Geodetic Base Line Sign Dedication

Great work. It's always good to remind the public that surveying is important.

Did anyone consider an EDM traverse to compare to the original distance measurement?

Mr. Penry,

That is a fascinating post - lots of historical significance. I read the associated article and have questions - I'll contact you through your web site for follow-up.

FYI, I was licensed as a surveyor in Nebraska nearly 40 years ago. Wonderful state.

It is my personal belief that there are no instruments available to the general surveying community today that could remeasure the geodetic baselines established by USC&GS more than a century ago. The reason is that they were measuring these baselines in microns with instruments such as the Ice Bar or Duplex Bars that required the use of microscopes to record the readings. They were also keenly taking into consideration atmospheric conditions upon every single measurement. The Duplex Bars were under a tent that was pulled by horses. Fifteen different men assisted in making each 5-meter measurement. Three separate thermometers were placed at each reading. Any measurement made today with an EDM or with GPS would only be a distance comparable to one made with a like instrument and not directly comparable to the length as established by USC&GS.

But, to answer your question, the lengths have not been remeasured and it would be an interesting experiment to see how far "off" a modern measurement would be compared to the exact measurement made by USC&GS.

Page Base Line - Then/Now

The Page Base Line was measured to be 8251.7569 meters in 1900. NGS has current NAD 83(1995) positions of:

NE Base
N 42°28'53.47527"
W 098°22'13.99939"

SW Base
N 42°25'25.39941"
W 098°26'00.79833"

The current values are adjusted from the original positions.

Does the distance between these two positions equal 8251.7569 meters when describing the length of this base line? Comments?

Page Base Line - Then/Now

Need some elevations

Page Base Line - Then/Now

NE Base - 580.1 meters NAVD88
SW Base - 626.231 meters NAVD88

Page Base Line - Then/Now

During coffee break I messed around a bit.

Using GRS80 and no heights;

Using Clarke 1866 and no heights;

Using 3D with (computed) ellipsoidal heights;

3D mark to mark distance is pretty close...

I'm thinking the 5 3/8" difference is in the sea level reduction of the original Datum (probably United States Standard Datum or North American Datum before NAD'27) and the adjustment of the triangulation net into NAD'83.

Page Base Line - Then/Now

Since the coordinates are NAD83:

Looks like about .006m just a bit over

I will go with the original measurement.;-)

maybe we could change it a bit using a different GEOID model, I used GEOID99 cause it's the first default in my program, I didn't think it would matter much in flat Nebraska, but look at the differences, .32m in such a short distance, quite a change

Page Base Line - Then/Now

I would think that in the NAD83 (1995) adjustment, which included conventional observations,that the measured baseline was used as a weighted constraint. Of course it probably has some (small) residual, depending on how it was weighted.

I used to be able to extract observations from the NGS IDB, but not anymore. So, I don't know what value is actually in the database. As some have alluded to, there are issues as to what exactly is that distance? "Horizontal" (depends on elevation), mark-to-mark (unambiguous), or ellipsoidal (depends on ellipsoid and the geoid model used in the reduction).

Page Base Line - Then/Now

The measured distance is what I understand we are trying to recreate, this would be an incrementally measured horizontal distance. At least, that's how I see it.

Page Base Line - Then/Now

Between two points there are theoretically an infinite number of different horizontal distances, it all depends on elevation. In reality there are two common ones-one at the low elevation and one at the high elevation.Iif reduced to the ellipsoid there is a chord distance and an arc distance, not a lot of difference between the two at 8.2 km but I am not sure how much, I am sitting in an airport waiting on a flight.

The NGS IDB (integrated database) stores measured distances as mark-to-mark, I believe exclusively but I may be wrong on that. It is what I always use in my databases and in processing, the m-to-m is unambiguous. It is the same as what most GPS software produces, a m-to-m vector. Certainly when they measured this baseline in 1900 they had limited knowledge of where the ellipsoid was with respect to the geoid, so the distance reduction to the ellipsoid probably has some bias due to this uncertainty.

Page Base Line - Then/Now

The high to low and low to high distance will be different of course, in some cases quite different, but those will be falling off a mountain.

This is a case where the measurement was done using short distances and a leveled device, set on tables. I will guess it was done both directions. Exactly how that will relate to modern day measurements, I will say it's the ground distance on my inverse slide. 🙂

After reading the interesting comments of others, I just could not resist adding my two cents. See http://www.globalcogo.com/Baseline-in-Nebraska.pdf The agreement is quite impressive.

Thanks for going through that, Earl. I think your comps have shown that it was indeed constrained quite strongly in the NAD83 (1995) adjustment, which is what one would expect.

I did inverses with the 4 sets of coordinates. I am surprised at the difference between NAD83(86) and NAD83(95), 4.5 mm. Note: these are ellipsoid distances:

NAD83(95): 8251.0138 m
NAD83(86): 8251.0183 m
NAD27: 8251.0058 m
USSD: 8251.0093 m (US Standard Datum, predecessor to NAD27, used same origin at Meades Ranch)

The NAD27 and USSD coordinates were only published to 3 decimals of latitude/longitude, which is about 3 cm on the ground. So, the above distances for those datums are really ± a few cm. I wonder how they reduced the NAD27 (and USD) to the ellipsoid? If they simply reduced it to "sea level", that would result in a difference of 3.9 ppm (24.7 m geoid height), which is 3.2 cm different. But, as I mentioned above, the positions are only to the nearest 0.001 seconds of latitude/longitude, so it is impossible to tell what the exact value used in those adjustments really was from this data.

An interesting thing happened around here (Pittsburgh). I have described the City of Pittsburgh triangulation (1920's) in past posts, and also posted the report of the survey. The old timers like me will remember the 1° X 2° control diagrams put out by the NGS showing triangulation, trilateration, traverse, and leveling. A measured baseline was shown on the diagram between two city triangulation stations which were later occupied by USCGS. I was pretty sure that they were never in the area during the EDM era, so I asked how that line was measured. It turns out that NGS used a computed distance from the City of Pittsburgh triangulation as a "measured baseline" in the NAD 1927 adjustment. The city network had 10 measured (taped) baselines, and the line used by NGS was computed based on these measured baselines and the triangulation network. I believe this was done because it would be very difficult to find a place for a long baseline in western PA. Later in 1975 they did a lot of trilateration south of Pittsburgh, and also they came through the area south (Uniontown) in 1969 with the Transcontinental Traverse (TCT). I believe this was the only place in the national network this was ever done. The person who created the Pittsburgh network went later to a career with USCGS, which probably explains it.

Earl, you seem to have shown that the "published" horizontal value is at the mean elevation. Interesting, but it is what one would expect.

I also note the rather large difference in geoid heights, 0.33 m, which corresponds to a deflection of the vertical of 8" in the azimuth (NE-SW) of the line. More than one would expect for the "flatlands" of Nebraska. I know there is some type of gravity high in Iowa, but I didn't expect it in NE. And of course the rockies are west, but pretty far away.

John,

Thanks for the feedback. After sleeping on it and reading your posts, I'll be posting a revised summary later today.

For example, I will re-compute using assumed standard deviations - 0.001 meter in n/e and 1.00 meter vertical (reason - is the NGS published elevation for the surface mark or the sub-surface mark). With a 1 meter standard deviation on vertical, it won't matter whether it is surface or sub-surface. We'll see what difference it makes.

I am not saying the NGS used the "mean" horizontal distance. I am just noting that the mean as shown/computed is very close to the value used.

I'll post a reply once I post the revised summary.

Like I mentioned, I used to be able to log on to a server at NGS and retrieve observations from the database. I believe distances are always stored there as mark-to-mark (M-to-M). This whole exercise shows why the M-to-M is the best way to process/adjust/archive values measured in the field. Probably many surveyors do not understand the M-to-M concept when it comes to distances and zenith angles. Nor do many realize that is what GPS gives as a baseline vector. In Trimble, the RTK data is stored in the data collector as an antenna-to-antenna vector, which then needs to be corrected to M-to-M. TBC, TGO, etc all output the M-to-M values. Once a distance and zenith angle is converted, it can be directly compared to GPS. People try to compare a measured EDM distance, converted to horizontal to GPS values, or SPC inverses, and then call me when they don't match.

I really liked being able to access the database because I could see how stations were connected, get astro observations, see what distances were measured, etc. Maybe not as important anymore with GPS, but it was very important to me back when we had to use triangulation stations as control (pre-HARN/pre-CORS). I was cutoff because the method of access (telnet) was not secure. Another casualty of hackers. They said they would develop another method of accessing the data, but I have not seen any way to do it. Last I heard they were transitioning to a more modern database. I hope they implement some way to extract observations instead of just datasheets. The problem with the way they had the database retrieval setup was that I could not get vertical angle observations. So, it was impossible for me to use the M-to-M data in an adjustment unless the points were leveled. I had wanted to do some analysis of very long TCT lines across the country versus high accuracy positions with GPS to see if they were really getting 1 ppm accuracy on the traverses they ran from 1961 to 1976 around the country.

Epping Base

Someone GPSed the Epping Base back in 1991 and wrote an article.
Anyone have a copy or link?

And notice how I skillfully avoided the wrath of NOLTON? 🙂

Epping Base

I couldn't find the article in a brief search, but did find a mention here of such an article by Andrew Kekacs.
http://cherryfieldhistorical.com/members/stories/baseline/baseline.shtml
A quote from there:
the GPS measuring project, spearheaded by Harold E. Nelson, ... found the Epping Baseline to be accurate to within a centimeter

This article has no GPS data, but mentions Epping Base and discusses early triangulation work and telegraphic determination of longitude.
http://www.pobonline.com/articles/89984-rebuilding-the-chimney

Data sheets:
http://www.ngs.noaa.gov/cgi-bin/ds_mark.prl?PidBox=PD1145 East
http://www.ngs.noaa.gov/cgi-bin/ds_mark.prl?PidBox=PD1170 West

Amateur recoveries w/ photos
http://forums.groundspeak.com/GC/index.php?showtopic=207678
http://www.geocaching.com/mark/details.aspx?PID=PD1145
http://www.geocaching.com/mark/details.aspx?PID=PD1170