This GIS dataset depicts the result of a geostatistical model of the bedrock topography in Alberta. This model is based on water-well litholog data, stratigraphic picks made on oil and gas petrophysical logs, data derived from digitized contour interpretations on bedrock topography maps, including bedrock outcrop locations and the 60 m grid-spaced Shuttle Rader Topography Mission digital elevation model.
These data comprise the raster surface of Alberta Geological Survey Map 550, Bedrock Topography of Alberta, Canada.
We modelled the bedrock top elevations from an Alberta geological mapping database and the bedrock outcrop locations in ArcMap. Based on geostatistical analysis, the bedrock topography in the Edmonton-Calgary corridor shows a clear trend, increasing to the southwest toward the Rocky Mountains. A global, first-order polynomial accounted for this trend. We evaluated higher order polynomials, but these provided no further improvement to the model and added unnecessary complexity.
The residual values (measurements minus the trend) were modelled using ordinary kriging, which accounted for local variations in the trend value. We produced the final model by adding the global trend to the estimated residuals. Determining bedrock topography involved a three-step process: kriging, assessing the quality and fit of the modelled surface to the data (QA-QC) and re-kriging, if necessary. Alberta Geological Survey Open File Report 2010-12 describes this process.
The digital data used to construct the model and resulting map were derived from a number of sources. These sources included stratigraphic picks made from oil and gas petrophysical logs, maintained by the Energy Resources Conservation Board, lithological picks made from water-well litholog records, obtained from Alberta Environment, data derived from digitized contour interpretations from bedrock topography maps, previously published by the Alberta Geological Survey, and the Shuttle Radar Topography Mission 60 metre grid-spaced digital elevation model.
The bedrock topography data were modelled using ArcGIS Geostatistical Analyst. We accounted for regional variations in the bedrock surface by using a local, first-order, polynomial trend model with a weight parameter of 100 kilometres. We modelled smaller-scale features with ordinary kriging using an exponential variogram with a nugget effect of 200 square metres, a sill of 1500 square metres and a range of 14 kilometres. A smooth-search neighbourhood minimized any potential discontinuities between the different data sources.