Spatial and Temporal Patterns of Seismic Anisotropy on Volcanoes and Geothermal Areas

Seismic anisotropy may reveal the state of stress in the crust, and its temporal changes have been attributed to deformation, seismicity, magmatic activity and geothermal extraction.  We review crustal anisotropy in volcanic and geothermal regions.  We compile the results to test hypotheses about the origin of anisotropy and about its utility for monitoring magmatic unrest or geothermal production. The majority of the articles that were published through 2019 (~100) examined shear-wave splitting.

Of the 88 studies examining the effects of stress vs. structure, the results were about evenly divided between causes related entirely to regional stress (16), local stress (10) or structure (11) alone or combinations of these possibilities. Delay times (a measure of anisotropy strength) increased with period and with depth in the two sets, but with much scatter.  Because geothermal areas tended to be studied at shallower depths (median 2.5 km), they yielded lower delay times (0.1 s) at shorter periods (0.1 s) than volcanoes (median 12 km depth, 0.25 s period, 0.19 s time delay and 6% anisotropy).

Surface wave studies of anisotropy have also become more common, and they are often interpreted in terms of radial anisotropy, i.e., the difference between horizontally polarised waves (S_H) and vertical polarisations (S_V). In volcanic areas, they can distinguish between magmatic storage in dykes, in which S_V >S_H , or sills, with  S_H >S_V. Because the lower crust in non-volcanic areas often has S_H >S_V, the presence of low absolute velocity should be used to confirm that magma is involved.

Time variations in shear wave splitting were examined in 29 studies, but few of these presented statistical tests.  Studies were divided between those that reported changes in delay times (12) or fast azimuths (8) alone, or both (8). Time variations were mostly reported to vary with the occurrence of eruptions or intrusions (19 volcanoes), seismicity or tremor rate changes (9), or deformation changes such as GNSS, tilt or strain measurements (10).  Focal mechanisms, b-value, isotropic velocity, Vp/Vs ratio, gas flux, coda Q, unrest level, geothermal activity, and fluid injection were also correlated with splitting in some studies. There is a clear need for studies that examine statistical relationships between anisotropy and other parameters to test monitoring capabilities.