MonitoringNovember 10, 2026· 9 min read

Is That Slope Moving? Continuous RTK GNSS Monitoring, Explained

Highwalls, pit slopes, embankments and retaining structures move before they fail. A survey every few months tells you where a point was. Continuous RTK GNSS monitoring tells you whether it is moving now, and whether you can trust the answer.

Why Continuous Monitoring Matters

A periodic survey is a snapshot. If a slope starts to creep a week after the surveyor leaves, nobody knows until the next visit, and by then the question is no longer whether it moved but how far. Continuous monitoring turns position into a trend, and a trend can raise an alarm.

It matters wherever ground or a structure holds something back or stands over people and plant: the highwall of an open pit or quarry, a tailings or water embankment, a cut slope along a road or railway, a retaining wall.

What RTK Adds

A GNSS receiver on its own is good to a few metres. RTK brings it to millimetres, relative to a reference. The receiver sits on a fixed monument and receives corrections from a base station. With them it resolves the carrier phase of the satellite signals — the state the receiver reports as “RTK fixed”.

On the example monitor in this article, that gives a position precision of 3.2 mm east and 4.1 mm north (1σ), and 8.7 mm vertical. Vertical is always the weakest axis, because the satellites are only ever above the horizon, never below it.

The screenshots come from one of our own test monitors, shown under a made-up name: Coyote Flats Highwall – Monitor 3.

Is the Data Trustworthy?

Before you read movement, check the measurement. A screen that shows a position without showing how good it is invites the wrong conclusion. Powoflow puts the evidence next to the number:

Fix rate — the share of each hour that was RTK-fixed. Only fixed positions count towards movement.
Corrections — how old the last correction was, and every time the correction stream dropped and for how long.
Satellites and geometry — how many satellites were used, how well spread they were (PDOP) and how strong each signal was.
A quality gate — positions that fail it are left out, for example PDOP above 3, fewer than six satellites, or a horizontal RMS above 15 mm.

One check matters more than it looks: what the receiver claims against what it did. On the example monitor the receiver claimed 4.6 mm per position, and the positions actually scattered by 8.4 mm over the last 24 hours, 1.82 times the claim. That is what a sound installation looks like. The receiver's own accuracy figure is never the headline, because while RTK-fixed it sits on a fixed floor and does not move with the real precision.

RTK solution quality over 24 hours, correction outages and monument stability — How much of each hour was RTK-fixed, when corrections dropped, and how still the monument held.

Mount mode, the monitor on the map and the satellites in view with their signal strength — The satellites in view, which ones are used, and how strong each signal is.

Movement or Noise?

Every position scatters a little. Movement is a trend that outgrows that scatter. Powoflow takes one RTK-fixed position per day and measures it against a reference: the mean of the first five counted days. From those it fits a trend on each axis, with its uncertainty, and states the smallest movement the record could have detected. That floor comes down as the days add up.

The movement alarm waits for persistence: a displacement beyond three standard deviations on three consecutive nights. Days that scatter far beyond the norm are left out rather than allowed to trigger it.

Is this structure moving: east, north and up rates with what each would detect — The movement trend on each axis, and the smallest movement it would have detected.

Underneath sit the details an engineer will want: hourly displacement on each axis with every loss of fix and correction outage marked, a drift trail of where the antenna went, and a map of where it spent its time.

Seven days of east, north and up displacement with the receiver's fix state — A week of displacement on each axis, with every loss of fix and correction outage marked.

The antenna's drift trail and where it spent its time over 30 days — Where the antenna went over 30 days, and where it spent its time.

Know What GNSS Cannot See

Good monitoring says where it stops being useful. Powoflow rates the precision it achieved against the accuracy classes in USACE EM 1110-2-1009 (Table 2-1). On the example monitor, slope and crest stability on an earth or rockfill embankment (20–30 mm) is met, and long-term movement of concrete (5–10 mm) is within range. Short-term concrete deflections and settlement, down to fractions of a millimetre, are marked as not credible for GNSS: use precise levelling, a total station or in-place instruments there.

Positions also carry the reference frame they are in. They are for measuring relative movement over time, not for setting out absolute survey coordinates.

The quality-control gate, precision per window length and the achievable accuracy class — The quality gate, the precision each window length achieves, and what that is good enough for.

Powoflow Sense Neva

The monitor in this article is a Powoflow Sense Neva node: a sensor we design and build ourselves, with RTK GNSS and vibration detection in one node, powered over Ethernet from Powoflow Sense Mesa. Neva is made to order for projects that need it. See it on the hardware page.

ConnectivityStarlink updates itself. Do you know when?

HardwareLoRaWAN sensors, from the field to the cloud

Have ground that needs watching?

Tell us about your use case, and we will tell you honestly what RTK GNSS can and cannot show you there.

Book a walkthrough Movement monitoring