Future Work

Plans for Future Work

"Paths are made by walking" - Franz Kafka

GeoSciML Update

The O&M based GeoSciML was last updated to V4.1 in 2017. This standard should be revisited in light of the developments in both O&M (updated to OMS) as well as the use of SensorThings API.

GroundWaterML2 Update

While GroundWaterML (OGC WaterML 2: Part 4 - GroundWaterML 2 (GWML2)) was last updated in 2019, this was before the introduction of OMS. Revisiting this standard in order to integrate the new concepts from OMS would be valuable.

JSON-LD Context for properties

Adding JSON-LD features to the SensorThings API has been a feature request for a long time, but making the entire API compatible with JSON-LD is not likely to happen very soon. A possible intermediate approach may be to add a JSON-LD context to the properties of each entity type. This won't make the entire response JSON-LD compatible, but may help with interpreting the free-form properties objects available in the SensorThings API.

Thing vs. FeatureOfInterest

In this edition of the GeoTech IE (assuming there will be more) borehole collars and trajectories are mapped to Things, while all samples are mapped to FeatureOfInterest. Although this split works, it is not ideal, and thus in the WaterQuality IE an experiment is being done with mapping all domain features into the FeatureOfInterest entity type, together with a new FeatureType entity type. This moves slightly further away from the STA core model, but results in a cleaner model and is currently considered as a basis for the SensorThings API version 2.0.

The Sampling, Sampler, SamplingProcedure, PreparationStep and PreparationProcedure entity types from this GeoTechIE have been taken up and improved by the WaterQualityIE, and are very likely to also be adopted as standard extensions to the SensorThings API v2.0.

The many extra self-relations that were added to many entities (RelatedBhCollarThings, RelatedDatastreams, RelatedObservations and RelatedFeatures) have also been improved upon in the WaterQualityIE and will most likely also be added as a standard extension to version 2 of the SensorThings API.

Geophysics, Profiles and 2D sections, 3D subsurface models

This work was mainly semantic focused with the principal intention to harmonize geotech concepts. A next focus should be made on the provision of geometry, including notably the capacity to share and query cross sections and 3D models. This topic, as well as Geophysics are in discussion as possible new activities of the OGC GeoScienceDWG.

Semantic transposition

OMS allows for provision of explicit information on observational or measurement data, providing detailed information on how these observations were created. Such additional measurement metadata is essential in order to determine if a specific dataset is fit for purpose within a specific use case. However, the complexity of the underlying data models can make interaction with OMS based systems difficult for inexperienced users, who can easily get lost in the many options. Based on these reflections, we came to the insight that within individual use cases, there are often two very different user groups interacting with OMS data:

• Thematic experts analyzing the data for usefulness within the UC. These experts require access to the full complexity of the underlying OMS data model to understand if the data being provided is truly fit for purpose.
• Data analysts or developers utilizing an already vetted dataset no longer need the details provide by the OMS model, for their purposes all they need is a geometry and some numbers, as shown in the example below.

{
    "type": "Feature",
    "geometry": {
        "type": "Point",
        "coordinates": [
            125.6,
            10.1
        ]
    },
    "properties": {
        "O3": {
            "value": 24.3,
            "uom": "mg/L"
        },
        "NO2": {
            "value": 16.8,
            "uom": "ppb"
        }
    }
}

The challenge in such a transformation pertains to how to transfer the content of the ObservableProperty (name="O3" or name="NO2") to the data structure of the simplified view, as we're transferring information from data content to structure. For this purpose, we've coined the term "Semantic Transposition", defined in View points on data points: A shared vocabulary for cross-domain conversations on data and metadata

Within the current work on updating SensorThings API to V2.0, integration of Semantic Transposition will be investigated.

Simple import options for complex models

In order to enable simple import to the complex models defined utilizing OMS and STA, once the structure for standard tests has been defined, simplified tools enabling data providers to easily provide their data must be investigated. Simple wizards could be of great help in enabling data provision.

Tabular options for complex models

Related to the points above on Semantic Transposition and simplified import, a further approach to simplified interaction with STA endpoints pertains to tabular views. This approach can be useful both in provided data as well as accessing data.

• On data provision, work is ongoing at the European Environment Agency (EEA) on how to enable provision of OMS data via tabular views, especially in cases where most of the observational metadata has default values, as common under reporting obligations (e.g. the UoM and measurement methodology are predefined.
• On data access, once the observational metadata has been checked by the relevant experts, a simple tabular download format would be valuable for further analysis as well as portrayal.

Visualization options for results

Within this IE, the focus was on utilizing the OMS model as well as the STA implementation for sharing of borehole data. A next step involved implementing various visualization options for better access to the retrieved data.