Jump to content
Sign in to follow this  
Followers 1

Background on Horizontal Datum realizations

Recommended Posts

In doing the experiment described in a related post, I've tried to sort through the various datums, datum tags, and other mysterious designations on the data sheets. This is more detailed information than most forum posters get into, so I thought I'd summarize what I think I know about datum realizations. Correct me where I'm wrong.


Some of this summary comes from:




You may also want to read articles such as:







Following regional triangulations begun by Hassler, the US Coast & Geodetic Survey worked through the 1800's and connected across the continent and formed the US Standard Datum. Adjustment of the data collected into the early 20th century resulted in the US Standard Datum in 1901 and then an update to the North American Datum of 1927 (NAD27). These used the Clarke ellipsoid model of the earth and were defined relative to Meades Ranch. NAD27 had a network accuracy of about 10 meters and local accuracy of 1:100,000.


After several decades of use and the gathering of more data (including pre-GPS satellite doppler data and VLBI vectors), a new datum was defined. It used the internationally adopted GRS80 ellipsoid earth model that fits the whole world better than Clarke and (with minor differences) is essentially the earth model used for GPS. It was defined as the North American Datum of 1983 and became available in 1986, hence NAD83(86). This realization had a network accuracy of about 1 meter and local accuracy relative to nearby stations of 1:100,000. That would be 2 arc seconds or 1 foot in 19 miles at 95% confidence.


The topographic maps of the country were prepared in NAD27 coordinates. You will find that even a recreational GPS can see the difference between that and NAD(83); in my area it is about 50 feet so you need to account for it when using topo map with GPS.


GPS operates with a datum that is used internationally, IGS08 which is for most purposes the same as the International Terrestrial Reference Frame (ITRF2008). The US military calls their version WGS84 and update it to match the international datum every several years with a datum tag of the GPS week, like (G1674).


NGS chose to fix NAD83 to the North American tectonic plate to minimize coordinate changes. The two datums drift apart by a couple centimeters per year. Additionally, it was learned that the initial realizations of those two datums were not as close as hoped, so the difference is more like a couple meters. Most recreational GPS units ignore the difference. For precise work the NGS horizontal time-dependent positioning program (HTDP) provides good estimates of the conversion at a given time epoch.


Since then additional GPS data has been used to improve the NAD83 realization. State-by-state updates were done in the 1990's under the name High Precision GPS Network (HPGN), renamed High Accuracy Reference Network (HARN) which you see on signs as you recover some stations. Some states had two versions, so the latest in any state is called the Federal Base Network (FBN), a label you see on data sheets. You also see Cooperative Base Network (CBN), which is a densification of the high accuracy stations.


The 199x FBN adjustments updated all horizontal control stations in the data base using additional VLBI, CORS, and other GPS measurements. Successive adjustments held the highest precision stations (VLBI and CORS) and adjusted other GPS stations to them, and then adjusted stations with no GPS data. Only a minority of the old tri-stations had HARN GPS data. Additional HARN stations resulted from GPS measurements on elevation bench marks that had no prior accurate lat-lon. The FBN stations improved accuracy by a factor of at least 10 relative to NAD83(86) but the tri-stations that were not occupied by GPS improved less.


HARN/FBN was still NAD83 because it used the same definitions, but the realization was different because it used additional measurements that shifted things a bit, hence names like NAD83(1995). I think of realizations like a fish net. After you tie down the lat-lon of some points on the net, there is still some sag or wandering by the rest of the net. Each realization gets its points better aligned, but not all points have equally good data.


NGS created NAD83(CORS96) which included only CORS stations and had an accuracy of about a centimeter.


The next realization (update) was NAD83(2007) with the goal to remove inconsistencies between the state adjustments and with CORS. It held fixed the NAD83(CORS96) positions. This time only GPS data was used, so triangulation stations that didn't have GPS data were omitted, and you still see their NAD83(199x) coordinates at the top of their data sheets. Those old coordinates typically have a fraction of a foot to even a few feet of local warpage relative to the stations with GPS data, which is not modeled because it is random error in the old measurements. Intersection stations are going to be less accurate than triangulation stations.


The latest is NAD83(2011, epoch 2010.0), which you see on data sheets having good GPS data. If a project design is based on an older realization, then measurements tied to the adjusted monuments can be converted accurately enough to 2007 and FBN, but stations that were not occupied with GPS will usually not match well.


(corrections welcome)

Edited by Bill93

Share this post

Link to post

The epoch is extremely important over here in California where we have significant velocity caused by plate tectonic motion and occasional large earthquakes.


I find that HTDP datum and epoch adjustments of existing CSRS 1991.35 HARN/HPGN monuments don't give me the same warm and fuzzy feeling that a new current measurement does. GEOCON adjustments are even worse. I'm talking about differences of two or three tenths of a foot when comparing an HTDP datum and epoch adjusted position to a newly measured position (24 hours of raw data static post processed through OPUS Projects using either rapid or final orbits). I gather HTDP just isn't capable of that level of accuracy.


I've been playing around with OPUS Projects lately. I think it's the bomb. If you haven't already done so, sign up for one of those online OPUS Projects webinars so that you can get access to OPUS Projects. If you set it up correctly it does a better job than either OPUS or OPUS-RS.


I haven't seen you post anything here yet about vertical height datums such as leveled NGVD29 vs. leveled NAVD88 vs. GEOID12B or the new gravity model based one coming out in 2022 or whenever.

Share this post

Link to post

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

Sign in to follow this  
Followers 1

  • Create New...