Unit 8 — Integrated stratigraphic research and applications

Construct, critique and communicate a multi-evidence basin reconstruction.

Integration requires more discipline than naming every plausible process. A basin reconstruction is a model of deposition, erosion, time and preserved geometry supported by incomplete observations. A good interpretation includes predictions about places that were not used to fit the original conceptual diagram. Hold back one section, test the reconstructed facies belt there and record failure honestly. Correlations should identify whether the line follows a rock unit, fossil datum, age marker or interpreted flooding surface. These lines need not coincide. The final report should include a sediment budget, a correlation panel, one quantified uncertainty and a ranked list of field observations most likely to settle the remaining disagreement. Applications to engineering, flood risk or subsurface resource decisions require separate site-specific professional methods. The purpose of a textbook model is to sharpen questions and identify evidence, not to license high-stakes operational decisions.

What is being explained?

Lateral facies shifts change rock character across synchronous surfaces; correlation demands distinguishing rock identity from time equivalence. Explain which visible feature would distinguish this process from later sedimentary reworking or burial modification.

The observational record

Mapped marker beds, fossils, dated ash and borehole logs support different kinds of connection between sections. Record the scale and instrument behind each observation; do not convert a field description into a claim about ancient environment without additional evidence.

Competing interpretations

Two sandstones with identical grain size may occupy different times if shoreline facies moved through the basin. Describe at least one predicted feature that differs between the rival and leading model, and identify a realistic observation that would distinguish them.

A discriminating investigation

Build a three-column panel with explicit lithologic and time lines; show where each must diverge. Specify the comparison material and observable decision rule so that a negative test outcome could change your preferred interpretation.

Field notebook and facies recognition. Write a two-column field record headed observation and inference. Enter grains, laminae, colour, contact shape and any measurable geometry in the observation column; reserve terms describing ancient rivers, shorelines or deep water for the inference column. Choose an outcrop reference scale suitable for the problem: centimetres for a ripple foreset, metres for a channel body, and kilometres for a basin map. Mark beds that cannot be observed because of cover. The same feature may be meaningful at one scale and misleading at another. Draw the simplest facies association consistent with the visible relationships and list one expected lateral equivalent. In your report do not call an exposure representative merely because it is accessible; accessibility itself may produce sampling bias.

Worked quantitative exercise 36: Time-equivalent thickness contrast

Hypothetical inputs: 68 m in a proximal section and 32 m in a distal section between identical ash markers. The transparent classroom expression is 68/32. Evaluating this expression produces 2.125 ratio. Identify the numerator and denominator when the calculation produces a percentage, and write each conversion between metres, kilometres, days, years or millions of years explicitly. If the result is a transport rate, do not describe it as an observed sedimentation rate unless storage and bypass are accounted for; if it is a thickness, identify whether erosion and compaction are represented.

Evidence limitations and a qualified conclusion

Fault offsets, unconformities and sparse wells create ambiguous links. Categorize the dominant uncertainty as sampling, analytical, chronological, mechanistic or preservation-related, and name the measurement that would reduce it.

State the property being correlated for every line and do not hide uncertain picks.

Practice and evidence checks

1. Describe correlating sections across facies transitions using three observations that do not yet assume a depositional setting.
2. State the physical reason the proposed mechanism would produce one of those observations, then identify a contrary observation that would count against it.
3. Recalculate Exercise 36 after modifying one stated parameter, document units and explain sensitivity.
4. Design an independent test using the protocol above; state a clear predicted outcome for the leading model and for its rival.
5. Write an 80-word report that explains what is known, what remains uncertain and why the distinction matters when interpreting preserved strata.

What is being explained?

A robust model predicts facies distributions and bounding surfaces in unexamined places rather than merely redescribing measured outcrops. Explain which visible feature would distinguish this process from later sedimentary reworking or burial modification.

The observational record

Cross sections, palaeocurrents, age constraints, process analogues and subsurface logs are potentially independent evidence streams. Record the scale and instrument behind each observation; do not convert a field description into a claim about ancient environment without additional evidence.

Competing interpretations

A visually elegant model may fit current sections yet make inaccurate predictions about a new well or outcrop. Describe at least one predicted feature that differs between the rival and leading model, and identify a realistic observation that would distinguish them.

A discriminating investigation

Reserve a synthetic section as a blind test and compare predicted facies occurrence with the withheld evidence. Specify the comparison material and observable decision rule so that a negative test outcome could change your preferred interpretation.

Measurement and dimensional reasoning. For the numerical component, write the governing relation and then translate every quantity into a unit consistent with that relation. State whether thickness is true stratigraphic thickness or apparent width, whether grain size is a mass-weighted or number-weighted distribution, and whether a flux represents sediment delivered or rock preserved. A point estimate may hide a large physical uncertainty. Recalculate after changing one assumed input by ten percent, and distinguish the mathematical sensitivity from the real variability of that input. Present units in the plotted axis titles rather than only in figure captions. A sensible arithmetic result is not sufficient evidence that the equation is valid at a particular outcrop. Discuss whether equilibrium, uniform flow, conserved grain mass or constant deposition has been assumed.

Worked quantitative exercise 37: Facies prediction accuracy

Hypothetical inputs: 36 withheld cases; 18 true positives and 11 true negatives, thresholds chosen for exercise. The transparent classroom expression is 100*(18+11)/36. Evaluating this expression produces 80.5556 %. Identify the numerator and denominator when the calculation produces a percentage, and write each conversion between metres, kilometres, days, years or millions of years explicitly. If the result is a transport rate, do not describe it as an observed sedimentation rate unless storage and bypass are accounted for; if it is a thickness, identify whether erosion and compaction are represented.

Evidence limitations and a qualified conclusion

Parameters may compensate for each other, making several basin histories observationally equivalent. Categorize the dominant uncertainty as sampling, analytical, chronological, mechanistic or preservation-related, and name the measurement that would reduce it.

Document predictive failures and revise hypotheses rather than treating every exception as noise.

Practice and evidence checks

1. Describe building and testing a depositional model using three observations that do not yet assume a depositional setting.
2. State the physical reason the proposed mechanism would produce one of those observations, then identify a contrary observation that would count against it.
3. Recalculate Exercise 37 after modifying one stated parameter, document units and explain sensitivity.
4. Design an independent test using the protocol above; state a clear predicted outcome for the leading model and for its rival.
5. Write an 80-word report that explains what is known, what remains uncertain and why the distinction matters when interpreting preserved strata.

What is being explained?

Alluvial processes, shoreline erosion and ground conditions have societal consequences but require site-specific engineering and monitoring. Explain which visible feature would distinguish this process from later sedimentary reworking or burial modification.

The observational record

Sediment budgets, historical shoreline maps, stratigraphic logs and uncertainty ranges can inform hazard screening. Record the scale and instrument behind each observation; do not convert a field description into a claim about ancient environment without additional evidence.

Competing interpretations

A teaching estimate of channel migration or subsidence is not sufficient to define a building setback or design flood. Describe at least one predicted feature that differs between the rival and leading model, and identify a realistic observation that would distinguish them.

A discriminating investigation

Present scenarios to a non-specialist audience while clearly separating screening-level evidence from regulated engineering analysis. Specify the comparison material and observable decision rule so that a negative test outcome could change your preferred interpretation.

Hypothesis competition and counterexample. Begin with a pair of incompatible explanations that both seem plausible from the first observation. Instead of collecting more features that merely illustrate the preferred story, identify the measurement on which the predictions disagree. Consider that one hypothesis may involve sediment transport while the other involves later erosion or cementation; these are not mutually exclusive unless the timing is explicitly constrained. Record the expected observation under each explanation before seeing any additional dataset. Evidence that fails to discriminate is still descriptive evidence, but it does not rank the competing models. Where a third process could imitate both, give that possibility a separate row in the reasoning table. The final verdict may be that present evidence supports only a family of environments or ages.

Worked quantitative exercise 38: Illustrative shoreline retreat

Hypothetical inputs: 370 m net shoreline change in 25 yr from maps of comparable accuracy; not a forecast. The transparent classroom expression is 370/25. Evaluating this expression produces 14.8 m/year. Identify the numerator and denominator when the calculation produces a percentage, and write each conversion between metres, kilometres, days, years or millions of years explicitly. If the result is a transport rate, do not describe it as an observed sedimentation rate unless storage and bypass are accounted for; if it is a thickness, identify whether erosion and compaction are represented.

Evidence limitations and a qualified conclusion

Rare events, future climate forcing and human modifications may be absent from historical observations. Categorize the dominant uncertainty as sampling, analytical, chronological, mechanistic or preservation-related, and name the measurement that would reduce it.

Communicate limitations and refer operational safety decisions to qualified local experts.

Practice and evidence checks

1. Describe sedimentary hazard and engineering communication using three observations that do not yet assume a depositional setting.
2. State the physical reason the proposed mechanism would produce one of those observations, then identify a contrary observation that would count against it.
3. Recalculate Exercise 38 after modifying one stated parameter, document units and explain sensitivity.
4. Design an independent test using the protocol above; state a clear predicted outcome for the leading model and for its rival.
5. Write an 80-word report that explains what is known, what remains uncertain and why the distinction matters when interpreting preserved strata.

What is being explained?

A repeatable study preserves raw measurements, metadata, analysis code and all decisions that change the interpretation. Explain which visible feature would distinguish this process from later sedimentary reworking or burial modification.

The observational record

Georeferenced logs, grain-size tables, sample photographs with scales and versioned correlation picks permit independent review. Record the scale and instrument behind each observation; do not convert a field description into a claim about ancient environment without additional evidence.

Competing interpretations

A polished graph can hide unreported measurement choices or a sample subset that reverses the trend. Describe at least one predicted feature that differs between the rival and leading model, and identify a realistic observation that would distinguish them.

A discriminating investigation

Provide a complete synthetic dataset, a calculation notebook and an audit trail from measurement to plot. Specify the comparison material and observable decision rule so that a negative test outcome could change your preferred interpretation.

Sampling design and data integrity. Prepare an explicit sampling protocol: what interval is the population of interest, how will representative sites be selected, which fraction of the sediment is included, and what precision is possible? Replicate critical counts because sedimentary deposits are patchy in both space and time. A sieve sample, thin section and borehole log observe different volumes and may not sample equivalent properties. Record locations and orientation without inventing precision that was not measured. For destructive sampling, respect collection permits and institution rules. Store the original readings separately from any normalized percentages so that the transformation can be inspected later. Treat missing data as missing, not as zeros, and write down every exclusion criterion before producing an attractive final plot.

Worked quantitative exercise 39: Replicate measurement range

Hypothetical inputs: largest logged-bed thickness 12.8 cm and smallest 11.6 cm across observers. The transparent classroom expression is 12.8-11.6. Evaluating this expression produces 1.2 cm. Identify the numerator and denominator when the calculation produces a percentage, and write each conversion between metres, kilometres, days, years or millions of years explicitly. If the result is a transport rate, do not describe it as an observed sedimentation rate unless storage and bypass are accounted for; if it is a thickness, identify whether erosion and compaction are represented.

Evidence limitations and a qualified conclusion

GPS error, changing analytical methods and inconsistent terminology hinder dataset merging. Categorize the dominant uncertainty as sampling, analytical, chronological, mechanistic or preservation-related, and name the measurement that would reduce it.

Reproducibility means another analyst can regenerate the result and diagnose disagreements, not simply obtain the same printed value.

Practice and evidence checks

1. Describe reproducible sedimentological data practice using three observations that do not yet assume a depositional setting.
2. State the physical reason the proposed mechanism would produce one of those observations, then identify a contrary observation that would count against it.
3. Recalculate Exercise 39 after modifying one stated parameter, document units and explain sensitivity.
4. Design an independent test using the protocol above; state a clear predicted outcome for the leading model and for its rival.
5. Write an 80-word report that explains what is known, what remains uncertain and why the distinction matters when interpreting preserved strata.

What is being explained?

Synthetic stratigraphic records can contain erosional gaps, diachronous shoreline facies and diagenetic overprints that resist a unique history. Explain which visible feature would distinguish this process from later sedimentary reworking or burial modification.

The observational record

A combined facies panel, two ash ages, a fossil datum and ambiguous isotope shift create multiple defensible interpretations. Record the scale and instrument behind each observation; do not convert a field description into a claim about ancient environment without additional evidence.

Competing interpretations

One model may prioritize relative sea level while another emphasizes shifting sediment supply and basin subsidence. Describe at least one predicted feature that differs between the rival and leading model, and identify a realistic observation that would distinguish them.

A discriminating investigation

Make a ranked list of proposed tests, include a sensitivity analysis and write a short manuscript with an uncertainty map. Specify the comparison material and observable decision rule so that a negative test outcome could change your preferred interpretation.

Comparative synthesis and communication. Move from a single log to a lateral comparison. Propose which properties ought to remain similar between neighbouring sections and which may change as environments migrate. A marker bed is only a reliable time line when evidence establishes its correlation; similar sandstone colour is not such evidence. Choose a visual representation that distinguishes observed boundaries from extrapolated ones, perhaps using solid versus dashed lines and a brief uncertainty legend. Present at least one alternative explanation in the figure caption rather than burying disagreement in the final paragraph. A coherent story should preserve contradictions instead of quietly removing inconvenient beds. In the conclusion identify the smallest claim that would remain valid if your preferred process model were overturned.

Worked quantitative exercise 40: Synthetic capstone average accumulation

Hypothetical inputs: aggregate 73 m from three preserved units over interpreted 1.5 Myr including unmeasured gaps. The transparent classroom expression is (23+31+19)/(1.5). Evaluating this expression produces 48.6667 m/Myr. Identify the numerator and denominator when the calculation produces a percentage, and write each conversion between metres, kilometres, days, years or millions of years explicitly. If the result is a transport rate, do not describe it as an observed sedimentation rate unless storage and bypass are accounted for; if it is a thickness, identify whether erosion and compaction are represented.

Evidence limitations and a qualified conclusion

Ambiguity may remain even after every available dataset is used; a model need not deliver a unique history. Categorize the dominant uncertainty as sampling, analytical, chronological, mechanistic or preservation-related, and name the measurement that would reduce it.

Defend the smallest set of conclusions that all credible models share and state which additional observation would decide among them.

Practice and evidence checks

1. Describe capstone: conflicting evidence in a changing basin using three observations that do not yet assume a depositional setting.
2. State the physical reason the proposed mechanism would produce one of those observations, then identify a contrary observation that would count against it.
3. Recalculate Exercise 40 after modifying one stated parameter, document units and explain sensitivity.
4. Design an independent test using the protocol above; state a clear predicted outcome for the leading model and for its rival.
5. Write an 80-word report that explains what is known, what remains uncertain and why the distinction matters when interpreting preserved strata.

Integrate three logs, two dated horizons, one mapped fault and a synthetic grain-size dataset into a self-contained report. Withhold one log until predictions are written. Compare observed and predicted facies, label correlation uncertainty and list three potentially decisive new observations. The conclusion should separate claims robust to all tested histories from claims conditional on a preferred reconstruction. Write an explicit limitations and responsible-use statement.

Practical workflow

A teaching dataset is a deliberately simplified set of observations with enough metadata to support a reproducible calculation. It is not evidence from a named real basin. Before graphing the numbers, decide which independent variable you think orders the processes and which observations might instead reflect changes in sampling position. Draw uncertainty markers or ranges where possible, even when the figures are hypothetical. A single good-fitting curve should be treated as a question: Which physical assumptions created it, and what independent data would disprove it? The submission must contain both a physical interpretation and a statement of one competing explanation. If a student uses a generative model or plotting program, they should preserve the original measurements and equations so the result can be independently reproduced.

The first deliverable is a clean measured-section or experimental-data table with a metadata line defining measurement units. The second is a graph or annotated cross-section that does not confuse elevation and stratigraphic thickness. The third is a two-model comparison in which each model must make at least one concrete additional prediction. The fourth is a brief calculation checked by an independent method and accompanied by a description of uncertainty. The fifth is a research note that states what cannot be resolved without additional evidence. Assign confidence to individual inferences, not to the diagram as a whole.

Data critique and transfer

A reproducible student submission must preserve a copy of the raw values, retain any derived normalized quantities separately and state which values were inferred rather than measured. Consider the possibility that the most accessible outcrop or easiest-to-analyse core was systematically unrepresentative. Plan an independent observation that would most sharply reduce uncertainty. If the best test is not feasible, present the honest limit of the inference rather than substituting a plausible-looking numerical value. This is especially important when educational methods are adapted to real land-use or subsurface questions.

Across these five topics, the central scientific skill is to distinguish a physically possible depositional interpretation from one that is specifically supported by evidence. Submit all five numerical exercises with units, the integrated research studio, and a correlation or process diagram that uses line styles to separate observed boundaries from inferred ones. Cite the source of external classification terminology when appropriate. A brief oral defence should explain which sedimentary observation would best falsify the preferred interpretation, how a missing interval affects the reconstruction, and whether the strongest evidence is geometric, compositional, chronological or process-based.

Self-assessment questions

• Which observations in this unit are primary depositional properties, and which could be modified by burial or structural deformation?
• Which quantities are measured directly and which are model-dependent estimates?
• Where would another measured section most effectively distinguish competing hypotheses?
• Does the proposed correlation connect similar rocks, synchronous time markers or a hypothesized environmental boundary?
• What remains valid if the preferred interpretation is wrong?