The hydraulic head in the Leduc HSU was first modelled separately in three regional areas of Leduc Formation carbonate platform and reef buildup complexes and then combined into one grid.
The following statistics describes the accuracy of the geostatistical layers used to create the final grid:
Region 1: Northwest Leduc (Peace River Arch platform and fringing reef complexes; Simonette, Sturgeon Lake, Gold Creek, Bigstone, Fir, Pine Creek, Obed, and Windfall reef complexes)
Samples: 134
Mean error: -0.14 m
Root-mean-square error: 31.3 m
Mean standardized error: 0.007
Root-mean-square standardized error: 0.98
Average standard error: 31.4 m
Region 2: Central Leduc (Rimbey-Meadowbrook reef trend; Cheddarville, Redwater, and Willingdon reef complexes)
Samples: 70
Mean error: -5.1 m
Root-mean-square error: 61.2 m
Mean standardized error: -0.035
Root-mean-square standardized error: 1.09
Average standard error: 37.1 m
Region 3: Southeast Leduc (Bashaw, Clive, Nevis, Stettler reef complexes; Killam barrier reef and southern Leduc shelf complex)
Samples: 276
Mean error: 0.2 m
Root-mean-square error: 38.7 m
Mean standardized error: 0.025
Root-mean-square- standardized: 0.97
Average standard error: 38.1 m
The hydraulic head grid was clipped to the data extent of both the hydraulic head and total dissolved solids data points.
The Leduc hydraulic head was modelled separately in three regions and then combined into one grid.
Selection criteria and the final data set used for interpolation of this grid are documented in the metadata for AGS DIG 2021-0028
Step 1 (Modelling the surface): An empirical Bayesian kriging algorithm was used in the ArcMap Geostatistical Analyst extension to interpolate the hydraulic values. This was done in three separate areas and then combined to create a final grid:
Northwest Leduc (Peace River Arch platform and fringing reef complexes; Simonette, Sturgeon Lake, Gold Creek, Bigstone, Fir, Pine Creek, Obed, and Windfall reef complexes)
Method Report:
Method: Empirical Bayesian kriging
Output type: Prediction
Transformation: None
Semivariogram model type: Power
Subset size: 100
Overlap factor: 1
Number of simulations: 100
Searching neighbourhood: Standard circular
Neighbours to include: 8
Include at least: 8
Sector type: 8
Radius: 20,279
Angle: 0
Central Leduc (Rimbey- Meadowbrook reef trend; Cheddarville, Redwater, and Willingdon reef complexes)
Method Report:
Method: Empirical Bayesian kriging
Output type: Prediction
Transformation type: None
Semivariogram model type: Power
Subset size: 100
Overlap factor: 1
Number of simulations: 100
Searching neighbourhood: standard circular
Neighbours to include: 8
Include at least: 8
Sector type: full
Radius: 32,840
Angle: 0
Southeast Leduc (Bashaw, Clive, Nevis, Stettler reef complexes; Killam barrier reef and southern Leduc shelf complex)
Method Report:
Method: Empirical Bayesian kriging
Output type: Prediction
Transformation type: None
Semivariogram model type: Power
Subset size: 100
Overlap factor: 1
Number of simulations: 100
Searching neighbourhood: standard circular
Neighbours to include: 8
Include at least: 8
Sector type: Four by 45 degree
Radius: 34,854
Angle: 0
Step 2 (Grid combinations): The three grids were combined using ArcMap's Add Rasters to Mosaic Dataset tool.
Step 3 (Grid Alterations): The final grid was clipped based on the extent of the Leduc HSU and the spatial distribution of data.