Nantua: reading slow subsidence with InSAR

The site was not chosen by chance. The area around Xavier Bichat high school is a remarkable geotechnical case: it lies near the lake outlet, in a glacial valley partly filled with several tens of metres of soft sediments.

Significant settlement has been documented there . The former high school reportedly experienced centimetric to decimetric settlement, locally very high, without this automatically implying failure or immediate danger. This is precisely what makes the case interesting: ground movement can be large, slow, differential, and still compatible with specific design choices.

The new high school was designed with this constraint in mind, using a lighter structure, box rafts and separation joints. The issue is therefore not only to detect motion, but to understand how it interacts with building design, courtyards, access areas and everyday use.

An InSAR dataset is more than a value displayed on a map. It is a time series: a sequence of radar observations acquired under a specific viewing geometry, affected by noise, uncertainty and interpretation limits.

Transforming these time series into a map therefore requires a series of choices. Such choices are unavoidable, but they should remain visible, transparent and open to discussion.

InSAR time series can reveal long-term trends, seasonal cycles or progressive changes. Yet to display this information on a map, the underlying observations are often reduced to a single indicator: mean velocity, median velocity, seasonal amplitude, coherence or variability.

This simplification improves readability, but it also compresses a complex temporal signal. Depending on the chosen metric, the same area may appear stable, active, noisy or highly contrasted.

InSAR measurements are often numerous and spatially discrete, particularly in urban environments. To make them easier to interpret at neighbourhood or municipal scale, they can be grouped into spatial cells.

This aggregation is not neutral. Cells that are too large may smooth local contrasts. Cells that are too small may create an illusion of precision. Excessive filtering can hide meaningful signals, while insufficient filtering can amplify noise.

The colours displayed here represent median radar line-of-sight velocities. They describe relative motion within the satellite observation geometry.

They should not be interpreted directly as vertical subsidence. Proper interpretation requires consideration of orbital geometry, viewing angle, radar coherence, point density and local context.

This map is not a field investigation, nor is it direct evidence of structural damage. It cannot replace site inspections, geotechnical studies or structural assessments.