National Academies Press: OpenBook

Toward an Integrated Arctic Observing Network (2006)

Chapter: 5 Designing the Network

« Previous: 4 Data Management
Suggested Citation:"5 Designing the Network." National Research Council. 2006. Toward an Integrated Arctic Observing Network. Washington, DC: The National Academies Press. doi: 10.17226/11607.

Designing the Network

This chapter describes the infrastructure and approach needed to create an Arctic Observing Network (AON), including ideas on types, number, and distribution of network components; where observations might be effectively made; and the role that remote sensing and novel technologies might play. It also describes processes that could be used to design the optimal mix of observations and test for data redundancies. The Committee adopts a philosophical approach in this chapter and reserves specific implementation ideas for the next chapter. There are many ongoing and planned observation programs and their design plans have details that relate to this chapter. The value added by this chapter is to raise considerations that are specific to building a network that can generate multipurpose, pan-arctic datasets.


In recent decades it has become clear that much environmental change in the Arctic is systemic. The change of state at any one place is likely to be strongly dependent on the state at locations far removed in space and time. For example, it is thought that seasonal sea-ice retreat that has been observed in the Chukchi Sea in recent years is at least in part dependent upon long-term, Northern Hemisphere atmospheric forcing (i.e., Arctic Oscillation shifts) that has brought younger and thinner ice that is more readily melted into the North American Arctic (Rigor et al., 2002; Rigor and Wallace, 2004). Connections to the rest of the globe also depend on pan-arctic systemic behavior. Two factors that are arguably among the most important in strongly coupling the Arctic with the rest of the Earth are ice-albedo feedback and control of global thermohaline circulation through export of freshwater to the subarctic seas. The Arctic is a key control point for these processes through changes in snow cover; glacier, lake, and sea-ice cover; and freshwater provided by arctic runoff and melted sea ice.

The AON will provide the critical data necessary to expand understanding of the ways the arctic system is connected and functions. Many existing observational mechanisms are research projects that have limited spatial and temporal scope. Continuity in time and space is rarely the result of a larger plan. Most existing science planning efforts cover specific realms, processes, time scales, or regions. The observing system needs for these can often be met with the traditional organizational and operational approaches. However, the overlay of the AON could supply the wide-area, long-term observations needed to track the state of the Arctic and understand how the system functions as part of the global environmental system.

Such an AON will not be unprecedented. Over the past several decades, there has been progress in recognizing the importance of time series from sustained observational efforts. For example, the beginnings of the Keeling record of continuous atmospheric carbon dioxide measurements at Mauna Loa was connected with the 1957 International Geophysical Year activities, but it was only maintained through the persistence of Keeling and colleagues (Keeling, 1998).1 This record is now critical for evaluating human use of fossil fuels and subsequent carbon cycling and sequestration, and the Mauna Loa collections have become an institutionalized resource that is the longest continuous record in the global carbon dioxide monitoring network. Requests for transfers of data from the Carbon Dioxide Information Analysis Center in Oak Ridge, Tennessee, which is the designated depository for this network, including the Mauna Loa record, typically exceed 10,000 per day.2 These requests come from almost every country on Earth. Yet, despite the demonstrated

Suggested Citation:"5 Designing the Network." National Research Council. 2006. Toward an Integrated Arctic Observing Network. Washington, DC: The National Academies Press. doi: 10.17226/11607.

value, sustaining these and other long-term measurements remains problematic and undervalued.

Network and observational systems require different funding strategies than shorter-term research projects. For the network to serve its purpose, a long-term commitment to network components is needed. The U.S. Long Term Ecological Research (LTER) and the National Oceanic and Atmospheric Administration (NOAA) trace gas monitoring programs are examples of a longer-term commitment to addressing science questions that are not posed as hypotheses that can be resolved in a field season or two. There have been many examples of long-time-series datasets from the LTER program that have helped guide research questions and set hypotheses, as well as evaluate regulatory and policy effectiveness. For example, in research conducted as part of the Bonanza Creek LTER project in Alaska, Grünzweig et al. (2004) showed that as agriculture expands at high latitudes, soil carbon losses are likely to be greater than those in other biomes. In another review, LTER data collected in the mid-1980s from Bonanza Creek were used to generalize about the long-term impacts upon nitrogen cycling of severe forest fires (Smithwick et al., 2005).

To be important and worthy of long-term support, networks should not need to be driven by hypotheses that can be tested over two-to-five-year funding cycles. It is clear that longer-term observations can directly help experimental research by providing the tools needed to address specific hypotheses. However, incentives and rewards for building long-term datasets are not always apparent, given the prevalence of two-to-five-year funding cycles. Protection of intellectual property rights to use data collected by individual observers, while still making them accessible to a larger community of users, will also remain a complex challenge for networks. And one of the key challenges for sustaining the AON beyond an initial effort like the International Polar year (IPY) is how to make the transition from pilot-type efforts like Keeling’s in the late 1950s to an institutionalized observing system that fulfills broad societal and scientific needs.

In addition to being founded upon a philosophy that values systematic, long-term, extensive measurements, the AON will also benefit if participants adhere to several other philosophical approaches. These include recognizing the value of measuring variables that are not currently changing in addition to variables that are changing dramatically; valuing open source, nonproprietary software and tools that allow data sharing across platforms and disciplines; valuing the ability to evolve with and embrace new technologies, opportunities, and demands; valuing the human dimension of the arctic system (see Box 5.1) and participation of local observers who make the AON more cost-effective and help make year-round observations; and valuing data management because of the efficiencies and overall cost-effectiveness that it can bring to the AON.

Box 5.1
The Human Dimension of the Arctic Observing Network: Perspectives from Human Dimension of the Arctic System

The following text is adapted from a brochure produced by HARC (Human Dimensions of the Arctic System) called “Designing the Human Dimensions into an Arctic Observing Network” (HARC, 2005).

Arctic environmental change is the set of biophysical transformations of land, ice, oceans and atmosphere, driven by an interwoven system of human activities and natural processes. Research on the human dimensions of arctic change addresses the coupled human-natural system and investigates how individuals and societal groups contribute to, are influenced by, and mitigate and respond to the changes that take place on a local, regional, and global level. Human dimensions science therefore encompasses may topics, approaches, methods, and disciplines.

Understanding how social systems interact with natural systems (both physical and biological) involves qualitative analyses and quantitative studies that rely on forms of hypothesis testing and analysis familiar to fields such as atmospheric science, terrestrial ecology, glaciology, or ocean biogeochemistry. When biophysical scientists study human-influenced phenomena such as ice roads, river flows, or fish catches, understanding human influences becomes critical. These are nontrivial challenges for biophysical-human dimensions research.

The human dimensions component of the AON could consist of a network of social scientists, citizens, and other observers who help make available and accessible arctic human dimensions data that are being collected in a common data structure with circumpolar scope. This part of the AON could also identify data gaps and fill them. Data might include the size, well-being, and livelihoods of arctic communities; demographic vital statistics, health and economic statistics; qualitative data such as historical accounts or life histories; and global economic and institutional trends (see also Table 2.1).

A key role for the human dimensions component of the observing network beyond collecting and organizing data could be to perform analyses needed to disseminate useful, useable, relevant, and timely data to researchers, policy makers, and the public through a single Web portal with multiple links.

Suggested Citation:"5 Designing the Network." National Research Council. 2006. Toward an Integrated Arctic Observing Network. Washington, DC: The National Academies Press. doi: 10.17226/11607.


A systems approach to understanding the Arctic defines needs that transcend national boundaries and the timeframes of individual science investigations. Thus, a key contribution of the AON will be to provide a framework within which existing programs can be linked and supplemented to achieve an interdisciplinary pan-arctic observation capability. As the overarching network, the AON will need to provide continuity across national boundaries into the foreseeable future. Challenges of achieving spatial coverage include identifying and filling observational voids and ensuring that observations made across AON’s component networks are equivalent and comparable. In the temporal domain, the network needs to be flexible to adapt to improved understanding and new technology, while at the same time providing continuity of certain key measurements across decadal periods or more. Also in the temporal domain, the observations will be needed at frequencies ranging from almost continuous (and near-real-time) to years. Finally, the data content of AON will span physical and biogeochemical variables and also include human dimensions measurements and observations to help understand how humans affect and are affected by the arctic system (Table 2.1). In discussing the idea of developing a human dimensions observing network in the Arctic, HARC (2005) notes that such a network would be “essential to understanding common patterns and local variations in the flow of Arctic social change, testing hypotheses and developing models about their causes, and developing credible, evidence-based future scenarios and policies useful for supporting decision making under conditions of escalating environmental and social change.”


Building Blocks

A variety of tools and platforms is available for arctic observations (Table 5.1). The compilation in Table 5.1 classifies observation platforms in broad, general categories of varying costs, duration, and range of coverage and measurement and provides some brief description of advantages and disadvantages of each platform. This list is not exhaustive and categorizations are subject to change as technology improves. Nevertheless, these are examples of platforms that will likely provide the infrastructural foundation for data acquisition in the AON.

Measurement tools and platforms are either automated or rely on direct human involvement. On the automated side, satellites provide remote sensing of a variety of parameters like temperature, albedo, vegetation cover, clouds, winds, ice extent and concentration, and glacier altimetry. Fixed unattended observation systems can be stationed on land, attached to ice, and moored to the seafloor, and depending upon location, flexibly equipped with in-situ environmental sensors for such parameters as salinity, temperature, current speed, wind speed and direction, photosynthetic active radiation, ice thickness, chlorophyll, nutrients, atmospheric gases, and contaminants. Automated devices can also drift with the ocean currents or winds, and self-propelled unattended systems are becoming available in the form of unmanned air vehicles (UAVs), autonomous underwater vehicles (AUVs), and underwater gliders.

Many measurements are sufficiently complex that there is no alternative to having experts on site, whether in an arctic community, on a ship or aircraft, at a field camp, or at a more permanent field station. The local residents of the Arctic constitute an invaluable resource as expert and intimate observers as well as potential collaborators for professional researchers who cannot be present on a year-round basis in remote high-latitude locations. A special investment will be needed to cultivate relationships with arctic communities and organizations and to collaborate on researching ways of linking local and traditional knowledge (LTK) and scientific data. Some general guidelines for incorporating LTK in the AON and for working with arctic communities were developed by Joan Eamer in her input to the Committee, and these, in addition to guidance from other sources (Box 5.2) provide a starting point for the AON with respect to collaborating with arctic residents.

The AON must leverage all of these capabilities—automated and manual—to maximize the value of the resulting data for the available resources.

Measurement Approaches

Observational approaches that can be used for the AON are categorized in the SEARCH implementation plan documents (SEARCH SSC and IWG, 2003; SEARCH, 2005) as intensive measurement programs at a few carefully chosen sites, broadly distributed or “extensive” measurements of more easily measured parameters, repeated surveys or sections giving detailed snapshots of spatial variability, and remotely sensed (e.g., satellite-based) measurements for achieving continuity across spatial domains and in situ data in-filling.

Intensive Measurement Programs

Flagship observatories that make intensive measurements in the terrestrial realm (Shaver et al., 2004) are situated in representative areas with access to a wide variety of arctic terrain. They have well-developed suites of comprehensive measurements and of time series for key variables throughout the year. Examples of sites that could fill this role include Abisko, Alert, Barrow, Bonanza Creek, Cherski, Kluane Lake, Ny-Ålesund, Pallas, Summit, Tiksi, Toolik Lake, and Zackenberg. Some of these intensive sampling sites are oriented toward specific disciplines. For example, Summit

Suggested Citation:"5 Designing the Network." National Research Council. 2006. Toward an Integrated Arctic Observing Network. Washington, DC: The National Academies Press. doi: 10.17226/11607.

TABLE 5.1 Comparison of Examples of Existing Arctic Observation Platforms











Surface, atmosphere

Time series on spatial grid

Comprehensive space-time coverage

Limited to surface and atmospheric information, continuity issues




Surface, atmosphere



Cost and weather limited

Manned Drifting Stations (ice camps)



Surface and subsurface

Lagrangiana time series, process studies

Provides access for many disciplines

Cost, ice, and weather limited




Surface and subsurface

Sections, process studies

Provides access for many disciplines

Cost, time and ice limited







Infrequent, sensor, and personnel limited

Gliders and AUVs






Frequency and duration under development

Surface Drifters




Lagrangian time series

Inexpensive, extensive temporal sampling

Limited to surface

Ice-Tethered Buoys



Surface and subsurface

Lagrangian time series

Extensive temporal and vertical sampling

Too few; Lagrangian time series

Bottom-Tethered Moorings




Eulerian time series

Extensive temporal and vertical sampling

Too few; surface and near-surface not sampled

Intensive Observatories



Surface, atmosphere

Eulerian time series

Temporal sampling

Limited geographical coverage

Extensive Measurement Sites


Short to Long-term

Surface, atmosphere

Eulerian and Lagrangian time series

Inexpensive, broad geographic coverage

Limited range of measurement parameters

Local Observers



Surface, atmosphere, socioeconomic, health

Eulerian and Lagrangian time series, sections

Year-round presence, inexpensive

Geographically focused (e.g., coasts), intercomparison challenges

aThe Lagrangian approach measures properties at a point whose geographic position changes with time. The Eulerian approach measures properties at a geographically fixed point.

specializes in atmospheric sampling including trace gases and atmospheric contaminants whereas hydrological and ecosystem-level variables are the focus at Toolik Lake. However, a broader goal of these sites is to provide co-located, multidisciplinary measurements that reveal the interdependence of physical, chemical, biological, and geological domains. For example, several sites (e.g., Alert, Barrow, and Tiksi) are co-located with rawinsonde stations, thereby contributing upper-atmosphere measurements to the comprehensive suite of surface and subsurface measurements.

In the Arctic Ocean and its marginal seas, marine observatories are already positioned in key areas to characterize variability and long-term changes of oceanic circulation and water properties year round. Examples of these sites include the North Pole Environmental Observatory, the Beaufort Gyre Observing System, and the Nansen and Amundsen Basin Observing System. Other internationally supported marine observational sites include the moored ASOF (Arctic-Subarctic Ocean Fluxes) transects in the straits and gateways of the Arctic Ocean. As is the case for the terrestrial observatories, these marine intensive sites also incorporate repeat surveys and suites of more widely distributed instruments.

Distributed Observations

Distributed observations are more limited in scope than those at intensive sites but nevertheless provide needed

Suggested Citation:"5 Designing the Network." National Research Council. 2006. Toward an Integrated Arctic Observing Network. Washington, DC: The National Academies Press. doi: 10.17226/11607.

Box 5.2
General Guidance on Incorporating Local and Traditional Knowledge into Observing Networksa

  • The AON must work in partnership with local and indigenous organizations from the outset to determine the type of information that will be linked to the AON, how it will be collected and used, and who will make decisions regarding control of information. There must be open and ongoing communication with communities as well as local and regional capacity building so that local people can engage with the project (Eamer, 2005; Loring, 2005). This process will take significant investment of time and resources. Existing guidelines and procedures for working with local communities and LTK should be examined and adapted. These include ethical and methodological guidelines from relevant regional or national organizations or funding agencies (e.g., Center for Arctic Cultural Research, 1989; Grenier, 1998; NRI and ITC, 1998; ARCUS, 2004; NSF, 2005).

  • When considering how LTK should be incorporated into the AON, there should be a distinction between

    i. incorporating LTK to improve science-based monitoring,

    ii. using LTK to help design and interpret monitoring, and

    iii. developing community-based monitoring systems that are linked to the AON. It is also important to distinguish between monitoring, repeated observations, and traditional knowledge studies that document local knowledge, beliefs, and observations of change over a past period. These are complementary, as are science-based research and monitoring (Eamer, 2005).

  • Issues of ownership, program management, processes for informed consent, processes for obtaining any necessary permits (e.g., research licenses), demonstration of respect for local experts, intellectual property rights, and information sharing protocols must be addressed early on (NRI and ITC, 1998; Eamer, 2005; Loring, 2005).

  • Methodological issues and best practices to forming research relationships with communities must be considered. These include appropriate methods for LTK collection, especially before implementing a standard approach, and methods in relation to the desired contribution and results (Eamer, 2005). LTK research should follow the “3 Rs”: respect, reciprocity, and relationship (Grenier, 1998).

  • The LTK component of the AON should not act to support or confirm scientific monitoring—it should stand on its own and make a unique contribution (Eamer, 2005).

  • The AON should take care to respect LTK as it is and keep the context and richness of LTK data (Eamer, 2005; Loring, 2005). Quantitative analyses are possible (e.g., creating indices, charting trends), but do not reduce LTK to a yes/no survey response (Eamer, 2005).

  • Environmental responses should not be separated out from social and economic responses because of the important interactions between these and because it will not make sense to local people to separate them (Eamer, 2005).


The guidelines are from input to the Committee (in particular, from recommendations submitted by Joan Eamer, UNEP-GRID, and Eric Loring, ITK) and other literature and online sources.

spatial and temporal context. They are essential for monitoring the evolution of the broader state of the Arctic. As a small-scale example, measurements in the Imnavait Creek microscale system (a 2.2 km2 headwater catchment) and the Kuparuk River basin supplement more intensively studied variables at Toolik Lake. At the pan-arctic scale, the Global Terrestrial Network (GTN)-Permafrost program3 measures active layer thickness and permafrost temperatures around the Arctic. In addition to fixed networks of distributed measurements, mobile platforms can also determine local synopticity or spatial context for intensive sites or larger-scale measurements.

In the marine setting, distributed observations will provide critical details on the temporal development of spatial gradients important to circulation and water mass transformations. Distributed marine observations will need to include mobile platforms such as drifting buoys, ice-tethered profilers, and AUVs, and also simple moored measurements and low-intensity manned observations such as for sea-level or sea-ice conditions. A combination of mobile platforms and fixed infrastructure (i.e., moored and cabled observatories), satellites, and underwater navigation and communication systems are key marine components of the AON (e.g., Proshutinsky et al., 2004; Coakley et al., 2005; SEARCH, 2005).

Repeat Transects, Sections, Surveys, and Hotspots

In the marine environment, repeat sections or transects by ship or aircraft are critical means of determining spatial gradients responsible for ocean circulation, biological

Suggested Citation:"5 Designing the Network." National Research Council. 2006. Toward an Integrated Arctic Observing Network. Washington, DC: The National Academies Press. doi: 10.17226/11607.

community composition, species distributions, and ice conditions. In many cases, there are no practical alternatives to transects for collecting water samples for key chemical and biological analyses.

On land, transects provide periodic sampling of areas that are particularly remote and not represented in existing networks. By including several sites along the climate gradient (e.g., the five bioclimate subzones portrayed on the Circumpolar Arctic Vegetation Map [CAVM Team, 2003]), for example, it is possible to reduce logistics costs of visiting these areas. Furthermore, by focusing several different projects along the same transect, the climate, soil, active layer, permafrost, vegetation, and biodiversity data can be compared among projects.

The atmospheric rawinsonde network, by virtue of its fixed nature, can be considered a source of repeat surveys of the atmosphere, albeit at frequent (12-hourly) intervals. The assimilation of rawinsonde soundings into operational analyses and reanalyses adds a vertical atmospheric dimension to surface transects for which projects from various disciplines have converged. Preservation of the rawinsonde network will allow the provision of temporal context to repeat transects, sections, and surveys performed at the surface.

In some disciplines, particularly biology, improved understanding of the arctic system requires studying “hotspots” where biological activity is high and observations are dependent upon intermediate (decadal) efforts that fall between the observing needs of field programs (annual efforts) and multidecadal time series.

Remotely Sensed Observations

On the broadest spatial scales, satellites provide a powerful synoptic observation capability for a wide variety of parameters and a means of obtaining systematic, repetitive coverage for many arctic applications. Sun-synchronous polar orbiting satellites (e.g., ERS, Landsat, NOAA polar orbiting meteorological satellites, SPOT—see Annex Table 3A.3) pass over the Arctic on every orbital period, and complete coverage of the Arctic is acquired except directly at the pole. The advantage of satellite remote sensing in polar regions is that large and inaccessible areas can be studied easily and even under darkness and through fog, cloud, or rain in the case of non-optical sensor systems. This is particularly important in the Arctic, which is frequently cloud-covered and has months of darkness.

Satellite instrumentation platforms provide key observations for understanding the Arctic such as the extent of sea ice and snow. There are also technologies available for determining freshwater altimetry; bottom water pressure; sea surface heights; ice type, ice motion, and perhaps sea ice thickness; land cover; leaf phenology (e.g., date of leaf out); species changes; tundra-to-bush transition; treeline locations; and some aspects of photosynthetic activity. Satellite observations provide relevant data for almost all of the Committee’s physical key variables, many of the biogeochemical key variables, and one of the human variables (Annex Table 3A.3). In addition to the growing range of measurement capabilities, there is also a growing archive of untapped remotely sensed data that presents excellent opportunities for innovative uses and synergistic studies.

In some cases, entire satellite missions are developed to support specific arctic science initiatives. These missions often include multiple instruments on a single satellite platform. These instruments provide different but complementary measurements to one another. In other cases, the mission may be collecting global measurements or have only a subset of the instruments collecting relevant measurements for arctic science. Because of the complexity and size of each satellite program, they are typically designed and operated under the direction of large government agencies such as Canadian Space Agency and ESA (European Space Agency), or NASA (National Aeronautics and Space Administration), Department of Defense, or NOAA in the United States.

Satellite systems are reliant on ground observations for calibration and validation. Ground stations are also critical because satellite missions have finite duration and may not be replaced immediately or at all, as is the case for Landsat. Consequently, when follow-on missions are scheduled, the sensing systems are unlikely to be identical. And although these differences are part of a healthy improvement in observational capability, they create problems in continuity of measurements. One unfortunate result of the advent of satellite remote sensing is that its promise has caused the closure of many ground-based observing systems that are still needed.


Basic Principles and Strategies

The proposed or planned global observing networks, in general, have poor coverage in the Arctic and typically do not treat the Arctic as a separate region with unique observational challenges. For example, the Global Ocean Observing System has no specific Arctic Ocean component, although there are internationally coordinated groups focused on regional seas such as the Black Sea, Baltic Sea, and Mediterranean.4 Gap identification and prioritization of gap filling will therefore constitute an early phase of the AON.

In constructing the AON, it will be necessary to identify and incorporate existing observing systems because of the logistics and data that can be used as a base for future expan-

Suggested Citation:"5 Designing the Network." National Research Council. 2006. Toward an Integrated Arctic Observing Network. Washington, DC: The National Academies Press. doi: 10.17226/11607.

sion and development of an effective network. For this reason, terrestrial locations such as Abisko, Barrow, Cherski, Ny-Ålesund, Pasvik river basin, Summit, Tiksi, and Toolik Lake have obvious advantages as network nodes because scientific data collection has been under way for many decades in some cases. Although the scope of ongoing data acquisition, particularly in Russia, is uneven, the small set of existing long-term research stations identified as potential flagship observatories (Shaver et al., 2004) (that includes those listed above) provides a starting point for multidisciplinary understanding of ecosystem processes. These 6 to 12 key arctic terrestrial observation nodes would be underpinned by data and observations from supporting networks of smaller field stations and individual research projects.

This model of centralized data collection at a small number of representative sites is not applicable to all fields of scientific observation that could be affiliated with the AON. This model may also prove challenging to link with LTK that is widely dispersed and often extends over large spatial scales during seasonal food gathering cycles. In addition, some scientific disciplines have used locations that do not coincide with flagship observatories for other research fields. For example, the space physics and aeronomy research community has used locations such as Alomar, Eureka, Heiss Island, Poker Flat, Resolute, Sondrestrom, and Svalbard for observing the aurora and other upper atmosphere processes, but there may be few data available from these sites for other fields of study. Many of these intensive sites were chosen because of their location relative to the auroral oval, geomagnetic pole, or proximity to rocket ranges that provide an opportunity for in situ rocket measurements. In recent years, there has been an increased thrust within the space physics and aeronomy community to install distributed autonomous instruments in the polar regions such as magnetometers, passive optical sensors, and small radars that support its intensive observatories.

It therefore should be recognized at the outset that it is difficult to designate geographical locations of the intensive site components of the AON that would be ideally shared by all disciplines. Reanalysis of existing datasets could identify discipline-by-discipline gaps, but in the end, a practical implementation of the AON will incorporate a suite of existing and expanded sampling sites that are not likely to be ideal or shared for all disciplines. There are regional precedents among arctic networks in which such a mix can work (Box 5.3).

Assessment and Design of Observing Networks

There will be an initial assessment and design phase of the AON that leads into an operational design and optimization phase. Several steps will be required in the initial assessment and design phase, namely forming partnerships across disciplines and arctic communities, agreeing on the objectives of the network, defining the type and scale of observa-

Box 5.3
European Network for Arctic-Alpine Environmental Research

The emerging AON strategy of linking and expanding existing research infrastructure to connect environmental observations over the Arctic has regional precedents. For example, the European Network for Arctic-Alpine Environmental Research (ENVINET; was a European Union-supported “infrastructure co-operation network” that incorporated environmental observation activities at 17 research stations. The objectives of ENVINET were to

  • exchange experience and information,

  • prepare existing data for research across the stations,

  • identify needs for new data collection,

  • improve the methods for obtaining data through dissemination of “best practice” in the network and projects for scientific/technical development, and

  • coordinate activities to foster common research projects on existing datasets, sampling of new data, data management, or improved methods.

The research stations participating in the network are diverse in orientation, including the Kiruna Observatory operated by the Swedish Institute of Space Physics, the Kristineberg Marine Research Station operated by Gøteborg University, and national research facilities operated by Italy, Germany, the United Kingdom, and Norway at Ny-Ålesund in Svalbard.

Because of the diversity of research programs within the station network, as would be the case in the broader AON, not all data collected at any single station are of universal interest. As a result, several criteria were selected for the datasets targeted for ENVINET research infrastructure and network improvement. These data

  • are relevant for studying climate change, ozone depletion, long-range transport of pollutants and changes in biodiversity, and cover the needs of different sciences involved,

  • have high quality and are intercomparable across the stations, and

  • are large enough and have great enough geographic extent to study the changes of environmental phenomena in time and space, or their variation under changing environmental conditions.

European Union funds for ENVINET lasted three years and the network has not been fully functional for several years, despite its many good ideas.

Suggested Citation:"5 Designing the Network." National Research Council. 2006. Toward an Integrated Arctic Observing Network. Washington, DC: The National Academies Press. doi: 10.17226/11607.

tions, and using data assessment techniques to initially optimize the network. The assessment will need to determine the effectiveness of the AON for detecting the state of the Arctic, the speed of environmental change, the impacts of change, and net changes in fluxes.

Assessment of the adequacy of geographical density of sampling platforms will need to consider accuracy, sampling error, cost, coverage, overlap, and field reconstruction error. A key consideration is the optimal sampling distance for specific measurements, based upon cost/benefit analysis. For at least some system components, data assimilation into models will feed observing system sensitivity studies that assess key types of measurements and locations. The effectiveness of the measurement density can also be evaluated through process-oriented design, using variational methods to search for the most sensitive area for a given process. These processes will be assisted by end-users such as scientists and arctic residents. For other system components, the data fields are so undersampled that objective approaches to observing system design are not useful. In these cases, observations are a compromise between efficacy and technical, operational, and financial practicality.

One of the working assumptions in the AON design process is that 100 percent effective coverage is not likely to be necessary to establish the initial iteration of the network. For example, in the case of a sea-surface temperature network in the Baltic Sea, satellite coverage provides extensive information content and in situ data collection only improves accuracy by 20 to 30 percent (She and Høyer, 2004). Although these are probably unrealistically low limits for in situ sampling for many arctic variables, an iterative process can, in principle, reveal the sampling density needed to provide the required functionality for any particular variable. Cases where remotely sensed data can be validated directly will lead to synergistic benefits that can be applied in both synoptic understanding and modeling. In general terms, when the required sampling density is not initially clear, data can be collected and then sensitivity experiments performed to characterize the sufficiency of sampling density.

One of the strategies to optimize network development will be to employ the forementioned mix of intensive and distributed sites. This mix will vary for different themes, such as social sciences relative to physical oceanography, in part because of the varied representativeness of sites for different themes. Furthermore, improved understanding of the arctic system in some disciplines and themes requires creative combinations of transects, repeat surveys, hotspots, and integration of remotely sensed data and data from paleo studies. A flexible network geometry is therefore critical. Considering the likely variable funding situation for network elements, the network will have to be sufficiently flexible to accommodate the transient nature of some participants.

Recommendation: A system design assessment should be conducted within the first two years of AON development—that is, as a component of IPY—to ensure a pan-arctic, multidisciplinary, integrated network. This effort should be undertaken by a diverse team, with participation and input from multiple disciplines, stakeholder groups, and those involved in related international observing activities. The assessment should use existing design studies, models, statistical approaches, and other tools.5

The initial system design assessment, in conjunction with the pulse of new data from the IPY projects, will provide valuable guidance for enhancing the AON and maximizing its potential utility, as will information on specific performance metrics and user feedback.

Recommendation: The AON should be continuously improved and enhanced by taking advantage of the findings and recommendations in the system design assessment and performance metrics and data provider and user feedback that will become an enduring component of the network.

For any one discipline, many useful sampling locations can be specified (see, e.g., SEARCH [2005] or the distribution of Arctic Monitoring and Assessment Programme network sites). But, as stated at the outset, one set of locations cannot be expected to be ideal for all disciplines. Thus it is probably more useful to outline strategies for system design and incremental expansion. Individual networks that can be better networked are a starting point for such discussion, and this is probably a more practical approach than to specify “ideal” locations where sampling should be coordinated. While recognizing that specification of locations is ultimately affected by the variables to be measured, the Committee provides examples in the next section of locations that are crucial within specific disciplines and themes and could be early targets for preservation or gap filling.

Specific Geographic Considerations


Routine atmospheric observations are made at many locations for use in operational weather forecasting. These observations would be potentially valuable elements of the AON, provided that continuity, homogeneity, and spatial distribution are used to filter them for use in monitoring. Assimilation of data into models can provide one such filter. Recent losses of weather stations in Russia and Canada are potential concerns with respect to coverage, however. In addition, upper troposphere and lower stratosphere measurements from in situ instruments (e.g., rawinsondes) over the


See Chapter 6 for specifics relating to this recommendation.

Suggested Citation:"5 Designing the Network." National Research Council. 2006. Toward an Integrated Arctic Observing Network. Washington, DC: The National Academies Press. doi: 10.17226/11607.

Arctic Ocean are notably absent. Fortunately, in situ surface measurements from drifting buoys are valuable sources of information on sea-level pressures and near-surface temperatures over the central Arctic Ocean.

Fixed sites for comprehensive atmospheric sampling, including those monitoring pollution levels, are already well established at Alert (replacing Station Nord in Greenland that was operated until 2002), Amderma, Ny-Ålesund, Pallas, and Summit, although varied sampling regimes complicate intercomparison. Some sites provide an observational database for inorganic contaminants and trace metals, while others (e.g., Barrow) are part of trace gas (carbon dioxide) monitoring networks. In addition, there is a major observational gap from Alert (on Ellesmere Island, Canada) across arctic Canada, Alaska, and Siberia—or approximately 75 percent of the Arctic. Additional sampling locations might be needed at air mass boundaries or to detect specific sources such as Asian dust or air pollution (e.g., an Aleutian or Alaska Peninsula location such as Attu, Cold Bay, Dutch Harbor, or Shemya to sample the North Pacific boundary region) and a western Siberian site for industrial activities. Because of the need for sampling representative air masses and boundaries, a minimal atmospheric network would also include representative air mass sampling locations such as the central Arctic Ocean, the Canadian Arctic at Alert, and the Lena River delta at Tiksi. Sampling for trace gases like carbon dioxide and methane can be sustained with a lower sampling density than organic contaminants and heavy metals.

The upper atmosphere from the mesosphere through the thermosphere and into the magnetosphere plays a key role in the transfer of solar energy from the sun to the lower atmosphere and Earth’s surface.6 Measurements of upper atmosphere processes are therefore important for the AON. Many upper atmospheric process are global in extent and require a relatively uniform distribution of sensors in longitude (approximately 6 to 12) per latitude and with latitudinal gradients that require measurements at every 4 to 8 degrees. The ionospheric and magnetospheric processes, which are geomagnetically aligned, have different spatial distribution requirements and must consider the location of the auroral oval and geomagnetic polar cap. Examples of existing upper atmospheric networks are the SuperDARN (Super Dual Auroral Radar Network) and CANOPUS (Canadian Auroral Network for the Open Program Unified Study) magnetometer network. These networks include high-frequency radars for determining the auroral ionospheric convection pattern and a distributed array of autonomous magnetometers.

Major gaps in upper atmosphere measurements exist in Russia, where many sites were decommissioned, and over the Arctic Ocean, where the required infrastructure is not available. The lack of measurements in Russia inhibits study of the full extent of upper atmosphere processes that are inherently global.

Marine Systems

For the purposes of describing the AON, the arctic marine system includes the Arctic Ocean and subarctic seas, and the overlying sea ice cover and atmosphere. Key processes in arctic marine systems include ocean, ice, and atmospheric circulation, water mass distributions, chemical transport and fluxes, and biological variability.

The SEARCH Science Plan (SEARCH SSC, 2001), which is oriented toward understanding functional change in the Arctic on a system basis, called for a coordinated program of long-term, pan-arctic observations. This plan has guided development of implementation plans (e.g., SEARCH, 2005) and recommends locations for intensive and distributed marine observatories (DMOs). A number of these DMOs are already in place at some functional level (see Figure 4 in SEARCH SSC and IWG, 2003) although key observational gaps remain including the Makarov Basin and the northern part of the Canadian Basin, particularly along the Canadian Archipelago and the Siberian shelves.

DMOs may include many elements but these can be grouped into three categories: moorings, instrument systems tethered to the sea ice, and repeat sections. These activities can and should be done together if only to minimize costs and ensure optimum coverage. Moorings are targeted for ocean pathways along the continental slopes and mid-ocean ridges, cross-shelf exchange sites such as shelf canyons, basin sites that are good indications of large-scale circulation changes, gateways to the Arctic Ocean (e.g., Fram Strait, Bering Strait, and the Canadian Archipelago), and representative shelf sites. Ice-tethered instruments provide observational coverage of the upper ocean, ice, and atmosphere measurements, and are vital to measuring the conditions and fluxes at the air-ice-ocean interface that are critical to both the ice-albedo and global thermohaline circulation feedbacks on global climate. In contrast to moorings and fixed coastal observatories, ice-tethered ocean buoys provide a Lagrangian description of mixed layer changes and interactions with the ice and atmosphere, as well as hydrographic sections over drift paths. Repeat sections, which can be obtained by aircraft in addition to ships, will need to include sampling across the critical straits (e.g., ASOF sections),


Nearly half of the incident solar radiation on Earth is absorbed by the atmosphere before reaching the surface. Energetic particles that are output from the sun during large solar storms, called coronal mass ejections, impact the Earth’s magnetosphere and travel down the high-latitude magnetic field lines into the atmosphere and result in dramatic displays of the northern lights. These processes also cause a heating and thus an expansion of the atmosphere in addition to a redistribution of atmospheric constituents through transport processes and solar-related chemical processes. Although the variation in total solar irradiance is small, variations in specific wavelength bands such as the extreme ultraviolet can be dramatic over the course of the 11-year solar cycle or a short-term solar event such as a flare or coronal mass ejection.

Suggested Citation:"5 Designing the Network." National Research Council. 2006. Toward an Integrated Arctic Observing Network. Washington, DC: The National Academies Press. doi: 10.17226/11607.

sections radially across the major deep basins and key pathways, and on the arctic shelves. The Arctic Ocean Science Board has endorsed an integrated Arctic Ocean Observing System Shelf-Basin Exchange initiative as an IPY activity7 that proposes standard sections from shelf to basin and across key gateways on a pan-arctic basis. Continuation of these observational transects beyond the IPY could provide the basis for long-term pan-arctic DMO sections, as can sections that are now being regularly occupied as part of existing prototype marine observatories.

Finally, observational programs should be further elaborated and continued for areas of high biological productivity. Hotspots of productivity have been identified and monitored with varying temporal resolution in the Bering and Chukchi Seas, Barrow Canyon, the Boulder Patch (kelp-based community) in the Beaufort Sea, and the Gakkel Ridge spreading plate boundary. Other localized areas of high productivity exist throughout the Arctic, including within recurring polynyas, at Hat Island in the Canadian Archipelago, in Baffin and James Bay, in the White Sea, and along the Kola Peninsula. Changes in these biological communities can have a direct bearing on food webs and human communities as well as illuminating impacts of environmental change, particularly upon subsistence food gathering communities that are located in areas they long ago identified as having high biological productivity.

Coastal Zones

Coastal locations are home to most of the largest cities and highest density human populations in the Arctic (State Committee of the USSR on Hydrometeorology and Controlled Natural Environments, 1985). It would therefore be beneficial to centrally locate most coastal observatories close to or within human communities, particularly where communities are co-located with sites that are independently identified in scientific planning as important for arctic time-series and observations. Because of the importance of coastal zones for human use, runoff, and terrestrial ecosystem exchanges, some observation systems will need to be situated to monitor coastal zone processes that are often missed in separate marine and terrestrial research efforts (Cooper, 2003).

Important details of coastal waters, including fast ice thickness and the separation of chlorophyll pigments relative to river plumes, are not well resolved by satellites, so development of observational systems in coastal and estuarine waters may be particularly challenging yet critically important for biological systems used by human communities. There are several examples of emerging observing coastal networks. One example is called Arctic Coastal Dynamics (Box 5.4).

Box 5.4
Arctic Coastal Dynamics

Arctic Coastal Dynamics (ACD)a documents changes in coastline configuration and promotes research on causative factors for changes in coastal dynamics, including erosion. Sites that are currently incorporated in this network are documented on the ACD Web site and include Barrow, Kongsfjorden (Ny-Ålesund), and Zackenberg, where other marine and/or terrestrial observations and intensive research are ongoing.

Improving observation systems in Russian arctic coastal regions will be crucial because these zones are key areas for sea ice formation (e.g., Laptev and East Siberian Seas). The Russian Arctic also has the largest undersea permafrost deposits, which may be an important source of dissolved and particulate organic carbon, carbon dioxide, and methane in response to climate warming and increased coastal erosion as sea ice cover decreases. Because of these unique characteristics associated with the Russian Arctic, the AON without significant Russian participation will not meet the goal of providing an observational basis for monitoring the state of the Arctic or understanding arctic environmental change and connections with the global system.

Terrestrial Systems

The concept of focused intensive sites with multidisciplinary research activities is particularly appropriate in terrestrial systems. Vegetation maps can be used to identify areas that are underserved by existing observatories. Observational gaps will need to be identified and filled at major boundaries such as the current tree line, at plant community boundaries (e.g., tundra to shrub transitions), and at locations where soil moisture conditions are changing. LTK of plant communities can help in this effort.

Physiographic processes such as the lack of glaciation in eastern Siberia during the last glacial period have resulted in biological communities that are not present in other portions of the Arctic. Globally significant peat deposits are also located in Russia, particularly in the western Siberian low-lands, and these deposits are likely to be important in affecting global carbon cycling in many global warming scenarios (Frey and Smith, 2005). All of these factors illustrate the need for better observational coverage in the Russian Arctic.

The most remote and coldest areas of the Arctic are home

Suggested Citation:"5 Designing the Network." National Research Council. 2006. Toward an Integrated Arctic Observing Network. Washington, DC: The National Academies Press. doi: 10.17226/11607.

to terrestrial life at the extreme. These areas may show some of the most dramatic physical and biological changes because their present-day summer temperatures are close to freezing and large temperature changes are likely in these areas if sea ice distribution changes dramatically. Repeat transects are useful for sampling these terrestrial environments that are not represented by existing research facilities. One such transect is the North American Arctic Transect that connects sites along the Dalton Highway in northern Alaska (Toolik Lake to Howe Island) and sites in the western Canadian High Arctic. The Canadian sites include Inuvik, Green Cabin, Mould Bay, and Isachsen. This transect was established as part of a National Science Foundation (NSF)-funded Biocomplexity in the Environment project (Walker et al., 2004). Another such transect is being considered for the Yamal Peninsula and Novaya Zemlya in Russia. This transect could attract a variety of research and link to several ongoing circumpolar and global research projects such as the International Tundra Experiment (ITEX), the Thermal State of Permafrost (TSP) project, the Circumpolar Active Layer Monitoring (CALM) project, the FLUXNET8 project, and the Circumpolar Arctic Flora and Fauna Biodiversity Monitoring Program.


Ten percent of world runoff enters the Arctic Ocean. Recent observations show increasing runoff (Peterson et al., 2002), and arctic runoff variations potentially influence freshwater balance and ocean stratification, including vertical mixing in the North Atlantic. Despite the importance of river observations, there has been a widespread loss of hydrological monitoring networks over the past 15 years in the United States, Canada, and Russia (Shiklomanov et al., 2002). This decrease in monitoring increases the uncertainty in the estimates of water, chemicals, and sediments from the pan-arctic drainages (Shiklomanov et al., 2002). At a minimum, these observation systems need to be restored and enhanced for major Arctic rivers, including the Yukon, Mackenzie, Yenisey, Ob, and Lena. But there is an important caveat. These five largest arctic rivers have drainages that extend well into temperate latitudes,9 so these rivers are not necessarily representative of organic carbon, tracers, and other materials contributed to the Arctic Ocean from tundra systems drained by smaller rivers and from retreating shorelines. Furthermore, gauged discharge from the ten major arctic river basins captures only approximately one percent of pan-arctic glacier discharge to the Arctic Ocean (Pfeffer and Dyurgerov, 2005) and total pan-arctic glacier contributions to arctic runoff are 30 to 40 percent uncertain due to sparse and diminishing glacier mass balance observations (Pfeffer and Dyurgerov, 2005). In other words, basin-wide precipitation and glacier runoff not carried by major river systems is very poorly characterized by present-day programs. As a result, sampling only the largest rivers will not produce a representative view of all runoff contributions to the Arctic. Additionally, the freshwater inflow through Bering Strait also appears to have been underestimated (Woodgate and Aagaard, 2005), so an arctic river sampling program that is at least minimally coordinated with Bering Strait sampling is critical for observations of freshwater balance in the Arctic Ocean.

Renewing and improving hydrometeorological observation networks in Russia is crucial for the freshwater flux measurements within the AON. The recent memorandum of understanding between the Russian State Committee for Hydrometerology (Roshydromet) and the U.S. National Weather Service10 is an example of the wider multigovernmental participation that will be needed to fully develop the AON. In the case of this memorandum, which may lead to World Bank funding, Roshydromet is seeking support to upgrade and modernize its basic hydrometeorological capabilities. This agreement will have benefits beyond Russia and the Arctic, and is seen as a vital link in building the Global Earth Observation System of Systems.


Permafrost has received much attention recently because surface temperatures are rising in most permafrost regions bringing permafrost to the edge of widespread thawing and degradation. Permafrost thawing that already occurs at the southern limits of the permafrost zone can generate dramatic changes in ecosystems and infrastructure functionality (Jorgenson et al., 2001; Nelson et al., 2002). However, there is no global database that defines the thermal state of permafrost within a specific time interval. Internationally, borehole temperature measurements have been obtained at various depths and periods over the past five or more decades indicating changing permafrost temperatures at different rates for different regions. Several cryospheric observation systems are already under way and can serve as a geographical focus for permafrost monitoring and other related observational activities. The International Permafrost Association11 is responsible for developing and implementing the GTN–Permafrost12 network of the Global Climate Observing System. The CALM network is a component of GTN-Permafrost and is providing data on active layer thickness measurements using approved protocols (Brown et al., 2000) (Box 5.5).

Suggested Citation:"5 Designing the Network." National Research Council. 2006. Toward an Integrated Arctic Observing Network. Washington, DC: The National Academies Press. doi: 10.17226/11607.

Box 5.5
The Circumpolar Active Layer Monitoring Network: Lessons in Data Harmonization

The Circumpolar Active Layer Monitoring (CALM) network was initiated in the early 1990s. Its initial focus was on formalizing measurement programs at sites that were making active layer measurements or had done so in the past, creating new observatories, and creating coherent data archives. Data on active layer thickness are now available for 130 locations in the Arctic, the Antarctic, and at lower latitudes where permafrost occurs. Participants from 15 countries are involved with CALM, many of them on a voluntary basis.a

One of the initial concerns was that active layer measurements were collected using a wide variety of methods. Because there was a paucity of information about how observations collected by the different methods related to each other, the issue of sampling design (e.g., spatial sampling for probed data) was also a concern. Early efforts in CALM therefore focused on comparing methodologies and sampling designs (e.g., Fagan, 1995; Nelson et al., 1998, 1999; Brown et al., 2000; Gomersall and Hinkel, 2001).

Another component of CALM created a protocol for field measurements (Nelson et al., 1996; Nelson and Hinkel, 2003). CALM developed links with the International Tundra Experiment (ITEX) and its field measurement protocol was included in the 1996 ITEX field manual. Network representatives also made presentations on CALM and the need for standardized observations at a variety of meetings. Actions on these various fronts helped to create a system of standardized measurements as the network grew rapidly in the late 1990s.

One of CALM’s most useful roles has been to provide reliable, relatively unambiguous data for model validation (F.E. Nelson, personal communication, December 2005). The data available from the network through the Frozen Ground Data Center (a component of the U.S. National Snow and Ice Data Center) are well documented, and shortcomings are noted. Working with data center staff (specialists in data organization, storage, and access) was a critical step in making the data useful to modelers and others.

Improvements in data and network design continue. CALM network members have started to analyze shortcomings in the geographic coverage of the network and further analysis of the accuracy of data collected using different techniques and sampling designs is under way.

A number of candidate sites have also been identified for a global borehole network for monitoring the thermal state of permafrost (Romanovsky et al., 2002).13 As part of the IPY, the International Permafrost Association plans to develop a network of permafrost boreholes for long-term observations. This network will include hundreds of sites in both hemispheres and initially involve participants from 20 countries. Some of these sites are located where CALM sites have also been implemented. In the United States, several of the borehole sites are located near or adjacent to areas of concentrated terrestrial ecological research. For example, Galbraith Lake and Imanavait Creek, both near Toolik Lake, are borehole and CALM sites.

Glaciers and Snowcover

Observations of glacier ice mass at high latitudes are important for understanding potential sea-level fluctuations and arctic runoff but are being discontinued. Glacier mass balance programs, such as in Alaska (e.g., the nearly defunct Alaska mass balance transect), Canada (e.g., Ellesmere, Axel Heiberg, Devon, and Penny Ice Cap), and Russia (e.g., Severnaya Zemlya), are in need of revitalization, maintenance, or expansion to extend valuable long-term time series. Furthermore, glaciers that are located in the subarctic will need to be monitored and linked into the AON, not least because of their rapid retreat (Arendt et al., 2002). In addition to contributing to river flow and sea-level rise, the loss of glacier ice is endangering potential proxy paleoclimate records that could provide information on Earth system history. The AON will need to provide for the contingencies of collecting and analyzing paleoclimate records stored in glacier ice that might otherwise be destroyed with continuing widespread glacier retreat.

Airborne and satellite-based laser altimetery, historical photographs, and remote sensing technology are among the resources available to supplement ground-based observations and to document changes in arctic glacier ice cover more broadly (e.g., Cogley and Adams, 2000; Zeeberg and Forman, 2001; Thomas et al., 2003). Because glacier observations have become highly dependent on remotely sensed images and data, the geographical co-location of terrestrial, cryospheric, and coastal observatories with glacial observatories is not critical. However, seasonal snow cover observations are closely related to vegetation patterns and albedo (Liston et al., 2002), so coordination of some snow cover observation programs with other environmental observations will be critical. Seasonal snow cover is an important hydrologic resource as well as an essential component of the physical system, and knowledge of snow water equivalent and albedo is particularly valuable. To facilitate calibration of remote sensing of these variables, in situ

Suggested Citation:"5 Designing the Network." National Research Council. 2006. Toward an Integrated Arctic Observing Network. Washington, DC: The National Academies Press. doi: 10.17226/11607.

measurements in areas of large interannual variability will need to be placed to enable upscaling to watersheds and satellite pixels (for validation of remote sensing), and for incorporation into system models.

Arctic Residents

Many arctic residents are widely dispersed in small communities throughout the Arctic that are distant from the few centralized flagship observatories that are likely to emerge, at least for the terrestrial components of the AON. A somewhat different situation is present in Russia, where the largest arctic population centers are located, but there are only minimal observations for a number of important parameters of environmental change and human health. Also, the Russian portion of the Arctic has limited documentation of LTK compared to North America and Fennoscandia (Huntington and Fox, 2005). Providing linkages from local communities to the wider network and collecting data of value to arctic residents are key components for successful AON implementation. An illustration of the types of data that could arise from widespread, distributed community observations is provided by Riedlinger (1999).

Our first visit to Sachs Harbour substantiated and added to the initial observations that had provided the basis for the project. It was clear that environmental change had not gone unnoticed. In the first day of workshops, the community discussed the accumulating evidence of changes occurring in the landscape around them. They described freeze-ups that were three to four weeks late and severe storms with wind, thunder, lightning, and hail. They discussed intense, unpredictable weather and fluctuations in seasons. Hunters described not seeing ice floes in the summer anymore, umingmak (muskox) being born earlier, geese laying eggs earlier, and nanuq (polar bears) coming out from their dens earlier because of warming and thaw. They also described catching species of Pacific salmon (identified by the Department of Fisheries and Oceans as sockeye Oncorhynchus nerka and pink Oncorhynchus gorbuscha) in their nets, when traditionally such occurrences were unheard of. Too much open water in the winter was making harvesting animals difficult, as was the lack of snow in spring, the lack of sea ice in summer, increased freezing rain, and thinner ice.

Given this example of the plethora of climate-related observations that are available from one small community in the Canadian Arctic, a valid answer to the question as to where the AON should be geographically located is that it should be located everywhere that people live, hunt, fish, and work in the Arctic. Ultimately, a successfully implemented AON will link those widely distributed observations with intensive, intercomparable data collected at centralized, flagship sites.


There are many design considerations for positioning observations in the AON. Often these are discipline- or theme-specific. However, there are some generalizations that emerge for the initial phase of building the AON. For example, existing observatories and platforms will form the core of the AON if they are sustained. Additionally, even though observing systems are deteriorating in some parts of the Arctic—especially noticeable in Russia—there is still some possibility to reverse this process and recover some of the infrastructure that was available recently.

Recommendation: The first phase of AON development will require sustaining existing observational capabilities (including those under threat of closure) and filling critical gaps.


The arctic environment creates challenges to sensing in all realms. For example, the majority of the Arctic Ocean is covered by ice that makes measurements of key ocean variables (e.g., water temperature, salinity, currents) difficult without the use of ice breakers, submarines, or aircraft. Terrestrial and atmospheric instruments must contend with the extreme temperatures, high winds, ice accumulation, and damage from animals. Instruments that are remotely deployed must also contend with power generation issues during the polar night. And people working in the field need improved technology for such actions as downloading data and viewing computer screens under harsh conditions. These are only a few of the challenges that must be surmounted when developing sensing technology and supporting infrastructure for the Arctic.

Technology development and innovation must and will play a pivotal role in the continued development of the AON. There are two key areas where technological advancements can contribute: improving infrastructure that will benefit all realms of the AON and enhancing sensors that measure key variables in the Arctic.


It is often assumed that because a technology already exists it should not be difficult to put the pieces together and make things work. This is a fundamental misconception. For example, consider a medium-sized scientific satellite that costs $300 million. The technology integrated into the satellite has “flight heritage” and has been used elsewhere so the fundamental engineering effort is on the interfaces between the system elements to ensure they all work together and will achieve the mission goals. Because the satellite will be placed in a harsh and remote environment, significant efforts are made to identify and manage potential risks and increase the probability of success. This so-called “systems engineering approach” is fundamental for complex technology development projects like those that will support the AON. The single investigator approach where small numbers of measurement systems are designed and deployed individu-

Suggested Citation:"5 Designing the Network." National Research Council. 2006. Toward an Integrated Arctic Observing Network. Washington, DC: The National Academies Press. doi: 10.17226/11607.

ally—while useful for development of new sensors—is not a sustainable model in AON infrastructure development. The AON will have an element of “big science” that will require significant coordination among scientists who are experts in application and engineers who are experts in technology.

Four areas of infrastructure technology are key to the development of the AON: communications, navigation, power storage and generation, and automation.


Communication to instrument platforms is needed for retrieving data and determining instrument status. Where bidirectional communications are available this link can also be used to remotely control the instrument. The requirements on a communications link vary depending on the measurement and observational cadence of the instrument. For some long-term monitoring applications, for example, it may be possible to retrieve data by physically visiting the site each year and collecting the data. However, without a near real-time communications link one cannot be sure the instrument is running until the annual visit and a system failure can result in significant data loss. In many cases, data are required either in real time for nowcast or forecast applications or within a few weeks of collection for scientific applications. Such requirements necessitate the use of real-time wireless communications, typically via satellite using the Iridium or a similar system.14 Increased connectivity to remote instruments will improve reliability because data can be streamed directly back to data analysis centers where operators can monitor the instrument performance in near real-time. In the event of a problem, the system can be reconfigured remotely as soon as an error is reported. Furthermore, increased connectivity would reduce costs in situations where a device no longer needs to be physically recovered to download its data. One example is a data uplink from subsurface moorings by means of an automated underwater vehicle or glider serving as a data shuttle—thus removing the need to recover the mooring to obtain its data.

Communications are therefore an important focus for innovation and technology development. Standard interface and communication protocols will allow for ready integration of instrumentation into the AON. There is a fundamental need to ensure that the elements within the AON can exchange data among mobile platforms (e.g., undersea vehicles, gliders, and either floating or other ice-surface vehicles), vessels (both surface and submarine), and aircraft, using both acoustic and radio frequency, so that the data collected will be retrievable by multiple pathways. This capability needs to be managed to minimize power drain (e.g., through scheduling or when data volume reaches a specified level), while remaining reliable and robust so that the failure of a single antenna or transmitter does not cause data loss.


All mobile observing systems need navigational capabilities. It is therefore of high priority to develop standards for navigation beacons. In the marine setting, by way of illustration, the highest priority for acoustic navigation is to determine the sound frequency and to ensure that deployed sound sources are compatible with all potential systems that are now under development. Some of these problems may be resolved under activities of the NSF PLUTO concept (Polar Links to Undersea Telecommunications and Observatories), which is particularly versatile and would link acoustic tomography, cabled observatories, moorings, gliders, and UAVs to provide coordinated synoptic arctic datasets. In cases where acoustic sensors are not available, low-cost inertial sensors must be used. These sensors could be used in UAVs and AUVs.

Power Storage and Generation

The power requirements of systems can range from very low power (a few watts) for underwater gliders to kilowatts for autonomous land-based observatories. Gliders can operate for extended periods of time using energy stored in batteries. However, the higher power land-based observatories either require a combination of renewable energy sources, such as solar and wind power, or large quantities of stored chemical energy, such as propane. Continued research and development on efficient sources of power generation (e.g., fuel cells) and power storage will undoubtedly benefit the AON. Although it is doubtful that the AON will have the resources to drive the technology in this area, its members would need to be active in the discussion of technology requirements to ensure that potential solutions can operate in the harsh arctic environment. The AON could benefit from collaboration with other technical entities that have power systems development programs (e.g., ESA in Europe or NASA and the Department of Energy [DOE] in the United States).

Increased Automation

Automation can enhance all network components through increased efficiency in data collection and improved data quality. An ability to automatically measure variables is particularly appealing in the harsh arctic environment where the cost and risk to humans is high and where human observers cover a small geographic area.

The variety of autonomous platforms includes anchored ocean moorings, autonomous underwater vehicles, ocean


Over the continental United States, satellite television vendors offer real-time Internet access with data rates in excess of 1 megabit per second (compared with the 10 kilobits per second data rates for Iridium).

Suggested Citation:"5 Designing the Network." National Research Council. 2006. Toward an Integrated Arctic Observing Network. Washington, DC: The National Academies Press. doi: 10.17226/11607.

gliders, autonomous aerial vehicles, and autonomous observatories based on land, ice, and the ocean bottom. These instruments may operate unattended for extended periods of time and may need a self-contained power generation or storage system, a basic data logging capability, a mechanism for transferring data to a central processing location, and a sophisticated navigation system if they are mobile. Although simple systems can operate with little or no user input or control, more complicated systems such as autonomous unmanned vehicles require a robust supervisory control system. Automation is difficult to centralize because the requirements for each system can be vastly different. However, where there is overlap between instrument systems, coordinated development of resources will create efficiencies. Development of proprietary source code is detrimental to this approach. Instead, the distributed open source model or a more coordinated approach would benefit the AON.


The development and integration of sensors within the AON will need to include three components: (1) adapting existing sensor technology to arctic requirements, (2) identifying observational platforms that can benefit from common technology and develop the hardware to meet these needs, and (3) determining where the AON has gaps caused by a lack of sensor availability and fill these gaps by investing in next-generation sensor research and development.

Adapting Existing Sensor Technology to Arctic Requirements

Not all new technology will be directly transferable to the arctic environment because of special challenges for developing instrumentation for polar regions. A case in point is the emerging micro- and nano-fabrication capability that has revolutionized sensory systems. Micro- and nano-electromechanical systems (MEMS and NEMS) have many advantages over their macro-scale counterparts: lower cost, smaller volume and weight, and lower power consumption. And MEMS/NEMS sensors can play a major role in the next generation of measurement systems. To reliably operate in harsh conditions, however, special sensors must be designed. To date, such developments have been limited due to the lack of commercial interest resulting from the very small potential market. Adapting MEMS/NEMS and other sensors and developing lower power versions of existing proven devices will need to be supported within a broad initiative for adapting new technologies that spans all disciplines and interests.

Common Technology

The cross-cutting nature of the AON lends itself to the development of sensors that can be deployed from a variety of arctic platforms. Shared sensors will increase the cost-effectiveness and maintainability of AON observing systems. For example, a set of sonar transducers and electronics can be mounted on an AUV to map the sea floor, or alternatively, attached to an ice-based platform, or a moored platform in an upward-looking direction to measure changes in sea ice draft over a broad swath. Ice mass-balance buoys being developed at the U.S. Cold Regions Research and Engineering Laboratory can also be co-located with other moored ice profile sonars, on helicopter or submarine surveys, or on other drifting buoys. There are a handful of other common marine-oriented sensors that are also useful in a broad spectrum of research disciplines15 and the AON could benefit from their shared development. These sensors include sonars, current profilers, CTDs (conductivity, temperature, and depth sensors), fluorometers, transmissometers, in situ nutrient analyzers, and optical plankton counters.

Common atmospheric variables with broad applications include temperature, pressure, and relative humidity. A small and simple measurement package that senses these or other broadly applicable key variables could be deployed on UAVs, autonomous meteorological stations, or perhaps on snowmachines. These small packages could be deployed as sondes and dropped from manned or unmanned aircraft to measure the vertical structure of the atmosphere or distributed across the ice and land to create an integrated sensor network. By also equipping these packages with small, low-power global positioning system receivers, accurate time and position information can be obtained and, in cases where the sensor resides on a nonstationary surface such as sea or glacier ice, used to infer motion.

Filling Gaps Created by Lack of Appropriate Sensor Technology

Measurement requirements may not be satisfied by existing sensing technologies. The result is a measurement gap. In cases where measurement gaps are created, research and development will be required to create the sensor. The time and effort required to conduct sound engineering and development should not be underestimated, and the decision to invest in such efforts will be aided by the AON system design activity, which will be broad-based and have a strategic outlook. Because of the cost and time considerations, a prioritization effort is needed to identify sensor technologies that are applicable to multiple disciplines and are critical to the overarching goals of the AON. Where possible, such efforts would benefit from coordinating with entities that have overlapping interests (e.g., DARPA, DOE, ESA, NASA, NSF). For example, sensor technology that is being developed for sensing chemical and biological agents for


For example, acoustics; physical, biological, and chemical oceanography; marine geology and geophysics; and cryogenics.

Suggested Citation:"5 Designing the Network." National Research Council. 2006. Toward an Integrated Arctic Observing Network. Washington, DC: The National Academies Press. doi: 10.17226/11607.

security applications could be leveraged for arctic observing—either by directly applying these technologies or with some modification for arctic operation.

Novel Technology and Arctic Communitites

Arctic residents can bring valuable insights to the design and deployment of new technologies and they provide a year-round presence to supervise testing, provide maintenance, and help with other aspects of novel technology development.

In addition, technology will facilitate the participation of arctic residents (particularly educators and students) in the AON. For example, advances in health technologies may offer better ways for locals to monitor disease or contaminants. Local people have ideas about technologies that will help them make observations and enhance community-based observation and knowledge documentation programs that link into the AON. The AON will need to strive to understand the technological needs of arctic communities to participate in the network.

Issues for Technology Development

A number of issues need to be addressed to maximize the contribution of technology development to the AON. Four of these issues are illustrated with U.S. examples because the sponsor of the study is a U.S. funding agency, but the issues are more broadly applicable.

First, the present research funding paradigm (e.g., in NSF) does not promote technology development for an AON-type endeavor. Under the present system, a single or small group of investigators is allocated resources to address a specific scientific problem. To receive funding, investigators are often overambitious in their technical objectives and do not have the time or resources to conduct proper engineering studies. Instruments are deployed without proper testing and may or may not operate as intended. Such an approach leads to unreliable equipment that cannot be propagated into future projects and provide a solid technological foundation. A coordinated engineering approach is needed where centers of excellence in polar engineering technology foster technology development. This is not without precedent. From the late 1940s until the early 1980s, the Naval Arctic Research Lab in Barrow promoted technology developments that supported polar research.

Second, support for technology development, for example by the Office of Naval Research and the Air Force Office of Scientific Research, has declined significantly in recent years. Without this or similar investment into developing, integrating, and deploying the technology, new arctic technology development will decline and adversely affect the ability of the AON to achieve its goals.

Third, validation and calibration of sensors is often neglected and little if any support is available for such efforts even though they can be complicated and time consuming. Validation and calibration are critical for long-term measurements and are as important as science goals. Without these measurements, it is impossible to conduct long-term monitoring and science since the measurement techniques and technology will inevitably change.

Fourth, the complexity and rapid evolution of technology makes it difficult for individuals or small teams to stay abreast of all advances, become experts in particular fields (e.g., communications, power storage, and power generation), or to tackle broad challenges that could address the overarching needs of networks such as the AON. Without a process for tracking and considering the benefits of all potentially useful technological advances (and identifying critical technology gaps), the AON will fall short of its goals.


The potential for continued technological improvement to develop the AON is strong and will begin to be realized as these issues are addressed. Technology will continue to evolve and the state of technological advancement cannot be predicted. The AON will need to continuously evaluate the readiness levels of technology that could significantly enhance measurement of key variables in the Arctic. Technology that will improve current measurement quality and reliability will need to be continuously evaluated and strategically introduced. The AON will need engineering expert groups that track developments in sensor and infrastructure technology16 and weigh actions (such as modifications of non-arctic technologies or development of new technology) that address overarching network needs expressed by the AON community and its users.

In conjunction with and downstream of these groups, the AON will need centers of excellence to develop and adapt existing sensor and infrastructure technology. These centers could coordinate with small businesses and with experts on specific technologies. The centers could also coordinate with a technology incubator program based on a competitive peer review process that provides resources to individual investigators or small teams to develop critical new technologies for the AON. Finally, the AON will need to foster cultural change on two fronts. The first change places infrastructure and sensor development, and validation and calibration of instruments, on the same level of importance as the resulting science and operational value. The second change replaces the culture of small independent research groups reinventing


This tracking could be enhanced by creating linkages with other groups that have shared networking challenges but not necessarily an arctic focus. In the U.S., these three examples are the National Ecological Observatory Network, the Consortium of Universities for Advancement of Hydrologic Sciences, and the Alliance for Coastal Technologies.

Suggested Citation:"5 Designing the Network." National Research Council. 2006. Toward an Integrated Arctic Observing Network. Washington, DC: The National Academies Press. doi: 10.17226/11607.

the wheel to solve their arctic infrastructure problems with a coordinated pan-arctic approach. This latter approach could embrace the systems engineering paradigm that the private sector and NASA use to tackle complex multidisciplinary engineering problems.

Recommendation: The AON should support development, testing, and deployment of new sensors and other network-related technology. In parallel with recognizing the importance of systems engineering and instrument validation and calibration, this will require supporting (i) expert groups to track advances in technology that satisfy overarching network needs and (ii) centers of excellence and a technology incubator program to adapt and develop needed technology.17


See Chapter 6 for detailed implementation ideas that relate to novel technology.

Suggested Citation:"5 Designing the Network." National Research Council. 2006. Toward an Integrated Arctic Observing Network. Washington, DC: The National Academies Press. doi: 10.17226/11607.
Page 71
Suggested Citation:"5 Designing the Network." National Research Council. 2006. Toward an Integrated Arctic Observing Network. Washington, DC: The National Academies Press. doi: 10.17226/11607.
Page 72
Suggested Citation:"5 Designing the Network." National Research Council. 2006. Toward an Integrated Arctic Observing Network. Washington, DC: The National Academies Press. doi: 10.17226/11607.
Page 73
Suggested Citation:"5 Designing the Network." National Research Council. 2006. Toward an Integrated Arctic Observing Network. Washington, DC: The National Academies Press. doi: 10.17226/11607.
Page 74
Suggested Citation:"5 Designing the Network." National Research Council. 2006. Toward an Integrated Arctic Observing Network. Washington, DC: The National Academies Press. doi: 10.17226/11607.
Page 75
Suggested Citation:"5 Designing the Network." National Research Council. 2006. Toward an Integrated Arctic Observing Network. Washington, DC: The National Academies Press. doi: 10.17226/11607.
Page 76
Suggested Citation:"5 Designing the Network." National Research Council. 2006. Toward an Integrated Arctic Observing Network. Washington, DC: The National Academies Press. doi: 10.17226/11607.
Page 77
Suggested Citation:"5 Designing the Network." National Research Council. 2006. Toward an Integrated Arctic Observing Network. Washington, DC: The National Academies Press. doi: 10.17226/11607.
Page 78
Suggested Citation:"5 Designing the Network." National Research Council. 2006. Toward an Integrated Arctic Observing Network. Washington, DC: The National Academies Press. doi: 10.17226/11607.
Page 79
Suggested Citation:"5 Designing the Network." National Research Council. 2006. Toward an Integrated Arctic Observing Network. Washington, DC: The National Academies Press. doi: 10.17226/11607.
Page 80
Suggested Citation:"5 Designing the Network." National Research Council. 2006. Toward an Integrated Arctic Observing Network. Washington, DC: The National Academies Press. doi: 10.17226/11607.
Page 81
Suggested Citation:"5 Designing the Network." National Research Council. 2006. Toward an Integrated Arctic Observing Network. Washington, DC: The National Academies Press. doi: 10.17226/11607.
Page 82
Suggested Citation:"5 Designing the Network." National Research Council. 2006. Toward an Integrated Arctic Observing Network. Washington, DC: The National Academies Press. doi: 10.17226/11607.
Page 83
Suggested Citation:"5 Designing the Network." National Research Council. 2006. Toward an Integrated Arctic Observing Network. Washington, DC: The National Academies Press. doi: 10.17226/11607.
Page 84
Suggested Citation:"5 Designing the Network." National Research Council. 2006. Toward an Integrated Arctic Observing Network. Washington, DC: The National Academies Press. doi: 10.17226/11607.
Page 85
Suggested Citation:"5 Designing the Network." National Research Council. 2006. Toward an Integrated Arctic Observing Network. Washington, DC: The National Academies Press. doi: 10.17226/11607.
Page 86
Suggested Citation:"5 Designing the Network." National Research Council. 2006. Toward an Integrated Arctic Observing Network. Washington, DC: The National Academies Press. doi: 10.17226/11607.
Page 87
Next: 6 Detailed Implementation Ideas »
Toward an Integrated Arctic Observing Network Get This Book
Buy Paperback | $61.00 Buy Ebook | $48.99
MyNAP members save 10% online.
Login or Register to save!
Download Free PDF

Observable changes with regional and global implications, such as warming temperatures and reduced sea ice, are taking place across the Arctic. However, the record of Arctic observations suffers from incomplete geographic coverage and limited duration, and measurements are not well coordinated. This makes it difficult to comprehensively describe current conditions in the Arctic, let alone understand the changes that are underway or their connections to the rest of the Earth system.

The U.S. National Science Foundation asked for guidance to help design a pan-arctic observing network. This book outlines the potential scope, composition, and implementation strategy for an arctic observing network. Such an integrated, complete, and multidisciplinary environmental observing network will improve society's understanding of and ability to respond to ongoing systemic changes in the Arctic and its capability to anticipate, predict, and respond to future change both in the Arctic and around the globe. The network would build on and enhance existing national and international efforts and deliver easily accessible, complete, reliable, timely, long-term, pan-arctic observations. Because many potential components of the network already exist or are being planned, and because of the surge of activity during the International Polar Year, there is an immediate opportunity for major progress.

  1. ×

    Welcome to OpenBook!

    You're looking at OpenBook,'s online reading room since 1999. Based on feedback from you, our users, we've made some improvements that make it easier than ever to read thousands of publications on our website.

    Do you want to take a quick tour of the OpenBook's features?

    No Thanks Take a Tour »
  2. ×

    Show this book's table of contents, where you can jump to any chapter by name.

    « Back Next »
  3. ×

    ...or use these buttons to go back to the previous chapter or skip to the next one.

    « Back Next »
  4. ×

    Jump up to the previous page or down to the next one. Also, you can type in a page number and press Enter to go directly to that page in the book.

    « Back Next »
  5. ×

    Switch between the Original Pages, where you can read the report as it appeared in print, and Text Pages for the web version, where you can highlight and search the text.

    « Back Next »
  6. ×

    To search the entire text of this book, type in your search term here and press Enter.

    « Back Next »
  7. ×

    Share a link to this book page on your preferred social network or via email.

    « Back Next »
  8. ×

    View our suggested citation for this chapter.

    « Back Next »
  9. ×

    Ready to take your reading offline? Click here to buy this book in print or download it as a free PDF, if available.

    « Back Next »
Stay Connected!