This dataset is part of the three-dimensional (3-D) hydrostratigraphic model of the Sylvan Lake sub-basin in the Edmonton-Calgary Corridor, central Alberta. It represents the structure top of hydrostratigraphic unit (HSU) C2, which contains the majority of fine-grained materials in the Neogene-Quaternary succession of the Sylvan Lake sub-basin and is the top boundary of the model domain.
This dataset supplements Alberta Geological Survey (AGS) Open File Report 2014-10, which includes a full description of steps taken to produce the structure grid.
10338 water wells from the AWWID were used in the study to broadly define hydrostratigraphic units in the Neogene-Quaternary succession. The AWWID has abundant, relatively low quality data records, which include location inaccuracies resulting in up to 10 wells with duplicate locations. In cases of collocated data, one well per geographic coordinate was selected, based on completeness of lithological descriptions and total depth, and duplicate wells were omitted.
The original input data were classified using unpublished hydrostratigraphic assignments to create a laterally extensive and mappable geobody. Classification of lithological data into hydrostratigraphic units may incorporate assumptions and human error when simplifying complex geological data, as unit geometry, extent, and description are a product of the author’s conceptual understanding. The HSU C2 grid is for use in regional-scale groundwater studies, as these mappable hydrostratigraphic units may be more complex or absent at the scale of a local investigation.
To optimize the HSU C2 grid for input into a numerical groundwater flow model, the grid was extended across the Sylvan Lake model domain. Where HSU C2 is absent, the grid nodes’ elevation is equal to the underlying HSU S2 grid plus a constant 0.10 m. The continuous HSU C2 grid, along with other key hydrostratigraphic units, has been exported to ASCII grid format and can be used as a framework for the creation of a groundwater flow model.
All borehole data from the AWWID were sampled to the 25 m Provincial Digital Elevation Model (DEM) to provide a borehole top elevation (m asl) to correlate hydrostratigraphic units between boreholes to create hydrostratigraphic grids. This process was necessary as the elevation of AWWID water wells, in many cases, was not available/recorded or verified, and if reported was estimated from a topographic map, field investigation or handheld Global Positional System (GPS) device. GPS devices have accuracies reported by the AWWID of ~1-30 m, but are of varying device quality and cannot be assessed for accuracy between devices. Elevation values of moderate and high quality data were also checked with the DEM to verify elevation accuracy and modified to match the DEM, if necessary.
The vertical accuracy of the 25 m DEM is stated as 5 m @ 90% in the Sylvan Lake sub-basin derived from 1:60 000 aerial photographs using photogrammetric methods.
Methodology to create this grid dataset is fully documented in AGS Open File Report 2014-10.
Modelling Program used to create grid: RockWorks15
Number of data points available: 8532
Data type: A variety of low- to high-quality borehole data were used, including lithologs submitted by water well contractors and compiled by the Alberta Water Well Information Database, downhole geophysical logs collected from water wells, and AGS continuous cores.
Algorithm used to interpolate data points: Inverse distance weighting (IDW2) with a minimum search radius of 16 points
Cell Size: 400 m
Deterministic Grid Alteration Information: Limited manual manipulation of grids was performed, with the exception of: 1) truncating HSU C2 below the land surface (i.e., the 25 m Provincial Digital Elevation Model); 2) clipping HSU C2 where the unit is known not to exist (i.e., Sylvan and Gull lakes); and 3) where HSU C2 is absent, adding a constant thickness of 0.10 m to the z-value of the underlying grid node (HSU S2) to produce a continuous HSU C2 grid across the Sylvan Lake sub-basin model domain.
Root mean square error: 5.9 m
The RMSE compares how similar interpolated grid values are to the original input data. This is a global estimate of error for the HSU C2 grid and provides neither indication of how error is distributed spatially in the model domain, nor the accuracy of the input data.