GPS: GPS Q&A: Industry experts answer reader's GPS questions

Q. To what degree does foliage - considering different leaf types, heights of trees or shrubs, densities of plants - affect the absolute accuracy of positions determined by GPS receivers? -K.M. Tempe, Ariz.

A. John C. Bohlke, Sokkia Corp.: Foliage degrades the absolute accuracy of positions determined by GPS receivers because it attenuates or blocks satellite signals from reaching the GPS antenna. In some cases, the foliage will decrease the number of available satellites being tracked which often leads to less accurate positions. In other cases where the foliage is extremely dense, GPS signals may become blocked entirely. Density affects the GPS signal more than the type or height of the flora. Under deciduous trees, GPS users will usually achieve better accuracy during the winter when the leaves have fallen from the tree. Accuracy errors caused by multipath occur only when the foliage becomes wet. Proper mission planning and dry conditions will help to reduce the amount of inaccuracy caused by foliage cover.

Charles Branch, Ashtech: GPS accuracy is largely reflected in a number called PDOP (positional dilution of precision). PDOP isn't just an irritating acronym foisted on an unsuspecting public by the GPS industry; it also is an estimate of how GPS accuracy can be "diluted" by the uneven distribution of satellites around the sky. Accuracy is degraded if you calculate a position using the signals of satellites that are bunched up in one corner of the sky.
      Unfortunately, this is just what foliage (and buildings, mountains, etc.) can force you to do. By obscuring some satellites you can easily increase your PDOP and thereby your error. And this is only one way foliage can foul you up.
     Some receivers achieve greater accuracy by using the signals from more than four satellites (and nowadays there are usually six to eight in the sky at any one time). This is sometimes called an "overdetermined solution." How much the accuracy will be improved is nearly impossible to say because GPS manufacturers have been working on all fronts - not just overdetermined solutions - to improve receiver performance. I would guess overdetermined solutions could change a 5-meter receiver to a 2-meter receiver and a 2-meter receiver to a 1-meter receiver (all accuracies are AFTER differential correction). However, the extra satellites used for an overdetermined solution can easily be obscured by foliage and your accuracy will drop. You may even fail to pick up enough satellites to calculate a position.
      Attenuation of GPS signals by foliage can be easily predicted by anybody with a thorough knowledge of botany, biochemistry and electromagnetic wave mechanics, i.e., by nobody. A simple rule is that GPS signals will be attenuated in inverse proportion to the amount of light reaching your eyes. If little light reaches the ground on a sunny day in a forest, GPS work will be nearly impossible. But if you can see a lot of sky through the leaves, branches, and trunks, you should be able to make do - particularly if your GPS receiver has a lot of channels and so can track satellites as they pop in and out of view as you walk through the forest.
      Studies conducted by the USDA Forest Service help to qualify GPS performance under canopy. Read up on these studies if you want to know how broadleaf deciduous trees stack up vs. conifers, how tree density is offset by tree height, etc. The problem with such studies, however, is that it is hard to relate the Forest Service's canopy conditions with what you might find in your area of operations. But don't worry - GPS signal attenuation is a lot like art: you may not be able to describe it but you'll know it when you see it. So before you buy anything, make sure you test it out in your neck of the woods.

Wendy Corcoran, NovAtel Communications: Because GPS signals are in the upper UHF range, they are line of sight and attenuate under tree branches and leaves. As a result, foliage causes multipath or in the worst case, blockage of the GPS signal. In a paper written by Gerard Lachapelle et al, foliage tracking was performed on several types of GPS receivers. There were three features of a GPS receiver that increased its effectiveness under foliage:
      ¥ fast reacquisition time (<3 seconds) when the signal is lost;
      ¥ all-in-view capability to track as many satellites as there are available above the horizon;
      ¥ Narrow Correlator tracking which results in lower noise code and better multipath rejection.
      All of these factors contribute to the quality and number of samples obtained which ultimately affects the absolute accuracy. In cases where Narrow Correlator receivers are used, the absolute accuracy is not significantly impacted.

Craig Hudson, II Morrow Inc.: A dense tree canopy is one type of "obstruction" that can limit your ability to receive and process GPS signals. In dynamic or mobile applications (i.e. capturing polyline or polygon data), foliage can impact your ability to track and maintain satellite lock. The moving tree branches and changing foliage conditions "modulate" the GPS signals and positional accuracy will be degraded. however, static point positioning does not suffer as much, until you encounter dense conifer foliage.
      A conifer tree canopy will have more of an effect when compared to deciduous canopies. GPS signals are able to penetrate the simple leaves of deciduous trees with less attenuation and reflection. Whereas, GPS signals can have difficulty in reaching the observer for applications in a dense cedar forest.
      Accuracy can be evaluated using PDOP data but will be degraded by the modulation effects of the canopy. When using differential GPS techniques, 2 to 5 meter positional accuracy can be achieved. The user needs to understand their application and remember that GPS is a line of sight positioning system and positional accuracy will be affected by all obstructions.

Art Lange, Trimble Navigation: Foliage affects the accuracy of a GPS receiver in two ways. The first is by absorbing and the other is by scattering some of the satellite signals. The foliage caused absorption reduces the signal amplitude which causes the GPS receiver to have more noise on the derived pseudoranges measurement. The foliage caused scattering results in a form of multi-path noise. Both of these effects are random in nature and cause the GPS receiver's derived position to suffer from random position noise of a few meters, depending on the PDOP. Averaging for a few minutes at each point of interest helps considerably in improving the differential accuracy because of the random nature of the foliage effects. For more information, refer to the May 1994 EOM article on page 50 by Chuck Gilbert titled "Using GPS in the Shade."
      In general, the denser or wetter the foliage, the greater the signal absorption and scattering that will occur. For dry lodgepole pine trees, about five branches of a mature tree between a satellite and a GPS receiver antenna completely blocks the signal from a GPS receiver on the forest floor. Two dry lodgepole pine branches has little effect on a GPS receiver, and two wet lodgepole pine branches causes measurable scattering of the GPS position.

Q. What are the minimum pieces required to use GPS for 2 meter accuracy? For sub-meter accuracy? For centimeter accuracy? -R.R. Denver, Colo.

A. Bohlke: It is possible to obtain 2-meter accuracy using a minimum of one or two GPS code-phase receivers. One GPS code-phase receiver requires the use of either a real-time correction receiver or access to community base station files that could be used for post-processing. Two GPS code-phase receivers provide 2-meter accuracy by operating one receiver as a base station reference recorder and then post-processing the data. The same combination of GPS equipment is capable of achieving sub-meter accuracy but it usually requires a GPS receiver(s) that collects both code phase and carrier phase data. Centimeter accuracy requires a minimum of two L1 or L1/L2 carrier phase receivers.

Branch: To achieve accuracy better than 40 meters, you need to "differentially correct" your GPS receiver's data no matter what type of receiver you have. This entails placing a "reference" GPS receiver over a point of precisely known coordinates and processing its data together with that of your receiver. You can either bring your data back to the reference receiver and process it ("post-processing differential") or you can broadcast the reference receiver's data out to your receiver in the field ("real-time differential"). The former is more common, more accurate in most scenarios, and more of a hassle. The latter offers many advantages but is more expensive because you have to have a transmitter attached to the reference receiver and a differential receiver attached to your GPS receiver in the field.
      Generally speaking the accuracy you obtain with differential correction will only be as good as the less capable receiver you are using. So-called "C/A code" receivers generally can give 2-meters of accuracy. If you have two C/A code receivers and good quality post-processing software made by the same manufacturer, you'll probably get 2-meters accuracy. If the receivers are of two different brands, they won't communicate with each other because of the bane of the GPS industry: proprietary formats. Try to mix and match receiver brands and you might become "collateral damage" in the format wars. (Don't worry, like all medieval conflagrations, the format wars will eventually die out and you won't even read about them in books.)
      Add a real-time differential correction transmitter and receiver and you can get 2-5-meters depending on things such as transmission latency. Commercial correction services are becoming increasingly widespread. With these you buy a correction receiver to attach to your GPS receiver, pay a monthly fee, and away you go. These services and free government broadcasts in some areas obviate the need to buy a reference receiver and transmitter.
      Submeter accuracy is available from some manufacturers who use normal C/A code receivers to record carrier phase data. ("Carrier phase processing" is what is used by "survey-grade" receivers to achieve centimeter accuracy.) By doing a "pretty good" job of processing carrier phase data, "mapping-grade" C/A code receivers can achieve submeter accuracy. The down side is that you must post-process the data (real-time is not available for these systems) and you must typically occupy the position for 10 minutes or more (no dynamic mapping).
      Some manufacturers provide submeter accuracy with more advanced C/A code receivers. These receivers use more accurate clocks, quieter RF circuitry, and more capable digital signal processing firmware to improve accuracy. Under some circumstances, these systems can even achieve submeter accuracy when moving. However, submeter will be difficult to impossible to achieve with real-time differential for a lot of complex reasons. But you'll probably get 1-2-meters real-time.
      Centimeter accuracy requires two survey-grade receivers mounted on tripods recording carrier phase data that is post-processed using special, usually expensive, correction software. Most surveyors post-process their data but real-time centimeter accuracy is increasing in popularity. This is typically achieved with low-power, line-of-sight radios set up at construction sites, open-pit mines, etc. Very slick, though for $60-70k it should be.

Corcoran: The following is a list of the minimum amount of equipment required to achieve the level of accuracy specified:

2 Meters 1 L1, C/A code GPS receiver with antenna and cables 1 Differential corrections receiver from J.E. Chance, U.S. Coast Guard Marker Beacons, DCI, CUE Paging, etc. 1 Battery

Sub-Meter 2 L1, C/A code GPS receivers, antenna, cables (one base, one remote GPS receiver) **GPS receivers must be capable of this accuracy 2 Radios/modems, antenna, cables (for real-time applications) 2 Batteries 1 Copy of post processing software (if application is only real-time, this is not needed)

Centimeter 2 L1, C/A code GPS receivers, antenna, cables < 25 km OR L1/L2, C/A, P code GPS receivers, antenna, cables > 25 km (one base, one remote GPS receiver) **GPS receivers must be capable of this accuracy 2 Radios/modems, antenna, cables (for real-time applications) 2 Batteries

Hudson: Differential GPS techniques are required to achieve 2 meter accuracy. A user can either store satellite range data (pseudoranges) for PC post processing or use a radio data link to transmit and receive satellite range corrections in real time.
      Post processing requires a GPS receiver, the Apollo GIS 940, which is capable of capturing and storing satellite pseudoranges. Once the data has been collected, it is processed on a PC with reference station data. The reference station data allows the PC software, ApolloPC, to generate differentially corrected positions from the pseudoranges.
      Real time differential techniques require a GPS receiver capable of interfacing with the standard RTCM S/C 104 formatted differential signal and a RTCM receiver. The Apollo GIS 940 and a RTCM receiver allow for real time (no post processing) 2 meter accuracy.

Lange: To obtain differential GPS accuracy requires two GPS receivers: a rover GPS receiver and a base GPS receiver. The base GPS receiver is used to compute the differential corrections. A GPS receiver will have different circuitry and will process the GPS signals differently, depending on the accuracy desired. For sub-meter to 2 meter accuracy the GPS receiver only needs to use the C/A code to derive a GPS position. For centimeter accuracy, the rover and base GPS receivers must be able to measure the GPS carrier phase in addition to the normal C/A code measurements. The differential computation may be either in real-time or as a post-processed operation. If real-time differential GPS is required, then a communication link is also required to communicate the GPS base station information to the GPS rover receiver. In the simplest case, with a C/A code receiver, the communication link at the rover site may be a commercial DGPS provider's pocket sized FM paging receiver. If post-processed DGPS is required then a computer program is required to process the base and rover data files.

About the Participants:
John C. Bohlke serves as GPS support manager at Sokkia Corp. in Overland Park, Kan. He may be reached at 913-492-4900 or 800-4-SOKKIA in the U.S.
Charles Branch is a GIS marketing manager at Ashtech in Sunnyvale, Calif. He may be reached at 408-824-1603.
Wendy Corcoran is manager, survey and mapping products at NovAtel Communications, Ltd. in Calgary, Alberta, Canada. She may be reached at 403-295-4789.
Craig Hudson serves as portable products manager at II Morrow Inc. in Salem, Ore. He may be reached at 503-391-3411 or 800-742-0011 in the U.S., or 800-654-3415 in Canada.
Arthur Lange is the GIS product manager for Trimble Navigation in Sunnyvale, Calif. He may be reached at 408-481-2994.

Back