Every official latitude, longitude and height published in the United States is going to change, by as much as four meters. The ground being described will not have moved at all. What is being replaced is the reference system those numbers are counted from.
The National Geodetic Survey is retiring the North American Datum of 1983 and the North American Vertical Datum of 1988. Those are the horizontal and vertical reference systems that currently underpin surveying, flood mapping, construction and aviation across the country. In their place will come four new terrestrial reference frames and one new geopotential datum. The agency has kept the national reference system since its predecessor, the Survey of the Coast, was founded in 1807.
The reference surface is what is changing
Nothing physical shifts. A property corner in Ohio will sit exactly where it sat before, and a mountain in Colorado will be exactly as tall.
A coordinate is not a property of a place; it is a measurement of that place against an agreed reference surface. The United States is changing the reference surface. The four meters is the gap between two descriptions of the same unmoved point.
That figure is not uniform either, and the agency does not present it as a single national number. Its guidance says the magnitude of change depends on which datum a user currently works in, where in the country they are, and which epoch the new coordinates refer to. Four meters is the outer edge of the range. NGS publishes separate maps for ellipsoid height change, orthometric height change, and horizontal change on the North American and Pacific plates.
Errors older than satellite navigation
NGS gives several reasons for the change, among them alignment with international standards and with satellite systems that orbit the Earth’s center. Two of the reasons are systematic errors built into systems that were defined before satellite positioning was routine.
The first is horizontal. NAD 83 was fixed when the center of the Earth could not be located with the precision satellites later supplied, and its origin does not sit where the center of the Earth is now understood to be. The agency’s summary puts the misalignment at about 2.2 meters.
NGS gives two different figures for that misalignment on the same page. Its opening summary says about 2.2 meters. A later section says the origin of the current NAD 83 frames disagrees with international estimates of the Earth’s center by two to four meters, depending on which frame is meant. There are three NAD 83 frames, which may account for the spread, though the page does not reconcile the two statements. The smaller figure is the one the agency leads with.
The second error is vertical. NAVD 88 is the surface American elevations are measured from. NGS describes it as both biased, by about one half meter, and tilted, by about one meter coast to coast, relative to the best global geoid models available today. An elevation in Maine and an elevation in Oregon are not being measured against quite the same zero. Elsewhere the page states what looks like the same comparison as a single range, with the zero-elevation surface implied by NAVD 88 disagreeing with current estimates of the geoid by about one half to two meters.
A gravity map flown over the whole country
Fixing the vertical error meant measuring gravity, because the surface elevations are counted from is defined by gravity, which nothing visible marks.
That was the work of Gravity for the Redefinition of the American Vertical Datum, a project that collected airborne gravity readings over the entire United States. Those readings went into the time-independent component of a geoid model called GEOID2022. That component is a static surface representing the geoid at the reference epoch 2020.00, while the model the modernized system relies on is itself time-dependent. GEOID2022 defines the zero-orthometric-height surface of the new vertical datum, the North American-Pacific Geopotential Datum of 2022, which will replace every existing geodetic vertical datum in the system, not only NAVD 88.
Frames that turn with their plates
The horizontal side is handled differently. There are four frames, one for each of four tectonic plates: North America, Pacific, Caribbean and Mariana. Each is named for its plate and will turn at that plate’s average rate, so latitude and longitude in the stable interior of a plate should drift less over time than they otherwise would. All four will be defined against the International Terrestrial Reference Frame of 2020.
From a million brass discs to a constellation
The current arrangement rests on physical objects. NGS describes the existing system as realized through more than 1,000 continuously operating satellite stations together with an aging network of more than 1,000,000 survey marks set into the ground. Those marks are the discs that hikers find on summits and surveyors hunt for in ditches. NGS’s own description is that they are often deteriorated, destroyed or unmaintained, which leaves the heights and positions attached to them outdated and inaccurate.
The modernized system moves control off the ground and into orbit. Access will run primarily through the NOAA reference station network and the International GNSS Service, and NGS calls this a paradigm shift away from reaching the system by physically occupying a survey mark. The marks do not vanish from the record. They will still carry published coordinates, and surveyors will still be able to contribute satellite data collected on them.
Published positions will carry an epoch, because a position without a date stops being meaningful in a system precise enough to see the continent move. NGS says it continues to plan for the first reference epoch to be 2020.00.
On sharing, the agency’s pages do not currently agree. One FAQ answer says shared user solutions, surveyors publishing their own computed positions, will not be supported, because those solutions were never integrated with the official coordinates. The same FAQ also says users will be able to share solutions, and the Track Our Progress page the FAQ points to says OPUS-Share will be retained with a caveat that a shared solution should not be used as geodetic control. Contribution itself does not end. An updated version of the agency’s static processing service will still accept user satellite data, and the coordinates that result will be computed and published by NGS.
The timeline has already been pushed back once
The switch has not happened yet.
In 2020 the agency published a Federal Register notice saying the rollout was expected between 2022 and 2025. In October 2024 it published another, revising that to 2025 or 2026. That notice also set out the sequence: components released for public testing, at least six months of testing after the last one lands, then a vote by the Federal Geodetic Control Subcommittee to approve the new system. It called that vote likely in 2026, and the agency’s guidance narrowed it to mid 2026.
Mid 2026 has passed, though the notice’s own window has not. NGS said it would mark the milestones with further Federal Register notices: one in late 2025 announcing that all components were available for testing, and one in 2026 after an approving vote. The Federal Register can be searched for exactly that. Notices from the National Oceanic and Atmospheric Administration mentioning the National Spatial Reference System, published since June 2024, come back as three documents. Only one of the three concerns the reference system, and it is the October 2024 notice itself. Neither promised notice appears to have been published. Until the post-vote notice is, NAD 83 and NAVD 88 remain the official datums of the United States.
The names will not be adjusted to match the year the system actually arrives. Every component carries 2022 in its title, and NGS has said it intends to keep those names, on the grounds that the year carries no scientific meaning and the terms have been public for years. It points out that this has happened before. NAD 83 was released in 1986, and NAVD 88 in 1991.
The deadline already on the calendar
Geodetic survey projects that create or update control in the current system must reach NGS by January 13, 2027, and the agency will accept no new submissions after that date. The existing software remains available until the modernized system is fully implemented, but no new work enters the old record. The agency’s explanation is procedural rather than dramatic: the modernized system requires new databases, workflows, products and software services that differ from the current ones.
What follows is a slower legal unwinding. NGS’s FAQ reads the Geospatial Data Act of 2018 as requiring federal civilian agencies that collect, produce, acquire, maintain or disseminate geospatial data to use the new reference frames within five years of their adoption by the Federal Geographic Data Committee. The same FAQ says any state agency or organization receiving federal funds to generate or process geospatial data must conform; others, it says, should follow applicable state or local laws or contract terms. Some states may need their statutes changed, and for those, surveying organizations and NGS jointly drafted template legislation, written loosely enough to refer to the most recent datum and name none, so the same fight need not happen again. Flood insurance rate maps sit downstream of the vertical datum, and NGS ran a pilot with FEMA in North Carolina as far back as 2010 to find out what a new height surface would do to them.
What is being replaced is an agreement about where zero is, and about which point in space every published number in the country is counted from. The agency’s own notice calls NAD 83 and NAVD 88 the last significant modernization, and says no significant update was possible until now. The discs in the ground will stay where they are, still carrying coordinates, but the system will no longer be reached by kneeling on top of one.
