Feature: Working Under Canopy
Professional Surveyor Magazine - January 2008
Jim Gillis, NSLS, CLS, RPLS, Joe Pileci, and Mark Strickland
While GPS has proven a panacea and taken surveying to new levels, surveyors know that it does not perform well under tree cover. Those of us who have worked with GPS from the early days—in our case, that goes back to the late 1980s—came to understand this early on.
Yet under certain circumstances, we found that once a full constellation of GPS satellites was available, it was possible to get acceptable results from static observations in light tree cover. This worked if the operator employed rigorous checks and balances to ensure that the fixed integer solutions obtained did not result from incorrect ambiguity selection. In simpler terms, we knew GPS had the potential to work under heavy canopy as technology advanced and with the development of some specialized techniques.
In the specialized field of geophysical exploration surveying, commonly referred to as seismic, it is not necessary to establish survey point positions to the accuracy commonly required for boundary, construction, or other types of surveying. Although it is preferable to know the three-dimensional coordinates of the points to an accuracy of a couple of tenths of a foot, geophysically, positional errors of less than five feet horizontal and one or two feet vertical are acceptable. This is because errors within these limits do not negatively affect the quality of the geophysical processing, and that dictates the survey needs.
As a data-gathering procedure, reflection seismic enables geologists and geophysicists to map underground strata and make educated guesses as to where hydrocarbon reserves may be located and exploited. The basic premise for the seismic surveyor is that you have two classes of points to lay out, one being the source points and the other the receivers. In 2D seismic, both source and receiver points are laid out on the same line. With 3D seismic, the points are usually laid out in some form of pattern or grid, with source and receiver points falling on separate lines that may be perpendicular to one another. A newer class of seismic has surfaced recently, 4D, that includes the time component. Here, in special cases such as where there are subsurface changes going on the client wishes to monitor, the same source and receiver points are used year after year.
How It Works in the Oil Patch
In helping to locate underground petroleum reserves, the science of reflection seismic involves an energy source, such as dynamite inserted into the ground or a huge weight dropped or shaken on the Earth's surface, which sends shock waves into the Earth. Listening devices called geophones, connected by wires or radio communication to a master recording device in the "doghouse," pick up and save the returning sound waves. This data, after processing, allows experienced interpreters to map the subsurface strata and determine in what areas petroleum resources are likely to be found. The seismic surveyor is required to locate the three-dimensional positions of these source and receiver points, and errors greater than the tolerances mentioned above will negatively affect the accuracy of the geophysicist's interpretation of the underground strata.
From the early days of seismic surveying until the advent of real-time GPS, conventional survey methods using transit and chain, stadia rods, plane tables, and later EDMs and total stations were employed in the layout and survey of the source and receiver points. When differential GPS (DGPS) was developed around 1991, it immediately became apparent that a real-time variation of GPS was ideal for open-sky seismic projects. The downfall of DGPS was that it was a code-only solution, meaning that at best, the horizontal position might be as good as two feet, and the vertical error would normally be twice the horizontal component. As the vertical part of the solution was the most important to the geophysicist, DGPS was not a good tool for seismic surveying.
But GPS technology continued to evolve, and by late 1993, Trimble Navigation had developed real time kinematic (RTK) GPS, that, under good conditions, could give the operator accuracies of close to one tenth of a foot in the vertical. In difficult conditions, such as in the bush or areas of excessive signal interference or multi-path, the GPS solution would revert to either a double difference or a code-only solution, and the accuracy would degrade to that few feet previously mentioned or in some cases up to 10 or 20 feet. Not good enough!
In the mid 1990s, some experimentation took place using RTK in the bush of northern Alberta, Canada with 4000 SSi receivers from Trimble Navigation, and the results proved acceptable in light canopy. GPS receivers from other manufacturers were also tested, with mixed results, but in general, moderate to heavy canopy proved the enemy of RTK GPS. However, the petroleum exploration industry continued to push the envelope because of the substantial cost savings and lower environmental impact over conventional survey techniques.
As a result of encouragement from a number of our clients, in 2000, Warren Plue, vice president of Wolf Survey and Mapping, a division of Destiny Resource Services Partnership, put together a team headed by Joe Pilieci and Mark Strickland, who would attempt to develop and perfect the techniques for using real-time GPS in light, moderate, and heavy tree cover. Over a period of nine months, extensive research and testing resulted in what is now known as under canopy GPS (UCGPS).
The basic theory of real-time GPS is that a base receiver broadcasts RTK "corrections" using some kind of radio or other wireless technology to the rover receivers, which use that data along with GPS data they receive independently from the satellites to compute a position. In seismic surveying, this computed position is used as an aid in navigating to pre-calculated locations of the source and receiver points. Ample tolerance is allowed in staking out those "pre-plot" locations as long as the final "as-surveyed" coordinates meet the client's accuracy specifications. Placing the point within a few feet or so of the pre-plot location is generally acceptable, and even under dense canopy, RTK can readily give that kind of accuracy.
The real challenge with RTK, due to the nature of GPS signals and the algorithms various manufacturers use to "fix the integers" and give a precise three-dimensional position, is that it is next to impossible, under moderate to heavy canopy, to correctly determine the number of carrier wavelengths from each satellite to the phase center of the GPS antenna. Signal blockage, multi-path, and inherently weak signal strength all contribute to a very confusing situation for the onboard firmware, which tries to compute these solutions in a few seconds.
Going under Canopy
With UCGPS, if the canopy is too dense or the signals too weak to allow the GPS receiver to determine the exact number of wavelengths from satellite to antenna, known in GPS parlance as "fixing the ambiguities," then the operator will stabilize the 15-foot antenna pole, normally using a lightweight bipod, and begin to log static GPS data. The amount of data logged is based on a formula developed after extensive and rigorous testing by Wolf Survey and Mapping that takes into account the number of GPS and GLONASS satellites and various other parameters.
All points that did not achieve a fixed integer RTK position in the field are post-processed, and using proprietary processing techniques Wolf has developed along with its unique field procedures, many of the positions that would not "fix" in the field can be "fixed" after the fact. However, depending on the density of tree cover, some points will exist that can't be "fixed," and in that case, we attempt to process until we get a "float" solution that meets the client's vertical specification, which is generally less than three feet. Occasionally, some points can't be brought into tolerance, and in those cases, the GPS field operator will be sent back to that location to survey it again. After seven years of testing, we find that approximately 98 percent of our data is acceptable on the first run.
We should explain that not all GPS receivers can be used effectively for UCGPS. First, the receivers must be dual-frequency and full-wavelength, capable of tracking all available signals. Extensive testing with several brands showed that receivers supplied by Leica Geosystems consistently yielded superior results in more difficult conditions, partly due to their ability to track the L2 frequency better than some competitors. Since the original tests in 2000, Wolf has employed only Leica Global Navigation Satellite System (GNSS) technology and currently owns 128 GX1230 receivers. These use the Russian GLONASS as well as GPS satellite signals. When the Galileo constellation becomes viable, we plan to make use of those satellites as well.
Often, in dense canopy, lidar has proven a cost effective method of improving the vertical component of the solution by replacing the non-fixed-integer GPS height with the height from a lidar-derived digital elevation model (DEM) and merging that with the horizontal component from the GPS position. Consequently, lidar has almost become a prerequisite on jobs where heavy canopy can prevent us from consistently achieving the vertical accuracies our clients require.
It is no longer correct to say that GPS cannot work under tree cover. For light to moderate canopy, Wolf Survey and Mapping has developed a series of field and office procedures that allow us to provide acceptable data using GNSS alone, and in dense canopy using lidar-assisted GNSS. These techniques will not work for all applications, but in the field of seismic surveying, the results are surprisingly good. This has allowed Wolf Survey and Mapping to establish a niche in the field of non-standard applications of GPS and GNSS.
About the Authors
Based in Calgary, Canada, Joe Pilieci, PE is vice president of Wolf Survey and Mapping. He has a degree in geomatics engineering from the University of Calgary and over 25 years of international surveying experience, specializing in GPS and related technologies.
Mark Strickland, BS of Calgary, Canada is general manager of Wolf Survey and Mapping. He has an honors degree in geomatics from R.M.I.T University in Melbourne, Australia and over 10 years of international survey experience, including eight years in the seismic industry.
Jim Gillis, NSLS, CLS, RPLS is U.S. operations manager of Wolf Survey and Mapping and lives in Atlanta, Georgia. He has associate degrees in land surveying and geophysical surveying from the College of Geographic Sciences in Nova Scotia, Canada and over 35 years of survey experience.
With offices in Calgary, Canada and Houston, Texas, Wolf Survey and Mapping is a division of Destiny Resource Services, a provider of survey, mapping, navigation, and other services to the geophysical exploration industry.
» Back to our January 2008 Issue