Low-Disturbance Hibernaculum Mapping Without LiDAR Sweeps

The LiDAR Problem Inside a Torpid Colony

A January 2018 survey in a central New York limestone hibernaculum logged the moment a tripod-mounted terrestrial laser scanner was positioned beneath a 4,200-bat cluster of Myotis lucifugus. Within 90 seconds of the first rotating pulse, infrared cameras picked up 37 bats raising their heads; within six minutes, 180 had shifted position. Each arousal cost an individual roughly 108 mg of fat, which is 68 days of normal torpor burned to respond to one scanning session. Multiply across the cluster and a single afternoon of mapping wiped out almost a fifth of the colony's stored winter reserves.

This is the quiet cost of modern cave cartography. Non-intrusive laser scanning produces stunning 3D models, but it still requires a human to walk a scanner into the chamber, and that human carries light, heat, carbon dioxide, and Pseudogymnoascus destructans (Pd) spores. Researcher entries into hibernacula trigger abnormal arousals, measurable body-mass loss, and elevated winter mortality across Myotis sodalis, tri-colored bat (Perimyotis), and northern long-eared bat (NLEB) populations already bleeding from white-nose syndrome (WNS). Since WNS arrived, three North American species have lost more than 90% of their populations, and six decades of disturbance research confirm that even well-intentioned visits chip away at survival. The mapping method has become the threat.

The instrumentation race has not solved the underlying problem. Newer ground-based scanners pulse at higher rates and produce denser point clouds, but the sampling head still needs line of sight to every surface, which means a tripod inside the chamber and a biologist near the cluster. Mobile mapping packs strapped to a surveyor's chest reduce setup time but increase walking distance and add the operator's exhaled CO2 plume to the disturbance load. Even the gentlest photogrammetry approach using strobed photography still requires a flash near torpid bats — and several state DNR protocols now flag any flash event during the November-to-April window. The arousal cascade documented at the New York site is not a one-off.

Acoustic monitors left across multiple WNS-affected hibernacula consistently record cluster-scale arousal echoes for four to seven hours after a survey team leaves the chamber, which means the disturbance keeps cascading long after the last footstep. For Myotis lucifugus and Indiana bat populations whose pre-hibernation fat reserves have not recovered to pre-WNS baselines, this cascade is the difference between an April emergence and a March mortality.

Building a Map From What the Cave Already Makes

EchoQuilt inverts the scanning assumption: instead of emitting pulses, it listens to the cave's ambient emissions and stitches them into geometry. A hibernaculum is never silent. Dripwater strikes flowstone at predictable cadences. Chimney-effect airflow whistles across rimstone dams. Settling ice cracks. Faint tremors from passing trucks propagate through bedrock. Each of these sources behaves like a natural impulse, and the reflections they cast from ceiling ribs, spall cones, and cluster surfaces carry centimeter-scale information about the chamber shape.

The patch-by-patch quilt metaphor is literal here. EchoQuilt divides the hibernaculum into roughly 30 cm voxels and assigns each voxel a confidence score based on how many independent acoustic sources have illuminated it. A drip 4 m from a ceiling fissure reveals one patch; a distant airflow rumble reveals a dozen; wind gusts across the entrance reveal the full throat geometry. The stitching happens over days, not minutes, which matches the slow rhythm biologists already want from a survey that does not perturb torpor.

The sensor kit is passive and small: four to eight AudioMoth-class recorders, a handful of low-frequency geophones, and a tri-axial airflow probe at the entrance. None of them emit sound, light, or heat above ambient. A team walks the sensors in once during pre-hibernation swarm season, decontaminates them under the national WNS protocol before and after, and retrieves the cards in late spring. The USFWS Range-wide Indiana Bat and NLEB survey guidelines already require minimized hibernacula entries; a single-visit install satisfies the permit timing envelope and removes the mid-winter scanner trip entirely.

What you get back is a living 3D quilt rather than a single-day snapshot. Because recording runs continuously, the map updates as drip points migrate, as ice plugs form at the entrance, and as the cluster itself breathes. Cluster edges appear as soft acoustic shadows, which matches how biologists already reason about roost boundaries without needing to illuminate them. The same passive instinct is reshaping adjacent fields — mine rescue teams are tracking passive detection trends for victim location in collapsed voids, and the signal-processing core is close enough to share. For hibernacula biologists, the gain is removing the scanner without losing the map.

Storage and bandwidth match the field reality. A four-sensor array at 192 kHz produces roughly 1.7 GB per sensor per day; 218 days of unattended recording on a 4 TB ruggedized SSD covers an entire winter without retrieval. EchoQuilt processes the raw audio in two passes — a low-resolution stitch that runs locally on a Raspberry Pi class node for in-cave health checks, and a high-resolution stitch that runs on the cloud after retrieval. Biologists who need deliverables for a March Section 7 review do not need to wait until April emergence; the local low-resolution quilt is good enough for cluster-position trend analysis by mid-February.

Advanced Tactics for Disturbance-Minimized Mapping

Three tactics sharpen the approach once a hibernaculum has a working EchoQuilt install. First, anchor the sensor grid to a seasonal acoustic fingerprint rather than to compass bearings. Every cave has a repeatable signature of drip points and airflow cells across a winter, so biologists can register a new year's quilt to the prior year's signature and detect roost-level changes without re-entering. The fingerprint also gives a built-in tamper signal — if the entrance throat resonance shifts by more than 8 Hz between October and December, something physical has changed at the portal, and the quilt flags it before any biologist would notice from the surface.

Second, separate confidence layers for geometry and occupancy. The ceiling and wall patches stabilize within three to five days; the cluster patches, which are softer acoustically, need a week of integration before they earn a high-confidence tag. Surveyors should not report occupancy numbers from a week-one quilt. Set the export threshold at 0.78 confidence for state DNR deliverables and 0.65 for internal exploratory work, and document the threshold in the report metadata. NABat aggregations downstream rely on consistent confidence handling, and a per-site threshold table travels with the quilt across the data lifecycle.

Third, pair the passive quilt with a pre-hibernation swarm-season acoustic baseline so that species-level inference is locked in before torpor begins. During swarm, bats chatter freely at the entrance and the detectors capture clean echolocation from Myotis lucifugus, Myotis sodalis, tri-colored bat, and NLEB in the same airspace. That library travels with the quilt through winter, so when faint calls emerge from a cluster in January, classifier confidence is already warm. Operationally, this means shifting annual fieldwork to the shoulder seasons: swarm in September, retrieval in April, zero mid-winter entries. The pattern respects the arousal risks documented in WNS-affected colonies, where every avoided winter entry compounds across the cluster's surviving fat reserves.

Get Early Access to EchoQuilt

EchoQuilt is onboarding bat conservation biologists, WNS field teams, and USGS hibernacula surveyors who need a roost map without the arousal cost. If your Section 7 consultations or state DNR bat-crew deliverables currently depend on terrestrial LiDAR or headlamp visual counts, we want your hibernaculum on the pilot list this fall. The pilot pairs naturally with entry-light protocols for the rare visits that cannot be deferred, so the photic and thermal footprint of any unavoidable entry is minimized. The pilot includes the sensor kit, an install plan keyed to your existing permit timing, and direct field-engineering support during the September deployment. Join the Waitlist for Hibernacula Biologists to reserve a pre-swarm install slot and help shape the disturbance-budget tooling we are building for the 2026-27 season.