Lessons From Karst vs Volcanic Sump Survey Deployments

Two Different Caves, One Hydrophone

Sheck Exley and Jim Fulghum's 1983 Atlantida expedition in Lanzarote mapped the first major submerged lava tube — a completely different animal from the Yucatán cenote systems that dominate karst cave-diving lore. Limestone solution caves grow by chemical dissolution along bedding planes, creating irregular passages, halocline transitions, and varied cross-sections. Volcanic lava tubes form uniform-diameter tunnels as molten lava drains from a flow, leaving smooth-walled, geometrically regular conduits. The National Cave and Karst Research Institute's cave-type taxonomy distinguishes the two morphologies at the level of first principles.

What matters for a sound-driven survey is that "water-filled cave" is not one acoustic environment. The USGS analysis of pseudokarst and volcanic caves separates limestone solution caves from volcanic tubes on lithological grounds; Halliday's pseudokarst taxonomy in JCKS refines the volcanic subclasses further. Each category produces a distinct acoustic fingerprint: karst halocline layers scatter sound in ways lava tubes don't; lava-tube wall roughness is periodic in a way karst bedding planes aren't; karst percolation adds a stippled signature absent from freshly drained volcanic tubes.

Iliffe's research on underwater caves spans both lithologies explicitly — lava tubes and karst — and underscores that the biology, hydrology, and acoustic character diverge at every level. A survey tool that works in one domain and fails in the other isn't solving the problem.

The lessons from one lithology often inform technique in the other. Multi-season siphon work in karst — documented in the siphon case study workflow that the Lot basin teams refined — translates surprisingly directly to volcanic submerged tubes once the signature library is swapped. Lava-tube exploration in Lanzarote and Iceland has matured along separate technical lines from cenote work, but the underlying patch-stitching and provenance-tracking primitives are common across both communities. Cross-pollination between the two has accelerated as more divers cross-train across the lithologies.

Stitching a Cross-Lithology Quilt

EchoQuilt was designed from day one for both environments. The core patch stitching works the same, but the signature library differs:

Karst patch signatures include halocline interfaces (typically a sharp density discontinuity around 12–18m depth in Yucatán cenotes), percolation stippling in the 2-6 kHz band, and bedding-plane reflections whose spacing encodes layer geometry. Reconciliation in karst depends on identifying these structured acoustic features consistently across patches.

Volcanic patch signatures include uniform-diameter tunnel resonances (lava tubes have near-sinusoidal radial modes that karst caves rarely show), smoother wall-reflection envelopes, and occasional lithification-transition echoes where the tube's wall thickness changes. Reconciliation in volcanic environments depends on tracking these periodic signatures against a much quieter percolation baseline.

The quilt handles both by tagging each patch's lithology at ingestion. A Yucatán patch stitches against the karst signature library; a Lanzarote patch stitches against the volcanic signature library. Hybrid systems (pseudokarst with subsequent dissolution, for example) get flagged for human adjudication because the signatures won't fit either library cleanly.

Three practical patterns across lithologies:

1. Halocline vs. thermal-gradient interpretation. In karst, an acoustic scatter zone is usually a halocline. In a volcanic tube, the same apparent scatter can be a thermal gradient where cold seawater intrudes. The quilt's lithology tag tells the surveyor which to interpret, preventing the "halocline" label from getting misapplied to volcanic passages. This connects to the broader phreatic scale scaling problem, where large systems often span multiple lithologies.

2. Wall-roughness encoding. Lava-tube walls carry a periodic roughness at 8–40 cm spacing from flow striations. Karst walls carry irregular roughness dominated by differential solution. EchoQuilt's wall-roughness classifier distinguishes them and refuses to interpolate across the boundary. The classifier was trained against datasets from both Yucatán cenote surveys and Lanzarote lava-tube expeditions, and it explicitly refuses to apply karst interpolation to a volcanic patch even when the operator manually overrides the lithology tag — because misapplied interpolation across the boundary produces silently corrupted geometry rather than a visible error. The refusal is a feature, not a friction point.

3. Cross-site survey continuity. Teams running comparative surveys across Hawaii, Iceland, and Lanzarote deploy the same EchoQuilt configuration with the volcanic signature library. Teams in Yucatán or Florida karst switch the library but keep all other configuration stable. The configuration discipline matters because subtle differences in calibration or filter parameters between teams will accumulate into systematic offsets that show up in cross-site comparisons.

Deployment Lessons From Both Domains

Survey teams deploying across karst and volcanic environments consistently report three operational lessons:

Signature library switching discipline. The single most common failure mode is forgetting to switch signature libraries when a team travels from Yucatán to Lanzarote (or vice versa). EchoQuilt now auto-detects the first patch's lithology from its acoustic profile and suggests a library switch, but teams working hybrid systems still need operator discipline. A patch recorded under the wrong signature library stitches incorrectly, and the error compounds through subsequent patches. Teams that travel between lithologies on consecutive expeditions develop their own pre-dive checklists that include a signature-library verification step, and several QRSS and GUE project leaders have started running the verification as a buddy check before splash to make the discipline social rather than individual.

Calibration environments per lithology. Hydrophone calibration in karst uses a different reference chamber than calibration in volcanic tubes, because ambient noise floors differ by 6-12 dB. Teams running both types of expeditions maintain two calibration baselines per unit and track which one applies to each patch. The lava-tube planetary-analog teams' cross-analog sites work has formalized a per-site calibration routine that transfers directly to mixed-lithology cave expeditions, and several QRSS-affiliated divers participate in those analog campaigns specifically to import the calibration discipline back into Yucatán cenote survey practice.

Publication-grade cross-lithology maps. When a publishable map spans both domains — some of Iliffe's comparative biology work does, for example — the reviewer needs clear lithology tagging on every leg. EchoQuilt's audit output includes lithology at the leg level, and publication editors can verify that the signature-library provenance matches the lithology tag.

Diver kit adjustments per lithology. Yucatán cenote work typically runs sidemount because the restrictions in cenote systems make backmount manifolds awkward. Lanzarote's submerged lava tubes are large enough to favor backmount or even DPV-towed configurations because the tube diameter is uniform and rarely restrictive. The kit choice changes the diver's acoustic signature, which changes the noise mask EchoQuilt has to apply. Teams crossing lithologies need to recalibrate their personal noise mask when they change rigging configuration, not just when they change cave.

Bailout planning per lithology. Lava tubes are often single-conduit with a clear path to the surface, while karst systems are often maze-like with multiple branches. The bailout calculation differs accordingly. EchoQuilt's exit-time estimator uses the lithology tag to apply the right bailout heuristics. NSS-CDS and GUE both teach lithology-aware bailout planning at the advanced level, and the EchoQuilt estimator is intended to formalize the planning that experienced divers do mentally on every dive.

Biological-survey overlay differences. Cave-fauna distributions differ markedly between karst and volcanic systems. Anchialine pool fauna in Lanzarote — remipedes, certain shrimp lineages — show up in volcanic systems and rarely in karst. EchoQuilt patches that include incidental biological signatures get tagged differently per lithology so a researcher cross-referencing the survey for biological work knows which fauna to expect in which patches.

Volcanic pseudokarst and limestone karst produce genuinely different acoustic worlds. The quilt only works when the signature library matches the rock.

Deploy EchoQuilt Across Both Rock Types

Teams running comparative surveys across Lanzarote lava tubes, Hawaii basalt systems, and Yucatán karst — or French sump-push teams, NSS-CDS volcanic-pseudokarst projects, and QRSS members spanning both lithologies — need a mapping tool that handles both. EchoQuilt's per-lithology signature libraries and cross-domain audit tagging make multi-rock surveys tractable. Join the Waitlist for Cave Diving Survey Teams and let us know your lithology mix — the first library expansions prioritize teams working genuinely mixed environments rather than single-rock pushes. Share your active site list across both rock types (Lanzarote tubes, Hawaii basalt, Yucatán cenotes, Florida springs, French Vercors karst), your typical rig per lithology (sidemount in cenote restrictions, backmount or DPV-towed in lava-tube diameter), your unit calibration baselines for each environment, your federation affiliation (NSS-CDS, GUE, QRSS, NACD, FFESSM, FFS), and any cross-analog campaign participation.

We will scope a per-lithology signature-library verification buddy-check protocol for splash-time discipline, prepare the dual-baseline calibration tracker for hydrophones working both rock types, set up the lithology-aware bailout estimator against your local exit envelopes, and configure the biological-survey overlay tagger so anchialine-fauna and karst-fauna patches stay distinguishable in cross-domain audit output.