RTK Baseline Error Budget and Vertical Uncertainty

An RTK specification is written as a fixed component plus a parts-per-million term, and both halves matter for different reasons. The fixed part is the receiver's own floor and it does not improve with a shorter baseline; the ppm part grows with distance from the base, because the atmosphere the correction models diverges between the two receivers as they separate. Eight millimetres plus one ppm means eight millimetres at the base and eighteen at ten kilometres. THE VERTICAL IS ROUGHLY TWICE THE HORIZONTAL AND THE REASON IS GEOMETRIC RATHER THAN ELECTRONIC. Satellites are all above the receiver and none below it, so the intersection geometry that fixes a horizontal position from many directions has only one side to work with in the vertical. That ratio holds at every baseline and with every receiver, and it means a job whose result depends on elevation should know before it relies on RTK heights. THE FAILURE THAT DWARFS BOTH IS THE BASE POSITION, and it is not a precision problem at all. If the base was set on an autonomous position rather than on a known control point, its coordinate can be off by a metre or more -- and every rover observation of that session inherits that error EXACTLY. It is systematic, not random: it does not average out over occupations, it does not shrink with a better receiver, and it does not show up in the internal quality figures the data collector displays, which describe the vector from the base and not the base itself. This adds it directly rather than combining it in quadrature for that reason. A float solution is not survey grade. Only a fixed integer solution carries the precision a specification quotes, and a data collector that reports float has not resolved the carrier ambiguities -- the coordinates it stores look identical to fixed ones in the file. And the check that costs two minutes: occupy a known point at the start and at the end of every session. It catches a wrong base coordinate, a wrong antenna height, a wrong datum, and a solution that drifted, and none of those is visible any other way. Antenna height is worth naming on its own, because a mis-measured or mis-typed antenna height is a pure vertical blunder of exactly that size on every point of the session. A precision budget, not an accuracy statement. It does not address multipath, which is site-dependent and can exceed everything here; satellite geometry and its dilution of precision, which varies through the day; ionospheric activity, which degrades long baselines badly during solar maxima; the network RTK case, where the correction is interpolated and the baseline concept differs; or datum and geoid model errors, which are systematic and often larger than any of it. Geoid models convert ellipsoid heights to orthometric ones and carry their own uncertainty, which this does not include. The receiver manufacturer's specification, the project's accuracy requirements, the control network, and the surveyor of record govern.

horizontal error = the fixed component + the ppm component x the baseline in km; vertical the same with its own coefficients, and roughly twice the horizontal; a base position error ADDS directly to both rather than combining in quadrature, because it is systematic.

The RTK error budget convention by name. The vertical is about twice the horizontal because satellites are all above the receiver and none below. A BASE POSITION ERROR IS SYSTEMATIC: every rover observation of the session inherits it exactly, it does not average out, and it does not appear in the data collector's internal quality figures. A float solution is not survey grade. A precision budget, not an accuracy statement: multipath, dilution of precision, ionospheric activity and geoid model uncertainty are not included. The receiver manufacturer's specification, the project's accuracy requirements, the control network, and the surveyor of record govern.

One multiplication on a manufacturer's published specification; no receiver data is reproduced.

Estimate. AHJ and licensed professional govern.

Field names used by the API: baseline_km, horizontal_fixed_mm, horizontal_ppm, vertical_fixed_mm, vertical_ppm, base_position_error_mm, target_vertical_mm, horizontal_ppm_mm, vertical_ppm_mm, horizontal_error_mm, vertical_error_mm, vertical_ratio, total_horizontal_mm, total_vertical_mm, base_error_multiple

• A precision budget, not accuracy multipath, DOP, ionosphere and geoid uncertainty are excludedthe project's accuracy requirements
• Base error adds, it does not combine it is a systematic shift on every observation of the sessionthe control network
• Only a fixed solution carries the specification a float solution stores coordinates that look identical in the filethe receiver manufacturer's specification