FINAL SCIENCE AND FIGURE REVIEW — 26 October 2026

Scope
Reviewed the 50-profile depositional study, the 20-interval model inventory and
interval hierarchy audit, and all 10 final original scientific figures. This is
a consistency and presentation review, not an independent re-interpretation of
raw logs, seismic, core or hydrocarbon test data.

Scientific distinctions confirmed
1. The proposed town corridor is treated as a geographic discussion corridor,
   not a proven continuous shoreline, reef, oil pool or migration route.
2. Present structure is distinguished from depositional paleotopography.
3. Gross interval thickness is distinguished from net reservoir and net pay.
4. Core observations and regional analogues are identified separately; oil
   stain, fluorescence and pyrobitumen are not equated with productive pay.
5. Carbonate grain banks, dolomitized muddy carbonates, mud mounds and framework
   reefs are not treated as interchangeable reservoir geometries.
6. Jurassic Moosomin valleys and younger pre-Cantuar drainage are separated.
7. Torquay saline-lake and older marine interpretations remain alternatives;
   F4 facies and stratigraphic Unit 4 are distinguished.
8. Bakken competing sequence models are acknowledged rather than forced into
   one unique timeline. A regional NE-proximal/SW-distal trend is time-specific.
9. The interval audit supports mapped Midale including Frobisher Evaporite,
   mapped Ratcliffe including Midale Evaporite, and Total Jurassic including
   Upper Watrous and therefore overlapping Total Watrous. Total Mississippian
   excludes the separately mapped Bakken. These are same-source hierarchy
   checks, not independent vertical-accuracy validation.

Corrections sent to root for the study build
1. Correct Atlas Chapter 26 URL:
   https://ags.aer.ca/publications/atlas-western-canada-sedimentary-basin/chapter-26-quaternary-geology-western-plains
   The previous guessed URL did not resolve. The replacement was verified
   against the official AGS page.
2. Quaternary temporal wording should say 'tens to hundreds of millions of
   years' rather than universally 'hundreds of millions' older Mesozoic strata.
3. Moosomin B 'water-wet/uncharged' conflates wettability and saturation.
   Replace with 'water-bearing and uncharged' or simply 'uncharged'.
4. Rierdon's stale implementation note about checking layer IDs should become
   a clear statement that no separate Rierdon geometry was added to the model.

Targeted primary-source checks
- Bates & Kendall (2008), Figure 2 / printed p.811, explicitly documents local
  caprock failure through anhydrite dissolution at Parkman12-16-9-33W1, with
  hydrocarbon migration into overlying Watrous. That study wording is supported.
  https://geoconvention.com/wp-content/uploads/abstracts/2008/094.pdf
- Lake (2023) is explicitly an alternative core-display interpretation: local
  removal of upper Souris Valley reservoir by erosion, with a proposed downdip
  Northgate–South Fillmore fairway. It is not proof of a newly mapped corridor.
  https://www.apegs.ca/assets/sgs-talk-notice-23-05-24-lake.pdf
- Official AGS Atlas Chapter 8, 9, 10, 11, 12 and 27 URLs in the source list
  were checked and resolve.

Figures and assets
All 10 final figures were visually reviewed. Labels, legends, sources, scale
limitations, conceptual qualifiers and current-geometry qualifiers are legible;
no overlap or clipping was found. PNG integrity was checked with Pillow.
Final PNG and SVG pairs have been copied to the site dist/figures directory.
All figure paths currently referenced by study.json resolve. Intermediate
*_qa.png and *_review.png files were excluded.
Checksums and image dimensions are in final_figure_audit.json.

Recommended inline selection
1. model_01_corridor_section.png — actual regional model geometry.
2. miss_02_frobisher_evaporite_cycles.png — small facies trap and evaporite-cycle
   distinctions.
3. meso_01_valleys.png — separate Jurassic and Cretaceous valley systems.
Optional fourth: older_01_carbonate_facies_and_pores.png — older carbonate
settings and the effect of later pore modification.

Automated consistency checks
Every one of the 20 mapped interval IDs has a study profile.
No profile references a missing source ID.
No study figure references a missing site asset.
Model counts: 4,074 regional well records; 31,794 selected-interval picks;
37,505 retained fields including two unconformity fields. Field counts are
not independent wells or independent geological observations.

Residual limits
This review did not test a physical VR headset. Regional 2 km cells and 50 m
structure contour spacing cannot establish small-pool closure or reservoir
connectivity. Unmodeled space does not itself demonstrate formation absence.
