Oxford University Cave Club

Expedition: "Gustuteru 1996"


A small pilot experiment with GPS

Tim Guilford

Locating cave entrances in the complex karst of the Picos is difficult enough. Relocating them can sometimes be well-neigh impossible. The deeply dissected, fractal nature of the landscape makes descriptions of entrance locations hard to interpret accurately to all but those who know the area very well. In an attempt to explore new methods of logging and relocating entrances, a small pilot study was conducted using a Garmin 38 GPS navigational aid.

In principle the advantages of using a global positioning system are that new entrances can be logged quickly and in a uniform format by prospectors, and relocated simply without any specialist area knowledge. Positions can also be interpreted on a topographical grid map, although GPS is probably not a good substitute for a properly surveyed fix in this regard. The Garmin 38 is one of a recent generation of hand held units, and cost me under œ200. The software is very friendly, and will allow simple display of track and distance to a logged position from any known point (e.g. camp), or current position when out looking for a cave entrance. Such features could also in principle be useful for navigating to a well known point in poor visibility, and for locating the positions of entrances known only from grid coordinates.

In terms of price, size, and potentially useful features, there is no doubt that GPS technology has now reached the stage at which it is a viable tool in cave exploration. The key question, then, is whether it is accurate enough, and consistent enough, for that potential to be realised. Theoretical limits suggest a horizontal accuracy of 10-15 metres, with occasional rogue readings up to 100 metres out. One cannot be sure, however, that such limits are achievable under the topographical and climatic conditions of the field. We therefore performed a small pilot experiment on the 1996 expedition to gauge the in-field consistency of a Garmin 38 GPS.

Readings were taken on six different days from a single position, the snow pole at top camp, a point enclosed substantially by low ridges on two sides, and a high peak (Verdelluenga) on a third. Nevertheless, it was normally easy to locate 6 or more satellites even under these conditions, and an initial fix took only a few minutes (subsequent fixes can be gained in a matter of seconds). Altitude estimates ranged from 1830 to 1895 metres. The known altitude is 1855m. I conclude that the GPS altitude estimate would not be a very useful navigational tool, certainly less so than a pressure based altimeter. The horizontal position estimates also showed substantial inconsistency, particularly in the Eastings where, even with this small sample, a range of 100 metres was encountered. A higher consistency (40 metres range) in the Northings may suggest that landscape features control the direction of inaccuracies (by limiting the visible satellites unevenly across the sky), but may also simply be sampling error. Nevertheless, even these ranges would allow a lost caver to whistle contact camp in thick clag, or employ a local search strategy to locate a cave entrance from the position fix that would, under many circumstances, aid rediscovery substantially. I conclude, therefore, that GPS is now probably just consistent enough to warrant use on major expeditions in terrain where navigation is difficult. I strongly suspect that in a very few years, consistency and accuracy will greatly improve, making GPS an invaluable tool not only for navigation but also for surveying.