CO 2 and H 2 S Degassing at Fangaia Mud Pool, Solfatara, Campi Flegrei (Italy): Origin and Dynamics of the Pool Basin

: The Fangaia mud pool provides a “window” into the hydrothermal system underlying the degassing Solfatara crater, which is the most active volcanic centre inside the restless Campi Flegrei caldera, Southern Italy. The present study aimed at unravelling the degassing dynamics of CO 2 and H 2 S ﬂushing through the pH 1.2 steam-heated Fangaia mud pool, an ideal ﬁeld laboratory as a proxy of an active crater lake. Our results from MultiGAS measurements above Fangaia’s surface show that H 2 S scrubbing, demonstrated by high CO 2 / H 2 S ratios, was most e ﬃ cient in the portions of the basin a ﬀ ected by di ﬀ usive degassing. Convective bubbling degassing instead was the most e ﬀ ective mechanism to release gas in quantitative terms, with lower CO 2 / H 2 S ratios, similar to the Solfatara crater fumaroles, the high-T end member of the hydrothermal system. Unsurprisingly, total estimated CO 2 and H 2 S ﬂuxes from the small Fangaia pool (~184 m 2 in June 2017) were at least two orders of magnitude lower (CO 2 ﬂux < 64 t / d, H 2 S ﬂux < 0.5 t / d) than the total CO 2 ﬂux of the Campi Flegrei caldera (up to 3000 t / d for CO 2 ), too low to a ﬀ ect the gas budget for the caldera, and hence volcano monitoring routines. Given the role of the rising gas as “sediment stirrer”, the physical and chemical processes behind gas migration through a mud pool are arguably the creating processes giving origin to Fangaia. Follow-up studies of this so far unique campaign will help to better understand the fast dynamics of this peculiar degassing feature.

The Solfatara-Pisciarelli magmatic-hydrothermal system, the most active spot inside the Campi Flegrei caldera (Figure 1a), is renowned for massive CO 2 degassing through fumaroles and diffuse soil degassing (up to 858 t/d and 3000 t/d of CO 2 [15][16][17][18][19]). Despite being in a non-eruptive state, The Solfatara crater hosts a muddy lake, Fangaia (Figure 1b), characterized by vigorous bubbling and evaporation, which serves as a "window" into the underlying hydrothermal system. Geophysical surveys revealed that Fangaia is, unsurprisingly, the top of the hydrothermal aquifer underlying the Solfatara crater [25]. The Fangaia mud pool is the surface expression of two different sources of water. The first one consists of liquid, produced by steam condensation in the fumarolic areas (mainly Bocca Grande and Bocca Nuova fumarolic system and Solfatara crypto dome, but also beneath the Fangaia itself), and driven towards the Fangaia topographical depression. The second source is meteoric water [26] flowing inside Solfatara crater that carries solutes from altered deposits and transports them towards the Fangaia basin. The combination of endogenous activity and the presence of altered deposits with peculiar features permit to distinguish the Fangaia subsoil from the surrounding ones within Solfatara crater [27]. From a geochemical point of view, Fangaia is a SO4-rich (~2000 mg/L) and Cl-poor (<30 mg/L), pH ~1.5 steam-heated pool with a temperature between 30 and 47 °C [28].
In this study we aim at revealing the CO2 and H2S degassing dynamics at Fangaia, based on MultiGAS measurements conducted right above the surface of the mud pool and executed on 6 June 2017. Although from a quantitative perspective, our determination of the CO2 output from Fangaia will not massively contribute to the total CO2 output of Campi Flegrei (i.e., through Solfatara and Pisciarelli fumaroles and diffuse soil degassing [19,20,22]), instead based on variations in CO2/H2S ratios above the Fangaia surface, we will provide insights into the kinetics of gases flushing through a hydrothermal system.

Data Sources: Field, Laboratory and Desk Procedures
We adapted a portable sensor-based gas analyzer (a.k.a. MultiGAS, see [22] for more details on the instrument) by connecting a syringe to a silicon tube sustained by a fishing rod held 2-20 cm above the surface of the Fangaia mud pool (Figure 2). Gas and steam are pumped through the tube into the MultiGAS at a rate of 1.5 L/min. The syringe was held at the same spot for about 20 s to permit the gas to reach the gas analyzer. The measurement procedure (approx. 50 min, Figure 3a) is The Solfatara crater hosts a muddy lake, Fangaia (Figure 1b), characterized by vigorous bubbling and evaporation, which serves as a "window" into the underlying hydrothermal system. Geophysical surveys revealed that Fangaia is, unsurprisingly, the top of the hydrothermal aquifer underlying the Solfatara crater [25]. The Fangaia mud pool is the surface expression of two different sources of water. The first one consists of liquid, produced by steam condensation in the fumarolic areas (mainly Bocca Grande and Bocca Nuova fumarolic system and Solfatara crypto dome, but also beneath the Fangaia itself), and driven towards the Fangaia topographical depression. The second source is meteoric water [26] flowing inside Solfatara crater that carries solutes from altered deposits and transports them towards the Fangaia basin. The combination of endogenous activity and the presence of altered deposits with peculiar features permit to distinguish the Fangaia subsoil from the surrounding ones within Solfatara crater [27]. From a geochemical point of view, Fangaia is a SO 4 -rich (~2000 mg/L) and Cl-poor (<30 mg/L), pH~1.5 steam-heated pool with a temperature between 30 and 47 • C [28].
In this study we aim at revealing the CO 2 and H 2 S degassing dynamics at Fangaia, based on MultiGAS measurements conducted right above the surface of the mud pool and executed on 6 June 2017. Although from a quantitative perspective, our determination of the CO 2 output from Fangaia will not massively contribute to the total CO 2 output of Campi Flegrei (i.e., through Solfatara and Pisciarelli fumaroles and diffuse soil degassing [19,20,22]), instead based on variations in CO 2 /H 2 S ratios above the Fangaia surface, we will provide insights into the kinetics of gases flushing through a hydrothermal system.

Data Sources: Field, Laboratory and Desk Procedures
We adapted a portable sensor-based gas analyzer (a.k.a. MultiGAS, see [22] for more details on the instrument) by connecting a syringe to a silicon tube sustained by a fishing rod held 2-20 cm above the surface of the Fangaia mud pool (Figure 2). Gas and steam are pumped through the tube into the MultiGAS at a rate of 1.5 L/min. The syringe was held at the same spot for about 20 s to permit the gas to reach the gas analyzer. The measurement procedure (approx. 50 min, Figure 3a) is filmed in time-lapse mode with a GoPRO camera (GoPro Inc., San Mateo, CA, USA, https://gopro.com/), creating a view of the entire area by using a fish-eye lens ( Figure 2). Minerals 2020, 10, x 3 of 11 filmed in time-lapse mode with a GoPRO camera (GoPro Inc., San Mateo, CA, USA, https://gopro.com/), creating a view of the entire area by using a fish-eye lens ( Figure 2).   The exact location of the syringe in the image (pixel coordinates) is consequently deduced using the open-source graphical software ImageJ v. 1.8.0_172 (https://imagej.nih.gov/ij/download.html) and MTRackJ v.1.5.1 (https://imagescience.org/meijering/software/mtrackj/) (Figure 4), and converted into metric coordinates with a reference frame established in the field by the clearly visible cones ( Figure 2, see [22] for further details). Images are extracted from the GoPRO time-lapse recording with a frequency of 5 s, and are afterwards synchronized to the frequency of data acquisition of the MultiGAS (2 Hz), obtaining a total of 1512 measurements ( Figure 3). CO2 concentrations, corrected for atmospheric background, varied from 320 to 75,800 ppmV, while H2S concentrations varied from 9.2 ppmV to a saturation value of 224 ppmV (with 36 or 2.3% of the measurement points > 224 ppmV) ( Figure 3). The CO2/H2S ratios were calculated using the opensource RatioCalc software [29] (Figure 3b). All data were processed to compute the concentration maps in Figure 5 using Kriging as the interpolation method.  Figure 4), (b) CO 2 versus H 2 S concentrations using RatioCalc software (https://sites.google.com/site/giancarlotamburello/home) [29].     [22] for further details). Images are extracted from the GoPRO time-lapse recording with a frequency of 5 s, and are afterwards synchronized to the frequency of data acquisition of the MultiGAS (2 Hz), obtaining a total of 1512 measurements ( Figure 3). CO 2 concentrations, corrected for atmospheric background, varied from 320 to 75,800 ppmV, while H 2 S concentrations varied from 9.2 ppmV to a saturation value of 224 ppmV (with 36 or 2.3% of the measurement points > 224 ppmV) ( Figure 3). The CO 2 /H 2 S ratios were calculated using the open-source RatioCalc software [29] (Figure 3b). All data were processed to compute the concentration maps in Figure 5 using Kriging as the interpolation method.
The maximum concentrations for both CO 2 and H 2 S, measured at <20 cm above the Fangaia surface, are below lethal threshold concentrations (8% and 250 ppmV for CO 2 and H 2 S, respectively; except for maximum 2.3% of the H 2 S concentrations >224 ppmV, [30]). Moreover, gas concentrations decrease rapidly with height above the surface and do not present any health risk at breathing height for humans.
During our survey in June 2017, the surface area of Fangaia was~184 m 2 . The bathymetry of Fangaia was obtained using 61 direct measurements of the lake depth (Figure 6), and elaborated graphically as above (Figure 4). An operational error of ±5 cm is estimated. The estimated volume of Fangaia was~75 ± 9 m 3 . . Graphical procedure (MTrackJ and ImageJ software) to mathematically transform pixels (a) in the picture into meters (b), based on the distances between the fluorescent cones measured in the field. The green (a) and grey (b) dots are the MultiGAS measurement points. The blue-circle marks the area included in diffuse CO2 flux measurements, the subaerial degassing area in June 2017 (see Figure 7).  The CO2 flux at the degassing (water-free) shores of Fangaia is directly measured using an accumulation chamber [31] equipped with an infrared detector LICOR-LI-820 (manufacturer, city, country) (Figure 7). These areas are included in the MultiGAS measurements to obtain CO2 and H2S concentrations and ratios over areas unaffected by scrubbing. The CO 2 flux at the degassing (water-free) shores of Fangaia is directly measured using an accumulation chamber [31] equipped with an infrared detector LICOR-LI-820 (LI-COR Biosciences, Lincoln, NE, USA) (Figure 7). These areas are included in the MultiGAS measurements to obtain CO 2 and H 2 S concentrations and ratios over areas unaffected by scrubbing. The CO2 flux at the degassing (water-free) shores of Fangaia is directly measured using an accumulation chamber [31] equipped with an infrared detector LICOR-LI-820 (manufacturer, city, country) (Figure 7). These areas are included in the MultiGAS measurements to obtain CO2 and H2S concentrations and ratios over areas unaffected by scrubbing.  The percentage of suspended solids (i.e., "muddiness") in the Fangaia water is obtained after centrifugation (5000 cycles/min), filtering, drying and weighing of the solid phase. Despite the muddy appearance, the small Fangaia contained only 2.85 wt% of solids, defining it as a proxy of a "miniature steam-heated lake".
A water sample was analyzed for its major element composition using ion chromatography (IC; manufacturer at INGV-OV, Napoli, Italy), after centrifugation and filtering in the lab. The pH was measured in the lab by a portable pH meter while the temperature was measured in situ. The steam-heated character (Cl-poor, SO 4 -rich) is confirmed by the chemical composition: Na + 17 mg/L, K + 20 mg/L, Ca 2+ 48 mg/L, Mg 2+ 3.9 mg/L, F -0.46 mg/L, Cl − 12 mg/L, SO 4 2− 1720 mg/L, NO 3 − 3 mg/L, with a pH of 1.2 and a T of 32 • C, similar to that reported by [28]. Figure 5a,b shows integrated concentration maps of CO 2 and H 2 S above Fangaia. The highest absolute concentrations for both species were detected above the bubbling areas near the eastern shore of Fangaia, suggesting that most efficient degassing for both species occurs via bubbling. Our expectation of higher CO 2 and H 2 S concentrations above the dark colored areas around the center degassing vents was not confirmed by the MultiGAS measurements (Figure 5a,b). Instead, the CO 2 /H 2 S map reveals low ratios (100-200) above bubbling degassing areas and generally high ratios above diffusively degassing areas (>300; Figures 5c and 8). The CO 2 /H 2 S ratios above bubbling degassing vents correspond to the ratios observed in the Bocca Nuova and Bocca Grande fumaroles (155 ± 52, for 33 analyses [23]), the high temperature end member (142-163 • C) representative of the Solfatara hydrothermal system, discharging 200 m east of Fangaia. The gas released at the bubbling degassing vents can hence be interpreted as a proxy of the Solfatara fumarole emissions. Remembering that pH 1.2 water is transparent to CO 2 , the practically identical CO 2 /H 2 S ratios of the high-T fumaroles and the bubbling degassing vents imply that H 2 S scrubbing in Fangaia is near-nil through bubble gas rise. The trend in varying CO 2 /H 2 S ratios agrees with a hypothesis of efficient H 2 S scrubbing along the water column during slower diffuse, and less efficient scrubbing associated with the relatively rapid rise of bubbling gas. The highest CO 2 /H 2 S ratios (>400) are detected near the southwestern steep-walled shore of Fangaia (Figures 5c and 8).

CO 2 and H 2 S Concentrations and CO 2 /H 2 S Ratios
degassing vents correspond to the ratios observed in the Bocca Nuova and Bocca Grande fumaroles (155 ± 52, for 33 analyses [23]), the high temperature end member (142-163 °C) representative of the Solfatara hydrothermal system, discharging 200 m east of Fangaia. The gas released at the bubbling degassing vents can hence be interpreted as a proxy of the Solfatara fumarole emissions. Remembering that pH 1.2 water is transparent to CO2, the practically identical CO2/H2S ratios of the high-T fumaroles and the bubbling degassing vents imply that H2S scrubbing in Fangaia is near-nil through bubble gas rise. The trend in varying CO2/H2S ratios agrees with a hypothesis of efficient H2S scrubbing along the water column during slower diffuse, and less efficient scrubbing associated with the relatively rapid rise of bubbling gas. The highest CO2/H2S ratios (>400) are detected near the southwestern steep-walled shore of Fangaia (Figures 5c and 8).  A recent study [9] argued that the oxidation state and chemical composition of the gas-receiving lake (El Chichón crater lake in their case) is decisive in the degree of H 2 S scrubbing. The very small size of Fangaia, the well-located degassing vents, and the shallow nature of the basin enable a clear differentiation in the response to H 2 S scrubbing with respect to the inert CO 2 along the lake surface. In other words, Fangaia is not large and deep enough to homogenize the composition of the gas flushing through the water, as generally is the case at large and deep crater lakes (e.g., Aso, Poás, Kawah Ijen [3,[6][7][8]). Considering Fangaia as a "window" into the Solfatara hydrothermal aquifer, variations in CO 2 fluxes and CO 2 /H 2 S ratios of soil gases can be affected by the spatial and temporal variations in water saturation levels of the aquifers beneath Solfatara's crater, as well as by the structural control.

CO 2 and H 2 S Fluxes
Three-point flux measurements of diffuse CO 2 soil degassing around Fangaia varied from 8400 g m −2 d −1 (above the southwestern shores, outside the basin of Fangaia) to 35,600-79,000 g m −2 d −1 (inside the dried-out basin, near the bubbling degassing vents, shown by the blue ellipses in Figures 4-6). If we tentatively scale these values to the Fangaia area (~184 m 2 ), we estimate total CO 2 emissions ranging from 6.6 to 14.6 t/d, and an H 2 S flux ranging from 0.05 to 0.12 t/d, obtained by dividing the CO 2 flux by the CO 2 /H 2 S weight ratio of~100. The latter ratio was measured near the dry sector of the basin (Figure 5c), where the diffuse gas flux measurements were performed.
We also calculated the gas fluxes by integrating the interpolated gas concentrations from Figure 5a and b (the integrated column amount, ICA [32]) and multiplying those values by a plume transport speed. The resulting ICA was 0.75 kg/m for CO 2 , and 0.0074 kg/m for H 2 S. The plume transport speed was estimated considering that the buoyant rise of a hot volcanic gas ranges from 1 to 3 m/s [32,33], and assuming a rising speed of <1 m/s for the colder gas of the Fangaia. This estimate results in an upper limit for CO 2 and H 2 S fluxes of <64 t/d and <0.5 t/d, respectively. The discrepancy between the gas fluxes derived from these two independent methods (i.e., accumulation chamber versus MultiGAS/ICA), suggests that (1) the diffusive gas flux per unit area inside the Fangaia basin might be much higher than the flux detected in the dried-out basin, and/or (2) the buoyant speed of the gas emitted by the Fangaia is much lower than our hypothesized upper limit of 1 m/s. However, our estimates provide an order of magnitude estimate of the gas flux from the Fangaia mud pool, that is comparable to the fluxes of the smaller fumaroles of Solfatara and Pisciarelli [15]. This finding suggests that the degassing of Fangaia does not significantly affect the total CO 2 and H 2 S budget, routinely measured during the geochemical monitoring of Campi Flegrei.

Origin and Dynamics of the Fangaia Basin
Heterogeneous degassing and steam condensation processes ongoing inside the mixed meteoric-hydrothermal aquifer, at the intersection with the Solfatara crater surface, affected by Minerals 2020, 10, 1051 8 of 10 fracture-controlled diffuse degassing [19], are probably causes of Fangaia's existence. The Fangaia basin was not formed after a (phreatic) eruption, and is hence, by definition, not a crater lake. The many funnel-shaped degassing vents rework sediments and bring them into suspension, creating a mud pool, and leads to an extremely dynamic sedimentary regime. The small basin is highly sensitive to direct rain fall contribution, causing drastic short-term variations in area and volume. E-W and NNE-SSW lineaments cross-cut the area [19]. These lineations in the location of bubbling sites at Fangaia (Figure 5), show it to be part of the Solfatara Diffuse Degassing Structure (DDS [19]), and evidence the structural control on degassing.
Despite the absence of massive CO 2 and H 2 S degassing, and the low probability for phreatic eruptions, fast changes in the morphology of the mud pool and its surroundings, and fracture degassing in the area, can lead to the opening of new fractures and fissures. Unfortunately, such rapid fracturing recently occurred after an anomalous rainfall event flooded Fangaia. Three days after, on 12 September 2017, three people died of asphyxiation after falling in a newly formed, poorly ventilated crevasse outside the restricted area of Fangaia. Our June 2017 "snapshot survey", three months prior to the accident, can only reveal the dynamic character of the Fangaia area, but, being first in its kind, could not possibly provide precursory signals in the degassing and morphological-sedimentary regime prior to the accident.

Conclusive Remarks
We describe a creative field-based method to measure CO 2 and H 2 S concentrations using a MultiGAS immediately above the mud pool Fangaia, in the emblematic Solfatara crater, the most active spot of the restless Campi Flegrei caldera. The small size of the pool (only 184 m 2 in June 2017) and easy access provided (1) an excellent test site for in-field measurements, and (2) insightful results on degassing dynamics of hydrothermal gases flushing through shallow acidic water bodies, a "miniature proxy" to active crater lakes. Diffuse degassing leads to more efficient scrubbing of H 2 S and hence higher CO 2 /H 2 S ratios, whereas bubbling degassing is an inefficient scrubbing mechanism, as indicated by low CO 2 /H 2 S ratios, to the degree that gasses emitted at the vents are chemically similar to the purest hydrothermal end member released in the Solfatara crater through the Bocca Nuova and Bocca Grande fumaroles. Variations in gas species ratios are diagnostic of changes in volcanic activity. Tracking such variations above the Fangaia bubbling vents can hence become a future monitoring focus.
Contrary to large active crater lakes that seem to show a chemically more homogeneous diffuse plume degassing without bubbling at the surface, the extremely shallow Fangaia highlights heterogeneous degassing through both vigorous gas bubbling and diffusion. Basin depth has arguably a quantitative (gas concentrations and fluxes) and qualitative effect (ratios between gas species) on degassing.
Despite being imposed on top of the hydrothermal system of the Solfatara crater, Fangaia is not a crater lake. The basin itself is formed by intense degassing and steam condensation processes through innumerable funnel-shaped vents, which, in combination with the high sensitivity to rainfall, leads to continuous remobilization of sediments. Rapid changes in morphology of the already weak area can lead to fracturing and destabilization of the basin. A follow-up of this pioneering study is recommended to track possible changes in Fangaia's dynamics, with a focus on (1) geomorphological and hydrological changes with time, on a short-(days) and long-term (months-years) basis, using a water mass balance approach, (2) the relationship between water and gas chemistry, to increase insights into gas scrubbing, (3) possible formation of secondary minerals due to changes in gas-water interactions, and (4) quantifying gas flux and tracking the possible migration of degassing vents, given their physical role of "sediment stirrer", all with the scope to better understand the evolution of the system. Author Contributions: D.R. field measurements, data analysis, main manuscript drafting; G.T. instrument development, field measurements, data analysis, manuscript drafting; T.R. field measurements, manuscript drafting; A.S. field measurements, data analysis; F.C. water sample analysis, manuscript drafting; S.C. manuscript drafting, responsible funding. All authors have read and agreed to the published version of the manuscript.
Funding: This study was funded by the Italian Dipartimento della Protezione Civile, Presidenza del Consiglio dei Ministri (DPC). This paper does not necessarily represent DPC official opinion and policies. The APC was funded by internal funding of INGV, Sezione di Bologna.