Reproducibility of geodetic source modelling and implications for the unrest at Campi Flegrei caldera (Italy)

Abstract

Surface deformation derived from satellite interferometry and GNSS observations represents a fundamental tool for investigating magma dynamics beneath active volcanoes. However, geodetic inversion results may depend on the quality, quantity, and accuracy of the datasets employed. Therefore, the assessment of model robustness is essential for deriving reliable interpretations of subsurface processes. Campi Flegrei caldera (Italy) has been experiencing progressive uplift since 2005, associated with increased seismicity and gas emissions. Modelling of the 2007–2023 deformation from COSMO-SkyMed and GNSS data suggested the presence of a shallow inflating source at 3.9 km depth since 2015 and a deeper deflating reservoir. This interpretation was obtained using a complex 3D heterogeneous numerical modelling framework. In this work, we evaluate the model robustness by repeating the geodetic inversion using multi-mission satellite datasets, including Sentinel-1 observations and a different COSMO-SkyMed analysis, together with GNSS measurements. We show that, although discrepancies affect part of the COSMO-SkyMed descending dataset, their impact on the inversion remains limited because the data are analysed over discrete time windows, the discrepancies remain moderate relative to the overall deformation, the uncertainties are accounted for, and the inversion integrates multiple geodetic datasets. The results show a stable geometry and depth of the shallow source, together with comparable estimates of volume change. These findings demonstrate that the source parameters are reproducible across alternative InSAR datasets, providing an explicit cross-validation. More broadly, they show that first-order constraints on magma-driven deformation at Campi Flegrei are robust and can support interpretations of volcanic unrest and hazard assessment.