key indicator variable. A key variable is a variable that is fundamental and related to questions that are important throughout the Arctic, is essential to an overall understanding of the arctic system, and is also relevant at the local scale. A key variable is a necessary component of integrated monitoring because changes in associated variables cannot be understood without knowledge of changes in the key variable. The key variable is a disaggregated driver of lower-level changes.3 An example is temperature.4

A key indicator variable is a response variable or index that can be conveniently measured to denote changes in one or more key variables. Examples are phenology (the timing of events) and indices of human activity. For example, the proportion of whales harvested in open water by an arctic community during autumn rather than spring can be an indicator of climate-driven changes in sea ice cover. Key indicator variables may manifest themselves differently at different locations.

In the framework to be used here, a key variable is a key variable whether its variations are obtained from direct measurements or by proxy methods. A proxy measurement of a key variable is an indirect estimation of the key variable, usually consisting of a measurement preserved in some physical manifestation. An example is tree ring width or density in fossil and modern wood samples: from the patterns of ring widths, one can infer variability in environmental variables such as, for example, temperature. The value of proxies to monitoring networks is that they provide historical context for the contemporary measurements and instrument records. They also improve data products that arise from reanalysis efforts. Beyond tree rings, other examples of proxies include ice and lake sediment cores and local and traditional knowledge of environmental history. Proxy records often share a common threat of being lost (perhaps through melting of an ice cap in which a climate record is preserved or through death of a village elder). Data in old formats are similarly at risk of being lost.


Variables can be grouped in many ways. One approach is to organize variables according to underlying concepts. These concepts define the basic state of the physical, biological, chemical, and human environment, and identify and characterize natural variability and anthropogenic change. These concepts represent the most basic approach to system understanding and to the identification of the causes of change (e.g., physical-chemical drivers, human and biological drivers, environmental impacts, human and biological responses). Integrated monitoring is a particularly effective approach for recording these variables, and field manipulation experiments can be performed to obtain relationships among them. Such an approach is particularly useful for testing process formulations in models.

This conceptual approach is the most consistent with the AON vision statement because this way of thinking provides an umbrella for the individual questions, hypotheses, and themes, and it facilitates a pan-arctic focus and long-term perspective. In addition, this approach is inclusive in its philosophy, cross-disciplinary, and represents a stable base that can evolve as key questions, access to the Arctic, technology, and human needs change. Although there are other ways of categorizing variables, such as organizing by problems or research theme, organizing by concept has the distinct advantage that the concepts are less likely to change over time than questions or themes. Thus, the Committee prefers an organization that is founded on state variables but presents other options to encourage discussion. These other options entail grouping by practicality and approach, grouping by synergies (i.e., creating benefits that only arise from a suite of measurements or combination of measurements and existing data), and grouping by timescale.

Grouping by What Is Practical

Groupings can be based on discipline, theme, and measurement approach. For example, variables can be classified by major disciplines: atmosphere, ocean, cryosphere, terrestrial. Themes that can provide the basis for classifications include biodiversity, land cover, and sustainable resource use. In terms of measurement approach, variables can be organized by platform (satellite, observatory, human observers, buoys, etc.), thereby maximizing the use of existing expertise and the cost-effectiveness of infrastructure. In addition, classifications can be responsive to initiatives that define their own information needs. Examples are the International Polar Year and the Arctic Climate Impact Assessment. These initiatives can provide structure and permanence to data. Finally, variables can be classified by stakeholder priorities such as the priorities of arctic communities.

Grouping by Synergies

A classification scheme can be based on one of a number of types of synergy. For example, synergies among existing and new monitoring activities can lead logically to groupings of variables. Synergies can also arise from better use of existing data (e.g., satellite images, photographs, biological samples that have untapped potential). For example, precipitation data collected by weather stations and then thrown


To be clear, this definition does not preclude a key variable from changing.


Consider the changing migration pattern of a caribou herd. The fundamental driver (key variable) could be temperature, which might be linked to changes in migration pattern because of its effect on vegetation phenology and composition, snow cover, and consequently the availability of food. These changes are lower-level responses that are driven by changes in the key variable temperature. Lower-level changes such as migration pattern are then indicators of changes in the key variable temperature.

The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement