**4. Discussion**

The key question in this paper is: How can we verify that plane-nested earthquakes are tracking magma injections in the shapes of dykes or sills beneath active volcanoes? The answer is in the form of a question, too: what other fluid except magma can demonstrate such hydrofracking capabilities? Superheated water and non-condensable gases (possibly CO2) are other candidates for such working fluid duties. However, superheated water in shallow permeable fracture reservoirs conditions is very sensitive to host rock temperatures and is easily converted into high-compressible two-phase conditions, forming geothermal production fields. Another fluid is CO2, especially magmatic CO2 having less compressibility as compared to superheated steam, and it may act as a working fluid coupling with magma. If so, then we should extend our term of magma fracking beneath active volcanoes to the term magma+CO2 fracking beneath active volcanoes. CO2 impact is also useful to explain traces of fracks in a slope of Koryaksky volcano (Figure 3), with no associated magma discharge on the surface at the same time. In this connection, it is worth noting that magmatic CO2 redistribution appears to be the key reason to explain the transfer of geysers activity from Geysers Valley to Uzon in Kamchatka and recent (2018) reactivation of Steamboat Geyser in Yellowstone.

Another important point is the accuracy of seismic hypocenter data and uniqueness of the Frac-Digger method for plane-oriented shapes definition. There is no uniqueness, since we found more fracks using less severe criteria of selections in Frac-Digger (see Section 2.1). Thus, we sugges<sup>t</sup> that 3D distributions of magma fracks be considered a plausible scenario of magma fracking beneath active volcanoes, but these may include more or less magma+CO2 dykes, depending on the Frac-Digger selection parameters. Nevertheless, some Frac-Digger selections pointed to 90%–95% of earthquake hypocenters belonging to plane-oriented clusters, meaning that fracking is a dominant process there.

High-temperature (HT) geothermal system formation due to magma fracking is also well explained in terms of the magma thermal-hydrodynamic modeling.

Conceptual 2D iTOUGH2-EOS1sc thermal hydrodynamic modeling of the Mutnovsky magmatic–hydrothermal system [7] reasonably explains its evolution over the most recent 1500–5000 years in terms of heat recharge (supplied by injected dykes from the active funnel Mutnovsky-4) and mass recharge (water injected through the dormant volcanic funnels Mutnovsky-3 and possibly Mutnovsky-2) conditions. We emphasize that the magmatic injection rate is approximately equivalent to the heat discharge rate (455 MWt). This is equivalent to approximately 455 kg/s of magma, which is in turn equivalent to from 15.3 to 25.6 km<sup>3</sup> of magma over 3000–5000 years. This volume value

is comparable to the available space for dyke accommodation under regional horizontal extension conditions [23].

Conceptual TOUGH2 modeling was used to understand and explain the mechanism of the formation of the hydrothermal system beneath northern slope of Koryaksky Volcano [4,5]. For this purpose, the following terms were found to be crucial in this model: (1) heat sources of 20 MW/km<sup>3</sup> (340 MW total in 17 km<sup>3</sup> of reservoir rocks) and gas (CO2) sources of 10 g/s/km<sup>3</sup> acting during 7000 years in the zones of magma injections; and (2) cold water recharge of 580 kg/s through the volcanic funnel to the deep dyke injection area. The modeling results reasonably match the Na–K geotemperature estimates of geothermal reservoirs (300 ◦C), the isotopic values (δD, δ18O) of high-elevation meteoric water recharge, the concentrations of magmatic CO2 (up to 4 g/kg) in the hot springs on the northern slope of Koryaksky Volcano, and the thermal reactions to the dyke injections recorded in Isotovsky and Koryaksky Narzan hot springs. This modeling also indicates that a hidden high-temperature geothermal reservoir is present also beneath the southern slope of Koryaksky Volcano (at an elevation of −1 km), which may become a subject of future drilling explorations.

Another interesting related issue is where magma came from to recharge volcanoes for magma fracking and consequent eruptions. The possible answer is that water release from the subduction plate (due to hydro-fracking in the plate itself) forms upflows into the mantle edge, which creates melting conditions for primary magma chambers at the depth of 150 km (Figure 7). This closed loop of water–magma–water circulation in the subduction zone may be responsible for strong earthquakes, when hydrofracking drives the opening of shear faults. A mini-loop of such water–magma interplay may be Karymsky Volcano, which has been working almost continuously since 1996 due to water recharge fed from Karymsky Lake into dip seismogenic faults #120 and #122 and from there into the magma chambers of the volcano (Figure 6).
