss of climate regulationthrough reduced sea ice), we can assess both the probability and the magnitude

of consequences in a geospatial context In combination, such an evaluationprovides guidance on both the relevant processes that need to be consideredand the specific regions to focus on

The following section concentrates on a limited range of sea-ice cesses and properties relevant for our understanding of sea ice as a naturalhazard in both polar regions At the same time, for geospatially explicitdiscussions of consequences, the focus is on the Arctic where both the nature

pro-of sea-ice hazards and the probability pro-of occurrence pro-of events impactinghuman activities are more substantial than in the Antarctic In Antarctica,human activity and vessel traffic are concentrated into a small subregion ofthe Antarctic Peninsula region, where research bases and tourism account forroughly 100 vessel passages per year (based on 2007/2008 season; Lynch

et al., 2010) In the Arctic, such vessel densities are typical of the entirecircum-Arctic with, for example, roughly 140 vessels passing through theCanadian Arctic maritime region in 2012 with close to half of these vesselsconsisting of bulk carriers, tankers, and cargo ships (Pizzolato et al., 2014).Moreover, in contrast with the Antarctic, (sub-)Arctic sea-ice hazards play aprominent role in offshore resource development and associated coastalinfrastructure

Equipment Environmental

hazards People Procedures Design/setting

Hazards & controlling factors

Prevention

Escalation

Initiating events

Consequences:

e.g., failure of structures,

damage to life & property,

disasters

Risk:

probability

of occurrence magnitude of consequences

383

Chapter j 13 Sea Ice: Hazards, Risks, and Implications for Disasters

13.2 GEOGRAPHIC DISTRIBUTION OF SEA ICE AND KEY PROCESSES AND PROPERTIES

Polar sea ice waxes and wanes with the seasons Ice fills the Arctic basin andoccupies roughly 15 106km2 at its maximum winter extent in March,shrinking to a seasonal minimum in September of <5  106km2in recentyears (Figures 13.2 and 13.3;Comiso, 2010) As is evident fromFigure 13.2,reductions in ice concentration, that is, the areal fraction of ocean surfacecovered by sea ice, in particular during summer, have exposed large parts ofthe Arctic shelf seas to open water In the Pacific Arctic sector, this trendtoward reduced summer ice is particularly pronounced and coincides withincreases in ship traffic and offshore oil and gas resource exploration(Brigham, 2010; Lovecraft and Eicken, 2011) Note that in contrast theBering Sea has experienced more extensive, though thinner and typically lessconcentrated ice cover in recent years (Figure 13.2, left)

During 1979e2014, Arctic sea ice has been found to form later in the falland break-up earlier in the year Consequently, the open water season has beenincreasing in length, in some regions such as the PacificeArctic sector(Figure 13.2) by as much as a month or more However, satellite imagery used

to obtain such sea-ice climate data records typically tracks ice only at centrations>10 percent and may not detect ice during summer months at lowconcentrations (Meier and Stroeve, 2008) Hence, Arctic shelf seas and coastalregions, such as those highlighted in Figure 13.2, undergoing longer openwater periods may still see lingering ice that can represent a hazard but is notcaptured at a sufficient spatial and temporal resolution (Eicken et al., 2011).Reductions in ice concentration over the course of summer in the NorthAmerican and Siberian Arctic (Figure 13.2, right) and a longer open waterperiod have resulted in substantial increases in the amount of solar heating ofthese waters (Perovich et al., 2007) Further, increased fetch during the fallstorm season has resulted in increases in wave amplitudes in fall (Overeem

con-et al., 2011) Thinner, more dynamic ice also appears to be associated withincreases in wave heights and deeper propagation of swell into the ice pack inspring (Francis and Atkinson, 2012) These changes work in concert to in-crease erosive action, thermal subrosion of coastal permafrost, and therebyaccelerate rates of coastal retreat with potentially negative impacts on coastalcommunities and infrastructure

A key property of relevance from a natural-hazard perspective is the chanical strength of sea ice Typically, ice strength varies disproportionatelywith age since desalination of sea ice during the first and subsequent meltseasons greatly reduces the porosity of the bulk of the ice cover, therebyincreasing bending and compressive ice strength by between one half to asmuch as a factor of three to four (Timco and Weeks, 2010) Similarly, brackish

me-or freshwater ice fme-ormed in coastal lagoons me-or estuaries as well as glacial ice inthe form of icebergs will exhibit greater strength than sea ice Mapping the

FIGURE 13.2 Monthly Arctic sea-ice concentration trends derived from passive microwave satellite data for the time period 1979e2013 (left: March; right: September) National Snow and Ice Data Center Sea Ice Index; Fetterer et al., 2013

distribution of sea-ice age, using a combination of remote-sensing and driftingbuoy data, provides a large-scale measure of the probability of encounteringolder, stronger ice (Maslanik et al., 2007) For the period 2005e2010, theNorth American half of the Arctic basin was filled with ice three years or older(Figure 13.3), though the mean age of Arctic sea ice continues to drop sub-stantially as a result of reductions in the extent of perennial ice surviving atleast one summer’s melt Here, ice age can also serve as a proxy for level-icethickness (Maslanik et al., 2007) First-year ice rarely exceeds 2 m and typi-cally ranges between 1 and 1.5 m in thickness across the (sub)Arctic shelf seas

in late spring In the Arctic basin, second-year ice ranges in thickness between

2 and 3 m, and only the oldest level ice is likely to exceed a thickness of4e5 m However, the Canadian Archipelago generates much older ice that canexceed a 10-m thickness even in level, undeformed areas and which may enterthe Arctic Ocean through some of the straits exiting the Archipelago (Johnston

et al., 2009; Barber et al., 2014)

The prevailing patterns of ice motion, largely driven by surface windforcing anddon time scales of months to yearsdby sea surface tilt, transportFIGURE 13.3 Mean sea-ice age for the time period 2005e2010 The map is based on ice age fields derived by tracking regions of ice as they drift around the Arctic ( Maslanik et al., 2007 ).

older ice from the high Canadian Arctic into the Pacific Arctic sector andwaters off Alaska (Figures 13.3 and 13.4) Along the North American and EastSiberian coastlines, ice motion is mostly shore parallel, directed toward theWest with the anticyclonic Beaufort Gyre In central and western Siberia, icemotion is directed away from the shore, whereas north of Greenland and theHigh Canadian Arctic, it is toward the shore As a result, some of the oldest,thickest ice is found in the latter regions, whereas some of the youngest iceprevails in the former (Figures 13.3 and 13.4) Recent studies have found

an increase in both ice speed and deformation rate (the latter by roughly

50 percent per decade since 1979), likely a result of changes in surface forcingand reduced ice thickness that foster increased ice mobility (Rampal et al.,

2009) As is evident inFigure 13.4, regions with the highest mean ice velocitiesalso appear to experience the greatest increase in ice speed Moreover, recentwork byBarber et al (2014)demonstrates that as a result of the combination ofthese changes, the complexity of ice motion patterns and hence the challenge topredict ice movement on short time scales has greatly increased

A measure of the geographic distribution of ice deformation processes can

be obtained from an analysis of winter ice movement obtained from radarremote-sensing imagery over the time period 1996e2008 (Figure 13.5;Herman and Glowacki, 2012) Here, the distribution of cells that fall into thetop 5-percentile of the mean total deformation rate provides an indication ofareas where shear or repeated divergent/convergent deformation events havethe potential to produce highly deformed ice As is apparent fromFigure 13.5,

FIGURE 13.4 Mean Arctic sea-ice velocity field for winter (OctobereMay; left) and summer (JuneeSeptember; right) for the time period 1982e2009 as derived by Kwok et al (2013) The color scale indicates whether ice velocities are exhibiting a trend toward higher (red) or lower (blue) velocities.

387

Chapter j 13 Sea Ice: Hazards, Risks, and Implications for Disasters

the high deformation regions are confined to the Arctic Ocean marginal seasand are likely the result of iceeland interaction in combination with strongsurface forcing by winds (reflected in the mean fields shown in Figure 13.4)and tidal currents (Lyard, 1997) North of the Canadian Archipelago, wheresome of the oldest ice in the Arctic resides, deformation rates are not as high.This is explained in part by the high strength and compactness of multiyear ice

in this region that does not accommodate significant strain Ice deformationcan result in ice thickness multiple times that of thick-level ice, by piling icefloe fragments into ridges or rubble piles An example of such deformationprocesses and the resulting features is shown in Figure 13.6 for rafted andridged ice forced onshore by motion converging on the coastline at Barrow,Alaska A case study below provides further insight into these processes Thecoastal topography and prevailing ice motion (Figures 13.4 and 13.5) favorsuch iceecoast interaction with potentially significant implications for coastalprocesses and hazards

In the Southern Ocean, sea ice girdles the Antarctic continent with a wintermaximum extent of roughly 18 106km2, decreasing to 3 106km2 in the

FIGURE 13.5 Occurrence probability of very strong Arctic sea-ice deformation events, sponding to the top 5-percentile of total deformation rates for the entire Arctic, as derived by

corre-Herman and Glowacki (2012) Data cover the winter (NovembereApril) from 1996 to 2008 Rectangle (solid black line) delineates the approximate extent of the data shown in Figure 13.8

austral summer (Comiso, 2010) Due to higher ocean-to-ice heat fluxes andlack of old multiyear ice, Antarctic sea ice is typically much thinner (maximumlevel-ice thickness well below 1 m in most regions) than Arctic sea ice Also,multiyear ice does not undergo the same degree of desalination as in the Arctic,such that the bulk strength of Antarctic multiyear ice is typically of the samemagnitude as that of first-year ice Deformation processes and sea-ice circu-lation are driven by a similar combination of factors as in the Arctic, creatingsimilar hazards for vessel entrapment in regions of persistent deformation andconvergence Interaction between the sea-ice cover and the Antarctic coastline

is substantially different, however, from that of the Arctic because of theprevalence of either ice shelves or rocky shores throughout Antarctica

13.3 SEA ICE AS NATURAL HAZARD

Building on an understanding of the services sea ice provides to people andecosystems (Eicken et al., 2009), and taking into consideration in particularthe Arctic ice cover’s recent transformation, we can evaluate the role of sea ice

as a natural hazard As illustrated in Figure 13.1, this problem cannot betreated as an assessment of, for example, geophysical processes in isolationfrom the broader setting Therefore, we recognize three principal aspects of theproblem: (1) broad, long-term hazards and associated risks associated with arapid reduction in summer ice extent, such as geographic shifts in marineecosystems and warming of submarine permafrost and adjacent land; (2) near-term hazards resulting from changes in sea-ice extent and dynamics such as

FIGURE 13.6 Sea ice forced onshore at Barrow, Alaska in an ice-push event in spring of 2001 ( Mahoney et al., 2004 ) Such events require a compact ice cover offshore and wind or ocean forcing acting over a sufficient fetch of ice Note how the parent ice sheet of a roughly 1.5-m thickness has been rafted onto the beach at right, while failing in a buckling mode, resulting in

an ice pressure ridge at left Coastal infrastructure, such as powerlines, harbor installations, or buildings (visible in the distance at the right) can be threatened by such ice-push events.

389

Chapter j 13 Sea Ice: Hazards, Risks, and Implications for Disasters

increased coastal erosion and broad threats to coastal infrastructure; and(3) immediate risks and the potential for disasters derived from the combi-nation of sea-ice hazards and human activities such as shipping or offshoreresource development that involve specific, localized assets The first type ofhazard can be thought of in terms of slow-onset events, typically at regional tohemispheric scale, that require a response in the form of mitigation andadaptation (e.g., IPCC, 2012) Hazards of type 3 are associated with rapid-onset events at the local scale (though with regional to global repercussions,e.g., in the case of the threat to major infrastructure or the potential of an oilspill of national or international significance), while type 2 hazards can fallsomewhere in between.

Type 1 hazards can be considered in the context of services provided by thesea-ice cover, such as the regulation of global climate and its central role inice-albedo feedback (Eicken et al., 2009) Rapid changes in ice extent andfeedback processes may then have significant impacts, ranging from changes

in the Earth’s heat budget, to impacts on midlatitude weather to threats to keyspecies of global significance With the science, for example, on potentiallinkages between Arctic summer sea-ice reduction and extreme weather eventsstill emerging (e.g.,Francis and Vavrus, 2012; Screen and Simmonds, 2013),causal attribution and identification of specific hazards are challenging at thisbroader scale Nevertheless, initial model-based attempts have been made toestimate the economic impacts of such hazards (Euskirchen et al., 2013).Future research will have to establish whether hazard mitigation can bedirectly translated into specific reductions in the risk of, for example, extremeweather events or associated disasters

For type 2 hazards, the link between sea-ice processes and their variabilityand change is better established Thus, the potential for greater significantwave heights and a longer open water period with potential for wave actionand coastal flooding has been identified in a number of regions, with majorimpacts in particular in the Siberian and Pacific Arctic sector (Atkinson, 2005;Overeem et al., 2011; Asplin et al., 2012; Francis and Atkinson, 2012) Underconditions of a more mobile ice cover (Rampal et al., 2009; Barber et al.,

2014), this recent development also increases the probability for iceecoastalinteraction during the open water and shoulder seasons For example, incoastal Alaska, revetments put in place to protect the shoreline from waveaction have been damaged by storm-induced ice action during the fall freeze-

up period It is not clear whether to expect more frequent occurrence of push and similar events (Figure 13.6) as a result of a changing ice cover.Anecdotal evidence suggests that reductions in the amount of well-anchored,stable shorefast ice have led to an increase in such events in northern andWestern Alaska

ice-Impacts of reduced sea ice on coastal permafrost exacerbate the bility to flooding and coastal retreat Thus, wave action and warming of sur-face waters promote the thermal subrosion of both terrestrial and marine