What the data is,
and how to read it
Every number in Telluric comes from public Sentinel-1 satellite radar, processed with open scientific tooling. This page explains the method and every metric we show — in the same plain language the product uses.
How the measurement works
- 01
Radar satellites measure distance
The Sentinel-1 satellites pass over every site on Earth every 6–12 days and bounce radar pulses off the ground. The returning signal carries a distance measurement precise to millimeters. This archive reaches back to 2015 and is public.
- 02
We compare visits
One image alone says nothing about movement. Comparing the radar signal between two visits reveals how much each patch of ground moved in between — a technique called interferometry (InSAR). We build a network of such comparisons across many visits.
- 03
A time series per point
Solving all comparisons together produces a movement history for every stable radar reflector — rooftops, rock, bare ground — inside your site. Surfaces that change (water, vegetation, active construction) can't be measured reliably and are filtered out rather than shown as fake certainty.
- 04
Screening-grade, honestly labeled
Every number ships with its uncertainty, its noise level, and its data quality — visible in the app, not buried in a footnote. Telluric shows you where to inspect and hands you a dated record. It does not replace geotechnical instrumentation or engineering judgement, and never claims to.
Every metric, explained
These are the exact explanations built into the product — tap the ⓘ next to any number in the app and you get the same text.
For every measurement cell
- Velocity (LOS)
- How fast this spot is moving, in millimeters per year, measured along the satellite's line of sight — the direction the satellite looks from, which is at an angle, not straight down. Negative means moving away from the satellite; for flat ground that usually means sinking.
- Trend ±1 SE
- A straight line fitted through this point's measurements, with its uncertainty. Read it as "the speed is probably within ± this much of the number". If the ± is nearly as big as the number itself, the record is too short or too noisy to trust the speed yet.
- Total displacement
- The net movement between the first and the last measurement in the record, in millimeters. The dates below show exactly which period it covers.
- Observations
- How many satellite passes went into this point's history, and how many days they span. More observations over a longer period = more reliable numbers. A short record can't tell slow real movement apart from noise.
- Noise floor (RMS)
- The typical scatter of this point's measurements around its trend, in millimeters. Think of it as the measurement's blur: movements smaller than this can't be distinguished from noise.
- Rate change
- Whether the movement is speeding up or slowing down: the speed in the second half of the record minus the speed in the first half. Accelerating settlement is usually more concerning than steady settlement.
- Coherence
- How consistently this spot reflected the radar across the whole record, from 0 to 1. High values (rooftops, rock) mean trustworthy measurements; low values (vegetation, construction) mean the readings are shakier. Points below the threshold are hidden by default.
- vs site mean
- How this point moves compared to the average of all points on the site. Even when a whole area sinks together, the dangerous thing is usually one part moving differently from the rest — that's what this highlights. The coordinates below are the point's exact location.
For every satellite pass
- Displacement
- How much this exact spot has moved since the first measurement, in millimeters. Negative numbers mean it moved away from the satellite — for flat ground, that usually means sinking. Every value is relative to a stable reference point nearby, not an absolute height.
- vs trend
- How far this single day's reading sits from the point's overall trend line. Small values mean the day fits the pattern; a big value means something unusual — often just weather interfering with the radar signal, not real movement.
- Coherence
- How clearly the satellite could "recognize" this spot on this day, on a scale from 0 to 1. Hard surfaces like rooftops score high; vegetation, water, and construction sites score low. Higher means the measurement is more trustworthy.
- Perpendicular baseline
- The satellite never flies exactly the same line twice — this is how far off its usual path it was on this pass, in meters. Small offsets are normal. Very large ones make the comparison between passes slightly less precise.
- Track
- The number of the satellite's fixed orbital path this measurement comes from. The satellite revisits the same numbered path every few days, and all measurements for a site use the same one so they stay comparable.
- Site RMS
- How scattered the measurements were across the entire site on this particular day, in millimeters. When this is high, the whole site looks noisy at once — the usual culprit is atmosphere (water vapor delays the radar), so readings from that day deserve less weight.
What we will not tell you
Satellite radar measures along the satellite's viewing direction, works best on hard surfaces, and needs a long record before slow movement separates cleanly from noise. Where the data is weak — vegetated ground, short archives, low coherence — the product says so instead of smoothing it over. Sites that can't be measured well are rejected at the free feasibility stage, before any charge.
Telluric is a screening tool: it shows you where to inspect and gives you a dated record you can cite. It does not replace geotechnical instrumentation or engineering judgement.