Marine GPS and GNSS Explained
What a GNSS receiver actually knows, why display precision is not accuracy, how course over ground differs from heading, and how to build position redundancy that survives a network failure.
GPS is so familiar that it is easy to underestimate it.
A modern chartplotter places a vessel symbol on a chart, calculates speed and course, feeds an autopilot, timestamps AIS messages and records a track — continuously, silently, without ever being asked. Because all of that happens quietly and never appears to fail, most boat owners treat position as a solved problem.
It is not solved. It is usually fine, which is a different thing, and the gap between those two states is where navigation accidents live.
A professional approach separates four questions:
- How accurately does the receiver know its own position?
- How accurately is the charted object positioned?
- Which position source is the network actually using?
- What happens if that source fails?
Most recreational navigation answers the first question well and the other three not at all.
#GPS is one GNSS constellation
GPS is the United States satellite navigation system. GNSS is the broader term.
A modern marine receiver may track:
- GPS — United States
- Galileo — European Union
- GLONASS — Russia
- BeiDou — China
- QZSS — Japan, regional, useful in the western Pacific
The practical benefit is not the number of logos on the specification sheet. It is improved satellite availability and geometry — more satellites visible means a better chance of a well-distributed set, which produces a stronger solution. That matters in a marina surrounded by masts, under a cliff, or with a bimini partly obstructing the sky.
#How position is calculated
Satellites broadcast precisely timed signals. The receiver estimates its range to each satellite from the signal travel time, and solves for the one location consistent with all those ranges at once.
From repeated positions it derives:
- speed over ground
- course over ground
Both are differences between successive fixes. That has a consequence people rarely think through: they are computed quantities, not measured ones, and they inherit the noise of the positions they came from. Near zero speed that noise dominates — which is the anchoring problem below.
These are not the same as heading or speed through water.
#Accuracy is not display precision
A screen can show coordinates to three decimal places of a minute. That does not mean the final digit is accurate.
Error sources include:
- satellite geometry — how well spread the satellites are
- atmospheric effects, particularly ionospheric delay
- multipath — signals arriving after reflecting off the water, a mast or a steel superstructure
- antenna obstruction
- radio interference, deliberate or incidental
- receiver design and quality
A good multi-constellation receiver with a clear sky view typically knows its position to within a few metres. That is excellent, and on many cruising vessels it is not the limiting factor.
Chart uncertainty is usually the larger practical hazard. If the receiver knows you to 3 m and the reef was surveyed to ±200 m, the precision that matters is not yours. See marine charts Australia.
#Satellite geometry
Satellites distributed widely across the sky provide a stronger geometric solution than satellites clustered in one direction.
Receivers may express this through HDOP or PDOP — dilution of precision figures, where lower is better.
Most recreational skippers do not need to monitor these continuously. What is worth understanding is that a fix has varying quality, and that the display does not usually tell you when quality has dropped. A receiver reporting a position under a steel bridge, beside a cliff, or with half the sky obstructed is reporting a worse position than the same receiver in open water, and the symbol looks identical.
#Antenna location
Internal MFD antennas work well with a clear view of the sky.
Problems appear when the unit sits:
- below deck
- under a metal hardtop or bimini with a reflective coating
- inside aluminium or steel structure
- beside interference sources — inverters, chargers, satellite terminals, some LED lighting
An external GNSS antenna can provide a single consistent position source to the whole network, which is preferable to several mediocre internal receivers competing. Mount it with a clear sky view, clear of the radar beam and clear of transmitting antennas. See sensor siting.
#Update rate
1 Hz means one position each second. 10 Hz means ten.
Higher rates matter most for:
- fast boats, where a second is 12 m at 25 knots
- racing, where start-line and layline calculations are sensitive
- autopilot control, which benefits from smooth input
- radar overlay, where lag shows as misalignment during turns
A 6-knot cruiser does not become dramatically safer because position refresh increases from 10 Hz to 20 Hz. Spend that money on heading instead.
#Position is not heading
A single GNSS receiver can calculate course over ground while moving.
COG is the direction the boat travels over the earth. Heading is where the bow points.
Heading: 040°
Course over ground: 057°Current or leeway is moving the vessel sideways. Seventeen degrees is a lot of water going past, and the boat is not misbehaving — it is being set.
This difference matters for:
- radar overlay, which must be oriented by heading
- MARPA, whose solutions depend on your own heading
- the autopilot, which steers a heading
- true wind, computed from apparent wind and heading
- laylines, derived from true wind
Five systems, one input. See heading sensors.
#At anchor
When the vessel barely moves, GNSS-derived COG becomes unstable, because tiny position changes create a wildly varying direction calculation. The receiver is not faulty; a direction derived from two nearly identical positions is meaningless.
A real heading sensor or a multi-antenna GNSS compass provides orientation at zero speed. That matters for anchor-watch interpretation, for radar orientation in a crowded anchorage, and for knowing which way the boat is actually lying. See anchoring electronics.
#Satellite compasses
A multi-antenna GNSS compass measures relative satellite carrier phase across a known baseline between two or more antennas, and derives orientation from the difference.
Advantages:
- stable zero-speed heading
- no local magnetic deviation — a steel hull stops being a problem
- high update rates
- pitch and roll output on advanced systems, enabling radar and sonar stabilisation
Useful on:
- steel vessels
- aluminium yachts
- large catamarans, where the sensor sits far from the centre of motion
- premium radar installations
- boats with difficult magnetic environments
They are not mandatory for every coastal cruiser. For a conventional fibreglass boat, a well-sited solid-state magnetic sensor is enough, and the money is better spent elsewhere.
The one installation trap: the antenna baseline must be aligned to the vessel's centreline within tolerance and the offset entered accurately, or you have bought a constant heading bias.
#Position-source management
A modern NMEA 2000 boat can contain:
- internal MFD GNSS
- an external antenna
- AIS GNSS
- autopilot GNSS
- a Furuno Marine Electronics 2026: Radar and Sonar Depth">satellite compass
Five position sources is normal, and the network chooses one.
Check which source is actually selected after a refit, a sensor change, or a firmware update — updates have been known to reset the selection. Then confirm the VHF is receiving position, because a DSC distress alert without one is materially worse.
The best sensor aboard is useless if the network quietly uses a weaker source. This is a standing item in commissioning and the pre-departure checklist for exactly that reason.
#Datum
Modern marine navigation normally uses WGS 84 or a closely compatible reference frame.
Old charts and old coordinates can use different datums, and the offset between datums can be hundreds of metres — enough to put a waypoint on the wrong side of a reef.
When entering coordinates from old cruising notes, a guidebook, or a historical source, verify the datum. A latitude and longitude without a datum is an incomplete piece of information.
#Failure modes
Possible GNSS failures:
- antenna fault or water ingress
- damaged or chafed cable
- power loss to the receiver
- NMEA network fault, which removes the data rather than the fix
- interference or jamming, which is increasingly reported in some regions
- receiver fault
- source-selection error, where the network switched to something worse
Note that several of those present as plausible but wrong rather than as an obvious absence. A receiver using a poor source, or affected by interference, may show a position that looks entirely normal.
For offshore navigation, assume the primary GNSS can fail.
#Practical redundancy
A sensible hierarchy:
Primary — installed network GNSS.
Secondary — a physically separate GNSS or MFD, ideally on a different circuit.
Backup — a tablet or phone with internal GNSS and fully downloaded offline charts.
Non-electronic orientation — a magnetic compass.
A second screen on the same failed network is not an independent backup. That is the most common mistake in the whole scheme, and it is why the redundancy guide argues in failure domains rather than device counts.
#The most important warning
GNSS accuracy does not make a chart accurate.
If the receiver knows the boat position to within a few metres but the reef was surveyed with far lower confidence, the screen creates false certainty — two objects drawn with identical crispness, one known to metres and one to hundreds of them.
Read marine charts Australia, and learn to display Zone of Confidence on your own system.
#Passage checklist
Before departure:
- confirm the primary GNSS source the network has selected
- compare position against a known harbour feature
- confirm the VHF is receiving position
- check backup chart coverage is downloaded offline
- verify tablet and handheld battery
- verify the magnetic compass
- check old imported waypoints and their datums where relevant
#Bottom line
GNSS tells you where the receiver is. The chart tells you where mapped objects are believed to be. Radar, depth, observation and judgement determine whether those two worlds agree — and knowing which source the network is using, and what happens when it stops, is the part that separates a navigator from a passenger.
Common questions
Short answers to the questions this guide raises most often.
How accurate is marine GPS?
A modern multi-constellation receiver with a clear sky view typically knows its own position to within a few metres. That is usually far better than the positional confidence of the charted hazard beside it, which is why Zone of Confidence matters more than receiver specification for most cruising decisions.
Why does my course over ground differ from my heading?
Because current and leeway push the boat sideways. The bow points one way while the vessel tracks another. The difference matters for radar overlay, MARPA, true wind and layline calculations, which is why a real heading sensor cannot be replaced by GNSS-derived course over ground.
Do I need a satellite compass?
Not for a conventional fibreglass coastal cruiser, where a well-sited solid-state magnetic sensor is usually enough. A multi-antenna GNSS compass earns its cost on steel and aluminium vessels, large catamarans, premium radar installations and any boat with a difficult magnetic environment.
Does a higher GPS update rate make navigation safer?
Rarely on a slow boat. Higher rates matter for fast powerboats, racing, autopilot control and radar overlay. A six-knot cruiser gains very little going from 10 Hz to 20 Hz.
Continue building the system
These guides pick up where this one stops.
Foundations
Australian marine charts, AusENC, ZOC and S-100
Why a beautifully rendered electronic chart can still be built on an old survey — and how to read Zone of Confidence, manage chart datum and plan around survey quality in Australian waters.
Read the guideFoundations
Heading sensors and satellite compasses
Heading is infrastructure. A single bad sensor location degrades the autopilot, radar overlay, MARPA, true wind and laylines at once — here is how to get it right.
Read the guideSeamanship & Safety
Navigation redundancy that actually works
Redundancy is not buying two of everything. It is preventing one failure from removing several critical capabilities at once — which means the best backup usually looks nothing like the primary.
Read the guideEquipment & Sensors
NMEA 2000, NMEA 0183 and marine Ethernet
A single bad connector can make a premium display, autopilot or sensor look defective. Treat the network with the same seriousness as the vessel's DC electrical system.
Read the guideFoundations
Marine navigation systems: the complete overview
The system-level guide: eight questions every navigation system has to answer, why sensors matter more than screens, and how to design a boat where technologies cover one another's blind spots.
Read the guideSeamanship & Safety
Passage planning, end to end
Route autogeneration is a planning assistant, not a plan. This is the four-stage method that turns a proposed track into a passage you can actually monitor and abort.
Read the guideReferenced by
Other guides that depend on the ideas on this page.
This guide is independent editorial information, not a substitute for official regulation. Equipment requirements, licensing and chart currency differ by jurisdiction and change over time — confirm the current position with the AMSA, the Australian Hydrographic Office, ACMA and your state or territory maritime authority before relying on it.