Thick overburden cover may hamper detection of kimberlite pipes and mineral deposits due to a lack of indicator minerals and/or geochemical signatures in till. To assist drift prospecting methods such as till geochemistry, hyperspectral surveys, and other geophysical techniques in areas with variable overburden thickness (up to 42 m), an overburden thickness model for the Slave Province is being developed as part of the Geological Survey of Canada's Remote Predictive Mapping project. The modelling process is based on the construction of a bedrock elevation database established from diamond drill-hole data, as well as data extracted from A-series surficial geology maps and from digital elevation models (DEM).

Extensive areas of the Slave Province are overlain by till, which is the most common deposit in the area as illustrated on the map displayed below. Till is divided into three subunits based on thickness and surface morphology: veneer (generally <2 m thick), blanket (8 m thick), and hummocky (10 m thick). Glaciofluvial deposits are geographically widespread but limited in extent, and range from a few metres thick to 10's of metres. Glacial lake sediments are only identified in a few locations, generally <4 m thick and also limited in extent.

Drift thickness maps represent the difference between the bedrock surface and a digital elevation model. In areas of outcrop the drift thickness is zero. In areas of glacial drift the bedrock surface is estimated as the elevation of an area, obtained form the digital elevation model minus the thickness of the glacial drift unit. In an area of till veneer the estimated bedrock elevation would by the elevation of the till veneer unit minus 2 meters, the thickness of the till veneer. Subtraction of the interpolated bedrock surface from the digital elevation model results in a drift thickness map. The bedrock elevation database is modelled in stages in order to establish the effects of each surficial unit on the drift thickness model.

The following steps where carried out to create a till thickness model.

1. Polygons representing each surficial unit are buffered inwards by 100 meters to avoid overlapping boundaries.
2. Polygons are converted to points that outline the perimeter of the polygon. The perimeter was used in order to keep the file sizes small. Additional data such as small individual outcrops and depth-to-bedrock obtained from diamond drill holes are combined with the polygon data.
3. The points are assigned an elevation determined from a point inspection of the nearest data point to a DEM elevation point. The 30 meter resolution DEM used for these two map sheets is displayed below.
   
4. Surface elevation point data is adjusted to account for till thickness to obtain data on the elevation of the bedrock surface.
5. The adjusted data set is surfaced using a natural neighbour interpolation with a cell size of 100 meters and an aggregation distance of 250 meters.
6. The resulting interpolated bedrock surface is merged with the DEM using the minimum value from both datasets to ensure that the interpolated bedrock surface is never at an elevation above that of the topographic surface elevation as determined from the DEM.
7. A drift thickness model is generated by subtracting the interpolated bedrock surface from the DEM.

The first dataset modelled incorporates depth-to-bedrock determined by diamond drill-holes (534 locations) combined with point elevations that correspond to the perimeter of outcrop polygons and individual outcrops as obtained from surficial geology maps. About 230 000 points where obtained from the information displayed on the map below.

A natural neighbour interpolation was carried out to create a model of the bedrock surface. Outcrop polygons are then 'stamped' with the corresponding clipped region of the DEM to provide an elevated surface for these regions that overrides the modelled surface. The resulting bedrock elevation model is subtracted from the DEM to establish a drift thickness model as displayed in Model 1 below.

Both the bedrock elevation and drift thickness models are modified by adding additional polygon elevation data extracted from surficial geology maps. Geographic distribution of till veneer polygons are displayed below.

After buffering, the polygons are assigned a 'best approximation' bedrock elevation by applying a point inspection from the DEM minus 2 meters to account for the thickness of the till veneer. Data from the point inspection of the till veneer polygons (approx. 580 000 points) is appended to the previous dataset containing outcrop and diamond drill hole depth to bedrock elevations for a total datset of about 820 000 points. The data set is re-modelled using a natural neighbour interpolation to establish a depth to bedrock surface and the resulting drift thickness model is displayed below.

Similar to till veneer the till hummocky polygons (displayed below) are buffered inwards by 100 meters and assigned a 'best approximation' bedrock elevation by applying a point inspection from the DEM minus 10 meters. Data from the point inspection of the till hummocky polygons (approx. 130 000 points) is appended to the Model 2 dataset containing outcrop and diamond drill hole depth to bedrock and till veneer elevations. The combined data set of about 950 000 points is remodelled using a natural neighbour interpolation to establish a depth to bedrock surface and the resulting drift thickness model as displayed in Model 3 below.

Similar to till veneer and hummocky till the polygons for blanket till (displayed below) are buffered inwards by 100 meters and assigned a 'best approximation' bedrock elevation by applying a point inspection from the DEM minus 8 meters. Data from the point inspection of the till hummocky polygons (approx. 740 000 points) is appended to the Model 3 dataset containing outcrop and diamond drill hole depth to bedrock, till veneer and hummocky till elevations. The combined data set of about 1 680 000 points is remodelled using a natural neighbour interpolation to establish a depth to bedrock surface and the resulting drift thickness model as displayed in Model 4 below.

Once the drift thickness model is completed cross-sections can be established in areas of interest. The cross section displayed below corresponds to the A-A' line displayed on Model 4. The green line represents the DEM and the yellow line represents the modelled bedrock surface. Drift thickness corresponds to the area between the two lines.