Practical guides written by a senior geotechnical engineer for engineers who receive deformation maps, velocity rasters or time-series plots and need to know exactly what the numbers mean before they land in a report. Every article states the geometry, the sign convention and the source of the data, and links to the site tools that turn an LOS velocity into a vertical settlement, validate it against GNSS and classify the hazard. The editorial rule is simple: if a number cannot be reproduced from a formula and a cited source, it does not get published.
The three current articles cover the core of the problem. The first explains why an InSAR velocity measured along the satellite line of sight is not a vertical settlement: the incidence geometry projects a 10 mm vertical displacement into roughly 8.2 mm along the LOS at 35°, and dividing by the cosine only recovers the truth when the motion is genuinely vertical — otherwise the correction manufactures an artifact with units of mm/yr. The second shows the verification checklist used in practice before a deformation map enters a report: colocation of the InSAR pixel with the GNSS antenna, alignment of the reference epoch, separation of the geometric components and an honest residual. The third teaches how to read the three shapes of an InSAR time series — steady, accelerating, stepping — and what each implies for hazard classification.
From satellite to report: the workflow behind the articles
Fig. 1 — The value chain of an InSAR monitoring project. The articles below explain each step and link to the site tools that turn a satellite measurement into a defensible engineering number.
Reading order
Step
Article
What it settles
1
Your InSAR velocity is not subsidence
LOS vs vertical: the geometry and when the cos θ correction is valid
Steady vs accelerating vs stepping: what each implies for hazard
The recommended reading order follows the workflow of a real project: first understand what the satellite measures and how to project it into vertical movement, then validate the result against ground truth, and finally interpret the time series. If you are short on time, start with the first article — it is the one that prevents the most common and most expensive misinterpretation: reading an LOS velocity directly as subsidence.
How to use this blog with the site tools
The blog is deliberately paired with the calculators on this site, so the theory in each article can be exercised on real numbers in minutes. After reading the LOS vs vertical article, open the LOS to vertical converter, paste the velocity of one of your monitoring points together with the local incidence angle, and reproduce the correction by hand before trusting any velocity map. The same point can then be checked against the GNSS validation tool, which implements the comparison checklist described in the second article: reference epoch alignment, spatial colocation and the residual between the satellite and the ground antenna.
When a time series shows one of the three shapes discussed in the third article, the time series reader helps you identify the break points — the moment a steady trend turns into acceleration, or the epoch of a step jump — and export the segments with their dates for the report. The subsidence hazard classifier then maps a vertical velocity and its gradient into the risk classes used in the region, so the interpretation step is explicit and reproducible instead of a subjective color call.
Two supporting tools close the loop for field integration. The coordinate converter enforces the datum and zone discipline described in the coordinate errors article, preventing the tens-of-metres shifts that appear when field GNSS points and satellite products are combined in different frames. Finally, the InSAR report generator assembles the validated values, the geometry parameters and the caveats into a draft technical appendix, so the engineer spends the time on interpretation rather than on formatting tables. This is the workflow the blog teaches: every number in an article can be traced to a tool on this site and to a cited source.
Who writes these guides
The articles are written by a senior geotechnical engineer with two decades of practice in civil engineering, geotechnical investigation and deformation monitoring — the same hands-on context in which InSAR products arrive on an engineer's desk and must be turned into a defensible decision. The perspective is deliberately that of the user of satellite data, not the radar scientist: the focus is on what a velocity value means for a slope, a road, a dam or a foundation, how to check it against field instruments, and how to present it in a way a reviewer can reproduce. That is why each guide states the geometry, the sign convention, the source of every number, and the link to the tool that performs the calculation. The articles are technical reference material; they describe methods and worked examples, and they never replace project-specific design performed by a licensed professional.
The core geometry lesson of the site, with a worked numerical example and a comparison table. Sentinel-1 looks sideways, so the incidence angle — roughly 30° near the inner swath to about 46° at the far edge — tilts the line of sight away from vertical. A purely vertical displacement of 10 mm therefore appears in the LOS product as 10·cos θ: about 8.7 mm at 30°, 8.2 mm at 35° and 6.9 mm at 46°. Inverting the relationship by dividing the LOS velocity by cos θ recovers the vertical component, and that is the entire mathematics of the LOS to vertical converter. The catch is that the division is only correct when the motion really is vertical: on slopes, embankments and creep-prone ground, the LOS vector also carries horizontal components, and dividing produces an artifact with units of mm/yr that is not the vertical component of anything. The article walks through the sign conventions (approach vs retreat, ascending vs descending), the systematic 20-30% understatement for thresholds and classification tools, and closes with the cases where the correction actually changes an engineering decision.
A deformation map without ground truth is a color ramp, and this article is the checklist used to turn it into an engineering product. The practical verification steps are covered in detail: colocating the InSAR persistent scatterer and the GNSS antenna within the same resolution cell (a 20-30 m pixel on Sentinel-1 descending stacks), aligning the reference epoch so both series start from the same date, separating the geometric components (the satellite is far more sensitive to vertical movement, while the GNSS antenna measures all three axes), and matching the temporal sampling when the satellite revisits every 12 days. The article also covers the honest handling of residuals — a root-mean-square difference of 2-4 mm is a healthy result; a correlation coefficient without the residual is a red flag — and how to report the validation in a format a reviewer can reproduce.
Not all mm/yr are equal, and the shape of the time series carries more information than the average velocity. A constant rate describes regional settlement or steady consolidation and allows a simple annualized projection; an accelerating trend is the signature of creep toward failure and should trigger an escalation of the monitoring frequency and a geotechnical assessment; a step-like jump between two acquisitions points to a sudden event — an earthquake, a rapid drawdown, an excavation step — that must be confirmed against the acquisition dates and the local records. The article explains how to separate the three shapes from noise, which quality flags to check first (coherence at the last epoch, atmospheric residual, orbital ramp), and what each shape implies for hazard classification, thresholds and the frequency of field inspection.