Overview
Global Positioning System (GPS) is gaining popularity among runners, and GPS receivers are wonderful tools for measuring run distances and monitoring running pace. But the information they provide about distance and altitude is approximate. To best use and enjoy this technology, it's important to understand its limitations and impact on running performance.
GPS receivers designed for runners are the least accurate grade of GPS now made. Currently, the two GPS receivers with features specifically designed for runners state they will provide a ground position that is accurate within 49 feet, and altitude location that is accurate within 75 feet, 95% of the time. This lack of precision results from several factors, and, importantly, may affect the distance and pace displayed by the GPS receiver during every run. Satellite geometry, atmospheric disturbances, electronic interference, and even running among tall buildings or under heavy tree canopy that can reflect or block GPS signals-all can throw off readings.
In the experience of the San Diego Track Club (SDTC) mapping team, it is not unusual for a GPS receiver to display run distances that are as much as 3% greater than the actual distances. For a 10-mile run in which the GPS error is 3%, the GPS receiver would display a distance of 10.3 miles. If the GPS distance is wrong, then the run pace (which is computed from distance and time) will also be wrong. Also, because of the greater margin of error for altitude location, elevation profiles produced by GPS receivers will usually be quite different from the actual terrain.
Other than measuring the distance of a run, GPS receivers are most commonly used for monitoring running pace. Instantaneous run pace is so erratic, though, it's virtually useless, which makes average pace the best tool to use for monitoring run pace.
Therefore, because there's the possibility that GPS errors will result in the wrong run pace being displayed, runners who use a GPS receiver for maintaining a desired pace during a race should be sure to compare GPS receiver split times with actual split times on the course and make "on the fly" pace adjustments as necessary.
GPS Introduction
The Global Positioning System is based on a series of 24 satellites, located 11,000 miles above earth, that work in conjunction with ground control stations and a GPS receiver to provide navigational information. This network of satellites was placed into orbit, and is maintained by, the U.S. Department of Defense. The GPS works 24 hours a day in all weather conditions.
There are three common grades of GPS receivers: survey grade, mapping grade, and recreational grade. Survey grade is the most accurate and is used in situations where accuracy is essential. Mapping grade receivers are used where accuracy is important, but not worth the added expense of owning or leasing a survey grade receiver.
The least accurate GPS receivers are recreational grade, used for outdoor sports or navigation. As of this writing, there are two GPS receivers with features specifically designed for runners, the Timex Ironman Triathlon Speed + Distance System and the Garmin Forerunner series (101, 201, etc.). These receivers, like others in the recreational grade, typically provide location information that is accurate to within 15 meters (49 feet), 95% of the time. That means it's less accurate the remaining 5% of the time. This loss of precision results from calculation errors caused by the configuration of satellites in the sky or by environmental factors here on earth. Survey and mapping grade receivers are subject to the same sources of error, but they incorporate additional technologies to correct most of them.
When a GPS receiver is turned on, it takes a few minutes to lock in on, or "acquire", the signals from GPS satellites. With three locked-in satellites, the receiver can calculate a two-dimensional location - latitude and longitude. If the GPS receiver can lock in on four or more satellites, it can calculate three-dimensional position information - latitude, longitude, and altitude. Once the position has been calculated, other information such as speed and distance can be determined.
GPS Receiver Errors
A GPS receiver determines its position based on low-power radio signals it receives from the GPS satellites. Those satellites are continuously moving in their orbits, so their position in the sky and their position relative to each other are not fixed. That the satellites move constantly gives rise to one of the types of GPS errors, the position error. Position errors result from the geometry of the satellites' positions between each other. Ideally, the satellites are at wide angles relative to each other. The greatest GPS position error occurs when satellites are in a line or clustered together so that there is little difference in the angle of their positions relative to the GPS receiver. Other types of GPS errors occur when there are atmospheric disturbances that delay GPS signals and electronic interference. Degradation in GPS accuracy also occurs when signals are blocked or reflected by running under bridges, among tall buildings or under heavy tree canopy. Hillsides, embankments, or even the runner's body also can block signals.
GPS Distance Accuracy
Depending on the above factors, I have observed that my GPS receiver (a Garmin Forerunner 201) varies in accuracy by as much as 3%, virtually always reporting a longer distance than is correct. On a 10-mile run, a variation of 3% would yield an error of 0.3 miles. This degradation in accuracy arises mostly from poor alignment of the GPS satellites, and is called Dilution of Precision (DOP). DOP varies constantly between hours and between days. Professional mappers and surveyors use software-planning tools (1) to help them identify the times with the lowest DOP, since these are the best times to collect GPS readings.
The Timex Ironman and the Garmin Forerunner series use the same technology and have the same expected degree of position accuracy, given factors already mentioned. So, with these units, at any given time on any given day, one may get GPS distance readings that differ from published route distances, such as those for the San Diego Track Club's Rockin' 'n' Runnin' marathon training long run maps. The SDTC mapping team measures route distances using mapping software and US Geological Survey photomaps that are corrected so they can be used for accurate measurements of ground distances.
GPS For Run Pacing
GPS receivers calculate speed from the distance they measure. Speeds in GPS receivers for runners are expressed in terms of minutes and seconds per mile, or run pace, based on when the runner starts and stops the timer in the GPS receiver.
Pace is computed by dividing the run time by the run distance. An accurate pace, of course, comes only when the GPS receiver accurately measures the distance.
If the GPS receiver measures a distance greater than the actual distance, the calculated pace will be faster than one is really running. An error of 3% in distance measurement equates to an error of 1.8 seconds for every minute of running pace. In this instance, a runner moving at an 8:00 pace will see the GPS receiver display a pace of 7:46! As a matter of racing strategy, if a runner knows that DOP will be high (i.e., bad) during the time of the race, she/he can plan to run at the incorrect pace that the GPS receiver will be showing. Using the above example of a +3% error, if the runner's goal pace is 8:00, he/she must maintain 7:46 on the GPS receiver. Otherwise, the runner will produce an actual 8:14 pace and most likely be disappointed with actual race results.
To help estimate the GPS receiver run pace for an upcoming race, it's possible to use planning software (1) to observe DOP for the day and time of the race. DOP values range from 1 (no satellite position error) to over 30 (very high error). If DOP is 4 or less for the duration of the race, the GPS receiver will probably display a sufficiently accurate pace figure (remember there are other non-DOP factors that can affect accuracy). If DOP is greater than 4, the GPS receiver run pace will be wrong. Just how wrong will depend on DOP peak value and the amount of time that DOP is greater than 4. For a DOP of around 10, for instance, consider using a pace adjustment of about 1.5%.
Because DOP may change during the time of a race, and errors other than satellite position may affect GPS receiver accuracy, runners should compare GPS receiver split times with actual split times on the course and make "on the fly" pace adjustments as necessary.
A final note about GPS receiver run pace: the GPS receiver models mentioned above (Timex and Garmin) display both instantaneous run pace and average run pace. Instantaneous run pace is so erratic as to be virtually useless, showing swings of as much as 30 to 45 seconds in a short period of time even when run pace remains constant.
Both GPS receiver manufacturers include GPS receiver configuration settings that allow the user to apply varying levels of "smoothing" to dampen these variances, but the "smoothing" does not work very well. The average run pace figure does not provide immediate feedback if the pace is increased or decreased by a small amount, but it does prove very useful for sticking to a goal pace for a race. If you see that your average pace begins to drop off, you can speed up until it returns to your goal pace. This makes average pace the best tool to use for monitoring running pace.
GPS Altitude Accuracy
GPS altitude numbers are subject to even greater error than are GPS horizontal distance measurements because all satellites are above the horizon. Generally, the altitude error for recreational-grade GPS receivers is 1.5 times the horizontal error specification, which means the altitude error is within ±23 meters (75 feet) 95% of the time. Elevation profiles that are created from GPS altitude recordings during a run should reflect the general nature of the course (the ups and downs), but don't be surprised to see altitude values that are drastically incorrect, or that are different for the same location at different times. Therefore, since DOP may vary considerably over the time of an out-and-back run, an elevation profile created for the run may not be symmetrical.
Accuracy of Other Common Distance-Measuring Methods
Though GPS isn't perfect, neither are other common methods of measuring distance, such as pedometers, bicycles, and automobiles. Pedometer calibrations require that the wearer measure his/her stride length. The pedometer then calculates distance by counting the number of steps taken and multiplying that number by the stride length. If the wearer's stride length varies (e.g., shorter steps going uphill), the distance figure will be wrong.
Bicycle and automobile odometers calculate distance by counting revolutions of the wheels. For one revolution of the wheel, the distance traveled is the circumference of the wheel. Total distance is the product of the number of wheel revolutions and the circumference. If the tire pressure differs from the manufacturer's recommendation, the tire circumference will not be correct and the distance will not be accurate. Inaccurate distance measurements will also occur if the owner buys a wheel and tire combination different in size from the original equipment.
More on DOP
Taking the measurement issue a step further, consider DOP has four standard error components related to satellite alignment. These components measure loss of precision in satellite clock timing signals, two-dimensional horizontal position, vertical position, and three-dimensional position. Combined, these four components are known as Geometric Dilution of Precision (GDOP). My experience has shown, for example, that when GDOP is around 10 (on a scale of 1 to 30), the error in the measured distance is about +1.5%. I have not found a theoretical formula that predicts the percentage of error from the GDOP.
The following chart shows DOP for San Diego for March 6, 2004.
This chart shows the value of DOP on the vertical axis and the time of day within a 24-hour period on the horizontal axis.
The different colored graph curves indicate the value of each type of DOP throughout the day. Satellite clock timing signal error, shown in yellow, has the lowest amount of error during this period. Other types of DOP are shown in different colors, and are usually greater than timing DOP. GDOP, the overall measure of position-related errors, is the red curve.
There are six times during this day when GDOP builds up to a peak greater than 6, during which the accuracy of GPS receiver distance measurements will be significantly degraded by poor satellite alignment. Those times are at midnight, around 1:30 am, about 6:30 am, again around 7:30 am, just before 6:00 pm, and lastly about 11:00 pm. At three other times GDOP peaks just above 4, the lower limit at which satellite position errors begin to affect GPS receiver accuracy. For all practical purposes, because these peaks are brief in duration and rarely above 4, distance measurements for runs spanning these three times will not be affected by satellite geometry.
A point to remember: Because satellites are constantly moving and changing their relationships to one another, this chart shows DOP values only for San Diego and only for March 6, 2004. This chart should not be used to estimate GPS distance accuracy errors for any other location or date.
When using the planning software (1) to examine DOP values, it is important to download a current ephemeris file before each use of the software. While some satellite orbits are known quite accurately, minor orbital ("ephemeris") errors arise from navigational pulls of the moon and the sun, and the pressure of solar radiation on the satellites. The Department of Defense uses very precise radar to track each satellite's exact speed, altitude, and position. The ephemeris file, which is updated several times daily, contains Department of Defense radar-tracking data to provide the most accurate information about the current orbit of each satellite. The ephemeris file can be downloaded from the same Web site that provides the planning software.
Summary
GPS receivers are wonderful tools for measuring run distances and monitoring running pace, but the information they provide about distance and altitude is approximate. New technologies that provide improved accuracy to within 3 meters (10 feet) are now available in GPS receivers intended for general route-finding use. These technologies will likely be incorporated into GPS receivers for runners soon, and have the potential to eliminate much of the error from satellite position. But the other factors - tree cover, buildings, rocks - will always be there to wreak havoc with the numbers. The message for GPS owners: enjoy your running and don't let the technology weigh you down!
Jim Clements
June 2004
(1) "Planning", copyright © 1997 - 2001 by Trimble Navigation Limited. Available for download at no cost at
trimble.com