The aerial photos of road crossings reduced the number of ground targets required.

Water has always been an important resource for Utah County. Located south of Salt Lake City, near Utah Lake, the region was once dominated by agriculture. As early as 1865, groups of farmers and private investors began forming the first irrigation districts to harness water from the Provo, Weber and Duchesne rivers and convey it through canals to the fertile farmland in the valley and outlying areas.

One canal originating from these efforts was the 22-mile-long Provo Reservoir Canal (PRC), also known as Murdock Canal, which carries water from the Provo River at the Murdock Diversion Dam located at the mouth of Provo Canyon to farmlands and communities in northern Utah and southern Salt Lake counties. A feature of the much bigger Provo River Project, the canal was expanded in the 1940s to carry up to 550 cubic feet per second (cfs) of water. Today, the canal is operated and maintained by the Provo River Water Users Association (PRWUA), a private nonprofit corporation, under an agreement with the Bureau of Reclamation. While it continues to provide agricultural irrigation, increasing urbanization in the Utah County communities through which the canal meanders--including Orem, Lindon, Pleasant Grove, American Fork, Highland, Alpine and Lehi--is changing water use and quality requirements.

Randy Miller makes final preparations to traverse into the 8-foot-diameter Provo Reservoir Canal siphon under I-15 at Point of the Mountain in Salt Lake City, Utah.

Several years ago, faced with growing municipal and industrial needs as well as safety and environmental concerns, the PRWUA requested that Reclamation authorize the Provo Reservoir Canal Enclosure Project, which would enclose the entire canal in a pipe or box culvert. The canal right-of-way grade would be restored leaving the pipe or culvert buried and unexposed. As part of the project, a nonmotorized trail would be constructed in the PRC right-of-way that would connect Utah Flight State Park (a launching place for hang gliders and paragliders) to the mouth of Provo Canyon and 14 miles of existing nonmotorized trail along the Provo River.

Reclamation authorized the changes, and the PRWUA hired CH2M Hill, a global full-service engineering, procurement, construction and operations firm, to manage the project.

A Topcon HiPer Lite + static GPS setup with snow-capped Mount Timpanogos in the background. Photo by Randy Miller.

Gathering the Data

In October 2008, CH2M Hill retained Aero-Graphics Inc., a full-service photogrammetry firm based in Salt Lake City, to verify and supplement surface topography and utility data from a 1997 survey. The two firms had worked together successfully on previous projects, and aerial photogrammetry was considered an ideal technique for the canal enclosure project because it would provide the ability to take measurements in restricted, private, hazardous, inaccessible or intermittently accessible areas. To provide redundancy and quality control, Aero-Graphics crews conducted ground survey work concurrent with the aerial data imagery capture.

Using plans developed in the office before the survey work began, a two- person crew set 11 aerial targets in the canyon and Murdock Diversion Dam areas. A combination of real-time network/real-time kinematic (RTN/RTK) GPS, static GPS and digital leveling techniques were used to derive coordinate values for the targets that would control the largest stereo model area in the project. The surveyors made static GPS observations on four of the 11 targets in the lower Provo Canyon area to verify the horizontal positions derived from RTN/RTK techniques. Where good elevations were critical to the accurate outcome of the photogrammetry process, the team ran differential levels as a check on the RTN/RTK values.

The Aero-Graphics plane and crew.

The RTN/RTK values for elevation were derived using the Geoid-2003 (GEOID03) file--a method that does not always produce trustworthy vertical values. Given the proximity of some of these values to some significant geologic masses, John Francis, PLS, senior surveyor, determined that making a level run up the middle of the model and turning through at least five targets and NGS station LO0500 was essential. “The mouth of Provo Canyon is a deep gorge,” Francis explains. “If there was ever a place to be skeptical of the geoid undulation, this was it.”

The crew used Topcon HiPer GPS receivers with internal memory set to 5-second intervals to gather the raw observations needed for OPUS solutions and GrafNet post processing. The static observations were made to provide a check on the RTN and differential level values as well as a check on the 1997 survey data. The team used a Leica model DNA10 digital level to perform differential leveling and a Topcon HiPer GPS receiver with Airlink modem and Survey Pro software by Tripod Data Systems to measure control points from the 1997 survey.

“Coming up with an adequate scale factor to match a control net that was run primarily with a total station in 1997 was a bit of a challenge,” Francis says. “We had to make a judgment call.” The team had to choose a scale factor that fit the most control points within an acceptable limit and reject the original coordinates based on current position likely due to surface motion. The resulting combined adjustment factor used to go from state plane to local ground was 1.000285814 (derived empirically rather than from a mean project elevation, which is the norm).

Lowering equipment down the ice-covered east entrance. Photo by Karl Jensen.

The team used the same RTN rover arrangement to gather hard surface data at most crossings. Since the canal was originally a Bureau of Reclamation project, several Reclamation bench marks existed in headwalls and gauging stations along the canal alignment. The survey team used the RTN rover to check into several of these bench marks using the GEOID03 model and made systematic corrections to RTK corrected measurements from the real-time network administered by subscription through Rocky Mountain Transit Instruments in Salt Lake City. Corrected vertical elevations differed from stamped elevations by an average of only 0.04 feet with minimum/maximum horizontal differences of 0.000 feet and 0.195 feet, respectively. “The 0.195-foot value was more than I wanted to see, but only one value was this far out, and I attributed the difference to monument stability and ignored the published value for that one station,” Francis says. “Other horizontal differences were less than 0.08 feet--not first order but within a tolerance we could live with for designing for large-diameter pipe placement.”

The team imported the ground survey data into commercially available stereo compilation software. These data were used to calibrate the stereo compilers to true ground and ensure accurate feature identification (water valves, hand holes, sewer and storm manholes, etc.) during the aerial survey.

The aerial crew used an UltraCam X digital camera by Microsoft Vexcel Imaging to collect aerial photos of road crossings. The digital technology provided superior image quality and reduced the number of required ground targets compared to traditional film-based cameras, and it also provided near-infrared frequency acquisition. (These data could be used after construction is complete to determine the differences in surface water saturation before and after the canal was enclosed.)

A POSAV 510 IMU by Applanix aboard a Cessna Stationair T206 was used to capture the imagery and correct camera geographic positions for each exposure frame. The RTN corrected rover measurements on photo-identifiable objects such as parking stripes and water valves. These points were used as quality control to validate and support the tightly coupled GPS/IMU solution used initially to orient the acquired images. “Airborne GPS and inertial measurement unit values are further refined by the availability and reliability of existing aerial targets or photo identifiable points,” says Casey Francis, vice president. “The IMU monitors and records accurate crab, pitch and roll angles of the camera platen in real time. Traditionally, these values were calculated in aerotriangulation, and for many applications, this is still the case. Using a tightly coupled GPS/IMU solution provides a highly accurate basis for aerotriangulation. In some cases, particularly high-altitude orthogonally rectified photography, the ground control requirement can be completely eliminated--the IMU values are that trustworthy.”

Using both ground surveying and photogrammetry allowed the team to verify the accuracy of each technique and ensure the quality of the data. “If any data were lacking in the ground survey, we could fill it in using photogrammetry,” Casey Francis explains. “Looking at survey points in the stereo model allowed us to safeguard against potential blunders or out-of-tolerance measurements in either the ground survey or compilation processes. Collectively, this process was very effective in ensuring a reliable product was delivered for each critical phase of the project.”

The Leica robotic total station perched in shoes at the 4:1 to 2:1 slope transition. Photo by Karl Jensen.

Traversing on a Slippery Slope

In January 2009, the CH2M Hill project team asked Aero-Graphics about the feasibility of taking measurements within an 8-foot-diameter siphon pipe under Interstate 15. “We had lots of questions and concerns about the access and safety issues for an undertaking like this,” says Karl Jensen, CP, data acquisition manager. “Our primary concerns were about ventilation, slippery slopes, and possible debris or water accumulation in the low points of the tunnel.”

Aero-Graphics contacted the canal maintenance team to discuss the safety issues. “They indicated that they enter the siphon annually for cleaning and inspection and that ventilation was not a problem aided by daily strong winds at the surface,” Jensen says. They were right--an incessant breeze blew through the tunnel on the day of the fieldwork. A canal maintenance team accompanied the survey crew and rigged ropes to assist the team in ascending and descending the pipe’s smooth, treacherous slopes. “The first descent from the outlet toward the inlet dropped 10 feet on a 4:1 slope followed by a 50 foot vertical descent on a 2:1 slope−scary,” Jensen says.

The team had originally considered using its prism-based total station to make the siphon measurements but decided that shuffling a backsight in the limited space would be too time consuming. Instead, the surveyors rented a Leica TCRP 1201 1-second robotic total station and an Allegro data collector with Carlson software from Rocky Mountain Transit Instruments. “Renting a robotic [instrument] turned out in our good fortune in a big way,” Jensen says. “We had headlamps and flashlights, but the amount of illumination they provided was much too small for an instrument operator to sight a prism/pole. Fortunately, a robotic uses a beam of sufficient intensity to track a prism, even in the pitch dark.”

The team traversed in from west to east on control stations established by the RTN rover. Maintaining sure footing was crucial. “While standing at the top of the first 2:1 slope, I let a 2-foot piece of a two-by-four slide down the slope and watched it pick up speed then stop at the toe,” Jensen says. “The maintenance crew had already rigged the rope and ensured that it reached the toe. Holding the rope was advisable for the descent but essential for climbing back up.”

The team set a control point a comfortable distance back from the 4:1 to 2:1 slope transition where the total station had line of sight to the bottom but could still rest on the 4:1 slope without slipping. Average snowfall and moderately cold temperatures added an element the surveyors had not counted on--ice. Perched at the precipice, the team used a pair of boots to give the tripod some stabilizing friction on the walls of the pipe. “After taking the backsight shot, we had to determine how we were going to physically pass the total station while holding the lifeline without sending the station on a ride to the bottom,” Jensen says. “We gingerly contorted under and around the tripod, descended the 2:1 slope, set and measured a control point, recovered the total station--and breathed a sigh of relief to have the total station off of the slope.”

Post-processing work included converting the GPS baselines from state plane to local ground and importing total station data into AutoCAD Civil 3D 2009 to make the coordinate geometry calculations. The points were then exported for inclusion with the other local ground coordinate files. The final deliverable for the data was a .csv file in PNEZD format (point number, northing, easting, zenith and description).

Randy Miller prepares to collapse another section of ice shelf. Photo by Karl Jensen.

Precise Planning

Construction on the canal is scheduled to begin in 2009 and will take place in three to five segments each winter while the canal is not in use. When finished, the enclosure is expected to save roughly 8,000 acre feet of water per year by reducing seepage and evaporation loss. These water savings will increase instream flows in the Provo River, which will benefit the June Sucker Recovery Implementation Program. Other benefits of the canal enclosure include enhanced public safety, improved water quality, and greater efficiencies in the operation and maintenance of the canal corridor.

Prior to the survey, CH2M Hill suspected that the construction plans from the 1970 retrofit differed substantially from what was actually built--particularly a 573-foot radius horizontal curve that appeared to go left on the plans but right in actual construction. The survey data confirmed this fact. “Construction plans will now accurately reflect not just the location of the siphon but, more importantly, the quantity of welded-in-place steel liner that will be required to reinforce the siphon for an increase in pressure,” Jensen says. “The ability to capture the precise geometry of the underground siphon is a big money saver for the construction project.”