Identifying deep potash at ~1.5 km — tested semi-blind against drilling
A Saskatchewan programme with Gensource Potash used ADR Virtual Boreholes to distinguish the Prairie Evaporite, halite and potash-bearing sylvinite members, then tested trained relationships on two semi-blind locations before comparison with existing well data.
Can surface ADR distinguish potash-bearing sylvinite from surrounding halite at kilometre-scale depth?
The target sits within the Prairie Evaporite Formation, where the zone of interest comprises alternating halite and sylvinite layers. The programme sought to identify the formation boundaries, differentiate the three major potash members, delineate sub-members and investigate a preliminary high/low KCl grade classification.
The scientific paper reports that 13 Virtual Boreholes were acquired across the Vanguard Area. Detailed interpretation focused on four V-bores: H1 and H2 as training locations with well/core information available, followed by H3 and H4 as semi-blind tests interpreted independently before Gensource supplied the corresponding well data for comparison.
Train → predict semi-blind → reveal → compare
TRAIN
H1 (C1) and H2 (C2) were interpreted with existing drilling, lithology and assay information available to establish ADR relationships.
PREDICT
The trained relationships were applied to H3 (SB1) and H4 (SB2), with interpretation completed independently of the corresponding well data.
REVEAL
After interpretation, Gensource supplied the existing well information so the predicted and actual subsurface results could be compared.
COMPARE
The study compared lithological zonation, potash-zone boundaries, individual members, thicknesses and a high-level KCl grade classification.
More than detecting the evaporite package
Potash stratigraphy
Patience Lake, Belle Plaine and Esterhazy were delineated in the interpreted V-bores. The paper also reports differentiation of some upper and lower sub-members.
Semi-blind depth comparison
The potash zone was identified in both semi-blind V-bores, with reported approximate discrepancies between interpreted and well-data top/base depths of 2.8% for H3 and 1.4% for H4.
Unit-top correlation
The paper reports an R² of 0.8695 between interpreted and physical drilling unit-top depths, with stronger correlation in the higher-grade Patience Lake and Belle Plaine sylvinite beds.
Can ADR responses say anything about KCl grade?
Training
Gensource assay data were used to establish relationships in H1 and H2. For this study, intervals were classified as high (>30% KCl) or low (<30% KCl), using relationships involving WMF, E-Gamma and F-Gamma.
Semi-blind application
The trained relationship was subsequently applied to H3 and H4. This is best presented as a high-level grade classification demonstrated in this study, rather than as a substitute for laboratory assay.
A strong quantitative validation case for pre-drill potash exploration
Depth
The Prairie Evaporite occurs at approximately 1.5 km below ground surface in the study area, providing a deep-target test of the pulsed-radar approach.
Discrimination
The study reports differentiation of halite and sylvinite and delineation of individual potash members rather than merely identifying a broad evaporite reflector.
Evidence design
The training/semi-blind structure is commercially important: interpretations at H3 and H4 were made before the corresponding well information was released for comparison.
Evidence status: validated historical ADR case study. The scientific paper is an accepted/published GEOPHYSICS case history. Results are specific to this project and geology and should not be treated as a universal performance guarantee.
Bring VEXRAD one target.
We can assess target depth, geology, existing ground truth and whether a staged training-to-blind-test programme is appropriate.
Scientific source
Stove, G., Robinson, M., Fourie, L., Neufeld, P. & Ferguson, M. “Identification and delineation of potash deposits in Saskatchewan, Canada using pulsed radar technology.” GEOPHYSICS, Society of Exploration Geophysicists, 2020. The study describes 13 V-bores, detailed analysis of two training and two semi-blind V-bores, comparison with Gensource well/lithology/assay information, and the quantitative results summarised above.
This page is a VEXRAD presentation of a legacy ADR programme. Historical client and project names are retained for factual provenance.