Your browser is no longer supported

For the best possible experience using our website we recommend you upgrade to a newer version or another browser.

Your browser appears to have cookies disabled. For the best experience of this website, please enable cookies in your browser

We'll assume we have your consent to use cookies, for example so you won't need to log in each time you visit our site.
Learn more

Glossop Lecture: From concept towards reality: Developing the attributed 3D geological model of the shallow subsurface.

Professor Martin Culshaw began by referring to Sir John Knill's 2002 1st Hans Cloos Lecture in which he set out to define engineering geology's core values.

Knill suggested the 'standard setting' phase is nearing completion and that what is now needed is analysis and synthesis of existing engineering geological information.

He said standards and classification systems have been developed and there now exists a large volume of data that comply with those standards and classification systems.

Developments in information technology hardware and software now make the storage, management, validation, analysis and synthesis of large quantities of engineering geological data possible. The key role in this process of geological surveys around the world in designing, populating and managing national databases was highlighted.

Culshaw also referred to the 1st Glossop Lecture by Peter Fookes (1997) on the 'Conceptual ground model' as an example of the classic approach to engineering geology.

He questioned whether the volume of available engineering geological data is now sufficient to consider moving from conceptual ground models to real ground models. This was to be the key question of the evening.

The intricacies of developing an engineering geological model were discussed in relation to the engineering geological triangle.

To build the model there is a need to identify and locate significant engineering geological unit boundaries in three dimensions and create a spatial volume model of the subsurface.

The model then needs to be attributed with information on the material and mass properties of the engineering geological units.

While the ability to do this exists and is used regularly in other applications, further work is required to model the effect of geological processes, Culshaw said.

Data acquisition and management issues were presented as the principle obstacles to the creation of 3D subsurface models. What also hinders the process is current site investigation practice, as defined in BS 5930, which focuses on the acquisition of new data through physical investigation.

If 3D subsurface models based on existing data are to be created and used in everyday applications there must be a change in the way information is obtained and managed. The availability of site investigation data is critical to the development of the models and to any subsequent improvements to them.

Culshaw felt that there is, therefore, a need for originators of geological data to make that data available, accompanied by sufficient information to put it into context, to national data holders (Talking point, GE January 05).

He put forward a good case for geological surveys, as publiclyowned, non-commercial bodies, to be seen as the ideal holders of national geological information.

A number of issues have restricted this function in the past, including a lack of any legislation concerning the national archiving of site investigation data.

Despite these obstacles, there now exists a sufficient density of coverage of site investigation data in many urban areas to allow the production of attributed 3D models of the shallow subsurface.

Creation of a 3D digital map requires three main data types;

surface geological mapping, reliable borehole logs and a digital terrain model. The digital terrain model provides the spatially defined surface from which the subsurface geological information will be hung.

The first step is to create a generalised vertical section;

effectively defining the engineering geological units and boundaries. Cross-sections are created to build up a fence diagram - a 3D representation of a network of interlinked cross sections. Where possible, the sections should be regularly spaced, intersect valleys and structures at right angles, have added minor sections to deal with local variations and anomalies, and deal with linear bodies.

The desired scale of the model will determine the number of boreholes required and the spacing of the sections. The top surface of each geological unit is created from the surface geological map and the fence diagram. The surfaces are then spatially combined to produce a 3D model 'stack'.

Once created, the model can generate synthetic borehole logs, synthetic sections (vertical sections and horizontal slice sections), contoured surfaces, isopachytes of single or combined surfaces, domain maps, and sub- and supra-crop maps. When these outputs are combined with other variables the list of potential applications is endless.

The next stage in building the 3D geological model is to attribute geotechnical and other point data to the geological volumes. Data for geotechnical parameters invariably show a range of results for any given geological unit. For attribution of parameters, ways to summarise the geotechnical properties or to visualise the variation need to be found.

It was suggested that for summarised geotechnical data, statistics appropriate to the data set are used. Generally, for geotechnical parameters this means non-parametric rather than parametric statistics.

However, both assume that all the data values are 'good' and equally reliable which is rarely the case. Robust statistics attempts to allow for the fact that most real data sets do contain a proportion of poor or bad data values and recognises that virtually all data sets do have some underlying structure.

Visualisation of geotechnical property variability is a new area of research and therefore the reliability of the model output needs to be tested. The variability of SPT N values for glacial till in the Firth of Forth was used as an example and shown as a video slowly moving down sequence.

The model output gave an immediate sense of the spatial variability which could not be obtained from graphical plots.

In terms of modelling geological processes, good progress has been made in producing hazard databases.

There is a National Landslide Database, Natural Cavities Database and a National Karst Database.

The British Geological Survey has also undertaken geohazard susceptibility mapping, initially as a package for insurers to determine risk by postcode.

Every mapped geological formation was assessed in terms of its susceptibility to a range of geohazards and attributed a geohazard rating. Subsequently, national geohazard susceptibility maps have been produced.

Culshaw reiterated that industry should anticipate major changes to site investigation practices as the transition from the 2D interpretative map to the interactive, digital attributed 3D model is made.

He suggested that in future, the first objective of a ground investigation will be to test the 3D engineering geological model created in the early stages of site investigation, the uncertainty associated with it and the risks resulting from that uncertainty.

For this vision to be taken forward there is clearly a need for greater cooperation between the geological data custodians and the data originators, and for a general willingness within the industry to share data.

Have your say

You must sign in to make a comment

Please remember that the submission of any material is governed by our Terms and Conditions and by submitting material you confirm your agreement to these Terms and Conditions. Please note comments made online may also be published in the print edition of New Civil Engineer. Links may be included in your comments but HTML is not permitted.