Developing a geophysical source model for 3D surface displacements above storage cavern fields with InSAR time series at the example of Epe (NRW, Germany)

Monitoring surface displacements that occur above storage caverns in salt bodies is important to estimate risks for infrastructure in the immediate vicinity. As the pressure inside a storage cavern is usually lower than in the surrounding rock, the cavern converges continuously. This volume loss leads to a prominent subsidence bowl at the surface. Often this subsidence is approximated by linear velocities, However, in particular for gas storage caverns, fields of multiple caverns in close proximity can cause displacement fields which are spatio-temporally complex. The amount of convergence and subsequent surface subsidence depends on the pressure inside the caverns, which for gas caverns changes with the cavern filling levels.  This causes seasonal displacement patterns, but also different total subsidence in subsequent years. Moreover, it can result in different superposing displacement patterns from neighboring caverns. Monitoring such a displacement field therefore requires geodetic measurements with dense spatial coverage and high temporal resolution.

Multi-temporal SAR interferometry (MT-InSAR) analysis can fulfill these requirements in optimal conditions, with the Sentinel-1 mission providing free C-band SAR-data with a revisit time of 6 to 12 days. However, as MT-InSAR depends on temporally stable backscattering characteristics of the ground targets, there are often many areas in practice without sufficient data available. Also, current SAR-satellite missions have low sensitivity to north-south directed displacements, which complicates the analysis of the 3D-displacement field with InSAR alone. A geophysical source model derived from surface displacements, that describes the relation between cavern filling levels and surface response, can help with both of these issues.

We derive such a geophysical source model for the storage cavern field Epe in NRW Germany, which consists of 114 caverns, more than 50 of them storing natural gas, from time series of eight years and four tracks of InSAR Sentinel-1 data. We use Persistent and Distributed Scatterers to process the data and validate our results with data of three permanent GNSS stations and annual leveling measurements. As parts of the cavern field in Epe experience other strong displacement effects such as the surface response to groundwater level changes that superpose the cavern signals, we use Independent Component Analysis to separate displacements from different sources. We combine a Kelvin-Voigt body with the Norton power law to relate the pressure differences in each cavern to volume change in the viscoelastic salt body. Then, we use a Sroka-Schober-model to translate this volume change through elastic layers to the surface. There, the effects of all caverns are superposed. We use the cavern related InSAR displacement data to optimize for local parameters and to obtain spatio-temporally high-resolution model predictions for 3D surface displacements.

Not only does this model provide displacement estimates for areas with no measurement data, with a causal relation to the cavern usage, but it also can give more insights to the dynamic convergence of the caverns, as cavern volumes are usually only measured every few years.