Handheld GPS Receivers: Results From Experiment to Assess Which Variable Most Affects Coordinates Accuracy

This is the third instalment in a series about our Oldest GeoKid’s grade 5 science fair project, undertaken when we were still relatively new to geocaching and had a keen interest in better understanding the optimal conditions for taking accurate coordinates for our geocache hides. Part 1 outlined the many variables he planned to evaluate. Part 2 outlined his actual experiment and shared the most high-level finding: tall buildings, trees, canyons, chain-link fences and other sources of multi-path error result in very inaccurate coordinates.

Here are the key findings relative to the other variables he studied:

Allowing the GPSr to warm up before marking coordinates had a noticable effect. At every location, the best readings generally occurred when the GPSr warmed-up for at least 20 minutes. At BCP211, two-thirds of the readings that were less than 2.5m away from the known location were taken when the GPSr had warmed up at least 20 minutes. For the two locations with multi-path error problems, the best readings were taken following a warm up period of 20 to 25 minutes. Based on readings and research undertaken after his experiments, Oldest GeoKid concluded that this makes a lot of sense: certain information coming from the GPS satellites (almanac and ionospheric data) are transmitted every 12.5 minutes, so if your GPSr is on for at least 25 minutes, the receiver will have time to fine-tune its calculations incorporating 2 full sets of this extra data.
Weather conditions also had an observable effect on accuracy, with clear skies generally resulting in better accuracy. More than half of the readings taken when it was sunny showed less than a 2.5 meter variance from the “truth” location. Snow seemed to decrease accuracy, with just 8% of readings taken during snowfall landing within 2.5 meters of the “truth” location. As mentioned in a previous post, since he did his project in November – December – January, he didn’t have a chance to test the effect of rain on accuracy.
Oldest GeoKid’s observations and analysis did not identify a relationship between Dilution of Precision (DOP) and accuracy, probably because there wasn’t much variation in DOP over the weeks he did his experiment. But his backgroundr research suggest that a DOP level > 10 will likely have a significant effect on accuracy.
Because our location in Calgary is almost 900 km from the nearest WAAS ground station, Oldest GeoKid couldn’t analyze the effect of “WAAS on” vs. “WAAS off”. His research suggests that if you are within 370 km (about 250 miles) of a WAAS ground station, you would probably see improved accuracy by turning “WAAS on”. If anyone out there has first hand experience with this, we’d love to hear about it.
And finally, throughout his test period, the level of atmospheric interference remained low. Again, he researched this question and found that this is one of the most significant variables: plasma density in the ionosphere can affect a GPSr’s calculated coordinates by up to 10 meters (32 feet).

Any good scientist attempts to identify potential sources of error and other limitations around their findings. Oldest GeoKid gave this quite a bit of thought and came up with the following points for consideration:

He used a Garmin MAP60CSx to record all of his observations. Other GPSrs may not be affected quite the same way by the different variables he assessed.
His conclusions about WAAS, atmospheric conditions and DOP were based primarily on research: he was too far away from a WAAS ground station to test this variable and there were not enough different variable states for atmospheric conditions and DOP to allow for effective testing of those two variables.
He also would have liked to take readings in rainy weather, to see whether he could identify a rain effect on accuracy.