Audit Protocols for Publishable Cave Maps From Sound Data

From Field Audio to Peer-Reviewed Cartography

The British Cave Research Association defined its two-part grading system — BCRA grades 1a through 6d — precisely because "this map looks right" was not good enough for publication. Grade 5c means compass readings accurate to ±1°, clinometer readings to ±1°, and distances to ±0.2 m, with detailed passage cross-sections. The UIS mapping grades technical note formalizes the cross-federation mapping standards that govern publication in the NSS Journal of Cave and Karst Studies and peer journals worldwide.

What BCRA doesn't spell out — and what turns out to be the harder problem — is how to audit a map derived from passive acoustic data. A traditional survey leg comes with a surveyor's notes, a measured tape, and a compass reading. An EchoQuilt patch comes with 8 minutes of hydrophone recording, 10 Hz IMU telemetry, and a derived geometry. The IntechOpen analysis of cave data in GIS makes the point explicitly: documentation is more valuable than the map itself. Without audit-grade provenance, a sound-derived map can't meet Grade 5c, no matter how accurate its underlying data.

BCRA's canonical reference on cave surveying frames survey grades as a publication prerequisite. PubMed's literature on promoting open research culture extends the principle to transparency standards. EchoQuilt's audit protocol was designed so a sound-derived map could be held to the same standard as a traditional compass-and-tape survey — auditable leg by leg, provenance chain intact. The grade-locked audit fits directly into the federated data sharing protocol used by QRSS, NSS-CDS, and other multi-team consortia, so a publication that draws patches from multiple federations carries a consistent grade interpretation across all contributing teams.

Stitching Audit-Grade Provenance Into the Quilt

Every patch in EchoQuilt carries a survey-grade tag that auto-computes from six inputs:

1. Hydrophone calibration age. Patches recorded within 30 days of calibration qualify for Grade 5 or higher; patches recorded up to 90 days earn Grade 4; anything older drops to Grade 3.

2. IMU heading confidence. The inertial unit's heading error grows with time since the last GPS fix. Patches recorded within 8 hours of surface fix can earn Grade 5c; longer elapsed times downgrade the leg.

3. Acoustic signal-to-noise. Patches with SNR above 18 dB on the primary flow band qualify for Grade 5; patches below 12 dB cap at Grade 4a regardless of other inputs.

4. Reconciliation agreement. If the patch conflicts with neighboring patches beyond the grade's spatial tolerance, the patch downgrades until the conflict is human-adjudicated.

5. Instrument provenance. Patches recorded on calibrated, serialized equipment with full maintenance logs qualify; anything recorded on untagged equipment caps at Grade 3.

6. Diver certification level. Patches recorded by divers with documented training on the quilt protocol qualify for higher grades; patches from untrained divers cap at Grade 4.

These six inputs compose into a per-leg grade visible in the audit dashboard. When a team submits a cenote map for publication, the reviewer sees every leg with its assigned grade and can drill into the provenance chain behind any individual leg.

Three operational patterns for publication workflow:

Pre-submission audit sweep. Before any map goes out for peer review, the team runs EchoQuilt's audit sweep. Legs below the target grade get flagged for re-survey; legs at or above the target get batched for publication. This workflow connects directly into deco-stop QC — decompression stops are where the audit-grade re-surveys happen without consuming push time.

Federated peer review. When multiple teams contribute to the same system map, the audit grade runs across all contributing teams. A system that's Grade 5c on one team's patches and Grade 3 on another's publishes at the lower grade unless the federation re-surveys the weaker sections. Reviewers see the contribution graph and can drill into which team's patches drove which grade decision.

Cross-domain audit alignment. Mapping standards in adjacent communities have converged on similar provenance requirements, and EchoQuilt's audit schema is intentionally portable across them.

Anomaly handling in audit grades. Some patches will always sit in a gray zone — a piece of passage where the SNR is borderline, the IMU drift is at the limit, and the diver's certification is just-current. The audit dashboard does not auto-promote these patches into a higher grade, but it does log the borderline status explicitly so a reviewer can make the call rather than the engine. This is a deliberate design choice. Auto-promotion at the borderline produces grade inflation across the federation; explicit borderline flagging keeps the grades meaningful even when individual patches are ambiguous. WKPP and QRSS reviewers have reported that the borderline log is one of the most valuable parts of the audit dashboard because it surfaces exactly the patches where judgment matters most.

Publication-Grade Tactics

Teams targeting the NSS Journal, Cave and Karst Science, or International Journal of Speleology need three operational disciplines that go beyond the automatic grading:

Pre-publication calibration certification. Within 14 days of submission, every piece of equipment used to record patches in the submitted map gets recalibrated. The certification goes into the map's appendix. Reviewers checking provenance can verify that the calibration dates are recent enough to back the claimed grades. NSS-CDS and GUE both maintain pre-publication calibration sheds at known sites where teams can run pre-submission verification without reinventing the test setup.

Reviewer-side replay tools. A modern peer reviewer should be able to load the patch-level provenance directly into their own EchoQuilt environment and replay any audio segment under question. Publication submissions include a reviewer access token that grants read-only access to the underlying audio for the review period, then expires. This is a higher transparency bar than traditional cartographic submission ever achieved, and the early adopters at the Journal of Cave and Karst Studies have noted that the replay tools change the character of the review from prose-only critique to direct evidence inspection. A reviewer can hear the JJ-CCR's solenoid noise floor and judge whether the published noise mask was applied correctly, rather than taking the author's claim on faith.

Independent replication pass. For high-grade publications (5c or above), a team member who didn't record the original patch replays the quilt and verifies the geometry independently. This is comparable to the peer-verification standard in laboratory science — the audit isn't just the dashboard score, it's a human check on the signal-to-geometry translation. Lava-tube planetary-analog teams face a parallel problem with their traverse geometry QC workflow, where mission-grade verification protocols require independent replication for any geometry that will guide robotic operations. EchoQuilt's audit schema maps cleanly across both domains because the underlying provenance question is the same.

Raw-data deposition. Publications grade-locked above 5 include a deposition of the underlying audio and telemetry streams in an archival repository. This is the FAIR "Reusable" commitment — future researchers can re-run modern filters against the original recordings and extract detail that wasn't visible at publication time. The open-research culture literature treats this as a baseline expectation for publishable science.

Publication standards written for compass-and-tape survey can apply, with discipline, to sound-derived maps. The audit is the mechanism that makes the translation defensible. Teams targeting top-tier journals should expect that reviewers will exercise the replay tools rather than skim the appendix, and they should design their data submissions to survive that scrutiny from the first patch onward.

Publish Maps That Survive Peer Review

NSS Cartographic Salon submitters, BCRA grade-5c projects, and teams targeting Journal of Cave and Karst Studies publication need audit-grade provenance on every patch. EchoQuilt tags each leg with its computed grade and exposes the provenance chain to reviewers. Join the Waitlist for Cave Diving Survey Teams and tell us your target publication venue — first audit-dashboard deployments prioritize teams with maps going to peer review inside the next 18 months. Share your target journal or salon (NSS Journal, Cave and Karst Science, International Journal of Speleology, NSS Cartographic Salon), your team's current grade ceiling per system, your CCR and DPV inventory feeding the submitted patches, your independent-replication staffing plan, your federation affiliation (NSS-CDS, GUE, QRSS, BCRA, FFS, IUCRR, NACD), and any pre-publication calibration shed access.

We will scope the within-14-day calibration certification template, prepare the reviewer-side replay token workflow that the Journal of Cave and Karst Studies early adopters already use, set up the FAIR-Reusable raw-data deposition schema for grade-5+ submissions, and configure the borderline-flag review queue so judgment calls surface to your reviewers explicitly rather than auto-promoting at the threshold.