Magnetic imaging of subseafloor hydrothermal fluid circulation pathways

Abstract

Hydrothermal fluid circulation beneath the seafloor is an important process for chemical and heat transfer between the solid Earth and overlying oceans. Discharge of hydrothermal fluids at the seafloor supports unique biological communities and can produce potentially valuable mineral deposits. Our understanding of the scale and geometry of subseafloor hydrothermal circulation has been limited to numerical simulations and their manifestations on the seafloor. Here, we use magnetic inverse modeling to generate the first three-dimensional empirical model of a hydrothermal convection system. High-temperature fluid-rock reactions associated with fluid circulation destroy magnetic minerals in the Earth's crust, thus allowing magnetic models to trace the fluid's pathways through the seafloor. We present an application of this modeling at a hydrothermally active region of the East Manus Basin.

Fig. 1. A map of the East Manus Basin in the region of the Tumai and Bugave Ridges, along with this study’s total magnetic field anomaly maps.

(A) A 35-m resolution bathymetric map of the East Manus Basin, centered about the Tumai Ridge and Bugave Ridge intersection. The inset map shows the major geological features near Susu Knolls (50) and the locations of the known active and inactive hydrothermal vent sites along the Tumai Ridge (22, 30), with 50-m bathymetric contour lines. Dashed gray and white lines represent the extents of the magnetic surveys. (B to D) Total magnetic field anomaly maps for the regional survey and both deposit-scale surveys, respectively. The black dots mark the positions of the measurement locations.

Fig. 2. Map and cross sections of effective magnetic susceptibility from the regional 3D magnetic inversion model.

A-A′ displays two prominent magnetic lows at ~2 km below the seafloor, which we interpreted to be magma chambers. The southwest magma chamber is the primary heat source driving hydrothermal circulation at Solwara 5 and Susu Knolls. B-B′ shows subseafloor high-temperature fluid pathways that feed the venting sites along the Tumai Ridge. C-C′ and D-D′ are two cross sections that pass through the hydrothermally active Kaia Natai volcano. The depth to the Curie isotherm is included on the plan view image as a contour map with orange lines, mapping the geometry of the underlying magmatic bodies.

Fig. 3. Map views and cross sections of the two deposit-scale 3D magnetic models.

(A) The 2007 data inversion model, with cross sections: A-A′ shows a vertical alteration column associated with ascending hydrothermal beneath Solwara 1; B-B′ shows variation in near-surface effective magnetic susceptibility along the Tumai Ridge, with focused low susceptibility zones positioned at Solwara 5 and the three venting sites of Susu Knolls. An additional site of possible mineralization, southeast of Solwara 5, is interpreted from its magnetic characteristics, which match those of nearby known hydrothermal venting sites. (B) The 2006 AUV data inversion model, with a single cross section C-C′ showing partitioning of the fluid pathways within Susu Knolls leading to known vent sites. All symbols from the legend of Fig. 2. bsl, below sea level.

Fig. 4. A 3D model of the high-temperature hydrothermal upflow column below the Tumai Ridge.

The shown cross section is B-B′ from Fig. 2, with the alteration column visualized with a 0.12 SI threshold of the regional model’s effective magnetic susceptibility. (A) View of the column facing northeast. (B) View of the same column facing northwest. All surface hydrothermal feature symbols and the color scale follow the legend in Fig. 2.

Fig. 5. The three meshes used for our inverse modeling, shown in relation to each other.

At the top is the mesh for the regional inversion model, from the 2016 M/V Miss Rankin cruise, with the two deposit-scale meshes below it in descending order of survey size, from the 2007 M/V Genesis then the 2006 R/V Melville cruises. All meshes are viewed from the south.

Fig. 6. The normalized data residuals for the three inverted datasets.

(A) Normalized data residual for the inverted 2006 AUV magnetic dataset, (B) for the inverted 2007 deep-tow magnetic dataset, and (C) for the inverted 2016 deep-tow magnetic dataset.