OR/19/003 Methods

From MediaWiki
Jump to navigation Jump to search
Kearsey, T, Lee, J R, and Gow, H. 2019. Buried valleys (onshore) - Version 1: scientific report and methodology. British Geological Survey Internal Report, OR/19/003.

The Buried Valleys (onshore) dataset contains two elements:

  1. Historic references to Buried Valleys
  2. Modelled Thickness of Buried Valleys

These were created using the methods described below.

Historic references to buried valleys

The BGS historic literature was searched using dtSearch the BGS text searching facility. This returned 120 BGS publications plus 25 technical reports and 8 open reports that contained the phrases ‘buried valley’, ‘tunnel valley’, ‘sub-drift topography’ and ‘drift filled channel’. Of these 96 had figures or maps that had sufficient geographic information that enabled them to be georectified. Buried valleys that are shown on the marginalia of published geological maps and within BGS digital mapping data (GB Geology 1:50 0000) were also captured within the dataset. The source and type of each feature are recorded within the table of attributes of the shapefiles.

The geographic location of buried valleys from historic sources were used to create two types of data (lines and polygons), with features identified either using a centre line of the feature (for original features drawn as linear features) or the margins of the feature (for original features drawn as polygons).

File:OR19003fig4.jpg
Figure 4    An example of historical buried valleys dataset showing centre lines and margins. Note the margins and centreline data in this example is drawn from different sources as a result there is difference in the precise location of the features. Contains Ordnance Data © Crown Copyright and database rights 2019. Ordnance Survey Licence no. 100021290 and NEXTMap Britain elevation data from Intermap Technologies. Created using ArcGIS. Copyright © Esri.

Modelled thickness of buried valleys

The BGS’s digital borehole database postdates most of the historical references to buried valleys. Therefore, they may be able to identify, and resolve more buried valleys, when compared to those discovered through traditional geological mapping activities. Some of these features are visible in the Superficial Deposits Thickness Model. However, it has been noted that the irregular distribution of boreholes and the gridding algorithm used in the interpolation can create apparent gaps, especially in buried valley features. Kearsey et al. (2018)[1] identified that a discrete-smoothing interpolation (DSI) was better at reducing the amount of apparent gaps caused by the algorithm within buried valleys. This method was used to create a modelled thickness of buried valleys based on current onshore BGS borehole dataset. To create this dataset two different inputs were used.

  • The 2018 borehole dataset which proves the depth to geological rockhead which is currently being prepared for the new SDTM. Geological rockhead is an inferred surface that separates bedrock from superficial deposits. For the purpose of this dataset, superficial deposits encompass all naturally-occurring deposits that are of Neogene or Quaternary age. This includes deposits of the Plio-Pleistocene Crag Group which historically have been classified on older geological maps as bedrock.
  • Rock at surface from GB Geology-50 superficial geology map version 8.

These datasets were imported into SKUA-GOCAD™ version 17. The input data was then modelled using a GOCAD Structural Workflow to create the discrete-smoothing interpolation. This is a two stage process:

  1. First the computer creates triangular mesh based on the average data distribution (401 m) and the surface was interpolated using a DSI.
  2. Secondly the error between created triangular mesh and the input data was calculated. Where the error exceeded 4 m the mesh surface is locally densified and locally re-interpolated until the error between the surface and the input data is reduced to >4 m (Figure 5).

This process creates an irregular mesh which honours the data in areas of high data density yet produces an output which has a relatively small file size.

File:OR19003fig5.jpg
Figure 5    Modelled thickness of buried valleys discrete-smooth interpolation (DSI) created in GOCAD. The mesh has been densified in areas where the error between the final surface and the input data exceeded 4 metres.

The resulting data was then contoured at a 10 m interval and all those areas with less than 10 m of superficial deposits were deleted. The contour set was imported into ArcGIS and converted to polygons. The dataset was then manually inspected and compared to topography and input data and screened with the rules (below) to remove bull’s eyes and erroneous polygons:

  1. Features that were not centred on more than 5 boreholes which contain 20 m or more of superficial deposits were removed.
  2. All occurrence of superficial greater than 40 m were merged in to the ‘greater than 40 m’ polygons.

References

  1. KEARSEY, T I, WHITBREAD, K, ARKLEY, S, MORGAN, D, BOON, D, and RAINES, M. 2018. How accurate is your model between boreholes? Using shallow geophysics to test the best method to model buried tunnel valleys in Scotland, UK. Three-Dimensional Geological Mapping — Workshop Extended Abstracts. Vancouver, Illinois State Geological Survey 39.