Backpacker Going Lightweight

Many GPS receivers will show you the "birds" (GPS satellites) that have their signals reaching the receiver. If there is a high number, that gives you a more dependable feel for the accuracy of the position result. If you see a high number before entering a box canyon, and then you see a low number when within the canyon, that tells you something. Many receivers will give you an Estimated Position Error (EPE). If EPE is high, then the position result is iffy. Some receivers will give you a Vertical Dilution Of Precision (VDOP), and that is based on the satellite geometry. My point is that many receivers will tell you how far you can "take it to the bank" on position results. If you never bother to consider such output, then you may find yourself with strange position results.

–B.G.–

"Bob, what 2oz gravity setup do you use?

Frontier Pro.

I would like to find the guy who said that he could buy them for $4."

Bob, I remember that post as well. Let me know if you ever find the guy or the $4 filters.

I asked the poster about that and vaguely remember it being a local sale (i.e. not available online).

Hi, Bob,

Thanks for your comments.

>Once you've used GPS for a while longer, you will learn how to interpret those results better.
Well, I've been using GPS since 2002, but I do always try to keep my mind open to learning more.

>For example, sometimes they are caused purely by lack of user training, like by using an incorrect datum, letting a barometric reading override a GPS reading, entering data wrong, etc.
Mmm…in the case where it had me 3000 feet underground and a mile away, no…I don't think so. I've seen other cases where, e.g., it had me levitating a thousand feet above a nearby lake. The vertical coordinate is the least reliable.

>Sometimes they are caused by a poor receiver situation, like if you are in a box canyon with multipath interference bouncing all around.
Interesting thought about multipath — hadn't thought of that before. But the case I was describing above was on a mountain peak.

>Commonly, old ephemeris data will cause all sorts of screwy results.
This is a good point, but I had been careful to check that my GPS was set to NAD27, the same as my map. Anyway, NAD27 versus NAD83 is only a horizontal difference of about 80 meters in the area I was in.

>At the end of the day, GPS is actually extremely reliable,
Sorry, just can't agree with that. In my experience it's the least reliable thing in my navigation toolkit. I get about 80% success, 10% failure because I'm in a canyon, and 10% gross errors like the one described above. But that doesn't mean it's not useful. Any navigation method can fail, and you just want to have multiple methods so that if one fails, the other one won't fail in the same way, and you'll get a check on your results.

-Ben

"The vertical coordinate is the least reliable."

The very best vertical accuracy will be 1.5x the horizontal accuracy, but again that depends on the satellite geometry.

">Commonly, old ephemeris data will cause all sorts of screwy results.
This is a good point, but I had been careful to check that my GPS was set to NAD27, the same as my map. Anyway, NAD27 versus NAD83 is only a horizontal difference of about 80 meters in the area I was in."

You mentioned the datum, which relates to map coordinates, but that has nothing at all to do with ephemeris data, which is in the downlink signal.

">At the end of the day, GPS is actually extremely reliable,
Sorry, just can't agree with that. In my experience it's the least reliable thing in my navigation toolkit. I get about 80% success, 10% failure because I'm in a canyon, and 10% gross errors like the one described above."

That confirms my suspicions about what is going on with lack of training.

–B.G.–

Bob wrote: "You mentioned the datum, which relates to map coordinates, but that has nothing at all to do with ephemeris data, which is in the downlink signal."
I see. Thanks for pointing out my misunderstanding. I'd be very glad to learn more about this. Can you explain more about the ephemeris data?

"That confirms my suspicions about what is going on with lack of training."
If there's an error I'm making, I'd love to learn what it is. Any information you can be provide will be much appreciated. If you have a hypothesis about what *user* error could lead to the GPS saying I'm at 5300 ft elevation, when actually I'm at 8985, that's a type of error that I need to know about. My actual location was at the top of Telegraph Peak, near Mt Baldy. The elevation of that peak is 8985 ft. My altimeter read 8700 ft. I had calibrated my altimeter at the Icehouse Canyon trailhead earlier that day, at the known elevation of that trailhead, 4960 ft. When I was on top of Telegraph Peak, the GPS read my elevation as 5300 ft, which is about 3700 ft lower than my actual elevation. I later confirmed that I had actually been in the place I believed myself to be, because shortly thereafter I *descended* to Mt Baldy Notch, at 7840 ft. (Baldy Notch is a big parking lot at a ski area, so I know for sure that that's where I was then.)

I'm sorry to be the bearer of bad news, but there are all sorts of screwy map database errors in many GPS receivers. That has to do with the way that the receivers have all of that data registered in the map database. In other words, if you have good GPS signals and lat/long, then those numbers are probably very good numbers. If you take those lat/long numbers and put them on a paper USGS map, they are still probably very good, and the indicated position is something that you can take to the bank. However, if the same GPS receiver gives you lat/long and then shows you that on its own electronic map (with errors), then I am not surprised if you see big errors, especially with altitude.

When a military mission planner plots out the course for a fighter jet to fly over some war zone, there is no way that they could tolerate having a 3700 foot altitude error. The jet might auger into the side of a mountain. The military people simply have accurate maps, and some of the very ordinary consumer-type GPS receivers have very inconsistent accuracy in the map database.

If you think about it, it is not hard to believe. The earth is a 3-D sphere (not quite a perfect one). The receiver's map database has to convert this for display in 2-D plus possibly contour lines. Errors enter in this conversion. You could probably get an accurate map database like the military has, but it would cost you only $50,000 extra.

I enjoyed it years ago when I was backpacking in Kings Canyon, and our group was heading cross-country from one pass to Granite Pass. Prior to the trip, I had used a paper USGS map to pick the exact lat/long for Granite Pass since there is a benchmark there. I had entered those lat/long numbers into my receiver as a GOTO. Once we got onto open land, I hit the GOTO and put my head down so as not to walk toward what I saw for landmarks. The distance decreased until the GOTO was 0.00 miles. Then I looked up, and the benchmark was on the ground seven feet in front of me. That was all the precision I needed. If I had been using the map database in a GPS receiver, I would probably still be hunting for the benchmark.

Ephemeris data is a list of the "fudge factors" for each individual GPS satellite. Some are flying more perfect orbits than others. If your GPS receiver has been running continuously for quite a while, and it has a current almanac, and then it gets good ephemeris data in the downlink, then it tends to be a little more accurate on position. Instead, if you just turn it on, get a fix, and then wonder why it has extra error, this could be the reason. As time goes by, the Air Force Space Command (who runs the satellites) is getting better and better at keeping the satellites orbiting and performing more accurately. That can only help, so positioning should get better in the future.

Tight canyons with lots of rock surfaces can be bad places to try to get a good fix. If you have doubts about what a receiver is telling you, hike up out of the canyon until your receiver sees more "birds."

–B.G.–

B.G. wrote: "USGS map, they are still probably very good, and the indicated position is something that you can take to the bank."
My GPS does not have electronic mapping. I use UTM coordinates and find my position on a paper topo map.

"Ephemeris data is a list of the "fudge factors" for each individual GPS satellite. Some are flying more perfect orbits than others. If your GPS receiver has been running continuously for quite a while, and it has a current almanac, and then it gets good ephemeris data in the downlink, then it tends to be a little more accurate on position. Instead, if you just turn it on, get a fix, and then wonder why it has extra error, this could be the reason."
Aha! This is very useful information. Thanks!

I've used UTM before, but I am kind of a traditionalist with latitude/longitude. It doesn't make much difference. Lat/long is what I learned as a kid.

There is a pretty fair chunk of technology going on inside each GPS receiver. Some are better than others. There is another big chunk of technology going on inside each GPS satellite, and then there is a worldwide network of monitoring stations for the satellites.

GPS receiver manufacturers are not going to bother to try to instruct each user on all of the technical bits and bytes. There are, however, some excellent tutorials on the web, if you are interested.

Back in the 1970's, when the government first said, "Let's have GPS!", they contracted with Rockwell to build lots of the first satellite systems. When I first started working in the GPS field, I was privileged to meet and talk with the principal engineer on the Rockwell project. He knew absolutely all there was to know about GPS.

–B.G.–

Ephemeris data is a list of the "fudge factors" for each individual GPS satellite. Some are flying more perfect orbits than others. If your GPS receiver has been running continuously for quite a while, and it has a current almanac, and then it gets good ephemeris data in the downlink, then it tends to be a little more accurate on position. Instead, if you just turn it on, get a fix, and then wonder why it has extra error, this could be the reason.

Is this the sort of data that is pretty permanent, and that once the GPS has acquired it will get remembered? Can I leave a GPS on at home to get "trained" and then just turn it on as needed when out on the trail?

Some people have been quite clear that they only turn on GPS as needed, in order to conserve batteries. Are you saying that they are likely to have substantial errors and not know they do?

–MV

There are more than 24 GPS satellites in space, at roughly half of the geosynchronous altitude, and they fly oblique polar orbits in one of six orbital planes. Some are moderately new, and they are flying in almost-perfect orbits. Others are in a slightly elliptical orbit, but still pretty good. Some satellites are nearing end of life, and their maneuvering fuel is almost gone, so they are getting moderately bad. If the Air Force can place more replacement satellites up quickly, often they will. Then, they remove the bad one from active service and use all of its remaining fuel to goose it straight out into a high orbit where it will be effectively out of the way for the next 50 years or something. So, there are all sorts of satellites in many different states of perfectness. Some are changing over time, and others are quite constant. The fudge factors are monitored and updated often, then put on the uplink to the satellites, and then later on it comes down to the receiver in the downlink, assuming that you leave the receiver on for a half-hour at a time (maybe 15 minutes).

First of all, it does not work good if you leave your receiver depowered for six months at a time, and then try to rush it into service. If you have had it depowered for a long time, it is good to point it toward the sky and let it soak for an hour while you are at home or before you start your trip. Then, when you get to the trailhead, or any time when it is turned on after being moved 300 miles or more (because you might have a different combination of satellites overhead), you really want to let it soak up some more signal before you thrust it into active use. Then, before you leave the trailhead, maybe change the batteries if you want.

You don't have to leave it powered up all the time on the trail. Many people do. However, as you make each nightly camp, you might want to turn it on and let it soak again, just to make sure that its tables are all updated.

Now, if you don't do that, I will not say that you will have substantial errors. The problem with GPS is that there are so many different factors that can screw up accuracy here and there, so if you really need it, you want to have as many of the factors eliminated as possible.

Some GPS receivers show Estimated Position Error (EPE) in feet. If you watch these things for hours at a time, as I have, you often see it start at 30 feet, then it drops to 20 feet, and then maybe 15 feet over the course of some minutes. The EPE algorithm is "wishful thinking" by the manufacturer, and often the 30 feet really means 50 feet.

If the satellite geometry is poor, you may have fair horizontal accuracy but very bad vertical accuracy. Some receivers have a screen to show you this.

I was on a high peak one time, and I had been running my receiver, so when I got to the summit, I set it down on the top rock for a few minutes. It stabilized with 12 satellites in view, because there was no other spot higher in the entire western hemisphere, and nothing could possibly block any reception. It stabilized with an altitude reading that was one meter off the official altitude. Close enough for me.

–B.G.–

Sorry, Chris, about having hijacked your thread with this long discussion of GPS! I don't know how the etiquette works on this site — should we start a new thread for this?

As a physicist, I just think it's amazing that I have this device that is so accurate that it needs to take Einstein's theory of general relativity into account in order to work. What a cool toy!

B.G. wrote: "Then, when you get to the trailhead, or any time when it is turned on after being moved 300 miles or more (because you might have a different combination of satellites overhead) […]"
The satellites are in highly inclined semisynchronous orbits, so even if you stay in one place, you're not going to have the same set of satellites all the time. I think the actual issue is that you might have a different combination of satellites overhead — *compared* to the combination you would have expected to have based on the almanac and your last known position.

The issue of initializing the GPS is an interesting one, and I spent some time this morning doing some reading online to learn about it.

Ephemeris data are broadcast every 30 seconds by each satellite, and are valid for 30 min. If the ephemeris data are out of date, the receiver will simply refuse to give a position until it's gotten it from all the satellites it has in view. So out of date ephemeris data are only a possible source of delay, not a possible source of error. On my unit, there is a progress bar that pops up after it's found a set of satellites. What the progress bar is showing is the progress in downloading the ephemeris data.

Almanac data are valid for several months, and are updated any time the receiver can get a continuous signal from one satellite for at least 12.5 minutes. Almanac data are coarse data, not used in calculating actual positions, so, like ephemeris data, they are not a source of error, only delay.

So, Bob, I think the long and the short of it is that the only reason you might want to leave your GPS powered on for half an hour or so at the beginning of a trip is that if you don't, you might experience an inconvenient delay when you turn the unit on later and want to use it.

B.G. wrote: "However, as you make each nightly camp, you might want to turn it on and let it soak again, just to make sure that its tables are all updated."
Assuming I've correctly educated myself this morning, and am not misunderstanding something important, I think this is bad advice. It will not improve your accuracy, and it will not even reduce delays. All it will do is run down your battery.

Re battery life, I did a crude test last night by leaving my GPS powered on outside all night. After 10 hours of being powered on, the LCD indicator on my Eneloop batteries showed a reduction of 1/8. Of course the LCD indicator is probably a lousy indicator of how much energy is actually left in the batteries, but this is at least some kind of indication of what you can expect. So it appears to me that even on a very long trip, you don't need to worry much about leaving the GPS on for 10 minutes versus leaving it n for 1 minute, but leaving it on continuously will certainly run down the batteries. Leaving it on for an hour every day, as you suggested, will probably not drain the batteries all by itself on a trip of any reasonable length, but it will contribute to draining them.

I did some reading about sources of errors in GPS. Ionosphere and troposphere effects are worst when the satellites are low on the horizon, because the signals have to punch through the most atmosphere. Apparently some units have a user setting that allows you to set a minimum angle above the horizon for a satellite to be used. Multipath is a big issue in urban environments, such as in 911 service, but is usually small in rural ones. Multipath through trees exists, but is too small to matter except for surveyers, etc. Multipath in canyons or off of cliff walls might be more significant.

Random errors in GPS are reported in a bunch of different ways. CEP is the horizontal radius that you have a 50% probability of being within. SEP is the same thing but in 3-d. Some manufacturers report a 2-sigma value, which would be 95% confidence for a Gaussian (normal) distribution. None of these, however, say much about how fat the tails of the distribution are. A google search will show you that outlers in GPS are a topic that's been studied by a lot of people. I found one paper on urban applications that defined an outlier as 1 km. So it seems clear to me that gross errors are possible with GPS, even without any human error. In the urban example, they're probably due to multipath from tall buildings. I haven't found anything to quantify how frequent outliers are in handheld backcountry applications.