Every published control coordinate has a provenance. Somebody occupied the mark, the observations were processed against something, the result was adjusted into a network, and a value was published with an accuracy statement attached. Knowing that chain is what lets you judge whether a coordinate is worth holding fixed or worth treating as a starting guess.
In the modern United States the chain runs through the CORS network, and for most surveyors the entry point is OPUS.
The CORS network
- CORS
- Continuously Operating Reference Station — a permanently installed GNSS receiver on a stable mount, logging data continuously and submitting it to a coordinating agency. The NOAA CORS Network aggregates stations from federal, state, academic and private operators into a single managed system.
CORS stations are the active control that realizes the national reference frame. Their coordinates and velocities are determined from long time series of continuous data processed in the international frame and then expressed in the national datum. Because they observe continuously, their motion is measured rather than assumed, which is what makes it possible to publish a velocity as well as a position.
That is a genuine reversal of the historical model. Under NAD 27 and the early years of NAD 83, control was passive: bronze disks in the ground, coordinates determined once and assumed to hold. Under the modern system the primary control is active and continuously monitored, and the passive marks are secondary, checked against the active network rather than defining it. The modernized frame carries this further still.
OPUS
- OPUS
- The Online Positioning User Service — a free NGS service that takes a submitted GNSS data file, processes it against nearby CORS stations, and returns a position in the national datum together with an assessment of the solution quality.
OPUS is not a black box that produces truth. It is a specific processing recipe with known behaviours, and reading its output properly is a skill.
| Service | Typical occupation | How it works | Realistic accuracy |
|---|---|---|---|
| OPUS Static | 2 hours to 48 hours | Dual-frequency carrier phase against three CORS, double-difference baselines | A couple of centimetres horizontally, worse vertically |
| OPUS Rapid Static | 15 minutes to about 2 hours | Uses a larger set of nearby CORS and atmospheric modelling | Centimetres, degrading quickly with short occupations and sparse CORS |
| OPUS Projects | Multiple sessions, whole campaigns | Managed multi-session network adjustment with user control | Best available; supports formal submission of results |
How a coordinate becomes published
- The mark is set in a stable setting and described so that someone else can find it decades later.
- It is occupied with a dual-frequency receiver for long enough to resolve the atmosphere, with the antenna height measured redundantly and the antenna type recorded.
- The observations are processed against CORS stations whose own coordinates and velocities are known.
- The results from multiple sessions and multiple marks are combined in a least-squares network adjustment, which redistributes error and yields formal uncertainties.
- The adjusted coordinates are propagated to the reference epoch of the datum realization.
- The result is published with its datum, realization, epoch, accuracy estimate and a description of how it was derived.
Step four is the one most often skipped in ordinary practice, and it is the one that turns a position into control. A single OPUS solution on a single mark is an observation. Several sessions on several marks, adjusted together with redundancy, is a control network with defensible accuracy statements.
Judging a coordinate you did not create
- Which realization and epoch is it on? A coordinate labelled only NAD 83 is incompletely specified.
- Was it adjusted or scaled? A datasheet distinguishes adjusted values from scaled ones, and scaled values can be off by tens of meters.
- How old is the determination, and is the mark in ground that moves?
- How many independent occupations support it?
- What accuracy is stated, and is it a network accuracy or a local accuracy relative to nearby marks?
- Has the mark been recovered recently, and by whom?
The distinction between network accuracy and local accuracy is worth internalising. Network accuracy describes how well the point is positioned relative to the datum as a whole. Local accuracy describes how well it is positioned relative to its neighbours. A pair of marks can have mediocre network accuracy and excellent local accuracy, which is precisely the condition in which they are ideal for setting up a project and useless for tying to a national dataset.
Real-time networks are a different animal
A commercial or state real-time network gives you an RTK position in seconds, which is a different product from a processed and adjusted coordinate. The network solves in whatever realization and epoch its base coordinates are published on, and that is not always the same as your project. Ask, and record the answer.
The practical rule is the ordinary one for redundancy: RTK is excellent for topographic work and for setting points once the control is established, and it is not by itself a way to establish control. Establishing control means static observations, several of them, adjusted together, with published marks held and checked. That workflow has not changed with the technology; only the instruments have.