*4.5. Phase V (26 December 2018–15 January 2019): Summit Explosive Activity*

The end of the short-lived lateral effusion was followed by an intense and powerful explosive activity mainly localized at BN/VOR sector as recorded by the integrated geophysical measurements and supported by space-based thermal data. The thermal output detected by MODIS-MIROVA persisted at moderate levels (10–100 MW), as typical of high-explosive summit activity, while the SENTINEL-2 images make it possible to confine thermal anomalies within the Central crater sector, showing an extensive and intense anomaly rising from BN sector (Figure 3l,m). Moreover, since 29 December, the VOR crater was "thermally reactivated", as shown in Figure 3l and in the thermal profiles (Figure 4a,b) retrieved by high-resolution images. Since early January 2019 all the monitored

parameters suggest a general decrease and a return to moderate explosive and sustained degassing activity, particularly from BN and NE craters (Figures 2 and 3m,n).

#### **5. Discussion**

The datasets presented in the previous chapter outline the variations of thermal, infrasonic and seismic activity continuously monitored at the Etna volcano by space- and ground-based sensors. The trends preceding and following the 24 December fissural eruption are well represented by plotting the cumulative curves of the above-mentioned parameters as recorded by SENTINEL 2 images, MVT infrasonic array and ETN seismic station, respectively (Figure 7).

**Figure 7.** Cumulative trend of: (**a**) number of hot-detected pixels retrieved by SENTINEL-2 images, (**b**) the number of hourly infrasonic detection multiplied per the hourly-average infrasonic pressure of the detected events, and (**c**) the hourly-mean seismic tremor amplitude counts In the Phase II the increasing of explosive activity marked by infrasonic measures was accompanied by a similar trend in thermal behavior track by SENTINEL-2 images. During the Phase III the thermal activity show a clear increase related to the NSE overflow episodes that cause the separation in infrasonic and thermal signals. In (**a**) the grey dotted line highlights the hypothetical contribution of explosive activity on the whole thermal activity following the infrasonic activity. Black dotted vertical line marks the onset of the lateral effusive activity as recorded by the seismic tremor amplitude.

Firstly, it can be noted that the increase in thermal activity at the end of November 2018 (Phase III in Figure 7a), was preceded by a slight increment in both thermal and infrasonic activity started since early November 2018 (Phase II; Figure 7a,b). Notably, SENTINEL 2 images allowed us to recognize the contribution of two distinct thermal sources: (i) the overflows at the eastern base of the NSE crater; (ii) the Strombolian activity localized at BN, NE and NSE vents. In contrast, the cumulate in tremor amplitude highlights that before the sharp increase preceding the 24 December event, seismic tremor stays at background level and showing any significant variations during September-November timespan (Figure 7c).

In our view, the occurrence of strombolian and outflows activity at different summit craters, coupled by sustained and growing infrasonic activity, likely represented the surface expression of a slow but continuous increase in the magma supply within the shallow portion of the magma column.

In fact, by considering Time Averaged Discharge Rates (TADR, retrieved by MODIS heat flux; see Method Section), the Phase III was characterized by an overall value of 0.5–0.9 m3s−1, for a total volume outpoured during the summit overflow episodes of 0.9–1.7 Mm<sup>3</sup> (Figure 8). On a whole, the ongoing summit effusive regime in relation with the increase in thermal flux, seems to indicate a value of 100 MW (0.3–0.5 m3 s<sup>−</sup>1), as a possible threshold for the transition between open-vent conditions to effusive activity, as already suggested by Coppola et al. [57].

**Figure 8.** TADR and Volume trend retrieved from MODIS data, measured during the effusive phase experienced by Mt. Etna on November–December 2018 by using a proposed crad of 2.5 <sup>×</sup> 108 J m−3. Blue and magenta dotted lines represent ranging estimates of plotted parameters as obtained by crad equal to 2 <sup>×</sup> <sup>10</sup><sup>8</sup> J m−<sup>3</sup> and 3.6 <sup>×</sup> 108 J m−3, respectively (see Method section).

On 24 December from the 06:00 UTC and the 11:13 UTC we record: (i) the abrupt interruption of infrasonic activity at BN crater; (ii) the beginning of a sudden increment in tremor amplitude that (iii) peaked at about the 11:13 UTC concurrently with the reawakening of a high-energetic infrasonic activity localized along the eruptive fissure (Figure 9). We interpret the three hours-long increase in tremor amplitude as a tracer of the magma- dyke injection and its propagation from the central conduit along the 2 km-long eruptive fissure extending on the western rim of the Valle del Bove. As a result, the timing of the increasing tremor let us suggest a velocity of the intrusion of 0.15–0.2 m s<sup>−</sup>1. The end of this excited short-term behavior of seismic and infrasonic activity substantially marked the onset of the fissural activity (Figure 9). This effusive episode was mainly characterized by the violent explosions testified by the high infrasonic activity (Figure 9) and by eruptive vents feeding several lava flow units, extending up to about 2.5 km (see Figure 6). Notably, the activity along the eruptive fissure caused the interruption of the overflows from the NSEC, suggesting the drainage of the shallow portion of the conduit, because of the opening of the lateral dyke. Moreover, a similar scenario may also explain the lack of infrasonic summit activity as the result of the lateral magma migration from the central conduit to the feeder dyke.

MODIS thermal data acquired after about 10 h after the beginning of this eruptive episode (24 December at 21:15 UTC) suggests that the lava discharged from the fissure vent(s) with a TADR of 9.2 (±3) m<sup>3</sup> s−<sup>1</sup> (Figure 8). However, on the early 25 December (01:25 UTC) this rate declined to a value of 2.4 (±0.7) m3 <sup>s</sup><sup>−</sup>1, indicating a rapid decrease in the effusive activity. This decline is supported by the SENTINEL-2 image of the 26 December 10:00 UTC which tracked the cooling of the lava flow, occurring after less than 48 h from the beginning of the effusive activity. As a whole, during this fissural eruption we measured an erupted lava volume of about 0.8 (±0.2) Mm3, to give a total lava volume erupted during the November–December phases of 2.1 (±0.6) Mm3 (Figure 8).

We must remind that because of the MODIS revisit time (4 image per day), all the short-lived eruptions (such as the 24 December eruption) may be severely underestimated. In this case, our estimate was mainly based to a unique peak thermal value that, however, has been recorded about 12 h after the onset of the effusive/explosive fissural eruption. Moreover, we know that Etna paroxysmal events such as effusions may be characterized by eruption rate during the first, most energetic phases, largely above the 10 m3 s−<sup>1</sup> [44,71,72]. Taking into account that a field report highlighted the high-energetic behavior of the 24 December eruption onset, we stress that the proposed whole erupted volume of 0.8 Mm<sup>3</sup> for the fissure effusive phase may be significantly underestimated.

As previously cited, the activity that followed the onset of the 24 December episode was marked by the resumption of high infrasonic activity at the BN/VOR craters accompanying strombolian explosions. Notably, this quick renewal of summit explosive activity was coupled with a tremor amplitude that remained well above the pre-eruptive level, also after the complete cessation of the effusive activity (Figure 9).

This feature seems to imply that the shallow magma drainage, occurred during the lateral effusion, enhanced a perturbation in the conduit feeding the Central Craters. Possibly, this fast emptying may have induced a decompression of the active magma column (unloading), with a consequent gas exsolution promoting a resumption of the summit infrasonic detections.

Previously, this interpretation has been invoked to explain sudden explosive behaviors ("paroxysms") at effusive basaltic systems promoting the rapid ascent of a deep undegassed magma batch [72–76]. Here the process has many similarities, with a partial drainage of magma stored in the central conduit [73–75] and a consequent downward depressurization created by the flank effusion. However, the small volume involved during the 24 December short-lived event may fails to trigger the unloading of deeper magma portion but rather cause a strong gas separation promoting related infrasonic and seismic activity.

**Figure 9.** Timeseries of the (**a**) VRP, (**b**) Infrasonic Pressure, (**c**) Infrasonic back-azimuth and (**d**) Tremor Amplitude focused on the behavior pre-during and after the 24 December Etna fissural activity. See text for details.

#### **6. Conclusions**

The combined acquisition of ground and space-based data during the September 2018–January 2019 timespan allowed us to track the changes on the Etna volcanic activity and, particularly, the longand short-term precursor signals preceding the 24–26 December fissural eruptive episode.

Our measurements showed that this short-lived activity has been heralded by monthly-long thermal precursor, tracked by MODIS and SENTINEL-2 satellite data, which clearly detect the overflow episodes at NSE crater sector. We link this activity as the surface expression of an upward migration of the active magma column that, moreover, promoted a concurrent increase in summit explosive activity measured by infrasonic arrays. By using MODIS data, we estimate that during this phase about 1.3 Mm<sup>3</sup> (±0.4 Mm3) was erupted from summit vents. The continuous increase in the supply rate pushing in the shallow portion of the magma conduit drive the intrusion of a feeder dyke producing a 2-km long NW–SE direction eruptive fissure. Acquired datasets suggest that the propagation and the emplacement of the dyke occurred at a rate of 0.15–0.20 m s−<sup>1</sup> during the three-hour long sudden increase in seismic tremor amplitude. This event was the prelude to the take place, along the fissure, of an activity characterized by violent strombolian ash-rich explosions and multiple lava flows and the concurrent interruption of overflows at NSE crater. Notably, the shift from summit and lateral activity has been revealed by the interruption of infrasonic detections at Central Craters and by a subsequent reawakening localized along the opened fissure. During the 24–26 fissural eruption, heat flux reach value typical of effusive activity (2294 MW at 24 December 21:15 UTC), such as the 26 December SENTINEL-2 image acquired at 10:00UTC capture the ending phase of this activity. The resumption of a sustained degassing activity at BN/VOR sector, recorded by the infrasonic data, seems the result of a

decompression mechanism acting on the central conduit likely due by the magma lateral drainage feeding the fissural activity.

Results highlight that satellite, infrasonic and seismic tremor data represent an invaluable support for the hazard assessment of the wide and complex Etna eruptive scenarios with an impact on the forecasting of its most energetic eruptive episodes. Finally, we stress how the integration of thermal, infrasonic and seismic data make it possible to track the shallow magma movements causing migration of the surface activity and, in general, to detect the main volcanic changes occurred at open-vent volcanoes.

**Author Contributions:** All of the authors contributed to the manuscript. In detail: M.L. and M.R. conceive the paper. M.L. and F.M. wrote the main sections of the manuscript. D.C., E.M. and F.M. prepare the Method section. M.L., F.M. and D.C. prepared the figures. M.L. and D.C. analyzed the MODIS data. F.M. and D.C. analyze the SENTINEL-2 images. E.M. and G.L. analyzed infrasonic data. E.M., G.L., L.I., M.D.S., M.C.S. and R.G. worked on processing of infrasonic and seismic data. C.C., D.C., M.R. and E.M. supervised the preparation of the manuscript.

**Funding:** Part of the research leading to these results was performed within the EUROVOLC project (https: //eurovolc.eu) and received funding from the European Community's Horizon 2020 program (grant agreement 731070). The ETN array and infrasound based early warning procedure of eruptive activity at the Etna volcano is performed by UNIFI for the Italian Civil Protection and received funding under the DPC-DEVNET project

**Acknowledgments:** MIROVA is a collaborative project between the Universities of Turin and Florence (Italy). We acknowledge the LANCE-MODIS data system for providing MODIS Near Real Time products. We acknowledge Copernicus Open Access Hub (https://scihub.copernicus.eu/) and the storage service of Amazon Web Service S3 (AWS-S3, https://registry.opendata.aws/SENTINEL-2/) for the availability and accessibility of SENTINEL-2 dataset. The comments and suggestions of two anonymous reviewers help us to improve the quality of the manuscript.

**Conflicts of Interest:** The authors declare no conflict of interest.
