In southwestern Saskatchewan, four formation names disappear into one coloured map unit.
Not because the rocks become one rock. Not because a mapmaker has found a continuous wall of identical material below the prairie.
On the Saskatchewan Geological Survey’s provincial legend, Eastend, Whitemud, Battle and Frenchman appear together as an undifferentiated Maastrichtian succession of less than 130 m.
The four names are still there.
The separation is not.
This is not a failure of the map. It is one of the most honest things a map can say.
The provincial bedrock-and-subcrop compilation is drawn at 1:1,000,000. Its metadata gives the physical translation: 1 cm on the page represents 10 km on the ground.
A contact thin enough, covered enough, uncertain enough, or merely too complex to show at that scale cannot be made more real by drawing it as a crisp line.
The polygon is an argument with a declared resolution.
It is not a promise of continuity.
A formation answers a question about rock
Formations are imagined as chapters in a geological calendar.
Frenchman comes after Battle.
Battle comes after Whitemud.
The names read as dates with thickness.
They are not.
A formation is a formal lithostratigraphic unit. A mappable rock body defined chiefly by lithology and stratigraphic position. Its name says that a particular package can be distinguished reproducibly from the packages around it.
It can include more than one lithology. It need not have formed in one environment. It need not have one constant thickness. It need not persist unbroken across every kilometre between observations.
The provincial compilation describes the Deadwood Formation as Upper Cambrian to Lower Ordovician. Its age label helps place it in Earth history. Its named green-red glauconitic clastic package is the formation.
Those are different statements.
They are held together in one map entry because a usable geological description needs both.
The provincial service keeps the distinction in its data model. It stores seven lithostratigraphic ranks, from major unit down through group, formation and member to bed. It separately stores eon, era, period, numerical age, chronology, lithology, rock code and geographic region.
The 1,814 Ma label on the Manitou Falls Group is an age attribute. It is not a rank.
Containment is not time.
Time is not lithology.
Lithology is not a map polygon.
The distinction is severe enough to survive familiar phrases. Campanian Bearpaw Formation combines a time assignment with a lithostratigraphic name. It does not make Campanian a member of Bearpaw, and it does not make every Campanian rock Bearpaw.
The same legend gives Bearpaw a Campanian-to-Maastrichtian assignment and a lithologic description that runs to grey claystone, shale, siltstone, minor sandstone, concretions, bentonite and fossils.
One label cannot carry all that without being unpacked.
Time correlation uses other evidence
Stratigraphy does not merely attach ages to formations. It asks how separated observations are related. Whether a horizon is the same physical bed, approximately the same age, or in an equivalent depositional position.
Those are not interchangeable successes.
Lithostratigraphy classifies rock bodies.
Chronostratigraphy classifies intervals of Earth history. Biostratigraphy classifies intervals by fossil occurrence. A shift in lithology, fossil assemblage, magnetic polarity, depositional geometry or inferred time need not occur at the same surface.
There is a simple reason. Lithology responds strongly to environment. Shore, basin, channel, reef and floodplain can each produce different rocks at the same time. The same environmental conditions can also recur at different times.
So a lithostratigraphic boundary can cross a synchronous surface laterally. Similar age does not guarantee similar rock. Similar rock does not guarantee the same age.
This is why a sequence framework and a formation map can both be correct while their boundaries fail to coincide. A sequence framework groups evidence around depositional surfaces, stacking patterns and changing accommodation. It is a model for correlation and geometry. A formation preserves a named mappable body.
The first can cut across the second without exposing an error.
The Alberta Geological Survey found several third-order sequences in western Athabasca core. That is useful basin interpretation. It is not a replacement hierarchy for Saskatchewan’s formal units. A sequence can organize a history across several rock bodies. A formation can remain the proper name for one of those bodies.
The names are not rival answers.
They are answers to unlike questions.
The observations have unequal powers
Every regional polygon begins as a collection of observations that cannot see the same things.
Outcrop can show a contact directly, where the contact is exposed. Core and cuttings sample a vertical interval directly, but only where a borehole exists. Wireline logs give continuous response down a hole, and a log motif becomes a lithology claim only after calibration to core or cuttings.
Seismic profiles connect laterally.
But a reflector is not a knife-edge rock surface. It is a wavelet and a tuning interval, tied to well information with known uncertainty.
No instrument cancels the limitations of the others.
In one USGS correlation release, synthetic traces made from sonic-log velocity information were repeated side by side to resemble a reflection profile.
The operation is practical.
It ties a well observation to a seismic event. It also names the gap. A direct fit is not perfect, because velocity modelling, bandwidth and lateral heterogeneity stay in the chain.
Another study calibrated well to seismic using velocity data from a set of borehole-compensated sonic logs. The result is not that seismic correlation fails. The result is that a correlation line needs a basis and a resolution.
That basis has to travel with the line. A core is not a map. A log is not a core. A seismic horizon is not a core. A coloured polygon is not a vertical section.
The words are close enough to be dangerous.
Below cover, the blank is not blank
The southern Phanerozoic part of the provincial product is largely a subcrop representation. It often shows bedrock below unconsolidated cover rather than exposed rock at the surface.
The cover hides the contact. It does not make the contact known.
Below cover, formations may be recognized through core, cuttings, wireline logs and well control. Where surface exposure is lacking, the International Stratigraphic Guide permits a type section in a well if samples and logs are adequate. That is not a relaxation of evidence. It acknowledges that rock can be classified from a subsurface observation, provided the observation and its provenance are clear.
But a well is a line through a volume. It says what occurred at that location.
The inference between wells adds correlation, structural interpretation and an account of uncertainty. Thickness can change through deposition, erosion, fault repetition, dissolution, compaction, and the place at which it is measured. A mapped footprint can be a subcrop extent without being the original depositional extent, or proof of every surviving buried remnant.
The map does not claim more than it can bear.
Readers often do.
The Meadow Lake Formation entry gives a thickness ceiling and an extent restricted northwest of the Meadow Lake Escarpment. That is a statement appropriate to a provincial compilation. It is not a borehole measurement at every coloured point.
The Prairie Evaporite Upper Salt map has a different evidentiary footing. Thousands of public well logs, a reference log and distinct top and base definitions let it record depth and thickness in a way a surface legend cannot.
Scale is not merely how much paper an object occupies.
It is the kind of statement that can be made honestly.
Undifferentiated preserves what it cannot draw
The Frenchman, Battle, Whitemud and Eastend entry has an easily missed virtue. It does not erase the units it groups. It keeps the fact that the succession contains several named formations, while declining to manufacture their individual contacts at provincial scale.
Undifferentiated is not a geological verdict that the units are alike. It is a cartographic verdict that the available scale or evidence does not support separation.
The distinction matters because a line looks more certain than a polygon. A solid contact can represent an observed exposure. A dashed line can represent a contact projected beneath cover. A queried line can record a plausible interpretation that is weakly constrained.
The convention does not turn one category into another. It makes the difference visible.
At the coarser scale, one combined colour may be the truer representation. A detailed core study can divide the succession more finely. Neither map invalidates the other. They occupy different places in the chain from rock to inference.
The same holds elsewhere.
A marker bed can be valuable in a local cross-section and too thin or discontinuous for a regional map. A group contains two or more formations with diagnostic properties in common, and grouping is still a communication choice. It is not permission to forget the individual rock bodies.
Resolution is not ignorance alone.
It is disciplined omission.
Correlation is useful because it can fail
The value of stratigraphy is not that it converts sparse observations into certainty. It converts them into a spatial hypothesis that can fail visibly.
Superposition supplies a robust local constraint when the succession is structurally understood. Overlying strata are younger than underlying strata. But faults can repeat or omit packages, erosion can truncate them, and structure can disturb the apparent order.
In one Florida study, high-resolution seismic profiles produced a stratigraphic and structural framework in which an inferred reverse fault offsets a formation by a few metres.
The figures do not belong to southwestern Saskatchewan.
The logic does.
A framework has to keep discontinuities in view rather than smooth them into a continuous section because continuity makes a prettier picture.
Once correlated surfaces and packages are in place, they can predict geometry away from direct observation. Where a body may thicken, pinch out, change facies or hold a certain stacking pattern.
That prediction is conditional.
A new core, well, exposure or seismic line can cross the expected surface, reverse the proposed trend, or reveal a fault. Then the interpretation has to change.
A map earns trust in precisely this way.
It says enough to be tested.
It does not claim to have already won.
What don’t we know?
The unanswered questions below cover are not a small residual around a finished provincial picture. They are built into the kind of evidence a map compiles.
Between sparse control points, contacts can be projected but not touched. Thickness can be interpolated but not made constant. A correlation can be plausible at map scale while its identity, age relation or depositional equivalence stays unresolved at core scale. A boundary may be real and displaced by faulting, removed by erosion, obscured by cover, or too close to another boundary for the chosen representation.
Even absence has more than one meaning. A bed may not have been deposited. It may have been eroded. It may be present and unpreserved, unexposed, unsampled, hidden by cover, or not recognized by the available method.
A visible gap in a section can record nondeposition, erosion, condensation, faulting or a failure of observation. The rock record does not supply the missing distinction automatically.
This makes stratigraphic mapping an inverse problem. Geologists observe surviving rock, logs, cuttings and reflected waves, then infer a three-dimensional arrangement and a history sufficient to relate them.
More than one history can fit a limited set of observations. Better calibration and more control points narrow the possibilities. They do not create the missing core or expose the covered contact.
The coloured unit in southwestern Saskatchewan is therefore not an empty simplification. It is the visible edge of an argument.
Eastend, Whitemud, Battle and Frenchman remain distinct names. At provincial scale they are shown as one succession because a finer division would say more than the map can demonstrate.
The map is useful because it tells the reader where the argument holds.
Its restraint is part of the evidence.