RTK vs DGPS in Marine Survey: Physics, Accuracy, and Which Survey Type Each Serves Best

Differential GNSS is the foundation of marine survey positioning, but "differential GNSS" describes a family of techniques rather than a single method. The two techniques most often compared in hydrographic and construction survey are DGPS, which applies corrections to code-phase pseudorange measurements, and RTK, which resolves the integer ambiguities of the carrier phase. The difference is not simply one of accuracy. It is a difference in the physical observable being processed, and that difference determines which survey types each technique serves best.

The Physics: Code Phase and Carrier Phase

A GNSS satellite broadcasts a signal that carries two distinct measurable components. The first is a coded modulation, the pseudorandom noise sequence, which allows a receiver to align a locally generated replica with the received code and measure a pseudorange. The second is the underlying radio carrier wave itself, a continuous sinusoid at the L-band frequency.

Code phase measurement aligns a code sequence whose chip length corresponds to a coarse distance. For the GPS C/A code, the chip length is approximately 293 metres, and the receiver can resolve the alignment to a small fraction of a chip, typically producing pseudorange noise at the metre level or below. Carrier phase measurement, by contrast, tracks the phase of the carrier wave. At the GPS L1 frequency of 1575.42 MHz, the wavelength is approximately 19 centimetres, and a receiver can resolve the phase to roughly two millimetres.

Carrier phase measurement is far more precise than code phase measurement because the carrier wavelength is roughly four orders of magnitude shorter than the code chip length. However, the carrier phase observable is ambiguous by an unknown integer number of whole cycles between the satellite and the receiver. Resolving that integer ambiguity is the defining problem of RTK.

DGPS avoids the ambiguity problem entirely. Because the code wavelength is long, the receiver does not need to determine how many whole cycles have elapsed. It measures a pseudorange directly, and a reference station at a known location computes a pseudorange correction that is transmitted to the rover. The rover applies the correction and obtains a position. The fundamental limitation is the noise and multipath susceptibility of the code observable itself, which bounds DGPS accuracy to the metre to sub-metre level in typical marine conditions.

RTK resolves the ambiguity. Once the integer number of cycles is fixed, the carrier phase becomes a precise range observable, and position accuracy improves by two to three orders of magnitude. The trade-off is that ambiguity resolution requires favourable satellite geometry, low multipath, a short baseline to the reference station, and reliable correction communication. When ambiguity is fixed, RTK delivers centimetre-level positions. When it is not fixed, the solution may degrade to a float ambiguity state or fail entirely.

Accuracy Envelopes in Practice

Reported accuracies vary by equipment, baseline length, atmosphere, and multipath environment, but the orders of magnitude are consistent.

• DGPS horizontal accuracy: typically metre to sub-metre, with values around two metres or better at 95 percent confidence in many marine configurations. Some systems achieve better than one metre under favourable conditions.

• RTK horizontal accuracy: typically two to seven centimetres at 95 percent confidence, with the best results over short baselines and clean satellite visibility.

• Vertical accuracy: generally weaker than horizontal for both techniques, but RTK vertical accuracy remains substantially better than DGPS vertical accuracy. This matters for any survey that uses ellipsoidal height to derive water level or seabed elevation.

These figures are not interchangeable. A DGPS solution that achieves one metre in open water may degrade near steel structures or in areas of poor satellite geometry. An RTK solution that achieves two centimetres on a short baseline may lose ambiguity fix at longer ranges or under radio interference. The practical engineering question is not which system is more accurate in isolation, but which accuracy is compatible with the survey objective and the physical scale of the feature being measured.

The Beam Footprint Argument: Why DGPS Is Often Sufficient for Single Beam

A single-beam echosounder transmits one acoustic pulse and measures the two-way travel time along the vertical or near-vertical path. The acoustic beam is not infinitely narrow. A typical shallow-water single-beam transducer may have a beam opening angle of several degrees, and the ensonified footprint on the seabed grows with depth. At a beam angle of six degrees and a depth of twenty metres, the footprint diameter is approximately two metres. At thirty metres, it is approximately three metres.

The footprint defines the spatial scale of the depth measurement. A single-beam sounding does not represent a mathematical point; it represents an average or dominant return from an area of the seabed whose size is comparable to the footprint. Positioning the vessel to a precision substantially finer than the footprint does not necessarily improve the accuracy with which the depth is attributed to a location, because the depth itself is already an area measurement.

In hydrographic survey, the achievable horizontal resolution of a sounding is limited by the acoustic footprint of the echosounder, not solely by the positioning system. Positioning precision finer than the footprint does not produce a corresponding improvement in feature definition.

This is the central physical reason why DGPS remains a practical and often appropriate choice for single-beam survey. If the objective is Order 1 hydrographic survey to IHO S-44, with a wide-beam single-beam echosounder, metre-level positioning is compatible with the resolution of the depth measurement. RTK would position the vessel more precisely than the echosounder can resolve the seabed. The additional precision is not wasted in absolute terms, but it does not translate into a proportional improvement in the survey product.

The same argument does not apply to multibeam. A multibeam echosounder forms many narrow beams across a swath. Modern shallow-water systems produce nadir footprints of decimetres at moderate depth, and outer-beam footprints grow with angle. When the beam footprint is comparable to or smaller than the positioning uncertainty, horizontal positioning error becomes a dominant error source. A DGPS horizontal error of one to two metres, combined with a decimetre-level beam footprint, degrades the bathymetric surface in a way that cannot be recovered by processing. This is why high-resolution multibeam survey is generally specified with RTK or equivalent centimetric positioning, while single-beam survey can often be conducted with DGPS to the applicable order.

Survey Types and Positioning Fit

Single-Beam Bathymetry and Reconnaissance

Single-beam echosounding with DGPS is well suited to reconnaissance survey, order-level hydrography where the beam footprint is broad, channel condition checks, and volume reconnaissance where the required accuracy is consistent with metre-level positioning. The operational advantages of DGPS in these roles are significant: correction services are widely available, base station infrastructure is not required on site, communication links are simpler, and the system tolerates longer baselines and less favourable conditions without losing the solution entirely.

Multibeam Bathymetry and High-Resolution Feature Detection

Multibeam survey with narrow beams, object detection requirements, or Special Order specifications generally requires RTK or a comparable centimetric positioning service. The reason is geometric: the beam footprint can be smaller than the positioning error, so position error propagates directly into the bathymetric surface. Pipeline route clearance, berth pocket inspection, scour monitoring, and pre-engineering seabed mapping fall into this category.

Dredging and Construction Control

Dredging control and marine construction guidance require vertical and horizontal positioning consistent with the tolerance of the works. RTK is commonly used where direct ellipsoidal height is applied to compute dredge depth without independent tide reduction, and where cutter or draghead position must be resolved at decimetre level. DGPS may be adequate for progress monitoring or reconnaissance, but not for final payment survey or tight vertical control.

Vertical Referencing and the RTK Advantage

DGPS vertical accuracy is generally unsuitable as a standalone vertical reference for hydrographic survey. A DGPS vertical error of two to four metres cannot be used to reduce soundings to a chart datum without independent water level observation. DGPS-based survey therefore depends on a tide gauge, a water level model, or a separately established vertical reference.

RTK changes this dependency. With centimetric vertical accuracy, RTK can provide an instantaneous ellipsoidal height at the transducer, which, after application of a geoid model and reduction to the project vertical datum, supports tide-independent bathymetric reduction. This is particularly valuable in areas without a nearby tide gauge, in dynamic shallow water where tide is spatially variable, and in construction control where direct elevation is required. The physics of carrier phase, in other words, does not only improve horizontal precision; it changes the vertical referencing workflow.

Operational Trade-Offs Beyond Accuracy

• Infrastructure: DGPS corrections are available from radio beacon networks, satellite subscription services, and regional networks with broad coverage. RTK typically requires a local base station or a network RTK service, with a communication link that must remain reliable throughout the survey.

• Baseline and range: DGPS accuracy degrades gradually with distance from the reference station because atmospheric and orbit errors become less correlated. RTK ambiguity resolution also degrades with baseline, but over shorter ranges, and can fail abruptly if the baseline is too long or the communication link is lost.

• Robustness: DGPS tends to degrade gracefully. Accuracy worsens, but the solution remains available. RTK can transition from a fixed ambiguity solution to a float solution or lose the correction link entirely, producing a step change in accuracy that must be detected and managed.

• Cost and mobilisation: DGPS receivers and correction services are generally less expensive and simpler to mobilise. RTK requires additional base station equipment, calibration, and communication infrastructure, which may not be justified for a survey whose resolution is limited by a broad single-beam footprint.

Optimization Strategies and Common Pitfalls

• Good practice: Match positioning technique to beam footprint and survey order, not to a default preference for the highest available accuracy.

• Good practice: For single-beam Order 1 survey, use DGPS with documented correction source, baseline length, and accuracy statement, and verify against independent control.

• Good practice: For multibeam, object detection, or Special Order survey, use RTK or an equivalent centimetric service and monitor ambiguity fix status continuously.

• Good practice: Apply water level reduction from a tide gauge or validated model when using DGPS, and document the vertical referencing method.

• Good practice: When using RTK for vertical referencing, verify the geoid model and tidal datum transformation independently.

• Poor practice: Using DGPS for high-resolution multibeam survey where the beam footprint is smaller than the positioning uncertainty.

• Poor practice: Assuming RTK is automatically better for every survey without considering footprint, survey order, logistics, and the graceful degradation characteristics of DGPS.

• Poor practice: Using DGPS vertical height as a tide-independent reference without independent water level verification.

• Poor practice: Operating RTK at long baselines or in radio-interference environments without monitoring fix quality and without a fallback plan.

• Poor practice: Reporting accuracy from manufacturer specifications rather than from field verification on the day of survey.

Practical Implications for Geomak Projects

Geomak selects positioning technique according to the physical scale of the measurement and the contractual accuracy requirement. Single-beam reconnaissance and order-level hydrography in Gulf coastal waters are frequently served well by DGPS, where correction coverage is reliable and the beam footprint is broad relative to positioning uncertainty. Multibeam survey, dredge control, pipeline clearance, and berth inspection are specified with RTK or an equivalent centimetric service, supported by base station calibration, ambiguity monitoring, and independent check points.

The engineering principle is straightforward. Positioning precision should be compatible with the spatial resolution of the sensor and the tolerance of the decision the data must support. DGPS is not obsolete; it is appropriately matched to single-beam survey physics. RTK is not universally required; it is essential where the beam is narrow, the tolerance is tight, or the vertical reference must be derived directly from GNSS.