I N T E R A G E N C Y A R C T I C R E S E A R C H P O L I C Y C O M M I T T E E

A R C T I C R E S E A R C HVOLUME 19 SPRING/SUMMER 2005

O F T H E U N I T E D S T A T E S

This document has been archived.

Cover The medusa Sminthia arctica, the most common species identified by video in the transition between the Pacific andAtlantic water layers of the Canada Basin, measured during an expedition to the Arctic Ocean in 2002. This gelati-nous species is less than a millimeter in width.

Aboutthe

Journal

The journal Arctic Research of the UnitedStates is for people and organizations interestedin learning about U.S. Government-financedArctic research activities. It is published semi-annually (spring and fall) by the National ScienceFoundation on behalf of the Interagency ArcticResearch Policy Committee (IARPC). TheInteragency Committee was authorized under theArctic Research and Policy Act (ARPA) of 1984(PL 98-373) and established by Executive Order12501 (January 28, 1985). Publication of the jour-nal has been approved by the Office of Manage-ment and Budget.

Arctic Research contains• Reports on current and planned U.S. Govern-

ment-sponsored research in the Arctic;• Reports of IARPC meetings; and• Summaries of other current and planned

Arctic research, including that of the State ofAlaska, local governments, the private sec-tor, and other nations.

Arctic Research is aimed at national and inter-national audiences of government officials, scien-tists, engineers, educators, private and publicgroups, and residents of the Arctic. The emphasisis on summary and survey articles covering U.S.Government-sponsored or -funded research rath-er than on technical reports, and the articles areintended to be comprehensible to a nontechnicalaudience. Although the articles go through thenormal editorial process, manuscripts are not

refereed for scientific content or merit since thejournal is not intended as a means of reportingscientific research. Articles are generally invitedand are reviewed by agency staffs and others asappropriate.

As indicated in the U.S. Arctic Research Plan,research is defined differently by different agen-cies. It may include basic and applied research,monitoring efforts, and other information-gatheringactivities. The definition of Arctic according to theARPA is “all United States and foreign territorynorth of the Arctic Circle and all United Statesterritory north and west of the boundary formedby the Porcupine, Yukon, and Kuskokwim Rivers;all contiguous seas, including the Arctic Oceanand the Beaufort, Bering, and Chukchi Seas; andthe Aleutian chain.” Areas outside of the bound-ary are discussed in the journal when consideredrelevant to the broader scope of Arctic research.

Issues of the journal will report on Arctictopics and activities. Included will be reports ofconferences and workshops, university-basedresearch and activities of state and local govern-ments and public, private and resident organiza-tions. Unsolicited nontechnical reports onresearch and related activities are welcome.

Address correspondence to Editor, ArcticResearch, Arctic Research and Policy Staff,Office of Polar Programs, National ScienceFoundation, 4201 Wilson Boulevard, Arlington,VA 22230.

1

O F T H E U N I T E D S T A T E S

SPRING/SUMMER 2005

A R C T I C R E S E A R C HVOLUME 19

SPECIAL ISSUE ON THENATIONAL OCEANIC AND ATMOSPHERIC

ADMINISTRATION’S RESEARCHIN THE ARCTIC

The Role of the National Oceanic and AtmosphericAdministration in the Arctic Region ........................... 2

NOAA’s Arctic Ocean Exploration Program ................. 3

Arctic Sea Ice and Ocean Observations ..................... 11

Correlated Declines in Pacific Arctic Snowand Sea Ice Cover ................................................... 18

Acidifying Pollutants, Arctic Haze, andAcidification in the Arctic ......................................... 26

The Barrow Atmospheric Baseline Observatory ......... 34

NOAA and the Alaska Ocean Observing System ....... 41

Marine Mammals in the Bering/Chukchi Sea .............. 50

Ocean Climate Changes and theSteller Sea Lion Decline ........................................... 54

Status of Alaska Groundfish Stocks ............................. 64

Status of Alaska’s Salmon Fisheries ............................. 66

Russian–American Long-term Census of the Arctic ... 73

On the Creation of Environmental Data Setsfor the Arctic Region ................................................ 77

Interagency Arctic Research PolicyCommittee Staff ....................................................... 93

INTERAGENCY ARCTIC RESEARCHPOLICY COMMITTEE

DEPARTMENT OF AGRICULTURE

DEPARTMENT OF COMMERCE

DEPARTMENT OF DEFENSE

DEPARTMENT OF ENERGY

DEPARTMENT OF HEALTH AND HUMAN SERVICES

DEPARTMENT OF HOMELAND SECURITY

DEPARTMENT OF THE INTERIOR

DEPARTMENT OF STATE

DEPARTMENT OF TRANSPORTATION

ENVIRONMENTAL PROTECTION AGENCY

NATIONAL AERONAUTICS AND SPACE ADMINISTRATION

NATIONAL SCIENCE FOUNDATION

SMITHSONIAN INSTITUTION

OFFICE OF MANAGEMENT AND BUDGET

OFFICE OF SCIENCE AND TECHNOLOGY POLICY

Managing Editorial Committee

Charles E. Myers, National Science Foundation—Editor

John Haugh, Bureau of Land Management—Associate Editor

David W. Cate, Cold Regions Research and EngineeringLaboratory—Consulting Editor

Editing and production: Cold Regions Research andEngineering Laboratory, Hanover, New Hampshire

2

This issue of the Arctic Research of the UnitedStates profiles Arctic research carried out by theNational Oceanic and Atmospheric Administration(NOAA). NOAA, an agency of the Department ofCommerce, has four mission goals:

• To protect, restore, and manage resources inthe oceans and the atmosphere;

• To understand climate change and variability;• To fulfill weather and water information

needs; and• To support the commerce and transportation

needs of the United States.The breadth of science carried out at NOAA is

spread between the NOAA National Weather Ser-vice (NWS), the primary source of weather data,forecasts, and warnings for the U.S.; the NOAAOcean Service (NOS), responsible for the observa-tion, measurement, assessment, and managementof the nation’s vast coastal and ocean areas; theNational Environmental Satellite and Data Informa-tion Services (NESDIS), which provides timelyaccess to global environmental data from satellitesand other sources to promote, protect, andenhance the nation’s economy, security, environ-ment, and quality of life; NOAA’s National MarineFisheries Service (NMFS), which is dedicated tothe stewardship of living marine resources throughscience-based ecosystem management; andNOAA’s Oceanic and Atmospheric Research(OAR), which provides unbiased information tobetter manage the complex systems of the atmo-sphere, the climate, and ocean and coastalresources.

The Role of the National Oceanic and AtmosphericAdministration in the Arctic Region

This article wasprepared by Kathleen

Crane, of NOAA’sArctic Research Office.

This issue of Arctic Research of the UnitedStates presents to the Arctic community slices ofNOAA’s Arctic research life. Articles cover Arcticatmospheric, ocean, ice, and marine life research,in particular, the topics of Arctic Haze, the BarrowAtmospheric Baseline Observatory, declines inPacific Arctic snow and sea ice cover, Arctic seaice and ocean observations, the Alaska OceanObserving System, the Arctic Ocean explorationprogram, the Russian–American Long-term Cen-sus of the Arctic, ocean climate changes andthe Steller sea lion decline, the status of Alaskagroundfish stocks and salmon fisheries assess-ments, and the status of marine mammals in theBering/Chukchi Seas. In addition, there are articlesfrom individuals who receive external researchfunding from NOAA.

To wrap up these timely Arctic research topics,the National Snow and Ice Data Center presents anarticle on the creation of environmental data setsfor the Arctic. Calling on the history of formerInternational Polar Years, this article raises criticalquestions about the future role of data centersduring the upcoming International Polar Year (IPY).

As the United States rapidly approaches theIPY of 2007–2008, NOAA is poised to make impor-tant, innovative, and far-reaching inroads into thefurther exploration and understanding of the Arc-tic region and to help create a legacy of polarresearch and observational platforms vitallyimportant for the present-day and future under-standing of the Arctic’s influence on the climateof the earth.

3

NOAA’s Arctic Ocean Exploration Program

This article was preparedby Kathleen Crane, of

NOAA’s Arctic ResearchOffice; Jeremy Potter, ofNOAA’s Office of OceanExploration; and RussellHopcroft, of the Univer-sity of Alaska Fairbanks.

The Arctic Ocean is largely unexplored, espe-cially those aspects not visible to the human eyefrom a surface ship or to a satellite sensor. Somedata collected for national defense purposes arenow available, and they provide a better pictureof the bathymetry and circulation of the ArcticOcean. In addition, the International Arctic BuoyProgramme contributes data on ice drift trajecto-ries and surface meteorology. While there havebeen intensive research campaigns, such as theSurface Heat Budget of the Arctic (SHEBA)program that lasted about a year, the deeperportions of the Arctic Ocean and areas far fromland-based facilities remain mostly unmapped,from the seafloor to the life and the currentswithin the sea.

Exploration ApproachNOAA and its domestic and international part-

ners decided in 2001 to undertake expeditions ofexploration and discovery in the Arctic Ocean,with initial emphasis on missions of discoverynear the Pacific Gateway to the Arctic, the CanadaBasin, and the Mendeleev Basin. Additionalemphasis has grown to include a census of marinelife plus the mapping and imaging of previouslyunmapped seafloor, including the continentalshelves, major ridges, and deeper basins.

The original exploration plan set forth in 2001stated that the most appropriate autonomousunderwater vehicles (AUVs), remotely operatedvehicles (ROVs), and acoustic technologies would

Locations of NOAA’s past,present, and proposed

Arctic Ocean ExplorationMissions, 2002–2008.

4

be employed to complement more traditional wire-line sampling of the seafloor and water column.The use of aircraft and buoys was also proposedto extend the spatial and temporal range of theexpeditions. Steps were also planned that wouldcontribute to the search for new products from thesea. Since the summer of 2002, four Arctic Oceanexploration expeditions have taken place under theguidance of NOAA’s Ocean Exploration Program.

The Hidden Ocean:Canada Basin

In 2002 the first of these expeditions focusedon exploring the deep Canada Basin, located inthe Arctic Ocean, together with the Department

of Fisheries and Oceans, Canada, on the Louis St.Laurent, a Canadian Coast Guard icebreaker. Inaddition, both China and Japan carried out signifi-cant programs onboard the vessel.

This international team of 50 scientists from theU.S., Canada, China, and Japan participated in acollaborative effort to explore the frigid surfaceto the depths of the Canada Basin. Because ofthe region’s heavy year-round ice cover, researchin this region had been extremely limited. Thisexpedition was the first of its kind. With the aidof an ROV called the Ocean Explorer (speciallydesigned to operate under ice and at great depth),scientists examined the hidden world of life in theextreme Arctic conditions.

The Canada Basin may be geologically isolatedfrom all the other deep-sea basins in the world’soceans because there may be no deep pathwaysconnecting it to other regions of the Arctic. Thislack of connectivity could severely limit the ex-change of biota from the Atlantic and the otherparts of the Arctic deep sea into the Canada Basin.What are the consequences of this biological iso-lation? Perhaps relict species of life still thrive inthis remote basin. Perhaps new species haveevolved here, isolated by the millennia. The expe-dition’s scientists were eager to bring back sam-ples and image the life in this ocean, from withinthe ice to the deep seafloor below.

From intricate microscopic organisms found inthe brine channels that run through the ice to thecreatures that make the sea bottom their home, thescience team studied the relationships betweensympagic (ice-associated), pelagic (water-column),and benthic (bottom-dwelling) communities. Theyinvestigated the manner in which food energy istransferred from the surface of the ice, through thewater column, and to the bottom of this harshenvironment. In addition, they analyzed bottomsediments to determine their chemical makeup, aswell as help reconstruct the climatic history andpaleo-environmental events that formed the region.

Ice cores were taken at a total of ten stations(on the Louis St. Laurent and the Xue Long, aChinese research vessel) and analyzed for icetemperature, salinity, chlorophyll, and ice faunalabundances. Ice fauna were mainly located inthe bottom 10 cm of the sea ice. These life formsincluded turbellarians, copepods, and nematodes.The abundances of these in-ice fauna in theCanada Basin were about two orders of magnitudebelow estimates for coastal fast, first-year ice.

Carbon and nitrogen productivity of the watercolumn was measured at 13 locations, including

Ship track of the Arctic2002 expedition, with

station numberssuperimposed.

5

three ice stations. A chlorophyll maximum wasobserved at many stations at 50–60 m. Primaryproduction under the ice is about an order of mag-nitude lower than in open water. The occurrenceof amphipods and Arctic cod was also studied.Amphipods were less abundant than reportedfrom other parts of the Arctic, occurring at meanabundances between 1 and 23 per m2 at each sta-tion. Small schools of Arctic cod were discoveredin narrow wedges along the ice edges, which wereArctic cod under the ice.

R. Hopcroft examining afreshwater melt pond onthe sea ice of the Canada

Basin, 2002.

Qing Zhang, a member of the Chinese science team,removing a long cylindrical core of ice from the thickArctic icepack. After the core had been drilled andextracted, scientists with the Primary ProductivityGroup could begin their sampling procedures.

6

documented for the first time as important fishhabitat. In addition, an unexpectedly high abun-dance of small-bodied copepod species was foundin the water column under the ice, the importanceof which has not yet been pursued in the Arctic.

Water column observations suggested a moreabundant assemblage of gelatinous taxa thanexpected, with many species having distinct depthranges, some extending to the bottom of the basin.The gelatinous zooplankton of the Canada Basinobserved with the ROV fell into four main groups:cnidarians, ctenophores, chaetognaths, andpelagic tunicates. The vertical distributions ofthese gelatinous zooplankton showed severaltrends related to the physical properties of thewater and the geographic locations within thebasin. The most common gelatinous organismsin the surface waters were the ctenophoresMertensia ovum and Bolinopsis infundibulum.These two species were found in very large num-bers in the near-surface mixed layer, immediatelyabove a layer characterized by the large jellyfishChrysaora melanaster. In the mesopelagic zone,siphonophores were common at the top of theAtlantic water layer, while below the transitionbetween the Pacific water layer and the Atlanticwater layer, the most common species wasSminthea arctica. Surprising numbers of thescyphomedusa Atolla tenella were found in thedeep waters of the basin, along with an unde-scribed species of narcomedusa. Larvaceanswere common in surface waters, were broadly

distributed throughout the water column, andrepresent the most abundant holoplanktonic taxabelow 2000 m. Chaetognaths were observedprimarily in the upper 500 m, where plankton netcollections revealed that the biomass of the mostabundant chaetognath, Eukrohnia hamata, wasexceeded only by the two species of Calanuscopepods that classically dominate Arctic zoo-plankton collections.

The benthic infauna were sampled using boxcores between 640 and 3250 m. A total of 90 benthicinvertebrate taxa were identified from four biogeo-graphic regions in the Canada Basin. At least three

Vertical distribution ofobserved Sminthea arcticain the water column from

the top of the ChukchiPlateau (left), the edge of

the Northwind Ridge(center), and the central

part of the Canada Basin(right). Temperature pro-

files are superimposed.Because the time spent at

each depth varied (all sta-tions were less than one

hour), the number of totalobserved specimens are

indicated in light barsand the numbers of

specimens standardizedto one hour are indicated

in dark bars.

The small medusa Sminthia arctica, the most commonspecies identified by video in the mesopelagic realm dur-ing the 2002 Canada Basin expedition.

7

new species of isopods were discovered. However,the benthic abundance of life and the biomass arevery low in the Canada Basin. Total abundancesand biomass were highest in the shallow Amundsen

Seamount discovered on the northern boundary of the Chukchi Plateau, 2003.

High-priority multibeam mapping sites on the Chukchi Cap and Northwind Ridge, 2003.

Gulf and lowest in the deep basin. Polychaetesand crustaceans were most abundant in the sam-ples, while polychaetes and mollusks dominatedthe biomass. ROV surveys revealed epifauna (lifeon the seafloor) where hard substrate was avail-able for attachment.

Abyssal and midslope Arctic benthic fisheswere sampled by still photography and videogra-phy using the ROV Global Explorer. The speciesdiversity of the observed fishes was very low,with only six species; the diversity varied amongstations sampled. Qualitative ROV video analysissuggests that demersal fish may be selecting habi-tats based on the presence or absence (or density)of other benthic animals. Stable isotope analysissuggests that most of the primary production isconsumed by water column grazers and that thebenthos primarily relies on food taken from sink-ing grazers and their products. This informationsuggests that there is a long food web of fourtrophic levels, which suggests low food availability.

Overall, the 2002 Canada Basin explorationrevealed fundamental discoveries about the distri-bution and types of life that inhabit this ocean.

Mapping the Arctic:Exploring the ChukchiPlateau

In September 2003, Arctic and hydrographicresearchers, led by Dr. Larry Mayer from theCenter for Coastal and Ocean Mapping at theUniversity of New Hampshire, embarked on a 10-day Arctic Ocean mapping expedition along theChukchi Plateau and Northwind Ridge. Thismission focused on creating detailed bathymetricmaps in a unique area located in the U.S. ExclusiveEconomic Zone (EEZ) north of Alaska.

The team sailed aboard the U.S. Coast Guardicebreaker Healy. The vessel was equipped witha hull-mounted Seabeam multibeam sonar capableof sensing the ocean floor at great depths. Thebathymetric and backscatter imagery data createdby the multibeam sonar provided important infor-mation about the tectonic processes affecting theocean basin.

In particular, the research team addressed ques-tions regarding the extent of grounded ice on theChukchi Plateau. Confirmation of the extent of icegrounding in this region is of great importance forunderstanding the history of Pleistocene glacia-tion in the Northern Hemisphere.

8

In addition to mapping the 2500-m contoursurrounding the Chukchi Plateau, the team discov-ered a seamount taller than Mount Rainier anduncovered evidence suggesting the outgassingof methane from the top of the Chukchi Plateau.

Russian–American Long-term Census of the Arctic

On July 23, 2004, a Russian research ship,the Professor Khromov, left Vladivostok, Russia,packed with U.S.- and Russian-funded scientists.It marked the beginning of a 45-day collaborativejourney of exploration and research in the RussianArctic. It was also a historic day for Russian–U.S.relations.

Stemming from a 2003 Memorandum of Under-standing for World Ocean and Polar RegionsStudies between NOAA and the Russian Academyof Sciences, this cruise was the first activity underthe Russian–American Long-term Census of theArctic (RUSALCA).

The expedition took place in the Bering andChukchi Seas. These seas, and the life within,are thought to be particularly sensitive to globalclimate change, because they are centers wheresteep thermohaline and nutrient gradients in theocean coincide with steep thermal gradients in theatmosphere. The Bering Strait acts as the onlyPacific gateway into and out of the Arctic Ocean.As such, it is critical to the flux of heat betweenthe Arctic and the rest of the world.

Monitoring the flux of fresh and salt water, aswell as establishing benchmark information aboutthe distribution and migration patterns of the sealife, is particularly important prior to the installa-tion of an Arctic climate-monitoring network.

The cruise was divided into two integratedlegs. Both included sampling and instrumentdeployment in U.S. and Russian territorial waters.The cruise objectives supported the U.S. inter-agency Study of Environmental Arctic Change(SEARCH) program and the NOAA Ocean Explora-tion Program.

The Hidden Ocean II:Canada Basin

In June–July 2005 a second exploration expedi-tion to the Canada Basin took place, building onthe exploration of the deep Canada Basin in 2002.This program also included collaboration with

Multibeam bathymetry of a section of the Chukchi Plateau, revealing a field of pock-marks and deep iceberg plow marks. The data were collected during the 2003 USCGicebreaker expedition.

9

Russian and Chinese scientists. Like the previousexpedition, it utilized an ROV, the Ocean Explorer,and sub-ice divers in conjunction with a deep-ocean camera package and more traditionalsampling tools.

Education and OutreachEducators, media specialists, and data manag-

ers have been involved from the beginning of theArctic Ocean Exploration Initiative. National Geo-graphic magazine participated during the 2002Canada Basin Expedition and was an integral partof the ROV development and operations. In 2004a reporter from Reuters participated onboard theRUSALCA expedition. Data management for the2004 RUSALCA expedition will be coordinatedwith the National Oceanographic Data Center, theNational Climate Data Center, the National Geolog-ical Data Center, the National Snow and Ice DataCenter, and the World Data Center at Obninsk,Russia, along with universities and other relevantinstitutions.

PartnersNOAA has involved partners from other Federal

agencies (such as NSF, DOD, DOI, and HomelandSecurity), universities, its joint institutes, and theNational Ice Center in the Arctic Ocean Explora-tion expeditions. NOAA also cooperates withappropriate international Arctic research institu-tions, such as those in Canada, Germany, Norway,Sweden, China, Korea, Japan, and the Russian

Federation. In addition, private sector organiza-tions were invited to participate, especially tofacilitate international collaboration and the pub-lic outreach aspects of the Ocean ExplorationOffice.

Arctic Ocean DiversityAt the same time as NOAA was developing

its plans for an Arctic Exploration Program, theinternational Census of Marine Life (CoML) pro-gram was also in development. Officially launchedin 2003, the program established an Arctic projectin 2004. The Arctic Ocean Diversity (ArcOD)project aims to increase our basic knowledge ofbiodiversity in the three major biological realmsby consolidating existing biological information,filling gaps in our knowledge by new samplingefforts, and synthesizing what is known by 2010.The steering group contains representatives fromCanada, Denmark, Germany, Norway, Russia, andthe U.S. to foster international collaboration anda pan-Arctic perspective on marine biodiversity.Within the U.S., NOAA’s Offices of ArcticResearch and Ocean Exploration have been theprimary supporters of this initiative because oftheir overlapping and complementary interests.ArcOD-associated scientists have been majorparticipants on the 2002 and 2004 NOAA expedi-tions, as they will be on the upcoming 2005expedition as well. Their attention to ecosystem-wide biodiversity will help lay the foundationfor assessing the impacts of climate change onbiological communities within all Arctic marinerealms.

Location of theRUSALCA leg 2

stations in the Beringand Chukchi

Seas, 2004.

A copepod, Paracuchaeta, with its late summer egg mass.This species is abundant in the Canada Basin.

10

International Polar YearDuring 2007 and 2008, the Ocean Exploration

Program plans to support multiple projects in boththe Arctic and the Antarctic in conjunction withthe International Polar Year. The type and level ofsupport provided will be project-specific. The Officeof Ocean Exploration may coordinate ship and sub-mersible time for multiple Ocean Exploration-fundedprojects, as part of one or more larger explorations.

ReferencesBluhm, B.A., I.R. MacDonald, C. Debenham, and

K. Iken (2005) Macro-and megabenthic commu-nities in the high Arctic Canada Basin: Initialfindings. Polar Biology, Vol. 28, p. 218–231.

Gradinger, R., and B.A. Bluhm (2005) Arctic Oceanexploration, 2002. Polar Biology, Vol. 28, p.169–170.

Gradinger, R.R., K. Meiners, G. Plumley, Q. Zhang,and B.A. Bluhm (2005) Abundance and compo-

sition of the sea-ice meiofauna in off-shorepack ice of the Beaufort Gyre in summer 2002and 2003. Polar Biology, Vol. 28, p. 171–181.

Hopcroft, R.R., C. Clarke, R. J. Nelson, and K.A.Raskoff (2005) Zooplankton communities ofthe Arctic’s Canada Basin: The contribution bysmaller taxa. Polar Biology, Vol. 28, p. 198–206.

Iken, K., B.A. Bluhm, and R. Gradinger (2005) Foodweb structure in the high Arctic Canada Basin:Evidence from d13 C and d15 N analysis. PolarBiology, Vol. 28, p. 238–249.

Lee, S. H., and T. E. Whitledge (2005) Primary andnew production in the deep Canada Basin duringsummer 2002. Polar Biology, Vol. 28, p. 190–197.

Raskoff, K.A., J.E. Purcell, and R.R. Hopcroft(2005) Gelatinous zooplankton of the ArcticOcean: In situ observations under the ice.Polar Biology, Vol. 28, p. 207–217.

Stein, D.L., J.D. Felley, and M. Vecchione (2005)ROV observations of benthic fishes in theNorthwind and Canada Basins, Arctic Ocean.Polar Biology, Vol. 28, p. 232–237.

The Arctic food web.

11

Arctic Sea Ice and Ocean Observations

This article was preparedby Ignatius G. Rigor, of

the Applied Physics Labo-ratory (APL), University

of Washington, Seattle;Jackie Richter-Menge, of

the U.S. Army ColdRegions Research and

Engineering Laboratory;and Craig Lee, of APL.

The Arctic and sea ice play several importantroles in the global climate system, includingeffects on the surface heat budget and the globalthermohaline circulation. Excess latent and sensi-ble heat from the sun absorbed at lower latitudesis transported poleward by the atmosphere andocean, where it is radiated back out to space.Sea ice over the Arctic Ocean insulates the atmo-sphere from the ocean, thus reducing the amountof heat lost to space. Sea ice also has a higheralbedo (reflectivity) than the darker ocean, reduc-ing the amount of sunlight absorbed by the sea-ice-covered ocean. A decrease in sea ice wouldincrease the exposed area of the darker ocean,increasing the amount of sunlight absorbed, thuswarming the ocean, melting more sea ice, andamplifying the initial perturbations (this is calledice–albedo positive feedback). However, the mois-ture fluxes into the atmosphere are also higherover open water than over sea ice, which mayincrease the areal coverage of fog and low clouds,increasing the albedo near the surface and damp-ing the initial perturbations (this is called cloud–

radiation negative feedback). These opposingfeedbacks underscore the complexity of the Arcticand global climate systems.

Understanding the changes in Arctic climateand sea ice is important, since these changes havesignificant impacts on wildlife and people. Manyspecies and cultures depend on the sea ice forhabitat and subsistence. For example, Inupiat huntfor bowhead whales, and polar bears hunt andraise their young on the sea ice. The lack of sea icein an area along the coast may expose the coast-line to ocean waves, which may threaten low-lyingcoastal towns and accelerate the rate of erosion.From an economic viewpoint, the extent of Arcticsea ice affects navigation from the Atlantic to thePacific through the Arctic along the Northern SeaRoute and Northwest Passage, which are as muchas 60% shorter than the conventional routes fromEurope to the west coast of the U.S. or Japan.

NOAA plays an important role in monitoringthe Arctic and supporting research to understandand predict these changes. This article describessome of the many programs that NOAA supportsto monitor Arctic sea ice and the ocean, such asthe International Arctic Buoy Programme (IABP)and the Study of Environmental Arctic Change(SEARCH).

International ArcticBuoy Programme

In 1974 the U.S. National Academy of Sciencesrecommended the establishment of a network ofautomatic data buoys to monitor synoptic-scalefields of sea level pressure, surface air tempera-ture, and ice motion throughout the Arctic Ocean.As a result, the Arctic Ocean Buoy Program wasestablished by the Polar Science Center, AppliedPhysics Laboratory (APL), University of Wash-ington, in 1978 to support the Global WeatherExperiment. Operations began in early 1979, and

Arctic connections toglobal climate. Excess

heat (temperature, T, andhumidity, q) absorbed atlower latitudes is trans-ported poleward by theatmosphere and ocean,

where it is radiated backout to space (longwave

radiation, L). Sea ice overthe Arctic Ocean insulates

the atmosphere from theocean, thus reducing the

amount of heat lost tospace, but it also has ahigher albedo than the

ocean, which reduces theamount of heat absorbedby the ice-covered ocean(shortwave radiation, S).Most of the heat from the

ocean escapes into theatmosphere through the

cracks in the sea ice(heat flux, F).

12

Current locations of International Arctic Buoy Programme (IABP) buoys and the IABP deployment plans for 2005. The current locations offundamental meteorological buoys (gray dots), ice mass balance (IMB) buoys (red dots), and automated drifting stations (ADS) (red stars) areshown. The planned deployments of buoys are shown in blue. More details can be obtained from http://iabpl.apl.washington.edu/Deploy2005/.

13

the program continued through 1990 under fund-ing from various agencies. In 1991 the IABP suc-ceeded the Arctic Ocean Buoy Program, but thebasic objective remains: to maintain a network ofdrifting buoys on the Arctic Ocean to providemeteorological and oceanographic data for real-time operational requirements and research pur-poses, including support to the World ClimateResearch Programme and the World WeatherWatch Programme.

The IABP currently has 33 buoys deployed onthe Arctic Ocean. Most of the buoys measure sealevel pressure and surface air temperature, butmany buoys are enhanced to measure other geo-physical variables, such as sea ice thickness,ocean temperature, and salinity. This observational

Observations from anIABP ice mass balance

(IMB) buoy and a JapanAgency for Marine-Earth

Science and Technology(JAMSTEC) compactArctic drifter (JCAD),which were deployed

together on the driftingArctic sea ice. These

buoys measure sea levelpressure, surface air

temperature, ice thick-ness and temperatures,snow depth, and ocean

temperatures and salinity.

array is maintained by the twenty participantsfrom ten countries, who support the programthrough contributions of buoys, deploymentlogistics, and other services. The U.S. contribu-tions to the IABP are coordinated by the U.S.Interagency Arctic Buoy Program (USIABP),which is managed by the NOAA/Navy NationalIce Center. Of the 33 IABP buoys currently report-ing, 13 buoys were purchased by the USIABP, and18 buoys were deployed using logistics coordi-nated by the USIABP. The USIABP also funds thecoordination and data management of the IABPby the Polar Science Center, at the University ofWashington. The observations from the IABP areposted on the Global Telecommunications Systemfor operational use, are archived at the World DataCenter for Glaciology at the National Snow andIce Data Center (http://nsidc.org), and can beobtained from the IABP web server for research(http://iabp.apl.washington.edu).

The observations from the IABP have beenessential for:

• Monitoring Arctic and global climate change;• Forecasting weather and sea ice conditions;• Forcing, assimilating, and validating global

weather and climate models; and• Validating satellite data.

As of 2005, over 500 papers have been writtenusing the observations collected by the IABP. Theobservations from IABP have been one of thecornerstones for environmental forecasting andstudies of climate and climate change. Many ofthe changes in Arctic climate were first observedor explained using data from the IABP.

USIABP CONTRIBUTORS

U.S. Coast Guard

International Arctic Research Center, University ofAlaska Fairbanks

National Aeronautics and Space Administration

National Oceanic and Atmospheric Administration(NOAA), Arctic Research Office

NOAA, National Environmental Satellite, Data andInformation Service

NOAA, Office of Global Programs

Naval Oceanographic Office

Naval Research Laboratory

National Science Foundation

Office of Naval Research

14

Study of EnvironmentalArctic Change

SEARCH is a coordinated U.S. interagencyprogram established in recognition of the impor-tant role of the Arctic region in global climate.SEARCH’s focus is to understanding the fullscope of changes taking place in the Arctic andto determine if the changes indicate the start ofa major climate shift in this region. NOAA hasinitiated its contribution to the SEARCH programwith seed activities that address high-priorityissues relating to the atmosphere and the cryo-sphere. One element of the NOAA SEARCHprogram is an Arctic Ocean Observing System(AOOS).

The SEARCH AOOS is envisioned to includesix categories of in situ observations: ocean path-way moorings, cross-shelf exchange moorings,basin moorings, gateway moorings, repeatedhydrographic sections, automated drifting sta-tions, and drifting buoys. Enhancement of theIABP and deployment of automated drifting sta-tions (ADSs), like the one at the North Pole Envi-ronmental Observatory (NPEO: http://psc.apl.washington.edu/northpole/), have been identifiedas two of the key components of the SEARCHAOOS. These enhancements and the deploymentof drifting buoys have been the initial focus ofNOAA’s efforts. More specifically, the focus hasbeen on establishing a network of instrumentationto monitor and understand changes in the thick-ness of the ice cover.

The “Vision” for theSEARCH Arctic Ocean

Observing System(AOOS), which includes

six categories of in situobservations: ocean

pathway moorings, crossshelf exchange moorings,basin moorings, gateway

moorings, repeatedhydrographic sections,

automated driftingstations, and

drifting buoys.

15

Data CollectionCentral to the progress that has been made in

establishing a network to monitor changes in thethickness of the ice cover has been the develop-ment and employment of autonomous ice massbalance (IMB) buoys. An IMB buoy is equippedwith thermistor strings that extend through thethickness of the ice cover, acoustic sensors thatmonitor the position of the top and bottom surfacesof the ice, a barometer, a GPS, and a satellite trans-mitter. These buoys provide a time series of sealevel pressure, surface air temperature, snow accu-mulation and ablation, ice mass balance, internal

Components of the ArcticOcean Observing System

(AOOS).

ice temperature fields, and temporally averagedestimates of ocean heat flux. Together, these datanot only provide a record of changes in the icethickness, but equally important they provide theinformation necessary to understand the sourceof these changes. This is critical to extending theresults from these individual sites to other regionsof the Arctic. The buoys are installed in the icecover and, hence, drift with the ice cover. Monitor-ing the drift of the buoys also provides informa-tion on the circular automated drifting stations(ADSs). These sites will be established throughcollaboration with the Alfred Wegner Institute inGermany, the NPEO, the Japan Agency for Marine-

16

Earth Science and Technology, and Woods HoleOceanographic Institution’s Arctic Group. Thesestations provide critical atmospheric, ice, andupper ocean hydrographic measurements thatcannot be obtained by other means.

Since the drifting buoys move with the ice andsurface currents and thus cannot reliably samplethe main subsurface currents of the Arctic, theseobservations must also be complemented bymoorings and hydrographic sections across theArctic Ocean. In August 2003 a new mooring sitewas established in the northern Chukchi Sea. Themooring is equipped with an ice profiling sonar(IPS) to measure the ice draft and velocity as theice drifts overhead, providing a measure of the icethickness distribution. This site was located usingthe results from a coupled ice–ocean sea icedynamics model. The model was used to estimatethe basin-wide mean annual thickness using a 52-year window: 1948–1999. Using these estimates, acorrelation analysis was applied to investigate theeffectiveness of establishing a second seafloor-moored IPS to monitor changes in the annual meanthickness of the Arctic sea ice cover. The analysisrecognized and was dependent on the existenceof the IPS at the NPEO. The results of the analysisindicated that a moored IPS located in the north-ern Chukchi Sea, coupled with the results from theestablished NPEO site, could explain 86% of thevariance of the basin-wide annual mean ice thick-ness. The location of a second mooring signifi-cantly improves the data collected from a singlemoored IPS at the North Pole, where the explainedvariance is estimated to be 65%. Data from themooring sites are only available after the mooringis recovered. The first recovery of the mooring inthe Chukchi Sea is scheduled for September 2004.

A potential AOOS design requires a range ofcomplementary platforms with combined capabili-ties that address SEARCH requirements. TheAOOS includes drifting buoys, moorings, autono-mous platforms (floats, gliders, and propeller-driven autonomous underwater vehicles), andhydrographic sections occupied by ships andaircraft landings. Drifting buoys and moorings,already employed for Arctic observing, would pro-vide sites for acoustic navigation and communica-tions beacons and serve as data repositories andlinks across the ice interface for satellite communi-cations. New autonomous platforms, especiallyfloats and gliders, will provide unprecedentedaccess to Arctic and subarctic regions. Theseplatforms have seen significant successes in mid-latitude oceans and are currently being adapted

for use in ice-covered environments as part of theNSF-supported Freshwater Initiative (see http://iop.apl.washington.edu for additional informa-tion). The first missions beneath the ice will inves-tigate freshwater exchange through Davis Strait.Broader high-latitude application of autonomousplatforms will require the development of long-range acoustic navigation and communicationssystems. An ideal system would supply long-range acoustic navigation to all platforms (Arcticunderwater GPSs), allow mobile platforms to hometo targets (drifting ocean buoys and moorings),and provide short-range, high-bandwidth commu-nications for rapid data exchange. In this vision,moored platforms provide time series, datastorage, and a relay through the ice to satellites;autonomous platforms provide broad spatialcoverage, data shuttling, and satellite links (whenin ice-free waters); and ship- and aircraft-basedhydrography provide critical tracer measurementsthat cannot be obtained by autonomous sensors.This combination of platforms offers comprehen-sive coverage and creates a “store-and-forward”network to improve the timeliness and reliabilityof data return.

ConclusionsRecent progress has been made in establishing

components of an AOOS. The initial focus is on anetwork of instruments to monitor and understandchanges in the thickness of the ice cover and innear-surface ocean characteristics. Central to thesuccess of this network is the coordination ofthese efforts with other national and internationalprograms. The use of sea ice dynamics models hasalso been important, helping to optimize the loca-tion of instrumentation and the allocation of limitedresources.

Data are only just beginning to be receivedfrom the recently deployed sites, but regional andinterannual variability in the changes in the thick-ness of the ice cover and upper ocean is alreadyapparent. This observation is consistent with otherhistorical observations.These data will be madegenerally available to the scientific community foruse in validating satellite-derived products; forforcing, calibration, and assimilation into numeri-cal models; and for forecasting weather and iceconditions. Maintaining and further developingthis network will provide a more consistent recordof change, necessary for improving understandingof this complex and important component of theglobal climate system.

The establishment ofthe components of the

AOOS discussed in thisarticle would not have

been possible without thecommitted collaboration

among many internationalinstitutions and

programs, such as theparticipants of the

International Arctic BuoyProgramme. The authorsexpress their appreciation

to the scientific partici-pants and crew members

onboard the Canadianicebreakers Louis St.

Laurent and Sir WilfredLaurier, the German ice-

breaker Polarstern, theRussian icebreaker

Kapitan Dranitsin, theSwedish icebreaker Oden,

and the U.S.icebreaker Healy.

17

ReferencesArctic Climate Impacts Assessment (2004) Impacts

of a Warming Arctic: Arctic Climate ImpactsAssessment. Cambridge University Press, 139 p.

Lynch, A.H., E.N. Cassano, J.J. Cassano, and L.R.Lestak (2003) Case studies of high wind eventsin Barrow, Alaska: Climatological context anddevelopment processes. Monthly WeatherReview, Vol. 131, No. 4, p. 719–732.

Melling, H., and D.A. Riedel (1996) Developmentof seasonal pack ice in the Beaufort Sea duringthe winter of 1991-1992: A view from below.Journal of Geophysical Research, Vol. 101, No.C5, p. 11,975–11,991.

Perovich, D.K., B.C. Elder, and J.A. Richter-Menge(1997) Observations of the annual cycle of sea

ice temperature and mass balance. GeophysicalResearch Letters, Vol. 24, No. 5, p. 555–558.

Perovich, D.K., T.C. Grenfell, J.A. Richter-Menge,B. Light, W.B. Tucker, and H. Eicken (2003)Thin and thinner: Sea ice mass balance mea-surements during SHEBA. Journal of Geophys-ical Research (Oceans), Vol. 108, No. C3.

SEARCH Science Steering Committee (2001)SEARCH: Study of Environmental ArcticChange, Science Plan. Polar Science Center,Applied Physics Laboratory, University ofWashington, Seattle, 91p.

Zhang, J., D. Thomas, D.A. Rothrock, R.W. Lind-say, Y. Yu, and R. Kwok (2003) Assimilation ofice motion observations and comparisons withsubmarine ice thickness data. Journal of Geo-physical Research, Vol. 108, No. C6, p. 3170.

18

Correlated Declines in Pacific Arctic Snowand Sea Ice Cover

This article was preparedby Robert Stone, of the

Cooperative Institute forResearch in Environmen-tal Sciences and NOAA’sClimate Monitoring and

Diagnostics Laboratory;David Douglas, of U.S.

Geological Survey’sAlaska Science Center;

Gennady Belchansky, ofthe Institute of Ecology,

Russian Academy ofSciences; and Sheldon

Drobot, of the ColoradoCenter for Astrodynamics

Research.

Simulations of future climate suggest that glo-bal warming will reduce Arctic snow and ice cover,resulting in decreased surface albedo (reflectivity).Lowering of the surface albedo leads to furtherwarming by increasing solar absorption at the sur-face. This phenomenon is referred to as “tempera-ture–albedo feedback.” Anticipation of such afeedback is one reason why scientists look tothe Arctic for early indications of global warming.

Much of the Arctic has warmed significantly.Northern Hemisphere snow cover has decreased,and sea ice has diminished in area and thickness.As reported in the Arctic Climate Impact Assess-ment in 2004, the trends are considered to be out-side the range of natural variability, implicatingglobal warming as an underlying cause. Changingclimatic conditions in the high northern latitudeshave influenced biogeochemical cycles on a broadscale. Warming has already affected the sea ice,the tundra, the plants, the animals, and the indige-nous populations that depend on them.

Changing annual cycles of snow and sea icealso affect sources and sinks of important green-house gases (such as carbon dioxide and meth-ane), further complicating feedbacks involving theglobal budgets of these important constituents.For instance, thawing permafrost increases theextent of tundra wetlands and lakes, releasinggreater amounts of methane into the atmosphere.Variable sea ice cover may affect the hemisphericcarbon budget by altering the ocean–atmosphereexchange of carbon dioxide. There is growing con-cern that amplification of global warming in theArctic will have far-reaching effects on lower-latitude climate through these feedback mecha-nisms. Despite the diverse and convincing obser-vational evidence that the Arctic environment ischanging, it remains unclear whether these changesare anthropogenically forced or result from naturalvariations of the climate system. A better under-standing of what controls the seasonal distribu-tions of snow and ice is fundamental to the problem.

Why Be Concerned? TheTemperature–Albedo Effect

Anticipation of a temperature–albedo feedbackis the primary reason observers look to polarregions for early indications of global warming.This effect relates to the dramatic contrast in thereflectivity of sunlight by snow or ice comparedwith open tundra or seawater. Fresh snow reflectsup to 90% of incoming solar radiation, while tun-dra and open water reflect only 15–20% and 6–8%,respectively. Changes in albedo affect the netenergy balance at the surface, which in turn influ-ences air temperature. An increase in net energycan raise temperatures, depending on how thatenergy is redistributed. Most of the excess energyis stored in the ground initially but is released asheat over time to warm the air above.

It has been rightfully stated that “the most sig-nificant factor influencing the magnitude of theyearly net radiation total is the date when snow-melt is completed.”* The timing of snow and icemelt over vast regions of the Arctic can affect cli-mate on a hemispheric scale. Thus, it is importantto monitor changes in Arctic snow and sea icecover and understand what causes them to vary.

Earlier Snowmelt inNorthern Alaska

Since 1940 the spring melt at the NOAA/CMDLBarrow Observatory (BRW) has advanced byabout 10 days. Most of the advance occurred after1976, however. The break in the record coincideswith regime shifts in many climatic and biologicalindicators of change, mainly a consequence of

* Maykut, G.A., and P.E. Church (1973) Radiationclimate of Barrow, Alaska. Journal of AppliedMeteorology, Vol. 12, p. 620–628.

19

changing atmospheric circulation patterns in theNorth Pacific. The year 2002 had a record earlymelt. The 2003 melt was again early, followed by amoderately early melt in 2004. The past threeyears, when combined with 1990, 1996, and 1998,are unprecedented in the 64-year record. Theseanomalously early events statistically drive thelong-term trend.

Variations in the annual snow cycle of northernAlaska are attributable, in large part, to changes inatmospheric circulation related to the position andintensity of two pressure systems: the AleutianLow (AL) and the Beaufort Sea Anticyclone (BSA).On this basis, an empirical model was developedto predict melt dates at BRW. About 75% of theinterannual variability in snowmelt dates at BRWis explained by the model’s input parameters:changes in snowfall during winter and variationsin springtime temperatures and cloudiness.

Annual cycles of daily average radiative energy at the NOAA/CMDL Barrow Observa-tory (BRW). The surface albedo (the bottom panel) is the ratio of the upwelling to thedownwelling shortwave irradiance (middle panel), where irradiance is a measure ofsolar energy. Historically, a daily average albedo threshold of 30% is used to define thefinal day of snowmelt at BRW; i.e., when the snow cover essentially disappears. Duringthe final week of melting, the albedo falls rapidly from about 75% to about 17%. Theimpact that this change has on the net all-wave radiative energy balance is illustrated inthe top panel, where net all-wave radiation is simply the sum of the net shortwave andlongwave components. Values of net short- and longwave irradiance are determined asdownward minus upward components, indicated by red and blue spikes in the top panel.The dramatic decrease in albedo at the time of snowmelt results in a large increase inthe net radiative energy (black curve) because of increased solar absorption by thesurface. In this example, the rise can be likened to illuminating every square meter of thetundra by a 120-watt light bulb (net irradiance increases by about 120 W m–2). Evenmore dramatic increases will occur over ocean areas, because seawater has a loweralbedo than tundra. It is because the melt season coincides with the peak of the annualsolar cycle that the albedo effect is so large.

The timing of snowmelt perturbs the net sur-face energy budget through the albedo effect.Using continuous radiometric measurements madeat BRW, the effect can be quantified. A 10-dayadvance in snowmelt results in a 14–16% increasein net surface radiative energy during the season,equivalent to more than 2 W/m2 of thermal forcingon an annual basis. While this sounds like a smallperturbation, an increase of only 1.0 W/m2 in netsurface radiation can increase the air temperatureby more than 0.5°C. The additional energy is redis-tributed in complicated ways that involve groundstorage, sensible and latent heat exchangesbetween the surface and atmosphere, and air flowthat distributes the energy gain to other regions.In addition to contributing to warming, the recentthawing of permafrost in the region is probablyattributable in large part to earlier snowmelt. Manybiogeochemical cycles are also influenced by thelength of the growing, or snow-free, season athigh latitudes.

Time series of snowmelt dates (as day of year) construct-ed for the NOAA/CMDL Barrow Observatory. Three lin-ear regressions are plotted: an overall fit for 1941–2004(thin black line), one for all years prior to 1977 (green),and a third beginning in 1977 (red). The results of anempirical model are also shown (dashed blue line). Thetime series was compiled from direct snow depth obser-vations, proxy estimates using daily temperature records,and, beginning in 1986, surface albedo measurements.

20

Diminishing Sea Ice in theWestern Arctic Ocean

One of the most alarming indicators of globalclimate change is the continuing decline of Arcticsea ice. Sea ice distributions have been trackedhistorically by direct observations from ship and

aircraft and since the late 1970s from satellite plat-forms. Passive microwave (PMW) satellite sensorsprovide daily data for evaluating sea ice condi-tions, including surface melt, age, ice concentra-tion, and derived extents. From such analyses weknow that the maximum retreat (or minimum extent)of ice occurs in September each year. Since thelate 1970s, September sea ice extent has decreasedby nearly 20%. The minimum occurred in 2002,coincident with the record early snowmelt innorthern Alaska. Last year (2004) marked the thirdconsecutive year of anomalously extreme iceretreat in the Arctic.

As a consequence of accelerated warming athigh latitudes caused by the temperature–albedoeffect, some climate simulations predict ice-freesummers in the Arctic by the end of this century.Even approaching this scenario will gravely affectArctic ecology. The zooplankton, fish, and marinemammals that depend on perennial sea ice, such aspolar bears and a variety of seals, will be directlyimpacted.

Although differing explanations are given forthe decline in Arctic sea ice, there is consensusthat patterns of change derived from satellitePMW data are qualitatively reliable. As in the caseof snowmelt at BRW, since 1989 several anoma-lous years drive the overall decline in ice extent.Changes in multiyear ice (MYI), which is sea icethat has survived at least one summer melt sea-son, have varied geographically. Since 1979, MYIhas decreased by more than 60% over large areas ofthe western Arctic Ocean. The most pronounceddecline began in 1989 in the east Siberian region,followed a few years later by an area north ofAlaska in the Beaufort–Chukchi region. After arebound in 1997, the downward trend resumed.These dramatic declines in the Pacific Arctic domi-nate the trends for the entire basin.

Interannual changes and trends in January multiyear ice area within six longitudinalsectors of the Arctic Ocean (top), and 26-year (1979–2004) linear trend in Januarymultiyear ice concentrations within the western Arctic (bottom).

Time series and linear trend in September Arctic sea iceextent, 1979–2004. The calculation is made for areashaving an ice concentration greater than 15.0%. Theregression yields a decrease of about 7.7% per decade.

21

Correlations of Meltat Sea and on Land

The snowmelt at BRW and the sea ice trends inthe Pacific Arctic appear to be related. The relation-ship is made clearer through cross-correlationsbetween time series of melt onset dates over seaice and snowmelt dates at BRW. A large region ofhigh positive correlation exists between melt onsetover sea ice and the BRW melt record. Much of thisregion also experiences a longer duration of melt.

forcing mechanisms. Several processes couldunderlie these variations, including changes inatmospheric dynamics, synoptic-scale patterns,and related effects.

Atmospheric DynamicsThere is general consensus that the overall

decline in Arctic sea ice (since the late 1970s) isdue to processes associated with dynamicalchanges in the atmosphere. On a pan-Arctic scale,the dynamical state of the atmosphere is oftenexpressed in terms of an Arctic Oscillation (AO)index that relates to sea-level pressure variations.High indices indicate lower-than-average sea-levelpressure over the central Arctic. During a recentand predominately positive phase of the AO(1989–1995), MYI concentrations declined rapidlyin a broad region of the western Arctic Ocean. It ishypothesized that during these years the BeaufortGyre, a region of recirculating ice north of Alaska,weakened. The weakened gyre allowed more ice tobecome entrained in a dominant current called theTranspolar Drift Stream, which exports sea ice intothe North Atlantic east of Greenland. The thickmultiyear ice that was exported was replaced byyounger, thinner ice that is more vulnerable tomelting during the summer months. In recent yearsthe AO index has been neutral or negative, how-ever, and sea ice continues to decline. This sug-gests that factors not well represented by thephase of the AO are at play. Below, an examinationof environmental conditions during years ofanomalous ice retreat reveals+ how regional circu-lation patterns contribute to observed snow andice variations in the Pacific Arctic.

Synoptic-Scale InfluencesOn a more regional scale, shifts in atmospheric

circulation patterns also influence the annual

Maps contouring correla-tion coefficients between

the date of snowmelt at theNOAA/CMDL Barrow

Observatory (indicated bythe red dot) and the onset

date of snowmelt over seaice (left) and the melt sea-son duration (days) over

sea ice (right). The circlesindicate an area of high

correlation.

Time series of standard-ized anomalies of BRWmelt dates (blue) and a

small area of melt onsetover ice (pink) located

within the region of highcorrelation. Standardiza-tion normalizes interan-

nual variations to a com-mon scale to facilitate

comparisons.

To better visualize these results on a scale ofequal variance, standardized anomalies (normal-ized differences) were computed for each timeseries. The results demonstrate that anomalousmelt onset dates over sea ice in the Chukchi–Beaufort region are reflected by similar anomaliesin the snowmelt dates at BRW.

Factors that Influence Snowand Sea Ice Distribution

The fact that melt chronologies over land andat sea have co-varied for many years over a largeregion suggests linkages to common physical

22

cycles of snow and sea ice. In the Pacific Arctic,variations in air temperatures and clouds correlatewith the frequency and intensity of southwesterlywinds. Airflow in this region varies with the juxta-position of the Aleutian Low (AL) and the Beau-fort Sea Anticyclone (BSA). It is the persistenceof clockwise winds within the BSA that drives theBeaufort Gyre. An examination of these synopticfeatures provides insights on what may underliethe regional anomalies in snow and sea ice cover.

For years with minimum sea ice retreat, it is typ-ical for the BSA and AL to be strongly coupledduring March, April, and May, forming a dipolepattern. The BSA effectively blocks Pacific airfrom flowing into the Arctic. Such a pattern keepsnorthern regions cold and relatively dry and con-strains the circulation of ice within the BeaufortGyre. Climatologically, in late May the North Slopeof Alaska and eastern Siberia remain covered insnow. Melt onset over sea ice does not commenceuntil the first week in June north of Alaska and notuntil late June north of Siberia. Under these condi-

The main currents of the Arctic Ocean. During periods of high AO index, the BeaufortGyre weakens, and divergent ice is entrained into the Transpolar Drift Stream, where itis exported from the Basin east of Greenland.

Environmental conditions over the Pacific Arctic averaged for years with minimum (left) and maximum (right) sea ice retreat. The extent of latesummer ice retreat, defined as the southern limit of more than 50% mean ice concentration during late September, is shown as a bold white line. Thinblue lines depict 10-m contours of mean March–May 850-hPa geopotential heights from the NCEP/NCAR 40-year Reanalysis Project. Thesesynoptic patterns for spring represent a critical transition period in the annual cycles of snow and sea ice. Geopotential is commonly used as avertical coordinate when describing large-scale flow patterns, where larger values at a prescribed pressure level (e.g., 850 mb) indicate higheratmospheric pressure.Generalized circulation patterns are shown with bold vectors. Mean melt onset dates over sea ice are color-shaded for areaswhere ice concentrations averaged more than 50% during the second half of May. Vegetation greenness is depicted by the mean maximum NDVI(Normalized Difference Vegetation Index), also during the last two weeks in May, derived from GIMMS NDVI-d and NDVI-n16 data sets.

23

tions the pack does not retreat very far north ofthe coastlines in September, leaving stretches ofthe Siberian coast ice bound.

In contrast, for four recent years of extrememinimum ice extent, the spring BSA is poorlydefined. Instead, a ridge of high pressure persistsover eastern Alaska, and the AL is shifted west-ward. This pattern sets up a strong west-to-eastgradient in the pressure field that favors the trans-port of warm, Pacific air northward. From the large,anomalously warm pool south of the AleutianIslands, a “jet” transports warm, moist air well intothe Arctic Basin. A region of intense southerlywinds coincides with large pressure gradientssouth of the Bering Strait. This incursion of warmair accelerates snowmelt over Alaska and easternSiberia and ice melt over adjacent ocean areas. Bylate May, Pacific air is further warmed as it flowsover bare tundra (with a low albedo) being irradi-ated continuously by intense sunlight. This fur-ther contributes to an early onset of melt over seaice. An early, and thus prolonged, melt seasonamplifies late-summer ice retreat, especially in thePacific Arctic, where the ice pack has becomeyounger and thinner.

Other FactorsThe Role of Clouds

During years with minimum ice extent, the influxof moist air associated with the circulation pattern

described above increase spring cloudiness. Thepresence of clouds have a profound impact on thesurface temperatures in polar regions, where para-doxically, they warm rather than cool the surface.The warming is due to net cloud-radiative forcingwhereby infrared (thermal) emissions from cloudsexceed their cooling effects caused by increasedsolar reflectivity. Snow and ice surfaces absorbabout 99% of this energy and reradiate much of itback into the atmosphere, which raises air temper-ature. Although the effect is greatest during win-ter, significant warming during March and April isassociated with enhanced cloudiness. Empirically,a 5% increase in cloud cover leads to about 1.0°Cof warming over the course of the season, with arange from about 0.7°C in spring to more than1.4°C in winter. In summer, net cloud radiative forc-ing is slightly negative at BRW but is not statisti-cally significant.

Prolonged effects of warm air incursions, aug-mented by cloud radiative forcing during earlyspring, are thought to modify the microphysicalstructure of the overlying snow. This “ripening”may precondition the snow so that melting accel-erates during May and June, when solar insolationreaches its annual peak.

Mean March–May differ-ence field of 850-mb tem-

peratures for (1998,2002–2004) minus (1985–

1988). A jet of air trans-ports anomalously warm

air from the North Pacificinto the Arctic Basin dur-ing years of early snow-

melt and large ice retreat.The southerly winds with-in the circle are more than

3.5 m s–1 more intenseduring years of early melt

onset than for years oflate onset. Temperatures

aloft, over a broad region,are warmer by more than2oC during years of early

melt onset. The results arederived from the NCEP/

NCAR 40-year ReanalysisProject.

Effects of Snowfall VariationsThe depth of snow prior to the onset of melt is

also important. Over land, for average conditions,if there is less snow on the ground when the meltbegins, the snowpack will melt more quickly. Andsignificant sea ice melt cannot begin until the

Empirical relationships of air temperature and total skycover at Barrow, Alaska, averaged over 33 years, illus-trating how the presence of clouds tends to warm thesurface. Regressions for monthly mean values show cor-relations, given in brackets, and temperature sensitivityto a 5% increase in sky cover, given in parentheses.

24

insulating surface layer of snow melts first. If lesssnow accumulates on sea ice in winter and condi-tions favor an earlier ripening of the snowpack,the snow cover will melt more rapidly, advancingthe onset of ice melt.

Historically, direct observations of snowfallover sea ice have been made at Russian drift sta-tions. Analyses of these data indicate reductionsin winter snowfall up until the early 1990s, when,unfortunately, measurements were indefinitelysuspended. Even over the terrestrial Arctic, snowdepth observations are sparse and difficult tointerpret because of wind-induced measurementbiases.

Despite these limitations, an analysis of Arcticsnow cover variations made in 2000 showedthat the Northern Hemisphere snow cover haddecreased by about 10% since the mid-1980s. AtBRW, a 36% decrease in October–February snow-fall from 1966 to 2000 was identified as a major fac-tor in the trend toward earlier snowmelt. BecauseBRW has been shown to be regionally representa-tive, it is likely that the western Arctic Ocean hasexperienced recent declines in snowfall as well, acondition that would exacerbate the early onset ofsummer ice melt.

The Albedo Effect RevisitedFinally, as a consequence of early melt onset

and the longer duration of the melt season, atemperature–albedo feedback occurs that further

accelerates ice melt. The albedo effect is very sig-nificant over land areas when surface reflectivitydecreases rapidly at the time of snowmelt. Mea-surements indicate that a two-week advance in thedate of snowmelt can lead to an increase of morethan 1.0°C in air temperature. Because open waterhas a significantly lower albedo than tundra, it willabsorb 7–12% more solar energy under equal illu-mination. The resultant increase in solar absorp-tion by seawater warms the water column as wellas the air above. Higher air temperatures feedback immediately to enhance melting, while higherwater temperatures increase lateral and subsurfaceice melt and extend the melt season by delayingthe onset of autumn freeze. Ultimately, the icepack will become thinner if the duration of meltbecomes longer and the period of freeze shorter.Without winter recruitment of ice that is thickenough to survive the subsequent melt season,the MYI fraction will continue to decline.

SummaryFactors that influence the distributions of snow

and sea ice are many and complex. Both dynamicand thermodynamic processes have contributedto the earlier onset of snowmelt and reductionsof ice concentration in the Arctic. These mecha-nisms occur naturally but may be exacerbated byincreasing global temperatures caused by green-house forcing. Significant regional and temporalvariations in snow and ice cover are observedbecause forcing mechanisms interact throughcomplicated pathways.

Recent trends appear to be related to shifts inplanetary wave patterns, either on a hemisphericscale as in the case of the Arctic Oscillation or onmore regional scales. The dramatic decrease in MYIin the Pacific Arctic dominates the trend observedfor the entire basin. In the Pacific Arctic, the juxta-position of the BSA and AL during spring affectsthe dynamic and thermodynamic processes thatultimately dictate the distributions of snow andice.

In summary, factors contributing to the loss ofsnow and ice in the Pacific Arctic include the fol-lowing:

• A breakdown of the BSA diminishes thestrength of the Beaufort Gyre. Older ice ismore readily entrained into the TranspolarDrift Stream and exported through the FramStrait, reducing the overall age and thicknessof the icepack and its resilience to summermelt.

An August sun casting awarm glow over the

northern coastline of Ells-mere Island, northeastern

Canada. Rocky ridges ofthe United States Range

rise above low-lying stra-tus clouds. There is aglint of sunlight from

exposed areas of brokensea ice. The image illus-

trates the complexity ofthe dynamic and thermo-

dynamic processes thatmodulate snow and sea

ice cover in the Arctic. Weface a great challenge to

gain sufficient under-standing of this system inorder to forecast the eco-

logical and sociologicalchanges that will accom-

pany rising globaltemperatures.

25

This review was madepossible with support of

the NOAA Arctic ResearchOffice, Study of Environ-

mental Arctic Change(SEARCH) program. R.

Stone is grateful for addi-tional support from

NOAA/CMDL, Boulder,Colorado, and appreci-

ates the efforts of its staffand field personnel whoprovided essential data

and technical supportduring these investiga-tions. The authors alsoacknowledge the great

scientific value in the dataarchives and ancillary

analyses supplied by theNational Snow and Ice

Data Center, ClimateDiagnostics Center, and

Goddard Space FlightCenter.

• A westward displacement of the AL, coupledwith a high-pressure ridge over Alaska duringspring, favors incursions of warm, moistPacific air into the Arctic, promoting earlierand more rapid melt because of thermodynamicand radiative preconditioning of the snow-pack. These involve sensible and latent heatexchanges at the surface through enhancedturbulence and net cloud-radiative heatingcaused by increased cloudiness.

• Earlier melt prolongs the melt season byenhancing the temperature–albedo feedback,exacerbating summer sea ice retreat anddelaying the onset of freezing in autumn.Reductions in the mean sea ice thicknessresult as this cycle repeats.

• Under conditions of reduced snowfall, snow-melt and subsequent icemelt occur morerapidly and the albedo feedback is furtherprolonged, enhancing processes describedabove. Snowfall variations are also a functionof varying modes of atmospheric circulation(now under investigation).

The question arises as to whether the recentretreat of Arctic sea ice is a manifestation of natu-ral, low-frequency climate oscillations or an earlysignal of anthropogenic forcing. Are these mecha-nisms now self-propagating as a consequenceof increasing global temperatures? Stroeve et al.(2005) questioned whether “the extreme ice minimaof 2002–2004 represent the crossing of a threshold,”in which case thinner ice cannot survive longersummer melt seasons. Lindsay and Zhang (2005)suggested the system may have already “tipped”into a new equilibrium state, in which summers willbe characterized by vast regions of open water.The mechanisms underlying major shifts in plane-tary circulation and their influence on regional-scale dynamic and thermodynamic processes mustbe better understood to determine the likelihoodsof future scenarios.

ReferencesArctic Climate Impact Assessment (2004) Impacts of

a Warming Arctic: Arctic Climate Impact Assess-ment. Cambridge University Press, New York.

Belchansky, G.I., D.C. Douglas, I.A. Alpatsky, andN.G. Platonov (2004) Spatial and temporal multi-year sea ice distributions in the Arctic: A neuralnetwork analysis of SSM/I data, 1988–2001.Journal of Geophysical Research, Vol. 109, p.C10017.

Belchansky, G.I., D.C. Douglas, and N.G. Platonov(2004) Duration of the Arctic sea ice melt sea-son: Regional and interannual variability, 1979–2001. Journal of Climate, Vol. 17, p. 67–80.

Lindsay, R.W., and J. Zhang (2005) The thinningof Arctic sea ice, 1988–2003: Have we passed atipping point? Paper no. JP2.19, Proceedings ofthe American Meteorological Society EighthConference on Polar Oceanography andMeteorology, San Diego, California.

Stone, R.S., D.C. Douglas, G.I. Belchansky, S.D.Drobot, and J. Harris (2005) Cause and effectof variations in western Arctic snow and seaice cover. Paper no. 8.3, Proceedings of theAmerican Meteorological Society EighthConference on Polar Oceanography andMeteorology, San Diego, California.

Stone, R.S, E.G. Dutton, J.M. Harris, and D. Longe-necker (2002) Earlier spring snowmelt in north-ern Alaska as an indicator of climate change.Journal of Geophysical Research, Vol. 107.

Stone, R.S. (1997) Variations in western Arctic tem-peratures in response to cloud radiative andsynoptic-scale influences. Journal of Geophys-ical Research, Vol. 102, p. 21,769–21,776.

Stroeve, J.C., M.C. Serreze, F. Fetterer, T. Arbetter,W. Meier, J. Maslanik, and K. Knowles (2005)Tracking the Arctic’s shrinking ice cover:Another extreme September minimum in 2004.Geophysical Research Letters, Vol. 32, No. 4.

26

Acidifying Pollutants, Arctic Haze,and Acidification in the Arctic

This article was preparedby Patricia K. Quinn, ofNOAA’s Pacific MarineEnvironmental Labora-

tory, Seattle, Washington;Betsy Andrews, of NOAA’s

Climate Monitoring andDiagnostics Laboratory,

Boulder, Colorado;Jesper Christensen, of the

National EnvironmentalResearch Institute,

Roskilde, Denmark;Ellsworth Dutton, of

NOAA’s Climate Monitor-ing and Diagnostics

Laboratory, Boulder,Colorado; and

Glenn Shaw, of theUniversity of Alaska

Fairbanks.

The Arctic HazePhenomenon

It has been more than 50 years since observa-tions of a strange atmospheric haze, of unknownorigin, were reported by pilots flying in the Cana-dian and Alaskan Arctic. Based on measurements atMcCall Glacier in Alaska, Shaw and Wendler (1972)noted that the turbidity in the air reached its peakin spring. The first measurements of the verticalstructure of the haze were made using an Alaskan“bush” airplane with a hand-held sunphotometer.

At that time the origin of the haze was uncer-tain and was attributed to ice crystals seeded byopen leads in the ice or blowing dust from river-beds. It was only through “chemical fingerprint-ing” of the haze that its primary anthropogenic(man-made) source in Eurasia was revealed. By thelate 1970s the anthropogenic origin was clear butsurprising, since it was widely believed that aero-sols were generally not transported more than afew hundred kilometers from their source regions.Experts from Europe and America convened at thefirst Arctic Air Chemistry Symposium at Lillestrom,Norway, in 1978, and an informal measurementnetwork was agreed upon. Data soon showed thedirection of flow and the anthropogenic cloud ofpollution. A combination of intensive field pro-

grams and long-term measurements extendingover the past thirty years confirmed the early con-clusions that the haze is anthropogenic and origi-nated from Eurasia. It also became known thatthese atmospheric contaminants were transportedto and trapped in the Arctic air mass during thewinter and early spring.

Arctic Haze has been the subject of more thanthree decades of study because of its potential tochange the solar radiation balance of the Arctic,affect visibility, and provide a source of contami-nants to Arctic ecosystems. The near-surface con-centration of aerosols at most places in the Arcticis about an order of magnitude lower than thosefound at more polluted and industrialized loca-tions. At the same time, however, the affectedareas are much larger, and the affected ecosystemsin the high Arctic are thought to be quite sensitiveto gaseous and aerosol contamination.

The haze is composed of a varying mixture ofsulfate, particulate organic matter (POM), and, toa lesser extent, ammonium, nitrate, dust, black car-bon, and heavy metals. The identification of par-ticular heavy metals allowed industrial sourcesto be identified. Particles within the haze are wellaged, with a mass median diameter of 0.2 micronsor less. This particle size range is very efficient atscattering solar radiation because the peak in theparticle surface–area size distribution is near themaximum efficiency for Mie scattering. The hazealso is weakly absorbing because of the presenceof black carbon. These scattering and absorptioneffects lead to the well-documented reduction inArctic atmospheric visibility (sometimes down toonly a few kilometers or less). During transportfrom Eurasia to the Arctic, the pollutant-containingair masses have a high probability of reaching sat-uration and nucleating and precipitating clouds.Both the clouds and the Arctic Haze may also sig-nificantly affect climate because the haze weakensthe reflectivity of the white snow and ice, loweringthe albedo of the earth in the process.

Soot covering a statue inKrakow, Poland, one

of the source regions ofArctic Haze.

27

Several seasonally dependent mechanisms arethought to contribute to the formation of ArcticHaze:

• Strong surface-based temperature inversionsform in the polar night, causing the atmo-sphere to stabilize.

• This cold and stable atmosphere inhibits tur-bulent transfer between atmospheric layers,and it also inhibits the formation of cloudsystems and precipitation, the major removalpathway for particulates from the atmosphere.

• In addition, transport from the mid-latitudesto the Arctic intensifies during the winter andspring.

The combination of these factors results in thetransport of precursor gases and particulates tothe Arctic and the trapping of the pollutant hazefor up to 15–30 days.

Aircraft and lidar measurements collectedthroughout the 1980s and 1990s revealed that thehaze occurs primarily in the lowest 5 km of theatmosphere and peaks in the lowest 2 km. Through-out the haze season, the pollution layers are highlyinhomogeneous, both vertically (tens of meters to1 km thick) and spatially (20–200 km in horizontalextent).

Recent aircraft measurements of sulfate aerosolrevealed how the haze develops its vertical struc-ture between February and May (Scheuer et al.2003). During early February, atmospheric sulfateaerosol is transported from cold regions in north-

ern Eurasia and accumulates in a 2-km-thick layeron the land and ocean surface. As the haze seasonprogresses, transport from warmer regions in Eur-asia (the source region of the haze) occurs athigher altitudes (up to at least 8 km). Since verticalmixing is prevented by the persistent Arctic low-level inversion, the haze layers remain stratifieduntil spring. During early April, sulfate layersbelow 3 km begin to dissipate because of thebeginning of solar heating and the resulting mix-ing near the surface. However, more stable isen-tropic transport (transport that follows constanttemperature lines) continues at higher altitudes.By the end of May, both the lower- and higher-altitude sulfate enhancements are significantlydecreased because of the continued break-up ofthe inversion and the return of rain and wet snow,which removes the haze from the atmosphere anddeposits it to the ground.

Occurrence and Trendsof Arctic HazeTrends in Chemical Composition

Arctic Haze is marked by a dramatic increase inthe concentrations of several key particulate pol-lutants during winter. Stations where Arctic Hazeis monitored include Alert in the Canadian Arctic(82.46°N), Station Nord in Greenland (81.4°N),Spitsbergen on the island of Svalbard (79°N), Bar-row, Alaska (71.3°N), Karasjok (69.5°N) and Svan-vik (69.45°N) in northern Norway, Oulanka innorthern Finland (66.3°N), and Janiskoski (69°N)in western Russia.

Each site undergoes a similar winter/earlyspring increase in sulfate, with maximum concen-

Range of the Arctic airmass in winter

(January, blue) andsummer (July, orange).

Arctic Haze sampling station locations.

28

trations reaching up to 2.5 μg/m3. Summertimemonthly average concentrations are generally lessthan 0.1 μg/m3. Nss (non-sea-salt) sulfate makesup about 30% of the submicron mass during thehaze season. The time series of particulate nitrateat Alert and Barrow show clear seasonal patternsfor this chemical species. Maximum concentra-tions approach 0.15 μg/m3.

Other species sampled in the haze have theirorigins from biomass burning and dust from Eur-asia (ammonium and nss potassium from biomassburning and magnesium and calcium from dust).The concentration of these particles in the hazereaches a maximum in the winter and spring.

Natural aerosol chemical components haveseasonal cycles that are quite different fromanthropogenic components. Sea salt is a natural

aerosol component that is transported to theNorth American Arctic from the North Pacific andthe North Atlantic Oceans between Novemberthrough February.

Another natural aerosol component is methane-sulfonic acid (MSA). Concentrations of MSAbegin to increase in late June as the ice melts andrecedes from the shoreline and phytoplanktonproductivity in surface waters begins. Dimethyl-sulfide that has been trapped under the ice isreleased to the atmosphere, where it oxidizes toform MSA.

A third aerosol component that has a naturalsource is particulate organic matter (POM), whichmakes up, on average, 22% of the total fine aero-sol mass. POM reaches a maximum in the atmo-sphere during the summer, probably because ofsummer biogenic emissions and/or enhancedoxidation processes. There is a small increase inorganic acids as early as February and March thatmay be an indication of photooxidation at polarsunrise.

The longest record of sulfate concentrationsin the Arctic (1980 to present at Alert, Canada)revealed no change in sulfate concentrationsduring the 1980s. These stable concentrationsare attributed to little change in emissions in theformer Soviet Union between 1985 and 1990.Beginning in 1991, sulfate and other measuredanthropogenic constituents (lead, zinc, copper,

Time series of monthlyaveraged particulate

sulfate and nitrate concen-trations in μg/m3 forBarrow, Alaska, and

Alert, Canada. The datawere made available for

Alert by the CanadianNational AtmosphericChemistry (NAtChem)

Database and AnalysisSystem and for Barrow by

NOAA PMEL (https://saga.pmel.noaa.

gov/data/).

Time series of monthlyaveraged particulate

sulfate concentrations inμg/m3 for eight Arctic

monitoring sites. The datawere made available for

Alert by the CanadianNational AtmosphericChemistry (NAtChem)

Database and AnalysisSystem, for Barrow byNOAA PMEL (Pacific

Marine EnvironmentalLaboratory; https://

saga.pmel.noaa.gov/data/), and for the otherstations by EMEP (Co-

operative Programme forMonitoring and Evalua-

tion of the Long-rangeTransmission of Air

Pollutants in Europe;http://www.emep.int/).

29

excess vanadium and manganese, ammonium,and nitrate) began to decline, suggesting that thereduction of industry in the early years of the newEurasian republics had an effect in the Arctic. Infact, sulfate concentrations measured at StationNord in northern Greenland and at several othersites in the Arctic decreased significantly through-out the 1990s, coinciding with a reduction in emis-sions from the former Soviet Union.

Based on a linear fit to monthly averaged Aprilconcentrations, the decrease ranges from 0.1 to

0.5% per year. In contrast, nitrate appears to beincreasing at Alert at the rate of about 0.3% peryear. The decoupling of the trends of nitrate andsulfate also are evident. Nitrate concentrationspeak in early winter and then again in spring inBarrow, while sulfate concentrations do not.

Trends in Optical PropertiesThe seasonality and trends of Arctic Haze are

clearly seen in time series data of light absorptionand scattering by aerosols measured at the sur-face and in total column aerosol optical depthmeasurements.

Bodhaine and Dutton (1993) reported that aero-sol scattering, optical depth measurements, andsulfate concentrations at Barrow and Alert were ata maximum in 1982, reducing twofold by 1992. Thedecrease was apparent during March and April,the usual maximum haze period. They suggestedthat the reduction in the output of pollution aero-sols by Eurasia and stricter pollution controls inWestern Europe caused the decrease in the haze.The decreases in aerosol scattering and opticaldepth at Barrow during this ten-year period arenot equal to the known reductions of sulfate emis-sions, however, indicating that other factors suchas changes in transport could have played a role.

Monthly averaged con-centrations of sulfate and

nitrate in μg/m3 for April.The dashed lines are

linear fits to the data. Thedata were made availablefor Alert by the Canadian

National AtmosphericChemistry (NAtChem)

Database and AnalysisSystem, for Barrow byNOAA PMEL (https://saga.pmel.noaa.gov/

data/), and for the otherstations by EMEP (http://

www.emep.int/).

Monthly averaged lightscattering and light

absorption at 550 nm bysub-10-μm aerosols at

Barrow, Alaska, andblack carbon mass con-

centration at Alert,Canada. The data were

made available forBarrow by NOAA CMDL(Climate Monitoring andDiagnostics Laboratory)

and for Alert bythe Canadian National

Atmospheric Chemistry(NAtChem) Database and

Analysis System.

30

From 1992 through the late 1990s, light scatteringat Barrow continued to decrease.

An update of the monthly averaged light scat-tering data analysis shows that, for the months ofMarch and April, scattering increased from the late1990s to 2004. This increase is not apparent in theMarch and April monthly averaged values of lightabsorption at Barrow. Sharma et al. (2004) reporteda decrease of 56% in black carbon (the main lightabsorber in the Arctic atmosphere) during the win-ter and spring from 1989 to 2002 at Alert, Canada.A monthly average of the data for April suggestsan increase in black carbon since 1998 at Alert.

An extension of the Barrow aerosol opticaldepth (AOD) data through 2002 shows a contin-ued decrease through the late 1990s. Monthlyaveraged values of AOD anomalies (relative to abase of non-volcanic years) for March show acontinued decline through 2002. However, theAOD anomalies for April indicate an increasebetween 1998 and 2001, where the currentlyavailable data record ends. In contrast to the Bar-row trend through the 1990s, Herber et al. (2002)reported a slightly increasing trend in AOD (1%per year) at Koldewey station in Ny-Alesund,Spitzbergen, between 1991 and 1999.

Effects of Aerosols on theClimate System in the ArcticDirect Effect

The direct effect of aerosols on the radiationbalance in the Arctic is due to the absorption andscattering of radiation by the aerosol. The Arcticis thought to be particularly sensitive to changesin radiative fluxes imposed by aerosols because ofthe small amount of solar energy normally absorbedin the polar regions. Arctic Haze is present as alayer of light-absorbing material over a highlyreflective ice and snow surface.

Shaw and Stamnes (1980) first realized that theabsorbing nature of Arctic Haze would have a sig-nificant impact on the energy balance of the Arc-tic. Several early calculations estimated that thediurnally averaged atmospheric warming due tothe layer ranged between 2 and 20 W/m2. Theseestimates agreed with direct measurements fromwideband sun photometers. Valero et al. (1989)measured heating rates of about 0.1–0.2 K/dayduring AGASP (Arctic Gas and Aerosol SamplingProgram) II. The AASE (Airborne Arctic Strato-spheric Expedition) II flights in the winter of 1992

Monthly averaged concentrations of light scattering and light absorption at 550 nm forsub-10-μm aerosol at Barrow, Alaska, and black carbon for Alert. The averages forMarch and April are shown. The dashed lines are third-order polynomial fits to thedata. The vertical lines represent one standard deviation of the monthly mean. The datawere made available for Barrow by NOAA CMDL and for Alert by the CanadianNational Atmospheric Chemistry (NAtChem) Database and Analysis System.

Monthly averaged aerosol optical depth anomalies at Barrow, Alaska, for March andApril. The anomalies are relative to a base of non-volcanic years. The data from 1992and 1993 were removed because of stratospheric aerosol influx from the Mount Pinatuboeruption in 1991. The vertical lines represent one standard deviation of the monthlymean. The data were made available by NOAA CMDL.

31

revealed soot-contaminated Arctic aerosols at alti-tudes of 1.5 km. Pueschel and Kinne (1995) calcu-lated that this layer of aerosols could heat theearth–atmosphere system above surfaces of highsolar albedo (ice and snow), even for single scat-tering albedos as high as 0.98. Hence, a modestamount of black carbon in the haze layers canresult in a measurable contribution to diabaticheating.

MacCracken et al. (1986) estimated that thecooling of the surface because of absorption ofsolar radiation by the haze layers would be bal-anced by infrared emission from the atmosphereto the surface. The large reflection from the high-albedo ice and snow surface may enhance aerosolsolar absorption. This aerosol-induced atmosphericheating would result in increased infrared emis-sion from the atmosphere to the surface, produc-ing surface cooling. During the dark winter, infra-red emissions from the haze may heat the surface,but this amount of heating is expected to be smallbecause the haze particles are predominantly inthe submicron size range and therefore are an orderof magnitude smaller than the characteristic wave-length of infrared radiation. Deliquescent sulfatesalts, however, may cause the particles to growand become cloud droplets or ice crystals, therebyenhancing their impact in the longwave. In addi-tion, since the haze is present throughout theArctic night, the integrated effect may modify theradiative budget. The vertical distribution of theabsorbing haze layers does not affect the radiationbudget at the top and bottom of the atmospherebut may impact atmospheric circulation and climatefeedback processes.

Indirect EffectsThe indirect effect of aerosols on radiative

fluxes in the Arctic results from the impact aero-sols have on the microphysical properties of

clouds. Aerosols modify cloud optical propertiesby changing the concentration, size, and phase ofcloud droplets. An increase in the number of pol-lution aerosol particles that act as cloud conden-sation nuclei (CCN) will affect Arctic stratus andstratocumulus clouds by increasing the clouddroplet number concentration, which results inmore radiation being reflected back to space. Atthe same time, cloud droplet size will decrease,reducing drizzle formation and increasing cloudcoverage and lifetime. Garrett et al. (2004) showedthat low-level Arctic clouds are highly sensitive toparticles that undergo long-range transport duringthe winter and early spring. The sensitivity wasdetected as higher cloud droplet number concen-trations and smaller cloud droplet effective radiicompared to summertime clouds exposed to parti-cles nucleated in the Arctic from local biogenicsources. In addition, Arctic stratus clouds appearto be more sensitive to pollutant particles thanclouds outside of the Arctic. The most significanteffect of the change in cloud properties caused byArctic Haze may be on cloud emissivity. A decreasein droplet effective radius in these optically thinclouds will increase the infrared optical depth andthus the infrared emissivity. The result is expectedto be an increase in downwelling infrared fluxesfrom the cloud and an increase in the rate ofspringtime snowpack melting.

Pollution aerosol within Arctic Haze also isthought to impact ice nucleation. Anthropogenicsulfate is a large component of the haze. Modelsestimate that aerosols containing sulfuric acid pro-duce fewer ice nuclei than nearly insoluble aero-sols. Measurements corroborate this finding.Borys (1989) reported that Arctic Haze aerosolshad lower ice nuclei concentrations, a lower ice-nuclei-to-total-aerosol fraction, and slower icenucleation rates than aerosols from the remoteunpolluted troposphere. The reduction in icenuclei leads to a decrease in the ice crystal number

Impact of soot depositedon snow and ice surfaces

in the Arctic. Polar icereflects light from the sunback to space (left panel).As the ice begins to melt,less light is reflected andmore is absorbed by theoceans and surrounding

land, leading to anincrease in overall

temperature and furthermelting. Darker, soot-

covered ice reflects evenless light and, thus,

enhances the warming(right panel).

32

concentration and an increase in the mean sizeof ice crystals. As a result, the sedimentation andprecipitation rates of ice crystals increase, leadingto an increase in the lower troposphere dehydra-tion rate and a decrease in the downwelling infra-red fluxes from the cloud. Girard et al. (2005) foundthat a cloud radiative forcing of –9 W/m2 at Alertmay occur locally as a result of the enhanceddehydration rate produced by sulfate aerosol. Ifthis applies to much of the Arctic, it could explainthe cooling tendency in the eastern high Arcticduring winter.

Because of the combination of the static stabil-ity of the Arctic atmosphere, the persistence oflow-level clouds, and the relatively long lifetimeof aerosols during the haze season, the impact ofaerosols on cloud microphysical and optical prop-erties may be larger in the Arctic than elsewhereon earth.

Surface AlbedoSurface albedo affects the magnitude and sign

of climate forcing by aerosols. Absorbing sootdeposited on the surface via wet and dry deposi-tion impacts the surface radiation budget byenhancing absorption of solar radiation at theground and reducing the surface albedo. Hansenand Nazarenko (2004) have estimated that sootcontamination of snow in the Arctic and the corre-sponding decrease in surface albedo yields a posi-tive hemispheric radiative forcing of +0.3 W/m2. Theresulting warming may lead to the melting of ice andmay be contributing to earlier snowmelts on tun-dra in Siberia, Alaska, Canada, and Scandinavia.

Clearly, the radiative impacts of pollutant aero-sols in the Arctic are complex. Complex feedbacksbetween aerosols, clouds, radiation, sea ice, andvertical and horizontal transport processes compli-cate the impact, as do potentially competing effectsof direct and indirect forcing. As a result, the mag-nitude and sign of the forcing are not yet wellunderstood for the Arctic.

SummaryMeasurements of sulfate aerosol—a main

constituent of Arctic Haze—and light scatteringand extinction show that the amount of the hazeimpacting the Arctic was either nearly constantor decreasing between the 1980s and early 1990s.The updated trends in light scattering presentedhere show indications of an increase in the haze atBarrow, Alaska, since the late 1990s. There also is

evidence, although not as strong, of an increasingtrend in light absorption during this same periodat Barrow and in black carbon at Alert, Canada.Sulfate appears to still be declining at all sitesexcept Barrow, where the trend is unclear. On theother hand, nitrate appears to be increasing atAlert and Barrow. Continued measurements cou-pled with chemical transport models are requiredto better define emerging trends and assess theircauses.

Arctic Haze is generally understood to consistof anthropogenically generated material and hasoften been attributed to sources in central Eurasia.There may be, however, increasing amounts ofmaterial entering the Arctic from growing naturaland anthropogenic sources in eastern Eurasia,particularily from China. Examples of Asian dustentering the Alaskan sector of the Arctic weredocumented as long ago as the mid-1970s. Therapid industrialization of China and increasingamounts of pollution being transported over longdistances are likely to impact the Arctic. Moreresearch is required to document the contributionof this increasing source to Arctic Haze and todetermine its impact on the Arctic.

Other key atmospheric species have a distinctseasonality in the Arctic. There is evidence of theenrichment of halogens in Arctic air masses in latewinter and spring. Since these compounds tendto peak later in the year, it is thought that theyare produced photochemically. More research isrequired to determine their sources (for example,anthropogenic, especially coal combustion, vs.marine), to investigate their numerous and com-plex chemical pathways, and to assess their envi-ronmental impacts. Of special note is iodine, whichshows a bimodal seasonal behavior, peaking inboth spring and autumn.

The direct radiative effect of Arctic Haze hasbeen estimated with one-dimensional radiativetransfer models, which find a warming in the atmo-sphere because of absorption of solar radiationand a concurrent cooling at the surface. Theseestimates are highly sensitive to the assumedproperties of the aerosol in the haze. Despite themany research activities devoted to the character-ization of Arctic Haze since the 1970s, measure-ments of Arctic aerosols are not extensive or welldistributed in space or time, which limits the accu-racy of the estimates of both the direct and indi-rect radiative forcing. Treffeisen et al. (2004) havedesigned an approach based on cluster analysisfor integrating aircraft, ground-based, and long-term data sets for use in three-dimensional climate

33

models. The accurate evaluation of climate forcingby Arctic Haze requires such data sets coupledwith three-dimensional climate models that considerboth direct and indirect effects. In particular,three-dimensional models are required to assessthe complex feedbacks between aerosols, clouds,radiation, sea ice, and dynamic transport and toquantify climate forcing caused by Arctic Haze.

ReferencesBodhaine, B.A., and E.G. Dutton (1993) A long-term

decrease in Arctic Haze at Barrow, Alaska. Geo-physical Research Letters, Vol. 20, No. 10, p.947–950.

Borys, R.D. (1989) Studies of ice nucleation byArctic aerosol on AGASP-II. Journal of Atmo-spheric Chemistry, Vol. 9, p. 169–185.

Bowling, S.A., and G.E. Shaw (1992) The thermo-dynamics of pollution removal as an indicatorof possible source areas for Arctic Haze. Atmo-spheric Environment, Vol. 26, p. 2953–2961.

Garrett, T.J., C. Zhao, X. Dong, G.G. Mace, andP.V. Hobbs (2004) Effects of varying aerosolregimes on low-level Arctic stratus. Geophysi-cal Research Letters, Vol. 31.

Girard, E., J.-P. Blanchet, and Y. Dubois (2005)Effects of Arctic sulphuric acid aerosols onwintertime low-level atmospheric ice crystals,humidity and temperature at Alert, Nunavut.Atmospheric Research, Vol. 73, 131–148.

Hansen, J., and L. Nazarenko (2004) Soot climateforcing via snow and ice albedos. Proceedings

of the National Academy of Sciences, Vol. 101,No. 2, p, 423–428.

MacCraken, M.C., R.D. Cess, and G. L. Potter(1986) Climatic effects of anthropogenic Arcticaerosols: An illustration of climatic feedbackmechanisms with one- and two-dimensional cli-mate models. Journal of Geophysical Research,Vol. 91, p. 14445–14450.

Pueschel, R.F., and S.A. Kinne (1995) Physical andradiative properties of Arctic atmospheric aero-sols. Science of the Total Environment, Vol.161, p. 811–824.

Scheuer, E., R.W. Talbot, J.E. Dibb, G.K. Seid, L.DeBell, and B. Lefer (2003) Seasonal distribu-tions of fine aerosol sulfate in the North Ameri-can Arctic basin during TOPSE. Journal ofGeophysical Research, Vol. 108, No. D4, p.8370.

Sharma, S., D. Lavoue, H. Cachier, L.A. Barrie, S.L.Gong (2004) Long-term trends of the black car-bon concentrations in the Canadian Arctic.Journal of Geophysical Research, Vol. 109.

Treffeisen, R., A. Herber, J. Ström, M Shiobara, T.Yamanouchi, S. Yamagata, K. Holmén, M.Kriews, and O. Schrems (2004) Interpretation ofArctic aerosol properties using cluster analysisapplied to observations in the Svalbard area.Tellus, Vol. 56B, p. 457–476.

Valero, F.P.J., T.P. Ackerman, and W.J.R. Gore(1989) The effects of the Arctic Haze as deter-mined from airborne radiometric measurementsduring AGASP II. Journal of AtmosphericChemistry, Vol. 9, p. 225–244.

34

The Barrow Atmospheric Baseline Observatory

This article was preparedby Russ C. Schnell andDavid J. Hofmann, of

NOAA’s Climate Monitor-ing and Diagnostics

Laboratory, Boulder,Colorado.

NOAA’s Climate Monitoring and DiagnosticsLaboratory (CMDL) operates the manned Atmo-spheric Baseline Observatory six miles east of Bar-row, Alaska (71.3°N; 156.6°W). The observatorymeasures changes in atmospheric climate forcingagents such as carbon dioxide (CO2), methane(CH4), nitrous oxide (N2O), sulfur hexafluoride(SF6), climate-forcing and stratospheric-ozone-depleting chlorofluorocarbons, air pollution fromEurasia known as Arctic Haze, and solar radiation,to name only a few of the more than 200 measure-ments conducted at the facility. NOAA estab-lished the observatory in 1973 in an 800-ft2 build-ing in use today. In recent times the observatoryhas taken on the support of 16 cooperative

research projects from other agencies and univer-sities, with a concentration of projects from theUniversity of Alaska. Two permanent CMDL staffoperate the facility.

The observatory is situated near the center ofan 80-acre parcel of Federal land one mile southof a DOD radar facility, through which the obser-vatory access road passes. The CMDL acreageis bounded on the west by an 80-acre parcel ofFederal land that is home to the U.S. GeologicalSurvey’s Barrow Magnetic Observatory and onthe south and east by the 7,500-acre Barrow Envi-ronmental Observatory (BEO), land preserved forscientific research. Adjacent to the main buildingis a two-vehicle garage with gas cylinder storagespace. Additional facilities consist of a 60-ft-tallwalk-up sampling tower, three elevated platformsfor instrument support, and a number of smallertowers and instrument installations on the tundra.The observatory site is host to a Department ofEnergy Atmospheric Radiation and Monitoring(ARM) facility, two NOAA National Environ-mental Satellite, Data, and Information Service(NESDIS) polar-orbiting satellite downlink anten-nas, and a NOAA Climate Reference Network(CRN) station.

The observatory facility does not have anyinternal combustion sources or volatile chemicalsthat might contaminate the trace gas measure-ments conducted within the facility. Vehicle trafficnear the station is controlled, and there are noroads upwind of the observatory. Some observa-tory measurements require a view of the naturalsurface that is unaffected by anthropogenic influ-ences and will be maintained in such a state for atleast a century or more. The upwind clean-airsector (45° through 100°) is expected to remain aclean-air sector in perpetuity unless oil productionfacilities are installed in the Arctic Ocean within a100-mile radius upwind of Barrow.

The Barrow Atmospheric Baseline Observatoryis the farthest north of the five manned observato-

NOAA/CMDL Barrow Atmospheric Baseline Observatory, viewed from the east nearthe base of the 20-m sampling tower. The former Naval Arctic Research Laboratory ison the upper right horizon. The DOE ARM facilities are above and to the left of thegarage (center right), with the USGS Magnetic Observatory above and to the left of thewhite DOE building. The Barrow Atmospheric Baseline Observatory is in the center ofthe view, and the Dobson ozone spectrophotometer dome is at center left. Winds persis-tently blow from the point of the photograph towards the main observatory building.

35

NOAA National Environmental Satellite, Data, and Information Service (NESDIS) 3-m (left) and 4-m (right) HighResolution Picture Transmission (HRPT) downlink antennas for the NOAA Polar Operational Environmental Satel-lite (POES) at the observatory site. These antennas allow for near-real-time downloads of ice and cloud conditionsover the Arctic Ocean.

The NOAA/CMDL global Carbon Cycle Greenhouse Gas monitoring network. Carbon dioxide is measured continuously at the observatoriesshown by the blue squares, and weekly, paired glass flask samples of air are collected at the sites represented by the red dots. The marine samplesare taken at 5°-latitude intervals from regularly scheduled ships. Weekly to biweekly vertical profiles of trace gases are sampled with light aircraftat the blue- starred locations.

36

ries operated by CMDL; the others are at TrinidadHead, California; Mauna Loa, Hawaii; AmericanSamoa; and South Pole. Carbon dioxide has thelargest total cumulative direct radiative climateforcing in the atmosphere—about three timesgreater than CH4, six times greater than the CFCs,and 15 times greater than N2O. Sulfur hexafluoride(SF6), although a strong greenhouse gas andincreasing in the atmosphere, is present in lowconcentrations and has a much smaller net effecton total radiative forcing than the other green-house gases.

Carbon Dioxide andMethane Measurements

Most of the landmass on earth is in the north-ern hemisphere, as is vegetation and the humanpopulation. Anthropogenic activities at lower lati-tudes produce combustion effluents that result inhigh background CO2 concentrations at Barrow

and other locations in the Arctic thousands ofkilometers north of the sources because of north-ward transport of the gases. Similarly, eventhough Barrow has very low annual plant growth,forests and agriculture in regions such as inRussia and Canada, well south of the latitude ofBarrow, dominate the summer CO2 drawdown inthe high Arctic. Combined, these factors producethe largest annual background CO2 cycle on earth.

Methane measurements at Barrow show thatthe Arctic also has the highest annual CH4 cycleamplitude on earth, as well as the highest winterconcentrations of the gas compared to lowernorthern latitudes and the southern hemisphere.This is caused by complex interactions between CH4sources and sinks and transport of mid-latitudeair into the Arctic. Global average atmosphericCH4 concentrations increased from 1625 ppbduring 1984 to 1751 ppb during 1999 and thenhave remained essentially constant for the pastfive years. This is a large decrease in the methanegrowth rates compared to 1983, when the growth

Monthly mean carbondioxide concentrations

determined from continu-ous analyzers at four

NOAA/CMDL baselinestations. Note the large

relative amplitude of theCO2 cycle at Barrow,

which is a combination oftransport of fossil fuel

combustion effluents fromlower latitudes and highprimary productivity in

the northern hemisphereand mid-latitude forests

and agricultural regions.

37

Three-dimensional viewsof the global distribution

of carbon dioxide (top)and methane (bottom)

over time, showing thelarge amplitude of both

cycles in the northernhemisphere. The 10°-lati-

tude band that encom-passes the Barrow data is

indicated in red on thethree-dimensional graph.

38

Concentrations intropospheric mixing

ratios of thechlorofluorocarbons

(CFCs) controlled by theMontreal Protocol and

measured at five CMDLmeasurement sites from

Barrow, Alaska, in theArctic to South Pole,

Antarctica. Note that theBarrow concentrations,

shown in red, are for themost part higher than

measurements at lowerlatitudes.

rate was 13.5 ppb. It is not yet known if thischange in the atmospheric methane growth rateis a temporary pause or a new, and as yet unex-plained, steady state.

CFCs, Nitrous Oxide,and Sulfur Hexafluoride

Concentrations of anthropogenic chlorofluoro-carbon-11 (CFC-11), CFC-12, methyl chloroform(CH3CCl3), and carbon tetrachloride (CCl4) inthe atmosphere have decreased since 1991 inresponse to the international treaty known asthe Montreal Protocol to Reduce Chemicals thatDeplete the Ozone Layer and its subsequentamendments. The amount of the decrease foreach compound, after production was reduced orceased, is related to how quickly the compoundis naturally destroyed in the atmosphere; methylchloroform has an atmospheric lifetime of 5.5 yearsand is decreasing fast compared to CFC-11, whichhas a lifetime of 45 years, and CFC-12, with a life-time of about 100 years. The Arctic typically hashigher levels of these anthropogenic gases thanat lower latitudes, as atmospheric transport moves

air pollution from heavily populated northern mid-latitudes into the Arctic, where the gas concentra-tions build up, especially in winter.

The concentrations of nitrous oxide and sulfurhexafluoride at Barrow continue to grow. Nitrousoxide is produced in conjunction with fossil fuelcombustion and oxidation of agricultural fertilizers,whereas sulfur hexafluoride is produced by a limitednumber of manufacturers mainly supplying theelectricity transmission industry. N2O has a sea-sonal cycle at Barrow, with a clear wintertime peakin February. SF6 is steadily increasing in the atmo-sphere at a rate of 5% per year, and its sources arein the northern hemisphere.

Advance of SnowmeltDate and Springtime

The date on which the last snow melts in theBarrow area, defined as the last day when oneinch of snow can no longer be measured, has beenmonitored by the National Weather Service (NWS)since 1940 and by CMDL since 1986. The snow-melt date at the Barrow Atmospheric BaselineObservatory has advanced by 10.2 days in 64

39

Global tropospheric mix-ing ratios for N2O (top)

and SF6 (bottom).. Globalmeans were determined

from measurements madeby on-site instrumentation

and from air collected inflasks in the NOAA coop-erative sampling network

that were subsequentlyanalyzed by NOAA/CMDL in Boulder,

Colorado.

years. This is a significant advance at the begin-ning of the snow-free period, which is in the rangeof 85–135 days in Barrow. The causes for thisincrease in snowmelt date are believed to be acombination of decreased snowfall in winter andhigher spring temperatures brought on by changesin both winter and spring Arctic air-flow patterns.

Whatever the causes of the advancing snow-melt date, one species of bird that nests on theobservatory property is taking advantage of theearlier spring. In the summer of 2002, snow bun-tings raised two clutches of chicks in the Barrowarea, the first time this has ever been observed,according to the local Inupiat hunters and elders.

40

ReferencesDlugokencky, E.J., S. Houweling, L. Bruhwiler,

K.A. Masarie, P.M. Lang, J.B. Miller, and P.P.Tans (2003) Atmospheric methane levels off:Temporary pause or a new steady-state? Geo-physical Research Letters, Vol. 30, No. 19.

Stone, R.S., E.G. Dutton, J. M. Harris, and D.Longenecker (2002) Earlier spring snowmeltin northern Alaska as an indicator of climatechange. Journal of Geophysical Research, Vol.107, No. D10, p. 4,089.

Observed dates of snow-melt at Barrow, showingan advancing trend over

the past 64 years.

Baby snow bunting (top) hatched near the Barrow obser-vatory and a parent (bottom) looking for insects to feedthis second-hatch chick. In 2002 the early spring andconsequent longer summer allowed snow buntings to suc-cessfully raise two sets of fledglings on the North Slopetundra, the first time this has been observed in Barrow inliving memory.

41

NOAA and the Alaska Ocean Observing SystemContributions to the National Backbone and Regional Needs

This article was preparedby Allen Macklin, of

NOAA’s Office of Oceanicand Atmospheric

Research; Gary Hufford,of NOAA’s National

Weather Service; BernardMegrey, of NOAA’s

National Marine FisheriesService; Rebecca Smyth,

of NOAA’s NationalOcean Service; and Molly

McCammon, ExecutiveDirector, Alaska Ocean

Observing System.

The Alaska Ocean Observing System (AOOS)will improve Alaska’s ability to detect changes inmarine ecosystems and living resources, predictfuture changes and their consequences for thepublic, and enable stakeholders to make betterdecisions about use of the marine environment.AOOS partners include:

• Federal agencies, such as the National Oceanicand Atmospheric Administration (NOAA), theU.S. Geological Survey, the Minerals Manage-ment Service, and the U.S. Coast Guard;

• Federal–State agencies, such as the ExxonValdez Oil Spill Trustee Council;

• State agencies, such as the Alaska Depart-ment of Fish and Game, and state academicinstitutions, including the University of Alaska;

• Research organizations, such as the NorthPacific Research Board, the Alaska SeaLifeCenter, the Prince William Sound ScienceCenter, the Arctic Research Commission, and the Barrow Arctic Science Consortium; and

• Industry groups, including fisheries andmarine navigation associations.

AOOS is one of 11 regional associations devel-oping across the country to ensure that observingsystems meet regional needs as part of the U.S.Integrated Ocean Observing System (IOOS).Nationally, the effort to establish IOOS is led bythe Ocean.US Office under the National Oceano-graphic Partnership Program. Legislation creatingthe national system and associated regional sys-tems, such as AOOS, is currently pending in theU.S. Congress. IOOS, in turn, is part of the GlobalOcean Observing System and ultimately will bethe U.S. ocean contribution to the Global EarthObserving System of Systems.

Implementation of AOOS started in 2002. Apilot project, employing elements of the PrinceWilliam Sound Ocean Observing System thatincludes NOAA platforms, will be the first on-lineoperational element, delivering information thisyear. When fully developed, AOOS will:

• Serve as the Alaska regional node for thenational network of observing systems(IOOS);

• Systematically deliver real-time informationand long-term trends about Alaska’s oceanconditions and marine life;

• Provide public Internet access to cost-freedata and information on coastal conditions;and

• Supply tailored products to meet the needsof mariners, scientists, industry, resourcemanagers, educators, and other users ofmarine resources.

Implementing AOOS presents an enormouschallenge because of the vastness of the region.Alaska’s nearly 44,000 miles of coastline consti-tute about two-thirds of the total U.S. coastlineand support a wide variety of habitats and usercommunities. NOAA, with a strong statewidepresence in research, monitoring, and forecasting,is well positioned to help establish AOOS.

To make the challenge of implementation moretractable, AOOS’s first approach is to organizealong three large marine ecosystem boundaries:Arctic, Bering Sea/Aleutian Islands, and Gulf ofAlaska. These regional classifications tend to benatural divisions that are differentiated by physi-cal and biological characteristics, managementschemes, and use by stakeholders. Even the sizeof these three regions, however, poses challenges.

Because of Alaska’s remoteness and extremeweather conditions (frigid temperatures, precipita-tion, storms, high sea state, and sea ice), designing,installing, and operating an ocean observing sys-tem throughout the three Alaska regions is moredifficult than in any other shelf area in U.S. waters.The extremely long distances render any plan forperiodic servicing or unscheduled maintenanceand repairs of observing arrays very costly andlogistically often impractical. The dearth of nearbyinfrastructure, such as villages or other semiper-manent settlements, makes power availability,

42

real-time data retrieval, and routine equipmentmaintenance extremely demanding for almostevery installation. Winter conditions challengeinstrument capabilities because of the extremetemperature changes and the high winds andseas, ice, snow, and fog that accompany them.Extensive cloud cover associated with frequentpassage of storms also contributes to the lackof ocean color, AVHRR, and other visible remotesensing products that are typically available inother coastal areas.

In spite of the challenges of establishing anintegrated ocean observing system in Alaska, theopportunities and needs warrant national atten-tion. Presently, the Alaska fisheries provide morethan 40% of the U.S. and about 5% of the worldharvest of fish and shellfish; Bristol Bay supportsthe world’s largest sockeye salmon fishery; andthe snow crab fishery is currently the largest crus-tacean fishery (by weight) in the U.S. In additionto supporting a large portion of the nation’s fish-ery production, Alaska waters also support morethan 80% of the U.S. seabird population. Anothercrucial point for implementing AOOS is thatgreenhouse-gas-related global warming is thoughtto be amplified in polar regions, making Alaskaconditions a harbinger for climate change.

NOAA’s MissionNOAA also shares many of Alaska’s concerns.

NOAA envisions an informed society that uses acomprehensive understanding of the role of theoceans, coasts, and atmosphere in the global eco-

system to make the best social and economic deci-sions. NOAA’s mission is to understand and pre-dict changes in the earth’s environment and con-serve and manage coastal and marine resources tomeet our nation’s economic, social, and environ-mental needs.

To achieve its mission, NOAA’s focus through2010 will be on four mission goals and a missionsupport goal:

• Protect, restore, and manage the use of coastaland ocean resources through an ecosystemapproach to management;

• Understand climate variability and change toenhance society’s ability to plan and respond;

• Serve society’s needs for weather and waterinformation;

• Support the nation’s commerce with informa-tion for safe, efficient, and environmentallysound transportation; and

• Provide critical support for NOAA’s mission.In an effort to build specific core strengths,

NOAA has selected five cross-cutting prioritiesfor the 21st century that it recognizes as essentialto support its mission goals. Three that pertainparticularly to efforts to develop AOOS are:

• Integrating global environmental observa-tions and data management;

• Ensuring sound, state-of-the-art research; and• Promoting environmental literacy.

StakeholdersBy partnering, NOAA and AOOS can address

common themes and provide benefits to their

AOOS’s three regions(left), which are similar tothe Large Marine Ecosys-

tems of the area (right).

43

stakeholders. The stakeholders cover a broadrange of subsistence, commercial, cultural, andeconomic interests. User groups include a widearray of commercial and recreational fishers; sub-sistence hunters and fishers; marine transporta-tion interests such as barges, ferries, cruise ships,and oil/gas tankers; oil and gas developers; coastalcommunities and their residents; and resourcemanagers, including the U.S. Coast Guard perform-ing its full range of regulatory, safety, and securitymissions.

The user groups have a wide range of needsfor data and information products. For example,some of these groups require precise navigationand real-time information, yet others need onlyrudimentary knowledge of currents and watermasses. While these needs exist today, others liein the future, such as possible Northwest Passagetransits under reduced Arctic ice cover. Increasedsurveillance, security, and safety of transportationand commercial shipping activities (offshore, inports, and in sea lines of communication betweenAlaska and the continental U.S.) are recent andemerging areas of concern for the U.S. that will beaddressed by many of the proposed AOOS activi-ties. All of the above information needs are closelytied to forecasting weather and oceanographicconditions, as most weather systems, includingextreme events, transit across marine watersbefore entering Alaska.

The use of AOOS observations and productsfor science applications is also important, espe-cially for developing a better understanding ofthe variability of Alaska’s ocean waters and thediverse ecosystem dynamics that produce thenation’s most abundant fish and shellfish har-vests, as well as important bird and marine mam-mal populations. Many of the science applicationsare directed toward the sustainability of commer-cial and subsistence fishing, especially in theBering Sea/Aleutian Islands and Gulf of Alaskaregions. Other examples of how AOOS productswill contribute to scientific understanding includeaddressing the need to better understand thebiophysical processes (for example, wind mixing,upwelling, and eddy formation) that contribute tothe sustained high productivity of the shelf andcontinental slope waters, as well as improvedassessment of biota. Weather and climate fore-casts will benefit greatly from a much larger setof real-time observations in coastal areas wherethey are presently missing. Modeling long timeseries data would result in an improved andmore comprehensive understanding of icing

phenomena, shelf currents, shoreline erosion,tsunami hazards, and the evolution of catastrophicspill trajectories. In addition, longer-term climatechange scenarios will become more “testable”given a more comprehensive and complete setof observations.

Statewide Priority NeedsThere is a need throughout Alaskan waters

for a system to acquire, process, integrate, andpresent remote sensing products, some of NOAAorigin, that incorporate wind, sea surface height,sea ice cover, ocean color, wave height and direc-tion, water column current, water column salinity,and water column temperature data. An immediaterequirement that NOAA will address is the needto obtain a density of data buoys comparable toat least half that along the rest of the U.S. coast.

Additionally, Alaska needs data managementand communications systems that provide real-time data for use by Alaska stakeholders. Thesystems must include the assimilation of data intomodels that provide information products suchas ocean circulation patterns (taking into accountwaves, eddies, and fronts) and improved near-shore forecasts to minimize impacts of coastalerosion on development. These data systems alsomust store the data and metadata from the observ-ing network in formats that provide ready accessto researchers, regulators, educators, and publicand commercial users.

Finally, Alaska must develop:• Models that assimilate data to simulate circu-

lation, predict wave heights and storm surges,and nowcast/forecast changing sea ice condi-tions;

• Systems that connect marine data and modelswith terrestrial data, especially given theimportance both of freshwater input into themarine system and anadromous fish suchas salmon that rely on both freshwater andmarine waters;

• Comprehensive coastal and offshore mappingand charting; and

• Shore-side capabilities to develop, stage,deploy, operate, and maintain observingsystems, including AUVs, cabled and mooredsystems, and ground- and air-based remotesensors throughout Alaska.

Besides these statewide needs, AOOS also willaddress specific requirements of the Arctic, BeringSea, and Gulf of Alaska regions. These are docu-mented on the AOOS web site (www.aoos.org).

44

NOAA’s Role in AOOSBecause of NOAA’s long-term involvement in

Alaska, NOAA is already a major contributor tothe development of AOOS. NOAA efforts rangefrom service on the Governance and the DataManagement and Communications committees toprovision of funds, observations, and products.Existing and planned activities by NOAA compo-nents are detailed in the subsections below.

National Environmental, Satelliteand Data Information Service

To properly understand the Arctic environ-ment, an observing system must consist of bothspace-based and in situ observations. The back-bone of present space-based observations is theoperational system of polar-orbiting satellites,

such as NESDIS’s Polar Operational Environmen-tal Satellite (POES) series. The five AVHRRsatellite-borne sensors offer a cost-effectivemeans of gaining large-scale information from thesynoptic to mesoscale in a systematic, repetitivemanner over remote, data-sparse, polar regions.With two operating POES satellites, a pass over aportion of the Arctic can be obtained about everytwo hours. The POES series of satellites providesa long-term (more than 30 years), consistent data-base to detect and monitor spatial and temporalvariability, necessary for distinguishing climatetrends from natural “noise.” The POES series willcontinue until 2012, when a new generation ofsatellites called the National Polar-Orbiting Opera-tional Environmental Satellite System (NPOESS)will be launched.

NPOESS will consist of a number of advancedsensor arrays to provide higher resolution and

Contributions of NOAA’s major line organizations to the Alaska Ocean Observing System. Early,direct contributions are designated by a check. With ongoing commitment, all boxes will eventuallybe checked.

AOOS ActivityNOAA Line Modeling and Data Management Education Governance

Organization Observations Analysis and Communications and Outreach and Planning Funding

NESDIS √NMFS √ √ √ √ √NOS √ √ √ √ √ √NWS √ √ √ √OAR √ √ √ √ √

A NOAA Polar Opera-tional Environmental

Satellite (POES), whichcollects global data on

cloud cover; surfaceconditions such as ice,snow, and vegetation;atmospheric tempera-

tures; and moisture,aerosol, and ozone

distributions. They alsocollect and relay

information from fixedand moving data

platforms.

45

more accurate measurements of the atmosphere,clouds, aerosols, the earth radiation budget, clear-air land/water/ice surfaces, sea surface tempera-ture, ocean color, ocean surface wind speed anddirection, ocean surface topography, and tempera-ture and moisture profiles. (See htpp://npoess.noaa.gov/index.html for detailed information onsensor type and expected performance.) The majorchallenge will be to integrate the satellite sensorinformation with the in situ observations, includ-ing calibration and verification of sensor data tothe surface observations.

Another area in which NESDIS is committed toAOOS is the implementation of Climate ReferenceNetwork (CRN) observing stations across Alaska,including coastal sites. The CRN stations willreduce the uncertainty in the observed climate sig-nal for surface temperature to less than 0.1°C percentury and precipitation to less than 1% per cen-tury on regional scales. Approximately 29 CRNsites will be located in Alaska, with about 10 sitesalong the coast. Two sites, at Fairbanks and Bar-row, are already operating. Four more sites will beinstalled during the summer of 2005, and it is antic-ipated that four sites will be installed each summeruntil the installation is complete. The coastal CRNsites will provide an important tie to the ocean–land system.

National Marine Fisheries ServiceNOAA’s National Marine Fisheries Service

(NMFS) conducts biological, ecological, and eco-

nomic research to provide information for theneeds of regional fishery management councils,interstate and international fishery commissions,fishery development foundations, governmentagencies, and the general public.

NMFS, through its research and monitoringactivities, seeks to understand and predictchanges to marine ecosystems and their sub-systems affecting living marine resources, fisher-ies, habitats, ecosystem condition, productivity,aquaculture, and the generation of net nationalbenefits. NMFS develops the scientific informa-tion base required for fishery resource conserva-tion, fishery development and utilization, habitatconservation, protection of marine mammals andendangered species, and the impact analyses andenvironmental assessments for management plansand international negotiations. It also pursuesfisheries oceanographic research from a marineecosystem standpoint to answer specific needs inthe subject areas of population dynamics, fisheryeconomics, fishery engineering, food science, andfishery biology.

The Alaska Fisheries Science Center (AFSC)conducts ecosystem-based research and assess-ments of living marine resources, with a focus onthe North Pacific, to promote the recovery andlong-term sustainability of these resources and togenerate social and economic opportunities andbenefits from their use. Since the early 1970s, theAFSC has conducted annual scientific fishery sur-veys to measure the distribution and abundanceof approximately forty commercially important fish

National Marine FisheriesService survey of com-mercially valuable and

associated fish, shellfish,and marine mammals in

the Gulf of Alaska,eastern Bering Sea,

and Aleutian Islands.

46

National Ocean Servicewater level observing

station at Valdez. Stationssuch as this one arelocated at 18 sites inAlaska, and six new

stations are planned.

and crab stocks in the Gulf of Alaska, eastern Ber-ing Sea, and Aleutian Islands. Surveys in the Ber-ing Sea are conducted annually on a regular sam-pling grid, while surveys in the Gulf of Alaskaoccur every other year using a stratified randomsampling approach to estimating abundance.

The research surveys utilize a wide range ofsampling techniques, mensuration equipment, andfishing gear, including underwater video systems,autonomous submersibles, hydroacoustic tech-nology, and midwater, bottom trawl, ichthyoplank-ton, longline, crab pot, and pot sampling gear, aswell as tagging studies. Often physical oceano-graphic measurements are taken concurrently withbiological samples.

Data derived from these surveys and othersampling programs are analyzed by AFSC scien-tists, and the results and outcomes from theseactivities are supplied to fishery managementagencies and to the commercial fishing industry,where they are used in making resource manage-ment decisions.

NMFS and the AFSC, through their regularexecution of large-scale fisheries surveys, will bean important source of biological information forAOOS. In turn, AOOS will provide much of thephysical and chemical information needed byNMFS for ecosystem-based fisheries management.

Besides the valuable survey information thatNMFS will provide to AOOS, NMFS is a majorpartner in the development of AOOS, providingfunds and personnel who serve in governanceand data management capacities.

National Ocean ServiceNOAA’s National Ocean Service (NOS) works

to balance people’s use of the coast with conser-vation of the nation’s coastal and ocean resources.Thus, NOS’s mission is to manage society’s usesof coastal ecosystems to sustain their naturalresources and services. This mission is undertakenin a variety of ways: supporting commerce andmarine transportation navigation; protecting,restoring, and managing coastal and marineresources; and building the capacity of regional,state, and local partners to undertake both theseactivities. A key component of these activities isto observe coastal and ocean conditions andresources, either directly or by building the capac-ity of regional, state, and local partners.

NOS is conducting or has planned a number ofactivities that will expand or enhance IOOS-relatedactivities in Alaska. NOS will expand the multi-mission National Water Level Observation Net-work (NWLON) in Alaska (18 existing locations)with six new NWLON stations, primarily to helpstrengthen the U.S. Tsunami Warning System.NOS will continue to conduct tidal current sur-veys in Cook Inlet and southeast Alaska to updatetidal current predictions, as well as hydrographicsurveys using in-house and contract capabilities.NOS also will deploy a high-frequency surfacecurrent mapper during the summer of 2005 to com-plete its data collection requirements in the CookInlet area.

NOS also provides financial and technicalassistance to Alaskan partners to aid in the devel-opment of regional observing capabilities. Grantsto AOOS have helped fund the development ofthe regional governance needed to establish theregional association, including outreach and datamanagement. Additional grants starting this fallwill aid the Alaska region in implementing datamanagement, visualization, and pilot observingsystems around the state and in developing edu-

47

cation, outreach, and business plans. Over thepast two years, NOAA has provided resources toaid in developing coastal observations along theGulf of Alaska. The system-wide monitoring pro-gram of the National Estuarine Research Reserves,part of the national backbone, which includes long-term data on water quality and weather at frequenttime intervals, provides resources for the Kache-mak Bay Reserve to participate in the system-wideprogram and the regional efforts. In addition, NOSis working with other parts of NOAA and withtheir partners, including AOOS, to coordinate andprovide technical assistance for data management.

National Weather ServiceThe National Weather Service (NWS) has

nearly 100 land observation stations acrossAlaska that report hourly. NWS also has over 150

cooperative observation sites (many near thecoast) that provide daily minimum and maximumtemperatures and total precipitation. In addition,Coastal-Marine Automated Network (C-MAN)stations are being installed along the coast, espe-cially in the Gulf of Alaska. C-MAN was estab-lished by the NWS in the early 1980s to continuemeteorological observations previously made bythe U.S. Coast Guard until automation of manyCoast Guard navigational aids ended that practice.Over the last few years, the number of fixed oceandata buoys has increased from 3 to 13, providingnew information in data-sparse areas. Several ofthese buoys will be instrumented with subsurfaceocean instruments in the coming year as part of anational effort to increase the data generated bythis system. The data from all these sites form thebackbone for the long-term surface observationsin Alaska.

Locations of nearly 115land and marine stationsoperated by NOS, NWS,and OAR. The observa-tions from these stations

represent NOAA’sprimary contribution to

AOOS. In addition, about150 more stations (notshown) operated by the

Federal Aviation Adminis-tration, the Department ofDefense, and commercial

interests report throughthe NWS network.

48

Biophysical mooring,operated jointly by OAR

and NMFS, that measuresa suite of environmental

variables (pressure, wind,radiation, humidity, air

and sea temperature,current, salinity, nutrients,and indicators of primary

and secondary produc-tivity), some of which are

reported in real time.

As funding becomes available, NWS will con-tinue its efforts to expand marine observationalsites, especially for the northern waters of theBering and Chukchi Seas and the Arctic coast, astechnology provides a buoy that can withstandsea ice conditions. With climate change occurringaround Alaska, forecasters are observing increas-ing frequencies and intensities of ocean storms,shifts in storm track, and more formation of stormsin the Arctic. The recession in sea ice cover is pro-ducing larger areas of open water, which is result-ing in greater air–sea interaction. Large wavesfrom these storms are not only affecting vessels atsea, but they are also creating increased coastalerosion and coastal flooding. In addition, subsis-tence activities are being disrupted. Longer leadtimes are required for short-term forecasts andwarnings so that affected towns and villages can

prepare the best they can. There is a need for moreand better Arctic atmospheric and oceanic obser-vations, both in situ and remotely sensed, andimproved numerical weather and ocean predictionmodels that will incorporate the observations andbetter handle advances in high-latitude meteorol-ogy and oceanography.

Office of Oceanic andAtmospheric Research

The major contributors to AOOS from NOAA’sOffice of Oceanic and Atmospheric Research(OAR) are the Pacific Marine Environmental Labo-ratory (PMEL) in Seattle and the Arctic ResearchOffice. PMEL helped fund early planning forAOOS and has provided representation and lead-ership to the AOOS Data Management and Com-

49

munications Committee. The Arctic ResearchOffice also has been instrumental in planningAOOS and is represented on the GovernanceCommittee. OAR marine observations and prod-ucts will come largely from PMEL. PMEL, togetherwith the Alaska Fisheries Science Center, is a leaderin the deployment of biophysical moorings in con-tinental shelf and slope waters of the Arctic andsubarctic. Several of these are planned for AOOS,including moorings in Cross Sound (southeasternAlaska), Shelikof Strait (western Gulf of Alaska),across the Alaska Stream south of the AleutianIslands, in passes of the Aleutian Islands, and inthe eastern Bering Sea. Biophysical moorings mea-sure a suite of environmental variables (pressure,wind, radiation, humidity, air and sea temperature,current, salinity, nutrients, and indicators of plank-ton biomass), some of which are reported in realtime. These observations may be reported directlyby AOOS or incorporated into marine productsthat are disseminated through AOOS. PMELalso is working cooperatively with AOOS to planand produce ocean circulation models of coastalAlaska.

NOAA funds the U.S. portion of the ArgoProject. The Argo Project is building an array of3,000 profiling CTD floats that, when completed,will measure the temperature and salinity of theupper 2 km at a spacing of roughly 3° latitude by3° longitude and at 10-day intervals. The U.S. ispresently contributing about half of the floats for

Locations of Argo floats,which provide a new

source of data from thetop 2 km of the ocean.

There were 1894 activefloats as of June 13, 2005.

The robotic floats spendmost of their lives at

depth but surfaceregularly to make temper-

ature and salinity profilemeasurements. Many

countries contribute thefloats, and all data are

freely available. Achallenge for the Argo

program will be thedeployment of floats in

Arctic waters. NOAAfunds the U.S. Argo

component.

the array, which was over 60% complete as of mid-June 2005. North of 45°N and east of 180°W in thePacific at that date, there were about 49 Argofloats, and about 15 of these were U.S. floats. Nofloats have been deployed in the Pacific Arcticregion. OAR’s PMEL and Atlantic Oceanographicand Meteorological Laboratory are key contribu-tors to the U.S. Argo effort.

SummaryNOAA and AOOS is a strong partnership.

NOAA’s line organizations are present and operat-ing in each of the three AOOS regions, offering arich mixture of marine and atmospheric measure-ments. AOOS and its constituents benefit fromthe information and services that NOAA supplies.Many of these will now be delivered to and dis-seminated by AOOS, as well as through existingNOAA channels. NOAA benefits from AOOSthrough the enriched flow of Alaska marine infor-mation that NOAA scientists and managers willhave at hand for formulating analyses and deci-sions concerning regional marine issues. More-over, discussions and exchanges between NOAAand AOOS foster an increased understanding ofmutual problems and aspirations that can furtherimprove cooperation between the organizations.Because of ongoing cooperation, NOAA andAOOS will be stronger and better able to serve theregion and nation.

50

Marine Mammals in the Bering/Chukchi Sea

This article was preparedby Sue Moore, of the

Alaska Fisheries ScienceCenter, NOAA.

The National Marine Mammal Laboratory(NMML), Alaska Regional Office, and the ProtectedResources Management Division are responsiblefor research on and management of 22 species ofmarine mammals that commonly occur in Alaska,including five endangered cetacean species(bowhead, fin, humpback, North Pacific right, andsperm whales); one pinniped species (Steller sealion), which is threatened in one portion andendangered in another portion of its range; andtwo depleted species (Cook Inlet beluga whaleand northern fur seal). Field research by theNMML staff on marine mammals off central andnorthern Alaska focused on two pinniped and sixcetacean species during 2002 and 2003: Steller sealions, harbor seals, Cook Inlet beluga whales, killerwhales, and large cetaceans (fin, blue, humpback,and North Pacific right whales) in the Bering Sea.

Steller Sea LionsNOAA is the lead agency responsible for the

management and recovery of the endangered west-ern and threatened eastern populations of Stellersea lions. The western population has declined bymore than 80% in the last two decades, but it mayhave stabilized over much of its range during thelast two years. Conversely the eastern populationappears to be recovering from severely reducedlevels in the early part of this century and hasexhibited consistent growth over the past threedecades. Factors hypothesized for the dramaticdecline in the western population include reducedprey availability leading to nutritional stress, poorjuvenile survival, and decreased reproduction;disease; pollution; predation by killer whales;incidental mortality in groundfish fisheries; and

Fin whale.

51

legal and illegal shooting. The Steller sea lionresearch program at the NMML conducts scientificresearch on each of the potential factors thatcould have contributed to the decline of the west-ern population. The core research program includesvessel and aerial surveys to quantify abundance,molecular and genetic studies to elucidate stockstructure, assessment of predator–prey dynamicsand foraging distributions to determine foragingecology, and individual identification and trackingto provide the foundation of mortality and lifehistory.

Killer WhalesTo investigate the potential role of killer whales

in the decline of the western population of Stellersea lions, a vessel-based survey extending fromthe Kenai Fjords to the central Aleutian Islandswas initiated in 2001. The DART (Distribution andAbundance of Residents and Transients) surveysare designed to estimate the abundance of killerwhales by ecotype. Three killer whale ecotypeshave been identified in Alaskan waters: the pisciv-orous (resident) ecotype; the mammal-eating(transient) ecotype; and the “offshore” ecotype,which apparently preys mostly on fish. Biopsysamples are taken whenever possible, to providedata for molecular genetic, prey isotopic, and fattyacid and contaminant analyses. When conditionspermit, photographs and biopsies of sperm, fin,humpback, and Baird’s beaked whales are alsotaken. These data augment sighting and biopsysampling conducted in collaboration with the AlaskaFisheries Science Center’s Resource Assessmentand Conservation Engineering groundfish surveys.

CetaceansResearch on the Cook Inlet beluga whale stock

has been conducted annually since 1993. Thisstock was designated as depleted under theMarine Mammal Protection Act in 2000. Scientistsfrom NMML, in cooperation with the AlaskaBeluga Whale Committee, the Cook Inlet MarineMammal Council, the Alaska Native Marine Mam-mal Native Hunters Committee, the Alaska Depart-ment of Fish and Game, and NMFS’s AlaskaRegional Office, have estimated the abundance ofthis relatively small and isolated population eachyear since 1994. Analysis of sighting data fromaerial surveys indicated that the abundance ofCook Inlet beluga whales declined by nearly 50%between 1994 and 1998. Distribution and abun-dance estimates from annual aerial surveys in 2002and 2003 indicated that the population was stablebut low in number. In 2002, research was directedtoward catching whales and outfitting them withradio and satellite tags to determine seasonalmovement patterns and correction factors foraerial surveys. A Cook Inlet beluga habitat model,which is in development, is based on satellitetracking data and fatty acids analyses of blubbersamples used to determine diet and contaminantburdens.

Since 1999, line-transect surveys for cetaceanshave been conducted periodically in the southeastBering Sea in association with the AFSC/RACEgroundfish stock assessment survey. Provisionalestimates indicate that fin whales are the mostcommon large whale and that Dall’s porpoises arethe most common small cetacean in these waters.Fin whales are common on the Middle Shelf (50–100 m) and Outer Shelf (100–200 m) domains of thesoutheastern Bering Sea, with large feeding aggre-gations noted near canyons along the Bering Seaslope. Dall’s porpoises, too, are most abundantalong the Bering Sea slope. Cetaceans are generallygood indicators of oceanographic productivitybecause to feed efficiently they must locate denseprey aggregations. Thus, baleen whale harvestsduring the commercial whaling era have been usedto document hydrographic patterns associatedwith areas of zooplankton and forage fish abun-dance. Alternatively, odontocete distribution likelyreflects patterns of higher-order productivityassociated with nektonic prey. Overall, the distri-bution and abundance estimates available from theline-transect surveys in the southeastern BeringSea suggest that baleen whales are reoccupyingproductive hydrographic zones in patterns similar

Feeding gray whale.

52

to those depicted in summaries of commercialwhaling harvests. Observations of cetacean distri-bution and estimates of cetacean abundance fromsurveys completed to date offer a beginning for theincorporation of whales and porpoises in ecosystem-based research plans for the Bering Sea.

Patterns of habitat change in the northernBering and Chukchi seas have been investigatedfor gray and bowhead whales. In the 1980s the

Chirikov Basin north of Saint Lawrence Island inthe northern Bering Sea was considered a primegray whale feeding area, but an unusual mortalityevent in this species in 1999–2000 precipitatedconcern that this area no longer supported aviable benthic forage community. In 2002 a provi-sional five-day survey for gray whales revealedrestricted distribution in the basin and a 3- to 17-fold decline in sight rates. To put these data incontext, a retrospective summary of gray whaleand benthic fauna distribution was undertaken.Available measures of biomass suggest a down-turn in productivity from 1983 to 2000, when esti-mates of gray whale population size suggestedthat the population began to expand. Sightingrates for gray whales were highest north of BeringStrait during the 2002 survey, suggesting thewhales may simply be moving north followingprey availability.

Bowhead whales are ice-adapted baleenwhales, the only species endemic to Arctic waters.To examine how sea ice changes may be affectingbowhead whale habitat, trends in sea ice coverwere examined over a 24-year period (1979–2002)in habitats identified as important to migration,feeding, and over-wintering. Significant increasesin open-water areas were identified for smallregions associated with feeding opportunitiesbut not in areas used during migration and over-wintering. High interannual variability, togetherwith consistent shifts to earlier and longer (i.e.,June to November) ice-free or light-ice conditions,may alter foraging opportunities or prey availability.The evaluation of sea ice cover at spatial andtemporal scales linked to bowhead whale naturalhistory provides a first step towards developingconservation insights regarding the potentialeffects of climate change on this pagophilic species.

PinnipedsFive species of pinnipeds are associated with

sea ice in Alaskan waters, including walrusesand four species of “ice seals”—bearded, ringed,spotted, and ribbon. NOAA has managementresponsibility for the ice seals, each species ofwhich depends at least in part on sea ice to rest,give birth, and molt. Ribbon and spotted seals arethought to prefer the loose ice in the Marginal IceZone (MIZ), which occurs between nearly solidfloes and open sea water. Conversely, beardedseals, ringed seals, and walruses are commonlyfound in zones where ice covers over 50% of thesea surface.

Gray whale distribution inthe northern Bering and

Chukchi Seas in the early1980s and in 2002. Thecolors indicate different

days and years of aerialsurveys.

53

Results of aerial surveys conducted south ofSt. Lawrence Island found that both ringed sealsand walruses preferred large ice floes (larger than48 m in diameter), while spotted seals were found

Bowhead whales.

Bearded seal.

on smaller ice floes (smaller than 20 m in diameter)near the MIZ. Ringed seals were found in areaswith more than 90% sea ice cover and beardedseals preferred 70–90% ice cover. All species,except spotted seals, were associated with aregion where benthic biomass was especiallyhigh.

In recent decades, Alaska harbor seals havedeclined dramatically in some regions, while theirnumbers have increased in some other regions.The primary objectives of NMML’s research onthis species are to obtain data on the abundanceof the species throughout Alaska and to collectinformation on haulout patterns that can be usedto better interpret abundance information. In 2002and 2003 the NMML produced peer-reviewedpapers describing the abundance of harbor sealsin the Gulf of Alaska and the stability of harborseal haul-out patterns. In addition, research wasundertaken to determine the response of harborseals to cruise ships and to determine the geneticrelatedness of harbor seals via molecular genetictechniques. Obtaining information on Alaskaharbor seals is critical, as they are an importantcomponent of the Alaska Native subsistenceharvest. A co-management agreement, signedby the Alaska Native Harbor Seal Commissionand NMFS, has charged the Harbor Seal Co-management Committee to prepare an AnnualAction Plan for this culturally important species.

54

Ocean Climate Changes andthe Steller Sea Lion Decline

This article was preparedby Arthur J. Miller, of the

Scripps Institution ofOceanography; Andrew

W. Trites, of theUniversity of British

Columbia; and HerbertD.G. Maschner, of Idaho

State University.

Steller sea lion populations declined by over80% between the late 1970s and early 1990s in thewestern Gulf of Alaska and the Aleutian Islands.Concurrent declines also occurred farther west inthe Russian coastal waters. Yet population trendswere reversed along the coasts of southeast Alaska,British Columbia, Washington, and Oregon, wheresea lions increased through the 1980s and 1990s.The cause or causes of these population changeshave not been resolved and have been the subjectof considerable debate and research because theirpreferred prey often coincides with economicallyimportant fisheries.

Much of the search for an explanation of theSteller sea lion decline in western Alaska hasfocused on trying to identify a single cause, ratherthan recognizing that many of the proposed theo-ries are interrelated. The leading hypotheses ofepidemic diseases, predation by killer whales,ocean climate change (regime shifts), and nutri-tional shifts in types of prey available to sealions (the junk food hypothesis) may all be linked

through bottom-up processes. Conceptually,changes in water temperatures, ocean currents,and other oceanographic variables can influencethe survival and distribution of assemblages ofspecies that are consumed by predators such assea lions. This in turn will affect the quantity,quality, and accessibility of the prey that sea lionsconsume. Individuals that consume sufficientenergy are typically fat and large and experiencereduced levels of oxidative stress at a cellularlevel. In contrast, inadequate nutrition canincrease oxidative stress (and susceptibility todisease), reduce body fat (and pregnancy rates),and increase rates of predation (as a function ofreduced body size or increased vulnerability whilespending longer times searching for prey). Suchchanges to the health of individuals ultimatelytranslate into changes in numbers at a populationlevel through decreased birth rates and increaseddeath rates.

A major change in both the physical state andthe ecology of the North Pacific Ocean occurredSteller sea lions.

55

during the mid-1970s, with basin-wide changes intemperature, mixed layer depth, primary productiv-ity, fisheries, and other variables. This linkagebetween the physical climate and the oceanic eco-system provided the impetus for the CooperativeInstitute for Arctic Research (CIFAR) to fund asuite of studies that addressed the hypothesisthat large-scale changes in the physical environ-ment of the North Pacific Ocean influenced Stellersea lion populations directly or indirectly. Theinvestigations covered a variety of topics, includ-ing physical and biological oceanographic dataanalysis, ocean modeling experiments, and archae-ological evidence.

CIFAR also sponsored a synthesis workshopin December 2004 that resulted in a detailed publi-cation in Fisheries Oceanography, which is brieflysummarized here. It had two primary goals. Thefirst was to determine whether any spatial andtemporal patterns in the physical and biologicaloceanographic data corresponded with observeddifferences in the diets and numbers of sea lionssince the late 1950s. The second was to put therecent decline in context with similar changes thatmay have occurred over the past 4000 years.

Characteristics ofSteller Sea Lions

Steller sea lions are restricted to the NorthPacific Ocean and range along the Pacific Rim fromCalifornia to northern Japan. Genetically there aretwo distinct population segments that are splitat 144°W near Prince William Sound, Alaska. Thesharp decline of the larger western populationthrough the 1980s was mirrored by populationgrowth in the smaller eastern populations insoutheast Alaska, British Columbia, and Oregon.

Counts of Steller sea lions in Alaska began in1956 and continued sporadically through the 1960sand 1970s. They suggest that sea lion numberswere relatively high and increased slightly throughthe 1960s and 1970s. Trouble was not noted untilthe mid-1970s, and it appeared to spread east andwest from the eastern Aleutian Islands in followingyears. The frequency and thoroughness of sea lioncensuses increased through the 1980s and 1990sand showed an overall rapid decline of sea lionsthrough the 1980s, with an inflection point andslowing of the decline around 1989. Recent counts

Habitats of the westernand eastern stocks of the

Steller sea lion. The graphshows the estimated num-

bers of Steller sea lions(all ages) in Alaska from

1956 to 2000. The dashedline shows the division

between the declining(western) and increasing

(eastern) populations.

56

(2002) suggest the possibility that some breedingpopulations in the eastern Aleutian Islands and Gulfof Alaska may have increased slightly since 1999.

Steller sea lions regularly haul out on shore atbreeding (rookeries) and nonbreeding (haulout)sites. They typically spend one to two days at seafollowed by one day resting on shore. Principalprey species include Atka mackerel, walleye pol-lock, Pacific cod, squid, octopus, salmon, Pacificherring, sand lance, and arrowtooth flounder. Themost complete set of dietary information for sealions was collected during the 1990s. It also sug-gests distinct geographic clusterings, with thesplit points centered on major Aleutian passes(Samalga Pass and Unimak Pass during summerand Umnak Pass during winter).

Significant correlations between rates of popu-lation decline and the diversity of diets suggestthat a relationship may exist between what sealions eat and how their populations have fared.Sea lions living in regions with the highest ratesof declines, such as the western Aleutian Islands,consumed the least diverse diets with the lowestenergy prey. In contrast, the increasing popula-tions of sea lions in southeast Alaska had themost energy-rich diet and the highest diversity ofprey species of all regions studied during summer.During the 1990s, sea lion diets were dominatedby Atka mackerel in the Aleutian Islands and bywalleye pollock in the Gulf of Alaska. Little isknown about what sea lions ate prior to theirpopulation decline.

The National Research Council review of thecauses of the Steller sea lion decline noted that“levels of groundfish biomass during the 1990swere large relative to the reduced numbers of sealions, suggesting that there has been no overalldecrease in prey available to sea lions.” They alsoconcluded that “existing data on the more recentperiod of decline (1990–present) with regard to thebottom-up and top-down hypotheses indicate thatbottom-up hypotheses invoking nutritional stressare unlikely to represent the primary threat torecovery” of sea lions.

The available data support the NationalResearch Council’s conclusion that gadid popula-tions were indeed abundant during the populationdecline and that Steller sea lions did not starveand incur “acute” nutritional stress. However, his-toric data and more recent studies do not supporta conclusion that no form of nutritional stressoccurred. Instead it appears that sea lions mayhave experienced “chronic” nutritional stressassociated with the high abundances of low-quality species of prey that were present duringthe 1980s and 1990s. This conclusion is based ona growing body of research that includes bloodchemistry comparisons, dietary analyses, popula-tion modeling, and captive feeding studies.

Shifting from a high-energy diet (dominatedby fatty fishes) to one dominated by lower-energyfish (such as walleye pollock) may have signifi-cantly affected young sea lions by increasing theamount of food they would have had to consumeto meet their daily energy needs. Bioenergeticmodels indicate that a yearling sea lion requiresabout twice the relative energy compared to anadult. Recent feeding experiments with captive sealions suggest that it may be physically impossiblefor young sea lions to meet their daily energyrequirements if low-energy prey species dominatetheir diet. Adults who have finished growing andhave lower metabolic needs than young animalsare not similarly constrained and have the stom-ach capacity to consume sufficient quantities ofprey to meet their daily needs.

The overall abundance of Steller sea lion preymay have changed in the mid-1970s because ofchanges in ocean productivity, fishery removals,and/or other ecosystem interactions. Decreasedprey availability could potentially have increasedforaging times and thus the risk of predation. Simi-larly, abundant prey located farther from shorecould also increase foraging times and exposure tokiller whales, which are the principal predators ofsea lions. Survival and reproduction would have

Conceptual model show-ing how sea lion numbersmight be affected by ocean

climate through bottom-up processes. This

hypothesis suggests thatwater temperatures,

ocean currents, and otherclimatic factors determinethe relative abundances of

fish available to eat,which in turn affects sealion health, as measured

by the proportion of bodyfat, rates of growth, and,at a cellular level, oxida-

tive stress. These threeprimary measures of indi-

vidual health ultimatelydetermine pregnancy

rates, birth rates, anddeath rates (through

disease and predation).Also shown are the effects

of human activities thatcould have directly or

indirectly affectedsea lion numbers.

57

ultimately been compromised if sea lions wereunable to efficiently acquire sufficient prey tomaintain normal growth and body condition. Adietary shift to low-energy prey could have furtherexacerbated any effects of decreased prey avail-ability by increasing food requirements.

Differences in diet and relative prey abundanceappear to be associated with pronounced changesin Steller sea lion numbers. The ocean climatecould account for these geographic and temporalpatterns. However, the spatial and temporal pat-terns associated with the available ocean climatedata have not been previously explored in thecontext of Steller sea lion dynamics and foodwebs. The following section therefore beginsevaluating the ocean climate hypothesis by con-sidering the changes that occurred in the oceanichabitats of sea lions in Alaska.

PhysicalOceanographic Data

Physical oceanographic data for the NorthPacific are generally sparse in time and space, andthis is especially true in the Gulf of Alaska. Broad-scale changes over recent decades have beenidentified in sea surface temperature (SST), whichis the most complete set of oceanographic dataavailable. The Gulf of Alaska was predominantlycool in the early 1970s and warmed in the late1970s and throughout the 1980s. There is substan-tial evidence that this was part of a basin-wideregime shift of the North Pacific that commencedduring the winter of 1976-77. These physicalchanges have been linked to a number of responseswithin the ecosystem of the Gulf. For some vari-ables, especially biological ones, the mid-70s tran-sition was not a sharp change, and the durationof the stable time periods before and after the shiftmay have ranged from six years to more thantwenty years.

Besides the issue of detecting significantregime changes from short time series, a greaterproblem lies in identifying the mechanisms bywhich the large-scale physical environmentalchanges drive associated biological regime shifts,which are highly uncertain. Some detailed mecha-nisms have been proposed, but none have yetbeen truly tested and validated with field studies.The large-scale, surface-derived indices such asthe Pacific Decadal Oscillation (PDO), the firstprincipal component of SST north of 20°N in theNorth Pacific, provide little information on how

large-scale climate affects local populations. Theregional dynamics of climate regimes and the tran-sitions between them need to be understoodbefore ecologically relevant, mechanistic-basedindicators of climate state can be developed.

Regional and depth-dependent differences inthe timing and amplitude of important oceanclimate events in the eastern subarctic Pacificcould have caused local differences in ecosystemresponse. Statistical patterns in SST reflect impor-tant large-scale climate impacts in the Gulf of Alaskaassociated both with El Niño events and the 1976regime shift. Moreover, the patterns were of suffi-cient magnitude and duration to potentially fosterchanges in lower trophic productivity and struc-ture. But there is also significant spatial heteroge-neity in long-term SST patterns across the region.An analysis of SST time series reveals five distinctregions, with common variability within the east-ern Gulf of Alaska and the western Gulf of Alaska,as well as the transitional zone to the south. Otheranalyses also revealed this robust east–westasymmetry.

The ocean temperature data show temporal andspatial patterns that are visually correlated withsome of the observed differences in sea lion num-bers and diets. Changes in the seasonality (phaseand amplitude of the seasonal cycle) of importantenvironmental processes may have a large eco-system impact by leading to mismatches in bio-physical coupling. Unfortunately, the availabletemperature data are on a much coarser spatialscale than the fine scales over which sea lions for-age, making it difficult to draw firmer conclusionsin the context of the Steller sea lion decline.

Important changes were observed across theunique mid-70s temporal boundary for winter SST,sea level pressure, and surface wind anomaliesbefore and after 1976-77. The timing of this majorregime shift corresponds to the start of the sealion decline. Comparing ocean climate conditionsacross the 1999 temporal boundary also showssimilarities between the latest period of sea lionstability (and possibly recovery) and the earliercool regime (before 1977). This is noteworthy,given some of the early indications that positivechanges in sea lion diets and numbers in the Gulfof Alaska may have begun with the start of the1999 regime shift. However, the 1999 regime shiftmay not be a reversal to earlier conditions. Signifi-cant differences between regimes (1970–1976 and1999–2002) are evident, such as a strong, displacedAleutian Low with a strengthened North PacificHigh. This suggests that more than two stable

58

climate states may exist, and it adds support to thearguments that a second SST pattern has becomemore important than the PDO in recent years.

The east–west patterns of sea lion populationdynamics are therefore associated with the east–west asymmetries in key physical oceanographicobservations, such as SST and thermocline depth.Atmospherically controlled oceanic forcing func-tions, such as Ekman pumping patterns andstreamflow discharge into the Alaska Current, alsoindicate that basin-scale ocean changes may haveoccurred after the 1976-77 climate shift when thepopulations decreased significantly in the westernGulf. These results imply a linkage between thedistinct observed climate changes and sea lionpopulations, but the specific physical forcingmechanisms affecting the animals are unclear.Numerical simulations can be used to gain furtherinsight into how these physical changes may haveinfluenced the sea lions and other species.

Ocean ModelingBecause of the sparseness of oceanographic

observations in space and time (especially beforethe 1976-77 climate shift), a number of modelingstudies were designed to decipher the physical

processes that may have led to changes in the sealion food web. These studies included hindcastsforced by observed atmospheric variations todetermine the magnitude of phasing of oceanicevents in the water column. They also involvedprocess studies in which the effects of eddies,such as eddy interactions with topography andmean conditions, were explored. Coarse-resolutionmodels allow a broad-scale perspective of thephysical oceanographic changes induced by cli-mate forcing, while eddy-permitting models cansuggest roles for eddies in altering the mean back-ground states of the ocean and driving fluxes ofnutrients across the shelf–slope system.

A coarse-resolution hindcast of the Gulf ofAlaska over the period 1958–1997 showed a shoal-ing of the pycnocline in the central part of the Gulfof Alaska after the mid-1970s and a deepening in abroad band that follows the coast. The deepeningwas particularly pronounced in the northern andwestern part of the Gulf of Alaska, to the south-west of Kodiak Island, where the pycnocline deep-ened by 25–30 m after 1976. The surface forcingresponsible for these changes was the localEkman pumping, which can account for a largefraction of the pycnocline depth changes as alocal response.

Sea surface temperatureanomalies (top panels)and sea level pressureanomalies and surface

wind anomalies (bottompanels) for winter periods

(November–February)before and after the 1976-77 regime shift and for the

most recent period.

59

Changes in the distribution of mesoscale eddiesin the Gulf of Alaska after the 1976-77 regime shiftwere studied in a regional eddy-permitting oceanmodel over the 1950–1999 time period. After theshift, mesoscale eddy variance changed dramati-cally in the western Gulf of Alaska. The conse-quences of this change included altering thecross-shelf/slope mixing of water masses of theopen-ocean and shelf regions. Since mesoscaleeddies provide a mechanism for transportingnutrient-rich open-ocean waters to the productivenear-shore shelf region, the fundamental flow ofenergy through the food web may have beenaltered by this physical oceanographic change.This mechanism may have altered the relativelyabundances of key prey species available toSteller sea lions prior to and following the 1977regime shift.

In contrast to the western Gulf of Alaska, themodel surface velocity variance in the eastern Gulfwas only weakly altered. Hence, an east–westasymmetry occurred in the Gulf of Alaska circula-tion response to the strengthened Aleutian Low.This is consistent with eastern populations ofSteller sea lions in southeast Alaska continuingtheir steady increase across the temporal bound-ary of the 1976-77 climate shift.

Basin-scale models designed to study oceanicprocesses are not of sufficient resolution to inves-tigate coastal ecosystem dynamics. Instead, limited-domain models of ocean circulation with higherresolution allow focused, regional studies of criti-cal processes and circulation. Such an approachallows for proper representation of the complexflow–topography interactions and their influenceon exchanges between adjacent water masses andthrough the Aleutian Island passes.

A pan-Arctic coupled sea ice–ocean model pro-vides insight into the circulation and exchangesbetween the subarctic and Arctic basins, particu-larly on the exchange between the North PacificOcean and the Bering Sea through the passes ofthe Aleutians, which can influence biological pro-ductivity along the Aleutian Island chain. Model-simulated eddies along the Alaskan Stream havesignificant influence on both the circulation andthe water mass properties across the eastern andcentral Aleutian Island passes. Eddy-relatedupwelling of salty water along the southern slopeaffects the water column down to about 1000 m.Given the high correlation between salinity andnutrient content at depths, the increased salinityin the upper ocean over the pass can representnutrient input for enhanced and/or prolonged pri-

mary productivity. Since modeled eddies along theAlaskan Stream occur throughout a year, theircontribution to high surface nutrient concentra-tions within the Aleutian Island passes could beespecially significant during otherwise low pri-mary productivity seasons. This effect would bemost important during years in which mesoscaleeddies frequently propagate along the AlaskanStream.

There is therefore evidence that climate-forcedchanges occurred in both the strength of the meancurrents of the Alaskan Stream and the spatialdistribution of the mesoscale eddy field of theAlaskan Stream after the mid-70s climate regimeshift. These changes are strongest in the westernGulf, where sea lion populations experienced pre-cipitous declines during the same period. Oceanmodels also demonstrate that mesoscale eddiesprovide an important mechanism for mixing nutrient-rich waters with nutrient-depleted waters alongthe Alaskan Stream and across the Aleutian Islandpasses. Hence, the flow of energy through theecosystem in the Gulf may have been fundamen-tally altered by changes in these basic physicaloceanographic changes. There are also other indi-cations that the broad suite of concurrent foodweb changes that occurred at basin and regionalscales were influenced by the effects of physicaloceanography.

Ecosystem andBiogeographic Links

The oceanographic studies described thusfar provide evidence of medium- and long-termchanges in the physical dynamics in the northernGulf of Alaska and Aleutian Islands. It is thereforereasonable to expect these changes to be reflectedin observations of the broad-scale ecosystem andthe biogeography of the regional fauna. Severalstudies have addressed these issues.

A nonlinear analysis of a multivariate data setalso captures the pattern of decline of Steller sealions. The data contained time series for such vari-ables as annual salmon landings for five speciesand three regions in Alaska, rockfish and herringrecruitment indices, herring biomass, and zoo-plankton biomass estimates for subregions of theGulf of Alaska and Bering Sea. The main result isa pattern with all positive scores from 1965–1979and all negative scores from 1980–2000.

Research cruises to the passes of the easternand central Aleutian Islands revealed a number of

60

intriguing biogeographic features of the region thatcorrespond to the population and dietary divi-sions of sea lions. Sharp fronts in surface salinitywere found at Unimak and Samalga Passes thatcoincided with demarcation points for sea lion dietsand population dynamics. Samalga Pass appearsto be a boundary between shelf waters to the eastand open-ocean waters to the west, with the AlaskaCoastal Current influencing the waters east of thepass and the Alaskan Stream water influencing thewaters farther west. The difference in source watersin the two regions likely affects the distributionsof nutrients and biota around the different passes.

Changes in the abundance and composition ofzooplankton species are associated with seasonalchanges in water mass and other physical proper-ties along the island chain. Declines in the abun-dance of Neocalanus plumchrus and N. flemingeriat Akutan and Unimak Passes in June resulted asthese species left the surface waters and migrateddown to depths over 300 m. Elevated abundancesof Calanus marshallae and Acartia spp. atUmnak, Akutan, and Unimak Passes were dueto their preference for warmer neretic conditions.Abundances of two species of euphausiid alongthe islands showed a preference by Thysanoessainermis for the neretic waters of Akutan, Unimak,and Samalga Passes and a preference by Euphau-sia pacifica for the open-ocean conditions of thepasses west of Samalga Pass.

In addition to zooplankton and fish, the west-ern extent of the Alaska Coastal Current also oper-ates as a biogeographical “boundary” for seabirdsaround Samalga Pass. Seabirds depending oncoastal food webs (shearwaters and puffins) are

more abundant east of Samalga, whereas seabirdsdepending on oceanic food webs (fulmars andauklets) are more abundant west of Samalga. Ful-mars and shearwaters consume oceanic copepodsand shelf-break species of euphausiids west ofSamalga, while both of these seabirds consumeshelf species of euphausiids east of Samalga.

At-sea surveys of seabirds during cruises in2001 and 2002 showed large-scale patterns of avi-fauna that suggest that the central and westernAleutian Islands support the vast majority ofbreeding seabirds dependent on oceanic zoo-plankton, whereas the eastern Aleutian Islandssupport the majority of piscivorous seabirds.These data support the premise that the easternAleutians provide a very different habitat thanregions of the archipelago west of Samalga Pass.At smaller spatial scales, the passes of the Aleu-tian Islands are the focal points of much seabirdforaging activity.

Changes in the benthic and pelagic fish com-munities within the Gulf of Alaska and AleutianIslands in response to the regime shift of 1976were dramatic. Population increases occurred inflatfish, gadids, and salmonids as a result of anincrease in the frequency of strong year classesafter 1976. At around the same time, decreasesoccurred in shrimp and crab stocks. Small-meshtrawl surveys conducted near Kodiak Islandbetween 1953 and 1997 documented the “commu-nity reorganization” in the Gulf of Alaska. Thecatch composition of the trawl catches in this sur-vey prior to 1977 was dominated by forage speciessuch as capelin and shrimp. Following the regimeshift, the catches were primarily high-trophic-levelgroundfish.

Further up the food chain, concurrent changeswere also noted in the few populations of marinemammals that have been counted since the 1960sand 1970s. At Tugidak Island, for example, thelargest population of harbor seals in Alaska beganan unexplained decline in the mid-1970s, falling toless than 20% of its peak abundance by the mid-1980s. Steller sea lions numbers at the nearbyrookery on Chowiet Island also fell rapidlythrough the 1980s. Similarly the Pribilof Islandspopulation of northern fur seals, which accountsfor about 80% of the world population, alsodeclined unexpectedly from the late 1970s to mid-1980s. All three populations of pinniped speciesappear to have declined at about the same time,coincidental with the 1976-77 regime shift.

These broad-scale ecological changes acrossall trophic levels are generally coincident in time

Nonlinear principalcomponent (NLPC)

analysis results from amultivariate data set of 45biotic indices (fishery andsurvey records) that span

broad regions over theBering Sea, Gulf of

Alaska, and North PacificOcean from 1965 to 2001.

The top panel shows thefirst (and dominant)

characteristic time seriesthat emulates the pattern

of sea lion decline and the1976-77 climate shift. The

bottom panel shows thesecond characteristic time

series that exhibitsinterannual variability.

61

and are widely believed to be driven by differencesand changes in the oceanic environment. This isnot to say that the other primary force affectingfish populations (fishing) is without impact. Fish-ing can, and does, affect community dynamics.The effect of fishing is added to natural sources ofvariability. Archaeological studies have repeatedlydemonstrated wide swings in the abundance offish species long before the development of large-scale fisheries. Generally, fishing impacts the adultportion of fish populations. An important linkbetween climate and population size occurs at thelarval and juvenile stages for many marine animals.Making the transition from egg (marine fishes) orsmolt (salmon) to successful recruit requires oce-anic and ecological conditions conducive to sur-vival. Under the regime-shift hypothesis, certainspecies are favored under one set of ocean condi-tions while other species flourish when conditionschange abruptly.

The effects of broad-scale changes in oceanclimate on Steller sea lion habitat appear to bemoderated through a number of indirect mecha-nisms. For example, increased storm activity mighthave reduced the suitability of certain hauloutsand rookeries, while bottom-up effects mediatedthrough at least three trophic levels (such as phy-toplankton, zooplankton, and forage fish) have thepotential to affect the distribution of Steller sealion prey species. In light of the spatial distribu-tions of different species in the food web, and thepotential foraging distances of individual sealions, further range-wide studies encompassingareas of both decreased and increased habitat

suitability will be required to fully elucidate theeffects that changing climate can have on apexpredators.

In summary, a suite of changes occurred acrossall trophic levels of the Gulf of Alaska and Aleu-tian Islands ecosystems that corresponded to thetiming of the 1976-77 regime shift and the declineof Steller sea lions in the western Gulf of Alaskaand Aleutian Islands.

The detailed regional influences of these cli-mate changes are difficult to pinpoint among thesparsely observed populations, but they appearto be modulated by the effects of biogeographicfeatures such as the Samalga Pass transition fromcoastal influence to open-ocean conditions andthe fine structure of island distributions. Thesetransition points delineate distinct clusterings ofprey species, which are in turn correlate with dif-ferential population sizes and trajectories of Stellersea lions. The results support the idea that a fun-damental change in the ecosystem occurred afterthe mid-70s, which may have cascaded up throughthe food web to influence the regional diets andhealth of sea lions. Other studies suggest thatsuch changes were not unique to the 20th century.

PaleoecologicalPerspective

Paleoecological studies provide a long-termperspective on changes seen in recent decades.Indicators of oceanic productivity in two sedimentcores, one from the central Gulf of Alaska shelfand one from the Bering Sea (Skan Bay), showedthat considerable variability has occurred in oceanproductivity over the last 150 years for eachregion. In the Gulf of Alaska, two productivityproxies increased since the 1976-77 regime shift,while the two proxy signals were mixed in the Ber-ing Sea. The Bering Sea data imply significantchanges in the phytoplankton community. Suchchanges in productivity could have affected theflow of energy up the food web and altered thefavored species upon which Steller sea lions feed.This paleoecological record averages out seasonalchanges, which may be an important effect inaddition to total productivity. The regional differ-ences in the paleoecological records may beimportant in explaining regional differences inthe numbers and diets of Steller sea lions.

Long-term changes in the North Pacific andsouthern Bering Sea ecosystems have also beenthe subject of intensive investigations using

Conceptual modelshowing how regime

shifts might havepositively or negatively

affected sea lion numbersthrough bottom-up

processes that influencedsuites of species and

subsequently affected sealion health and numbers.

62

archaeological and anthropological data. Thearchaeological data indicate that significant varia-tions occurred in the distributions of key speciesover the last 5000 years. Correlations betweenchanges in the relative abundances of speciessuch as Steller sea lions and regional climaticregimes are only suggestive at this time, withcooler periods having near-average harvests ofsea lions by Aleuts and warmer periods havingbelow-average harvests. Notably, the greatestabundance of Steller sea lions occurred during theLittle Ice Age, which may be significant. The endof the Little Ice Age also appears to have coincidedwith a reduction in Steller sea lion populations.

Decadal-scale changes in the marine ecosystemspanning nearly 150 years are identifiable usingboth ethnohistoric data and traditional ecologicalknowledge of local Aleut fishermen. Based onRussian and early American accounts of the region,there have been two periods in the last 250 years—one in the 1870s, coinciding with a warming periodas observed in the Sitka air temperatures, andanother in the 1790s—when there were few or noSteller sea lions in many areas of the North Pacific,leading to widespread starvation for the indige-nous peoples who depended on them. Thesedepressions in the population levels cannot becorrelated with any human-based harvesting ofeither the sea lions or their food sources.

Traditional knowledge of local fishermen indi-cates that the North Pacific ecosystem underwenta series of disruptions over the last 100 years thatmay or may not have been caused by commercialfishing. For example, the North Pacific was heavilyfished for cod between the 1880s and the mid-1930s, when the fishery collapsed. Cod appear tohave been completely absent in many areas southof the Alaska Peninsula between 1945 and 1970,during which time shrimp and crab were dominantcomponents of the ecosystem. The extent towhich these changes were mitigated by predator–prey interactions, fishing, or changes in ocean cli-mate is not known. However, it is interesting thatthe Aleut term for codfish can be rendered intoEnglish as “the fish that stops,” meaning that itdisappears periodically. It is also noteworthy thatthe major shifts in species abundances line up rea-sonably well with the major documented regimeshifts recorded over the past century.

In summary, the archaeological and anthropo-logical analyses provide data on time scales thatare currently not available in any other form ofanalysis. They demonstrate that the North Pacificand southern Bering Sea have been dynamic and

volatile and subject to great fluctuations over thelast hundreds to thousands of years. This requirescareful evaluation of current models to determinewhere sea lion populations are currently posi-tioned within the large-amplitude swings in popu-lation sizes that are evident from the past. Theresults also provide additional evidence that cli-mate may very well underpin ecosystem restruc-turings that can be manifest as large, regionalchanges in Steller sea lion abundance.

SummaryWe examined the hypothesis that the decline

of the Steller sea lion populations in the AleutiansIslands and Gulf of Alaska is a consequence ofphysical oceanographic changes caused by the1976-77 climate shift. The available data suggestthat ocean climate can affect the survivorship ofkey species of prey consumed by Steller sea lions.It is therefore conceivable that a change in climaticconditions following the 1976-77 regime shift mayhave enhanced the survivorship and distributionof leaner species of prey (such as pollock, flatfish,and Atka mackerel), which in turn negativelyaffected the survival of young sea lions from 1977to 1998. Thus, physical environmental changescould have had consequential effects on thehealth and fecundity of Steller sea lions. Highertemperatures appear to be associated with anincreased abundance of cod and pollock, while areturn to cooler temperatures would favor Stellersea lions.

In broad terms, the suite of studies that havebeen undertaken to investigate the temporal andspatial differences in ocean climate in the NorthPacific have identified ocean climate patterns thatare consistent with the patterns of sea lion distri-butions, population trends, numbers, and diets.The oceanic response to climate forcing after1976-77 has an east–west asymmetry, with strongerchanges occurring in the western Gulf of Alaska.The geographic clustering of sea lion diets andpopulation trajectories, and their correspondencewith key biogeographic and oceanographic fea-tures of the Gulf of Alaska and Aleutian Islands,add credence to the view that there is a linkagebetween Steller sea lions and the physical envi-ronment. However, additional studies will berequired on finer spatial scales to draw firmer con-clusions, particularly in regions closer to shore,where sea lions spend more time foraging.

Our assessment of the ocean climate hypothe-sis does not discount the other leading hypothe-

63

ses that have been proposed to explain the declineof Steller sea lions, such as the nutritional stresshypothesis, the fishing hypothesis, the diseasehypothesis, and the killer whale predation hypoth-esis. Instead, the ocean climate hypothesis pro-vides a holistic framework within which each ofthe alternative hypotheses can be aligned.

ReferencesAlverson, D.L. (1992) A review of commercial fish-

eries and the Steller sea lion (Eumetopias juba-tus): The conflict arena. Reviews in AquaticSciences, Vol. 6, p. 203–256.

Benson, A.J., and A.W. Trites (2002) Ecologicaleffects of regime shifts in the Bering Sea andeastern North Pacific Ocean. Fish and Fisher-ies, Vol. 3, p. 95–113.

Calkins, D.G., D.C. McAllister, and K.W. Pitcher(1999) Steller sea lions status and trend insoutheast Alaska: 1979–1997. Marine MammalScience, Vol. 15, p. 462–477.

DeMaster, D., and S. Atkinson (ed.) (2002) StellerSea Lion Decline: Is It Food II? University ofAlaska Sea Grant, Fairbanks, 80 p.

Hare, S.R., and N.J. Mantua (2000) Empirical evi-dence for North Pacific regime shifts in 1977and 1989. Progress in Oceanography, Vol. 47,p. 103–146.

Loughlin, T.R. (1998) The Steller sea lion: A declin-ing species. Biosphere Conservation, Vol. 1, p.91–98.

Marzban, C., N. Mantua, and S. Hare (2005) Retro-spective study of climate impact on AlaskaSteller sea lion: A report. Technical Report No.485, Department of Statistics, University ofWashington. Available at http://www.stat.washington.edu/www/research/reports/.

Merrick, R.L., M.K. Chumbley, and G.V. Byrd (1997)Diet diversity of Steller sea lions (Eumetopiasjubatus) and their population decline in Alaska:A potential relationship. Canadian Journal of

Fish and Aquatic Sciences, Vol. 54, p. 1342–1348.

Miller, A.J., D.R. Cayan, T.P. Barnett, N.E. Graham,and J.M. Oberhuber (1994) The 1976-77 climateshift of the Pacific Ocean. Oceanography, Vol.7, p. 21–26.

National Research Council (2003) The Decline ofthe Steller Sea Lion in Alaskan Waters: Untan-gling Food Webs and Fishing Nets. NationalResearch Council, Washington, DC, 216 p.

Peterson, W.T., and F.B. Schwing (2003) A newclimate regime in northeast Pacific ecosystems.Geophysical Research Letters, Vol. 30, ArticleNo. 1896.

Rosen, D.A.S., and A.W. Trites (2000) Pollock andthe decline of Steller sea lions: Testing the junkfood hypothesis. Canadian Journal of Zoology,Vol. 78, p. 1243–1250.

Savinetsky, A.B., N.K. Kiseleva, and B.F. Khas-sanov (2004) Dynamics of sea mammal and birdpopulations of the Bering Sea region over thelast several millennia. Palaeogeography,Palaeoclimatology, Palaeoecology, Vol. 209, p.335–352.

Stabeno, P.J., N.A. Bond, A.J. Hermann, N.B.Kachel, C.W. Mordy, and J.E. Overland (2004)Meteorology and oceanography of the North-ern Gulf of Alaska. Continental Shelf Research,Vol. 24, p. 859–897.

Trites, A.W., A.J. Miller, H.D.G. Maschner, M.A.Alexander, S.J. Bograd, J.A. Calder, A. Capoton-di, K.O. Coyle, E. DiLorenzo, B.P. Finney, E.J.Gregr, C.E. Grosch, S.R. Hare, G.L. Hunt, J.Jahncke, N.B. Kachel, H.-J. Kim, C. Ladd, N.J.Mantua, C. Marzban, W. Maslowski, R. Men-delssohn, D.J. Neilson, S.R. Okkonen, J.E.Overland, K.L. Reedy-Maschner, T.C. Royer,F.B. Schwing, J.X.L. Wang, and A.J. Winship(2005) Bottom-up forcing and the decline ofSteller sea lions in Alaska: Assessing the oceanclimate hypothesis. Fisheries Oceanography,submitted.

64

Status of Alaska Groundfish Stocks

This article was preparedby Anne B. Hollowed and

James N. Ianelli, ofNOAA’s National Marine

Fisheries Service.

The Alaska Fisheries Science Center of theNational Marine Fisheries Service produces stockassessments for major groundfish and shellfishstocks in the Alaskan waters on an annual basis.Stock assessment and fishery evaluation reportsare prepared for the North Pacific Fishery Man-agement Council meetings and use the assess-ments to recommend levels of acceptable biologi-cal catch (ABC).

The Alaska groundfish management system isbased on extensive data available from the NationalMarine Fisheries Service observer program anddedicated research cruises. Catches of target andprohibited species (such as salmon, crab, herring,and Pacific halibut) are estimated at sea or in pro-cessing plants to provide real-time information toensure that fisheries do not exceed total allowablecatches (TACs) or violate other fishery restric-tions (such as time–area closures). Dedicatedresearch cruises coupled with observer data makeit possible to build detailed population dynamicsmodels. The results of these modeling activitiesare used to determine the status of individualspecies.

The first step in determining the TAC beginswith the preparation of stock assessment andfishery evaluation reports. These reports containanalyses summarizing the information about theindividual stocks and groups and include ABCand overfishing level (OFL) recommendations forfuture years. The authors of these reports (gener-ally National Marine Fisheries Service scientists)present their findings to the North Pacific FisheryManagement Council’s groundfish Plan Teamseach September and November. At these meetingsthe reports are reviewed and recommendations forABC levels are compiled into two stock assess-ment and fishery evaluation report volumes (oneeach for the Bering Sea/Aleutian Islands and theGulf of Alaska regions). In addition, the Plan Teamrecommendations for ABCs are presented. Thecompiled reports are then submitted to the North

Pacific Fisheries Management Council’s Scientificand Statistical Committee for further review. Thiscommittee makes the final ABC recommendation tothe Council, and the Council’s Advisory Panel ofindustry representatives makes TAC recommenda-tions. Finally, the recommended TAC levels areadjusted (for some species) by the Council toensure that other constraints (for example, limitingthe sum of all allowable catches in the Bering Seaand Aleutian Islands to be less than 2 million tons)are met. The following rule applies for all federallymanaged groundfish species in a given year:Catch < TAC < ABC < OFL.

In practice, catch is often much less than theTAC, and the TAC is often much less than theABC. The multispecies management system is,therefore, based on the premise that no individualcomponents are overfished or below stock sizesthat are considered detrimental to the ecosystem.Stock assessments can be obtained at www.afsc.noaa.gov/refm/stocks/assessments.htm.

A change in the timing requirements for con-ducting assessments was implemented in 2004.Based on an analysis conducted by scientists atthe NOAA Alaska Fisheries Science Center incoordination with the NOAA National MarineFisheries Service’s Alaska Regional office, it wasfound that for longer-lived species, managementadvice on quotas could be based on biennialassessments. This cycle was designed to coincidewith the current Alaska Fisheries Science Centersurvey regularity.

Presently, the main species of groundfish are allabove their target stock size, and 2004 catch levelswere below the maximum permissible ABC levels.During 2001–2003, fisheries for these groundfishspecies yielded 2.1 million metric tons annually,valued at $615 million. The abundances of themajor stocks of Alaska pollock and Pacific cod arehigh but subject to variability because of recruit-ment fluctuations. Virtually all flatfish resources(for example, rock sole, yellowfin sole, Alaska

65

plaice, and arrowtooth flounder) are at high andhealthy levels. Atka mackerel abundance isincreasing and above average levels. Rockfishspecies comprise 5–8% of the groundfish complexbiomass and are at healthy and stable levels. Forthe main stocks with age-structured analyses, thebiomass trends for the Bering sea and AleutianIslands regions suggest that stock conditions arefairly evenly split between those that are aboveaverage and those that are below in the past fewyears.

Data limitations make it difficult to assess less-abundant (minor) rockfish species. Together withother non-target species (such as sharks, skates,sculpins, and octopus), accurately assessing thevulnerability of these species represents a majorchallenge for NOAA. Efforts to monitor the statusof non-target species have improved, and stepshave been taken to ensure that adequate data col-lection programs are in place in advance of directedfishery development.

Relative 2005 spawning stock size compared to the target stock size versus relative2004 catch levels compared to 2004 maximum permissible Acceptable Biological Catch(ABC) levels for Bering Sea–Aleutian Island stocks.Values below the red line indicatethat the catch levels in 2004 are less than the ABC estimated for that year. Values to theright of the blue line indicate that the spawning stock biomass projected to 2005 isgreater than the level that would theoretically provide the MSY (maximum sustainableyield).

Biomass trends for Bering Sea and Aleutian Islands stocks relative to their mean level,1978–2004.

66

Status of Alaska’s Salmon Fisheries

This article was preparedby William R. Heard, of

NOAA’s Alaska FisheriesScience Center, Auke

Bay Laboratory,Juneau, Alaska; and

Loh-Lee Low, of NOAA’sAlaska Fisheries Science

Center, Resource Ecologyand Fisheries Manage-

ment Division.

Pacific salmon have played a pivotal role in thedevelopment and history of Alaska. Many Alas-kans still depend on salmon for recreation, food,and employment. The fishing industry has providedAlaska with its largest private sector employment,and subsistence use of salmon is still an importantpart of the life of rural Alaskans.

Monitoring the status of salmon involves mea-suring salmon run sizes, including both catch andescapement information. However, other factorsare also used to measure the well being of salmonstocks, such as fish sizes, fish quality, disease andinfestations, and run timing. All of these character-istics are difficult to quantify and standardize. Evenrun size data (particularly data on escapement sizes)are hard to obtain because the salmon escapementroutes into different river systems are usually notwell known. The primary goal of salmon managementis to allow adequate escapement so that spawningand replenishment of the salmon stock will occur.Thus, the yearly variation of salmon catch, afterallowing escapement, may be used as the mostconsistent measure of the status of salmon stocks.

levels in many regions of the state. The record-high commercial landing of 218 million salmon in1995 was 17% higher than the previous record of196 million in 1994. Throughout the mid- to late1990s, recreational and subsistence fishermenharvested 2–3 million salmon annually.

Significant declines in commercial catches dur-ing the three years following the peak harvest in1995 were thought by many to indicate that amajor downturn in productivity of Alaska salmonmight have started. The history of commerciallandings shows a distinct cyclic pattern of highand low harvest levels often lasting decades.Much of this fluctuation is now believed to becaused by interdecadal climate oscillations in theocean environment that affect the marine survivalof juveniles. A major climatic regime shift thatoccurred in 1977 helped Alaska stocks reboundfrom the previous low-cycle years, while anotherregime shift in 1989 may portend a future down-ward trend in Alaska’s salmon resources.

An interesting correlation associated withAlaska’s cyclic salmon harvest is an inverse pro-duction regime with abundance levels of WestCoast salmon. Recent increases in the numbers ofWest Coast salmon, therefore, may also suggest adeclining trend for Alaska salmon. However, thecorrelation is still debatable. Alaska’s commercialcatch did decline for three years after the record1995 harvest, but it rebounded to 217 million fishin 1999, essentially matching the peak harvest yearcatch. Landings in 2000 fell to 137 million salmon,increased to 175 million fish in 2001, dropped to131 million in 2002, and then increased to 173 mil-lion in 2003. All of these recent Alaska harvestswere well above the long-term norm for Alaska,while West Coast salmon runs have continued torebound from their lows.

A number of factors have contributed to thecurrent high abundance of Alaska salmon:

• Pristine habitats with minimal impacts fromextensive development;

Spawning salmon.

Overall Harvest LevelsAlaska commercial salmon harvests have gen-

erally increased over the last three decades. Afterreaching record low catch levels in the 1970s, mostpopulations have rebounded, and fisheries inrecent years have been at or near all-time peak

67

• Generally favorable ocean conditions thatallow high survival of juveniles;

• Improved management of fisheries by stateand Federal agencies;

• Elimination of high-seas drift-net fisheriesof foreign nations;

• A well-managed hatchery production system;and

• Reductions of salmon bycatch in fisheries forother species.

ManagementAlaska’s 44,000-mile coast is nearly two-thirds

the length of the coastline of the United States.Along the Alaska coastline, over 14,000 waterbodies support populations of five species ofsalmon. Salmon management over such a vast arearequires a complex mixture of domestic and inter-national bodies, treaties, regulations, and otheragreements. Federal and state agencies cooperatein managing salmon fisheries. The Alaska Depart-ment of Fish and Game (ADF&G) manages salmonfisheries within state jurisdictional waters, wherethe majority of the harvest occurs. ADF&G’s prin-cipal salmon conservation policy uses escapement-based management, providing adequate spawningescapement into natal streams over any preseasonharvest goals. Management in the Exclusive Eco-

nomic Zone (EEZ), 3–200 nautical miles offshore,is the responsibility of the North Pacific FisheryManagement Council (NPFMC), which hasdeferred specific regulations to the State ofAlaska. Management of Alaska’s salmon fisheriesis based primarily on regional stock groups ofeach species and on time and area harvestingusing specific types of fishing gear.

Over 25 commercial salmon fisheries in Alaskaare managed by a special limited-entry permit sys-tem that specifies when, where, and what type offishing gear can be used in each area of the state.The commercial gear used includes drift gillnets,set gillnets, beach seines, purse seines, hand troll,power troll, or fish wheel harvest gear. Sport fish-ing is limited to hook and line, while subsistencefishers may use gillnets, dip nets, or hook and line.Some subsistence harvesting of salmon is alsoregulated by special permits, and in some rivers,fish wheels can be used for subsistence fishing.

Management of some Alaska salmon fisheriesis also negotiated with Canada under the PacificSalmon Treaty. Some major issues of concernoccur in Southeastern Alaska, while others occurin the Yukon River area.

On a broader international scope, the manage-ment of salmon harvest in the high seas of theNorth Pacific Ocean from 1957 to 1992 was autho-rized by the International North Pacific FisheriesCommission (INPFC) and via bilateral and multilat-eral agreements and negotiations with Taiwan andthe Republic of Korea. In 1993 the North PacificAnadromous Fish Commission (NPAFC) wasformed to replace the INPFC. Membership of theNPAFC has now expanded from the original threecountries (Canada, Japan, and the U.S.) to includethe Russian Federation and the Republic of Korea.

The NPAFC Convention prohibits high seassalmon fishing and trafficking of illegally caughtsalmon. Coupled with United Nations GeneralAssembly Resolution 46/215, which bans large-scale pelagic driftnet fishing in the world’s oceans,legal harvesting of Pacific salmon on the high seasno longer exists. This allows for effective manage-ment control to fully return to the salmon-producingnations.

Because salmon are anadromous species thatspend up to seven years of their lives at sea andthen return to freshwater streams, rivers, and lakesto spawn and die, the well-being of salmon inAlaska, in addition to harvest management prac-tices, is also directly influenced by land manage-ment practices. The quality of freshwater habitatsdetermines the success of reproduction and initial

Alaska salmon catch ofall species combined, innumbers of fish and in

metric tons.

68

rearing of juveniles. Several entities have signifi-cant influence on the quality of freshwater spawn-ing and rearing habitats for salmon throughoutAlaska. Among these are the U.S. Forest Service,the Bureau of Land Management, the NationalPark Service, the U.S. Fish and Wildlife Service,Alaska state parks and forests, Alaska Nativeregional and village corporations, municipalities,boroughs, and private landowners that controlwatersheds used by salmon.

Species and StatusAll five species of Alaska salmon—pink, sock-

eye, chum, coho, and chinook—are fully utilized,and stocks in most regions of the state generallyhave rebuilt to near or beyond previous high lev-

els. Although in recent years there has been ahigh statewide abundance of salmon in Alaska,there are also issues of serious concern for theseresources, especially with some species in certainregions. For example, stocks in western Alaska,especially chinook and chum salmon, generallyhave been at very depressed levels since the mid-1990s. Also, some of the same factors affectingdeclines of salmon in the Pacific Northwest—issues associated with overfishing, incidental takeas bycatch in other fisheries, and losses of spawn-ing and rearing habitats in freshwater and in near-shore ocean environments—are of concern insome areas of Alaska.

Recreational fishing for salmon continues togrow and is an important component of the Alas-kan lifestyle. This is partly because many Alaskan

Alaska Salmon Catch by Species in Numbers of Fish, 1970–2003.

Pink Sockeye Chum Coho ChinookYear Salmon Salmon Salmon Salmon Salmon Total

1970 31,096,000 27,622,000 7,476,000 1,524,000 645,000 68,363,0001971 23,539,000 14,177,000 7,679,000 1,444,000 661,000 47,500,0001972 15,913,000 6,999,000 6,655,000 1,834,000 554,000 31,955,0001973 9,805,000 4,448,000 5,928,000 1,455,000 550,000 22,186,0001974 9,857,000 4,789,000 4,698,000 1,860,000 559,000 21,763,0001975 12,987,000 7,458,000 4,323,000 1,014,000 455,000 26,237,0001976 24,755,000 11,779,000 5,924,000 1,432,000 531,000 44,421,0001977 28,647,000 12,465,000 7,326,000 1,789,000 620,000 50,847,0001978 53,852,000 18,140,000 6,677,000 2,821,000 836,000 82,326,0001979 50,137,000 28,696,000 5,608,000 3,122,000 779,000 88,342,0001980 63,304,000 33,295,000 9,603,000 3,115,000 675,000 109,992,0001981 60,089,000 36,348,000 12,613,000 3,416,000 823,000 113,289,0001982 64,859,000 28,954,000 10,994,000 6,040,000 877,000 111,724,0001983 60,359,000 52,875,000 10,222,000 3,636,000 828,000 127,920,0001984 76,343,000 38,450,000 13,096,000 5,405,000 667,000 133,961,0001985 90,335,000 38,983,000 10,570,000 5,749,000 721,000 146,358,0001986 77,320,000 32,208,000 12,510,000 6,293,000 616,000 128,947,0001987 46,493,000 35,431,000 10,527,000 3,493,000 680,000 96,624,0001988 50,358,000 30,038,000 15,105,000 4,473,000 589,000 100,563,0001989 96,869,000 44,139,000 7,896,000 4,650,000 572,000 154,126,0001990 88,208,000 52,693,000 8,010,000 5,478,000 666,000 155,055,0001991 128,336,000 44,646,000 9,769,000 6,153,000 613,000 189,517,0001992 60,597,000 58,283,000 10,223,000 7,095,000 606,000 136,804,0001993 109,631,000 64,314,000 12,238,000 6,050,000 667,000 192,900,0001994 116,720,000 52,816,000 16,135,000 9,551,000 640,000 195,862,0001995 128,333,000 63,532,000 18,796,000 6,471,000 663,000 217,795,0001996 97,310,000 50,270,000 21,856,000 6,150,000 500,000 176,086,0001997 71,280,000 30,910,000 15,620,000 2,900,000 650,000 121,360,0001998 104,770,000 22,720,000 19,070,000 4,680,000 580,000 151,820,0001999 145,990,000 45,120,000 20,480,000 4,590,000 430,000 216,610,0002000 74,800,000 33,500,000 24,290,000 4,200,000 360,000 137,150,0002001 127,620,000 26,520,000 15,400,000 4,950,000 370,000 174,860,0002002 87,310,000 22,211,000 16,210,000 5,059,000 584,000 131,374,0002003 121,696,000 31,013,000 15,931,000 4,105,000 599,000 173,344,000

69

households use sport fishing as a convenientmethod to collect seafood for the table. Some partof the total sport fish harvest of salmon in Alaska,therefore, might more appropriately be included insubsistence fishery statistics. Much of the recentgrowth in sport fishing is because of increasedguided recreational fishing by tourists visitingAlaska. A total of 392,980 Alaska sport fishinglicenses were issued in 2002, with 71% issued tononresident anglers. More nonresident sport fish-ing licenses have been sold in Alaska than resi-dent licenses since 1990. Sport fishing for salmonis a vital part of the recent rapid growth in Alaskantourism. Coho salmon were the most popularsport-caught salmon in Alaska, representing 38%of the 3.2 million salmon caught by recreationalfishermen in 2002, followed by pink salmon (25%),sockeye salmon (21%), chum salmon (7%), chi-nook salmon (5%), and non-anadromous land-locked salmon (4%).

Pink SalmonPink salmon are the most abundant of Pacific

salmon in Alaska, accounting for 40–70% of thetotal harvest each year. During the past 33 years

(1970–2003), pink salmon made up 58% of theaverage annual commercial harvest of salmon inAlaska. Pink salmon are harvested mostly in thesoutheastern, southcentral, and Kodiak Islandregions of the state.

Unique among Pacific salmon, pink salmonhave a fixed life-history cycle whereby the speciesalways matures and spawns at two years of age.This cycle is genetically fixed, so spawners ineven-numbered years are always separate and dis-tinct from spawners in odd-numbered years.

Throughout much of its range, the specieshas viable populations in both odd- and even-numbered years; however, in some areas, pinksalmon only occur in one or the other cycle year.In Bristol Bay and western Alaska, for example,pink salmon, near the effective limit of theirnorthern range, essentially occur mostly in even-numbered years, whereas in the Pacific Northwest,near the effective limit of their southern range,they occur primarily in odd-numbered years.

Sockeye SalmonSockeye salmon, second in abundance among

species caught in Alaska fisheries, generallyaccounted for 27% of the harvest in recent years.Sockeye salmon, however, provide greater dollarvalue to fishermen than all other commerciallycaught salmon in Alaska combined, usually yield-ing 60–70% of the ex-vessel value of the annualharvest. In more recent years, however, worldsalmon prices have declined significantly but haverebounded a little in 2004.

The largest fisheries for sockeye salmon occurin Bristol Bay, Cook Inlet, the Alaska Peninsulaand Aleutian Islands, and the Kodiak regions, butthere are also significant fisheries for this salmon

Alaska pink salmon catch, 1970–2003.

Pink salmon fry.

Alaska pink salmon catch,1970–2003.

70

in the southeastern, Prince William Sound, andChignik regions.

The most common sockeye salmon life-historypattern is for juveniles to rear in lakes for 1–2years before migrating seaward as smolts. Thelarge lake complexes on Bristol Bay rivers providethis necessary life-history component and form acritical part of the important fishery in this region.The Bristol Bay fishery is concentrated in a nar-row window of time from late June until mid-July,when millions of returning adult sockeye salmonpour into Bristol Bay rivers from the ocean.

During the five-year period from 1992 to 1996,returns to Bristol Bay ranged from 29.6 to 44.4 mil-lion and averaged 36.5 million sockeye salmon peryear. The return to Bristol Bay in 1997, however,was only 18.9 million, with a fishery harvest of 12.1million. This unexpectedly low return of sockeyesalmon created a serious shortfall in the catch andin the incomes of fishermen and communities.

As bad as the 1997 sockeye salmon harvest inBristol Bay was, commercial landings in 1998 wereeven worse, with a harvest of only 10.0 millionfish. Returns improved somewhat in 1999 and2000, with commercial catches of 26.1 and 20.5million sockeye salmon, respectively, but morerecently have declined again, with commercialcatches of only 14.2, 10.7, and 14.9 million fish,respectively, in 2001, 2002, and 2003. All of theserecent harvest levels of sockeye salmon in BristolBay are well below previous decadal averages.

Several hypotheses have been suggested toexplain recent shortfalls of sockeye salmon return-ing to Bristol Bay. Unusually warm, calm weatherduring summers has resulted in high water temper-atures, which may have caused high mortality andchanges in the migration behavior of adult salmonentering Bristol Bay. Other suggested causesinclude changes in freshwater or ocean rearing

conditions that affect the growth and survival ofjuveniles or immature adults, increased predationat sea, interception by other fisheries, disease,and, in some instances, over-escapements onspawning grounds. The true causes of theseshortfalls, which likely involve a combinationof many factors, remain unknown. A paramountunanswered question, however, is whether or notcyclic changes in oceanic environmental condi-tions have occurred that portend lower survivalrates and smaller sockeye salmon returns to Bris-tol Bay in future runs.

Chum SalmonOver the 33 years from 1970 to 2003, chum

salmon have accounted for 10% of Alaska’s salmonharvest. Over the past eight years (1996–2003), theaverage annual chum salmon harvest across Alaskawas 18.6 million fish, with the 2000 harvest wellabove this average at a record 24.3 million fish.Currently 60–70% of the commercially harvestedchum salmon in Alaska occur in the southeasternregion, where hatcheries produce a significantportion of the catch.

Chum salmon runs in southwestern and west-ern Alaska, similar to sockeye salmon in BristolBay, were well below long-term averages. Manag-ing chum salmon fisheries in western Alaska iscomplicated by another commercial fishery atFalse Pass in the Aleutian Islands. Western Alaskachum salmon may spend part of their ocean life inthe Gulf of Alaska. These salmon, as maturingadults on their return migration, funnel throughAleutian passes into the Bering Sea. The FalsePass fishery, targeted primarily on sockeye salmonreturning to Bristol Bay, must be managed so as tonot overharvest chum salmon destined for theKuskokwim and Yukon Rivers in western Alaska.

Alaska sockeye salmoncatch, 1970–2003.

71

The Alaska Board of Fisheries has placed majorrestrictions on the False Pass fishery in an effortto help rebuild depleted chum salmon resourcesin western Alaska. Chum salmon in western Alaskanot only are an important part of commercial fish-eries in that region but are also a significant sub-sistence resource for local residents.

Coho SalmonCommercial catches of coho salmon across

Alaska in 2003 totaled 4.1 million fish, a half millionless than the recent six-year average but still wellabove the record low catches of the 1970s. Thisrelatively high commercial harvest was due to gen-erally favorable returns in the southeastern region,where 3.0 million or more coho salmon from hatch-ery and wild stock production were caught in fourout of the last six years. Coho salmon, along withsockeye and chinook salmon, are a popular targetspecies in recreational fisheries throughout Alaska.

Chinook SalmonThe annual commercial harvest of chinook

salmon in Alaska has ranged between 300,000 and700,000 fish over the past two decades. The state-wide 13-year (1991–2003) average annual harvestwas 559,000 fish. In general, chinook salmon arethe first species each year to begin spawningmigrations into Alaskan rivers. Only in a few Bris-tol Bay and western Alaskan rivers are fisheriespermitted to directly target these early returningruns of chinook salmon. However, in fisheriestargeting other salmon, chinook salmon are oftentaken incidentally.

The region-wide chinook salmon harvest insoutheastern Alaska, where significant numbers ofnon-Alaska-origin fish are caught, is normally reg-ulated by a quota under provisions of the PacificSalmon Treaty. This annual harvest quota is thenre-allocated among various fisheries by the AlaskaBoard of Fisheries.

Alaska chum salmoncatch, 1970–2003.

Alaska coho salmoncatch, 1970–2003.

72

ReferencesAlaska Department of Fish and Game (2001) Alaska

Subsistence Fisheries 1999 Annual Report.Division of Subsistence, Juneau, Alaska.

Alaska Department of Fish and Game (2003)Catch, Effort, and Value Statewide Harvest;Recent Years Harvest Statistics. Reported atweb site: http://www.cf.adfg.state.ak.us/.

Byerly, M., B. Brooks, B. Simonson, H. Savikko,and H. Gieger (1999) Alaska CommercialSalmon Catches, 1878–1997. Regional Infor-mation Report No. 5J99-05, Alaska Departmentof Fish and Game, Division of Commercial Fish-eries, Juneau, Alaska.

Burger, C.V., and A.C. Wertheimer (1995) Pacificsalmon in Alaska. In Our Living Resources:A Report to the Nation on the Distribution,Abundance, and Health of U.S. Plants, Ani-mals, and Ecosystems (E.T. La Roe, G.S. Farris,C.E. Puckett, P.D. Doran, and J.J. Mac, ed.). U.S.Department of the Interior, National BiologicalSurvey, Washington, D.C., p. 343–247.

Hare, S.R., N.J. Mantua, and R.C. Francis (1999)Inverse production regimes: Alaska and WestCoast Pacific salmon. Fisheries, Vol. 24, p. 6–15.

Hare, S.R., and N. Mantua (2000) Empirical evi-dence for North Pacific [climatic] regime shifts

in 1977 and 1989. Progress in Oceanography,Vol. 47, p. 103–145.

Howe, A.L., R.J. Walker, C. Olnes, K. Sundet, andA.E. Bingham (2001) Participation, Catch, andHarvest in Alaska Sport Fisheries during1998. Fishery Data Series No. 99-41 (revisededition), Alaska Department of Fish and Game,Anchorage.

Mantua, N.J., S.R. Hare, Y. Zhang, J.M. Wallace,and R.C. Francis (1997) A Pacific interdecadalclimate oscillation with impacts on salmonproduction. Bulletin of the American Meteoro-logical Society, Vol. 78, No. 6, p. 1069–1079.

Minobe, S., and N. Mantua (1999) Interdecadalmodulation of interannual atmospheric andoceanic variability over the North Pacific.Progress in Oceanography, Vol. 43, p. 163–192.

North Pacific Anadromous Fish Commission(1998) Statistical Yearbook, 1994. NPAFC,Vancouver, B.C.

Wertheimer, A.C. (1997) The status of Alaskasalmon. In Proceedings of the Symposium onPacific Salmon and their Ecosystems; Statusand Future Options (D.J. Strouder, P.A. Bisson,and R.J. Naiman, ed.), Seattle, Washington, Jan-uary 10–12, 1994. Chapman Hall, New York, p.179–197.

Alaska chinook salmoncatch, 1970–2003.

73

Russian–American Long-term Census of the ArcticInitial Expedition to the Bering and Chukchi Seas

This article was preparedby Kathleen Crane,

of NOAA’s ArcticResearch Office.

July 23, 2004, marked a historic day in Arcticresearch and exploration, as well as in Russian–U.S. relations. On this date the Russian researchship the Professor Khromov left Vladivostok,Russia, packed with U.S.- and Russian-fundedscientists to begin a 45-day collaborative journeyof exploration and research in the Arctic.

Stemming from a 2003 Memorandum of Under-standing for World Ocean and Polar RegionsStudies between NOAA and the Russian Academyof Sciences, this cruise was the first activity underthe Russian–American Long-term Census of theArctic (RUSALCA). RUSALCA means “mermaid”in Russian. In November 2003 a RUSALCA plan-ning workshop was held in Moscow to outline the

biological, geological, chemical, and physicaloceanographic sampling strategies to be pursuedin the Bering Strait and Chukchi Sea.

This initial cruise was a collaborative U.S.–Russian Federation oceanographic expedition tothe Arctic seas regions shared by both countries:the Bering and Chukchi Seas. These seas and thelife within are thought to be particularly sensitiveto global climate change because they are centerswhere steep thermohaline and nutrient gradientsin the ocean coincide with steep thermal gradientsin the atmosphere. The Bering Strait acts as theonly Pacific gateway into and out of the ArcticOcean and as such is critical for the flux of heatbetween the Arctic and the rest of the world.

Stations undertakenduring the voyage of the

Professor Khromov, aRussian research vessel

engaged in the RUSALCAexpedition. The colored

area in the Arctic Oceanindicates a region ofenhanced ice melting

between 1970 and 2001.

74

Monitoring the flux of fresh and salt water andestablishing benchmark information about thedistribution and migration patterns of the life inthese seas are also critical tasks that must becompleted prior to the placement of a climate-monitoring network in this region.

The cruise was divided into two legs, whichincluded sampling and instrument deployment inU.S. and Russian territorial waters. The cruiseobjectives were to support the U.S. interagencyStudy of Environmental Arctic Change (SEARCH)program and the NOAA Ocean Exploration Program.

Many Russian Federation agencies participatedin the planning and execution of the RUSALCA

Vice-AdmiralLautenbacher (NOAA)

and Vice-PresidentLaverov (RAS) signing

the Memorandum ofUnderstanding betweenNOAA and the Russian

Academy of Sciences.

2004 mission. These included the Ministry ofDefense, Roshydromet, the Ministry of NaturalResources, the Ministry of Science, and theRussian Academy of Sciences, the initial partnerof NOAA. Group “Alliance,” a private companyregistered in Moscow, Russia, facilitated the inter-national agency support.

Leg 1: Piips VolcanoThe first leg was in the Bering Sea, with the

ship leaving Vladivostok, Russia, on July 23 andarriving in Nome, Alaska, on August 6. The U.S.chief scientist was Kevin R. Wood, of NOAA’sPacific Marine Environmental Laboratory. Thechief of the expedition was Captain VladimirSmolin, of the Ministry of Defense, RussianFederation.

Bordering the Bering Sea at its southern termi-nus with the Pacific Ocean is the Aleutian VolcanicArc. Waters entering and exiting the Bering Seafrom and to the Pacific Ocean transit through thisarc and are most likely chemically and dynamicallymodified by their interaction with the intensehydrothermal activity resulting from the mid-watervolcanoes. Quantifying the flux from this relativelyshallow volcanic arc and its influence on thewaters and atmosphere above are important fac-tors when considering the relationship betweenearth processes, the ocean, and greenhouse gasexchanges.

Russian Federation scientists have exploredthis region in the past. However, this was the firstopportunity for scientists from the U.S. and Rus-sia to work together to map the volcanic featuresand search for fluxes of methane, mercury, andother hydrothermal fluids and gases.

The first leg of the RUSALCA expeditionfocused on the hydrothermal activity and relatedgeological and biological processes associatedwith the Piips volcano, which lies at a depth of 300m in the southern part of the Komandorskayadepression of the western Aleutian Arc.

Russian marine geologists discovered the Piipshydrothermal field in 1987. Temperatures of up to130°C were measured, and hydrothermal depositscomposed of sulfates, carbonates, amorphoussilica, and other materials were discovered. Inaddition, large fields of bacterial mats and numer-

Dirty ice near Herald Island, as seen from the Khro-mov. Arctic scientists are continuously finding new andinteresting, sometimes even alarming, informationabout changes in the sea ice over time.

75

ous hydrothermal fauna were detected. However,very little is known about the influence of hydro-thermal activity on the regional biochemical andphysical environment of the neighboring ocean.

During the RUSALCA expedition, the combinedRussian side-looking sonar–CTD–methane sen-sors were lowered over the volcano. However,upon retrieval after the second launch, the sonarwas lost at sea. Atmospheric monitoring of mercuryalong this leg from Vladivostok revealed high con-

centrations near the Asian coastline and againover the volcano. Because of the relatively shal-low depth of the volcano, the data suggest thatmercury is released to the water column and theatmosphere above as a result of vigorous hydro-thermal activity at this site.

Leg 2: Bering Strait throughthe Herald Canyon

The second leg was in the Chukchi Sea, leavingNome, Alaska, on August 8 and returning to Nomeon August 24. The U.S. chief scientist was Dr.Terry Whitledge, of the University of Alaska.The chief of the expedition was Captain VladimirSmolin, of the Ministry of Defense, RussianFederation.

Because of the reduction of ice cover in theArctic and the possibility of permanent loss ofthe seasonal ice cover in the Chukchi Sea studyregion as shown by climate models, it is thoughtthat this area might be subject to significant eco-system change. A program of ecosystem-orientedexploration was planned for Leg 2 of the RUSALCAexpedition to provide a foundation for detectingfuture ecosystem indicators of climate change.

Twelve scientific programs examined benthicprocesses, a census of Arctic zooplankton, bio-diversity of adult and juvenile fish, nutrient andprimary productivity, marine chemistry, physicaloceanography, microbial reactions and fluxes,side-looking sonar and video imagery of the sea-floor, paleoceanography, and atmospheric contam-inants. The primary study area lay between Wran-gel Island and Herald Canyon in Russia Federationterritorial waters to Cape Lisburne in Alaska toPoint Barrow and south to the Bering Strait. Aseries of hydrographic transects were taken toallow sampling of all water masses during thissummer period. A high priority of the hydrographicsurvey was to collect samples across the Bering

Starfish from the Russianwaters of the Bering

Strait.

Bathymetry of the Piipsvolcano.

Gas plume venting from the Piips volcano .

76

Strait in support of the Russian and Americanmoorings in the western Bering Strait, to collect aseries of high-speed CTD transects across HeraldCanyon, and to enhance the knowledge of faunaldistributions for the Arctic census of marine life.The long-term goal in this region is to obtain con-tinuous and comprehensive monitoring within theBering Strait for several years, which will requireroutine access to the eastern and western portionsof the study area for scientific operations.

High-density CTD stations also examined therole, rates, and rhythms of Pacific water transportthrough the Herald Canyon and analyzed the dis-persion of this water into the greater high Arctic

beyond. Until recently the transport pathways ofwater into and out of this region have been onlypoorly mapped. The degree to which these watersmix with newly invasive Atlantic waters over theChukchi Plateau and the Mendeleev and CanadaBasins is also not well known. Understandingthese physical pathways and the consequentnutrient pathways is critical for mapping the distri-bution of biota and its migration routes throughthis region of the Arctic.

Leg 2 resulted in the following:• 77 CTD and nutrient casts;• Two moorings (both Russian and U.S.) placed

in the western waters of the Bering Strait;• 87 phytoplankton samples;• A comprehensive survey and census of zoo-

plankton species at 33 stations in the BeringStrait through the central Chukchi Sea;

• Benthic grabs at 11 stations;• Benthic epifauna sampled at 17 stations;• Larval and juvenile fish collected at 17 sites;• 31 species of fish sampled;• 27 trawls for adult fish, collecting 24 species;

and• Eleven remotely operated video lowerings

from the Bering Strait into the Herald Canyon.During Leg 2, the hydrographic, biochemical,

and productivity sampling was integrated from allstations sampled, and the data from U.S. and Rus-sian collaborators will be combined for the jointassessment of climate change, water mass proper-ties, and a census of marine life in the Arctic.

Future ProgramsThe RUSALCA program anticipates returning

annually to the Bering Strait to service the twoRussian–American moorings left at their sites inthe western part of the strait in 2004. Negotiationsare developing to use a Russian Navy hydro-graphic research vessel to carry out the BeringStrait operations in 2005–2007.

NOAA anticipates that it will expand the breadthand range of the RUSALCA program during theInternational Polar Year if funding permits. Effortsare underway to design ship-based traverses fromthe Chukchi and East Siberian Sea shelves northinto the deeper Makharov Basin to investigateecosystem indicators of climate change, examinethe physical and chemical properties of the oceanin the region where the greatest amount of thaw-ing of Arctic sea ice has taken place, and carry outa census of marine life in this poorly explored andmapped region.

RUSALCA Leg 2station locations.

RUSALCA mooringlocations in the western

Bering Strait.

77

On the Creation of Environmental Data Setsfor the Arctic Region

This article was preparedby Florence Fetterer,

NOAA Liaison, NationalSnow and Ice Data

Center/World Data Centerfor Glaciology, Boulder,

Colorado, and IgorSmolyar, of NOAA’s

Ocean ClimateLaboratory, NationalOceanographic DataCenter, Silver Spring,

Maryland.

When Karl Weyprecht proposed better coordi-nation of research in 1874, leading to a series ofcoordinated synoptic observations in the Arctic,little did he think that his ideas would producescientific data that remains of intense interest toresearchers 130 years later. And still less would hehave imagined that his proposals, and the result-ing International Polar Year of 1882–1883, wouldinform goals for creating and managing scientificdata in the 21st century. Because of advances inobservation and data technologies, the questionsthat Weyprecht addressed have only increased insignificance. What constitutes useful environmen-tal data? How are data both a product of researchand a catalyst for new research? How should databe managed to ensure continued accessibility andusefulness? The NOAA Arctic research activitiesdescribed elsewhere in this edition of ArcticResearch of the United States both use and pro-duce data. This article examines the process of

ronmental data sets can be found in almost everyNOAA line office, but it is the NOAA NationalEnvironmental Satellite, Data, and InformationService Data Centers that share a mission of datamanagement. The NOAA National Data Centers’commitment to long-term data management pro-vides institutional support for producing exemplaryenvironmental data sets. Each center has a partic-ular research focus and expertise that adds valueto its data management results. After a brief profileof these centers, we will discuss what generalcharacteristics make certain data sets especiallyvaluable and what elements come into play duringthe production of these data sets, highlightingenough of them here to provide a sense of thebreadth of NOAA’s Arctic data production activi-ties. An atlas, the Climatic Atlas of the ArcticSeas 2004, serves as a case study. We also cite anumber of NOAA operational, research, and mod-eling products as examples of particular aspectsof data product creation.

National Data CentersNational OceanographicData Center

Located in Maryland, the National Oceano-graphic Data Center (NODC) is a repository anddissemination facility for global ocean data.Researchers from NODC’s Ocean Climate Labora-tory (OCL) announced in 2000 that the worldocean has warmed significantly over the past 40years. Just as the atmosphere has a climate, withvariability on different time scales, the ocean’stemperature, salinity, and other characteristicschange over time. OCL researchers based theirconclusions on data laboriously collected, qualitycontrolled, and assembled into a special form ofenvironmental atlas called a climatology. To facili-tate comparisons of the past with the present,and to investigate interannual-to-decadal ocean

IPY meteorologicalstation, 1882.

creating environmental Arctic data sets and thesymbiosis of research, data, and data management.These data sets may have value beyond that ofadvancing Arctic research objectives: they maybe, for example, monitoring tools for change detec-tion, or they may underlie decision support appli-cations.

This focus on data and data management hasbecome a proper discipline at NOAA in the 130years since Weyprecht’s call for better coordina-tion of research and data resources. Arctic envi-

78

climate variability, many thousands of raw obser-vations acquired from ships were interpolated toa regular spatial grid and combined over annual,seasonal, and monthly compositing periods. Adefinitive statement about oceanic warming wouldnot be possible without these climatologies.

In addition to supporting scientific studies,OCL’s International Ocean Atlas and InformationSeries (currently nine in number) exemplify inter-national cooperation. Much of it is taking placethrough the OCL’s World Data Center (WDC) forOceanography, in Silver Spring, Maryland. Inter-national collaboration is an absolute necessity foracquiring a sufficient number of observations forclimatologies. The Global Oceanographic DataArchaeology and Rescue (GODAR) Project, forexample, has added over six million historicalocean temperature profiles to the archives, as wellas a large amount of other data. Initiated by theNODC and WDC, this OCL-directed project wassubsequently endorsed by the UNESCO Inter-governmental Oceanographic Commission.

National Climatic Data CenterAmong the hundreds of climate data compila-

tions housed at the National Climatic Data Center

(NCDC) in Asheville, North Carolina, are Arcticstation data from the Global Historical ClimateNetwork, the most comprehensive homogeneouscollection of station temperature data available.“Homogeneous” means consistent over the yearsand from place to place, through changes ininstrumentation, acquisition method, and sitecharacteristics, so that scientists may look fortrends in the data. Homogeneous data sets requirecareful quality control. Historical data are madehomogeneous with present-day observations byadjusting for non-climatic discontinuities, such asa jump in precipitation that might be caused by achange in instrumentation. An important part of thequality control process is compiling station inven-tories that detail the history of each station, includ-ing changes in instrumentation, changes in loca-tion, and changes in surroundings. If a town growsup around a formerly rural station, for example, aheat island effect may be present in the data record.

NCDC also operates the World Data Center forPaleoclimatology (WDC Paleo), located in Boulder,Colorado. Paleoclimatology puts the relativelyrecent changes in Arctic climate, apparent in theinstrumented record, in long-term context. Proxydata from tree rings, ice cores, and lake and marinesediments available from the WDC Paleo wereused by an international team of scientists for acircum-Arctic view of surface air temperaturechanges over the last 400 years.

The WDC Paleo web site provides interpreta-tions of the record: A steep increase in warmingbetween 1850 and 1920 was most likely due tonatural processes. Warming since 1920 is moredifficult to ascribe to natural forcing alone. For aneven longer view, ice core data are valuable. WDCPaleo and the National Snow and Ice Data Centerjointly maintain the Ice Core Gateway. Proxyclimate indicators from ice cores such as oxygenisotopes, methane concentrations, dust content,and other parameters stretch the record back morethan 1000 years.

National Geophysical Data CenterThe National Geophysical Data Center (NGDC),

Boulder, Colorado, contributes significantly toArctic science through participation in the devel-opment of the International Bathymetric Chart ofthe Arctic Ocean (IBCAO). IBCAO bathymetryprovides a detailed and accurate representationof the depth and morphology of the Arctic Oceanseabed. This dynamic database contains allavailable bathymetric data north of 64°N. It is

Arctic temperatureanomalies

Ice core samples.Ice cores are taken from

ice sheets or ice capsand are used by

paleoclimatologists as arecord of past climate.

79

maintained as a gridded database, and a versionhas been published in map form. The IABCO teamremapped the Lomonosov Ridge, showing it to bemore segmented in structure, wider, and shallowerthan had previously been mapped. The Lomon-osov Ridge is an important topographic barrierthat influences deep water exchange between theeastern and western basins of the Arctic Ocean.An accurate seafloor is important for applicationsincluding ocean modeling, mapmaking, and otherresearch endeavors. The IABCO effort involvesinvestigators from eleven institutions in eightcountries. It has been endorsed and supported bythe Intergovernmental Oceanographic Commission,the International Arctic Science Committee, theInternational Hydrographic Organization, and theU.S. Office of Naval Research.

National Snow and IceData Center

Operational products, such as sea ice charts forshipping interests from the NOAA/Navy/CoastGuard National Ice Center, are often laboriouslyproduced by manually interpreting and synthesiz-ing data from many sources, both satellite and insitu. They are generally more accurate than similarproducts from single sources. The National Snowand Ice Data Center (NSIDC) works with opera-tional groups within NOAA to make these prod-ucts available to a different user base by archivingoperational data, making data available online,providing documentation, and fielding questionsfrom researchers about the data.

Originally founded to manage scientific datafrom the International Geophysical Year of 1957–1958 (the follow-on inspired by the IPY of 1882–1883), the World Data Center (WDC) for Glaciologyis operated by NSIDC in Boulder, Colorado. Today,NSIDC is a NOAA-affiliated data center, designatedby NOAA in 1976, affiliated with NGDC, and part ofthe University of Colorado’s Cooperative Institutefor Research in Environmental Sciences (CIRES).

The NOAA program at NSIDC supports theWDC and emphasizes data rescue and data fromoperational sources that can be used for climateresearch and change detection. NOAA-fundedactivities complement the activities of NSIDC’sDistributed Active Archive Center (DAAC) and itsNSF-funded Arctic System Science Data Coordi-nation Center. The latter centers handle large vol-umes of satellite data (about 80% of NSIDC’sfunding comes from NASA for operation of theDAAC) and data from individual scientists.

NSIDC has 522 data products in its on-linearchive. Excluding satellite data sets, 104 of theseare Arctic data sets, and of these about 20% arethe more evolved compilations termed Arctic envi-ronmental data products.

What Makes a GoodEnvironmental DataProduct?

To make a good environmental data product,one starts with raw data and then processes orpresents them in such a way that they becomeinformation. Data are transformed into a productthat can advance a user up a hierarchy from datato information, to knowledge, to wisdom, shorten-ing the user’s path from data to knowledge or tothe why and how of environmental interactionsand change. Following is a summary of some ofthe data management practices that effect thistransformation.

ContextSimply presenting data systematically is some-

times enough to transmit any underlying meaning.That is, even a well-organized collection of rawdata can be an environmental data product. Usually,though, data products are more sophisticated.Presenting data in context is important, and thedata product creator must decide what “context”means for the particular data under consideration.Temporal context usually means having as long arecord as possible, while spatial context may meancovering as much of the Arctic as possible at anappropriate resolution or station density. Contextmay mean including population data for both pred-ator and prey in an Arctic species survey or includ-ing as complete a set of oceanographic hydro-chemical parameters as possible. The point is tomake significant patterns evident, while committingno “sins of omission” in choosing what to include.Methods are important, as is documenting uncer-tainty. For example, if the product is a gridded clima-tology of snow depth on Arctic Ocean sea ice, someanalysis should be done to ensure that enoughobservations are included in each grid cell for anacceptable level of accuracy. Sometimes the wayin which data are gridded, interpolated, and pre-sented implies a certain level of precision. For exam-ple, a two-dimensional quadratic function fit to snowdepth on sea ice tells the user immediately that thesnow depth information is not very precise.

80

Data products may include the results of ananalysis, such as an empirical orthogonal functionanalysis of surface pressure data that shows theArctic Oscillation pattern (that is, the tendency ofpressure near the pole to act counter to pressureat mid-latitudes) or the addition of a trend line orother model fitting to observations. In such cases,the product creators have an extra responsibilityto explain the limitations of their method of analy-sis, since these methods, if appropriately used,draw the pattern in the data rather than leave it toinference.

DocumentationWords are often the only way to provide appro-

priate context. The heat island effect, known onlyif the weather station history is known, is but oneexample of the importance of clear and completedocumentation. Good documentation is writtenwith the user in mind. For example, if the usersare scientists, they will need to know about anyknown biases in the data record. If the data prod-uct is created for the general public, this informa-tion is just as important because it influences whata user infers from the data, but the informationmust be given in non-technical language. Opera-tional users often need today’s data irrespectiveof historical biases and may require little or nodocumentation.

For example, the NOAA National Weather Ser-vice’s National Operational Hydrologic RemoteSensing Center (NOHRSC), located in Minneapo-

Variation in snow depth on Arctic sea ice, depicted by fitting a two-dimensional qua-dratic function to available data by month. There are very few measurements of snowcharacteristics on Arctic sea ice. This representation, from the Environmental WorkingGroup’s Meteorology Atlas, is the best possible in the absence of dense station coverage.While it appears unrealistic, product documentation explains why a more sophisticatedgridding method for the available data is not appropriate.

Examples of the Arctic Oscillation in its positive phase (left) and negative phase (right) from the NOAA NationalWeather Service’s Climate Prediction Center web site. Blue indicates negative pressure anomalies, and orangeindicates positive. The Arctic Oscillation is a large-scale atmospheric circulation pattern. Variability in the AO hasbeen implicated in changes such as the recent steep decline in ice extent.

81

lis, Minnesota, provides snow information in avariety of products and formats to meet opera-tional forecasting needs. The NOHRSC web site(http://www.nohrsc.noaa.gov/) is designed toserve these users efficiently with interactive prod-ucts and brief documentation. NSIDC archivesassimilation model output (http://nsidc.org/data/g02158.html) from NOHRSC and gives the researchcommunity access to this unique data set. Exten-sive documentation on the NSIDC site, not neededon the NOHRSC site, covers alternative products,data quality and value, and potential researchuses of the data. Similarly, the NOAA NESDIS Sat-ellite Services Division Operational Daily NorthernHemisphere Snow Cover Analysis is made avail-able to the operational community at http://www.ssd.noaa.gov/PS/SNOW/ and is archived for theresearch community at NSIDC (http://nsidc.org/data/g02156.html). Likewise, the NODC web site(http://www.nodc.noaa.gov/) is designed to pro-vide users access to various products with ahigher level of documentation than would beneeded for operational users.

Graphics and Site DesignGraphical presentation and site design are

aspects of information architecture that are espe-cially important for complex environmental datasets that are viewed or used through a web page.Though not an environmental data set per se, theNOAA Arctic Theme Page (http://www.arctic.noaa.gov/) offers an example of good site design.

Careful attention must be paid to the graphicalpresentation of data. Gridded data to which a colortable has been applied, or data smoothed by inter-polation, are often subject to misinterpretation.The display resolution of pixels should faithfully

represent the underlying resolution of the data,lest a scientist infer regional relevance that is notsupported. Color tables should not have suddenjumps in intensity or hue that can draw the eye toa sea surface temperature difference, for example,that is an arbitrary point in a continuum, thus sug-gesting a pattern that is not there.

Data IntegrityThree main components ensure environmental

data product integrity:• The product must have scientific integrity;

peer review of the data and a citation for thedata set are needed to accomplish this.

• The data repository must be trustworthy.• The data must not have been altered since the

data were acquired or produced (or any alter-ation must be well described). The data man-agement concepts of fixity, provenance, andsource authentication come into play here.

Often it is the reputation of an individual scien-tist that imbues his or her data product creationwith an aura of integrity. Data centers work withscientists to ensure that the reality measures up.Though it is common in the U.S. for investigatorsto manage their own data, this is rarely successfulover the long term because scientists rarely havethe requisite data management background neededto keep their data useful and accessible to the nextgeneration of scientists or the resources to dealwith technical issues that keep data secure, suchas media migration and off-site backups.

What Makes a Well-UsedEnvironmental Data Set?

Certain attributes will ensure that a data setwill have many users. In such cases it is especiallyimportant to follow the design precepts above.Data products that include unique data, that arecomprehensive collections, that offer continuouscoverage over a long time period, that are easyto use, and that provide a synthesis of availableinformation are characteristics of the most popularArctic data products.

UniquenessUpward-looking sonar data from submarines

provide the only measurements of ice thicknessover a large portion of the Arctic. Ice thicknessestimates are critical to estimating the mass bal-

Sea ice extent trends.When data are presented

with a trend line, as inthis example, data provid-

ers should include errorbars and document the

limitations of the method(linear regression in this

case) when it comes toproviding information

from the raw data.

82

ance of ice in the Arctic Ocean—ice extent andconcentration are only two dimensions of a three-dimensional problem. One difficulty in workingwith original records is that almost all submarinedata are classified. Investigators at the U.S.Army’s Cold Regions Research and EngineeringLaboratory (CRREL) in Hanover, New Hampshire,and the Polar Science Center, Applied PhysicsLaboratory, University of Washington, workedwith the U.S. Navy’s Arctic Submarine Laboratoryto find a way to “fuzz” the submarine track posi-tions so that the data could be cleared for releaseby the Chief of Naval Operations. NSIDC distrib-utes the data set, Submarine Upward LookingSonar Ice Draft Profile Data and Statistics (http://nsidc.org/data/g01360.html). It has been the basisof a number of research articles on the controver-sial topic, “Is Arctic ice thinning?”

ComprehensivenessComprehensive data products offer more value

than data sets that must be combined with othersin order to have enough data of a single type fora scientific study. It takes a well-funded project, amulti-year commitment, and many individual andinstitutional partners to assemble, for example,“all” surface marine reports from ships, buoys,

and other platform types. As such, the Interna-tional Comprehensive Ocean Atmosphere Data Set(ICOADS, http://www.cdc.noaa.gov/coads/) ofquality-controlled data dating from the late 18thcentury is a remarkable achievement. An entirebody of literature has grown up around topicsrelated to the quality control of historical ship datain ICOADS. For example, sea surface temperaturesacquired by throwing a bucket over the side andmeasuring the temperature of the retrieved waterare not the same as temperatures acquired fromthe engine intake. A “bucket correction” must beapplied. This correction is based on modeled heatloss for water in a bucket on deck and should takeinto account ship speed (and its uncertainty) andthe material of the bucket (wood or canvas). Clearly,quality controlling the millions of observations ofvarious types so that they are homogeneous overtime is a Herculean task.

ICOADS began as a U.S. project (COADS) in1981 as a partnership between the NOAA Office ofOceanic and Atmospheric Research’s Environmen-tal Research Laboratories and NCDC, CIRES, andNSF’s National Center for Atmospheric Research.The NOAA portion of ICOADS is currently sup-ported by the NOAA Climate and Global ChangeProgram. NSIDC makes an Arctic subset available(http://nsidc.org/data/nsidc-0057.html).

Continuous Spatial andTemporal Coverage

Products that are as close as possible tocontinuous in space and time are often desirablebecause, for example, the danger of aliasing isminimized (that is, there is a smaller chance of

Temperature anomaliesfor October through

December 2002 in theArctic. This illustration

assisted in attributing thecauses of the 2002 and

2003 sea ice extent mini-ma to, in part, anoma-

lously warm air tempera-ture. The NOAA-CIRES

Climate Diagnostic Center(http://www.cdc.noaa.gov)

display tool allows usersto choose the month andyear to display from the

NCEP/NCAR re-analysisproducts. The color

table is intuitive (warmcolors are warmer thanaverage temperatures),

and the resolution ofone degree shown in

the color bar isappropriate to

the data set.

A submarine surfacing through sea ice. This photo comesfrom SCICEX (Scientific Ice Expeditions), a collaborationbetween the U.S. Navy and civilian scientists for environ-mental research in the Arctic.

83

missing a significant event or pattern in the data).The enormously popular “reanalysis” productsare an example.

Reanalysis projects take data such as those inICOADS and assimilate them through a numericalweather prediction model to produce a series ofanalyses in which parameters such as surfacepressure and temperature fields are physicallyconsistent with one another. The fields cover alarge area (the Northern Hemisphere, for instance)without gaps and are available at regular timeintervals over a long record, making reanalysisoutput more useful than observations for manyapplications. One example is studying the spatialand temporal variability of large-scale atmosphericcirculation patterns, such as the Arctic Oscillation.NOAA’s Arctic Research Office is planning a cou-pled atmosphere–sea ice–ocean–terrestrial reanal-ysis optimized for the Arctic region. The descrip-tion of the climate system it will produce can be

used to detect Arctic change and assist in attribut-ing change to specific causes.

Ease of Access and UseMany valuable data sets lie unused in archives

simply because they are not easily accessible. For-merly, NSIDC’s Glacier Photograph Collection ofthousands of historical glacier photographs datingfrom the 1880s saw only a handful of users eachyear because users had to travel to NSIDC to viewthe fragile collection of prints. Now, thanks toNOAA’s Climate Database Modernization Programfunding for scanning the photos, many of the photo-graphs can be viewed on line, and high-qualitydigital images can be downloaded (http://nsidc.org/data/g00472.html). As a result, the number of usershas climbed to about a thousand each month.

Improving access can broaden the user basefor a data set. NSIDC’s satellite passive microwave

Cushing Glacier, Alaska.This photograph, taken in1967, is one of thousands

that are part of theGlacier Photograph

Collection, created by theNational Snow and Ice

Data Center andavailable on line at

http://nsidc.org/data/g00472.html.

84

sea ice concentration data are popular among sci-entists but not geared toward the general public.The data are voluminous, and some technical andscientific sophistication is needed to simply readand interpret the data. To mitigate these issues,NSIDC developed the Sea Ice Index, which pro-vides an easy way to visualize the satellite data.The Sea Ice Index web site lets any user trackchanges in sea ice extent and compare conditionsbetween years. About 3,000 users visit the Sea IceIndex site every month.

Similarly, the OCL web site (http://www.nodc.noaa.gov/OC5//SELECT/dbsearch/dbsearch.html)allows users to extract data from the World OceanDatabase 2001. While contributors to an environ-mental data product often prefer that the data becompiled on CD-ROM or DVD for reasons of fixityand attribution, data sets are much more likely to beused if they can be easily browsed or manipulatedon line with a selection tool to facilitate access.

SynthesisData products that offer synthesis—a “big pic-

ture” version of the information in the data—arerare because they are difficult to construct. Syn-thesis products are built by distilling information

from multiple sources. An exciting and successfulexample of this kind of product is NOAA’s NearRealtime Arctic Change Indicator web site (http://www.arctic.noaa.gov/detect/), which summarizes“the present state of the Arctic climate and eco-system in an accessible, understandable, andcredible historical context.” Designed for decisionmakers and the general public, it presents asophisticated 30-year principal components analy-sis (the synthesis) of 19 climate, land, marine,and ecosystem “indicator” time series, such as thelength of the travel season over tundra, the BeringSea pollock population, the number of extremelycold days each year in cities such as Minneapolis,and the extent of Arctic sea ice.

Taken alone, any one of these time serieswould not present a compelling account of Arcticchange. Taken together, the big picture emerges.The site tracks the rate and extent of changes inthe Arctic to facilitate informed decisions concern-ing the impacts that result. Web pages for eachof the indicators give a succinct but completeanalysis of the data record in non-technical terms.Changes are given in context, including the con-text of the human dimension. Links to reports andmore detailed data make it a useful resource forscientists as well.

Sample result from theSea Ice Index, which

displays anomalies in iceextent and other ice

parameters going back to1979. Here, recent sum-mer ice extent, which has

been the lowest in the datarecord, is displayed with

the median extent (pinkline) to give climatologicalcontext to the information.

85

of collaborating with Russian institutions to addmore historical data to the OCL World OceanDatabase Project. The most recent result is theClimatic Atlas of the Arctic Seas on DVD, withmeteorology, oceanography, and hydrobiology(plankton, benthos, fish, sea birds, and marinemammals) data from the Barents, Kara, Laptev, andWhite Seas, collected by scientists from 14 coun-tries during the period 1810–2001. The MurmanskMarine Biological Institute of the Academy of Sci-ences of the Russian Federation and OCL/NODCprepared the atlas with support from NOAA NESDISand the Climate and Global Change Program.

The atlas provides historical context for itsobservations by including a written history ofoceanographic observations in the Arctic, as wellas scanned copies of selected rare books and arti-cles. A gallery with photos and drawings gives theuser some idea of what historical data collectionplatforms and expeditions looked like.

As is often the case with projects involvingdata rescue, libraries provided much of the materialand documentation; the NOAA Central Library

Selection of time series representing Arctic change. The combined indicators are theresult of a mathematical analysis (principal component analysis) that resolves thetrends in all the time series into two major components. Series noted by an asterisk havebeen inverted. Red indicates large changes in recent years.

Marine biologists P. Savitsky and I. Molchanovsky sam-pling plankton in the Kara Sea as part of an expedition ofthe Murmansk Marine Biological Institute on the nuclearicebreaker Sovetsky Soyz in April–May 2002. The Cli-matic Atlas of the Arctic Seas weds early observationswith contemporary observations in a seamless package.

The Arctic Change Indicator web site wasdeveloped by NOAA’s Arctic Research Officeunder the stewardship of investigators at theNOAA Pacific Marine Environmental Laboratory.It draws on the work of hundreds of investigatorsaround the globe. A major challenge will be tokeep the site updated. As NOAA looks towardbuilding new observing systems, it will be criticalto maintain data flow from existing observingstations.

The Climatic Atlasof the Arctic Seas

With these principles in mind, we now turn toa case study of active, collaborative data manage-ment that produces new knowledge and dissemi-nates this potential across a wide community ofresearchers.

The WDC for Oceanography in Silver Spring ,Maryland, and OCL/NODC have a long history

86

(Silver Spring, Maryland), the Slavic and BalticBranches of the New York Public Library, the NewYork Museum of Natural History Library, the Dart-mouth College Library (Hanover, New Hampshire),the Slavic Library (Helsinki, Finland), and the pub-lic libraries of Moscow, Murmansk, and St. Peters-burg (Russian Federation) all contributed.

Assembling an atlas on this scale presents anumber of challenges. A surprisingly difficult oneis the elimination of duplicate stations from differ-ent data sources. As databases or parts of data-bases are shared, metadata are altered. For exam-ple, one database may have a station location indegrees, minutes, and seconds, and another mayconvert to decimal degrees. Rounding errors maygive the appearance that these are two stationsseparated by as much as a few miles. Values ofparameters may be presented at observed levelsin one data set but interpolated, often by anunknown method, to a standard level in anotherdata set. Another source of uncertainty is convert-ing units of measurement. As a result, the samestation data from different sources may differ incoordinates, time of measurement, and values ofthe parameters themselves. To help choose whatstation records to include, the atlas authors useda system of priorities: cruise reports, ship logs,and expedition diaries (all original sources) weredeemed more reliable that data sources where thedata apparently were repeatedly transformed.Elimination of duplicates and “near duplicates”brought the number of stations down from aninitial 1,506,481 to a still sizeable 433,179.

Users have two ways of accessing raw obser-vations: by oceanographic cruise or through 1°squares. For every month, a distribution map ofstations is generated that allows a user to accessdata from a chosen square. Data may be easilyimported into Excel or other database applications.Access to the actual observations is importantfor many users. Other users are likely to prefer aclimatological presentation, since climatologiesprovide a convenient representation of averageconditions, such as monthly or decadal means.The atlas satisfies both by including mean monthlytemperature and salinity distribution fields at fivestandard depths, using an objective data analysismethod.

A Long Journey from thePast: The InternationalPolar Year

The International Polar Year serves as animportant milepost for assessing our efforts andestablishing stronger standards to carry the valueof observations and research far into the future.As we look back over past IPY/IGY efforts andforward to those coming in 2007–2008, those ofus who create Arctic environmental data sets haveobserved some lessons over the years.

Many of us know the tragic story of the GreelyExpedition, an American venture sent into theArctic in 1881 to establish an IPY station thatended in starvation for most of the party, but

One of the first Russianresearch vessels, Andrey

Pervozvanny, on an expe-dition at the beginning of

the 20th century, in theBarents Sea.

87

fewer of us are aware of the sad tale of their scien-tific data. Their observational records should havebecome a legacy to their efforts and sacrifice, butthis is not the case. As Kevin Wood of NOAAwrites, narrating the story of their data:

“Perhaps the most compelling aspect of the Greelytragedy is the utter commitment of these men topreserve their scientific work. Aware that if relief

didn’t arrive in time they would be left to retreat ontheir own, Greely began making copies of their sci-entific work (amounting to some 500 observationsper day). When they were forced to abandon FortConger in August 1883 they took with them—inlieu of extra rations—these copies sealed in three tinboxes of 50 pounds each, all of the daily journals,70 pounds of glass photographic plates, and all ofthe standard thermometers and several other impor-

Average Septembersurface temperature (top)

and salinity (bottom),from the Climatic Atlas of

the Arctic Seas. Note thelow salinity at the river

mouths. Scientists areworking to understand the

role of freshwater inputfrom rivers on Arctic

Ocean (and global ocean)circulation. Climatologies

provide a picture ofaverage conditions

against which toevaluate changes.

88

tant instruments. They also continued a program ofscientific observation, in the face of starvation, untiljust 40 hours before they were rescued.”

Surely, Greely and his men hoped that theirwork would lead to important advances in science.Greely expressed this sentiment when he wrote inhis official report, “The conviction that at no dis-tant day the general laws of atmospheric changeswill be established, and later, the general characterof the seasons be predicted through abnormaldepartures in remote regions, causes this work tobe made public…in the hope that it may contributesomewhat to that great end.” The desire to be ableto “predict the character of the seasons” still moti-vates researchers today.

Unfortunately, the research program of the firstIPY was never completed as Weyprecht had origi-nally planned. Each nation issued an individualreport over the ensuing years, but no systematicstudy of the simultaneous observations—theheart of the IPY program—was undertaken. The

International Polar Com-mission dissolved, andthe data collected atsuch cost during thefirst IPY soon fell intoobscurity. Today the originalrecords of the first IPYare widely scattered invarious libraries andarchives and are oftenin a perilous state ofpreservation. Some ofthe published reportsare extremely rare andare very difficult toobtain. The fate of thefirst IPY records, gainedat such high cost, under-utilized both then andnow, and scattered over

the course of time, highlights how important it isto provide for the effective preservation and man-agement of such extremely valuable data.

The scientific legacy of the Greely Expeditionand the other expeditions of the first IPY has onlywith difficulty been preserved. NOAA has recentlymade meteorological data from the first IPY avail-able in digital format, along with an extensive col-lection of documentary images (see http://www.arctic.noaa.gov/aro/ipy-1).

There is another kind of legacy that we can

create from the experience of the first IPY. As welook forward to a new International Polar Year, wemust remain focused on these key lessons aboutdata management.

Lesson 1: Applying the Right Kind and RightAmount of Effort at the Right Time is Imperative

While data rescue is difficult, tedious, andoften expensive, it is crucial. The only way toreduce uncertainty in our estimates of past, andpredictions of future, Arctic environmental changeis to incorporate more, older, and better data intoour analyses. NOAA’s Climate Database Modern-ization Program, now in its sixth year, has keyedor scanned and placed on line over 45 million envi-ronmental records. More needs to be done, espe-cially in documenting and quality controllingthese records, because these last steps requirecapturing the knowledge of people who know the“rescued” data best, often a cadre that are beyondretirement age.

Lesson 2: Structures that Enable InternationalCollaboration can Dramatically Increase Value

As a result of Arctic geography, the most com-prehensive data sets result from internationalcooperation. GODAR and ICOADS are models forthis cooperation. International data-sharing agree-ments are essential. In contrast to the NationalData Centers, the World Data Center system pro-vides a structure within which data sharing canoccur with a minimum of diplomatic overhead.WDCs in the U.S. that share Arctic data interna-tionally are the WDC for Glaciology, Boulder (co-located with NSIDC), WDC for Oceanography,Silver Spring (co-located with NODC), the WDCfor Marine Geology and Geophysics, Boulder(co-located with NGDC), the WDC for Meteorol-ogy, Asheville (co-located with NCDC), and theWDC for Paleoclimatology (affiliated with NCDC).

Lesson 3: Good Data Stewardship is Superior toUntimely Data Rescue

We can avoid expensive and possibly fruitlessdata rescue efforts in the future by heeding thelessons of the past. The International Polar Year,2007–2008, will be a catalyst for reinvigoratingprofessional data management. The IPY promisesnew international collaboration and the potentialfor synthesis of knowledge under the headingsof cross-disciplinary research themes. Good datastewardship will help ensure that this major under-taking will not shortly become a dimly recedingspot on the horizon behind us.

Sgt. Jewell recording tem-perature, Fort Conger,

during the Greely expedi-tion, 1881–1883.

89

This effort requires not only attention to thedata, but also to capturing the “data about thedata” that enables continuing understanding andvalue. A disciplined effort to define and organizethe metadata will enable other researchers tolocate, understand, and interpret data for yearsto come, providing the foundation for long-termcoordination and synthesis. In the instance of thefirst IPY, Weyprecht’s vision of coordinated syn-optic observations led to the acquisition of a dataset that serves as a snapshot of climatic condi-tions as they existed in that now long-past year.Data collected then can now be compared withconditions as they are today. Making that compar-ison, Wood and Overland (2005) found thatmonthly mean air temperatures at IPY-1 stationswere generally within recent climatological limits,and spatial patterns in temperature anomalies(departure from the long-term mean) were consis-tent with Arctic-Oscillation-driven patterns of vari-ability. In a nod to the value of documentationand metadata, Wood and Overland noted that “thequalitative logs are particularly useful in validatingclimate information.”

NOAA will focus its strength in environmentalobservations and analysis on the polar regionsduring IPY. NOAA’s Arctic Research Office hasendorsed a fundamental goal for IPY data manage-ment: to securely archive a baseline of dataagainst which to assess future change, and toensure that IPY data are accessible and preservedfor current and future users.

What will this IPY snapshot look like, and howwill data be preserved? In contrast to Weyprecht’sIPY, most data from the coming IPY will be “borndigital.” Station logbooks from Weyprecht’s daycould be preserved in libraries, where they had tobe physically protected from destruction by fire,insects, and chemical decomposition of paper andink. One might think it is easier to preserve digitalinformation, but digital data are not immune fromphysical destruction, and they require a host ofmeasures to ensure their usability into the future:“digital objects require constant and perpetualmaintenance, and they depend on elaborate sys-tems of hardware, software, data and informationmodels, and standards that are upgraded orreplaced every few years.”*

During the coming IPY, hundreds of investiga-tors and agency programs will produce raw obser-

vations, satellite data, and environmental dataproducts in a number and of a complexity thatwould have been hard to imagine in the late 1880s.To ensure preservation,

• NOAA’s Data Centers and the Arctic ResearchOffice will work to advance standards andtechnologies that support this goal. NOAAadvocates the use of the Open Archival Infor-mation System (OAIS) Reference Model formetadata. Work on the OAIS model and ontechnological advances such as GRID com-puting and interoperable catalogs is happen-ing now at NOAA’s National Data Centers.

• Cross-agency support for IPY data manage-ment is needed. Because of the international,distributed nature of IPY activities, the datathey produce will necessarily be archivedand made accessible through distributed datamanagement. This distributes the burden ofdata management but imposes additionalcoordination challenges. Within the U.S.,the National Academy of Sciences’ PolarResearch Board has endorsed the conceptput forward by the International Council ofScientific Unions’ IPY Planning Group of acoordinating IPY Data and Information Ser-vice (IPY-DIS). Cross-agency support of theDIS at a national level will ensure that the U.S.leaves a secure IPY data legacy.

• Adequate funding is needed. Funding forthe management of data acquired throughresearch programs is often difficult to obtain,either because the importance of data man-agement as a discipline is not recognized orbecause there are simply not enough dollarsto go around. Currently, for every $30 dollarsspent nationally on Arctic research, about $1is spent on Arctic data management.

In the end, it is important to remember thattechnological advances and digital archives willsecure data for future generations of researchersonly to the extent that they are successful in cap-turing what people know about the data. We mustalso keep today’s equivalent of the IPY-I station’s“qualitative logs.” With them, future researcherswill have the appropriate contextual material toturn the coming IPY data into information-filledArctic environmental data products.

ReferencesArzberger, P., P. Schroeder, A. Beaulieu, G. Bowker,

K. Casey, L. Laaksonen, D. Moorman, P. Uhlir,and P. Wouters (2004) An international frame-

Jane Beitler and RuthDuerr, NSIDC, assisted in

editing this article. KevinWood, PMEL, provided

the material on thefirst IPY and the

Greely expedition.

* From It’s About Time: Research Challenges in DigitalArchiving and Long-Term Preservation, final report of aworkshop sponsored by NSF and the Library of Congress,August 2003.

90

work to promote access to data. Science, Vol.303, No. 5665, p. 1777–1778.

Levitus, S., J.I. Antonov, T.P. Boyer, and C.Stephens (2000) Warming of the world ocean.Science, Vol. 287, No. 5461, p. 2225–2229.

National Science Foundation and Library of Con-gress (2003) It’s About Time: Research Chal-lenges in Digital Archiving and Long-termPreservation, Final Workshop Report. Spon-sored by the National Science Foundation’sDigital Government Program and Digital Librar-ies Program, Directorate for Computing andInformation Sciences and Engineering, and TheLibrary of Congress’s National Digital Informa-tion Infrastructure and Preservation Program.

Overpeck, J.T., K. Hughen, D. Hardy, R. Bradley, R.Case, M. Douglas, B. Finney, K. Gajewski, G.Jacoby, A. Jennings, S. Lamoureux, A. Lasca, G.MacDonald, J. Moore, M. Retelle, S. Smith, A.Wolfe, and G. Zielinski (1997) Arctic environ-mental change of the last four centuries.Science, Vol. 278, No. 5341, p. 1251–1256.

Peterson, T.C., and R.S. Vose (1997) An overviewof the Global Historical Climatology Networktemperature database. Bulletin of the AmericanMeteorological Society, Vol. 78, No. 12, p.2837–2849.

Wood, K.R., and J.E. Overland (2005) Climatelessons from the First International Polar Year1881–1884. Poster presented at the AmericanMeteorological Society 8th Conference on PolarMeteorology and Oceanography, San Diego,California.

Data Set CitationsBerger, V. J., A.D. Naumov, N.V. Usov, M.A. Zuba-

ha, I. Smolyar, R. Tatusko, and S. Levitus (2003)36-Year Time Series (1963–1998) of Zoo-plankton, Temperature, and Salinity in the WhiteSea. NESDIS Atlas 57, NOAA, Washington, D.C.

Fetterer, F., and K. Knowles (2004) Sea Ice Index.Updated from 2002, digital media, NationalSnow and Ice Data Center, Boulder, Colorado.

Fetterer, F., and V. Radionov (Ed.) (2000) Environ-mental Working Group Arctic Meteorologyand Climate Atlas. CD-ROM, Arctic ClimatologyProject, National Snow and Ice Data Center,Boulder, Colorado.

Lappo, S., Y. Egorov, M. Virsis, Y. Nalbandov, E.Makovetskaya, L. Virsis, I. Smolyar, and S.Levitus (2003) History of the Arctic Explora-tion 2003: Cruise Reports, Data. CD-ROM,International Ocean Atlas and Information

Series, Vol. 8, World Data Center for Oceanog-raphy, Silver Spring, Maryland.

Markhaseva, E., A. Golikov, T. Agapova, A. Beig,and I. Smolyar (2002) Zooplankton of the ArcticSeas 2002. CD-ROM, International OceanAtlas and Information Series, Vol. 6, World DataCenter for Oceanography, Silver Spring, Mary-land.

Matishov, G., A. Zuyev, V. Golubev, N. Adrov, V.Slobodin, S. Levitus, and I. Smolyar (1998)Climatic Atlas of the Barents Sea 1998:Temperature, Salinity, Oxygen. NESDIS Atlas26, NOAA, Washington D.C.

Matishov, G. P., Makarevitch, C. Timofeyev, L.Kuznetsov, N. Druzhkov, V. Larionov, V. Golubev,A. Zuyev, V. Denisov, G. Iliyn, A. Kuznetsov, S.Denisenko, V. Savinov, A. Shavykin, I. Smolyar,S. Levitus, T. O’Brien, and O. Baranova (2000)Biological Atlas of the Arctic Seas 2000:Plankton of the Barents and Kara Seas.NESDIS Atlas 39, NOAA, Washington D.C.

Matishov, G., A. Zuyev, V. Golubev, N. Adrov, S.Timofeev, O. Karamusko, L. Pavlova, O. Fady-akin, A. Buzan, A. Braunstein, D. Moiseev, I.Smolyar, R. Locarnini, R. Tatusko, T. Boyer, andS. Levitus (2004) Climatic Atlas of the ArcticSeas 2004: Part I. Database of the Barents,Kara, Laptev, and White Seas - Oceanographyand Marine Biology. NESDIS Atlas 58, NOAA,Washington, D.C.

National Operational Hydrologic Remote SensingCenter (2004) SNODAS Data Products atNSIDC. Digital media, National Snow and IceData Center, Boulder, Colorado.

National Snow and Ice Data Center (1998) Subma-rine Upward Looking Sonar Ice Draft ProfileData and Statistics. Digital media, NationalSnow and Ice Data Center/World Data Centerfor Glaciology, Boulder, Colorado.

National Snow and Ice Data Center/World DataCenter for Glaciology, Boulder (Compiler) (2002)Online Glacier Photograph Database. Digi-tized subset of the Glacier Photograph Collec-tion, National Snow and Ice Data Center, Boul-der, Colorado.

NOAA/NESDIS/OSDPD/SSD (2004) IMS DailyNorthern Hemisphere Snow and Ice Analysisat 4 km and 24 km Resolution. Digital media,National Snow and Ice Data Center, Boulder,Colorado.

Serreze, M. (Compiler) (1997) ComprehensiveOcean–Atmosphere Data Set, LMRF ArcticSubset. Digital media, National Snow and IceData Center, Boulder, Colorado.

91

92

Page 3, 9 (top), 26, 27 (top and bottom), 73, 75 (top left), and 76 (top and bottom): K. Crane, NOAA. Page4: Bon van Hardenberg, Institute of Ocean Sciences, Canada. Page 5 (top): I. Macdonald. Page 5 (bottomleft): R. Gradinger and B. Bluhm. Page 5 (bottom right): Emory Kristof, National Geographic Society. Page6 (middle) and 9 (bottom): R. Hopcroft. Page 6 (bottom): Data from Raskoff, K.A., J.E. Purcell, and R.R.Hopcroft (2005) Gelatinous zooplankton of the Arctic Ocean: In situ observations under the ice. PolarBiology, Vol. 28, p. 207–217. Page 7 (top) and 8: University of New Hampshire and NOAA. Page 7 (bot-tom): Larry Mayer, University of New Hampshire, and K. Crane, NOAA. Page 10: R. Hopcroft, R.Gradinger, and B. Bluhm. Page 11: Norbert Untersteiner. Page 12: U.S. Army Cold Regions Research andEngineering Laboratory and Japan Agency for Marine-Earch Science and Technology. Page 13 and 14:Ignatius Rigor, Applied Physics Laboratory, University of Washington. Page 15: Craig Lee, AppliedPhysics Laboratory, University of Washington. Page 19 (left and right), 21 (top and bottom), 23 (top andbottom), and 24: Robert Stone, Cooperative Institute for Research in Environmental Sciences andNOAA’s Climate Monitoring and Diagnostics Laboratory. Page 20 (top): J. Stroeve, National Snow andIce Data Center. Page 20 (bottom): After Belchansky, G.I., D.C. Douglas, I.A. Alpatsky, and N.G. Platonov(2004) Spatial and temporal multiyear sea ice distributions in the Arctic: A neural network analysis ofSSM/I data, 1988–2001. Journal of Geophysical Research, Vol. 109, p. C10017. Page 22 (top): P.D.N.Hebert (2002) Canada’s Polar Life (www.polarlife.ca). Page 22 (bottom): Bulletin of the American Meteo-rological Society (2005) State of the Climate in 2004. Page 42: M. Ott, Alaska Ocean Observing System,and K. Sherman, Large Marine Ecosystem Program, Narragansett Laboratory, NOAA Fisheries. Page 44:NOAA Satellite and Information Service. Page 45: S. Barbeaux, Alaska Fisheries Science Center, NOAAFisheries. Page 46: R. Edwing, Center for Operational Oceanographic Products and Services, NOAAOcean Services. Page 47: A. Macklin, Pacific Marine Environmental Laboratory, NOAA Research. Page48: P. Stabeno, Pacific Marine Environmental Laboratory, NOAA Research. Page 49: M. Belbeoch, ArgoInformation Centre. Page 50: Lori Mazzuca. Page 51 and 53 (top): Sue Moore. Page 52 (top and bottom):J. Clarke. Page 63 (bottom): B. Christman. Page 54: Andrew Trites. Page 55, 56, and 61: After Trites et al.(2005) Bottom-up forcing and the decline of Steller sea lions in Alaska: Assessing the ocean climate hy-pothesis. Fisheries Oceanography, submitted. Page 58: After Peterson, W.T., and F.B. Schwing (2003) Anew climate regime in northeast Pacific ecosystems. Geophysical Research Letters, Vol. 30, Article No.1896. Page 60: After Marzban, C., N. Mantua, and S. Hare (2005) Retrospective study of climate impacton Alaska Steller sea lion: A report. Technical Report No. 485, Department of Statistics, University ofWashington. Page 66: Office of Oceanic and Atmospheric Research, National Undersea Research Pro-gram, NOAA. Page 74 (top), 80 (bottom), and 82 (bottom): NOAA. Page 74 (bottom): K. Wood, NOAA.Page 75 (top right): A. Astakhov, Pacific Institution of Oceanography, Vladivostok, Russia. Page 74 (bot-tom): V. Gladish, SONIC. Page 77 and 88: K.R. Wood and J.E. Overland. Page 78 (top): From the web site“A Paleo Perspective on Global Warming,” http://www.ncdc.noaa.gov/paleo/globalwarming/overpeck.html. Page 78 (bottom): Mark Twickler. Page 80 (top) and 84: National Snow and Ice Data Center. Page 81:From the NSIDC Sea Ice Index (http://nsidc.org/data/ seaice_index). Page 82 (top): SCICEX (ScientificIce Expedition). Page 83: Austin Post, from the Glacier Photograph Collection, National Snow and IceData Center. Page 85 (top): From the NOAA Arctic Change Indicator web site (HTTP://www.arctic.noaa.gov/detect/indicators.shtml). Page 85 (bottom) and 86: Archive of the Murmansk Marine Biological Insti-tute. Page 87: Climatic Atlas of the Arctic Seas.

IllustrationCredits

93

The following individuals are the principal staff representatives for the Interagency Arctic ResearchPolicy Committee. Additional staff support is provided by the Federal agencies for specific activitiesthrough working groups, as necessary.

Interagency Arctic Research Policy Committee Staff

James DevineU.S. Geological SurveyDepartment of InteriorReston, Virginia 22092jdevine@usgs.gov

Waleed AbdalatiNational Aeronautics and Space AdministrationWashington, DC 20546wabdalat@hq.nasa.gov

Charles E. MyersNational Science FoundationArlington, Virginia 22230cmyers@nsf.gov

Igor KrupnikSmithsonian InstitutionWashington, DC 20560krupnik.igor@nmnh.si.edu

Ann V. GordonDepartment of StateWashington, DC 20520gordonav@state.gov

Richard VoelkerMaritime AdministrationDepartment of TransportationWashington, DC 20590richard.voelker@marad.dot.gov

Luis TupasDepartment of AgricultureWashington, DC 20024ltupas@csrees.usda.gov

John CalderNational Oceanic and Atmospheric AdministrationDepartment of CommerceSilver Spring, Maryland 20910john.calder@noaa.gov

John StubstadDepartment of DefenseRosslyn, VA 22209John.Stubstad@osd.mil

Wanda FerrellDepartment of EnergyWashington, DC 20545wanda.ferrell@science.doe.gov

Douglas SteeleU.S. Environmental Protection AgencyWashington, DC 20460Steele.Doug@epamail.epa.gov

Natalie TomitchNational Institutes of HealthDepartment of Health and Human ServicesWashington, DC 20892tomitchn@ficod.fic.nih.gov

Thomas WojahnU.S. Coast GuardDepartment of Homeland SecurityWashington, DC 20593twojahn@comdt.uscg.mil

[Blank Page]

NSF 05-39

NATIONAL SCIENCE FOUNDATION ARLINGTON, VA 22230

OFFICIAL BUSINESS

RETURN THIS COVER SHEET TO ROOM P35 IF YOU DO NOT WISH TO RECEIVE THIS MATERIAL , OR IF CHANGE OF ADDRESS IS NEEDED , INDICATE CHANGE INCLUDING ZIP CODE ON THE LABEL (DO NOT REMOVE LABEL).

PRESORTED STANDARD

U.S.POSTAGE PAID National Science Foundation