Applied Geodesy

Nobody's drone map gets rejected because the photogrammetry was bad. It gets rejected because it's in the wrong reference frame — WGS84 ellipsoid heights delivered to a client who lives in State Plane with NAVD88 elevations. The geometry was beautiful; the basis was wrong; every point is shifted or floats a geoid height above the ground.

Applied geodesy is the discipline of saying, precisely, where your coordinates live: which datum, which projection, which height system, which epoch. Four questions, asked before fieldwork, that make every deliverable fit.

A datum is the reference frame your coordinates are expressed in — an ellipsoid (the smooth mathematical earth) plus a definition of where it sits and how it moves.

• NAD83. The North American datum — what State Plane zones, NGS control, and effectively every US survey product is built on. Pinned to the North American tectonic plate, so US monument coordinates stay nearly constant year to year.
• WGS84 / ITRF. The global frames GNSS natively positions in. Defined for the whole earth, so they don’t ride the North American plate — relative to NAD83 they drift a couple of centimetres a year.
• The offset. NAD83 and current WGS84/ITRF disagree by roughly 1–2 metres in CONUS, and the difference changes slowly with time (hence epochs like 2010.00). If your base coordinates are in one frame and the client expects the other, the whole job shifts by that amount — consistently, invisibly, and wrong.
• The practical rule. Corrections inherit the frame of their source. OPUS and state CORS networks hand you NAD83 — most US workflows never need a manual transformation. The danger zone is mixing sources: a base set by OPUS with a drone PPK’d against a global PPP service now has two frames in one project.

GNSS measures height above the ellipsoid. Water flows, and clients design, against gravity — orthometric height above the geoid. Confuse the two and everything is off by the local geoid undulation: up to ~30 m of error in CONUS, the single largest mistake available in this field.

• The tell. A surface that is self-consistent but ~25–35 m above or below known ground is a geoid mistake, not a measurement mistake.
• The check. Shoot one benchmark with a published NAVD88 elevation before production. If your H matches, the whole height pipeline — antenna, geoid model, datum — is proven at once.

Plans are flat; the earth isn't. A map projection trades a controlled amount of distortion for the convenience of working in feet of northing and easting. State Plane (SPCS) is the US system of ~120 zones, each kept small enough that the distortion stays near 1:10,000.

• Two projection types. Zones that run east–west (Tennessee, North Carolina) use a Lambert Conformal Conic; zones that run north–south (most of California’s, Illinois East/West) use a Transverse Mercator. Same idea either way: a developable surface touching the earth where the zone lives.
• Grid distance ≠ ground distance. Project the curved earth to a flat grid and distances change slightly. The scale factor says by how much at your location; the elevation factor accounts for your height above the ellipsoid. Multiplied together they’re the combined factor.
• Why a 500 ft wall measures 499.95 on the plan. A combined factor of 0.9999 means every grid distance is 1 part in 10,000 shorter than ground. Trivial on a lot survey; 0.5 ft per 1000 ft on a highway job — real money in construction staking.

Many site jobs sidestep the factor by working at ground on a local or low-distortion projection — in which case the deliverable must say so loudly. A drawing whose coordinates are “State-Plane-ish but scaled to ground” with no metadata is a trap for every surveyor who touches the site after you.

• 1. Horizontal datum + epoch? NAD83 (2011) epoch 2010.00 for nearly all US work. Write it on the deliverable.
• 2. Projection? Which State Plane zone — or which local/low-distortion system — and grid or ground? If ground: what combined factor, scaled about what point?
• 3. Vertical datum + geoid? NAVD88 via GEOID18 today. (NGS’s modernized frames will eventually replace both NAD83 and NAVD88 — one more reason metadata matters.)
• 4. Units? US survey foot vs international foot differ by 2 ppm — about 0.1 ft per 50,000 ft of coordinate. Wrong-unit State Plane coordinates land tens of feet off, a famously confusing failure.

Answer those four in the project file and on the drawing, and your data fits everyone else's — this is the entire trick behind “professional-grade” deliverables.