ITRF2014, WGS84 and NAD83

ITRF is Geocentric and NAD83 is not Geocentric

Source: GPS for Land Surveyors

The North American Datum of 1983 (NAD83) is used everywhere in North America except Mexico. The latest realization of the datum as of this writing is NAD83 (2011) epoch 2010.0. This realization in the coterminous United States and Alaska is available through the National CORS (Continuously Operating Reference Stations). Their number of National CORS and Cooperative CORS sites continuously grows with the addition of several new stations each month.

As mentioned earlier, in the past we did not have to be concerned with the shift between NAD83 (1986) and WGS84 as introduced in 1987, because the discrepancy easily fell within our overall error budget. NAD83 and WGS84, originally, differed by only a centimeter or two. That is no longer true. In their new definitions—NAD83 (2011) and WGS84 (G1762)—differ up to one or two meters within the continental United States. On the other hand, ITRF08, ITRF2014 and WGS84 (G1762) are virtually identical if their epochs (moment in time) are the same. The typical standard epoch for both are WGS84(G1762) and ITRF08 is 2005.0.  The typical standard epoch for both ITRF2014 is 2010.0; NGS has developed a program called Horizontal Time Dependent Positioning (HTDP) to transform positions from one epoch to another.  In other words, this program allows the movement of positions from one date to another, transformation from one reference frame to another and supports the recent realizations of the NAD 83, ITRF and WGS84. This underlines the fact that ITRF and WGS84 systems are global, and their realizations take into account that the earth is in constant motion due to the shifting of tectonic plates around the world. However, NAD83 is fixed to one plate, the North American plate, and moves with it. Consequently, NAD83, in the continental United States moves approximately 10 to 20 millimeters per year in relation to the realizations of ITRF and WGS84 reference frames 

The Management of NAD83

Since geodetic accuracy with GPS depends on relative positioning, surveyors continue to rely on NGS stations to control their work, just as they have for generations. Today, it is not unusual for surveyors to find that some NGS stations have published coordinates in NAD83 and others, perhaps needed to control the same project, only have positions in NAD27. In such a situation, it is often desirable to transform the NAD27 positions into coordinates of the newer datum. But, unfortunately, there is no single-step mathematical approach that can do it accurately. The distortions between the original NAD27 positions are part of the difficulty. The older coordinates were sometimes in error as much as 1 part in 15,000. Problems stemming from the deflection of the vertical, lack of correction for geoidal undulations, low-quality measurements, and other sources contributed to inaccuracies in some NAD27 coordinates that cannot be corrected by simply transforming them into another datum.

Transformations from NAD27 to NAD83

Nevertheless, various approximate methods are used to transform NAD27 coordinates into NAD83 values. For example, the computation of a constant local translation is sometimes attempted using stations with coordinates in both systems as a guide. Another technique is the calculation of two translations, one rotation and one scale parameter, for particular locations based on the latitudes and longitudes of three or more common stations. Perhaps the best results derive from polynomial expressions developed for coordinate differences, expressed in Cartesian or ellipsoidal coordinates using a 3-D Helmert transformation. However, besides requiring seven parameters (three shift, one scale, and three rotation components) this approach is at its best when ellipsoidal heights are available for all the points involved. Where adequate information is available, software packages such as the NGS program NADCON can provide coordinates.

Even if a local transformation is modeled with these techniques, the resulting NAD27 positions might still be plagued with relatively low accuracy. The NAD83 adjustment of the national network is based on nearly 10 times the number of observations that supported the NAD27 system. This larger quantity of data, combined with the generally higher quality of the measurements at the foundation of NAD83, can have some rather unexpected results. For example, when NAD27 coordinates are transformed into the new system, the shift of individual stations may be quite different from what the regional trend indicates. In short, when using control from both NAD83 and NAD27 simultaneously on the same project, surveyors have come to expect difficulty.

In fact, the only truly reliable method of transformation is not to rely on coordinates at all, but to return to the original observations themselves. It is important to remember, for example, that geodetic latitude and longitude, as other coordinates, are specifically referenced to a given datum (reference frame) and are not derived from some sort of absolute framework. But the original measurements, incorporated into a properly designed least-squares adjustment, can provide most satisfactory results.

Densification and Improvement of NAD83

The inadequacies of NAD27 and even NAD83 positions in some regions are growing pains of a fundamentally changed relationship. In the past, relatively few engineers and surveyors were employed in geodetic work. Perhaps the greatest importance of the data from the various geodetic surveys was that they furnished precise points of reference, to which the multitude of surveys of lower precision could then be tied. This arrangement was clearly illustrated by the design of state plane coordinates systems, devised to make the national control network accessible to surveyors without geodetic capability. However, the situation has changed. The gulf between the precision of local surveys and national geodetic work is virtually closed by GPS, and that has changed the relationship between local surveyors in private practice and geodesists. For example, the significance of state plane coordinates as a bridge between the two groups has been drastically reduced. Today’s surveyor has relatively easy and direct access to the geodetic coordinate systems themselves through GPS. In fact, the 1- to 2-ppm probable error in networks of relative GPS-derived positions frequently exceed the accuracy of the NAD83 positions intended to control them.

High Accuracy Reference Networks

Other significant work along this line was accomplished in the state-by-state super-net programs. The creation of High Accuracy Reference Networks (HARN) were cooperative ventures between NGS and the states, and often include other organizations as well. The campaign was originally known as High Precision Geodetic Networks (HPGN). A station spacing of not more than about 62 miles and not less than about 16 miles was the objective in these statewide networks. The accuracy was intended to be 1 part-per-million, or better, between stations. In other words, with heavy reliance on GPS observations, these networks were intended to provide extremely accurate, vehicle-accessible, regularly spaced control point monuments with good overhead visibility. These stations were intended to provide control superior to the vectors derived from the day-to-day GPS observations that are tied to them. In that way, the HARN points provide the user with a means to avoid any need to warp vectors to fit inferior control. That used to some time happen in the early days of GPS. To further ensure such coherence in the HARN, when the GPS measurements were complete, they were submitted to NGS for inclusion in a statewide readjustment of the existing NGRS covered by the state. Coordinate shifts of 0.3 to 1.0 m from NAD83 values were typical in these readjustments, which were concluded in 1998. The most important aspect of HARN positions was the accuracy of their final positions. 

The original NAD83 adjustment is indicated with a suffix including the year 1986 in parentheses, that is, NAD83 (1986). However, when a newer realization is available, the year in the parentheses will be the year of the adjustment. The most recent realization is NAD83 (2011).