Variables

The proposed variables (Table 2) have been selected to be:

• the minimum set needed to observe, understand and predict changes in critical life supporting processes;

• integrated internally consistent and mutually supporting and enhancing;

• globally applicable and standardisable as possible;

• linked to related variables in higher and lower tiers via models, permitting interpolation in space and time to cover gaps in the system.

Models will play an essential role in filling the gaps in space and time unavoidably left by an incomplete sampling system. In general, tiers 1 and 2 generate models, which are run, validated and interpolated using data from tiers 3, 4, and 5.

Organizations joining the GCOS/GTOS will choose the tier at which they should operate. They must collect all the variables appropriate to the system which they study within, to allow integrated data sets to be developed. In a few cases this is not efficient, and a specialized agency may be responsible globally for certain variables within one or more tiers. For example, sulphur aerosols emitted from volcanoes are important for the global radiation balance, but the emissions are localised and intermittent, and therefore best measured by volcanologists.

The variables are specified in terms of definition, units of measurement, spatial resolution, temporal frequency and required accuracy. In most cases, the methods to be used to obtain them are not prescribed, although harmonisation of methods will form part of the programme. This is because the surface of the world is so heterogeneous that a single method is seldom globally applicable. However it is important to assure that methods are comparable and compatible.

The area of the global surface sampled at tiers 1 and 2 is very small, and cannot be statistically unbiased. The approach is for tier 1 to span major global environmental gradients, and for tier 2 to have at least one centre in each major crop, terrestrial biome, freshwater system and cryosphere system. The number of locations at tier 3 is approaching a statistically useful sample within broad types, but because it is to be made up largely of existing facilities, does not represent an unbiased sample. It does define the range of variation within a type. Tier 4 provides a statistically unbiased sample. The sample design for this tier is left to individual participating nations, and may be systematic or stratified. Tier 5, where the coverage is complete, provides a globally consistent basis for stratification.

The temporal frequency of sampling is defined by the rate of change of the variable. Tier 2 centres are equipped with automatic data loggers, which can record very rapid events, such as the eddies which carry trace gases between the earth's surface and the atmosphere. Tier 3 has permanent staff, who can record events on a daily to monthly basis, such as snow depth, lake level or leaf cover. Tier 4 is visited once a year to once a decade, and is used for observations of variables such as soil organic carbon content, ice sheet thickness, and depth to the water table. The required accuracy is determined by what is needed for useful interpretation, and what is possible with affordable current technology.

Proposed Climate-related Variables

Table 2. Recommended Variables.