- Email: [email protected]
Physical properties of tuffs from a scientific borehole at Alban hills volcanic district (central Italy)
Tectonophysics 471 (2009) 161–169
Contents lists available at ScienceDirect
Tectonophysics j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / t e c t o
Physical properties of tuffs from a scientific borehole at Alban hills volcanic district (central Italy) S. Vinciguerra a,⁎, P. Del Gaudio a, M.T. Mariucci a, F. Marra a, P.G. Meredith b, P. Montone a, S. Pierdominici a, P. Scarlato a a b
Istituto Nazionale di Geofisica e Vulcanologia, Sezione di Roma1, I-00143 Rome, Italy Department of Earth Sciences, University College London, WC1E 6BT London, UK
a r t i c l e
i n f o
Article history: Received 24 April 2008 Received in revised form 8 July 2008 Accepted 7 August 2008 Available online 22 August 2008 Keywords: Scientific borehole Volcanic rocks Physical properties
a b s t r a c t Recent seismic swarms and hydrothermal activity suggest that the Quaternary volcanic complex of the Alban Hills may pose a threat to the city of Rome. A 350 m scientific borehole was therefore drilled into this volcanic area to elucidate its inner structure for the first time. Wire-line logs were run in the borehole in order to characterize the physical properties of the rocks and their variations with depth. In particular, a detailed sonic log was run to measure the P-wave velocity from the well-head down to 110 m. To further investigate velocity changes, we carried out laboratory measurements of P and S elastic wave velocities and fluid permeability at effective pressures up to 70 MPa during both increasing and decreasing pressure cycles on selected core samples representative of the main volcanic units. Specifically, we studied samples from two pyroclastic units representative of two classes of volcanic deposits that are representative of the whole succession: (i) a coarse-grained, well-lithified facies (Pozzolane Rosse unit), containing abundant mm-to-cm lava clasts and crystals; and (ii) a fine-grained, matrix-supported pyroclastic deposit (Tufo Pisolitico di Trigoria unit), with rare lithic lava clasts and sparse pumice. Elastic wave velocities reveal significant differences between units and indicate how, within the same lithology, the different degree of lithification and presence of clasts can affect significantly physical property values. The mean laboratory value of the Pwave velocity for Pozzolane Rosse and Tufo Pisolitico di Trigoria units is respectively of 3.75 and 3.2 km/s at an effective pressure equivalent to that at the depth at which the sonic velocity was measured. Under increasing effective pressure a profound influence on the transport properties is observed. Permeability ranges from the order of 10− 18 m2 for the Pozzolane Rosse unit to the order of 10− 15 m2 for the Tufo Pisolitico di Trigoria unit, in good agreement with the shallow aquifer circulating in the shallower units. © 2008 Elsevier B.V. All rights reserved.
1. Introduction The Alban Hills form part of the 200-km-long, Roman Magmatic Province (Washington, 1906), a chain of ultra-potassic, mainly explosive volcanic districts and small eruptive centres that developed along the Tyrrhenian margin of central Italy since the Middle Pleistocene. The regional structural trend follows the main NW–SE oriented extensional faults characterizing the Apennines and the Tyrrhenian Sea margin, and it is crossed by a local N–S, right-lateral fault system developed behind the Olevano–Antrodoco thrust front (Fig. 1a). This latter feature represents a major geodynamic boundary, along which the Umbria–Marche–Sabina pelagic/transitional series of the northern Apennines over thrusts the Latium–Abruzzi carbonate platform of the central Apennines. The Alban Hills are located where
⁎ Corresponding author. Tel.: +39 06 51860478; fax: +39 06 51860507. E-mail address: [email protected] (S. Vinciguerra). 0040-1951/$ – see front matter © 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.tecto.2008.08.010
recognized N–S strike–slip structural trends meet the NW–SE extensional ones. The present-day stress field in the area is not well constrained, and its interpretation is the subject of current debate (e.g. Faccenna et al., 1994; Montone et al., 1995; Marra, 1999). In particular, it is not clear whether a still-standing extensional regime characterized by a vertical σ1 and a NE-oriented σ3 affects this area, or a substantial similarity of σ1–σ2–σ3 causes the frequent permutation of the stress tensor, allowing for repeated superimposition of two competitive tectonic regimes (Marra, 2001). Enhanced deformation (Amato and Chiarabba, 1995; Salvi et al., 2004) and seismic activity (Amato et al., 1994; Feuillet et al., 2004) have both been reported in the area of the most recent volcanic activity. In particular, instrumental records of the 1989–1990 seismic swarm, that struck the western flank of the volcano, suggest that the main faults are active under a NE extensional regime (Fig. 1b). In contrast, both recent and historical deformations of the surrounding Acque Albule travertine basin (Fig. 1a) have been interpreted as the effect of strike–slip tectonics under a NE trending σ1 (Faccenna et al.,
162
S. Vinciguerra et al. / Tectonophysics 471 (2009) 161–169
S. Vinciguerra et al. / Tectonophysics 471 (2009) 161–169
1994; Marra et al., 2004). Furthermore, two main fracture systems, SW and NW dipping, have been recognized from recent down-hole measurements in the volcanic units close to the Tuscolano–Artemisio caldera rim (Mariucci et al., in press), and interpreted as occurring from the superimposition of the two tectonic styles in the area. However, analysis of borehole breakouts defined an active stress field with an approximately E–W oriented minimum horizontal component, which may be interpreted as evidence of strain-partitioning along the local N–S structures due to their re-activation under a NEstriking extensional regime (Mariucci et al., in press). The scientific community has recently increased its efforts to understand the internal structure of the volcano through tomographic studies based on inversion of seismicity from local earthquakes (Feuillet et al., 2004; Chiarabba et al., 1994). While high velocity zones have been interpreted as either solidified magma reservoirs or uplifted basement, the observed low velocity anomalies on the western flank suggest either the presence of molten material or fluid overpressure (Chiarabba et al., 1997). The documented unrest episodes of enhanced seismic activity and deformation are also accompanied by anomalous gas emissions (CO2, 222Radon, H2S) (Pizzino et al., 2002; Carapezza et al., 2003) controlled by the interplay between a shallow aquifer (less than 1 km of depth), hosted within the volcanic pile, and a reservoir buried in the carbonate basement, separated by low permeability Plio-Quaternary deposits (Boni et al., 1995; Di Filippo and Toro, 1995; Capelli et al., 2001). As a consequence, unraveling the velocity structure and the conditions controlling fluid flow within the volcanic complex, and their evolution with depth, have been primary drivers of scientific investigation. Hence, a 350 m deep scientific borehole was recently drilled in the western sector of the volcanic complex, allowing both wire-line logging and continuous core recovery to be carried out. Here we present results from a joint study in which (i) the results of laboratory measurements of porosity, P and S elastic wave velocities and fluid permeability on recovered cores under simulated in situ conditions are integrated with (ii) in situ measurements of P and S wave velocities from down-hole sonic logs. This integration of multi-scale measurements of physical properties allows us to constrain the velocity structure of the volcanic complex and highlight the physical mechanisms controlling the fluid circulation within the different volcanic lithologies. 2. Volcanic stratigraphy The volcanic history of Alban Hills may be roughly divided into three main phases, marked by different eruptive mechanisms and magma volumes (De Rita et al., 1988, 1995; Giordano et al., 2006). The early Tuscolano–Artemisio Phase (c. 561–366 ka; Karner et al., 2001; Marra et al., in press) was the most explosive and voluminous, and is characterized by five large pyroclastic-flow forming eruptions with volumes of the order of tens of km3 and minor effusive activity. Each one of these cycles ended with caldera collapse followed by intra- and peri-caldera effusive and strombolian activity. The last cycle originated the so-called Tuscolano–Artemisio caldera and its ring of scoria cones in the period c. 366–351 ka. After a c. 40 kyr-long dormancy, the less energetic Faete Phase of activity occurred (c. 308–250 ka; Marra et al., 2003). This phase started with peripheral effusive eruptions coupled with subordinate hydromagmatic activity, and encompassed the formation of the Faete central edifice and several minor scoria cones within the older intracaldera area. A further c. 50 kyr-long dormancy then preceded the start of the Late Hydromagmatic Phase (c. 200–36 ka; Marra et al., 2003; Freda et al., 2006; Giaccio et al., 2007), which was dominated by pyroclasticsurge eruptions with the formation of several monogenetic or multiple maars and/or tuff rings clustered southwest of the Mt. Faete edifice.
163
The products of the Tuscolano–Artemisio phase represent by far the largest portion of the erupted deposits and constitute the greater portion of the volcanic substrate of the area of Rome. These deposits may be grouped into two main classes, based on their petrographic characteristics and their eruptive mechanism (Marra et al., in press). The earlier eruptive units (Tufo Pisolitico di Trigoria, 561 ±1 ka; Tufo del Palatino, 528 ±1 ka; Tufo di Acque Albule, 526 ±1 ka; Marra et al., in press) are the result of hydromagmatic eruptions and are characterized by prevalently fine-grained, pyroclastic ash-flow deposits. The later units (Pozzolane Rosse, 456±3 ka; Pozzolane Nere, 407±2 ka; Villa Senni, 365 ±4 ka; (Marra et al., in press) are the result of dry eruptions and are characterized by coarse-grained, pyroclastic scoria and ash-flow deposits. 3. The drilling project As reported above, a 350 m scientific borehole was drilled in the western sector of the Alban Hills area (Fig. 1a), where seismic swarms (Fig. 1b), enhanced ground deformation and high gas concentration in water had been observed. The aim was to elucidate the internal structure and determine the stress field by means of down-hole measurements (see Mariucci et al., in press). From a tectonic point of view, data from the borehole allowed two main fracture systems (dipping SW and NW) to be recognized, indicating a prevalent strike–slip component of fault movement. By contrast borehole breakout analysis defined an active stress field with an approximately E–W oriented minimum horizontal component (Fig. 1a). Geochemical analysis of deep fluids from the borehole revealed water of meteoric origin which had a long residence time in the aquifer, and gas dominated by CO2 content. The gas exhibited one of the highest magmatic components from the Alban Hills, which was partly deriving from a magmatic source and partly from the thermal decarbonization of carbon-rich basement rocks (Mariucci et al., in press). The borehole was drilled using the wire-line coring technique that allowed us to retrieve a complete core record and hence establish a detailed stratigraphy (Fig. 2a). Wire-line logs (Fig. 2b) were also run over selected intervals in order (i) to characterize the in situ physical properties of the rocks (acoustic wave velocities, electrical resistivity, natural gamma radiation, magnetic susceptibility) and their variations with depth, and (ii) to define the structural setting and assess the present-day stress field (by borehole televiewer and caliper log). All the down-hole logs were run by the ICDP Operational Support Group (GFZPotsdam), and the data was analysed using WellCad software (ALT, Luxembourg). A more detailed description can be found in Mariucci et al. (in press). The detailed borehole stratigraphy allowed us to correlate the geophysical logs with specific rock features (Fig. 2a,b). The main lithology cored was the pyroclastic-flow deposits (tuffs) of the Tuscolano–Artemisio explosive phase of activity (561–365 ka), which exhibited a wide variability in grain-size and cohesion. We have chosen two units for further study (Pozzolane Rosse and Tufo Pisolitico di Trigoria), which are representative of the two main lithologic classes of erupted products described earlier. For clarity, we will simplify the names and use PR for the Pozzolane Rosse unit and TPT for the Tufo Pisolitico di Trigoria units. In outcrop, the PR deposit is described as a single, massive, poorlysorted pyroclastic-flow deposit. It is characterized by a coarse-ash matrix, poorly consolidated or locally lithified due to vapour-phase zeolite crystallisation (De Rita et al., 1995). It includes reddish-purple to dark grey, poorly to moderately vesicular, scoria lapilli and blocks, abundant lava and thermally metamorphosed sedimentary lithic lapilli and blocks, and scarce granular (leucite+ clinopyroxene) inclusions. In contrast to what is commonly observed in outcrops, in the recovered cores the PR displays a very well-lithified facies, resulting from a pervasive
Fig. 1. Overview of the geological and geophysical characters of the area. a) Regional tectonic structures and stress field (modified after Marra, 1999). b) Geological sketch of volcanic district with CA1 borehole location and seismicity (modified after Mariucci et al. in press): 1) Holocene alluvial deposits; 2) Plio-Quaternary sedimentary deposits; 3) “Hydromagmatic phase” (200– 36 ky); 4) deposits of “Faete phase” and related lava flows (308–250 ky); 5) “Tuscolano–Artemisio phase” and related lava flows (561–351 ky); 6) crater rims; 7) recent seismic swarms (after Feuillet et al., 2004).
164
S. Vinciguerra et al. / Tectonophysics 471 (2009) 161–169
Fig. 2. a) CA1 borehole stratigraphy of the volcanic units, belonging to the Tuscolano–Artemisio Phase (ca. 561–351 ky). Legend: P — “Pozzolanelle” (366 ± 4 Ka); PN — “Pozzolane Nere” (407 ± 4 Ka); FL — “Fioranello Lava”; PR — “Pozzolane Rosse” (457 ± 4 Ka); VL — “Vallerano Lava” (460 ± 4 Ka); AFS — “Air Fall Sequence” (488 ± 2 Ka); TAA — “Tufo di Acque Albule” (526 ± 1 Ka); TP — “Tufo del Palatino” (528 ± 1 Ka); TPT — “Tufo Pisolitico di Trigoria” (561 ± 1 Ka). Small grey intervals (not to scale) are paleosols. Ages are after Karner et al. (2001). b) P-wave velocity from down-hole sonic log.
zeolitization of the ash matrix. Abundant mm-to-cm lava clasts and mmsized pyroxene and biotite crystals are embedded within the coarse-ash matrix. The observed strong cementation that occurs in the intermediate, scoriaceous volcanic deposits (Pozzolane Rosse) is the likely result of secondary process of transformation of primary volcanic particles into authigenic minerals, due to the chemical action of the ground water. The outcropping deposits of the TPT occur as a light gray, indurate, matrix-supported, stratified coarse-ash deposit (Palladino et al., 2001). It includes altered leucite and abundant mm-to-cm-sized, poorly vesiculated gray scoria lapilli, and subordinate lava, carbonate and tuff lithics. Up to cm-sized accretionary lapilli (pisoliti) are frequent. In the recovered cores, the TPT occurs as a fine-grained, matrix-supported, lithified pyroclastic deposit. It contains rare mm-sized lithic lava clasts, sparse pumice and carbonatic clasts, and accretionary lapilli. 4. Laboratory petrophysical investigation 4.1. Methodology All measurements were carried out on cylindrical sub-core samples 38 mm in diameter by 40 mm long, taken from the recovered
borehole cores. Initial porosities were determined gravimetrically by weighing samples of known volume both dry and water saturated. Initial elastic wave velocity measurements were made on dry samples under ambient laboratory conditions. Measurements were made radially at azimuthal steps of 10°. In order to understand the variation of rock properties with increasing effective pressure, we then carried out simultaneous measurements of ultrasonic P and S wave velocities and fluid permeability in a servo-controlled, steady-state-flow permeameter. The permeameter comprises a hydrostatic pressure vessel, using silicone oil as the pressure medium, equipped with two 70 MPa servocontrolled fluid pressure intensifiers (volumometers) to provide pore fluid pressure independently to each end of the sample and piezoelectric ultrasonic transducers for P and S wave velocity measurements (Fig. 3, plus see also Benson et al., 2003, for further details). Rubber-jacketed samples are placed inside the pressure vessel, located between two steel end-caps. Measurements were made at effective pressures from 5 to 80 MPa, during both increasing and decreasing pressure cycles. The effective pressure was increased in steps of 5 MPa from 5 MPa to 40 MPa and then in steps of 10 MPa from 40 MPa to 70 MPa.
S. Vinciguerra et al. / Tectonophysics 471 (2009) 161–169
Fig. 3. Schematic sketch of the servo-controlled permeameter equipped for Vp, Vs measurements, sample set up and pressure vessel.
A 900 V pulse generator was used to excite a 1 MHz resonant frequency piezoelectric transmitting transducers for measurements of both P and S wave velocities. Waveforms were captured using an identical receiving transducer, and were pre-amplified by 40 dB before being recorded and displayed on a digital storage oscilloscope. Accuracy is estimated as 1% for P waves and 2% for S waves. In order to measure fluid permeability, a fixed pore fluid pressure gradient is set across the sample and the steady-state volume flow rate recorded as a function of time. The permeability can then be determined directly from the measured volume flow rate and the sample dimensions using Darcy's law. Microstructural analysis was carried out on both as-received and experimentally pressurized samples using a high-resolution thermal field emission scanning electron microscope (FESEM) in order to study changes in the microstructure due to pore collapse or crack damage induced during the pressurization and de-pressurization cycles. The FESEM generates images with much lower electrostatic
Table 1 Porosity, P and S wave velocities at room pressure and dry conditions. Sample
Porosity (%)
Vp (m/s)
Vs (m/s)
PR1 PR2 PR3 PR4 PR5 PR6 PR7 PR8 TPT1 TPT2 TPT3 TPT4 TPT5 TPT6
31.12 11.00 21.60 10.80 11.00 17.00 10.20 19.40 14.10 31.00 15.00 17.20 13.80 15.10
3601 3732 3610 3715 4130 3836 3728 3702 3226 3175 2930 3415 3159 3453
2590 2590 2604 2550 3102 2833 2703 2856 2345 1960 2355 2465 2789 2466
Fig. 4. Radial Vp for (a) PR and (b) TPT.
165
166
S. Vinciguerra et al. / Tectonophysics 471 (2009) 161–169
distortion than a conventional SEM, and with a spatial resolution that is 3 to 6 times higher (i.e. less than 2 nm). 4.2. Results The lithologies investigated exhibited initial porosities ranging from around 10% to over 30% (Table 1), with an average porosity of 14.2% and 17.8% for PR and TPT respectively. Axial P and S wave velocities for dry samples under ambient conditions range from 3.6 to 4.1 km/s and 2.5 to 3.1 km/s, respectively, for PR. P and S wave velocities are found to be lower for the TPT samples at 2.9 to 3.4 km/s and 1.9 to 2.7 km/s, respectively (Table 1). Radial P-wave velocity measurements (Fig. 4a) show that PR exhibits a considerable velocity anisotropy of up to about 25%, while the Tufo Pisolitico (Fig. 4b) exhibits a very much lower velocity anisotropy of only about 3%. We consider that the differences in mean porosity and axial wave velocities are due to the much more developed lithification of the PR with respect to the relatively poorly consolidated pyroclastic deposits that make up the TPT. The abundance of mm-to-cm size lava clasts in the PR, which are almost entirely absent in the TPT matrix, likely determines both the higher porosity variability and the higher degree of anisotropy observed in the PR samples. This suggests that even within the same tuff lithology, such variations can significantly affect rock physical properties. Measurements of P and S wave velocities (Fig. 5a,b) carried out on water-saturated samples (wet) under elevated effective pressures confirm that the absolute values of Vp and Vs are also higher for PR than for TPT. Furthermore, the relative increases with increasing effective pressure are also significantly different. For PR, the P-wave velocity increases by only about 3%, from an average of around 3.75 km/s at an effective pressure of 5 MPa to around 3.87 km/s at
Fig. 6. Velocity vs. effective pressure during increasing and decreasing effective pressure. Permanent increase of velocity originated from compaction due to pressurization.
70 MPa. Similarly, the S wave velocity increases less than 2%, from an average of around 2.66 km/s at an effective pressure of 5 MPa to around 2.70 km/s at 70 MPa. By contrast, the velocities for TPT show much greater changes, with the P-wave velocity increasing by 17% from an average of around 3.20 km/s at 5 MPa to 3.80 at 70 MPa, and the S wave velocity increasing by 16% from an average of around 2.30 km/s at 5 MPa to an average of around 2.70 at 70 MPa. Overall, we observe higher absolute values of both P and S wave velocities and the much lower velocity increases with increasing effective pressure in PR samples than in samples of TPT. This is entirely as expected, and can be explained by the much higher degree of cementation in PR and the abundance of elastically strong mm-to-cm lava clasts. In contrast, the TPT is relatively poorly cemented and almost entirely lacks elastic clasts in its matrix. These observations are further supported by the velocity hysteresis data shown in Fig. 6. Following cycles of pressurization and de-pressurization, the P-wave velocities for PR exhibit little hysteresis and return to within 2% of their initial values. However, TPT exhibits much greater hysteresis, with the P-wave velocity after the same pressurization/de-pressurization cycle remaining approximately 10% higher than its initial value. Furthermore, only about 20% of the velocity increase during the pressurization part of the cycle is recovered on de-pressurization. Taken together, these observations all imply that some form of inelastic compaction of the weak matrix of TPT occurs during pressurization to 70 MPa, while the stronger PR deforms primarily elastically. The presence of mm-to-cm lava clasts in PR samples, and the rapid degradation of TPT samples with increasing pressure, meant that it was challenging to achieve the steady-state fluid flow conditions necessary for accurate permeability measurements (Fig. 7). Nevertheless, steadystate conditions were achieved for a number of samples, and the changes in permeability with increasing effective pressure are shown in Fig. 8. Unsurprisingly, we found a significant difference in the permeability values for PR and TPT samples. Initial values for PR at an effective pressure of 5 MPa were of the order of ~10− 18 m2, and these values decreased by one order of magnitude as effective pressure was increased to 70 MPa. By contrast, initial values for TPT were three orders of magnitude higher (~ 10− 15 m2), and these values decreased by two orders of magnitude as effective pressure was increased to 70 MPa. These measurements provide independent evidence that the relatively poorly consolidated TPT microstructure is much more permeable than the more lithified PR. It further suggests that the two different mechanisms of emplacement exert a profound influence on the corresponding transport properties. 4.3. Microstructural observations
Fig. 5. P (a) and S (b) wave velocities for PR and TPT vs. increasing effective pressure.
FESEM images of both PR and TPT taken before and after pressurization to 70 MPa are shown in Fig. 8. The image of the unpressurized PR (Fig. 8a) shows the presence of mm-to-cm lava clasts and
S. Vinciguerra et al. / Tectonophysics 471 (2009) 161–169
Fig. 7. Fluid permeability for PR and TPT vs. effective pressure.
sub-spherical voids of the order of a micron. Many of the voids are in-filled with zeolites. After pressurization and de-pressurization some reduction in void space is observed (Fig. 8b), however, the
167
majority of voids remain open, implying that the combination of a strongly-cemented matrix and stiff, sub-spherical voids is sufficient to support the applied stress. Importantly, no significant crack damage is observed, indicating that little inelastic deformation occurs during pressurization. This is entirely consistent with the observed changes in P and S wave velocities, which exhibit only very limited increases during pressurization and low hysteresis after depressurization. Fig. 8c shows that unpressurized TPT has a much more porous matrix, with large voids of mm-to-cm scale. These are thought to be related to the rapid degassing that occurred during emplacement of the pyroclastic flow. After pressurization and de-pressurization, significant reduction of porosity is observed (Fig. 8d). This is associated with pervasive crack damage affecting both the matrix groundmass and single crystals, and provides independent evidence of significant inelastic deformation during pressurization, such as inelastic pore collapse and crack growth. Again, these observations are completely consistent with both the significant increases in P and S wave velocities measured during pressurization and the significant hysteresis observed after de-pressurization.
Fig. 8. Microstructural observation of intact and pressurized (a,b) PR and (c,d) TPT samples.
168
S. Vinciguerra et al. / Tectonophysics 471 (2009) 161–169
5. Seismic velocity determination from in situ sonic logs A sonic log was run in the borehole to measure compressional and shear wave velocities in the shallow crustal layers down to 110 m for P waves (Fig. 2b). We used a slim tool of 43 mm diameter and about 6 m length, with a vertical resolution of 50 cm and a depth of investigation of 20 to 70 cm. The tool was lowered to the bottom of the uncased, water-filled hole, and then raised at a rate of 8 m/min. The transmitters generated a 20 kHz pulse every 10 cm, which propagated through the adjacent formation to the receivers where the full received waveforms were recorded digitally. In general, the units in the upper part of the borehole (i.e. down to 67 m) exhibit a higher and more consistent P-wave velocity, around 3.5 km/s, than the units in the lower part of the borehole, except for the interval between 38 and 49 m depth which exhibits an anomalously low velocity. Between 65 and 83 m P-wave velocity shows a strong variability and an average of 2.6 km/s for “Vallerano lava” and 2.1 km/s. P-wave velocity values of ~2 km/s were measured over the interval between 83 and 100 m depth, corresponding to the “Tufi Terrosi” and “Tufi di Acque Albule” units. An increase to 2.9 km/s is observed for the ‘Tufi del Palatino’ unit in the interval between 100 and 108 m depth. The variations observed in the recorded sonic wave velocities can be correlated with the different volcanic units. However, for this specific study, the units of interest are the Pozzolane Rosse (PR) and the Tufo Pisolitico di Trigoria (TPT) units. These exhibit a very consistent P-wave velocity of 3.5 and 2.9 km/s respectively for Pozzolane Rosse and Tufo del Palatino, the compositionally and texturally most analogous unit to TPT, throughout its depth. 6. Discussion and conclusions When comparing velocities measured in the kHz range from sonic logs with laboratory measurements made in the MHz range, it is necessary to take account of the frequency dependence of the velocities and the differences in volumes investigated. An earlier study on volcanic rocks (Vinciguerra et al., 2006 and references therein) demonstrated that dispersion processes, quantified through the Biot–Gassman equation, resulted in velocities measured at ultrasonic laboratory frequencies that were between 1 and 10% higher than velocities measured at seismic frequencies (i.e. in the Hz range). Overall, we would expect the dispersion effects to be more limited in our study, since we are dealing with only 3 orders of magnitude difference in frequency rather than the 6 orders of magnitudes in earlier studies (Vinciguerra et al., 2006; Zamora et al., 1994). The other major consideration is that it is also necessary to consider the in situ temperature and pressure conditions to which the rocks are subjected at depth in the borehole. Taking all these considerations into account, we find excellent agreement between the P-wave velocities from the sonic logs and the laboratory ultrasonic measurements for the Pozzolane Rosse unit. The mean laboratory value of the P-wave velocity for water-saturated samples of this material is 3.75 km/s at an effective pressure equivalent to that at the depth at which the sonic velocity was measured (i.e. 5 MPa). This is within about 5% of the mean sonic velocity of 3.55 km/s. On the same token the laboratory average of P-wave velocities for TPT (~3.2 km/s) is in good agreement with the sonic logs (~3 km/s) on the compositionally and texturally equivalent units, i.e. Tufo del Palatino. Overall the good match between laboratory velocities made on water-saturated samples and sonic log P-wave velocities suggests that the rock is likely to be water saturated under in situ conditions. This is in agreement with the reported saturation of the units with gas and fluid due to the presence of a shallow aquifer (Carapezza et al., 2003). The effect of increasing effective pressure on wave velocities is strongly influenced by the degree of lithification of the matrix and the presence of mm-to-cm scale lava clasts and crystals. Hence, for the coarser and more lithified PR unit, which responds essentially
elastically to the applied stress, both P and S wave velocities are seen to increase by only a few percent. By contrast, for the TPT unit, where the applied stress induces pore collapse and enhanced crack damage, velocity increases approaching 20% are observed. These differences in microstructure and its evolution under increasing effective pressure have a profound influence on the transport properties, which range from the order of 10− 18 m2 for PR to the order of 10− 15 m2 for TPT. We suggest that PR can act as a low permeability layer, enhancing the confinement of a shallow aquifer in the shallower units. These results are in good agreement with the regional hydrogeologic studies carried on in the volcanic district (Carapezza, 2003 and references therein), that show that this volcanic horizon represents the main regional aquifer within the pyroclastic succession. However the lithification and the pervasive zeolitization of the ash matrix can significantly alter the permeability properties within the volcanic units. Whether this feature is a local peculiarity or it should be considered the typical texture in the proximal volcanic area remains undetermined. The results obtained from laboratory measurements and their comparison with field determinations, such as sonic logs, provide crucial information for the interpretation of the internal structure of the volcanic district, and in turn suggest how applied stress can significantly change the rheology and permeability of the Alban hills tuffs, opening new perspectives for the interpretation of the volcanic district dynamics. It is also crucial to assess whether the observed features of the analysed cores (in particular the high degree of cementation affecting both permeability and P-wave velocities of the PR lithology) should be regarded as representative of the whole central volcanic area, or if they represent the effect of local conditions (e.g. chemical characteristic of the ground water). Future rock properties measurements at the laboratory and at larger scale joint to drilling studies in the volcanic area will allow at clarify this important issue. Acknowledgements This work has been developed within the Italian Dipartimento della Protezione Civile in the frame of the 2004–2006 Agreement with Istituto Nazionale di Geofisica e Vulcanologia (Project DPC V3.1) and FIRB-MIUR Project “Research and Development of New Technologies for Protection and Defense of Territory from Natural Risks”, RU C2 and C3 coordinated by P. Scarlato and P. Montone respectively. References Amato, A., Chiarabba, C., Cocco, M., Di Bona, M., Selvaggi, G., 1994. The 1989–1990 seismic swarm in the Alban Hills volcanic area, central Italy. Journal of Volcanology and Geothermal Research 61, 225–237. Amato, A., Chiarabba, C., 1995. Recent uplift of the Alban Hills volcano (Italy): evidence for a magmatic inflation? Geophysical Research Letters 22, 1985–1988. Benson, P.M., Meredith, P.G., Platzman, E.S., 2003. Relating pore fabric geometry to acoustic and permeability anisotropy in Crab Orchard Sandstone: a laboratory study using magnetic ferrofluid. Geophysical Research Letters 30 (19), 1976. doi:10.1029/2003GL017929. Boni, C., Bono, P., Lombardi, S., Mastrorillo, L., Percolo, C., 1995. Hydrogeology, fuid geochemistry and thermalism. In: Trigila, R. (Ed.), The Volcano of the Alban Hills, Rome, pp. 221–242. Capelli, G., Mazza, R., Giordano, G., Cecili, A., de Rita, D., Salvati, R., 2001. The Colli Albani Volcano (Rome, Italy): breakdown of the equilibrium of a hydrogeological unit as a result of unplanned and uncontrolled over-exploitation. Hydrogeologie 2000 (4), 63–70. Carapezza, M.L., Badalamenti, B., Cavarra, L., Scalzo, A., 2003. Gas hazard assessment in a densely inhabited area of Colli Albani Volcano (Cava dei Selci, Roma). Journal of Volcanology and Geothermal Research 123, 81–94. Chiarabba, C., Malagnini, L., Amato, A., 1994. Tree-dimensional velocity structure and earthquake relocation in the Alban Hills Volcano, central Italy. Bulletin Seismological Society of America 84, 295–306. Chiarabba, C., Amato, A., Delaney, P.T., 1997. Crustal structure, evolution, and volcanic unrest of the Alban Hills, central Italy. Bulletin Volcanolology 59, 161–170. De Rita, D., Funiciello, R., Parlotto, M., 1988. Geological Map of the Colli Albani Volcanic Complex. Progetto Finalizzato Geodinamica C.N.R., Rome, Italy. De Rita, D., Faccenna, C., Funiciello, R., Rosa, C., 1995. Stratigraphy and volcano-tectonics. The Volcano of The Alban Hills. Tipografia Sgs, Roma, pp. 33–71.
S. Vinciguerra et al. / Tectonophysics 471 (2009) 161–169 Di Filippo, M., Toro, B., 1995. Gravity features. Stratigraphy and volcano-tectonics. In: Trigila, R. (Ed.), The Volcano of the Alban Hills, Rome, pp. 213–219. Faccenna, C., Funiciello, R., Mattei, M., 1994. Late Pleistocene N–S shear zones along the Latium Tyrrhenian margin: structural characters and volcanological implications. Bollettino di Geofisica Teorica Applicata 36, 507–522. Feuillet, N., Nostro, C., Chiarabba, C., Cocco, M., 2004. Coupling between earthquake swarms and volcanic unrest at the Alban Hills volcano (Italy) through elastic stress transfer. Journal of Geophysical Research 109 (B02308). doi:10.1029/ 2003JB002419. Freda, C., Gaeta, M., Karner, D.B., Marra, F., Renne, P.R., Taddeucci, J., Scarlato, P., Christensen, J.N., Dallai, L., 2006. Eruptive history and petrologic evolution of the Albano multiple maar (Alban Hills, Central Italy). Bulletin of Volcanology 68, 567–591. Giaccio, B., Sposato, A., Gaeta, M., Marra, F., Palladino, D.M., Taddeucci, J., Barbieri, M., Messina, P., Rolfo, M.F., 2007. Mid-distal occurrences of the Albano Maar pyroclastic deposits and their relevance for reassessing the eruptive scenarios of the most recent activity at the Colli Albani Volcanic District, Central Italy. Quaternary International 171/172, 160–178. doi:10.1016/j.quaint.2006.10.013. Giordano, G., De Benedetti, A.A., Diana, A., Diano, G., Gaudioso, F., Marasco, F., Miceli, M., Mollo, S., Cas, R.A.F., Funiciello, R., 2006. The Colli Albani mafic caldera (Roma, Italy): stratigraphy, structure and petrology. Journal of Volcanology and Geothermal Research. doi:10.1016/j.jvolgeores.2006.02.009. Karner, D.B., Marra, F., Renne, P.R., 2001. The history of the Monti Sabatini and Alban Hills volcanoes: groundwork for assessing volcanic-tectonic hazards for Rome. Journal of Volcanology and Geothermal Research 107, 185–219. Mariucci, M.T., Pierdominici, S., Pizzino, L., Marra, F. and Montone, P., in press. Looking into a volcanic area: an overview on the 350 m scientific drilling at Colli Albani (Rome, Italy), Journal of Volcanology and Geothermal Research. doi:10.1016/j.jvolgeores.2008.04.007. Marra, F., 1999. Low-magnitude earthquakes in Rome: structural interpretation and implications for local stress-field. Geophysical Journal International 138, 231–243. Marra, F., 2001. Strike–slip faulting and block rotation: a possible triggering mechanism for lava flows in the Alban Hills? Journal of Structural Geology 23 (1), 127–141.
169
Marra, F., Freda, C., Scarlato, P., Taddeucci, J., Karner, D.B., Renne, P.R., Gaeta, M., Palladino, D.M., Trigila, R., Cavarretta, G., 2003. 40Ar/39Ar Geochronology of the recent phase of activity of the Alban Hills Volcanic District (Rome, Italy): implications for seismic and volcanic hazards. Bulletin of Volcanology 65, 227–247. Marra, F., Montone, P., Pirro, M., Boschi, E., 2004. Evidence of active tectonics on a Roman Aqueduct System (II–III Century A.D.) near Rome, Italy. Journal of Structural Geology 26, 679–690. Marra, F., Karner, D.B., Freda, C., Gaeta, M., and Renne, P.R., in press. Large mafic eruptions at Alban Hills Volcanic District (Central Italy): chronostratigraphy, petrography and eruptive behavior, Journal of Volcanology and Geothermal Research. Montone, P., Amato, A., Chiarabba, C., Buonasorte, G., Fiordelisi, A., 1995. Evidence of active extension in Quaternary volcanoes of Central Italy from breakout analysis and seismicity. Geophysical Research Letters 22, 1909–1912. Palladino, D.M., Gaeta, M., Marra, F., 2001. A large K-foiditic hydromagmatic eruption from the early activity of the Alban Hills volcanic district, Italy. Bulletin volcanologique 63, 345–359. Pizzino, L., Galli, G., Mancini, C., Quattrocchi, F., Scarlato, P., 2002. Natural Gas Hazard (CO2 , 222Radon) within a quiescent volcanic region and its relations with seismotectonics: the case of the Ciampino–Marino area (Colli Albani volcano, Rome). Natural Hazards 27 (3), 257–287. Salvi, S., Atzori, S., Tolomei, C., Allievi, J., Ferretti, A., Rocca, F., Prati, C., Stramondo, S., Feuillet, N., 2004. Inflation rate of the Colli Albani volcanic complex retrieved by the permanent scatterers SAR interferometry technique. Geophysical Research letters 31 (L12606). doi:10.1029/2004GL020253. Vinciguerra, S., Trovato, C., Meredith, P.G., Benson, P.M., Troise, C., De Natale, G., 2006. Understanding the seismic velocity structure of Campi Flegrei caldera (Italy): from the laboratory to the field scale. Pure Applied Geophysics 163, 2205–2221. Washington, H.S., 1906. The Roman Comagmatic Region. Publ no. 57. Carnegie Institute. Zamora, M., Sartoris, G., Chelini, W., 1994. Laboratory measurements of ultrasonic wave velocities in rocks from the Campi Flegrei volcanic system and their relation to other field data. Journal of Geophysical Research 99 (B7), 13553–13562.
Contents lists available at ScienceDirect
Tectonophysics j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / t e c t o
Physical properties of tuffs from a scientific borehole at Alban hills volcanic district (central Italy) S. Vinciguerra a,⁎, P. Del Gaudio a, M.T. Mariucci a, F. Marra a, P.G. Meredith b, P. Montone a, S. Pierdominici a, P. Scarlato a a b
Istituto Nazionale di Geofisica e Vulcanologia, Sezione di Roma1, I-00143 Rome, Italy Department of Earth Sciences, University College London, WC1E 6BT London, UK
a r t i c l e
i n f o
Article history: Received 24 April 2008 Received in revised form 8 July 2008 Accepted 7 August 2008 Available online 22 August 2008 Keywords: Scientific borehole Volcanic rocks Physical properties
a b s t r a c t Recent seismic swarms and hydrothermal activity suggest that the Quaternary volcanic complex of the Alban Hills may pose a threat to the city of Rome. A 350 m scientific borehole was therefore drilled into this volcanic area to elucidate its inner structure for the first time. Wire-line logs were run in the borehole in order to characterize the physical properties of the rocks and their variations with depth. In particular, a detailed sonic log was run to measure the P-wave velocity from the well-head down to 110 m. To further investigate velocity changes, we carried out laboratory measurements of P and S elastic wave velocities and fluid permeability at effective pressures up to 70 MPa during both increasing and decreasing pressure cycles on selected core samples representative of the main volcanic units. Specifically, we studied samples from two pyroclastic units representative of two classes of volcanic deposits that are representative of the whole succession: (i) a coarse-grained, well-lithified facies (Pozzolane Rosse unit), containing abundant mm-to-cm lava clasts and crystals; and (ii) a fine-grained, matrix-supported pyroclastic deposit (Tufo Pisolitico di Trigoria unit), with rare lithic lava clasts and sparse pumice. Elastic wave velocities reveal significant differences between units and indicate how, within the same lithology, the different degree of lithification and presence of clasts can affect significantly physical property values. The mean laboratory value of the Pwave velocity for Pozzolane Rosse and Tufo Pisolitico di Trigoria units is respectively of 3.75 and 3.2 km/s at an effective pressure equivalent to that at the depth at which the sonic velocity was measured. Under increasing effective pressure a profound influence on the transport properties is observed. Permeability ranges from the order of 10− 18 m2 for the Pozzolane Rosse unit to the order of 10− 15 m2 for the Tufo Pisolitico di Trigoria unit, in good agreement with the shallow aquifer circulating in the shallower units. © 2008 Elsevier B.V. All rights reserved.
1. Introduction The Alban Hills form part of the 200-km-long, Roman Magmatic Province (Washington, 1906), a chain of ultra-potassic, mainly explosive volcanic districts and small eruptive centres that developed along the Tyrrhenian margin of central Italy since the Middle Pleistocene. The regional structural trend follows the main NW–SE oriented extensional faults characterizing the Apennines and the Tyrrhenian Sea margin, and it is crossed by a local N–S, right-lateral fault system developed behind the Olevano–Antrodoco thrust front (Fig. 1a). This latter feature represents a major geodynamic boundary, along which the Umbria–Marche–Sabina pelagic/transitional series of the northern Apennines over thrusts the Latium–Abruzzi carbonate platform of the central Apennines. The Alban Hills are located where
⁎ Corresponding author. Tel.: +39 06 51860478; fax: +39 06 51860507. E-mail address: [email protected] (S. Vinciguerra). 0040-1951/$ – see front matter © 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.tecto.2008.08.010
recognized N–S strike–slip structural trends meet the NW–SE extensional ones. The present-day stress field in the area is not well constrained, and its interpretation is the subject of current debate (e.g. Faccenna et al., 1994; Montone et al., 1995; Marra, 1999). In particular, it is not clear whether a still-standing extensional regime characterized by a vertical σ1 and a NE-oriented σ3 affects this area, or a substantial similarity of σ1–σ2–σ3 causes the frequent permutation of the stress tensor, allowing for repeated superimposition of two competitive tectonic regimes (Marra, 2001). Enhanced deformation (Amato and Chiarabba, 1995; Salvi et al., 2004) and seismic activity (Amato et al., 1994; Feuillet et al., 2004) have both been reported in the area of the most recent volcanic activity. In particular, instrumental records of the 1989–1990 seismic swarm, that struck the western flank of the volcano, suggest that the main faults are active under a NE extensional regime (Fig. 1b). In contrast, both recent and historical deformations of the surrounding Acque Albule travertine basin (Fig. 1a) have been interpreted as the effect of strike–slip tectonics under a NE trending σ1 (Faccenna et al.,
162
S. Vinciguerra et al. / Tectonophysics 471 (2009) 161–169
S. Vinciguerra et al. / Tectonophysics 471 (2009) 161–169
1994; Marra et al., 2004). Furthermore, two main fracture systems, SW and NW dipping, have been recognized from recent down-hole measurements in the volcanic units close to the Tuscolano–Artemisio caldera rim (Mariucci et al., in press), and interpreted as occurring from the superimposition of the two tectonic styles in the area. However, analysis of borehole breakouts defined an active stress field with an approximately E–W oriented minimum horizontal component, which may be interpreted as evidence of strain-partitioning along the local N–S structures due to their re-activation under a NEstriking extensional regime (Mariucci et al., in press). The scientific community has recently increased its efforts to understand the internal structure of the volcano through tomographic studies based on inversion of seismicity from local earthquakes (Feuillet et al., 2004; Chiarabba et al., 1994). While high velocity zones have been interpreted as either solidified magma reservoirs or uplifted basement, the observed low velocity anomalies on the western flank suggest either the presence of molten material or fluid overpressure (Chiarabba et al., 1997). The documented unrest episodes of enhanced seismic activity and deformation are also accompanied by anomalous gas emissions (CO2, 222Radon, H2S) (Pizzino et al., 2002; Carapezza et al., 2003) controlled by the interplay between a shallow aquifer (less than 1 km of depth), hosted within the volcanic pile, and a reservoir buried in the carbonate basement, separated by low permeability Plio-Quaternary deposits (Boni et al., 1995; Di Filippo and Toro, 1995; Capelli et al., 2001). As a consequence, unraveling the velocity structure and the conditions controlling fluid flow within the volcanic complex, and their evolution with depth, have been primary drivers of scientific investigation. Hence, a 350 m deep scientific borehole was recently drilled in the western sector of the volcanic complex, allowing both wire-line logging and continuous core recovery to be carried out. Here we present results from a joint study in which (i) the results of laboratory measurements of porosity, P and S elastic wave velocities and fluid permeability on recovered cores under simulated in situ conditions are integrated with (ii) in situ measurements of P and S wave velocities from down-hole sonic logs. This integration of multi-scale measurements of physical properties allows us to constrain the velocity structure of the volcanic complex and highlight the physical mechanisms controlling the fluid circulation within the different volcanic lithologies. 2. Volcanic stratigraphy The volcanic history of Alban Hills may be roughly divided into three main phases, marked by different eruptive mechanisms and magma volumes (De Rita et al., 1988, 1995; Giordano et al., 2006). The early Tuscolano–Artemisio Phase (c. 561–366 ka; Karner et al., 2001; Marra et al., in press) was the most explosive and voluminous, and is characterized by five large pyroclastic-flow forming eruptions with volumes of the order of tens of km3 and minor effusive activity. Each one of these cycles ended with caldera collapse followed by intra- and peri-caldera effusive and strombolian activity. The last cycle originated the so-called Tuscolano–Artemisio caldera and its ring of scoria cones in the period c. 366–351 ka. After a c. 40 kyr-long dormancy, the less energetic Faete Phase of activity occurred (c. 308–250 ka; Marra et al., 2003). This phase started with peripheral effusive eruptions coupled with subordinate hydromagmatic activity, and encompassed the formation of the Faete central edifice and several minor scoria cones within the older intracaldera area. A further c. 50 kyr-long dormancy then preceded the start of the Late Hydromagmatic Phase (c. 200–36 ka; Marra et al., 2003; Freda et al., 2006; Giaccio et al., 2007), which was dominated by pyroclasticsurge eruptions with the formation of several monogenetic or multiple maars and/or tuff rings clustered southwest of the Mt. Faete edifice.
163
The products of the Tuscolano–Artemisio phase represent by far the largest portion of the erupted deposits and constitute the greater portion of the volcanic substrate of the area of Rome. These deposits may be grouped into two main classes, based on their petrographic characteristics and their eruptive mechanism (Marra et al., in press). The earlier eruptive units (Tufo Pisolitico di Trigoria, 561 ±1 ka; Tufo del Palatino, 528 ±1 ka; Tufo di Acque Albule, 526 ±1 ka; Marra et al., in press) are the result of hydromagmatic eruptions and are characterized by prevalently fine-grained, pyroclastic ash-flow deposits. The later units (Pozzolane Rosse, 456±3 ka; Pozzolane Nere, 407±2 ka; Villa Senni, 365 ±4 ka; (Marra et al., in press) are the result of dry eruptions and are characterized by coarse-grained, pyroclastic scoria and ash-flow deposits. 3. The drilling project As reported above, a 350 m scientific borehole was drilled in the western sector of the Alban Hills area (Fig. 1a), where seismic swarms (Fig. 1b), enhanced ground deformation and high gas concentration in water had been observed. The aim was to elucidate the internal structure and determine the stress field by means of down-hole measurements (see Mariucci et al., in press). From a tectonic point of view, data from the borehole allowed two main fracture systems (dipping SW and NW) to be recognized, indicating a prevalent strike–slip component of fault movement. By contrast borehole breakout analysis defined an active stress field with an approximately E–W oriented minimum horizontal component (Fig. 1a). Geochemical analysis of deep fluids from the borehole revealed water of meteoric origin which had a long residence time in the aquifer, and gas dominated by CO2 content. The gas exhibited one of the highest magmatic components from the Alban Hills, which was partly deriving from a magmatic source and partly from the thermal decarbonization of carbon-rich basement rocks (Mariucci et al., in press). The borehole was drilled using the wire-line coring technique that allowed us to retrieve a complete core record and hence establish a detailed stratigraphy (Fig. 2a). Wire-line logs (Fig. 2b) were also run over selected intervals in order (i) to characterize the in situ physical properties of the rocks (acoustic wave velocities, electrical resistivity, natural gamma radiation, magnetic susceptibility) and their variations with depth, and (ii) to define the structural setting and assess the present-day stress field (by borehole televiewer and caliper log). All the down-hole logs were run by the ICDP Operational Support Group (GFZPotsdam), and the data was analysed using WellCad software (ALT, Luxembourg). A more detailed description can be found in Mariucci et al. (in press). The detailed borehole stratigraphy allowed us to correlate the geophysical logs with specific rock features (Fig. 2a,b). The main lithology cored was the pyroclastic-flow deposits (tuffs) of the Tuscolano–Artemisio explosive phase of activity (561–365 ka), which exhibited a wide variability in grain-size and cohesion. We have chosen two units for further study (Pozzolane Rosse and Tufo Pisolitico di Trigoria), which are representative of the two main lithologic classes of erupted products described earlier. For clarity, we will simplify the names and use PR for the Pozzolane Rosse unit and TPT for the Tufo Pisolitico di Trigoria units. In outcrop, the PR deposit is described as a single, massive, poorlysorted pyroclastic-flow deposit. It is characterized by a coarse-ash matrix, poorly consolidated or locally lithified due to vapour-phase zeolite crystallisation (De Rita et al., 1995). It includes reddish-purple to dark grey, poorly to moderately vesicular, scoria lapilli and blocks, abundant lava and thermally metamorphosed sedimentary lithic lapilli and blocks, and scarce granular (leucite+ clinopyroxene) inclusions. In contrast to what is commonly observed in outcrops, in the recovered cores the PR displays a very well-lithified facies, resulting from a pervasive
Fig. 1. Overview of the geological and geophysical characters of the area. a) Regional tectonic structures and stress field (modified after Marra, 1999). b) Geological sketch of volcanic district with CA1 borehole location and seismicity (modified after Mariucci et al. in press): 1) Holocene alluvial deposits; 2) Plio-Quaternary sedimentary deposits; 3) “Hydromagmatic phase” (200– 36 ky); 4) deposits of “Faete phase” and related lava flows (308–250 ky); 5) “Tuscolano–Artemisio phase” and related lava flows (561–351 ky); 6) crater rims; 7) recent seismic swarms (after Feuillet et al., 2004).
164
S. Vinciguerra et al. / Tectonophysics 471 (2009) 161–169
Fig. 2. a) CA1 borehole stratigraphy of the volcanic units, belonging to the Tuscolano–Artemisio Phase (ca. 561–351 ky). Legend: P — “Pozzolanelle” (366 ± 4 Ka); PN — “Pozzolane Nere” (407 ± 4 Ka); FL — “Fioranello Lava”; PR — “Pozzolane Rosse” (457 ± 4 Ka); VL — “Vallerano Lava” (460 ± 4 Ka); AFS — “Air Fall Sequence” (488 ± 2 Ka); TAA — “Tufo di Acque Albule” (526 ± 1 Ka); TP — “Tufo del Palatino” (528 ± 1 Ka); TPT — “Tufo Pisolitico di Trigoria” (561 ± 1 Ka). Small grey intervals (not to scale) are paleosols. Ages are after Karner et al. (2001). b) P-wave velocity from down-hole sonic log.
zeolitization of the ash matrix. Abundant mm-to-cm lava clasts and mmsized pyroxene and biotite crystals are embedded within the coarse-ash matrix. The observed strong cementation that occurs in the intermediate, scoriaceous volcanic deposits (Pozzolane Rosse) is the likely result of secondary process of transformation of primary volcanic particles into authigenic minerals, due to the chemical action of the ground water. The outcropping deposits of the TPT occur as a light gray, indurate, matrix-supported, stratified coarse-ash deposit (Palladino et al., 2001). It includes altered leucite and abundant mm-to-cm-sized, poorly vesiculated gray scoria lapilli, and subordinate lava, carbonate and tuff lithics. Up to cm-sized accretionary lapilli (pisoliti) are frequent. In the recovered cores, the TPT occurs as a fine-grained, matrix-supported, lithified pyroclastic deposit. It contains rare mm-sized lithic lava clasts, sparse pumice and carbonatic clasts, and accretionary lapilli. 4. Laboratory petrophysical investigation 4.1. Methodology All measurements were carried out on cylindrical sub-core samples 38 mm in diameter by 40 mm long, taken from the recovered
borehole cores. Initial porosities were determined gravimetrically by weighing samples of known volume both dry and water saturated. Initial elastic wave velocity measurements were made on dry samples under ambient laboratory conditions. Measurements were made radially at azimuthal steps of 10°. In order to understand the variation of rock properties with increasing effective pressure, we then carried out simultaneous measurements of ultrasonic P and S wave velocities and fluid permeability in a servo-controlled, steady-state-flow permeameter. The permeameter comprises a hydrostatic pressure vessel, using silicone oil as the pressure medium, equipped with two 70 MPa servocontrolled fluid pressure intensifiers (volumometers) to provide pore fluid pressure independently to each end of the sample and piezoelectric ultrasonic transducers for P and S wave velocity measurements (Fig. 3, plus see also Benson et al., 2003, for further details). Rubber-jacketed samples are placed inside the pressure vessel, located between two steel end-caps. Measurements were made at effective pressures from 5 to 80 MPa, during both increasing and decreasing pressure cycles. The effective pressure was increased in steps of 5 MPa from 5 MPa to 40 MPa and then in steps of 10 MPa from 40 MPa to 70 MPa.
S. Vinciguerra et al. / Tectonophysics 471 (2009) 161–169
Fig. 3. Schematic sketch of the servo-controlled permeameter equipped for Vp, Vs measurements, sample set up and pressure vessel.
A 900 V pulse generator was used to excite a 1 MHz resonant frequency piezoelectric transmitting transducers for measurements of both P and S wave velocities. Waveforms were captured using an identical receiving transducer, and were pre-amplified by 40 dB before being recorded and displayed on a digital storage oscilloscope. Accuracy is estimated as 1% for P waves and 2% for S waves. In order to measure fluid permeability, a fixed pore fluid pressure gradient is set across the sample and the steady-state volume flow rate recorded as a function of time. The permeability can then be determined directly from the measured volume flow rate and the sample dimensions using Darcy's law. Microstructural analysis was carried out on both as-received and experimentally pressurized samples using a high-resolution thermal field emission scanning electron microscope (FESEM) in order to study changes in the microstructure due to pore collapse or crack damage induced during the pressurization and de-pressurization cycles. The FESEM generates images with much lower electrostatic
Table 1 Porosity, P and S wave velocities at room pressure and dry conditions. Sample
Porosity (%)
Vp (m/s)
Vs (m/s)
PR1 PR2 PR3 PR4 PR5 PR6 PR7 PR8 TPT1 TPT2 TPT3 TPT4 TPT5 TPT6
31.12 11.00 21.60 10.80 11.00 17.00 10.20 19.40 14.10 31.00 15.00 17.20 13.80 15.10
3601 3732 3610 3715 4130 3836 3728 3702 3226 3175 2930 3415 3159 3453
2590 2590 2604 2550 3102 2833 2703 2856 2345 1960 2355 2465 2789 2466
Fig. 4. Radial Vp for (a) PR and (b) TPT.
165
166
S. Vinciguerra et al. / Tectonophysics 471 (2009) 161–169
distortion than a conventional SEM, and with a spatial resolution that is 3 to 6 times higher (i.e. less than 2 nm). 4.2. Results The lithologies investigated exhibited initial porosities ranging from around 10% to over 30% (Table 1), with an average porosity of 14.2% and 17.8% for PR and TPT respectively. Axial P and S wave velocities for dry samples under ambient conditions range from 3.6 to 4.1 km/s and 2.5 to 3.1 km/s, respectively, for PR. P and S wave velocities are found to be lower for the TPT samples at 2.9 to 3.4 km/s and 1.9 to 2.7 km/s, respectively (Table 1). Radial P-wave velocity measurements (Fig. 4a) show that PR exhibits a considerable velocity anisotropy of up to about 25%, while the Tufo Pisolitico (Fig. 4b) exhibits a very much lower velocity anisotropy of only about 3%. We consider that the differences in mean porosity and axial wave velocities are due to the much more developed lithification of the PR with respect to the relatively poorly consolidated pyroclastic deposits that make up the TPT. The abundance of mm-to-cm size lava clasts in the PR, which are almost entirely absent in the TPT matrix, likely determines both the higher porosity variability and the higher degree of anisotropy observed in the PR samples. This suggests that even within the same tuff lithology, such variations can significantly affect rock physical properties. Measurements of P and S wave velocities (Fig. 5a,b) carried out on water-saturated samples (wet) under elevated effective pressures confirm that the absolute values of Vp and Vs are also higher for PR than for TPT. Furthermore, the relative increases with increasing effective pressure are also significantly different. For PR, the P-wave velocity increases by only about 3%, from an average of around 3.75 km/s at an effective pressure of 5 MPa to around 3.87 km/s at
Fig. 6. Velocity vs. effective pressure during increasing and decreasing effective pressure. Permanent increase of velocity originated from compaction due to pressurization.
70 MPa. Similarly, the S wave velocity increases less than 2%, from an average of around 2.66 km/s at an effective pressure of 5 MPa to around 2.70 km/s at 70 MPa. By contrast, the velocities for TPT show much greater changes, with the P-wave velocity increasing by 17% from an average of around 3.20 km/s at 5 MPa to 3.80 at 70 MPa, and the S wave velocity increasing by 16% from an average of around 2.30 km/s at 5 MPa to an average of around 2.70 at 70 MPa. Overall, we observe higher absolute values of both P and S wave velocities and the much lower velocity increases with increasing effective pressure in PR samples than in samples of TPT. This is entirely as expected, and can be explained by the much higher degree of cementation in PR and the abundance of elastically strong mm-to-cm lava clasts. In contrast, the TPT is relatively poorly cemented and almost entirely lacks elastic clasts in its matrix. These observations are further supported by the velocity hysteresis data shown in Fig. 6. Following cycles of pressurization and de-pressurization, the P-wave velocities for PR exhibit little hysteresis and return to within 2% of their initial values. However, TPT exhibits much greater hysteresis, with the P-wave velocity after the same pressurization/de-pressurization cycle remaining approximately 10% higher than its initial value. Furthermore, only about 20% of the velocity increase during the pressurization part of the cycle is recovered on de-pressurization. Taken together, these observations all imply that some form of inelastic compaction of the weak matrix of TPT occurs during pressurization to 70 MPa, while the stronger PR deforms primarily elastically. The presence of mm-to-cm lava clasts in PR samples, and the rapid degradation of TPT samples with increasing pressure, meant that it was challenging to achieve the steady-state fluid flow conditions necessary for accurate permeability measurements (Fig. 7). Nevertheless, steadystate conditions were achieved for a number of samples, and the changes in permeability with increasing effective pressure are shown in Fig. 8. Unsurprisingly, we found a significant difference in the permeability values for PR and TPT samples. Initial values for PR at an effective pressure of 5 MPa were of the order of ~10− 18 m2, and these values decreased by one order of magnitude as effective pressure was increased to 70 MPa. By contrast, initial values for TPT were three orders of magnitude higher (~ 10− 15 m2), and these values decreased by two orders of magnitude as effective pressure was increased to 70 MPa. These measurements provide independent evidence that the relatively poorly consolidated TPT microstructure is much more permeable than the more lithified PR. It further suggests that the two different mechanisms of emplacement exert a profound influence on the corresponding transport properties. 4.3. Microstructural observations
Fig. 5. P (a) and S (b) wave velocities for PR and TPT vs. increasing effective pressure.
FESEM images of both PR and TPT taken before and after pressurization to 70 MPa are shown in Fig. 8. The image of the unpressurized PR (Fig. 8a) shows the presence of mm-to-cm lava clasts and
S. Vinciguerra et al. / Tectonophysics 471 (2009) 161–169
Fig. 7. Fluid permeability for PR and TPT vs. effective pressure.
sub-spherical voids of the order of a micron. Many of the voids are in-filled with zeolites. After pressurization and de-pressurization some reduction in void space is observed (Fig. 8b), however, the
167
majority of voids remain open, implying that the combination of a strongly-cemented matrix and stiff, sub-spherical voids is sufficient to support the applied stress. Importantly, no significant crack damage is observed, indicating that little inelastic deformation occurs during pressurization. This is entirely consistent with the observed changes in P and S wave velocities, which exhibit only very limited increases during pressurization and low hysteresis after depressurization. Fig. 8c shows that unpressurized TPT has a much more porous matrix, with large voids of mm-to-cm scale. These are thought to be related to the rapid degassing that occurred during emplacement of the pyroclastic flow. After pressurization and de-pressurization, significant reduction of porosity is observed (Fig. 8d). This is associated with pervasive crack damage affecting both the matrix groundmass and single crystals, and provides independent evidence of significant inelastic deformation during pressurization, such as inelastic pore collapse and crack growth. Again, these observations are completely consistent with both the significant increases in P and S wave velocities measured during pressurization and the significant hysteresis observed after de-pressurization.
Fig. 8. Microstructural observation of intact and pressurized (a,b) PR and (c,d) TPT samples.
168
S. Vinciguerra et al. / Tectonophysics 471 (2009) 161–169
5. Seismic velocity determination from in situ sonic logs A sonic log was run in the borehole to measure compressional and shear wave velocities in the shallow crustal layers down to 110 m for P waves (Fig. 2b). We used a slim tool of 43 mm diameter and about 6 m length, with a vertical resolution of 50 cm and a depth of investigation of 20 to 70 cm. The tool was lowered to the bottom of the uncased, water-filled hole, and then raised at a rate of 8 m/min. The transmitters generated a 20 kHz pulse every 10 cm, which propagated through the adjacent formation to the receivers where the full received waveforms were recorded digitally. In general, the units in the upper part of the borehole (i.e. down to 67 m) exhibit a higher and more consistent P-wave velocity, around 3.5 km/s, than the units in the lower part of the borehole, except for the interval between 38 and 49 m depth which exhibits an anomalously low velocity. Between 65 and 83 m P-wave velocity shows a strong variability and an average of 2.6 km/s for “Vallerano lava” and 2.1 km/s. P-wave velocity values of ~2 km/s were measured over the interval between 83 and 100 m depth, corresponding to the “Tufi Terrosi” and “Tufi di Acque Albule” units. An increase to 2.9 km/s is observed for the ‘Tufi del Palatino’ unit in the interval between 100 and 108 m depth. The variations observed in the recorded sonic wave velocities can be correlated with the different volcanic units. However, for this specific study, the units of interest are the Pozzolane Rosse (PR) and the Tufo Pisolitico di Trigoria (TPT) units. These exhibit a very consistent P-wave velocity of 3.5 and 2.9 km/s respectively for Pozzolane Rosse and Tufo del Palatino, the compositionally and texturally most analogous unit to TPT, throughout its depth. 6. Discussion and conclusions When comparing velocities measured in the kHz range from sonic logs with laboratory measurements made in the MHz range, it is necessary to take account of the frequency dependence of the velocities and the differences in volumes investigated. An earlier study on volcanic rocks (Vinciguerra et al., 2006 and references therein) demonstrated that dispersion processes, quantified through the Biot–Gassman equation, resulted in velocities measured at ultrasonic laboratory frequencies that were between 1 and 10% higher than velocities measured at seismic frequencies (i.e. in the Hz range). Overall, we would expect the dispersion effects to be more limited in our study, since we are dealing with only 3 orders of magnitude difference in frequency rather than the 6 orders of magnitudes in earlier studies (Vinciguerra et al., 2006; Zamora et al., 1994). The other major consideration is that it is also necessary to consider the in situ temperature and pressure conditions to which the rocks are subjected at depth in the borehole. Taking all these considerations into account, we find excellent agreement between the P-wave velocities from the sonic logs and the laboratory ultrasonic measurements for the Pozzolane Rosse unit. The mean laboratory value of the P-wave velocity for water-saturated samples of this material is 3.75 km/s at an effective pressure equivalent to that at the depth at which the sonic velocity was measured (i.e. 5 MPa). This is within about 5% of the mean sonic velocity of 3.55 km/s. On the same token the laboratory average of P-wave velocities for TPT (~3.2 km/s) is in good agreement with the sonic logs (~3 km/s) on the compositionally and texturally equivalent units, i.e. Tufo del Palatino. Overall the good match between laboratory velocities made on water-saturated samples and sonic log P-wave velocities suggests that the rock is likely to be water saturated under in situ conditions. This is in agreement with the reported saturation of the units with gas and fluid due to the presence of a shallow aquifer (Carapezza et al., 2003). The effect of increasing effective pressure on wave velocities is strongly influenced by the degree of lithification of the matrix and the presence of mm-to-cm scale lava clasts and crystals. Hence, for the coarser and more lithified PR unit, which responds essentially
elastically to the applied stress, both P and S wave velocities are seen to increase by only a few percent. By contrast, for the TPT unit, where the applied stress induces pore collapse and enhanced crack damage, velocity increases approaching 20% are observed. These differences in microstructure and its evolution under increasing effective pressure have a profound influence on the transport properties, which range from the order of 10− 18 m2 for PR to the order of 10− 15 m2 for TPT. We suggest that PR can act as a low permeability layer, enhancing the confinement of a shallow aquifer in the shallower units. These results are in good agreement with the regional hydrogeologic studies carried on in the volcanic district (Carapezza, 2003 and references therein), that show that this volcanic horizon represents the main regional aquifer within the pyroclastic succession. However the lithification and the pervasive zeolitization of the ash matrix can significantly alter the permeability properties within the volcanic units. Whether this feature is a local peculiarity or it should be considered the typical texture in the proximal volcanic area remains undetermined. The results obtained from laboratory measurements and their comparison with field determinations, such as sonic logs, provide crucial information for the interpretation of the internal structure of the volcanic district, and in turn suggest how applied stress can significantly change the rheology and permeability of the Alban hills tuffs, opening new perspectives for the interpretation of the volcanic district dynamics. It is also crucial to assess whether the observed features of the analysed cores (in particular the high degree of cementation affecting both permeability and P-wave velocities of the PR lithology) should be regarded as representative of the whole central volcanic area, or if they represent the effect of local conditions (e.g. chemical characteristic of the ground water). Future rock properties measurements at the laboratory and at larger scale joint to drilling studies in the volcanic area will allow at clarify this important issue. Acknowledgements This work has been developed within the Italian Dipartimento della Protezione Civile in the frame of the 2004–2006 Agreement with Istituto Nazionale di Geofisica e Vulcanologia (Project DPC V3.1) and FIRB-MIUR Project “Research and Development of New Technologies for Protection and Defense of Territory from Natural Risks”, RU C2 and C3 coordinated by P. Scarlato and P. Montone respectively. References Amato, A., Chiarabba, C., Cocco, M., Di Bona, M., Selvaggi, G., 1994. The 1989–1990 seismic swarm in the Alban Hills volcanic area, central Italy. Journal of Volcanology and Geothermal Research 61, 225–237. Amato, A., Chiarabba, C., 1995. Recent uplift of the Alban Hills volcano (Italy): evidence for a magmatic inflation? Geophysical Research Letters 22, 1985–1988. Benson, P.M., Meredith, P.G., Platzman, E.S., 2003. Relating pore fabric geometry to acoustic and permeability anisotropy in Crab Orchard Sandstone: a laboratory study using magnetic ferrofluid. Geophysical Research Letters 30 (19), 1976. doi:10.1029/2003GL017929. Boni, C., Bono, P., Lombardi, S., Mastrorillo, L., Percolo, C., 1995. Hydrogeology, fuid geochemistry and thermalism. In: Trigila, R. (Ed.), The Volcano of the Alban Hills, Rome, pp. 221–242. Capelli, G., Mazza, R., Giordano, G., Cecili, A., de Rita, D., Salvati, R., 2001. The Colli Albani Volcano (Rome, Italy): breakdown of the equilibrium of a hydrogeological unit as a result of unplanned and uncontrolled over-exploitation. Hydrogeologie 2000 (4), 63–70. Carapezza, M.L., Badalamenti, B., Cavarra, L., Scalzo, A., 2003. Gas hazard assessment in a densely inhabited area of Colli Albani Volcano (Cava dei Selci, Roma). Journal of Volcanology and Geothermal Research 123, 81–94. Chiarabba, C., Malagnini, L., Amato, A., 1994. Tree-dimensional velocity structure and earthquake relocation in the Alban Hills Volcano, central Italy. Bulletin Seismological Society of America 84, 295–306. Chiarabba, C., Amato, A., Delaney, P.T., 1997. Crustal structure, evolution, and volcanic unrest of the Alban Hills, central Italy. Bulletin Volcanolology 59, 161–170. De Rita, D., Funiciello, R., Parlotto, M., 1988. Geological Map of the Colli Albani Volcanic Complex. Progetto Finalizzato Geodinamica C.N.R., Rome, Italy. De Rita, D., Faccenna, C., Funiciello, R., Rosa, C., 1995. Stratigraphy and volcano-tectonics. The Volcano of The Alban Hills. Tipografia Sgs, Roma, pp. 33–71.
S. Vinciguerra et al. / Tectonophysics 471 (2009) 161–169 Di Filippo, M., Toro, B., 1995. Gravity features. Stratigraphy and volcano-tectonics. In: Trigila, R. (Ed.), The Volcano of the Alban Hills, Rome, pp. 213–219. Faccenna, C., Funiciello, R., Mattei, M., 1994. Late Pleistocene N–S shear zones along the Latium Tyrrhenian margin: structural characters and volcanological implications. Bollettino di Geofisica Teorica Applicata 36, 507–522. Feuillet, N., Nostro, C., Chiarabba, C., Cocco, M., 2004. Coupling between earthquake swarms and volcanic unrest at the Alban Hills volcano (Italy) through elastic stress transfer. Journal of Geophysical Research 109 (B02308). doi:10.1029/ 2003JB002419. Freda, C., Gaeta, M., Karner, D.B., Marra, F., Renne, P.R., Taddeucci, J., Scarlato, P., Christensen, J.N., Dallai, L., 2006. Eruptive history and petrologic evolution of the Albano multiple maar (Alban Hills, Central Italy). Bulletin of Volcanology 68, 567–591. Giaccio, B., Sposato, A., Gaeta, M., Marra, F., Palladino, D.M., Taddeucci, J., Barbieri, M., Messina, P., Rolfo, M.F., 2007. Mid-distal occurrences of the Albano Maar pyroclastic deposits and their relevance for reassessing the eruptive scenarios of the most recent activity at the Colli Albani Volcanic District, Central Italy. Quaternary International 171/172, 160–178. doi:10.1016/j.quaint.2006.10.013. Giordano, G., De Benedetti, A.A., Diana, A., Diano, G., Gaudioso, F., Marasco, F., Miceli, M., Mollo, S., Cas, R.A.F., Funiciello, R., 2006. The Colli Albani mafic caldera (Roma, Italy): stratigraphy, structure and petrology. Journal of Volcanology and Geothermal Research. doi:10.1016/j.jvolgeores.2006.02.009. Karner, D.B., Marra, F., Renne, P.R., 2001. The history of the Monti Sabatini and Alban Hills volcanoes: groundwork for assessing volcanic-tectonic hazards for Rome. Journal of Volcanology and Geothermal Research 107, 185–219. Mariucci, M.T., Pierdominici, S., Pizzino, L., Marra, F. and Montone, P., in press. Looking into a volcanic area: an overview on the 350 m scientific drilling at Colli Albani (Rome, Italy), Journal of Volcanology and Geothermal Research. doi:10.1016/j.jvolgeores.2008.04.007. Marra, F., 1999. Low-magnitude earthquakes in Rome: structural interpretation and implications for local stress-field. Geophysical Journal International 138, 231–243. Marra, F., 2001. Strike–slip faulting and block rotation: a possible triggering mechanism for lava flows in the Alban Hills? Journal of Structural Geology 23 (1), 127–141.
169
Marra, F., Freda, C., Scarlato, P., Taddeucci, J., Karner, D.B., Renne, P.R., Gaeta, M., Palladino, D.M., Trigila, R., Cavarretta, G., 2003. 40Ar/39Ar Geochronology of the recent phase of activity of the Alban Hills Volcanic District (Rome, Italy): implications for seismic and volcanic hazards. Bulletin of Volcanology 65, 227–247. Marra, F., Montone, P., Pirro, M., Boschi, E., 2004. Evidence of active tectonics on a Roman Aqueduct System (II–III Century A.D.) near Rome, Italy. Journal of Structural Geology 26, 679–690. Marra, F., Karner, D.B., Freda, C., Gaeta, M., and Renne, P.R., in press. Large mafic eruptions at Alban Hills Volcanic District (Central Italy): chronostratigraphy, petrography and eruptive behavior, Journal of Volcanology and Geothermal Research. Montone, P., Amato, A., Chiarabba, C., Buonasorte, G., Fiordelisi, A., 1995. Evidence of active extension in Quaternary volcanoes of Central Italy from breakout analysis and seismicity. Geophysical Research Letters 22, 1909–1912. Palladino, D.M., Gaeta, M., Marra, F., 2001. A large K-foiditic hydromagmatic eruption from the early activity of the Alban Hills volcanic district, Italy. Bulletin volcanologique 63, 345–359. Pizzino, L., Galli, G., Mancini, C., Quattrocchi, F., Scarlato, P., 2002. Natural Gas Hazard (CO2 , 222Radon) within a quiescent volcanic region and its relations with seismotectonics: the case of the Ciampino–Marino area (Colli Albani volcano, Rome). Natural Hazards 27 (3), 257–287. Salvi, S., Atzori, S., Tolomei, C., Allievi, J., Ferretti, A., Rocca, F., Prati, C., Stramondo, S., Feuillet, N., 2004. Inflation rate of the Colli Albani volcanic complex retrieved by the permanent scatterers SAR interferometry technique. Geophysical Research letters 31 (L12606). doi:10.1029/2004GL020253. Vinciguerra, S., Trovato, C., Meredith, P.G., Benson, P.M., Troise, C., De Natale, G., 2006. Understanding the seismic velocity structure of Campi Flegrei caldera (Italy): from the laboratory to the field scale. Pure Applied Geophysics 163, 2205–2221. Washington, H.S., 1906. The Roman Comagmatic Region. Publ no. 57. Carnegie Institute. Zamora, M., Sartoris, G., Chelini, W., 1994. Laboratory measurements of ultrasonic wave velocities in rocks from the Campi Flegrei volcanic system and their relation to other field data. Journal of Geophysical Research 99 (B7), 13553–13562.