Historical well logs preserve measurements that may remain useful long after the original logging equipment and software have changed. Paper prints, scanned images, and PDF records can provide valuable evidence for correlation and interpretation, but their plotted curves are not directly equivalent to reusable numerical data. Digitization bridges that gap by converting selected information into a defined digital structure.
GAEA’s geophysical well-log digitization service describes a workflow of source review, depth and track calibration, curve capture, metadata checking, visual comparison, exception review, and LAS 3.0 file checks. This article explains what those stages mean and the project decisions that help make the resulting files useful. Specific scope and acceptance requirements should be agreed during source assessment.
Understand what is being recovered
A plotted geophysical curve relates a measured property to depth through the scales printed on the log. Recovering it requires more than recognizing a line. The digitization process must establish which curve the line represents, how its horizontal position translates into values, and how its vertical position relates to the depth reference.
This differs from scanning a document for storage or transcribing descriptive borehole records. If the project also requires lithological descriptions, construction details, handwritten notes, or other tabular information, identify those separately. A curve-digitization scope should not be assumed to include every item visible on the page.
1. Define the intended output before production
Start with the task the recovered data must support. A team comparing a particular curve across several wells may need a different scope from one preserving all available tracks. List the requested wells, logging runs, curves, depth ranges, and metadata. Identify the receiving application so the delivery can be tested in the environment where it will be used.
GAEA’s published service specifies LAS 3.0 delivery for scoped curves and available header information. Confirm curve naming, engineering units, the depth increment, and the missing-value convention in the project specification. A file that opens successfully can still be unsuitable if it omits an important curve or represents the wrong interval.
Agree how multiple records for the same well will be organized. Separate runs, repeated intervals, and later annotations can have different meanings. Preserve enough identification to distinguish them rather than silently combining everything under one filename.
2. Review representative source records
The existing workflow begins with representative material so difficult conditions are identified before the full archive is processed. Include clear records and challenging examples: faint traces, handwritten annotations, overlapping curves, page joins, or changing scales. A quotation sample containing only the cleanest page may not represent the production work.
Provide complete images wherever possible. Headers, legends, track labels, depth markers, and page edges can contain information needed to interpret the plotted curves. A cropped image may preserve the line but remove the evidence required to calibrate it. Identify any known missing pages or alternate copies during the initial review.
Build a simple inventory linking each source file to the intended well or run. Record duplicates and uncertain identities as questions to resolve. This prevents the appearance of a complete numerical dataset from concealing a basic uncertainty about where the information belongs.
3. Prepare images while preserving the original evidence
The source image remains the reference for later checks. Working copies may need rotation, careful cropping, or adjustments that improve legibility. Preserve an unchanged original and distinguish it from any prepared image so a reviewer can examine what was actually supplied.
Preparation should not invent a missing trace or conceal an annotation. When discussing image treatment, ask whether fine lines, small labels, and grid markings remain visible. An image that looks cleaner at normal viewing size may have lost details needed for technical review.
A practical project record can identify the source filename, prepared copy, affected page or interval, and any significant processing decision. The level of documentation can be proportionate to the archive, but the relationship between the original evidence and the working image should remain clear.
4. Establish the depth reference
Depth calibration connects vertical image positions to the printed depth values. Review top and bottom references, intermediate markers, page overlaps, and discontinuities. A local alignment that looks correct near one label does not by itself demonstrate that the entire page is registered correctly.
For example, consider a hypothetical multi-page log with an overlapping interval at a page join. The overlap can help a reviewer compare the two sections, but the delivered file should not accidentally duplicate that interval. If pages disagree or a section is absent, the discrepancy needs an explicit treatment rather than a silent join.
Keep depth units and the stated reference identifiable. If a source uses a convention that is unclear, seek clarification or record the limitation. A numerical column with orderly increments does not prove that its depth reference was correctly understood.
5. Calibrate each curve track
Track calibration translates horizontal position into the curve’s values. The project must account for linear or logarithmic presentation, scale direction, and changes in the printed range. Different curves sharing a track may use different scales, so visual proximity alone cannot establish their numerical relationship.
A hypothetical curve changing scale partway through a record illustrates the issue. Tracing the line consistently while applying the original scale to the whole interval would produce plausible-looking but incorrect values. The scale change needs to be identified and reflected in the capture and checking process.
Where traces overlap, labels are incomplete, or a curve crosses a track boundary, document the question and the evidence used to resolve it. Do not select a line merely because it creates a smooth result. The objective is faithful recovery of the visible record.
6. Capture curves and scoped metadata
Requested curves are traced or sampled over the agreed intervals. Header information and other metadata are captured according to the approved scope. Standard naming and missing-value conventions help the archive remain consistent, while source references allow reviewers to understand what the normalized fields represent.
Choose the sampling increment for the intended use and the information available in the source. A finer numerical interval creates more rows, but it cannot recover detail that the printed line or scan does not preserve. Discuss any resampling requirement explicitly rather than treating the smallest increment as automatically the best result.
Preserve distinctions between an absent value, an unreadable interval, and a genuine zero. If a gap is represented by the agreed null convention, retain the explanation needed to interpret it. Interpolation across missing evidence should not be assumed unless it is separately specified and clearly identified.
7. Compare the digital result with the source
Visual comparison checks whether the recovered curves follow the evidence after calibration. Replotting the numerical data against the source can reveal shifts, scale problems, omitted sections, and unexpected transitions. Review difficult intervals as well as ordinary ones; a match in a clear section does not establish the quality of every page.
Numerical spot checks provide a complementary view. Select identifiable positions and compare depth and curve values with the printed scales. The acceptance plan should define how coverage and tolerances will be assessed for the particular source quality and intended use, rather than relying on an unexplained claim of accuracy.
- Identity: Does the output correspond to the correct well, run, and source record?
- Coverage: Are the requested curves and depth intervals present or explicitly accounted for?
- Calibration: Are depth, scale direction, units, and scale changes represented correctly?
- Continuity: Are page joins, overlaps, gaps, and abrupt transitions explained?
- Traceability: Can a reviewer locate the source supporting an uncertain value or interval?
8. Check file structure and downstream usability
The final file check addresses the organization of the delivered data: headers, curve definitions, columns, depth sequence, and null values. These checks serve a different purpose from visual review. A structurally valid file can contain incorrectly digitized values, while accurate values can be difficult to use if their labels or units are wrong.
Test a representative delivery in the receiving application before full production where practical. Confirm that the expected curves appear, units are understood, and missing intervals behave as intended. Record any required import settings so the next person can reproduce the result without guessing.
Ask how production and review responsibilities are assigned for the project, including whether any checks are performed independently. Independence and review coverage should be confirmed in the agreed process rather than assumed from the general term QA/QC.
Manage ambiguity as part of the deliverable
Digitization preserves information from the supplied record; it cannot establish facts that are absent or illegible. A useful exception note identifies the affected file, curve, interval, and nature of the issue. It also records whether the question was resolved, accepted as a limitation, or referred back to the customer.
For instance, if two possible traces cannot be distinguished across a damaged interval, the useful outcome may be a documented gap. Filling the interval with an unsupported line would create apparent completeness at the expense of traceability. Reviewers should be able to recognize those limitations when using the data later.
Approve a pilot and control the final handover
For a varied archive, a pilot batch can test the agreed naming, sampling, metadata, and exception rules. Include representative difficult material and identify who will approve the result. Update the production specification when the pilot reveals a requirement that was not clear at the start.
Agree the final delivery package, including digital files and any source index, exception notes, or review summary required by the project. Keep delivery revisions identifiable and specify how corrected files replace earlier versions. The handover should make it possible to understand both what was digitized and what remains uncertain.
To begin a historical well-log project, contact GAEA for a sample assessment and quotation. Describe the archive, the curves and intervals required, the destination software, and the intended use. Clear scope, source-based checking, and explicit treatment of uncertainty provide the foundation for reusable historical data.