Subsurface ice interpretation using joint seismic-electrical responses and rock physics diagnostic

The detection of liquid water and ice, as well as the discrimination between frozen and unfrozen geological materials, remains a major challenge in periglacial geophysics. Seismic refraction and electrical resistivity tomography are widely applied to address this problem (Hauck et al., 2011; Mollaret et al., 2020); however, similar velocity and resistivity anomalies may originate from distinct physical configurations, including pore-filling ice, massive ice bodies, or heterogeneous mixtures of frozen and unfrozen materials. This intrinsic non-uniqueness complicates the interpretation of field geophysical data and limits the ability to infer the elastic contribution of ice to the seismic response. In this study, we investigate how the assumed physical behaviour and spatial distribution of ice influence the joint interpretation of seismic velocity and electrical resistivity responses. Field geophysical datasets are analysed using a phase-based petrophysical joint inversion framework that estimates volumetric fractions of liquid water, ice, air, and solid matrix by coupling seismic refraction velocities and electrical resistivity following four-phases model scheme (Wagner et al., 2019). The seismic forward response is governed by effective-medium rock-physics formulations that explicitly account for the elastic contribution of ice (Mavko et al., 2009), whereas electrical resistivity is primarily controlled by the connected liquid water phase under Archie-type assumptions (Archie, 1942).To support the interpretation of the field inversion results, a suite of synthetic models is constructed to represent end-member and transitional ice configurations, including pore-filling ice, massive ice, and patchy distributions of ice-bearing and ice-free domains. Ice-related elastic properties are modelled using self-consistent approximation (SCA) effective-medium theory (Mavko et al., 2009), while electrical properties remain dominated by liquid water content following Archie-type relationships (Archie, 1942). Synthetic seismic and electrical datasets are generated and inverted using the same workflow applied to the field data, providing physically consistent reference scenarios for interpretation. Comparison between synthetic and field inversion results reveals systematic differences in the coupled seismic–electrical response associated with volumetric ice contributions versus elastically stiff ice contributions to the seismic response. While multiple ice configurations may reproduce either seismic velocity increases or resistivity anomalies independently (Hauck et al., 2011), only a limited subset of scenarios yields mutually consistent fits to both datasets when rock-physics constraints are considered. Mismatches between inferred phase fractions, seismic velocity enhancement, and resistivity contrasts serve as diagnostic indicators for rejecting physically implausible interpretations and avoiding interpretational pitfalls. Although a unique determination of ice type is not achievable, the combined use of rock-physics-informed joint inversion and synthetic reference models significantly reduces interpretational ambiguity. The results highlight the value of physically constrained joint inversion as a diagnostic tool for assessing the presence and elastic relevance of subsurface ice in periglacial environments.