How underwater caves are surveyed and studied

Every length figure in a cave list, every map in a dive shop and every scientific sample from a flooded passage depends on survey. This article explains, at an introductory level, how underwater caves are measured and documented, what modern tools add, and how survey data are shared or deliberately kept private. It describes methods; it is not training, and survey diving is taught in dedicated courses.

Why survey at all

A survey turns a series of dives into a map that others can use. Hydrologists need maps to understand where groundwater flows, archaeologists need them to place finds in context, and land managers use them to decide what to protect. The Quintana Roo Speleological Survey (QRSS) describes its archive of survey data as research by explorers, cartographers and scientists committed to conservation and scientific documentation ^[8].

The classic line survey

Underwater caves are usually surveyed along the guideline, the continuous line that divers lay from the entrance for navigation. The U.S. Deep Caving Team (USDCT) described the traditional method used at Wakulla Springs in 1987: a knotted line is tied to rock projections at each bend; at each tie-off the surveyor counts the knots, spaced about 3 m apart, records the depth and takes a compass bearing (azimuth) along the line to the next tie-off, and notes estimated passage width, height and floor sediment on a slate ^[1]. Each tie-off becomes a survey station, and the string of distances, bearings and depths between stations is the backbone of the map.

The same account names the method's weaknesses. It is slow, which matters when every minute at depth adds decompression time, and every measurement carries a small error. Because each station is positioned from the one before, these errors accumulate, so long penetrations can carry large errors at their far ends ^[1]. Software developer and cave instructor Sebastien Kister lists the usual culprits: parallax when reading a compass, misjudged line lengths and mistakes in copying numbers from the slate ^[4].

Back on land, the numbers are processed into line plots and maps. Free programs such as Survex, an open-source package designed for large and complicated cave systems ^[12], and dedicated cave-diving tools such as Ariane, which can build stick maps viewable in Google Earth ^[3], do this work.

Resurveys

Because errors accumulate, mature systems are often resurveyed. The QRSS reported in 2022 that a resurvey of the Blue Abyss area of Dos Ojos confirmed earlier survey work while finding about 340 m of new passage ^[9]. CINDAQ's 2025 report describes a resurvey of Sistema Ox Bel Ha that by then had re-measured almost all of the original 245 km of mapped passage, alongside more than 185 km of new exploration, and that integrates the results in a geographic information system (GIS) ^[7]. Resurveys like these are how long-cave figures become more reliable over time.

Electronic survey tools

Several tools aim to make survey faster and less error-prone. The MNemo, developed by Kister, is a small device that a diver clips onto the guideline; swimming at normal pace from one tie-off to the next, it records the length of line covered, the depth or tilt at the start and end, and the heading, and the data load directly into Ariane ^[3]. In a comparison on the maker's website, the first section of Ralph's Cave on Abaco was surveyed both by hand and with the MNemo, with a difference of 1.8%; this is the maker's own report rather than an independent test ^[4].

The 1998 Wakulla 2 expedition went further with its Digital Wall Mapper, a unit attached to a diver propulsion vehicle with 32 sonar transducers measuring distances to the walls, depth sensors and an inertial reference unit tracking position and orientation. Four times a second it fired its sonar and logged the readings, producing a three-dimensional point cloud of the passage ^[1]. To limit drift, the team used induction radio to fix waypoints inside the cave to surface coordinates; their calculations aimed to keep error within about 1 m per kilometre if a control point was set every 500 m ^[1].

At extreme depth, conventional survey may not be possible at all. Writing about the Pearse Resurgence in New Zealand, explorer Richard Harris noted that navigation data from the team's scooters, although not designed for mapping, gave useful information on the depth, direction and length of passages where traditional mapping could not be done ^[14].

Tying the cave to the surface

A cave survey is most useful when it can be placed accurately on a surface map. Radiolocation does this with low-frequency transmitters. The Bahamas Caves Research Foundation (BCRF) describes placing self-levelling transmitters on the ceiling of Ralph's Cave on Abaco, recording each one's depth and distance from the entrance, then tracking the signals on the surface with a receiver and taking GPS positions directly above them. The aim was to show the footprint of the passages beneath a proposed conservation area ^[5]. CINDAQ also records entrance positions with satellite positioning (GNSS) ^[7].

Imaging and 3D models

Photographs and video are now survey tools in their own right. At Hoyo Negro, the project team has used structure-from-motion photogrammetry, which builds 3D models from overlapping 2D photographs, despite total darkness, a complex shape and limited bottom time. Models at the scale of the site, individual features and single bones support anatomical and taphonomic study and let researchers annotate the imagery remotely before deciding what to sample ^[6]. The 2014 study of the Hoyo Negro skeleton noted that most of its information had come from video, photography, minimal sampling and 3D modelling ^[16]. CINDAQ has added mobile photogrammetry at dry entrances and shallow-water ground-penetrating radar in cenotes ^[7]. In the Czech Hranice Abyss, the local caving group is drawing plan maps of the dry and flooded parts at 3 m depth intervals and combining them into a 3D model ^[13].

Science as part of the dive

Survey dives often double as research dives. Divers measure water layers, collect water and rock samples and record the position of features. A 2021 dataset of submerged stalagmites from around Tulum, for example, records each sample's elevation measured with a dive computer to a stated precision of 0.1 m ^[17]. Equipment choice can matter too: a study at Bundera Sinkhole in Australia found that rebreathers disturbed the cave's layered water less than open-circuit scuba, although both had measurable effects ^[15]. Collecting samples is regulated. The QRSS points out that collecting biological material in Mexico requires a federal permit ^[18], and the Hoyo Negro team recovered bones under permits from Mexico's archaeology council ^[19].

Who owns survey data, and what gets published

Survey data generally belong to the teams who collect them, and they choose what to share. The QRSS archive depends on explorers contributing their data, and the survey says it does not reveal cave locations or where unexplored leads might be ^[8]. It also notes that only about 30 underwater caves or systems in Quintana Roo had been represented by complete maps of any kind, most of them outdated ^[8]. The National Speleological Society's conservation policy, adopted by its Cave Diving Section, says the society will oppose publishing cave locations where there is reason to think it would lead to vandalism before protection is in place, and asks that survey markings be few, small and removable ^[10].

Some projects publish openly. The USDCT posts plan and profile maps from its 1987 Wakulla expedition on its website ^[2], and the Hranice group has placed a profile of the abyss on an information board at the viewpoint ^[13]. Others share maps only with members or partners; the NSS Cave Diving Section's history notes that maps created by its explorers were made available to members ^[11]. For readers, the practical consequence is that a missing map is often a deliberate choice that protects a cave or respects a landowner, not an oversight.