Texas firm learns the many phases of a cave survey.

Until recently, few people knew that Texas harbored one of the largest and most diverse bodies of rock art in the New World. In fact, the Lower Pecos region of southwest Texas contains one of the most well-photographed pictograph sites known as the White Shaman. On its walls is encapsulated the basic religious principles of the prehistoric inhabitants of the region. The flight of the shaman to the land of spirits and his metaphorical death and rebirth are a message from the past about humanity’s quest for solutions to life’s mystery.1 Faced with the realization that the art was deteriorating at a rapid pace, The Rock Art Foundation (San Antonio, Texas) was formed to promote the conservation and study of the Native American rock art in the Lower Pecos region. After suspecting erosion of one of the rockshelters, The Rock Art Foundation commissioned the Texas Department of Transportation (TxDOT) to examine whether erosion had taken place inside the cave site. TxDOT in turn called on the engineering and surveying firm CDS/Muery Services and laser scanning firm 3D Laser Technologies Inc., both of Houston, Texas, in January of 2002 to begin the site survey. The rockshelter, upstream from the State Highway 90 Pecos River Bridge, is based on previously recovered and observed diagnostic artifacts and is believed to have been intermittently inhabited by a semi-nomadic people over the past 4,000 years.

The site is referred to as a cave, but the configuration is more of an overhang approximately 66 meters wide with an average depth of 12 meters. This type of geologic formation is the result of a solution cavity formed by water-eroding soft limestone. The majority of the cultural deposit within the cave appears to be intact. A major concern of The Rock Art Foundation is that suspected erosion, possibly caused by the bridge approach slope or some other undetermined cause, may be impacting cultural deposits within the cave and will continue to do so if not corrected.

The Scope of Work included verifying the relative position differences on existing control surveyed by other firms in the past and performing a topographic survey of the cave. We used a Cyrax 2500 laser scanner (Cyra Technologies, San Ramon, Calif.) to acquire a dense point cloud of the cave features, which was then used to create a 1 meter contour map based on the published values of the control stations found at the site and generated a comparison between a survey dated May 28, 1987 by McGray & McGray Surveying of Austin, Texas. This project was also a pilot survey to determine if the use of a three-dimensional laser scanning system was a viable option for this type of work.

Control checks were made to single prisms set on tripods with tribrachs where space permitted. Prism poles with bipods were used in other areas. Vertical determinations were made with trigonometric leveling techniques with a Trimble (Sunnyvale, Calif.) TTS 300 conventional laser total station. A number of shots were taken directly on the rock face of the cave site with the TTS 300 to provide quality control checks once the laser scan point cloud had been fully registered and placed into the horizontal and vertical reference frame of McGray & McGray’s survey. There were numerous scans taken around the cave site in order to obtain complete three-dimensional coverage of the survey area. In order to register these scans together into one common scene, scanner targets were placed in each scene. The horizontal and vertical positions of the scanner targets were determined by using the TTS 300.

Phase 1: General Requirements

The survey was tied to the existing control using the published horizontal and vertical values of the control points that TxDOT provided us. The horizontal datum used is the North American Datum of 1983, 1993 Adjustment (NAD83/93). The projection system used is the Texas State Plane Coordinate System, South Central Zone. The vertical datum was not yet confirmed but believed to be NGVD29. Meters (m) were the designated unit of measure.

Existing Control Verification
Conditions in the cave site area did not lend themselves to exacting techniques. Light levels were low and the soil had the consistency of powder. Control Point 2 (CP2) was a cotton spindle set under the cave overhang and had some movement to it. CP4 was a TxDOT aluminum cap set in rock 448.495 meters west of CP2. The maximum distance observed in the cave site was 31.373 meters to CP40 creating unbalanced sighting conditions. We used face 1 and face 2 observations to minimize the errors. The average target difference between positions obtained by conventional ties compared with our scanned position is 0.013 meters. A table depicting differences between published control values versus checked control values is presented in the table.

These results indicate an average difference of 0.031 m for the north coordinate and 0.030 m for the east coordinate. The average vertical difference was 0.026 m. These checks were of sufficient quality for the purpose of this survey.

Scanning the outside of the cave from the roadside.

Phase 2: Reconnaissance and Planning

The project team met in Del Rio, Texas, on Jan. 21, 2002. The team included Jimmy Cruse, Steve Schmidt and Michael Barker of the Austin TxDOT office ISD department; Allen Bettis of TxDOT’s Environmental Affairs Division; Corky Kuhlman of the Texas Parks and Wildlife (TPW); Alton S. Geisendorff Jr. of CDS/Muery Services Houston Office; and Dave Ross and Chance Alford of 3D Laser Technologies Inc.

We went to the project site after lunch and started the evaluation of the logistics of the survey. CDS/Muery Services and 3D Laser Technologies Inc. were confident that the scans and survey could be accomplished. The unknown factor was the transporting of the equipment to the cave site. The scan unit and the battery weighed 27.8 kg (61 lbs.) with an off-center weight distribution. This, along with the scan targets and conventional surveying equipment, presented a situation that resembled a trek to Mount Everest without the Sherpas. After exploring several different approaches, the route decided on was a direct descent down a 45˚ slope to a draw that dropped into a steeper descent to the final approach to the cave. The final approach was uphill across boulders through mesquite trees. Michael Barker of TxDOT is a trained climber and provided the necessary rope configuration that allowed us to rappel down the slopes.

Alton S. Geisendorff locates the scan targets at the south end of the cave.

Phase 3: Field Survey

On January 22, we arrived at the parking lot on the east side of the Pecos River Bridge at 7:30 a.m. All project members began to sort and inventory equipment to ascertain that return trips would be minimized considering the difficulty of the trip into the cave. Meanwhile, Mike placed the ropes for the descent. We then began the descent with all project members shouldering packs or strapping on tripods. Each climber also brought water and snacks since coming out for lunch was not an option. Officially, TxDOT and TPW representatives were observers but willingly became equipment bearers. The trip into the cave took about an hour.

Once there we divided into a scanning team and a conventional surveying team and began the surveying operation at 9:30 a.m. The surveying team checked the existing cave site control while the scanning team began their work in the south end of the cave. Scan targets placed by the scan team were surveyed after the control was verified. This procedure was carried from the south end of the cave to the north end. During this time, Allen Bettis of TxDOT conducted an archeological assessment of the cultural deposits and documented some amazing archeological features and artifacts such as rock etchings, known as pictographs and petroglyphs, and stone tools that had been used by the ancient inhabitants. The operation proceeded smoothly until 3:30 p.m. when the Dell laptop computer battery power indicator showed 37 minutes of operating time left. The database, which contained all of the previous scans, became a “read only” file and a moment of despair swept over the team. We decided to leave the site and conduct a scan from the roadway above the cave with a new battery since that area was needed to complement the cave site survey. After an arduous, hour-long ascent, we completed a scan of the adjacent cave site slope and made the necessary target ties to relate this scan to the interior cave site sight scans. We left the site at 6:30 p.m. and re-grouped at a restaurant to relax and review the day’s findings.

The general consensus was that TxDOT and TPW had seen enough to convince them that this was a viable procedure. During the day we had shown several individual 3D scans to the project team and they were impressed. If the database was lost, TxDOT did not feel it was necessary to duplicate the survey but was hopeful that we would be able to furnish a contour map as originally planned. The general consensus among CDS/Muery Services personnel and 3-D Laser Technologies Inc. personnel was that we had set out to deliver a complete project, so we would repeat the procedure.

On January 23 we went back to the cave area at 7:45 a.m. Everybody knew the routine and proceeded to load up equipment while Mike put the ropes back in place. The enthusiasm levels were a little lower, but we knew we had a job at hand. We carried out the same procedures as the day before and all the necessary information was obtained by 3:30 p.m. Extra laptop batteries were carried on this trip and we monitored the computer power levels closely. All was not lost from the proceeding days mishap, as TxDOT personnel were given an extra day to assess erosion damage and artifact findings and CDS/Muery Services and 3D Laser Technologies Inc. personnel were able to refine procedures and further evaluate this new tool. We left the site at 5:00 p.m. and once again all team members met for a final dinner gathering to relax and discuss the possibility of including 3D scanning as a realistic data gathering tool for everyday applications.

Michael Barker preparing ropes for the descent into the canyon adjacent to the cave.

Phase 4: Data Processing

The processing of the data began on January 28. The goal was to develop a 3D presentation drawing and a topographic map for erosion control design. This involved resolving the data in the scanning software, Cyclone to obtain a format that would be usable by Land Desktop 3 software. Difficulties can arise from not having enough information in almost all topographic surveys; the opposite is true for laser scanning. Surplus information needs to be decimated to obtain a manageable file size. Scanning software such as Cyclone is designed to handle large amounts of points. The objective, however, is to shape this information into something that technicians can proceed with as with any other project, using the software for which they’ve been trained. We delivered the final product to TxDOT in March.

A Lasting Impression

3D scanning combined with conventional surveying is a relatively new approach to the everyday survey techniques used by the surveying and engineering profession. Many articles have been written in professional journals about the use of this technology in roadway design and industrial retrofits over the past year or so. It is a specialized survey tool that needs to be balanced with a client’s needs and the cost of its use. It is still an “exotic” piece of apparatus that is rapidly evolving into a useful tool as we discover more and more applications for its implementation. We at CDS/Muery Services are continuing to explore the possibilities of this technology and look forward to being involved in other projects as challenging as the Pecos River Cave Site.