Handheld Coordinates To Ngs

[+m&h]

This is a pitch for special care in submitting handheld coordinates to NGS, with apologies for singling out one example. As we know, handheld coordinates can be useful when the datasheets give scaled coordinates. But if handheld coordinates are surprisingly far from the scaled ones, we might take the time to plot them on a map before submitting them.

 

The one example is SY0690, designated M 10, in Olympia, Washington. The scaled coordinates on the data sheet are 47 02 33 (N) 122 54 02 (W). These plot to within 60-80 meters of the described intersection of Legion Way and Washington Street. Not too bad for scaled coordinates; at this latitude the six-second margin of error measures roughly between 125 meters (east-west) and 190 meters (north-south).

 

The recovery note for 2005 gives handheld coordinates as 47 02 35.04 (W) 122 50 59.46 (N). The reversal of the latitude and longitude designations is a typo whose obviousness varies with the observer, but most of us will catch it. The 50 minutes in the (assumed) longitude reading is a less obvious error that puts the coordinates almost four kilometers west of the station, not far from Exit 107 off Interstate 5 in Olympia. A correction to 53 minutes brings the longitude well within the six-second margin of error for scaled coordinates, but who knows what the GPSr actually said?

 

Up toward the middle of the data sheet, we find that on the date of this recovery, the site was reported as not suitable for satellite observations. The 50 minute reading may not be a typo. In any case, it might have been better not to trust it.

 

We have ourselves learned this lesson from experience; it is not at all difficult to press the wrong key, or to put too much trust in a GPS receiver. As Paul says on the thread he started, several of us know how it feels to submit corrections to our own reports.

 

Cheers,

 

m&h

[+Spoo]

Allow me to speak to US amateurs. I do not think we should feed our hand held GPS co-ordinates to NGS at all.

 

I consider myself to be very good at navigation. I am a trained pilot and navigator. I go back to the days of VOR, DME, LORAN and ADF. I have a very good hand held GPS. I consider myself to be very good at land navigation with years of hiking experience. I own a top of the line GPS.

 

And from all this, I know that my errors on any given day and any given location CAN BE WRONG. We amateurs just do not have the equipment or the knowledge for the accuracy that professional surveyors demand. And I am not even addressing such issues as a wrong datum or inverted number when we type.

 

I know that most of us amateurs out here use our heads and do not try to second guess the professionals. I also know that at some time or another we have found what we think is an error on their part. But that does not put us in a position to do anything more than SUGGEST to the NGS that there MIGHT be an error.

 

I am probably preaching to the choir here, but just wanted to say that we non-professionals should not feel holier-than-thou about our fancy gadgets when compared to people that have spent their lives perfecting a science and art.

Edited March 15, 2006 by Spoo

[+Black Dog Trackers]

quoting Spoo:

Allow me to speak to US amateurs. I do not think we should feed our hand held GPS co-ordinates to NGS at all.

I think this is a bit of an over-reaction. We must remember that the handheld coordinates we report to the NGS are only part of the to-reach instructions. They are not assimilated by the NGS to replace scaled coordinates (even though they are much better 99.5% of the time). There will always be minor errors in to-reach instructions and we've all seen them.

I am probably preaching to the choir here, but just wanted to say that we non-professionals should not feel holier-than-thou about our fancy gadgets when compared to people that have spent their lives perfecting a science and art.

In submitting handheld coordinates to the NGS, I don't think we're doing that at all. Certainly I am not.

[+PFF]

Ditto to BDT's comments. I've found the SCALED coordinates to be off by 400 feet on numerous occasions.

 

It was suggested by some of our professional participants that it is a waste of time to report amateur GPS readings on disks with ADJUSTED coordinates. However, it is appropriate when the SCALED coordinates are significantly off.

 

For a long time, my notation in recovery reports was: THE COORDINATES ON A CONSUMER-GRADE GPS WERE......

 

Later, a participant in the Forum said we could shorten this to an abbreviation for "hand held", but if we did, we should express it as "HH2". This code indicates that the coordinates are from a consumer-grade GPS and not from a professional-grade GPS.

 

By the way, I recently priced the professional-grade units. Trust me. You'll want to stay with HH2. [grin]

 

-Paul-

[mloser]

I agree with BDT and PFF. We are just providing one more TO REACH for the mark. I am a bit flaky about whether I include the HH2 coords, sometimes because I have left my GPSr in the car and don't want to trek back to it, sometimes because the location is pretty obvious, such as on a very noticable bridge or structure. I am much more inclined to include the coords if I feel they would help the searcher get close enough to find the other clues. One of my best examples is KW0697. I looked for this one for well over an hour based on the scaled coords and the lack of usable reference points. I added some better references in the description as well as the coords. Just getting closer to the location of the mark would be helpful. If my coords are right on and theirs match, that would be great, but certainly mine will get them in the ballpark.

[joegeodesist]

We at NGS enjoy receiving GPS positions for our many SCALED bench marks, whose positions were manually derived from quad map plots and can be many meters in error.

 

NGS is hoping to add handheld GPS coordinate entry to our mark recovery entry form. We plan to include two caveats:

1. determine which type of GPS you use (normal handheld vs. survey grade differential)

2. ensure that your expected accuracy is superior to the existing accuracy.

 

Not to worry, manual data entries will never replace survey grade positions.

[Photobuff]

IMO, one has to use a bit of skill if including handheld GPS coordinates. Unless I have a minimum of 7 or so satellites being tracked, WAAS, and a low error reading, I wouldn't include the numbers. If all those things are present, I'm confident my error is within a couple meters, and usually far less. We have a lot of buried marks here, where all the original landmarks are gone, and having reasonable consumer grade GPS coordinates should still be a great help in getting close to the mark. One of the (missing) marks around here has a description that reads "900' from center of driveway", and that seems a bit unfriendly to one with a 100' tape! The only problem I've had with the abbreviated format is due to the way my Magellan reads. I don't have a tenth second digit, so maximum precision is with decimal degrees (5 places). I don't like the risk of making a mistake on a manual conversion, but the preferred format is still DD MM SS.S. I've previously provided decimal degrees and no one's complained, but in the future I'll just do very careful double checked conversions to hold to the standard. IMO, decimal degrees make far more sense from an engineering standpoint, but trying to change long held traditions is a foolish endeavour.

[+PFF]

Interesting about the missing digit on the GPS receiver. As an alternative, you can take your reading in decimal degrees and convert with an on-line calculator. The "missing" digit may suddenly appear. :-)

 

-Paul-

[Z15]

Listing seconds to 2 decimal places is not found on my hand-helds beause the are not relaibe to that precision. My Garmins only go to 000 00 00.0 (1 decimal place).

 

Survey grade is going to put you within a foot in most instances (and closer depending on factors), not 9 ft. I don't think anyone here should be listing postions to more than 1 decimal place in the seconds.

Edited March 16, 2006 by Z15

[Photobuff]

Just to clarify on the Magellan- the choices are DDD.DDDDD and DD MM SS. There is no option for DD MM SS.S and IMO the tenth digit is necessary and important. I'm not talking about hundredths of a second, which is pretty much worthless with a consumer grade device, though with a program like SA Watch one can easily achieve accuracy at that level given enough time (hours or observations on multiple days). The difference in the size of the "box" between seconds and tenths of a second, is huge, in terms of finding a mark, and the consumer grade units can deliver a couple tenths most of the time. Why Magellan chose not to display tenths in the DMS format (on the 210) is beyond me, as it's otherwise a pretty decent unit

[+Black Dog Trackers]

Photobuff -

 

0.00001 degree is about 1/3 of 0.1 second.

What I'd recommend is to experiment with your Magellan to see if it seems to be able to distinguish 0.00001 degree. My Garmin 'acts funny' at that setting, I can input coordinates to 0.00001 degree but the unit then 'snaps' to either a higher or lower 0.00001 amount some of the time. For instance, I can't always enter two waypoints 0.00001 degree apart to check the distance between them with the unit's distance measuring function. I am assuming that the unit is reacting to the fact that it's resolution is not totally down to 0.00001 degree. (It does read to 0.1 second and I can input 0.1 second without it 'snapping'.) Also move around a bit at the 0.00001 degree setting and see how the unit changes the last digit. If your experiments seem to indicate that your unit can read to 0.00001 degree, you could convert from the DDD.DDDDD format to the DDD MM SS.S format mathematically. Of course, you must round to 0.1 second. My GPSr says that it is accurate to around 13 feet when it has a good fix on lots of satellites. Does yours do that?

[Photobuff]

I fiddled with this a while back and concluded that the internal calculations are actually done to a higher precision than is displayed, based on the data that SA Watch extracts via the comm cable. Also, based on the fact that the distance to target reads out cleanly in 1' increments. OTOH, when the data is stored as a POI or waypoint, the file format limits the precision and accounts for a slight "snapping" higher or lower, if the data is stored and recalled. It's only the last digit, and not always that. My conclusion was that decimal degrees (DDD.DDDDD) is the preferred setting, and conversions using Geocalc are as good as can be done, if one starts with decimal degrees. The error indicator will typically be 7-13' if I have good tracking of 7 or more satellites, and I've seen it go as low as 3'. Based on local adjusted benchmarks, the error indicator is actually conservative. Using SA Watch with weighted averaging, and enough data collection, positions will repeat to well under a foot. Unfortunately I don't have a laptop, so I haven't tested absolute accuracy on a benchmark, using averaging. What I'd like to determine is how good one can do with elevation, given enough data processing, as I have several elusive local marks on steep terrain. If I could determine elevation within a foot or so, I'd have a far better chance of locating them, if they still exist. IMO, getting good elevation numbers is a lot tougher than horizontal. Remember that to get the performance I'm talking about requires perfect conditions; I don't even bother if I can't get a boatload of good strong signals and WAAS!

[Bill93]

I believe the indicated accuracy values on my Garmin unit are based on a 95% confidence circle; that means you will be further off than it says 1 time in 20. Magellan may be similar.

 

This assumes the only things degrading accuracy are things the unit can sense, mostly signal to noise ratio and amount of disagreement between its various measurements. If you get some combinations of reflections you could be further off than it indicates.

 

Another concern about averaged values is that I'm not sure whether my unit rounds or truncates to that last displayed digit. It clearly operates with greater precision internally, as you noted for your unit. I've been collecting data on some known points on various days and manually averaging them. So far it isn't clear which is happening. Like you, I need a notebook computer to run SA Watch or similar program to get the numbers read out to more precision.

[+Cyclometh]

That was my recovery report. I didn't realize I'd gotten the coordinates wrong. I was simply trying to be helpful.

 

I submitted the report on February 4, 2006. Several of my reports have gotten the date wrong.

 

Here's the coordinates in my GC.com report, which I recall I used to get the values in the NGS report.

 

N 47° 02.584 W 122° 53.991

 

As you can see, those are exactly correct. So either I made a mistake in converting the coordinates, or an error occured in entering that data in the datasheet.

 

I won't submit any more recovery reports to NGS for any marks that aren't exceptional (haven't been found in a long time, etc) from now on. And I'll certainly never include handheld coordinates again.

 

Sorry for the trouble.

 

Edited to add: And no, it's not suitable for sattellite observation. There's tall buildings all over the place and trees to well above the horizon in every direction from that location.

Edited March 17, 2006 by Cyclometh

[+Cyclometh]

I emailed Deb to ask that the coordinates be removed and the recovery date updated.

[+Cyclometh]

Actually, thinking about it, I'm pretty sure I was in the area of the bad coordinates that same day, looking for another mark that's long gone. I may have taken a waypoint and accidentally used the wrong coordinates in my report to NGS. That seems the most likely scenario.

 

Rest assured I'll never include coordinates on any report ever again, and I sincerely doubt I'll submit any more recovery reports to NGS, outside of something truly exceptional. The last thing the database needs is more cruft with poor quailty reports.

 

I'm truly sorry for the problem. In this case, I really doubt it's caused anyone any trouble, but it's still very disconcerting to find out I did this.

[+BuckBrooke]

Cyclometh,

I would advise you to report to the NGS if you want to. I've made half a dozen mistakes that I've corrected, in 700 or so logs to the NGS. I probably have a few more errors undetected. These range from minor (accidentally getting two reports submitted) to more serious, and embarrasing (my not finding a disk, reporting it as NOT FOUND, and then being emailed to go out and find it by a local surveyor who uses the station all the time). Cheryl is very willing to fix entries like this, and you've learned something. I try to check back over my reports a month or two after I enter them, when they've been entered in.

Edited March 17, 2006 by BuckBrooke

[+Black Dog Trackers]

Rest assured ......

another over reaction, I'd say. Cutting off the supply of information you can give to the NGS is worse than having to correct one handheld coordinates mistake. Just get past it.

 

I also submitted about 8 PID reports once (my first NGS submission) and discovered later that the handheld coordinates were all off by some paperwork transcription error, I forget what. So, I emailed Deb the corrected versions and she fixed them all. They are now in the NGS database. It's all part of the learning experience and becoming an expert (amateur) reporter. Fixing one is much less of a problem than than fixing 8.

 

As I indicated, I have seen several mistakes even by the pro's in to-reach descriptions. It happens.

[Photobuff]

Goodness, don't stop reporting, and certainly keep including coordinates. I haven't be at this very long, and each of the couple dozen reports I've made could certainly be improved a bit. You'll never achieve perfection in this lifetime, and the only way to get close is practice. Being a bit error prone myself, now you see why I'm so reluctant to do conversions, and why I triple check them when I do!

[+Cyclometh]

Rest assured ......

another over reaction,

 

Maybe, but this is important stuff we're dealing with and my mistake could cost someone thousands of dollars. Not in this case, because if you had seen the mark, you'd know it's impossible to miss- you can see it from the street and the building's freakin' huge.

 

I could give you to reach instructions that said "go to legion and 7th, look at the southeast corner and walk across the lawn. Now look down about knee height. Enjoy" and you wouldn't have any trouble.

 

But if I'd done something like that on a mark that's not so simple... not cool.

[Wintertime]

IMO, one has to use a bit of skill if including handheld GPS coordinates. Unless I have a minimum of 7 or so satellites being tracked, WAAS, and a low error reading, I wouldn't include the numbers. If all those things are present, I'm confident my error is within a couple meters, and usually far less.

I don't know about other folks here, but after I've done a benchmark hunting run, I download the newly made waypoints (plus the waypoints I made before the trip from the existing NGS data sheet) from my Garmin to my Mac and display them on a topo map in MacGPS Pro. I can then quickly see whether my waypoints are closer to the actual locations of the marks than are the coordinates given on the NGS data sheets. So that's a good "ground truth" check that folks can do if they aren't sure of the accuracy of their handheld coordinates. (Since I've been chasing a lot of USGS benchmarks, those are often very conveniently already marked on the topo maps!)

 

Of course, as others have said here, I only submit coordinates with my NGS reports when the existing data sheet has scaled coordinates. I found one in Yosemite last fall that was off by a good 1/4 of a mile!

 

Patty

[+m&h]

For Cyclometh's benefit, we'd like to reiterate our apology for singling out one example, and our ending note that most of us have had to correct ourselves occasionally. We second the responses of those who urge Cyclometh to keep reporting.

 

Cheers,