wiki:ProposedChanges

Proposed changes to the current version

A key element of the proposed changes is the use of the concept of observation (ISO 19156) - MOLES MO_Observation class is a specialisation of the OM_Observation class (ISO 19156).

In particular, proposed changes and discussion topics in respect with the current version MOLES v3.2 are:

1. Association of MO_Project with MO_Observation

MOLES v3.2 makes use of the Union Class MO_Deployment (either MO_Acquisition or MO_Computation) . This class is associated with the MO_Project having the role activity. An instance of MO_project can be associated with zero to any number of instances of MO_Deployment. Instances of MO_Acquisition and MO_Computation are associated with the MO_Project which has the role owner with respect to them.

The proposed version, on the other hand, does not make use of the Union Class MO_Deployment at all. As discussed at the Basic Formalism Section observations are actions of a project in order to obtain estimates of a property values of an identifiable object. Therefore, the MO_Project class is associated with the class MO_Observation which has the role activity. The MO_project has the role owner for the Observations which are under its auspices.

2. MO_Process Class

In the current version the MO_Process class ( at the bottom of the diagram) is a union (using either an instrument or software).

Following the discussion at the Basic Formalism Section it seems more suitable for MO_Process class to be a composition; the classes MO_Acquisition and MO_Computation are its containees ( See diagram). The MO_Process can contain zero to any number of MO_Acquisition or MO_Computation instances.

As depicted in the diagram, MO_Process is a specialization of OM_Process (ISO 19156) class. MO_Process , as OM_Process, is abstract, and has no attributes, operations or associations. Therefore, it seems that MO_Process can be replaced by the OM_Process class.

3. Discussion topic: Can the class GFI_PropertyType and its association with OM_Observation Phenomenon ( see diagram) model observations where the observed property can be considered as a vector ( with any number of elements)?

EXAMPLE 1: wind velocity u=(Ux, Uy,Uz)
EXAMPLE 2: Many Numerical models can simulate a quite large number of parameters in one go, therefore it seems reasonable to take: property0=(property1, property2 , …, propertyN) [e.g property0 =(temperature, pressure, humidity, wind velocity)] rather than having N different observations using the same procedure , at the same time (parallel observation) at property1,property2,…, propertyN.

Quotation 1:From ISO 19123:2007 , section 5.1.3
“The range of a coverage is a set of feature attribute values. This value set is represented as a collection of records with a common schema. EXAMPLE: A coverage might assign to each direct position in a county the temperature, pressure, humidity, and wind velocity at noon, today, at that point. The coverage maps every direct position in the county to a record of four fields.”

Quotation 2:From ISO 19156 ( document for review and comment)
“A property-type may be an operation or function such as a spatio-temporal coverage.”

From the above quotations the answer to the addressed question seems to be Yes, but I couldn’t manage to materialise a model from ISO 19123 and 19156.

According to Quotation 2 the property type can be a function ; for the EXAMPLE 2 its domain and range could be:
domain : values of property1,property2,…, propertyN
range: values of property0.

However this is not always correct as property0 may not exist in the nature (EXAMPLE 2). We define it. More accurately instead of property0=(property1, property2 , …, propertyN) we write property0 := (property1, property2 , …, propertyN). In orther words property0 is not a function but an aggregation of property1,property2,…, propertyN .

This leads to the model. Go here) for supporting examples.

4. Location/time attributes of MO_Process - Monitoring Facility/Platform?

....coming soon....

5. Handling Laboratory Specimens

See SpecimenHandling!

 Trac Powered
Site hosted at the
British Atmospheric Data Centre
for the