A 5.414Mb PDF of this article as it appeared in the magazine—complete with images—is available by clicking HERE
For surveyors keenly interested in history, it can be a passionate journey to recover monuments that have disappeared from the modern public record. Retracing the steps of the original surveyor and interpreting century old descriptions are often challenging tasks, at times both frustrating and rewarding. Obtaining knowledge of the original surveyor or the agency that placed the monument is often the key to recovering them.
The monuments of the Missouri River Commission (MRC), placed as part of a precise triangulation network along the Missouri River in the late 1800’s, passed on one vital piece of information–a published latitude/longitude position. These positions found in the 1891 Annual Report for the MRC were based upon their own datum which was not initially tied to the networks of other government agencies. (See The American Surveyor Vol. 10, No. 6, 2013).
When the U. S. Coast & Geodetic Survey (C&GS) expanded their triangulation network across the nation, some of the MRC monuments, when found, were occupied. One reason for using an existing monument is because it was often already located at a key location such as a hilltop or area that was visible from the surrounding area. The MRC triangulation monuments were ahead of their time and consisted of a 5½" diameter cast iron cap placed on a 4-inch diameter iron pipe 36" long. Below the pipe was a drill hole in an 18" square buried stone.
The early MRC surface monuments quickly became victims of theft by curio seekers due to their elaborate design. Unbeknownst to the thieves, the buried stones almost always remained. The lack of the visible surface monument, however, proved to be a dilemma for the C&GS surveyors who came later to the same hilltop knowing there should be a monument at the general location, but finding nothing. Probing for the buried stone was often time consuming. Calculating an accurate position could not yet be done since they were still in the process of creating their network which was also on a different datum. In many instances, C&GS therefore placed their own separate monument in the same general location. Such was the case for one particular triangulation station in Gregory County, South Dakota, located along the west side of the Missouri River.
When the MRC surveyors arrived in the area in 1889, they chose the name "Brickkiln" for their station due to a prominent nearby landmark that was a barren sandstone formation rising upwards in a red colored spire. Early residents in the area carved their names on the sides of the spire to leave their mark. The sandstone formation is located 900 yards from the shore of the Missouri River and is 2.0 miles north of the location where the MRC surveyors strategically placed their monument. Today, the natural landmark is locally known as "Red Rock".
Undoubtedly, the C&GS surveyors came to the same general location in 1934 in search of the 1889 MRC monument, only to discover that the capped iron pipe on the surface had been removed. A general search was probably made, but the area is on a large rounded grassy hill where many possibilities would be suitable for a monument. The C&GS surveyors therefore placed their own monument, a bronze disk in concrete, and named it "Brickkiln 2".
When the C&GS network was complete, "Brickkiln 2" was adjusted to the North American Datum of 1927 (NAD 27) with its precise position. There were then two monuments somewhere on the same hilltop with precise positions, but each was on a different datum. Finding the lost MRC monument would typically be just a matter of inversing the two positions to obtain the azimuth and distance between them. The situation, however, was more complex because the datum of the MRC of 1889 could not be directly related to the NAD 27. To accurately solve the dilemma of finding the older monument, a datum shift would first have to be determined that would then relate the MRC’s datum to the NAD 27.
One method that has proven successful in finding these older monuments is to find an initial monument and then obtain a modern GPS position for it. Comparing this modern position with the published position from the historical record then provides a shift that can be applied to aid in finding the other monuments in the same local area.
In this region of South Dakota, however, the initial monument on which to calculate a shift had not yet been found. Upon searching for and reading the National Geodetic Survey (NGS) datasheet for C&GS station "Brickkiln 2" (OR0886), it was discovered that there was also a datasheet available for the MRC monument "Brickkiln" (OR0887). A current position was listed for the MRC station even though this monument had probably not been found in well over a century, and as far as anyone knew, was not an existent monument.
I inquired with David Doyle at NGS as to why the old MRC monument was listed in the NGS database with a modern position. Doyle replied that he had endeavored to find many pre-NAD 27 positions for the stations of various agencies and to get them loaded into the NGS database. A rigorous least-squares adjustment was then performed on the older stations to arrive at the modern positions. A table of superseded positions was also developed dating back to the original MRC position. With two points on the same datum, I returned to the hill where the two monuments were located and determined the azimuth and distance from the 1934 C&GS monument to where the 1889 MRC monument should be located–a mere 17.79 feet.
Despite a snow-covered day in December 2012, the ground was not frozen, allowing a probe to sink in and strike the buried MRC stone three feet below the surface. I returned in July 2013 to uncover the stone and remonument the MRC position by placing a surface monument directly over the buried stone.
The remaining question was just how close had NGS calculated the position of the historic MRC monument that was originally on its own independent datum? In late 2015, I performed two static observations on the MRC monument one month apart and processed the data through the Online Positioning User Service (OPUS). The difference in the results were only 9 mm north-south and the same east-west, so I felt confident of the position. When the mean of the OPUS derived positions was then compared to the position listed on the NGS datasheet, the difference was 0.00633 seconds in latitude and 0.00397 seconds in longitude (19.5 cm south and 8.9 cm east). This is remarkable considering that NGS had never occupied or even seen the original MRC monument to begin with. The historic MRC position, precise to only two decimal places of a second, was on a different datum that had been done 45 years prior to the C&GS survey.
I decided to also obtain static observations on the nearby C&GS monument "Brickkiln 2" in order to compare what OPUS would show when compared to its published position. The results from the published datasheet position to the OPUS position was 0.9 cm south and 18.6 cm east. I was now comparing the results of a 1934 theodolite triangulation survey with that of 2015 contemporary GPS observation. The north-south distance seemed reasonable, but the east-west distance seemed a bit too far off from expected.
I once again sought the advice of the expert, now retired David Doyle, who explained that even First-Order stations, when compared to contemporary GPS observations, can have differences of 40-50 cm and it would not be unusual to even see a difference of one meter. Doyle further explained that the older triangulation in the country had been readjusted to fit each state’s High Accuracy Reference Network (HARN) in the 1990’s, but nothing had been done with them since. The positions on most triangulation stations have too much internal error to warrant an effort of chasing down a few centimeters.
According to Doyle, many parts of the C&GS triangulation networks across the nation have suffered severely from scale issues due to the lack of long and well determined baselines. With the introduction of EDM in the 1950’s, the issue began to improve. Older networks such as the one performed in South Dakota during the 1930’s were scaled with baselines that were seldom more than 10 miles long and were then used in triangulation schemes that were hundreds of miles long. One can then begin to understand the issues associated with the mathematical aspect, but until surveyors had EDM and GPS, they typically could not see these distortions.
This distortion can create a dilemma for surveyors who often rely upon the published positions of NGS as the end-all and final say, but then discover that their independent observations do not always match well when counting sub-centimeters. Doyle recollected that while still at NGS he and others gave many workshops stressing this issue to hundreds of surveyors, but that covered only a small percentage of all surveyors. Many of those in attendance had only a passing knowledge of geodesy and probably failed to grasp the entire situation.
The early triangulation networks established by USC&GS and other agencies such as the Missouri River Commission are not perfect, but when given the broadest understanding of what was achieved and with the constraints and equipment of those eras, the work was nothing less than outstanding.
Note: David Doyle contributed to this story.
Jerry Penry is a licensed land surveyor in Nebraska and South Dakota. He is a frequent contributor to the magazine.
A 5.414Mb PDF of this article as it appeared in the magazine—complete with images—is available by clicking HERE