(A) Tectonic setting of the Tofua volcanic arc, part of the Kermadec-Tonga subduction zone. Main volcanoes (red triangles), Hunga volcano (yellow circle), plate boundaries (black lines), and velocities relative to the Australian plate (black arrows, millimeters per year) . FT, Futuna; NI, Niuafo’ou; TO, Tonga; KE, Kermadec microplates. (B) Pre-2022 multibeam bathymetry of Hunga volcano and topography of the land extent as of September 2017.

(A and B) Marine gravity anomaly derived from satellite altimetry. Subaerial island outlines in gray. (C) Change in marine gravity anomaly between (A) and (B). (D and E) Vertical gravity gradient (VGG) data from satellite altimetry. (F) Change in VGG between (D) and (E). (G) Preeruptive bathymetry, (H) Posteruptive bathymetry, and (I) change in bathymetry between (G) and (H) attributed to the January 2022 Hunga eruption. Caldera outline (green dashed line) and Universal Transverse Mercator (UTM) coordinates (zone 1) tick marks (in kilometers) shown for reference in red.

(A) Residual Bouguer gravity anomaly calculated with a crustal density of 2750 kg/m3 and sea water density of 1027 kg/m3, after removing a quadratic polynomial regional trend. (B) Same as (A) but using posteruptive bathymetry and V32.1 of marine gravity anomaly. (C) Change in residual Bouguer gravity anomaly. (D to F) Residual Bouguer gravity anomaly and anomaly change calculated from the VGG data shown in (D), (E), and (F), respectively. Caldera outline (green dashed line) and UTM coordinates (zone 1) tick marks (in kilometers) shown for reference in red.

(A) Depth slices and (B and C) cross sections through the density model. Inversion results shown for model norm [0, 2, 2, 2]. Color scale shows the density contrast with respect to a reference density of 2.75 g/cm3. Caldera outline (green dashed line) and historical vents (stars). Solid lines are optimal sections through the preeruptive low-density anomalies, and dotted lines are optimal sections through the posteruptive low-density anomalies.

(A) Depth slices and (B and C) cross sections through the density model. Inversion results shown for model norm [0, 2, 2, 2]. Color scale shows the density contrast with respect to a reference density of 2.75 g/cm3. Caldera outline (green dashed line) and historical vents (stars). Solid lines are optimal sections through the preeruptive low-density anomalies, and dotted lines are optimal sections through the posteruptive low-density anomalies. Orange dashed line is the outline of the two main low-density anomalies for the preeruptive time interval.

Pre-2022 estimates (blue) and post-2022 estimates (yellow) presented for the inversion runs with three different model norm combinations ([1, 2, 2, 2], [1, 1, 1, 1], and [0, 2, 2, 2]) and three different isosurfaces for the central anomaly (A1), northern anomaly (A2), and northwestern anomaly (A3). Green dashed lines in the box plots indicate the samples median, and orange lines indicate the mean. Symbols not contained in the box plots are outliers. Gray squares on the last panel are the sum of the preeruptive median volume values of A1 and A2 anomalies from the three model norms.

Proposed architecture and dynamics of the Hunga magmatic system from the gravity models presented in this study. 3D volumes represent isosurfaces of constant density contrasts. Bathymetry (gray surface) is exaggerated vertically 2.4 times, and ash plume is not drawn to scale. Green arrows show the inferred movement of magmatic and/or hydrothermal fluids from one reservoir to another and replenishment from a deeper source. b.s.l., below sea level.