Tracking data sources and processing done to them is becoming increasing important as scientists seek to define trends and unexpected changes in the environment. Keeping track of data transformations and processing, generally termed lineage, is an important role of high-quality metadata. The ISO metadata standard provides a simple lineage model based on sources which are either used or produced in a series of process steps. This model can be helpful in many cases despite its simplicity. Sources and process steps are linked together to describe the lineage of a resource.
This Figure shows an overview of the ISO lineage model which links sources to process steps. Sources can be input to, or output (19115-2) from a process step. Each process step has associated processing and algorithm information (also added in 19115-2). These improvements make it important to use 19115-2 if you need good lineage descriptions. ISO 19115-2 Lineage UML. This Figure shows more detail in the UML model used by the ISO Standard to describe lineage. In some cases, a simple descriptive statement can describe the lineage effectively. In more complex cases, multiple sources and process steps might be required. The definitions of sources and processSteps are also shown in the UML. The capability to specify the spatial and temporal extent of the source and to describe the rationale for a process step are new in the ISO Standard. Note that each source can have any number of associated sourceSteps and that each processStep can have any number of sources (and outputs in ISO 19115-2).
Sources and Steps
The original ISO 19115 Source descriptions (LI_Source) were extended in 19115-2 to include several more elements. The LE_Source includes the following elements:
LE_Source + description[0..1]: CharacterString + scaleDenominator[0..1]: MD_RepresentativeFraction + sourceReferenceSystem[0..1]: MD_ReferenceSystem + sourceCitation[0..1]: CI_Citation + sourceExtent[0..*]: EX_Extent + processedLevel[0..1]: MD_Identifier + resolution[0..1]: LE_NominalResolution
and Process Steps include
LE_ProcessStep + description: CharacterString + rationale[0..1]: CharacterString + dateTime[0..1]: DateTime + processor[0..*] : CI_ResponsibleParty + processingInformation[0..*]: LE_Processing + report[0..*]: LE_ProcessStepReport LE_Processing + identifier: MD_Identifier + softwareReference[0..*]: CI_Citation + procedureDescription[0..1]: CharacterString + documentation[0..*]: CI_Citation + runTimeParameters[0..1]: CharacterString + algorithm[0..*]: LE_Algorithm LE_Algorithm + citation: CI_Citation + description: CharacterString LE_ProcessStepReport + name: CharacterString + description[0..1]: CharacterString + fileType[0..1]: CharacterString
The ISO Lineage model is simple but is probably sufficient for many common processing scenarios. It may only provide summary information in complex processing scenarios. This is facilitated by the use of CI_Citations in LE_Sources, LE_Processing, and LE_Algorithm. The resources referenced by these citations can provide more detail when necessary.