OR/14/013 Model datasets and workflow
|Monaghan A A, Arkley S L B, Whitbread K, McCormac M. 2014. Clyde superficial deposits and bedrock models released to the ASK Network 2014: a guide for users Version 3 . British Geological Survey Internal Report, OR/14/013.|
General caveats regarding BGS datasets and interpretations can be described:
- Geological observations and interpretations are made according to the prevailing understanding of the subject at the time. The quality of such observations and interpretations may be affected by the availability of new data, by subsequent advances in knowledge, improved methods of interpretation, improved databases and modelling software, and better access to sampling locations.
- Raw data may have been transcribed from analogue to digital format, or may have been acquired by means of automated measuring techniques. Although such processes are subjected to quality control to ensure reliability where possible, some raw data may have been processed without human intervention and may in consequence contain undetected errors.
The Superficial Deposits models were constructed in GSI3D using a NEXTMap® Digital Elevation Model ©Intermap Technologies, at 50 m resolution (Central Glasgow, North Glasgow, South Glasgow) or 25 m resolution (Paisley, Clydebank), the BGS digital borehole database, BGS 1:10 000 scale digital maps (DiGMapGB, 2009), BGS 1:50 000 maps (DiGMapGB, 2008), field slip scans, historic maps and scanned geological cross-sections. Other literature such as BGS regional geological guides and scientific papers influenced the correlation of geological units.
Borehole data were entered to the BGS corporate database BGS Borehole Geology. The spread of borehole data across the area was variable, from extremely closely spaced at site investigation locations to more widely spaced and isolated boreholes.
A standard GSI3D workflow for superficial geological models was followed (Kessler et al., 2008 http://nora.nerc.ac.uk/3737/1/OR08001.pdf) for the original component block models. The method principally involves construction of cross-sections between the best quality borehole data followed by envelope construction around the limits of the geological units. GSI3D model calculation then uses envelopes in combination with nodes on the geological surfaces along cross-sections to build geological surfaces by triangulation.
The original component block models and individual map sheets within the block models were modelled over a number of years, by a number of geologists and during a period of software development. As such, a number of steps were taken before release of these v2 (or v3 Central Glasgow) models to attempt to resolve inconsistencies between block models. Further details are documented in Whitbread (2013). The key steps included:
- Inclusion/exclusion of water bodies to a common standard
- Creation of a set of merged borehole data for each component block model
- Creation of a set of master envelopes ‘extents’ covering the modelled area
- Inclusion of higher resolution corridor model sections and envelope edits on SW Central Glasgow and Paisley (Monaghan and Whitbread, 2012)
- Revision to Lexicon and lithology codes describing the modelled units to current best practice and update GVS, legend and GSIPR files
- Using extended DTM’s to cover the edge matching zone for each model
- Edge matching the block models such that the geological interpretation was consistent across block boundaries. The main issues to be resolved were
- Marine-influenced Gourock Sand Member (GOSA) has been modelled flanking the Clyde in Clydebank and Central Glasgow, but the equivalent unit is modelled as fluvial Law Sand and Gravel Member (LAW) to the south-east of Glasgow, upstream of the tidal limit. This was resolved by taking the GOSA as far as the Uddingston-Bothwell area and then using the LAW upstream of this.
- The Paisley Clay Member was not modelled under Linwood Clay Member in the north-west quarter sheet of Paisley Model meaning that the Paisley Clay in the Clydebank Model ended abruptly at the model edge. Conversely the Linwood Clay was not modelled in the north-east of the Paisley Model. This was resolved by undertaking a revised interpretation of both units on both block models, such that they are now consistent.
- Edge matching issues in envelopes (XY) or sections (Z) viewed as linear jumps. These were resolved by revised interpretations, guided by borehole data in the problem areas.
The following inconsistencies were not attempted to be resolved at this stage:
- Insertion of additional cross-sections to achieve a more uniform cross-section coverage
- Revision of made ground (particularly on NS56NE) where some boreholes prove made ground that is not modelled.
The Central Glasgow and Clyde Catchment scale bedrock models were constructed from all borehole data entered to the BGS corporate database BGS_Borehole_Geology and reaching the modelled horizons, and mine plan, map outcrop and interpreted data.
The spread of borehole data across the area was very variable, from closely-spaced site investigations metres apart, to in extreme cases, boreholes more than a kilometre apart. Data points were concentrated around the outcrop of worked coals and were sparse on stratigraphic surfaces in deeper parts of the basin.
Most borehole data points have a reasonably good level of certainty. Boreholes with very poor quality records or very poorly known sites were not coded into the database. However, there can be uncertainty: in geologically coding short isolated site investigation boreholes; in a drillers record of a borehole (i.e. if not examined by a geologist); or sometimes in the siting of the borehole. However, these should result in errors in location being no greater than about 5–10 m in Z, and perhaps 20–50 m in terms of XY.
Mining data were compiled from all available mine abandonment plans. These consist of spot heights surveyed on the base of worked coal seams underground, and rarely of structure contour elevation data. The distribution of mining data points is variable. These data points have a high confidence level as they were systematically surveyed. Estimates of error on mining data points range from 0–5 m in Z and 0–25 m in XY.
The bedrock map represents the outcrop (or subcrop) of stratigraphic horizons at rockhead i.e. very commonly buried beneath superficial deposits. For the Central Glasgow bedrock model, the map data for NS66SW had been revised prior to this study and an updated map published (BGS, 2008). The map data for NS56SE and NS66NW were revised prior to the study but those revisions have not been published. That is the model contains more up-to-date map linework than does the current edition of the 1:10 000-scale maps (BGS, 1995, 1996). The errors in mapped outcrop line work are extremely variable — from 0–10 m in XYZ where seen at outcrop, to tens of metres where an interpretive outcrop was created from little constraining data.
For the Clyde Catchment scale bedrock model a variety of map scales have been used to constrain the modelled extent. 1:10 000 scale data was used for the Glasgow Ell Coal seam whereas 1:50 000 and smaller scale data has been used for the stratigraphic horizons (McCormac, 2013). For all modelled horizons the outcrop extent has been simplified to be appropriate for the scale and resolution of the regional scale of the model.
A standard BGS GOCAD® modelling workflow using the structural workflow was employed. GOCAD® (Paradigm, www.pdgm.com) calculates a triangulated mesh based on XYZ data points and then the geologist modeller undertakes various processes to aid geological interpretation in data poor areas. Expert geological interpretation was added during modelling as cross-sections, isopach maps, editing of fault-surface contacts and removal of overlaps during GOCAD® modelling.
- BRITISH GEOLOGICAL SURVEY [DIGMAPGB]. 2009. Digital Geological Map of Great Britain 1:10 000 scale (DiGMapGB-10) data. Version 2.18. Keyworth, Nottingham: British Geological Survey. Release date 15-01-2009.
- BRITISH GEOLOGICAL SURVEY. 2008. Digital Geological Map of Great Britain 1:50 000 scale (DiGMapGB-50) data. Version 5.18. Keyworth, Nottingham: British Geological Survey. Release date 20-05-2008.
- KESSLER, H, MATHERS, S J, SOBISCH, H-G and NEBER, A. 2008. GSI3D — The software and methodology to build systematic near-surface 3D geological models. (Version 2) British Geological Survey Open Report, OR/08/001 144pp.
- <WHITBREAD, K. 2012. A Methodology for Compiling the Clyde Superficial Deposits Models: recommendations for future multi-phase modelling programmes. British Geological Survey Internal Report, IR/13/030. 35pp
- MONAGHAN A A and WHITBREAD K. 2012. Shieldhall High Resolution 3D Model. British Geological Survey Commercial in Confidence Report, CR/12/076. 35pp
- BGS 2008 1:10 000 bedrock map of NS66SW
- BGS 1995 1:10 000 bedrock map of NS56SE
- BGS 1996 1:10 000 bedrock map of NS66NW
- MCCORMAC, M. 2013. Clyde Catchment GOCAD® bedrock regional model, 2013. British Geological Survey Internal Report, IR/13/014. 11pp