OR/15/002 Appendix 1: Detail of amendments to input datasets compared to Version 1

Ó Dochartaigh B É,Doce D D, Rutter H K and MacDonald A M. 2015. User Guide: Groundwater Vulnerability (Scotland) GIS dataset, Version 2. Revised Report. British Geological Survey Internal Report, OR/15/002.

The input datasets used to produce Version 2 of the groundwater vulnerability dataset are listed and described briefly in Table 2. More detail on how they have been amended from Version 1 of the groundwater vulnerability dataset is given here. The derivation/ownership of the input datasets is given in brackets after the name of each dataset. For more detail on the original derivation of these datasets for Version 1, see Ball et al. (2004).

• DiGMapGB-50 (BGS)

• DiGMapGB-50, the Digital Geological Map of Great Britain at 1:50 000 scale, for bedrock and superficial deposits, are the key input datasets. These have been significantly revised since Version 1 of the groundwater vulnerability dataset was produced. Version 5.18 was used during the creation of the vulnerability map. For a few parts of Scotland, DiGMapGB-50 is not available for superficial deposits, and in these cases, superficial geology data at 1:625 000 scale has been merged in.

• Permeability of bedrock and superficial geological units (BGS)

• This has been taken from the latest versions of BGS’s Permeability Datasets for Great Britain, Version 6 (2010) for bedrock and superficial deposits, which have been revised since Version 1. These datasets define both a minimum and a maximum permeability for each geological formation to encompass the likely range of permeability for any one formation, both because of the natural heterogeneity of geological formations (e.g., the presence of siltstone or mudstone bands within a dominantly sandstone sequence), and because of the lack of measured permeability data available for most formations, and therefore the uncertainty in ascribing a single permeability value. For the purposes of the vulnerability assessment, a cautionary approach was taken in that the maximum possible permeability is the most relevant, since this defines the worst case scenario in terms of groundwater vulnerability.
• For the purposes of the vulnerability map, a small number of changes were made to the bedrock permeability classifications in the latest version of BGS’s Permeability Dataset for Great Britain, in those cases where new information (e.g. from test pumping) has become available since the dataset was last revised, or where errors were noted.
• The superficial deposits permeability classifications in the latest version of BGS’s Permeability Dataset for Great Britain have been used directly in the production of the groundwater vulnerability dataset, with no changes. A number of classifications in this dataset are recognised as being potentially wrong on a local scale, either because of the level of available detail and accuracy in the geological mapping, or because of the natural heterogeneity of geological formations, in particular superficial deposits. As has been discussed in the user notes for the associated aquifer productivity maps (Scotland) (Ó Dochartaigh et al. 2015), classifying superficial deposits aquifers on a national scale is subject to much uncertainty, and must of necessity involve simplification, generalisation and therefore the masking of local variations. Because this groundwater vulnerability dataset is also made on a national scale, it is accepted that the same approach of simplification must be made. These points in particular, however, are made:

• The permeability of superficial deposits refers only to the superficial deposits mapped at the ground surface. The permeability of any different superficial deposits at depth in the superficial deposits sequence has not been taken into account, EXCEPT where it is known that at least 5m thickness of clay exists within the sequence.
• The permeability of superficial deposits has been used both to help define the productivity of superficial deposits aquifers, and as an input into the groundwater vulnerability map in its own right. However, aquifer permeability and productivity are different things, so that, for example, geological formations may have high permeability but do not form productive aquifers (perhaps because they are too thin or of too small a lateral extent). There is not necessarily a direct relationship between high permeability and high aquifer productivity.
• For Version 1 of the groundwater vulnerability dataset, much focus was put into the subdividing of till deposits according to their likely permeability. This approach has been simplified for Version 2.

• For Version 1, the permeability of till deposits was subdivided using HOST and supporting soil parent material data to define three till subclasses: highly permeable tills (derived from Precambrian rocks and classed as local aquifers); moderately permeable tills (derived largely from Devonian sandstones in Strathmore); and low permeability tills (largely derived from Carboniferous sedimentary rocks in central Scotland) (Ball et al. 2004).
• This approach has been revised partly because it is no longer thought that any tills are highly permeable, and partly because of the level of available detail and accuracy in superficial deposits mapping. Superficial deposits were historically often mapped in less detail and/or with less precision than bedrock formations. The level of detail and accuracy in the current maps of glacial deposits mean it is not possible to accurately distinguish between till, moraine, and hummocky/moundy glacial deposits, although recent evidence suggests that moraine is typically significantly more permeable than till.
• The lithology of most of these glacial deposits is described interchangeably as diamicton or some mix of diamicton, silt, clay, sand and gravel. Because of this, the permeability of all has been classed as ranging from low (minimum) to moderate (maximum). Only in areas which have been remapped using modern techniques are the lithologies of these deposits described in more detail, for example where hummocky/moundy glacial deposits are described as comprising gravel and sand only, and these deposits are classed as having high permeability.
• In Version 1, till was treated separately from moraine and hummocky/moundy deposits, but this set up a false division between mapped ‘till’ and ‘moraine’, which in reality cannot be distinguished. It therefore also gave a false impression of the accuracy of the map.
• To promote consistency as well as transparency, and given this available level of detail and accuracy in the available geological mapping, for Version 2 of the groundwater vulnerability dataset the permeability of all glacial deposits, as for all other superficial deposits, has been taken directly from the latest version of the BGS Permeability Dataset of Great Britain (Version 6, 2010).

• Soil permeability (James Hutton Institute (formerly the Macaulay Institute))

• Soil permeability derived from HOST data was used to refine the vulnerability classification for areas where BGS maps show no superficial deposits. Where no superficial deposits are mapped and the HOST classes is one of 9, 16, 18, 24, 26, 28 or 29, which all refer to low permeability soils, the vulnerability classification is reduced.

• Soil thickness (James Hutton Institute (formerly the Macaulay Institute), BGS)

• Soil thickness, derived from HOST data in combination with superficial deposits mapping, was used to distinguish areas where essentially bare rock is exposed at the ground surface. Where the HOST class is either 22 or 27, which both refer to very thin soils, the vulnerability classification is increased.

• Depth to groundwater level in superficial deposits (BGS)

• An updated map of depth to groundwater in superficial maps has been produced for Version 2 of the groundwater vulnerability map, using updated input data. This map is based primarily on a surface representing the depth from ground level to the river base level, which is modelled based on a digital terrain model (DTM) and the locations of major rivers. This surface is termed the River Head Space (Bloomfield et al. 2007). It was recreated for Version 2 using the NextMap DTM (instead of the Ordnance Survey DTM used for Version 1). The same digital river network, from SEPA, was used as for Version 1. The River Head Space surface was converted to a vector file with four depth to groundwater classes: <3m; 3- 10m; 10-30m and >30m.
• This was then combined with HOST soils data which show soils where the groundwater table is <2m (taken as HOST classes 10 and 12), including these areas in the River Head Space <3m category. Where the HOST data show the water table is <2m and the River Head Space data show it is >3m, the HOST data takes precedence and the final classification is <3m, as the HOST data are based on real observations, while the River Head Space map is based on an unverified model.

• Depth to groundwater level in bedrock aquifers (BGS)

• Depth to groundwater in bedrock aquifers is defined only for aquifers with significant intergranular flow. This has been changed slightly from Version 1 because the bedrock aquifer flow type classifications in the bedrock aquifer productivity map (Ó Dochartaigh et al. 2015) have been updated. Depth to groundwater has been added for one other bedrock aquifer unit, the Glenvale Sandstone Formation west of Dundee, which was previously classed as having mixed intergranular/fracture groundwater flow but has been reclassified as having significantly intergranular flow. Where no groundwater level data are available for a bedrock unit with significantly intergranular flow, the depth to groundwater map defaults to the shallowest category, <3 m, as the conservative scenario (i.e. worst case in terms of groundwater vulnerability).

• Superficial deposits thickness (BGS)

• This dataset is unchanged from Version 1: a version of BGS’s Geohazard interpolated model based on borehole data, which ahs been manually corrected by Quaternary geologists in those areas where no borehole data exist. This combination of modelling and geological knowledge has provide the most realistic superficial deposits thickness model for Scotland at 1:50 000 scale.

• Presence of thick clays in the superficial deposits sequence (BGS)

• This dataset is unchanged from Version 1.