FlagstoneReport
Does the work support the claim?
Assessment of an NI 43-101 technical report

Schaft Creek Project

Northwestern British Columbia, Canada (Cassiar/Liard Mining Division, ~61 km south of Telegraph Creek)
Study stage
FS · Feasibility Study
Published
Effective date January 23, 2013
Prepared by
Tetra Tech (Document No. 1053440200-REP-R0003-03.1); QPs include Hassan Ghaffari, P.Eng., Jianhui (John) Huang, Ph.D., P.Eng., and others; supporting consultants Moose Mountain Technical Services, Knight Piésold, McElhanney
Source
FS-PFS_Schaft_Creek_2013_compressed.pdf
Study-level anchor — not scored
FS Feasibility Study

Every judgment below is measured against what a FS is expected to demonstrate — not an absolute bar.

Detected from the filing

Title page and Section 1.1: 'Feasibility Study on the Schaft Creek Project, BC, Canada'; 'This feasibility study and National Instrument 43-101 (NI 43-101) technical report discusses scope, design features and economic viability of the Schaft Creek Project'. Mineral Reserves (Proven and Probable, 940.8 Mt) are declared, a full project execution plan with permitting/construction milestones is presented, and Section 21 references AACE-class capital estimation. A prefeasibility study was completed in 2008; this document supersedes it at FS level.

Peer Benchmarks

2 of the report's figures are benchmarked against FS peers — shown as the expected range for the stage and where this project sits within it.

2 within range
Cu Recovery – Flotation86.6%

Sits within the expected range for FS.

82%93.1%
Typical FS range 86.3%–90.9% · full peer span 82%–93.1%
Capex Intensity – Open Pit$69 /t·yr

Sits within the expected range for FS.

$21$157
Typical FS range $40–$76 · full peer span $21–$157
Internal consistency
Recovery claim vs. testwork86.6 → 86.2%

The claimed 86.6% carried into the economics is backed by the 86.2% testwork result.

Sitting outside the expected range is context to investigate, not a verdict — a project can legitimately sit apart for geological or metallurgical reasons. These positions inform the reader; on their own they do not decide the grade.

Flagstone Score

We grade two crucial dimensions, then determine a final score from them.

BMeets

Methods robustness

Does the disclosed work match what the study stage demands?

Judged against the Part 3 FS column: bench-scale flotation — present in abundance (twelve programs, 2004–2012). Variability composites — 'many, spanning grade and spatial variability' is satisfied: 34–36 individual interval samples in 2008 across Liard and Paramount, plus six 2012 composites explicitly selected to cover 'a significant portion of the mineralized zone and included different grade classes as well as lithology and alteration', plus five mine-plan composites representing the first five consecutive years of mill feed. Locked-cycle testing — required at FS and clearly present: ~26 bulk locked-cycle tests summarised in Table 13.25, plus five Cu-Mo separation locked-cycle tests. Pilot-scale — present (2007 and 2008 campaigns, 8,000 kg), exceeding what a conventional flowsheet strictly requires. Mineralogy→recovery linkage — partially satisfied: mineralogy is measured in detail (chalcopyrite dominant with bornite/chalcocite, pyrite 0.04–1.0%, copper-sulphide liberation 39–51% at 150 µm) and is used qualitatively to justify the 150 µm grind and to explain why concentrate grade is decoupled from head grade ('This indicates the significant effect of copper mineralogy on the copper concentrate grade'), but mineralogy is NOT translated into per-mineral or per-domain recovery equations. Flowsheet — deposit-specific and justified by measured data (SABC sizing simulated from the deposit's own SMC/DWi data at both 120 and 130 kt/d; regrind to ~20–25 µm; ferric chloride leach of Mo concentrate specifically to remove Cu/Pb identified as penalty elements in the multi-element assays). Cu-Mo separation is tested at FS level rather than assumed, which is the correct FS expectation. The material weakness is the recovery framework itself (see metallurgical_framework): a single undomained head-grade regression is applied to a deposit the report itself states shows 'a very poor correlation in the metallurgical performance among the samples tested' and unresolved inter-laboratory contradictions for Paramount/West Breccia. At FS this naive framework is a genuine deduction and is what holds the letter at B rather than A, despite the volume and rigour of the physical testwork clearly meeting every FS checklist item. Under Part 7.1, this profile matches B — every expected-for-stage item present (bench, extensive variability, locked-cycle, deposit-specific flowsheet, tested Cu-Mo separation) with the mineralogy→recovery linkage only partially delivered.

  • Report omissionaffects the gradeMineralogy is characterised in depth but is not converted into a per-mineral or per-domain recovery relationship; at FS the expectation is a required mineralogy→recovery linkage, and the delivered link is qualitative (grind-size justification and concentrate-grade explanation) rather than predictive.
  • Report omissionaffects the gradeThe QP explicitly notes the Paramount and West Breccia zones 'yielded different metallurgical performances' between PRA and G&T and that 'the reasons for the difference in the metallurgical performance are not clear', recommending further testwork — an unresolved metallurgical uncertainty carried into an FS.
  • Report omissionaffects the gradeNo gold search was performed on rougher or cleaner tailings; G&T recommended determining whether free gold is lost to those streams, so the gold deportment picture is incomplete.
  • Report omissionaffects the gradeThe pilot plant achieved only ~78% Cu recovery versus ~86% in locked-cycle work, attributed to elevated oxide copper in the pilot feed (~11% non-sulphide Cu vs ~7% in the bench sample); no oxide-copper domain or spatial model of oxide copper is presented to show how much of the reserve carries that elevated oxide content.
  • Data gapnot in the extracted text — does not affect the gradeDetailed derivation, sample-by-sample scatter statistics and residuals behind the Table 13.35 regression equations are referenced to Appendix D (full test reports), which is not in the provided text.
Metallurgical framework quality
Considers operating conditions beyond head gradepartial
Metallurgical domains defined, each with its own recoveryno
Range of lab results explained by the frameworkpartial

Explained vs. unexplained The delivered recovery model (Table 13.35) is a piecewise logarithmic function of head grade alone for each metal — Cu Rec = 6.2475·ln(Cu%) + 94.638 over 0.15–1.25% Cu, with flat values above and below. It is a single deposit-wide relationship: no separate equation for the Liard Zone versus the Paramount/West Breccia Zone, no separate treatment of oxide/secondary-copper-rich material versus primary chalcopyrite material, and no mineralogical variable in the equation. The framework therefore explains only the first-order grade trend. The report itself quantifies the residual: 'The variability test results show a significant variation in the metallurgical performances between the individual drill core interval samples. The variation can be traced to significant differences in mineralogy between the samples. The varying differences in the results have reflected a very poor correlation in the metallurgical performance among the samples tested.' Locked-cycle Cu recoveries in Table 13.25 span 70.0% (WLZ 2006) to 93.7% (Paramount 2007) — a ~24 point spread — while the model would predict roughly 88–90% across those head grades. Gold and silver scatter is described as more severe still, and the 2012 Sample 1 and Sample 2 underperformance is attributed post hoc to low copper content rather than captured by a domain. Positively: the report DOES diagnose the cause of the scatter as mineralogy (chalcopyrite:secondary-copper ratio averaging 2.7:1 but highly variable; pyrite ranging 0.04–1.0%; oxide copper ~3% weak-acid-soluble in 2010 composites and ~11% of total Cu in the 2008 pilot feed), and it DOES explain the pilot-plant shortfall by oxide content and the inter-laboratory Paramount/West Breccia divergence is at least flagged as unexplained rather than buried. Operating conditions are extensively tested (grind size, regrind size, pH, reagent dosage, and the locked-cycle grind-size series showing Cu falling 89→86% and Mo rising 71→87% between 109 and 173 µm) and one operating condition — the fixed 150 µm design grind and fixed 28% Cu / 50% Mo concentrate grades — is effectively baked into the model as a constant, but no operating-condition term varies within the recovery equations. Net: roughly the central trend is explained; the ~20-point locked-cycle spread and the mineralogy-driven per-sample scatter are acknowledged in prose but not carried into the model structure.

What the QP flagged in Section 13
  • riskParamount and West Breccia zone metallurgy'the samples from the Paramount Zone and the West Breccia Zone yielded different metallurgical performances. The metallurgical performance results obtained from PRA and G&T for the two zones were conversely different. The reasons for the difference in the metallurgical performance are not clear.' The QP notes the averages are similar to Liard and that 2012 Paramount samples behaved like Liard, but states 'Further test work is suggested for better understanding the metallurgical performances of the Paramount mineralization and the West Breccia mineralization.' An unresolved, laboratory-dependent metallurgical discrepancy on a zone that forms a substantial part of the resource, carried into an FS-level reserve using a single deposit-wide recovery curve.
  • limitationVariability scatter and mineralogical control'The variability test results show a significant variation in the metallurgical performances between the individual drill core interval samples. The variation can be traced to significant differences in mineralogy between the samples. The varying differences in the results have reflected a very poor correlation in the metallurgical performance among the samples tested.' The QP concedes the head-grade regression is fitted through scattered, mineralogically-driven results; block-by-block recovery prediction is weaker than the LOM average implies.
  • recommendationGold losses to tailings'No gold searches were conducted on the tailings streams, which include the cleaner flotation tailings and rough flotation tailings. G&T suggested determining whether there was free gold lost into the two tailings.' Gold recovery (73.0% LOM, a meaningful revenue contributor at 4.2 Moz LOM) rests on an incomplete deportment picture; possible upside or unquantified loss mechanism.
  • riskPilot plant underperformance vs locked-cycle'The metallurgical performance of the sample in the pilot plant tests was inferior to the results obtained from the locked cycle tests. G&T indicated that the inferior metallurgical performance was due to the elevated copper oxide content in the sample' (~11% non-sulphide Cu vs ~7% in the KM2136 bench sample); 2007 pilot results likewise attributed to 'difficulty maintaining control in small-scale tests'. The continuous-operation demonstration achieved ~78% Cu recovery against a design basis of 86.6%; the gap is explained by feed oxide content but the spatial distribution of high-oxide material is not modelled.
  • limitationGold recovery response to grind size'It is not clear why the gold recovery decreased with a finer primary grind.' The locked-cycle grind series showed Cu and Ag improving with finer grind while Au and Mo deteriorated. A counter-intuitive, unexplained response at the chosen design grind of 150 µm; the grind selection is justified partly on grinding-resistance economics rather than a fully understood metallurgical optimum.
  • opportunityMolybdenum concentrate copper penalty and rhenium'The copper content of the molybdenum concentrate should be higher than the penalty thresholds outlined by most molybdenum smelters. To avoid the penalty, the molybdenum concentrate should be leached by ferric chloride to reduce the copper level, which would increase molybdenum grade as well.' Also, rhenium in the 48–50% Mo concentrate 'is expected to be higher than 400 ppm… It is possible that the rhenium may be recoverable.' The chloride leach is designed into the flowsheet to meet the 50% Mo assumption; rhenium is an unbooked potential credit.
  • riskFluorine in copper concentrate'On average, the impurities in the copper concentrates produced from the mineralization should not attract smelting penalties as set out by most smelters. However, fluorine levels in some of the concentrates may be higher than the penalty thresholds.' Possible localised smelter penalty exposure not reflected in an average-concentrate assumption.
  • limitationConcentrate grade discount applied'Copper concentrate grade is assumed to be 28%, based on the average concentrate grade obtained from G&T's locked cycle test results adjusting down by 2.9% for potential inefficiency of industrial operations and for the improvement of metal recoveries.' A conservative, explicitly disclosed lab-to-plant adjustment — the trade-off between grade and recovery is stated rather than hidden.
  • recommendationHPGR wear characterisation'More ATWAL tests are recommended if the HPGR technology will be used for the Project.' HPGR remains a tested alternative rather than the selected route (SABC is selected); wear data would need completion before any flowsheet change.
  • limitationRecovery extrapolation beyond tested head grades'The metal recovery and concentrate grade estimates for the materials, having head grades beyond the range of the tested heads, are assumed based on the test results and experience.' Low-grade tail of the recovery curve (below 0.15% Cu, where the model steps to fixed 70%/45%/10%/0%) is judgement-based, relevant given the low LOM head grade of 0.271% Cu and stockpile blending in later years.
BAdequate

Result representativity

Are the reported numbers trustworthy as reported?

Against the Part 4 signals: (1) Variability behind averages — strongly disclosed. Table 13.25 lists every locked-cycle test individually with its own grind, regrind, head grades, concentrate grade and four metal recoveries, showing the full 70.0–93.7% Cu spread rather than only the mean; Table 13.6 lists all 36 variability head assays; Table 13.13 gives mineral composition sample-by-sample; Figures 13.14–13.18 plot the raw recovery-vs-head-grade scatter. (2) Test conditions — representative and honestly disclosed, including the poor performers. The report publishes the 2007 pilot results as low as 62% Cu recovery, the 2008 pilot at 78%, the 2006 WLZ locked cycle at 70.0%, and the 2012 Sample 2 at 78.5% — this is not a best-result-only presentation. It also discloses the graphite-contaminated 2006 PRA pilot as unusable rather than quietly dropping it. (3) Mineralogy grounding — present and used to explain results (secondary copper minerals explain high concentrate grades decoupled from head grade; low pyrite justifies the coarse 150 µm grind; oxide copper explains the pilot shortfall), though not built into the recovery equations. (4) Basis disclosure — clear and explicit: Section 13.6 states exactly how the projection was constructed (locked-cycle averages plus variability regressions, adjusted for Cu-Mo separation losses and bench-vs-proposed flowsheet differences), and explicitly flags that out-of-range head grades are assumption-based. The 28% concentrate grade is disclosed as the lab 30.9% discounted by 2.9 points for industrial inefficiency — a transparent, conservative adjustment. (5) Internal consistency (excluding the claimed-vs-measured recovery gap, which is handled in code) — the LOM 86.6% is consistent with the 86.2% locked-cycle average and with the block-model regression; Table 1.2, Table 1.7 and Table 17.2 are mutually consistent. Two real reservations hold this at B rather than A: first, the headline 86.6% is presented as a single LOM point with no confidence range or downside case despite the report's own admission of 'very poor correlation' between samples and a demonstrated 78% in continuous pilot operation; second, an unreconciled internal conflict is left standing — the Paramount/West Breccia inter-laboratory contradiction is acknowledged as unexplained yet a single deposit-wide recovery curve is nevertheless carried into the reserve and the NSR cut-off. Under Part 7.2 this is precisely the 'one or two real reservations while the bulk is traceable, mineralogy-grounded and basis-disclosed' profile of a B.

  • Report omissionaffects the gradeThe headline LOM 86.6% Cu recovery is carried as a single point estimate into the economics with no disclosed uncertainty band, sensitivity to recovery, or downside recovery case, despite locked-cycle results spanning 70–94% and pilot operation at 78%. The economic sensitivity analysis tests metal prices, exchange rate, capex and opex — but not recovery.
  • Report omissionaffects the gradeThe unexplained PRA-versus-G&T divergence for Paramount and West Breccia zone metallurgy is acknowledged but left unreconciled, while the resulting recovery curve is applied uniformly across both zones in the block model.
  • Report omissionaffects the gradeHead grades of the tested composites (typically 0.32–0.45% Cu, G&T locked-cycle average head 0.37% Cu) sit materially above the LOM reserve head grade of 0.271% Cu; the report does not explicitly demonstrate that the tested population is grade-representative of the full 21-year mill feed, relying instead on the regression to extrapolate downward.
  • Data gapnot in the extracted text — does not affect the gradeFull statistical treatment of the regression fits (residuals, prediction intervals, sample weighting) sits in the underlying laboratory reports in Appendix D, not in the provided text.
Flagstone Score
B

Work solidly supports the stage; minor reservations.

How the grade was reached

The letter is a deterministic roll-up of the two dimension grades and the peer flags — the same inputs always produce the same grade.

  1. Recovery: claimed 86.6% (carried into economics) vs measured 86.2% -> gap +0.4 pts
  2. Gap not material -> no representativity penalty
  3. Weakest link: methods=B, representativity=B -> take the lower = B
  4. Weighted penalty = -0 notches, capped at -3 and floored to whole notches -> -0
  5. Final grade (floored at D) = B

Methods letter B (Meets). Judged against the Part 3 FS column, this report clears essentially every expected-for-stage item and then some: twelve testwork programs over eight years across five laboratories; ~26 bulk locked-cycle tests (required at FS); two pilot plant campaigns including an 8,000 kg twelve-run Liard campaign (not required for a conventional SABC-flotation flowsheet, so this is above-stage work); a genuinely broad variability suite of 34–36 individual interval samples plus zone composites plus five-year mine-plan composites; comprehensive comminution characterisation (Bond suite, JK drop-weight, SMC, HPGR) on multiple samples feeding deposit-specific SABC simulations at two throughputs; and — importantly at FS — Cu-Mo separation was actually tested in five locked-cycle runs rather than assumed, which is the correct FS behaviour and removes what would be a legitimate FS gap. What prevents an A is the metallurgical framework. The recovery model carried into the reserve is a single, undomained, head-grade-only logarithmic curve per metal (Table 13.35), applied identically to Liard and Paramount/West Breccia, to primary chalcopyrite and to secondary/oxide-copper-bearing material. The report itself states the variability results show 'a very poor correlation in the metallurgical performance among the samples tested', traces that scatter to mineralogy, admits the Paramount/West Breccia inter-laboratory results were 'conversely different' for reasons 'not clear', and recommends further testwork — all inside a Feasibility Study whose reserve and NSR cut-off depend on that curve. At FS the expectation is that the metallurgy is understood, domains defined and residual variability minimal; a naive head-grade model with acknowledged unexplained scatter is a real, stage-scaled deduction. Mineralogy is measured extensively but only qualitatively linked to recovery, which is the one expected-for-stage item delivered partially. Volume of work at A level, framework at C level, netting to B. Representativity letter B (Adequate). The numbers are unusually well documented: every locked-cycle test is tabulated individually with its own conditions and results, the full 70–94% Cu recovery spread is visible, poor performers are published (62% and 78% pilot recoveries, 70% WLZ locked cycle, contaminated 2006 test disclosed as unusable), the basis of the projection is explicitly stated in Section 13.6 including the deliberate 2.9-point downward adjustment on concentrate grade for industrial inefficiency, and out-of-range extrapolations are flagged as assumption-based. This is honest reporting, not value-chasing. Two reservations keep it from A: the headline 86.6% is a single point with no recovery uncertainty band and recovery is absent from the economic sensitivity analysis; and the Paramount/West Breccia conflict is left unreconciled while a uniform curve is nonetheless applied. Peer position, FS/PFS bands only: claimed flotation Cu recovery of 86.6% sits below the [Cu Porphyry] green threshold of ≤90.5% — comfortably within peer norms and close to the PFS/FS peer median of 87.0% (p25 86.0). Capex intensity on the open-pit band computes from Cdn$3,256.6 M over ~47.45 Mtpa nameplate to roughly $69/annual t-yr (or ~$73 using the 44.8 Mtpa LOM average), landing inside the [Cu Porphyry] green range of $40–76/t-yr and near the PFS/FS open-pit p75 of 75.6. Grind P80 of 150 µm is at the lower edge of the green 150–200 µm band. Opex at Cdn$13.60/t milled is below the $15.1 green threshold, though note this is an all-in figure in Canadian dollars. There is no same-stage benchmark entry for leach recovery here (the project is flotation-only), so none is applied. On the claimed-versus-measured recovery signal, the 86.6% economic figure and the 86.2% locked-cycle average are effectively the same number and the report explains the derivation path from testwork to projection, so there is no unreconciled gap for the code to penalise — the pilot's 78% is lower but is explicitly reconciled by the QP to elevated oxide copper in that specific feed. Net picture: a thoroughly worked, transparently reported feasibility study whose physical testwork comfortably supports FS confidence, let down at the margin by a recovery framework that is simpler than the deposit's own demonstrated metallurgical heterogeneity warrants.