55 foot Discrepancy Between National Geodetic Survey Info and GPSr/Phone

[+Chipper3]

Yes, I realize that our hand held consumer devices have error but...

 

I decided to run an experiment on Accuracy and Precision of a couple of devices against a known National Geodetic Survey Marker.  I assumed that the location of the physical marker and its coords would be the Gold Standard of what a GZ actually is.

 

I used my iPhone running a coord averaging tool and collected 30 data points over time and by moving and returning to the actual NGS marker.  I used my Garmin 64ST in averaging mode and collected 4 100% readings for the same NGS marker over several hours.

 

MY iPhone and the Garmin GPSr were pretty much in agreement to the average coord location.  But when I plotted the results of my readings and included the published NGS coords for the marker, I found that the published coords for the physical marker were 55 feet way from 1.) the two device coords/locations and 2.) the actual in-the-ground marker. Let me repeat the NGS marker is embedded in concrete. The NGS published coords for that location are 55 feet away. My GPSr readings for that marker are 5 feet away. 

 

I checked the datum systems used by both devices and the NGS and the coords are all in agreement.  I am pretty sure that I have an apples and apples comparison.  In the spirit of full disclosure the NGS data is reported in NAD83  and my readings are in WGS84.  My conversion app shows the same coords for both datum.

 

Hmmmmm....

 

Advice and comments welcomed as I am out of my league here.  I thought I would learn something about the Accuracy and Precision of my devices against a known location but what I found was unexpected.  

 

 

Edited February 6, 2022 by Chipper3

[+baer2006]

This is strange. For consumer devices, NAD83 and WGS84 are indeed effectively the same - AFAIK the difference is less than 1 meter (will slowly change in the upcoming centuries because of continental drift).

 

However, 55 ft sounded reasonable for a difference between NAD27 and WGS84. I peeked into your profile, and you give your location as Anderson, SC. Assuming that you ran your test in your home area, I picked a random point in Anderson. Then I used an online NAD27-to-NAD83 converter to see, where I land when I use my point's WGS84 coordinates with the NAD27 datum. Result: 56 ft difference! IMHO that's too close to your 55 ft measurement to be a coincidence. Therefore my guess is that somewhere in your results there is a hidden NAD27/NAD83 datum confusion.

[+Chipper3]

1 hour ago, baer2006 said:

This is strange. For consumer devices, NAD83 and WGS84 are indeed effectively the same - AFAIK the difference is less than 1 meter (will slowly change in the upcoming centuries because of continental drift).

 

However, 55 ft sounded reasonable for a difference between NAD27 and WGS84. I peeked into your profile, and you give your location as Anderson, SC. Assuming that you ran your test in your home area, I picked a random point in Anderson. Then I used an online NAD27-to-NAD83 converter to see, where I land when I use my point's WGS84 coordinates with the NAD27 datum. Result: 56 ft difference! IMHO that's too close to your 55 ft measurement to be a coincidence. Therefore my guess is that somewhere in your results there is a hidden NAD27/NAD83 datum confusion.

Thank you for your response.  I hope you will investigate further for me as you seem to be on to something.  I am providing information on the marker in question:

ED3783

The associated Data Sheet is found at  https://www.ngs.noaa.gov/cgi-bin/ds_mark.prl?PidBox=ED3783

All the readings  from that document look like 

NAD 83(2011) POSITION- 34 34 03.42411(N) 082 37 31.08616(W)   ADJUSTED  

The link to the site I am using  - https://geodesy.noaa.gov/NGSDataExplorer/

 

I am exploring your suggestion by making a conversion of their stated coords from NAD27 instead of NAD83 and will post my findings.

[+Chipper3]

I am adding my device generated coords to the discussion for your use- 

Garmin 64 ST  with averaging   N34 34.0582  W 082 37.5172

iPhone Averaging Tool with 30 readings   N34 34.059     W 082 37.518

[Moun10Bike]

The datasheet present on Geocaching.com when that benchmark was imported back around 2002 shows different coordinates: https://www.geocaching.com/mark/datasheet.aspx?PID=ED3783. These come out to N 34° 34.062 W 082° 37.517 and are only about 18-23 feet off from what you recorded with your Garmin and iPhone. The coordinates for the benchmark moved about 30 feet between the two datasheets. I wonder if the recovery recorded in 2011 could have had an error.

[+Capt. Bob]

From the NGS forum FAQ (https://forums.geocaching.com/GC/index.php?/topic/95774-ngs-forum-faq/)

 

Q. I found a mark and its LAT/LONG coordinates were way off. What is going on?

A. This is a bit technical, so please read carefully.

NGS has two different types of marks: Horizontal control marks, and Vertical control marks. The marks used for vertical (i.e. elevation) control often have a “SCALED” position. You can tell if this is the case by looking at the datasheet. A “SCALED” position means that the horizontal (LAT/LONG) coordinates were estimated using a map. In this case (and ONLY in the case), your handheld LAT/LONG coordinates are potentially more accurate than the position listed on the datasheet. In these cases you can record your GPSr coordinates for the mark and include them in the notes area of the NGS recovery form. Your coordinates will be used to help people find the mark in the future. Keep in mind that submitting handheld GPSr coordinates is totally optional. If you do chose to submit handheld coordinates you should know that NGS uses the NAD83 datum and coordinates are recorded as DD MM SS.S. Recovery reports which include GPS readings should conform to those standards.

[Moun10Bike]

The datasheets indicate that benchmark is adjusted, not scaled.

[+baer2006]

9 hours ago, Chipper3 said:

NAD 83(2011) POSITION- 34 34 03.42411(N) 082 37 31.08616(W) ADJUSTED

Converting this to deg/min gives N34°34.0570685' W82°37.5181027'

 

8 hours ago, Chipper3 said:

I am adding my device generated coords to the discussion for your use- 

Garmin 64 ST  with averaging   N34 34.0582  W 082 37.5172

iPhone Averaging Tool with 30 readings   N34 34.059     W 082 37.518

The distance between the position in the NGS data sheet and your measurements is 2.5 meters (Garmin) and 3.6 meters (iPhione). Easily within expectations for consumer devices.

Therefore, it seems that I don't understand your original problem . Exactly what is "55 ft off"?

[+igator210]

Not for this specific situation, but I've also seen these type of markers get moved due to construction. 

[+Chipper3]

7 hours ago, baer2006 said:

Converting this to deg/min gives N34°34.0570685' W82°37.5181027'

 

The distance between the position in the NGS data sheet and your measurements is 2.5 meters (Garmin) and 3.6 meters (iPhione). Easily within expectations for consumer devices.

Therefore, it seems that I don't understand your original problem . Exactly what is "55 ft off"?

 

[+Chipper3]

The third point referenced in the original post is the one that showed up 55 feet away.  This was the location given by NGS as the location of the marker.   My computations resolved the posted marker a 55 feet away from what my devices read.

 

[+Chipper3]

Thanks to everyone that took the time to help me!

 The problem that I was having is solved. The error comes about as the first conversion tool I used to convert the NGS coords did not allow me to put in decimal readings in the seconds field for DMS entry to convert to the input my mapping program wanted as input. (Yes, Yes, I used a poor tool but I was ignorant at that point)  

 

So the difference of 1 second equals 100 feet.  the actual reading for Lat was 3.42411 seconds so by entering 3, I would be off by 42 feet.  When I add the same error for the Lon, I get the increased error which accounts for the 50 feet that I originally reported.

 

Digits Matter!

 

Edited February 7, 2022 by Chipper3

[Darwin473]

4 hours ago, Chipper3 said:

Digits Matter!

 

I read not that long ago that Nasa "only" uses pi to about ten places when doing orbital mechanics. That may be 100% false as I read it on the internet, but you do get diminishing returns for extra digits. Though you do get sent to math jail for rounding pi to 3.

[+barefootjeff]

8 hours ago, Chipper3 said:

The problem that I was having is solved. The error comes about as the first conversion tool I used to convert the NGS coords did not allow me to put in decimal readings in the seconds field for DMS entry to convert to the input my mapping program wanted as input. (Yes, Yes, I used a poor tool but I was ignorant at that point)

 

You really don't need a conversion tool to convert from DMS to DDM, any old calculator will do the job. Just divide the seconds by 60 and that becomes the decimal part of the minutes.

[+Chipper3]

Just now, barefootjeff said:

 

You really don't need a conversion tool to convert from DMS to DDM, any old calculator will do the job. Just divide the seconds by 60 and that becomes the decimal part of the minutes.

Thank you!