- Galdos Systems Inc. - https://www.galdosinc.com -

EarthCube – and a Global Parameter Registry

The U.S. National Science Foundation (NSF) has launched an ambitious program to unify data, models, and supporting software across all Geosciences, including atmospheric science, geomorphology and the solid earth sciences, polar science, and oceanography. This is to be achieved by the creation of a geosciences cyber-infrastructure. It is intended as a transformative initiative for the practice of geoscience in the United States and beyond (see http://earthcube.org/ [1] for further details).

Like other large data integration efforts, a key issue will be to develop common semantic models that can be shared across all of the geoscience domains. Such common semantic models are far more important than simply having catalogues of data, or focusing on syntactical data interchange mechanisms. The latter are, of course, required, but the real benefit is obtained with support for common semantics.

One possible component of the EarthCube cyber-infrastructure would be a global registry of physical parameters. Such parameters are used in measurements and theoretical models of all types and in all sciences, and include things like the rotation rate of the earth, shock wave blast radius, air density, viscosity of sea water, elevation, depth, pore-fluid pressure, and many, many more. Many of these parameters will carry units of measure which can, in turn, be expressed in terms of their fundamental quantities, for example, air density [ML-3]. Some, such as Reynolds, Eckman , Rosby, and Froude numbers, are non-dimensional. Still other parameters take on ordinal values (e.g. Rock hardness) or values from a list or taxonomy.

In such a global parameter registry, it will likely be important to distinguish observables, some of which may be measured quantities, from parameters that appear in theoretical models. The appropriate semantics of these global parameters needs to be elaborated; however, the work being done in the OGC (Geography Markup Language, Observations and Measurements, SensorML) and the ISO (ISO 19156) is a good start, together with predecessor documents like the Handbook of Mathematical Psychology in which Luce lays down a very clear theory of measurement.

Supposing such a global parameter registry existed – how would it be created, maintained, and used?

Clearly we would require some form of governance structure — some form of body, perhaps set up through the NSF or the ISO — that admitted or updated parameters in the registry, and modified their life cycle status. A global parameter just submitted to the registry might have the status “submitted” (meaning that it has been recorded in the registry but has not yet been looked at by an authority) and, upon review by the appropriate committee, its status might be changed to “approved”.

We would assume that each such parameter in the registry would be assigned a globally unique identifier, so that each parameter might have multiple names and symbols. In addition to the identifier, the registry would maintain one or more attributes such as name/alias, a text definition and description, possibly graphics or images, and other information sufficient to clearly identify the parameter. The registry would also contain associated symbols (classified by system of units, application domain, etc.) as well as maintaining relationships between the parameters.

It is assumed that there were would be many thousands, or perhaps tens of thousands, of parameters in the registry. To use the registry, creators of geoscience models, measurement programs, and processing software would map each of the parameters/variables/attributes in these items to the appropriate global parameter by referencing the parameter registry and the appropriate parameter’s globally unique identifier. In some cases, they may even use the parameter ID directly in their model, measurement program, or software.

Having a global parameter registry, based on a common semantics, would be a big step toward unification of the Geosciences.