What insurers should know about satellite-based subsidence monitoring

Subsidence-related losses share an awkward property: the movement that causes them usually develops slowly, over years, and is invisible during a single site visit. By the time cracking is reported, the informative part of the story — when the movement started, how fast it progressed, whether it is accelerating — has already happened, unrecorded.

Insurers feel this gap at three points. Underwriters price risk on location and construction attributes without knowing whether the specific site has been moving. Claims teams assess damage with little independent evidence about the site's movement history. Portfolio managers hold aggregations of exposure they cannot economically inspect.

Radar satellites such as Sentinel-1 image most land areas every 6–12 days, day and night, largely regardless of cloud. Interferometric processing (InSAR) compares the radar phase between acquisitions and converts it into an estimate of surface movement along the satellite's line of sight. Persistent Scatterer InSAR extends this across hundreds of images, producing a movement time series for every stable reflective point — typically buildings, hard standing, and structures.

The result, in built environments, is a dense grid of measurement points, each with a multi-year history. Under favourable conditions, velocity estimates reach millimetres per year. Crucially, the archive already exists: for most locations, movement history since 2014 can be reconstructed retrospectively.

The realistic applications are screening and evidence, not automated decisioning. In underwriting, historical movement context can inform referral decisions and inspection triggers for higher-value or higher-uncertainty risks. In portfolio management, consistent screening across thousands of locations produces a ranked list of assets that may deserve closer attention. In claims, an independent record of when and how a site moved can support — not decide — the investigation.

Three limits deserve particular attention. First, measurement is of surface movement, not structural condition: a moving point is not a damaged building, and a stable reading is not a guarantee. Second, coverage is uneven — vegetated ground, small structures, and rapidly changing sites may yield few reliable points. Third, causation is not observable from space: satellite data can show patterns consistent with certain mechanisms but cannot conclusively attribute movement to clay shrinkage, mining, groundwater, or construction without supporting information.

Used with those limits stated plainly, satellite subsidence data is what insurance workflows chronically lack: consistent, historical, asset-level movement evidence at portfolio scale.