A seismic archive audit answers one question: what do you actually have, and how reliable is it? Operators, data brokers and new owners all need to know before they can search, sell or report on data with confidence. This guide explains how to run a practical audit, from defining scope to exporting findings for remediation.

Step 1: Define the audit scope

Before you count anything, decide what you are counting. A seismic archive audit can cover:

  • Catalogue tables — the database records that describe each line, project and survey.
  • Media inventory — the physical tapes, cartridges and digital files that hold the actual data.
  • Supporting documents — observer logs, permit files, processing reports and maps.

Most audits start with the catalogue, because that is what users search. If the catalogue is incomplete or wrong, every downstream answer is suspect. Physical media verification is a separate pass, often done by sampling rather than a full shelf walk.

Define the scope in writing. Note which schemas or tables you will examine, which media types are in scope, and whether you will audit supporting documents or only line records.

Step 2: Inventory the archive

With scope defined, run a raw inventory. This is a count, not a quality check:

  • Line count — how many seismic line records exist in the catalogue?
  • Media count — how many tapes, cartridges or digital objects are on the shelf or in storage?
  • Catalogue-to-shelf match — do the catalogue entries point to media that actually exists, and does every piece of media have a catalogue entry?

Discrepancies here are common. A catalogue might list a line whose tape was never received, or a box might hold unlabeled reels that appear nowhere in the database. Record these gaps now; you will return to them in remediation.

For large archives, run the inventory in batches by project, client or date range. A single query that returns hundreds of thousands of rows is hard to act on. Smaller slices let you spot patterns, such as one acquisition year with consistently poor cataloguing.

Step 3: Measure completeness with a scorecard

Completeness is the first quality metric. For every field in the catalogue, ask: how many rows have a value, and how many are blank or null?

A completeness scorecard shows per-column fill rates, ranked from worst to best. If the Area field is 98% filled and the Contractor field is 12% filled, the scorecard makes that obvious at a glance. Worst fields first means you see the biggest problems first.

Useful thresholds:

  • Above 95% — acceptable for most operational searches.
  • 70–95% — usable, but users will hit blank results regularly.
  • Below 70% — field is unreliable; searches on this column will miss data.

Completeness is necessary, not sufficient. Step 4 is where you test whether the filled values are right.

Step 4: Apply quality rules to find errors

A high fill rate does not mean the data is right — see why completeness is not the same as quality. Quality rules test whether values make sense. Common rules for seismic archives include:

  • Zero coordinates — latitude and longitude of (0, 0, 0, 0) means the line has no real location. The fields are filled, but the data is unusable for mapping or spatial search.
  • Blank-but-required fields — some fields, like line name or project, should never be empty. A rule that flags any blank in these columns catches cataloguing failures.
  • Impossible dates — acquisition dates in the future, or before seismic work existed in the region, are data-entry errors.
  • Orphan media — a catalogue entry that points to a tape ID that does not exist on the shelf, or vice versa.
  • Duplicate line names — two entries with the same identifier but different metadata, or two different identifiers for the same physical line.

Each rule produces a count: how many rows violate it? Sort rules by violation count to focus remediation on the biggest problems.

Step 5: Drill through to offending rows

A rule that says "3,200 rows have zero coordinates" is only useful if you can see which rows. Drill-through lets you export or view the actual records that fail a rule or have a blank in a critical column.

For small violation sets, review each row and correct it directly. For large sets, export the list to a spreadsheet for offline review or batch correction. Include enough context in the export — line name, project, media ID — so that the person doing the fix can identify the record without returning to the database.

Sampling is sometimes practical for very large violation counts. If 40,000 rows fail a rule, reviewing 200 at random may reveal whether the problem is systematic (every row has the same bad value) or scattered (many different causes). Systematic problems are often easier to fix in bulk.

Step 6: Export findings for remediation priority

An audit produces a list of problems. Remediation is the work of fixing them. To hand off findings to the team that will do the fixes:

  1. Export the scorecard — a summary of fill rates by column, ranked worst first.
  2. Export rule violations — counts per rule, with drill-through lists for each.
  3. Tag priority — which problems block searches? Which affect billing? Which are cosmetic?

Priority is a business decision, not a technical one. A field that is 50% filled might be more urgent than a field that is 30% filled, if the first one is used for client-facing searches and the second one is internal metadata.

CSV exports work well for remediation handoff. They open in any spreadsheet application, can be filtered and sorted, and leave a paper trail of what was found and when.

Step 7: Re-profile after fixes

After remediation, run the audit again. This is not optional. A second pass confirms that fixes were applied correctly and did not introduce new problems.

Compare the new scorecard and rule counts to the originals. If a column was 72% filled and is now 94% filled, the improvement is measurable. If a rule that flagged 3,200 rows now flags 80, you know the batch correction worked and how much residue remains.

Re-profiling also catches regressions. If new data is being added to the archive, the audit should run periodically, not just once. A clean scorecard today does not guarantee a clean scorecard next quarter.

Frequently asked questions

How long does a seismic archive audit take?

It depends on archive size and the state of the catalogue. A small archive with a clean database might audit in a day. A large archive with decades of inconsistent cataloguing can take weeks, especially if physical media verification is in scope. Automated tools speed the process, but interpreting results and planning remediation still takes human time.

Do I need special software to audit a seismic archive?

You need access to the catalogue database and a way to run queries. Spreadsheets can calculate fill rates for small exports, but purpose-built profiling tools are faster and can apply quality rules automatically. Some archive management platforms, including WyseSearch, include quality profiling features.

What if my archive has no catalogue at all?

Then the first step is to build one. A physical inventory of media on the shelf, combined with reading tape headers or file metadata, produces the raw data. That data then needs to be structured into a searchable catalogue. This is cataloguing, not auditing. Once a catalogue exists, you can audit it.

Should I audit before or after migrating legacy tapes?

Both. Audit before migration to understand what you have and what identity information (line name, project, area) needs to travel with the data. Audit after migration to confirm that the new digital copies match the catalogue and that no data was lost or mislabeled in the process. See Legacy Seismic Tape Transcription for more on migration workflows.