209FEBRUARY 2004AMERICAN METEOROLOGICAL SOCIETY |

A

“When the heavy ice is way out, old timers say, ‘I hopewe don’t have that storm.’”

—KENNETH TOOVAK

striking aspect of anticipated global climatechange in response to increasing greenhousegases is that the largest warming is predicted to

occur in the Arctic (Houghton et al. 2001). The pre-

dicted Arctic warming, though spatially variable, issupported by both current climate observations(Serreze et al. 2000) and paleoclimate reconstructions(Overpeck et al. 1997). This observed warming hasaffected glaciers, sea ice, ecosystems, and permafrost.Increasing amounts of open water in the arctic seas,combined with rising sea level and the coastal geog-raphy, are contributing to increased harshness ofstorm impacts on coastal areas, resulting in damagesfrom high winds, storm surges, flooding, and shore-line erosion (BESIS 1997; Cohen 1997). Storm surgesand waves are greater when little or no sea ice ispresent. Further, less persistent landfast ice (icegrounded near shore) can reduce protection fromerosion. In some areas of the North American Arc-tic, the coastline has eroded by more than 1 m yr-1.Coastal wetlands and moist tundra regions are un-stable and easily changed by erosion, flooding, and theinvasion of saltwater. Villages are frequently locatedon the coast to provide proximity to marine resources.Flooding of low-lying coastal communities threatensinfrastructure such as harbors, water supplies, fuelstorage, and sewage disposal systems.

This paper introduces a project to apply an im-proved understanding of climate variability to informlocal decision making regarding coastal communities,

TOWARD AN INTEGRATEDASSESSMENT OF THE IMPACTSOF EXTREME WIND EVENTS

ON BARROW, ALASKABY A. H. LYNCH, J. A. CURRY, R. D. BRUNNER, AND J. A. MASLANIK

Involving local residents in the integrated assessment of the impacts of climate change on

their community redirects the project focus to extreme events.

AFFILIATIONS: LYNCH—Program in Atmospheric and OceanicSciences and Cooperative Institute for Research in the Environmen-tal Sciences, University of Colorado, Boulder, Colorado; CURRY—School of Earth and Atmospheric Sciences, Georgia Institute ofTechnology, Atlanta, Georgia; BRUNNER—Center for Public PolicyResearch, University of Colorado, Boulder, Colorado; MASLANIK—Cooperative Institute for Research in the Environmental Sciencesand Department of Aerospace Engineering Sciences, University ofColorado, Boulder, ColoradoCORRESPONDING AUTHOR: Judith A. Curry, School of Earth andAtmospheric Sciences, Georgia Institute of Technology, Atlanta, GA30332-0340E-mail: curryja@eas.gatech.eduDOI: 10.1175/BAMS-85-2-209

In final form 7 January 2003©2004 American Meteorological Society

210 FEBRUARY 2004|

management of marine mammals, marine transpor-tation, and offshore resource development for theNorth Slope of Alaska (Fig. 1). The North Slope Bor-ough (NSB) encompasses 89,000 square miles on thiscoastal plain, with a population of approximately8000. Permanent residents, the vast majority of whomare Inupiat Eskimos, are concentrated in eight vil-lages; the largest is the city of Barrow. The NorthSlope Borough was incorporated in July 1972, follow-ing an assessment of the implementation of the AlaskaNative Claims Settlement Act (Berger 1985). Thisincorporation provided “the self-government neededto better the life-style of Inupiat people on the NorthSlope” (Kaleak 1993). Approximately half of theInupiat adult workforce engage in a subsistencelifestyle (see sidebar 1). Another 5200 people work atleast half the calendar year in the Prudhoe Bay andother oil fields (NSB 1999). This region was selectedfor this study for several reasons: it is modest in geo-graphic scope and relatively isolated from remoteevents that cannot be adequately anticipated or con-trolled, there are significant current issues of eco-nomic and policy relevance associated with climate,and authority and control for governing the regionare concentrated in the NSB. Further, dramaticchange is already evident in this region, and hence itrepresents a relatively simple model of larger prob-lems and opportunities that may later become appar-ent in other regions of the globe.

The Alaskan nearshore region comprises the Beau-fort and Chukchi Seas, including Kotzebue Sound,

and is bordered by a broad coastal plain (Fig. 1). Thecoast is predominantly low-lying, wetland tundra,dotted by numerous thaw lakes. The continental shelfhas a gently sloping bottom with relatively flat terraininland. Offshore islands and shoals moderate the lo-cal influence of polar pack ice on the coast. Most ofthese islands are sand and gravel barrier islandsbounding shallow lagoons, while others are relics ofearlier coastal retreat processes and lie farther off-shore. The islands themselves, and mainland coastwhere unprotected by islands, are subject to consid-erable erosion by wave action. Aided by thermal ero-sion of the tundra, erosion rates average 1–3 m yr-1

along the mainland coastline, but in some locationsmay reach 38 m in a single severe storm. Because ofthis geography, much of the Chukchi–Beaufort coastis vulnerable to high storm surges. The fall stormseason is the most dangerous time of the year, whenhigh winds and wind-driven high tides place residentsin greater danger of coastal flooding than at any othertime of the year (P. Patterson 2002, personalcommunication).

Since the 1960s, the social, economic, and culturallife of the North Slope has changed in response to theAlaska Native Claims Settlement Act, oil and gas de-velopment, migration, the environment movement,and technological change. The growth of the NSB, theInuit Circumpolar Conference, the Alaska EskimoWhaling Commission, and other local institutionsreflects the increasing social complexity of the region.The interplay of federal, state, and local concerns in

coastal zone planning andenvironmental protectionhas also been importantduring this period of rapidchange. Permanent com-munity residents want toprotect the Inupiat culture,subsistence whaling tradi-tions and privileges, sub-sistence-resource harvestareas, and locally based de-cision making.

Climate variability andchange are accepted as factby residents of the NorthSlope because various signsare prominent in their ev-eryday experience. NorthSlope residents have madeappropriate adaptations inmany aspects of their lives(see Sidebar 1). Because of

FIG. 1. Map of the North Slope of Alaska, showing Barrow and the North SlopeBorough, topography, and bathymetry. Digital elevation model 100-m grid cre-ated from U.S. Geological Survey (USGS) data. Bathymetry obtained from1:250,000-scale National Ocean Service data. Inset shows a map of the Bering–Chukchi region, encompassing Alaska and eastern Siberia, Russia.

211FEBRUARY 2004AMERICAN METEOROLOGICAL SOCIETY |

the vulnerability to environmental change, the NSBhas become a leader in the enterprise to involve localpeople in the planning, management, and perfor-mance of environmental science (Brewster 1997).This is due in large part to the establishment in 1947of the Arctic Research Laboratory in Barrow in whichInupiat people were involved from the very begin-ning. Inupiat knowledge of the arctic environmenthas proved crucial to the success of many scientificprojects in the region and has led to new interdisci-plinary approaches. The North Slope residents arenevertheless very much in doubt about how to mini-mize the significant vulnerability of people and prop-erty to extreme events. Development of Barrow’sunderground water and sewer pipelines (called the“utilidor”), above ground structures, freshwater andsewage lagoons, and the coastal road contributes tothe increased vulnerability of the community to flood-ing and erosion. Barrow’s utilidor has been valued at

roughly half a billion dollars (M. Aamodt, Boroughassemblyman, 2002, personal communication). Tominimize vulnerability to extreme events, the NSBimplemented a beach nourishment program, but thishas since been terminated. The NSB has also consid-ered federally subsidized insurance, raising the heightof the road (see Fig. 6), and adding a concrete mat-tress revetment to the seaward slope of the coastalroad and bluffs.

An active partnership with residents is necessaryto understand how natural and social sciences re-search can be integrated with local knowledge to as-sist North Slope residents in adapting to climate vari-ability and change (Korsmo 1999). For this study tosucceed, we must understand residents’ perspectiveson climate variability and change in the region in or-der to focus our scientific research on their principalconcerns and eventually to advise them on possiblepolicy responses to priority problems. An active part-

The Inupiat rely on a variety ofmarine resources as significantsources of food. The harvesting,sharing, and consuming of subsis-tence resources form an impor-tant part of the traditional Inupiatculture and spiritual life. Subsis-tence activities and resources aredefined to be required for survivaland do not accumulate capital forits own sake.

Walrus, numerous species ofseals, and whales all follow the iceedge, taking advantage of theready access to food and (for sealsand walrus) the availability of icefor sunning, mating, and raisingpups (Tynan and DeMaster 1997).Threats to the marine subsistencefood supply arise, or are perceivedto arise, from

• changes to the coastal oceanenvironment associated withclimate change, such as sea iceretreat;

• mineral resource extractionactivities, including offshore oildrilling and the laying ofpipelines (NRC 1994); and

• political issues (e.g., the recentInternational Whaling Commis-sion revocation of a subsistence

THREATS TO THE SUBSISTENCE LIFESTYLEexemption for Inupiat nativepeople because of a concernwith Japanese subsistenceallowances).

Inuit hunters are noticingdiscolorations and thinning of seaice, changes in the leads and openwater areas, and the presence ofanimals not previously found inthe region. It has been suggestedthat early formation of extensiveice-free conditions along theNorth Slope in 1998 (Maslaniket al. 1999) prevented a substan-tial number of polar bears frommigrating onto the pack ice.Changes now being experiencedin the western Arctic are re-garded as “portents of increas-ingly frequent disruptions in seaice dynamics” (D. Norton 2002,personal communication). Fornow, however, native residentsappear to have adapted success-fully to climate changes that havemade whaling more dangerous(E. Itta 2002, personal communi-cation). Among other things, theyhave employed GPS, helicopters,and other modern technologies insearch and rescue operations, andlarger boats for hunting, fishing,

and transportation to outlyingcamps (E. Brower 2002, personalcommunication). However, theydo not yet have a satisfactorylaunching facility for the largerboats. (J. Vorderstrasse 2001,personal communication).

Calving of landfast ice, exacer-bated by strong winds, is aparticularly important issuebecause of the potential danger tospring whalers. In May 1997, theAnchorage Daily News reportedthat 142 Inupiat whalers had to beevacuated from drifting ice byhelicopter when landfast ice brokeaway from the shore near Barrow.At that time, the landfast ice wasthinner than generally observed(H. Brower Jr. 2003, personalcommunication). This eventprompted further research tobetter understand the factorsaffecting the stability of thelandfast ice, such as the installa-tion by NSB Department ofWildlife Management scientists ofsensors to monitor sea levelchanges under landfast ice insubsequent spring whaling seasons(C. George 2003, personalcommunication).

212 FEBRUARY 2004|

nership is possible because residents continue to beconcerned about issues of climate change and variabil-ity on the North Slope (C. Pungowiyi 2003, personalcommunication; Cochran et al. 1999). We have be-gun to construct an active partnership through a se-ries of meetings with a variety of local citizens andgroups, as well as public seminars and liaisons withlocal school teachers and students.

CLIMATE VARIABILITY. Chapman and Walsh(1993) and Serreze et al. (2000) have documentedincreases in the annual mean and seasonal tempera-ture trends in the northwest American sector of theArctic over the period 1966–95 (see Sidebar 2).Locally, annual trends exceed 0.5°C per decade, even

when allowance is made for the “urban heat island”effect (Jones 1994). On a seasonal basis, warming isconcentrated in the winter and spring, with areally av-eraged trends in summer close to zero. Stone (1997)analyzed surface air temperature at Barrow over theperiod 1965–95 and found that annual warming ismost pronounced during late winter, including a sta-tistically significant 30-yr increase of 5.5°C in Febru-ary. For the period 1949–98, Stafford et al. (2000)found the increase in mean daily temperature at Bar-row to be around 1.6°C for the annual mean and justunder 3°C for winter.

We have conducted a more extensive analysis of thedaily averaged surface air temperature record for Bar-row using data from the National Climatic Data Cen-

There is a clear perceptionamong residents of the NorthSlope that the climate of theregion has changed in livingmemory. Factors influencing thisperception include the fact thatauguring for foundations inBarrow has had to go deeper toreach permafrost (M. Aamodt2001, personal communication),snowmelt onset at hunting campsis becoming unpredictable (K.Toovak 2002, personal communi-cation), the sun is feeling hotter(K. Toovak 2002, personalcommunication), and the summermosquito population is increasing(M. Carroll 2002, personalcommunication).

Careful examination of theclimate records at Barrowgenerally supports these percep-tions. Directly associated with theincreasing air temperature is adecrease in the amount andperiod of winter snow cover andincreasing permafrost thaw depth.Winter total precipitation inBarrow has decreased by morethan half in the last 50 yr (Staffordet al. 2000). Snow cover onsetshows very little change, butsnowmelt onset shows a signifi-cant trend toward earlier springsnowmelt—almost 1 monthearlier over the last 50 yr.Further, the snowmelt date(defined as the date on which

PERCEPTIONS OF CLIMATE CHANGE BY LOCAL RESIDENTSsnow depth fell below 2.5 cm forthe first time) is showing in-creased variability over that sameperiod (Fig. SB1)—the windowrepresenting the 90% confidenceinterval for snowmelt onset datewould have to be increased bynearly 1 day every 3 yr. This isconsistent with May air tempera-tures, which show an abrupt andrapid increase in variability since1990.

Soil temperatures measured atBarrow since 1994 indicate thatthe end-of-season permafrostthaw depth has increased be-tween 1994 and 1998, followed bya decrease in 1999 and 2000(Hinkel et al. 2001). A strongcorrelation exists between themaximumannual thawdepth and theannual thawingdegree daysover the periodof record.

Data oncloud coverand radiativefluxes arerelevant toperceptionsthat “the sun isbecomingmore intense”but are insuffi-cient to

support long-term analyses in thisdata-poor region. Annually, lowcloud has decreased from anaverage of 6.5 to 5.5 tenths inBarrow (Curtis et al. 1998),although cloudiness has generallyincreased in winter and earlyspring (Stone 1997). Gurney(1998) presents an analysis ofnoontime ultraviolet irradiancemeasurements made at Barrowfrom 1991 to 1996. After elimi-nating the effects of clouds,increases between 3% and10% yr-----1 in the wavelengthsknown to cause biological damagewere found in all daylit monthsexcept June, with a strongcorrelation to total column ozoneamount.

FIG. SB1. Snow cover end date with linear trend (slope–0.529 days yr-----1) showing increasing variability for therecord over the last 50 yr.

213FEBRUARY 2004AMERICAN METEOROLOGICAL SOCIETY |

ter (NCDC), the NationalOceanic and AtmosphericAdministration (NOAA)Climate Monitoring andDiagnostics Laboratory(CMDL), and the Geophysi-cal Institute at the Universityof Alaska, Fairbanks. TheNational Weather Service(NWS) record extends from1921, with no missing data.The station was moved in1924, 1925, 1942, 1944, 1955,and 1966—all moves in-volved distances less than amile and elevation changesof less than 20 ft. Tempera-ture instruments were low-ered a few feet in 1977.Barrow has experienced asignificant warming overthe last 80 yr, but this warm-ing is not uniform or significant for all seasons, noris it uniform over the entire period (Fig. 2). A signifi-cant linear trend of increasing temperatures is evidentin winter and spring. Trends are positive regardlessof the time window chosen (from the present day) forcalculating slope, and the warming may be accelerat-ing. As Stone (1997) found, the most intense warm-ing was observed in late winter, with a statistically sig-nificant increase of 3.0°C over the 81-yr period inFebruary (Table 1). A comparison of temperaturesfrom the NWS and CMDL sites over the last 30 yrshowed no urban heat island effect within the limitsof the CMDL data record.

For the period 1953–98, 6 of the 10 yr of minimumsummer ice extent in the Arctic basin have occurredsince 1990, and the past decade has been character-ized by increased regional variability (Maslanik et al.1999). Gloersen et al. (1999) analyzed the spatial dis-tribution of trends in sea ice extent for the period1978–96 and found that annual ice concentrationshave decreased by between 3% and 9% for the regionof the Beaufort and Chukchi Seas, the trend increas-ing toward the west. Of specific relevance to thepresent study is the Barnett ice severity index (Drobotand Maslanik 2002, manuscript submitted to J. Cli-mate, hereafter DM). This is a combination of thedistance from Point Barrow northward to the ice edgeon 15 September, the number of days the entire searoute from the Bering Strait to Prudhoe Bay is ice free,and the length of the navigable season. The Barnettindex shows a significant trend toward lighter ice con-

ditions, but with considerable variability from year toyear (Fig. 3). Two key factors appear to be influenc-

Jan 0.0204 0.0176

Feb 0.0369* 0.0181

Mar 0.0204** 0.0120

Apr 0.0064 0.0134

May 0.0155** 0.0087

Jun 0.0119* 0.0057

Jul 0.0064 0.0064

Aug 0.0015 0.0092

Sep 0.0092 0.0093

Oct –0.0205 0.0150

Nov –0.0078 0.0184

Dec 0.0046 0.0145

Winter 0.0205** 0.0107

Spring 0.0141** 0.0076

Summer 0.0066 0.0052

Autumn –0.0064 0.0115

Annual 0.0087 0.0057

TABLE 1. Temperature trends (°C yr-1) andvariability in Barrow 1921–2001, by month,season, and annually.

Linear trend Std dev

*Significant at the 95% confidence level.

**Significant at the 90% confidence level.

FIG. 2. Record of Barrow temperatures since 1921 by season.

214 FEBRUARY 2004|

ing these changes in ice distribution. The first factoris the increase in surface air temperature describedabove. The second factor is a shift in arctic air circu-lation patterns (Maslanik et al. 1999; Lynch et al.2001). As a result, transport of thick, multiyear iceinto the southern Beaufort Sea during the precedingwinter and spring tends to be decreased, and relativelywarm air advected from arctic landmasses duringspring and summer melts some of the coastal sea iceand transports the remainder away from the shore

and toward the central ice pack or westward into theChukchi Sea.

Of particular consequence to the present study isthe climatology of extreme conditions that include cy-clones, high winds, and large fetch,1 which combineto produce coastal flooding and erosion at Barrow.McCabe et al. (2001) examined the frequency andintensity of winter extratropical cyclones in theNorthern Hemisphere for the period 1959–97 andfound a statistically significant increase in high-lati-

tude cyclone frequency and inten-sity. However, Curtis et al. (1998),who focused on Barrow and BarterIsland over the period 1949–98,found a decrease in the variability ofsea level pressure. This suggests thatintensity and/or frequency of cy-clones in this region of the Arctic hasdecreased. Curtis et al. (1998) sug-gest that this decrease in cyclonic ac-tivity is linked to a shift in the Pa-cific–North American (PNA)teleconnection pattern. The PNA isprimarily a winter pattern (Novem-ber through March). A positive PNAindex represents an amplified wavepattern with a strong Aleutian low,whereas a negative PNA representsa more zonally oriented pattern. ThePNA index had some tendency to-ward negative values in the early tomidseventies, and a tendency formore positive values in the late sev-enties through the late eighties, butshows no significant trend overallsince 1948.

To clarify the results of McCabe etal. (2001) and Curtis et al. (1998) forthis region, we analyze cyclone fre-quency and intensity in the Beaufort–Chukchi sector using the sameSerreze (1995) dataset that was ana-lyzed in the McCabe et al. (2001)study, updated to 2000. For thisanalysis, we restrict our domain tothe region bounded by 65°N to thesouth, 78°N to the north, 170°E to

1 Fetch is defined as the distance of openocean over which wind is blowing that de-termines the amplitude of wind-drivenocean waves.

FIG. 3. Barnett ice severity index from 1953 to 2001, calculated fol-lowing DM and shown as departures from the climatological mean.High numbers indicate light ice conditions.

Winter cyclone frequency –0.003 –0.33a

Winter cyclone intensity –0.009 –0.37b

Spring cyclone frequency –0.005 –0.44c

Spring cyclone intensity 0.017 0.08

Summer cyclone frequency –0.019 0.13

Summer cyclone intensity 0.048 –0.33a

Autumn cyclone frequency 0.000 –0.32a

Autumn cyclone intensity 0.023 0.16

Annual cyclone frequency –0.015 –0.41b

TABLE 2. Forty-year trends in cyclone frequency (number) andintensity (Pa m-----2; significant in bold), together with thecorrelation of these trends with the PNA teleconnectionpattern index, as defined by Wallace and Gutzler (1981).

aSignificant at the 97.5% level.bSignificant at the 99% level.cSignificant at the 99.5% level.

Linear trend Correlation to PNA

215FEBRUARY 2004AMERICAN METEOROLOGICAL SOCIETY |

the west, and 230°E to the east.Over the 41-yr time period 1959–2000, cyclone frequency and in-tensity in the Beaufort–Chukchisector have not significantlychanged on an annual basis(Table 2). Further, the proportionof storms that occur in each sea-son has not shifted significantly(Fig. 4a). The only significantchange is the intensity of summercyclones (Table 2), which showsan average increase of 0.5 Pa m-2

per decade. This increase in sum-mer cyclone intensity (shown alsoin Fig. 4b) is manifest as an abruptshift in cyclone intensity in 1971,an increase that has continuedover the following 30 yr.

From this analysis, it is clearthat trends in cyclone frequencyand intensity are generally insig-nificant for this region and hencemay not be strongly related toindicators of large-scale circula-tion, such as the Arctic Oscillation(AO) or Northern Hemisphereannular mode (Thompson andWallace 2001), which have shownstrong trends over the last severaldecades. Conversely, Stone (1997)found that warmer conditions inBarrow prevail when cyclonic ac-tivity increases in the North Pa-cific, resulting in advection ofclouds northward (see also Har-ris and Kahl 1994). If outflowfrom anticyclones centered in theBeaufort Sea is dominant, colder,drier air influences the North Slope. Stone (1997)therefore argues that Barrow may be highly suscep-tible to changes in central Arctic circulation. How-ever, analysis of these trends in the context of the AOshowed no relationship between cyclones in the Beau-fort–Chukchi region and this large-scale telecon-nection pattern, and hence the strong trend toward apositive AO index over the past 30 yr is not reflectedin cyclone frequency and intensity.

As expected, winter cyclone frequency and inten-sity is somewhat linked to the PNA index, as is springcyclone frequency (Table 2). A negative correlationindicates more frequent and stronger storms in theBeaufort–Chukchi region when the Northern Hemi-

sphere flow is strongly zonal. More surprising is thefact that summer cyclone intensity, the strongesttrend, is correlated with this winter pattern. The workof Zhang et al. (1998) suggests that this may be dueto the persistence of sea surface temperature anoma-lies from one season to the next.

The more intense storms during the summer in theBeaufort–Chukchi region, when sea ice is far fromshore, may cause more damage from flooding anderosion. However, not all high-wind events in Barroware associated with strong cyclones. High-wind eventsat Barrow can be linked to a range of different syn-optic conditions that may be characterized most gen-erally by a Beaufort–Chukchi cyclone, an extremely

FIG. 4. (a) Proportion of storms that occur in each season (winter, spring,summer and autumn) by decade. (b) Time series of summer cyclone in-tensity in the Beaufort–Chukchi region from 1959 to 2000.

b)

a)

216 FEBRUARY 2004|

large, strong Aleutian cyclone, or a strong Beaufort–Chukchi ridge (analysis of these patterns is ongoing).

We analyzed the occurrence of high-wind eventsin Barrow over the past 55 yr, using available NWSobservations. Analysis of CMDL wind data revealedsome differences, but did not change the conclusions,and hence CMDL wind data are not shown. The con-clusions are somewhat confounded by instrument andmeasurement changes over the time period in ques-tion, and this caveat should be borne in mind. Averagewinds for each season displayed significant increas-ing trends in winter (at the 99% confidence level) andspring (at the 95% confidence level), results that aredominated by strong trends in December, January,and March (Table 3). However, there is little appar-ent linear trend in the highest winds (Fig. 5a) in anyseason. The increase in summer cyclone intensity isreflected in a small increase in strong winds duringthe summer months. The storm of August 2000 (dis-cussed in Sidebar 3) is a clear outlier in the summerrecord, whereas the very damaging storm of October1963 does not represent such a strong departure fromautumn storm intensity.

From the record of highest winds, and throughinterviews with Barrow residents, we extracted a sub-set of significant high-wind events (Fig. 5b). Thisrecord of high-wind events is correlated significantlywith the AO (correlation coefficient of 0.3) and thePNA pattern (correlation coefficient of 0.45), eventhough cyclones in this region and average winds inBarrow are not correlated with either of these large-scale teleconnection patterns. Only three summerstorms are evident, in 1950, 1993, and 2000. Clearlya longer record will be necessary to determine if sum-mer storm intensity in Barrow is truly on the rise.Dramatically apparent in the wind event record,though, is a relatively quiet period in all seasons fromthe late sixties to the early eighties, with the lowestperiod in the late seventies (Fig. 5c).

Flooding at Barrow is caused by a combination ofstorm surges and waves (which depend on atmo-spheric pressure, sustained wind, and fetch). A gen-eral rule of thumb used by the NWS is that a 30-hPadecrease in surface pressure will cause a 0.3-m rise insea level. Hence, this effect on sea level is relativelysmall in the weak cyclones affecting the North Slopeof Alaska. At Barrow, winds from the west causestorm surge, whereas winds from the east cause a low-ering of sea level (Hume and Schalk 1967). A com-mon feature of many damaging storms is the presenceof large open-water fetch, which limits the dampingof waves and wind-driven storm surge by ice. Hence,the season of greatest potential danger from floodingis late summer and early autumn, when the sea ice ex-tent is a minimum and the open water fetch is thegreatest. Figure 5b shows the time line of significantstorms and their average fetch (calculated as the dis-tance in the direction from which the wind is blow-ing to the 50% ice concentration line). Of the threesummer storms, the storm of 1950 occurred when theice was completely in, and hence there was zero fetch.The storms of 1993 and 2000 both exhibited an aver-age fetch of 800 to 900 km over the lifetime of thestorm. Other storms exhibiting nonzero fetch, rang-ing between 560 km for the October 1963 storm to1200 km for the October 1977 storm, all occurred inthe months of September and October. Autumnstorms have historically been most feared and pre-pared for by Barrow residents.

Greatest fetch and strongest winds are not suffi-cient conditions to predict flooding, erosion, and theresultant damage to the town of Barrow. The stormsof October 1963 and August 2000 had comparablewinds from similar directions, comparable fetch, andsimilar duration, and yet the flooding and damage in1963 were much greater than in 2000. Moreover,

Jan 0.0151a 0.0073

Feb 0.0120 0.0082

Mar 0.0099b 0.0059

Apr 0.0091 0.0062

May 0.0067 0.0065

Jun 0.0018 0.0046

Jul 0.0086b 0.0046

Aug 0.0056 0.0058

Sep 0.0048 0.0056

Oct 0.0018 0.0076

Nov –0.0042 0.0085

Dec 0.0173a 0.0085

Winter 0.0158c 0.0056

Spring 0.0086a 0.0043

Summer 0.0047 0.0035

Autumn 0.0008 0.0047

Annual 0.0047 0.0035

TABLE 3. Average wind speed trends (m s-----1 yr-----1)and variability in Barrow 1921–2001 by month,season, and annually.

Linear trend Std dev

aSignificant at the 95% confidence level.bSignificant at the 90% confidence level.cSignificant at the 99% confidence level.

217FEBRUARY 2004AMERICAN METEOROLOGICAL SOCIETY |

more people and property were ex-posed in 2000 because of growth anddevelopment in Barrow over the in-tervening decades (Fig. 6). A majortask of this project is to determinethe necessary and sufficient condi-tions for flooding and erosion.

REGIONAL IMPACTS OF IN-TENSE STORMS AND COM-MUNITY RESPONSES. The im-pacts of intense storms on Barrowand the immediate vicinity rangefrom immediate impacts, such aswind damage, flooding, overriding ofthe shore by sea ice (known as ice-push or “ivu” events), to longer-termimpacts on the environment. Dam-age to the built environment, includ-ing buildings, roads, boat landings,airfields, utilities, and equipment, isan important impact. The NSB RiskManagement Department has notedthat NSB property is fully insuredand that damages from recent stormshave been relatively small (P.Patterson 2002, personal communi-cation). The impacts of flooding canbe diverse and widespread. Saltwaterintrusion into freshwater suppliescan contaminate drinking water for

b)

a)

c)

FIG. 5. Analysis of winds at Barrow.(a) Highest wind events by seasonfor every year since 1945, showingthe highest daily average windspeed reported for the season(blue); the highest sustained windspeed reported for the season(purple; 1965–79, 1984–2001); thehighest observation reported forthe season, which is a good esti-mate for sustained wind (pink;1945–64, 1980–83); and the highestwind gust reported for the season(orange; 1976–2001). (b) Time lineof extreme wind events (as definedby a peak wind gust greater than24 m s-----1, a sustained wind speedgreater than 20 m s-----1, or an aver-age daily wind speed greater than14 m s-----1, depending on data avail-ability) and associated fetch. (c)Ten-year moving sum of high windevents (number of events).

218 FEBRUARY 2004|

long time periods (see Sidebar 3). Flooding of septictanks can cause a significant public health hazard,particularly in locations where wells are relied upon

for freshwater supply. Airport flooding can isolateoutlying communities for several days. Barrow usesbarges for regular supplies but receives a variety ofgoods by air and sends medical supplies, food, and fuelby air to outlying communities. Further, if floodwa-ters breach the sewage lagoon or buried fuel dumps,significant pollution could occur. Residents are al-

THE EXTREME STORMS OF OCTOBER1963 AND AUGUST 2000The most severe recorded storm along theBeaufort Sea coast occurred from 3 to 5 October1963 and caused extensive damage at Barrow(Hume and Schalk 1967). Sustained winds of25 m s-----1 (based on the highest observation) wereobserved, with gusts unofficially reported at33 m s-----1. Water moved inland 122 m from theshore, and large chunks of sea ice were washedinland to about 4.6 m. Waves were estimated tobe 3 m high with a storm surge of 3 m (shown as ablue line in Fig. 6). Coastal areas in the vicinity ofthe Naval Arctic Research Laboratory wereflooded for distances up to 1200 m, and shorelinesbacked by lakes were flooded up to 1600 m inland,contaminating local drinking water. Damageincluded the destruction of 10 buildings, whilebluffs southwest of Barrow were eroded 3 m, andthe shoreline retreated 18 m. The damageestimate for the 1963 storm at Barrow was$3.25 M (about $19 M in 2001 dollars). Since 1963,additional development that includes undergroundutilities, a hotel, and fuel storage tanks would havesubstantially increased the losses associated withthis storm (Fig. 6).

More recently, the storm of 10–11 August 2000produced record wind gusts of 29 m s-----1 and 2-minsustained winds of 25 m s-----1, according to the NWSanemometer. This anemometer may have“pegged out”—a maximum wind of 33 m s-----1 mayhave been observed at the CMDL weather station(D. Endres 2002, personal communication)—thepeak 1-s wind at CMD was 31.3 m s-----1. Emergencymanagement teams had insufficient notice tomobilize heavy equipment to build up temporaryberms for protection along the coast. The stormeroded the beach to within 100 m of a mainfreshwater and sewage lift (or pumping) station ofthe utilidor, sank a dredge barge, washed out aboat ramp, and removed roofs from 40 buildings.A tidal gauge located in Prudhoe Bay showed thatthe storm surge was within 6 cm of the 100-yrevent (A. Morkill 2002, personal communication).The final cost of this storm, mostly damage to thedredge barge, was $7.7 M. Storm costs to theDepartment of Municipal Services, which main-tains and repairs roads and beaches among otherthings, was $829,772.

Importantly, there have been no reports ofdeaths or even serious injury due to the 1963 or2000 storms.

FIG. 6. Map of the Barrow area from 1964 and aerialphotograph of Barrow from 1997. Note the new roadsacross Isatquaq Lagoon to access Browerville. The topmap, which is preliminary, was created from a 14 Jul1964 1:10, 000 aerial photograph by the U.S. ArmyCorps of Engineers Cold Regions Research and Engi-neering Laboratory (CRREL). The 3.5-m flood line waschosen to emulate what flooding from the 1963 stormmay have looked like in Barrow. There are no recordsof where the flooding actually took place in Barrow. Thelocation map digital elevation model (DEM) was devel-oped by W. Manley from the 1955 1:63,360 USGS to-pographical map. The bottom aerial photograph wasborrowed from the NSB and was acquired by AeroMapon 16 Jul 1997.

219FEBRUARY 2004AMERICAN METEOROLOGICAL SOCIETY |

ready concerned about persistent organic pollutantsin their environment. These pollutants in turn affectmany species that people rely upon for subsistencehunting (e.g., Hurwich and Chary 2000).

There is a perception that dredging for militaryconstruction in the 1940s may have initiated coastalerosion problems. A dependable and comprehensiveunderstanding of such damages will be constructedfrom multiple sources, including NSB documents andinterviews, during this project.

There are no surface weather stations to the north-west of Barrow, the direction from which most severestorms come. Satellite imagery can alert forecastersof an incoming storm but can give only limited in-formation as to its strength. In any case, the NationalWeather Service does not forecast outside of histori-cal extremes, and the August 2000 storm set officialrecords for wind strength. Because the NSB had tomobilize equipment rapidly, private companies’equipment was used without contracts, contributingto the high costs. Another issue was the instructionthat people go home during the storm—subsequentreports suggest that people would have been safer ifthey had stayed where they were rather than attempt-ing to travel in the midst of the storm. Finally, therewas concern regarding the ability to evacuate peoplefrom the former Naval Arctic Research Laboratoryfacilities located along the coast, northwest of Barrow,if the coastal road was to be flooded or washed out. Itis important to note that the standard procedure out-lined in the Disaster Management Handbook for theNSB was followed during the August 2000 storm.Hence, the evidence suggests that minimizing risk anddamage from such storms depends critically, thoughnot soley, on storm prediction.

Development by the NSB Planning Department ofa coastal zone management plan began in the late1970s, was adopted in 1988 by the NSB Assembly, andis still in place. However, this plan is directed prima-rily toward protecting the environment amidst theheavy concentration of oil industry activities on theNorth Slope rather than vulnerability to extremeweather events or the effects of ongoing coastal ero-sion. Federally subsidized insurance for propertyalong the coastline was considered in 1997–98 but wasabandoned in the face of federal zoning requirements,high costs, and local opposition.

To minimize vulnerability to flooding and erosion,the NSB initiated a beach nourishment program inSeptember 1986, after two storms (Fig. 5b) drew at-tention to erosion problems. The dredge Qayuutaqwas sunk by the storm of August 2000, discussedabove, effectively ending the program. The informal

evaluations we have received to date from Barrowresidents note that the program may have maintainedthe shoreline, but at a higher overall cost than origi-nally planned. A $3–4 million Army Corps of Engi-neers feasibility study for storm damage reduction atBarrow was proposed, and supported by the NSBmayor, in 2001. The options being considered includeadding width to the beach (15 or 30 m), raising andperhaps hardening the beach road along 8 km ofcoastline, and adding a concrete mattress revetmentto the seaward slope of the road and bluffs. This is arelatively expensive approach and relies strongly onthe presence of appropriate material in nearby ElsonLagoon for building up the beach. Unless alternativefunds are found, options will be constrained by con-tinued decline in the borough property tax base inPrudhoe Bay (NSB 1999).

Improved understanding of the local climate vari-ability and the interplay of coastal geography withatmosphere, ocean, and sea ice dynamics can be usedto support the local decision-making processes. How-ever, vulnerability is a function not only of climatevariability and natural geography, but also of humangeography. Community vulnerability has increasedover the decades with community modernization,which includes increased structures, electricity andtelephone lines above ground, the underground wa-ter and sewage utilidor, and natural gas pipelines. Inaddition, the diversity of storms and their impacts onBarrow in the past implies that policies to reducecommunity vulnerabilities cannot anticipate in suffi-cient detail major storms of the future. Hence, resil-ience is necessary to expedite recovery from impactsof a severe storm, and flexibility is necessary to accom-modate what is learned from each storm to reducevulnerabilities over the long term.

METHODOLOGY FOR INTEGRATEDASSESSMENT. The process of integrated assess-ment is to analyze, synthesize, and communicateknowledge from a range of scientific disciplines inorder to provide useful information for policy(Rotmans and van Asselt 1996). As such, the processis highly participatory in nature. Our strategy for in-tegrated assessment has three main elements:

1) Description, analysis, and projection of climatevariability, including what has already occurred,why, and what may be reasonably expected for thenear future, on decadal time scales. This researchwill include the application of models to arrive atplausible scenarios of climate variations, use of tra-ditional knowledge, records, and remotely sensed

220 FEBRUARY 2004|

data to document location and rate of shorelinechanges, and, where appropriate, application ofmodels and data analysis to arrive at plausibleimpacts and response scenarios relevant to key en-vironmental issues affecting the coastal zone.

2) Analysis of the regional impacts of variability. Thisanalysis is supported by development of a regionalinformation system using a geographic informa-tion system (GIS) to integrate the climate and im-pacts data. We have already supplied data andanalysis products to the NSB GIS department andwill continue to do so through the life of theproject. This element of the strategy integratesscientific and local knowledge and informationinto an initial conceptual model of the natural andpolicy context. Relevant data include contempo-rary climate and geographical and human infra-structure maps.

3) Documentation and analysis of community re-sponses, including efforts to enhance North Slopedecision making by conducting requested analy-ses or providing requested information. For ex-ample, the synoptic climatology of the Barrowarea has been requested for use as a training toolby the Barrow National Weather Service office.

The policy research is being organized and coor-dinated through a continuing decision seminar inBarrow, with members of the community and projectpersonnel as core participants. A decision seminar(Lasswell and McDougal 1992) is designed to evalu-ate and integrate knowledge pertinent to a givenpolicy problem in order to articulate possible im-provements in policy decisions and decision pro-cesses. An important component of the research strat-egy is to engage the participants in evaluating andcorrecting the conceptual model of the policy contextand then using it to redirect further research. Forexample, what parts of the initial model of the con-text are problematic from the standpoint of scientificor local knowledge, or from a policy standpoint?What are the most important problems? Some under-standings of extreme wind events, erosion, and flood-ing in our initial model almost certainly will be calledinto question on empirical grounds when exposed tolocal knowledge. Such questions are candidates forfurther research.

This iterative strategy factors the community’soverall concern about climate change and variabilityinto a shifting focus on smaller, more tractable con-cerns. Meanwhile, the initial model reinforces com-munity interest and participation, leaves flexibility toredirect scientific and policy research as new priori-

ties emerge, and allows for interactions over a longenough time to build shared understandings andmutual trust among core participants and others withdiverse disciplinary, cultural, and policy perspectives.New priorities can emerge through improvements toour conceptual model; for example, an understand-ing that the distance to the ice edge in the Beaufort–Chukchi Seas region has little impact on cyclone in-tensity (Lynch et al. 2003). Alternatively, newpriorities can arise from evolving community con-cerns, such as the importance of precipitation phase,which arose during discussion at the last decisionseminar. At each iteration, we expect to show cumu-lative progress in these respects. At the end of theproject, success will be gauged by improvements inpolicy responses to climate impacts on the NorthSlope and by advances in knowledge that will informfuture scientific and policy research.

ACKNOWLEDGMENTS. Thanks to Elizabeth Cassanoand Mathew Rothstein for the extensive data analysis thatwent into the conclusions on climate variability reportedin this paper. William Manley and Leanne Lestak as-sembled geographic data and prepared the maps. GlennSheehan and Anne Jensen of the Ukpeagvik Inupiat Cor-poration and the Barrow Arctic Science Consortium pro-vided invaluable assistance in planning community meet-ings, arranging individual discussions with NSB residents,and providing insights into community factors and issues.The people of Barrow have been tremendously generouswith their time and information—particular thanks are dueto Mike Aamodt, Marie Carroll, Richard Glenn, and Ken-neth Toovak. This work is supported by NSF Grants OPP-0100120 and OPP-9732461 and through the InternationalArctic Research Center by the National Science Founda-tion under Agreement OPP-0002239 and the NationalOceanic and Atmospheric Administration under Agree-ment NA67RJO147.

REFERENCESBerger, T. R., 1985: Village Journey: The Report of the

Alaska Native Review Commission. Hill and Wang,New York, New York, 202 pp.

BESIS, 1997: The impacts of global climate change in theBering Sea region. Bering Sea Impacts Study (BESIS),International Arctic Science Committee, Oslo, Nor-way, 41 pp.

Brewster, K., 1997: Native contribution to Arctic scienceat Barrow, Alaska. Arctic, 50, 277–288.

Chapman, W. L., and J. E. Walsh, 1993: Recent varia-tions of sea ice and air temperature in high latitudes.Bull. Amer. Meteor. Soc, 74, 33–47.

221FEBRUARY 2004AMERICAN METEOROLOGICAL SOCIETY |

Cochran, P., and Coauthors, 1999: Traditional knowl-edge and radionuclides. Alaska Native Science Com-mission Rep.

Cohen, S. J., 1997: What if and so what in northwestCanada: Could climate change make a difference to thefuture of the Mackenzie basin? Arctic, 50, 293–307.

Curtis, J., G. Wendler, R. Stone, and E. Dutton, 1998:Precipitation decrease in the western Arctic, withspecial emphasis on Barrow and Barter Island.Alaska. Int. J. Climatol., 18, 1687–1707.

Drobot, S. D., and J. A. Maslanik, 2004: Interannual vari-ability in summer Beaufort Sea ice conditions: Rela-tionship to spring and summer surface and atmo-spheric variability. J. Climate, in press.

Gloersen, P., C. L. Parkinson, D. J. Cavalieri, J. C. Comiso,and H. J. Zwally, 1999: Spatial distribution of trendsand seasonality in the hemispheric sea ice covers:1978–1996. J. Geophys. Res., 104 (C9), 20 827–20 835.

Griggs, M. Noguer, P. J. van der Linden, and D. Xiaosu,Eds., 2001: Climate Change 2001: The Scientific Basis:Contribution of Working Group I to the Third Assess-ment Report of the Intergovernmental Panel on ClimateChange (IPCC). Cambridge University Press, 944 pp.

Gurney, K. R., 1998: Evidence for increasing ultravioletirradiance at Point Barrow, Alaska. Geophys. Res.Lett., 25 (6), 903–906.

Harris, J. M., and J. D. W. Kahl, 1994: Analysis of 10-dayisentropic flow patterns from Barrow, Alaska: 1985–1992. J. Geophys. Res., 99, 25 845–25 855.

Hinkel, K. M., F. Paetzold, F. E. Nelson, and J. G.Bockheim, 2001: Patterns of soil temperature andmoisture in the active layer and upper permafrost atBarrow, Alaska: 1993–1999. Global Planet. Change,29, 293–309.

Hume, J. D., and M. Schalk, 1967: Shoreline processesnear Barrow, Alaska: A comparison of the normaland the catastrophic. Arctic, 20, 86–103.

Hurwich, E. M., and L. K. Chary, 2000: Persistent or-ganic pollutants in Alaska. Circumpolar Conserva-tion Union Rep., 77 pp.

Jones, P. D. 1994: Hemispheric surface air temperaturevariations: A reanalysis and an update to 1993. J. Cli-mate, 7, 1794–1802.

Kaleak, J., Sr., 1993: Introduction to Taking Control: TheStory of Self-Determination in the Arctic by W. Hess.North Slope Borough Public Information Division,148 pp.

Korsmo, F. L., 1999: Indigenous peoples and the con-duct of the science in the north. 15th Biennial Conf.of the Association for Canadian Studies in the UnitedStates, Pittsburgh, PA, ACSUS.

Lasswell, H. D., and M. S. McDougal, 1992: Jurispru-dence for a Free Society: Studies in Law, Science and

Policy. 2 vols. New Haven Press and MartinusNijhoff, 1612 pp.

Lynch, A. H., J. A. Maslanik, and W. Wu, 2001: Mecha-nisms in the development of anomalous sea ice ex-tent in the western Arctic: A case study. J. Geophys.Res., 106, 28 097–28 106.

——, E. N. Cassano, J. J. Cassano, and L. Lestak, 2003:Case studies of high-wind events in Barrow, Alaska:Climatological context and development processes.Mon. Wea. Rev., 131, 719–732.

Maslanik, J. A., M. C. Serreze, and T. Agnew, 1999: Onthe record reduction in western Arctic sea-ice coverin 1998. Geophys. Res. Lett., 26, 1905–1908.

McCabe, G. J., M. P. Clark, and M. C. Serreze, 2001:Trends in Northern Hemisphere surface cyclone fre-quency and intensity. J. Climate, 14, 2763–2768.

NRC, 1994. Environmental Information for Outer Conti-nental Shelf Oil and Gas Decisions in Alaska. NationalResearch Council, National Academy Press, 254 pp.

NSB, 1999: Comprehensive annual financial report ofthe North Slope Borough, Alaska, July 1, 1998–June30, 1999. North Slope Borough Department ofAdministration and Finance, November 1999, 114 pp.

Overpeck, J., and Coauthors, 1997: Arctic environmen-tal change of the last four centuries. Science, 278,1251–1256.

Rotmans, J., and M. van Asselt, 1996: Integrated assess-ment: A growing child on its way to maturity. Cli-matic Change, 34, 327–336.

Serreze, M. C., 1995: Climatological aspects of cyclonedevelopment and decay in the Arctic. Atmos. Ocean,33, 1–23.

——, and Coauthors, 2000: Observational evidence of re-cent change in the northern high-latitude environ-ment. Climatic Change, 46, 159–207.

Stafford, J. M., G. Wendler, and J. Curtis, 2000: Tem-perature and precipitation of Alaska: 50 year trendanalysis. Theor. Appl. Climatol., 67, 33–44.

Stone, R. S., 1997: Variations in western Arctic tempera-tures in response to cloud radiative and synoptic-scale influences. J. Geophys. Res., 102, 21 769–21 776.

Thompson, D. W. J., and J. M. Wallace, 2001: Regionalclimate impacts of the Northern Hemisphere annu-lar mode. Science, 293, 85–89.

Tynan, C. T., and D. P. DeMaster, 1997: Observationsand predictions of Arctic climate change: Potentialeffects on marine mammals. Arctic, 50, 308–322.

Wallace, J. M., and D. S. Gutzler, 1981: Teleconnectionsin the geopotential height field during the NorthernHemisphere winter. Mon. Wea. Rev., 109, 784–812.

Zhang, Y., J. R. Norris, and J. M. Wallace, 1998: Season-ality of large-scale atmosphere–ocean interactionover the North Pacific. J. Climate, 11, 2473–2481.