Direct Ice Density Constraints from Multimode Surface Waves Using DAS

Constraining subsurface density from seismic data is challenging, although density is fundamental to quantifying mass and structure in both the solid Earth and glaciers. Empirical scaling relationships between seismic wave speeds and density are therefore widely used. In this contribution, we show that density can instead be constrained directly from surface-wave observations when multimode dispersion and fully nonlinear inversion are combined.

We analyze distributed acoustic sensing (DAS) recordings acquired in glaciated environments, where strong serendipitous anthropogenic sources generate coherent Rayleigh-wave overtones with high signal-to-noise ratio. These dense DAS measurements allow robust extraction of surface-wave multimode dispersion. We invert the data using a probabilistic Hamiltonian Monte Carlo (HMC) framework that accounts for nonlinearity, parameter trade-offs, and uncertainty, while avoiding biases introduced by subjective regularization choices.

Our results show that Rayleigh-wave overtones carry resolvable sensitivity to density structure down to depths of order 100 m, enabling direct density estimation from seismic data with quantified uncertainties. We further evaluate commonly used velocity–density scaling relationships for firn (the transitional layer between fresh snow and glacial ice) and find that their application can lead to density errors on the order of 10%, with direct implications for inferred mass estimates.

Overall, these findings demonstrate that overtone-based probabilistic inversion enables constraints on weakly sensitive parameters and highlight the potential of DAS for quantitative near-surface parameter estimation.