The Somma-Vesuvius Magma Chamber: a Petrological and Volcanological Approach F. BARBERI H. BIZOUARD R. CLOCCHIATTI N. METRICH R. SANTACROCE A. SBRANA [stituto di Mineralogia e Petrografia, University of Pisa, Italy. Laboratoire de Petrographie et Volcanologie, E.R. 45 CNRS, Universit~ Paris XI, Orsay, France. Istituto di Mineralogia e Petrografia, University of Pisa, Italy. AGIP S.p.A., Servizio Esplorazione Geotermica, S. Donato Milanese, Italy. ABSTRACT The volcanic history of Somma-Vesuvius indicates that salic products compatible with an origin by fractionation within a shallow magma chamber have been repeatedly erupted (~(Plinian~ pumice deposits). The last two of these eruptions, (79 A.D. and 3500 B.P.) were carefully studied. Interaction with calcareous country rocks had limited imp0~ance, and all data indicate that differentiated magmas were produced by crystal-liquid fractionation within the undersaturated part of petrogeny's residua system at about 1 kb water pressure. The solid-liquid trend indicates that the derivative magmas originated by fractionation of slightly but signfficantly different parental liquids. Some lavas of appropriate composition were selected as parental liquids to compute the entity of the fractionatien. Results suggest that in both bases a fractionation of about 70 weight % was needed to produce liquids with the composition of the pumice. The combination of all data indicates that the two Plinian eruptions were fed by a magma chamber (3-4 km deep) having a volume of approx. 2.0-2.5 km3. The temperature of the magma that initially entered the chamber was about ll00°C, whereas the temperature of the residual liquids erupted was Plinian pumice was 800 ° and 850°C respectively. There is no evidence that such a magma chamber existed at Vesuvius after the 79 A.D. eruption. These results have relevant practical implications for volcanic hazard and monitoring and for geothermal energy. Bull. Volcanol., Vol. 44-3, 198I INTRODUCTION Somma-Vesuvius volcano is character- ized by a dominant association of silica undersaturated and potassic basic products, ranging in composition from tephrites to leucitites, that were erupted during long periods of prevalently effusive activity with only minor associated pyro- clastic rocks. This relatively monotonous sequence of mostly basic lavas was repeti- tively interrupted by highly explosive eruptions of Plinian type, that produced pumice-fall deposits, very often associated with pyroclastic surges and flows. Seven eruptions of this kind occorred in the last 17.000 years (DELIBRIAS et al., 1979), the most famous of which being that of 79 A.D. which destroyed Pompei, Hercula- neum and Stabiae. These Plinian eruptions correspond almost always (6 times in 7 cases) to the beginning of a new eruptive cycle, follow- ing a long period of quiescence, as indicat- ed by the presence of paleosofls underly- ing the pumice deposits. Furthermore Plinian pumice is characterized by highly evolved (mostly phonolitic) composition and correspond to the most salic magmas erupted by the volcano. Because of this highly evolved nature and the long period of quiescence that preceded their erup-
Transcript
The Somma-Vesuvius Magma Chamber: a Petrological and Volcanological Approach
F. BARBERI
H. BIZOUARD R. CLOCCHIATTI N. METRICH
R. SANTACROCE
A. SBRANA
[stituto di Mineralogia e Petrografia, University of Pisa, Italy.
Laboratoire de Petrographie et Volcanologie, E.R. 45 CNRS, Universit~ Paris XI, Orsay, France.
Istituto di Mineralogia e Petrografia, University of Pisa, Italy.
AGIP S.p.A., Servizio Esplorazione Geotermica, S. Donato Milanese, Italy.
ABSTRACT
The volcanic history of Somma-Vesuvius indicates that salic products compatible with an origin by fractionation within a shallow magma chamber have been repeatedly erupted (~(Plinian~ pumice deposits). The last two of these eruptions, (79 A.D. and 3500 B.P.) were carefully studied. Interaction with calcareous country rocks had limited imp0~ance, and all data indicate that differentiated magmas were produced by crystal-liquid fractionation within the undersaturated part of petrogeny's residua system at about 1 kb water pressure. The solid-liquid trend indicates that the derivative magmas originated by fractionation of slightly but signfficantly different parental liquids. Some lavas of appropriate composition were selected as parental liquids to compute the entity of the fractionatien. Results suggest that in both bases a fractionation of about 70 weight % was needed to produce liquids with the composition of the pumice. The combination of all data indicates that the two Plinian eruptions were fed by a magma chamber (3-4 km deep) having a volume of approx. 2.0-2.5 km 3. The temperature of the magma that initially entered the chamber was about ll00°C, whereas the temperature of the residual liquids erupted was Plinian pumice was 800 ° and 850°C respectively. There is no evidence that such a magma chamber existed at Vesuvius after the 79 A.D. eruption. These results have relevant practical implications for volcanic hazard and monitoring and for geothermal energy.
Bull. Volcanol., Vol. 44-3, 198I
INTRODUCTION
Somma-Vesuvius volcano is character- ized by a dominant association of silica undersaturated and potassic basic products, ranging in composition from tephrites to leucitites, that were erupted during long periods of prevalently effusive activity with only minor associated pyro- clastic rocks. This relatively monotonous sequence of mostly basic lavas was repeti- tively interrupted by highly explosive eruptions of Plinian type, that produced pumice-fall deposits, very often associated with pyroclastic surges and flows. Seven eruptions of this kind occorred in the last 17.000 years (DELIBRIAS et al., 1979), the most famous of which being that of 79 A.D. which destroyed Pompei, Hercula- neum and Stabiae.
These Plinian eruptions correspond almost always (6 times in 7 cases) to the beginning of a new eruptive cycle, follow- ing a long period of quiescence, as indicat- ed by the presence of paleosofls underly- ing the pumice deposits. Furthermore Plinian pumice is characterized by highly evolved (mostly phonolitic) composition and correspond to the most salic magmas erupted by the volcano. Because of this highly evolved nature and the long period of quiescence that preceded their erup-
TABLE i - Chemical analyses and C.I.P.W. norms of the Somm,-Vesuvius lavas and Plinian pumice. Key - G. s Mechanically separated glasses; PT s phonolite-tephrlte; PH -- Phonolite; PP ~ Peral-
l~Rllne phonolite; LT -- Leucite-tephrite; LC m Leucitite; PTR -- Phonolitic trachyte. 1 ffi Scoriace- ous sample with altered groundm~ss erupted in the time interval 17,000-12,000 years B.P. (pAlms Campana). 2 ffi 1944 lava flow.
X-ray fluorescence analyses; FeO wet snAlyses; MgO by atomic absorption (analysts: G. CRISCI, M. SAIT~rA and M. IVIZ~C~nNI).
SiO 2
TiO 2
A1203
Fe203
FeO
MnO
MgO
CaO
Na20 I
K20
P205
1 .o.~.
or ab an ne lc wo
di hy ol mt il ap
a c
rock type
AVELLINO PUMICE
white grey facies facies
I18 I l18G I16 I II6G
52.8553.0555.8455.39
0.49 0.49 0.12 0.I0
18.91 19.44 21.86 22.08
2.26 2.36 1.07 1.06
1.97 1.89 0.69 0.50
0.11 0.12 O.lO 0.11
2.55 2.47 0.19 0.08
6.44 5.31 1.82 1.41
5.22 5.85 8.64 9.61
6.56 6.38 7.00 6.37
0.23 0.23 0.02 0.01
2.41 2.41 2.65 3.28 2.39 2.64 2.56 2.61 i
38.8 37.7 41.4 37.6 49.0 46.9 55.2 49.9
POMPEI PurIICE 11T SOMf.IA (pre-caldera)
white grey facies f~cies LAVAS
' 1 t 15811 158G 157 157G lO 30 51 68 !
52.88 53.53 54.92 54.92 50.40 47.70 52.19 54.47
0.56 0.53 0.22 0.21 0.96 1.02 1.02 0.79
18.84 19.67 21.29 21.49 17.59 17.04 18.47 ]9.66
2.70 2.60 1.50 1.42 3.?9 4.07 3.98 4.08
1.81 1.79 0.89 0.88 3.58 3.83 3.16 2.64
0.12 0.12 0.10 O.ll 0 .13 0.15 0.13 0.14
2.12 1.44 0.19 0.15 6.58 5.84 4.39 2.11
5.68 4.73 2.64 2.60 8.96 9.14 7.60 5.83
4.42 4.85 6.33 7.16 2.19 2.61 2.42 2.94
8.29 7.93 9.34 8.44 5.28 6.87 5.61 6.50
0.19 0.18 0.02 0.02 0.60 1.12 0.61 0.46
0.46 0.61 0.42 0.38
31.2 29.5 33.1 38.4
VESUVIUS (post-caldera)
LAVAS
1691[ 412
50.12 50.78
0,93 l .Of
]7.23 18.03
2.66 3.21
4.74 3.50
0.12 0.12
5.91 5.07
8.29 8.14
4.52 2.52
3.28 6.26
0.58 0.65
1.62 0.73
19.4 37.0
I0.6 13.9 21.6 21.1
8.8 8.0 0.2
I 18.2 19.3 27.9 30.9
I -
0.7 2.7 1.9
15.8 13.3 1.7 2.1
0.6
3.3 3.4 1.6 O.l
0.9 0.9 1.2
0.5 0.5 O.l O.l
- - 2.8
PT PT-[ PHl PP
4.2 I0.4 7.2 II.6
7.1 8.5 2.1 1.6 i
18.0 16.6 25.1 26.5
1.8 1.2 3.8 4.0l
112.1 8.6 1.4 1.4 I
3.9 3.8 2.2 2.1
I.L 1.0 0.4 0.4
0.5 0.4 O.l O.l
Pt P'I PHFPH
I0.7 19.9 24.9
22.6 14.5 23.0 21.2
4.2 12.0 0.3 -
8.7 - -
I 14.2 18.6 8.4 3.7
[ - 0 . 5
8.7 5.7 5.7 2.4
4.8 5.9 5.8 5.9
! 1 . 8 1.9 1.9 1.5
i i . 4 2.7 1.4 1.0 I
LT LC LT PTR
20.3 8.4
17.0 19.4
9.7 7.0
15.2 13.3
8.7 6.2
3.9 4.7
1.8 1.9
1.4 1.5
LC LC
THE SOMMA-VESUVIUS MAGMA CHAMBER 297
tions, Plinian pumice is theoretically compatible with an origin by fractionation of basic liquids, that rested for a long time before eruption in a shallow magma chamber beneath the volcano. The pres- ence of a shallow magma chamber, locat- ed at a depth not exceeding 5 km within the Mesozoic carbonate basement beneath Vesuvius, has already been suggested (RITTMANN, 1933) and confirmed by the study of xenoliths of contact-metamorphosed calcareous rocks erupted during some Plinian phases (BARBERI and LEONI, 1980).
The aim of this paper is to check such an hypothesis and to at tempt an estimate of the size and depth of the Somma-Vesu- vius magma chamber on a purely petrolog- ical and volcanological basis. Our method- ological approach to the problem consists of the following steps:
- - collection of data relevant to investi- gate the petrogenesis of pumice from some significant Plinian eruptions, with emphasis toward the estimate of tempera- ture, pressure (and hence depth) of the fractionation process;
- - selection of basic lavas with composi- tion suitable to represent parental magmas for the Plinian pumice, to be used for a quantitative estimate of the amount of fractionation;
-- estimate of the depth and size of the magma chamber using the previous results and the measured volume of erupted Plinian products.
The problem has several important practical implications. On one hand, the reconstruction of the petrogenetic process occurring within the magma chamber can help in formulating a model for Vesuvius Plinian eruptions that is relevant for volcanic hazard evaluation and volcano surveillance. On the other hand, the magma chamber can be considered as the heat source of a potentially economically exploitable geothermal field, and a geothermal model for Vesuvius can easily be constructed once the depth and thermal capacity of the heat source, and its hydrogeological environment, are known.
THE PLINIAN PUMICE
The two most recent huge pumice fall deposits were selected for this study, the so called <<Pompei>> (79 A.D.) and <<Avel- lino ~ (about 3,500 y.B.P.) eruptions (LrRER et al., 1973).
AveUino and Pompei pumice deposits have quite similar field features being characterized by a sudden colour change passing from white (lower part) to grey (upper part). Both facies contain volcanic and metamorphic xenolithic ejecta whose amount increases upwards. Pumice fall layers are generally covered by thin (up to 1.0 m) pyroelastic surge deposits followed, at Pompei, by pumice flows and pyroclast- ic flows (RosI et al., 1981).
Although almost identical in the field, the two pumice deposits differ in their mineralogy and chemistry. Both have a composition ranging from tephritic phono- lite (grey) to phonolite (white), but the Na]K ratio is distinctly higher in Avellino pumice (1.2-1.9) than in Pompei (0.8-1.0) (Table 1). This is also reflected by the different nature of the feldspathoid that is nepheline in Avellino and leucite in Pompei. In both deposits, phenocrysts content increases from the white to the grey pumice, but Avellino rock is more porphyritic.
Samples representing the white and grey facies of Avellino, and Pompei were selected for chemical and microprobe mineralogical analyses. The chemical analyses of these samples and their resid- ual glasses, obtained by mechanical sepa- ration of phenocrysts, are reported in Table 1.
MINERALOGY
All data reported have been obtained by microprobe analysis of phenocrysts and microlites. The liquidus phases are clinopyroxene, biotite and sanidine, joined by nepheline at Avellino and leucite at Pompei. Later in the crystallization se- quence other minerals like amphibole, magnetite, scapolite, cancrinite and garnet precipitated from the magma, probably as a result of interaction between the melt
2 9 8 B A R B E R I - B I Z O U A B D - C L O C C H I A T T I - M E T R I C H - S A N T A C R O C E - S B R A N A
TABLE 2 - Modal estimate of phenocrysts content (vol. %) of the AveUino and Pompei pumice-fall deposits.
-}- : present (less than 0.1 vol. %) - : absent
of each eruptiom Pompei microlites have the same composition of phenocrysts, whereas those of Avellino are surprisingly more potassic (Or 84) than the corre- sponding phenocrysts.
Sanidine
Clinopyroxene
B io t i t e
Nepheline
Leucite
Opaques
Amphibole
Scapolite
Garnet
Plagioclase
AVELLINO If8 I l l 5 greylwhite
7.4 12.7
7.1 0.6
1.2 0.3
+ +
0.4 0.4
O.l 0.5
0.9 0.3
0.2 +
0.6 0.4
POMPEI
158 I 157 greyl white
0.3 2.8
4.3 1.3
2.4 0.4
- ?
0.2 +
0~.7 0.6
+ +
+ +
+ +
+ +
Total 7.9 15.2 7.9 5.1
and the calcareous country rocks. Small crystals of olivine and calcic plagioclase are also randomly found. Counterpoint modal estimates (12 thin section for each sample) of phenocrysts content are report- ed in Table 2.
The main results are now briefly presented and commented.
K-feldspar (Table 3)
It is a common component of the pumice both as phenocryst and in the groundmass. It is more abundant in the Avellino than in the Pompei pumice and decreases in both cases from white to grey facies. The Pompei sanidine is more potassic (Or 86) than the Avellino one (Or 75), and the phenocryst composition is the same in both the white and grey pumice
Feldsphatoid
As previously noticed, the Pompei pumice contains leucite phenocrysts, whereas nepheline is found at Avellino. Average representative analyses of these minerals are reported in Table 4. Avellino nepheline shows a quite constant composi- tion (Ne 85). The leucite microlites of Pompei grey facies are significantly more sodic K](Na+K) = 0.87 than phenocrysts K/(Na % K) = 0.98. Leucite is abundant in the groundmass of the white pumice.
Clinopyroxene (Table 5)
It strongly decreases from the grey to the white pumice. Two clinopyroxenes, respectively green and colourless, are observed in the Avellino pumice. The green variety has a fassaitic composition (W54_~2 Enls.s0 Fasl.ls ) and exhibits sector zoning with large variation in the A1 content (from 3.7 to 7.7% in the same crystal) (see Fig. 1). The colourless pyro- xene is diopsidic (Wo4e~o En4B.s9 Fss.11 ) and evolves toward A1, Fe, Ti enrichment in both grey and white facies. A thin A1- rich (5-7%) colourless outer rim mantles the fassaite phenocrysts. The diopsidic pyroxene of late cryst~iliTation probably precipitated during interaction of the phonolitic magma with the calcareous wall rocks. The two clinopyroxene varieties are found also in the Pompei pumice, but without the diopsidic outer rim on the green phase. For Pompei the microprobe data refer to the grey pumice only (Table 5). Note however that data from the two pumice varieties largely overlap, with a tendency for the Ca-Fe richest terms to concentrate in the Avellino grey pumice (Fig. 2).
Large compositional variations also occur in clinopyroxene phenocrysts of Somma-Vesuvius basic lavas, as it will be shown later, and seem typical of the
THE SOMMA-VESUVIUS MAGMA CHAMBER 299
TABLE 3 - Represental~ve microprobe analyses of K-feldspar from the Pompei and AveUino pumices.
l . 893 I 1. 893 4.070 4. 069 0.016 0.047 0.056 0.020
,0.514 0.562 3.547 3.476
115.98 16.01 4.12 4,06
86.2 85.7 12.5 13.8 1.3 0.5
Quaternary potassic volcanic province of Italy (THOMPSON, 1972). They probably reflect a disequilibrium during crystalliza- tion and prevent the use of clinopyroxene composition as tracer of magmatic evolu- tion.
Biotite and Amphibole (Table 6)
Amphibole is a relatively rare compo- nent scattered in the groundmass of both the Avellino and Pompei pumices, being generally more abundant in the grey facies. It occurs also as inclusion in garnet and scapolite. It can be classified as a .potassic ferropargasite, distinctly K-richer m the Pompei pumice. Biotite is an ubiquitous mineral in the Pompei and Avellino pnmice. Analyses refer to micro- phenocrysts and microlites from the Avel- lino white facies and to phenocrysts from the Pompei grey one. Biotite also forms inclusions in garnet of the Pompei white
pumice. An increase of Fe/(Fe + Mg) ratio is observed passing from phenocrysts and microphenocrysts (0.25) to microlites (0.50) and inclusions in garnet (0.35). All these phases differ from those of thermo- metamorphic limestone xenoliths by their distinctly iron-richer compositions.
Scapolite and Garnet (Table 7)
They are common phases in the Avel- lino and rather rare in the Pompei pumice. The presence of glassy inclusions shows that they crystallized from the magma probably as a result of contami- nation by the calcareous wall rocks. Scapolite composition is intermediate between dypire and mizzonite, and is distinctly more sodic than the typical meionites of thermometamorphic marble ejecta (SCHERILLO, 1935). Garnet has a melanitic composition with nearly 70% of andrachte.
This minera l (Ne 94%) is found only in the AveUino p . m i c e as smal l crys ta ls a n d as inclusions in scapoli te (Tab le 4 ) .
Plagioclase and Olivine
Both phases a re very rare and occur as small grains. Plagioclase also forms inclu-
sions in the garnet, It has usually a very calcic composition, ranging from bytownit- ic to anorthitic. Occasionally more sodic compositions (labradoritic to andesinic) also are found. Olivine is very magnesian and s~mnox to that reported from the contact metamorphosed limestone xeno- liths (HERMES, 1977). Both these minerals have a probable xenocrystic origin and have been inherited from the reaction
A I / 6 0 :
0 . 4 ° o
0 .3
0.2
0 . 1
-0 .3
-0 .2
-0.t
0 .4 I
o •
• • o • o o
0 • o OOo~
eo
I A V E L L I N O &
LPOMPEi 8 o..oo':' .,.
~P a a
o : . o • ~ # O• •
0 0 0 m eO
• D O0
go ~ , o.. °
J 0
~'0.4
O 0 0
0
0.6 I
10
~-0"3° o • *~ 0 • .
! 3 0 41
o
" 0 . 4 0 •
; o o o
- 0 . 3
o: " . ~-0.2 ~ - 0 . 2
o . o , : o .
1 6 9
- 0 . 3
- 0 . 2
i 0 . 7 0 . 8 0 . 1 0 . 7
I I !
I ! I ~-0.3 •
i e •
5
I " 0.2• •
0.1 0 . 7
I
51
• • F 0 " 3
I O0 0
0 . 2
O o 0
0
0 o 0 0
o ° , ~ 0
o • .11 • 0 0 0
0 0
0 . 7 0 . 8 i i
6 8
o . . T +°..' o.7 o ; e . e I I I I
M ~ ; / M S * Fe
FIG. 1 - Mg](Mg + Fe) vs A1/content for 6 oxygens diagrams for the clinopyroxenes of the Somma-Vesuvius pumice (Avellino & Pompei diagram) and lavas.
Results of microprobe analyses of pumice residual glasses and of melt inclu- sions in Avellino nepheline are shown in Table 8. These data have not the same meaning as the residual glasses analyses reported in Table 1. In fact the latter were separated mechanically and repres- ent the residual liquids after phenocryst precipitation, whereas the glasses analyzed with the microprobe correspond to the last liquids remained after the crys- tallization of the groundmass microlitic phases. These (<final>> glasses show no significant compositional differences between the Avellino grey and white pumice. The melt trapped in the nephe- line from the white pumice has a Na/K ratio higher than that of both whole rock and final glass, and its composition is similar to that of the mechanically sepa- rated glass. The finn] glass of the Pompei white pumice is slightly more sodic than the corresponding residual glass, reflecting the groundmass crystallization of leucite.
COMPARISON BETWEEN AVELLINO AND POMPEI PLINIAN PUMICE AND PETROGENETIC IMPLICATIONS
The mineralogy of the two Plinian pumice deposits is summarized and compared in Table 9. The chemical varia- tions observed within the deposit can be interpreted using a plot in the petrogeny's residua system (Fig. 3). I t is evident from the diagram that the Avellino p-m~ce is closer than the Pompei pumice to the << phonolitic >> minimum of the system, in agreement with its more sodic character. The Avellino representative points lie in the proximity of the K-feldspar-nepheline boundary line for 1 kb PH20. This is consistent with their mineralogy, character- ized by the coexistence of nepheline and sanidine, and suggests that there is probably equilibrium between liquid and solid phases, and hence that the sequence grey pumice-white pumice-residual glass, comprising also the melt inclusions in nepheline, can represent a liquid line of descent toward the phonolitic minimum at appro~ 1 kb PH20. The position of resid- ual glasses is anomalous, as it falls in a
302 BARBERI - BIZOUARD - CLOCCHIATTI - M E T R I C H - SANTACROCE - SBRANA
T A B L E 5 - R e p r e s e n t a t i v e a n a l y s e s o f c l i n o
StO 2
Ti 02
A1203
FeO
HnO
Mgo
CaO
Na20
total
Fe01.
Fe20~3
51 A11¥
A1 VI
Fe" '
Ti
Fe' '
Mn
Mg
Ca
Na
M9/( M 9+Fe )
En
Fs
go
AVELLINO PURICE
118 grey facies
] zooe, .reeo. , . ,tte [ I P" core rim core rim I c°reHPHrim
1.= Fe203 calculated according to Ham and Vteten (1971).
higher temperature porUon of the system - i.e. farther from the minimum - with respect to the white pumice. The increase in the K/Na ratio of the final liquid has probably tha same meaning of the previously described anomsJous increase of K content in Avellino sanidine micro- lites. A parallel slight but significant increase in Ca is observed in some white pumice phases (sanidine, scapolite) togeth- er with slight desilication of the residual liqukL We think that these slight composi- tional anomalies can be the result of inter- action of the phonolitic magnm with the calcareous waU-rocks. This would have
induced the final precipitation of scapolite and cancrinite, with consequent enrich- ment in potassium of the final liquid, and the other reported slight chemical changes in the last cryst~lli~.ing phases. The final precipitation of more potassic sanidine crystals can also have been favoured by a fluid pressure increase due to the development of decarbonation reac- tion around the magma chamber. The same could explain the late diopsidic composition of clinopyroxene. If this inter- pretation is correct, we can conclude that the modifications of the Vesuvius Pliuian pumice composition, due to coni~minatJon
THE SOM~J_A-VESUVIUS ~GIv IA CHAMBER
p y r o x e n e s . S a m e s y m b o l s a s i n T a b l e 4.
TABLE 7 - R e p r e s e n t a t i v e m i c r o p r o b e a n a l y s e s o f s c a p o l i t e a n d g a r n e t . S a m e s y m b o l s a s in T a b l e 4.
SiO 2
AI203
FeO
CaO
Na20
K20
total
he%
I t A V E L L I N 0 P U M I C E I POMPEI PUMICE
m
G A R N
116 157
MPH PH
S C A P O L I T E
118 116 116 116
PH PH MPH core rim
49.36 52,55 49.46 51.98
24.75 24.70 25.60 24.51
0.27 0.13 0.25
12.41 10.52 13.96 II.88
5.40 6.99 5.14 6.80
1.67 1.54 1.49 1.51
93.86 96.43 95.35 96.93
51 42 56 46
157 157
GM GM
SiO 2 35.96 35.86 35.01 35.46
ITiO 2 2.46 2.45 2.95 3,12 I
!AI203 8.40 8.55 7.65 8.42
'FeO 20.I0 18.35 19.02 18.30
IMgO 0.08 0.24 '0.21 0.30
CaO 30.82 32.03 32.25 31.71
iMnO 1.82 0.78 0.74
total 99.06 98.26 97.83 97.31
and 72.7 67.9 71.1 68.4
gro 19,2 27.7 24,4 27.0
spe 4.4 1.8 1.8
alm 3.4 1.8 1.8
pyr 0.3 0.8 0.8
TABLE 8 - A v e r a g e m i c r o p r o b e a n a l y s e s a n d C . I .P .W. n o r m s s i o n s .
1 = G l a s s f i l l i ng f r a c t u r e s i n c l i n o p y r o x e n e s ; - = N o t d e t e r m i n e d ; G = G r e y f ac i e s ; W = W h i t e f ac i e s .
AVELLINO PUMICE
118 G
~esidual
116 W
residual inclusions in-- nephe I i ne
5L26 52.42 ).30 0.06
21.50 22.64 !.72 l .36 ).16 ). 20 O. 02 !.57 1.60 ~.22 8.27 ~.63 5.64
Plag~oc]ase (an ~1.1~) I 92 - F.$ ] 90 - 82 I 89 - 86
A.ph,bo,~ (F./F~) [0.8,-0.6, I e.,-O.6,S, 4' -- e~ 1 O.BO-O.7,
Canl:rtn4te (tie m l . t ) I + [ 94 8 iOt l te {Fe/Fe44~ ) I 0.48-0.26 O. ZB-O. Z6 * Garnet (andr~dtte mo .~} J • 73 ~ 8g - ~1
with the calcareous country rocks, have been only minor and do not mask the dominant magmatic evolutive process controlled by crystal-liquid equilibrium.
Pompei pumice and residual glasses result in the diagram of Fig. 3 in a position clearly different from that of Avellino samples. The Pompei grey pumice, its residual glass and the white pumice actually fall, for a PH20 = 1 bar, in the leucite field. As the mineralogy is charac- terized by the coexistence of sanidine and leucite, the corresponding liquid must lie, in the case of an equilibrium process, on the K-feldspar - leucite boundary line. This suggest that the Pompei pumice crys- tallized under a PH20 slightly higher than 1 kb. Considering that sanidine is very scarce in the grey pumice, it is probable that only the residual liquid of this pumice lies on the leucite-sanidine line, that would have been reached after an initial precipitation of some leucite. Also Pompei results appear therefore compatible with a genetic process dominated by a crystal- liquid fractionation and do not show the anomalous change in the composition of the final liquid found in the Avellino pumice.
The genetic link between grey and white pumice of both deposits can also be discussed in the light of collected data. We have shown that the two facies have an almost identical mineralogy. As a whole the observed differences refer to the quantity of phenocrysts that are higher in the grey facies, and to the composition of the residual liquid as shown in Table 1.
The grey colour of the last erupted p ,mice is due to a combined effect of an higher iron and titanium content in the residual liquid and an higher abundance of mafic microlites. All data are consistent with a differentiation process by crystal- liquid fractionation giving rise to composi- tional vertical zonations within a magma chamber. We can reasonably assume that the white pumice originated by crystal fractionation of a liquid having approxi- mately the composition of the grey pumice residual glass that occupied the upper par t of the magma chamber. In a deeper zone of the chamber, instead, a selective enrichment in mafic phenocrysts occurred producing the grey pumice, whereas me_tic cumulates were formed at the chamber bottom. The different evolu- tion of the residual liquids of the two pumice facies would be explained by. a (( buffering ~ effect in the grey purmce due to the excess of the solid to be equilibrated with the residual liquid, with the consequent reduction of the evolutive potential of the liquid that was instead enhanced in the white pumice by the removal of the same solid from the liquid.
Data from the Plinian pumice appear therefore compatible with the initial work- ing hypothesis and allow one to pass to the following step of the study, namely to identify a suitable basic parent magma for the Avellino and Pompei pumice and to a t tempt a quantitative reconstruction of the fractionation process that occurred within the Vesuvius magma chamber.
PARENT MAGMA OF THE VESUVIUS PLINIAN PUMICE
The search of the Somma-Vesuvius products having a composition suitable to represent the parent magma of the Avel- lino and Pompei pumice is quite difficult because of the lack of rocks with interme- diate composition. As mentioned in the introduction, most of the lavas erupted from this volcano are in fact basic and range in composition from leucitites to tephrites.
306 B A R B E R ] [ - B I Z O U A R D - C L O C C H I A T T I - M E T R I C H - S A N T A C R O C E - S B P J t _ N A
Go o eo o o
~ O AVELLt NO 4 0 /
dO POMPEI /
40/7 ~ " O O0 ° O j ~ '
6 0
\ \
30 I '" os-I
I0 / 51 50
4O
169 68
,o s/:. 6O
so/ 41 4 o / ""eq~° ~
Wo
40
En .ES
FIG. 2 - The Somma-Vesuvins clinopyroxenes in the 40-60 Wo portion of the Wo-En-Fe triangle. Symbols - Avellino pumice: open circles = grey pumice; solid circles = white pumice.; squares =
pyroxenes from skarn rocks ejecta (acc. to data from HERMBS, 1977). Pompei punnce = open circles. Lavas and scorias (the number at the left upper corner of each diagram refers to the sample number): ph = phenocrysts; gm = groundmass.
A clear chemical change is observed in time, as shown in the Ks/O and SiO 2 v s
MgO diagram (Fig. 4). The products of the historic cycle of Vesuvius are actually more potassic and silica-undersaturated then the older Somma lavas. The petroge- ny's residua system can be used to help in seleet~g basic samples with a composition appropriate t o represent the parent magma of the pumice (Fig. 3).
The basic rocks form two main groups:
i) a first is represented only by the Somm~ samples that fall on both sides of the K-feldspar- leucite line at 1 kb PHi/ O. I t includes two lava flows with interme- diate composition found at Somma that probably represent a fractionation trend toward trachytic residual liquids. Trachyt- ic products actually occur among the old Somma products, both as ejecta lava blocks and early Plinian pumice (NARI, 1979). The samples plotted in the proxim-
THE SOMMA-VESUVIUS MAGMA CHAMBER 307
i ty of the K- fe ld spa r - leucite boundary could ins tead lead, through fractionation, to the Avell ino pumice.
ii) a second includes all the s t rongly unde r sa tu ra t ed Vesuvius lavas and few Sornma samples , and falls in the silica- poor sector of the leucite field. T h e p a r e n t m a g m a of the Pompei pumice has to be ident if ied in this group. Considering the re la t ive posit ion of the Avell ino and Pompe i pumice (Fig. 3), a genet ic l ink be tween the Avell ino pumice with a m a g m a belonging to the same strongly
unde r sa tu r a t ed potass ic group cannot be excluded.
On the basis of the previous considera- t ions 5 lava samples (Tab le 1) were se lec ted for a microprobe s tudy in order to check minera logica l ly the different possibili t ies. T h r e e samples are S o m m a lavas represen t ing possible p a r e n t m a g m a s of the Avell ino (n. 51 and 10) and Pompei (n. 30) pumice. S a m p l e 68 is one of the in t e rmed ia te S o m m a lavas, appa ren t ly belonging to a t rachyt ic suite. Final ly, s amp le 41 is a lava from the l as t
50 / - *50
20 80 Q
40 /40 A ~ 00 ~ r60
\, / l l ~ - ~ 1 Ne ~'Ks
2 5 0 ~ x . .50
%7 4 4 o
• 5 / " ~ _- s20 . ~ ".,~ " ~ \
3 o t ~ ~ i - w * ' _ . - ' 7 0
20 ' 80 * g
Fro. 3 - Plot of the Somma-Vesuvius pumice and lavas in the petrogeny's residua system (HAMILTON and McKENzm, 1965) at 1 kb PH~O.
Symbols - I = Pompei (g = grey; w = white) pumice joined with the mechanically-separated glass; 2 = Avellino (g = grey; w ~ white) pumice joined with the mechanically-separated glass; 3 = glass inclusions in the nepheline phenocrysts of the Avellino white pumice; 4 = microprobe- analysed residual glasses of the Avellino grey and white pumice; 5 = Mr. Sornma pre-caldera lavas; 6 = Mt. Vesuvius post-caldera lavas; 7 = glass inclusions in the olivine and chnopyroxene phenocrysts of scoria no. 169; 8 = glass inclusions in the leucite phenocrysts of scoria no. 169; S = Somma lavas; V = post-79 A.D. Vesuvius lavas. The boundary lines for Pd,~y = 1 arm are after ~IDALI (1963).
FIG. 4 - SiO 2 and I~O vs MgO diagrams of the Somma-Vesuvius Iavas.
Symbols - Solid circles = Mr, Somma lavas; open circles = Mt, Vesuvius lavas.
1944 Vesuvius eruption, chosen for compar- ison. An attempt has also been made to infer the liquid line of descent from the basic magma by analysing the melt inclu- sions in phenocrysts of a quenched scoria- ceous sample (rL 169). This sample was selected in spite of its altered glassy groundrnRss, because of abundance and size of phenocrysts and the freshness of their melt inclusions.
MINERALOGY OF SOMMA-VESUVIUS LAVAS
The most common phenocrysts are represented by clinopyroxene, olivine, leucite and plagioclase. Only samples 51 and 68 show some relevant mineralogical differences, due to the lack of leucite phenocrysts, to the occurrence of alkali
feldspar in the groundmass and to pres- ence (sample 68) of biotite phenocrysts.
Olivine (Table 10)
Is present both as phenocrysts and in the groundm,ss of all selected lavas, except sample 68 where it is confined to the groundm-ss. A practically constant composition (Fo 77-72) characterizes the core of phenocrysts and olivine inclusions in clinopyroxene (sample 10). Fayalite content increases progressively passing to microphenocrysts and microlites (up to 56% in sample 10). The lack of olivine phenocrysts and the high Fe content (up to Fa 62) of sample 68 microlites confirms the rather evolved nature of this rock The lower forsterite content of olivine phenocrysts in basic lavas compared to that of olivine of pum;ce confirms the non- magmatic origin of the latter.
Cl inopyroxene
It is an ubiquitous phase in all selected lavas both as phenocrysts and in the groundmass, and generally has an augltic- salitic composition with limited variations tiFig. 2). The more calcic salitic composi- ons are restricted to Al-rich clinopyrox-
enes from the samples 30 and 41. This is in agreement with the more undersaturat. ed and potassic character of the host rocks (CARMICHAEL, 1974). Fe-rich fassaite compositions observed in pumice are not found in the clinopyroxenes of the lavas, that is well in agreement with the more evolved nature of the p,mlce. Clinopyrox- ene phenocrysts of sample 68 (phonolitic- trachyte) show nice hour-glass zoning and the diopsidic composition are related to Mg,Si preferential enrichment in the (111) sector (HOT.T.IaTER and GANOARZ, 1971; FERGUSON, 1973; WASS, 1973; LEUNG, 1974). Sector zoning is ac~ml!y a frequent feature of clinopyroxene pheno- crysts of many Vesuvius lavas (SAVELLI, 1969; THOMPSON, 1972; RAHMAN, 1975). In Fig. I the A1 content calculated on the basis of 6 oxygens is reported as a func- tion of Mg/(Mg q- Fe) atomic ratio. Figure 2 shows the usual negative correlation
THE SOMMA-VBSUVIU8 MAGMA CHAMBER 309
TABLE 10 - Representative microprobe analyses of olivine in the Somme-Vesuvius lavas.
310 B A R B E R I - B I Z O U A R D - C L O C G H I A T T I - M E T R I C I - I - S A N T A C R O C E - S B R A N A
between A1 and Mg/(Mg + Fe) and the higher Fe content of microlites with respect to coexisting phenocrysts. No A1 increase is observed passing from pheno- crysts to microlites, in contrast with what was stated by DOLFI and TI~GmA (1978).
Feldspar (Table 11)
The An content of plagioclases decreases progressively from anorthitic- bytownitic compositions (An 90) in the case of phenocrysts to acid labradorites (down to An 50) in microlites. All
• analysed samples exhibit such a wide compositional range of plagioclase (Fig. 5) apparently without any relation to the
degree of evolution of the host rock Note that the difference in Fig. 5 are more apparent than real, as sample 30 does not contain plagioclase phenocrysts, whereas the composition of plagioclase microlites of sample 51 could not be determinecL Soda sanidine (Or83 Aba2 Ans) is present in the groundmass of the most evolved lava sample Rnalysed (68) together with quenched Or-rich alkali feldspathic areas (found in sample 51 also). It occurs also in the groundmass of the scoria 169.
A biotite (Table 6) analysis (average of 7 analyses) is available for micropheno- crysts of sample 68 showing an intermedi- ate biotlte-phlogopite composition with a Fe/(Fe + Mg) ratio of 0.34.
An
A )r
Ab Or
FI6. 5 - Ab-Or-An molecular triangle of microprobe-analysed feldspars of Sorams-Vesuvius pumice (Avellino & Pompei diagram) and lavas.
Symbols - Avellino & Pompei: solid circles = Avellino white pumice; open circles = Avellino grey pumice; squares = Pompei pumice. The numbers in the diagrams refer to the lava sample number of Somma lavas, except for no. 41 that refer to the 1944 Vesuvius lavas.
THE SOMMA-VESUVIUS MAGMA CHAMBER 311
~ ! . T INCLUSIONS IN THE PHENO- CRYSTS OF THE SCORIA SAMPLE 169, AND PETROGENETIC IMPLICATIONS
The quenching of sample 169 allowed fresh melt inclusions in phenocrysts to be preserved, and these provide information on the chemical changes induced in the basic magmatic liquid by crystallization.
Microprobe analyses were made either on glass drops found in leucite, clinopyrox- ene and olivine and on glass Filling frac- tures in clinopyroxene (Table 8). Inclu- sions in olivine and clinopyroxene show rather similar composition with low SiO2 undersaturation (normative ne < 5.0%) and N a / ( N a + K ) ratios ( < 0.4). Melt inclusions in leucite are more undersatu- rated (he > 10%) and sodic (Na/Na H-K comprised between 0.54 and 0.61). Glass filling fractures in clinopyroxene pheno- crysts has an intermediate composition and it might in fact represent an interme- diate stage in the evolution of the residual liquid not yet affected by an important leucite precipitation. In the petrogeny's residual diagram (Fig. 3c) the composi- tions of melt inclusions fall along the leucite-K feldspar line for 1 kb PH20. The melt inclusions in olivine and clinopyrox- ene are, however, poorly representative in this diagram because of their relatively low content in SiO2, nepheline and kalsi- lite components. The melts in leucite slightly diverge from the lkb leucite- K feldspar boundary line towards sodic compositions. This could reflect a modifi- cation in the composition of the trapped melts due to a leucite crystallization on the inclusions walls, or to a PH20 lower than 1 kb during the crystallization of leucite outer rim where the melt inclu- sions occur. Their position in the diagram however suggests that the leucite outer rim formed at a lower temperature than olivine and clinopyroxene phenocrysts. It is therefore possible that the residual melt has reached the lc-Kfp line only in a rela- tively late stage of the crystallization, as it is also suggested by the presence of sanid- ine only in the groundmass of the scoria.
As the whole, the mineralogical and melt inclusions data of the Somma-Vesu- vius basic lavas suggest that:
i) sample 68, because of the lack of olivine and leucite phenocrysts, the high iron content of the groundmRSs olivine, and its position in the petrogeny's residua system, can indeed represent an interme- diate step of a fractionation trend leading to a trachytic residual liquid. Therefore it will not be further considered in this study that deals with phonolitic residua;
ii) the parent magma of the Pompei pumice has to be identified in the group of highly undersaturated and potassic lavas. The initial basic liquid at upper crustal levels could then have had a composition similar to that of sample 30;
iii) the composition of melt inclusions in scoria 169 phenocrysts indicates that the fractionation of basic liquids corre- sponding to Somma tephritic lavas (e.g. sample 51) can lead to phonolitic residua with a composition rather similar to the Avellino grey pnmice (a part a higher iron contents); data are qualitatively consistent with a fractionation process initially domi- nated by olivine, clinopyroxene and pIagioclase, then by leucite and finally by the simultaneous precipitation of leucite and sanidine (with clinopyroxene).
The lack of a complete rock-suite from basic tephrites to phonolites obviously prevents one to affirm the parentage between the selected Somma lavas and the Plinian p,mices of Avellino and PompeL Neither do we want to say that the << intermediate >~ melts trapped in the scoria phenocrysts represent a necessary intermediate step towards the phonolitic Avellino pumice. We want simply to stress that the entire data body is compatible with the envisaged petrogenetic process
TABLE 12 - Heating stage thermometry.
Sample
PFSV 116 Avellino white pumice
I#
PFSV 157 Pompei white pumice
PFSV 169 Scoria
I Host ~tfieral
Sanidine
Scapalite
Nepheline
Clinopyroxene
Sanidine
Olivine
Clinopyroxene
Leucite
T ° C
780-820+10
780-820+10
780-820+10
950-1050+15
850-870+10
1200+15
1080-1120+15
I080-I120+15
312 B A R B E R I - B I Z O U A R D - C L O C C H I A T T I - M E T R I C H - S A N T A C R O C E - S B R A N A
and that compositions like those of samples 51 and 30 can indeed correspond to the initial liquids that entered the Vesuvius magma chRmber and produced, mostly by crystal fractionation, the phono- litic pumice of Pompei and Avellino.
OPTICAL THERMOMETRY
c~nnot be compared with isotherms in the petrogeny's residua system.
Optical thermometry therefore confirms petrological data, and suggests that basic liquids could have entered the magma chamber at - 1200°C and that the frac- tionated residual liquids were erupted as Plinian pumice respectively at approx. 800°C (Avellino) and 850°C (Pompei).
Thermometric measurements under the heating stage (method described by CLOCCH~TTI, 1975) were carried out on melt inclusions occurring in phenocrysts of the Pompei and Avellino white pumice, and in the scoria 169 (Table 12). The identical melt homogenization tempera- tures shown by inclusions in nepheline, sanidine and scapolite phenocrysts of the AveUino white pumice confirm that the corresponding liquid was on the nephe- line-alkMifeldspar boundary and precipi- tated scapolite by endomorphic reaction with calcareous country rocks. Further- more, there is an excellent correspond- ence between the experimentally deter- mined temperatures and those inferred by the plot of the Avellino white pumice in the 1 kb PH20 petrogeny's residua system (Fig. 3). Melt inclusions in clinopyroxene have a distinctly higher homogenization temperature (~ 1000°C) suggesting the earlier apparence of this mineral that~ however, on petrographical evidences, persisted to crystallize when sanidine and nepheline precipitated. The crystall/zation temperature of the Pompei white p,,mlce sanidine is significantly higher - out of the analytical error - than that of the AveUino sanidine. Again, the results perfectly agree with the plot of the Pompei white pumice in an higher temperature portion of the residual diagram and are also consistent with the higher or-content of the Pompei san/dine.
Crystallization temperatures obtmned from the glassy inclusions in scoria 169 phenocrysts indicate that olivine was the first crystallizing phase at 1200°C while clinopyroxene and leucite included melt at a somewhat lower temperature, but still above 1000°C. These high temperatures are obviously related to the basic composi- tion of the host liquid and therefore
ESTIMATE OF THE FRACTIONATION DEGREE AND OF THE VOLUME AND SHAPE OF THE MAGMA CHAMBER
The results presented in the previous chapters allow to estimate the quantity of solid that must be removed from the selected initial liquids in order to produce residual melts having the composition of the Avellino and Pompei pumice. The lack of intermediate compositions makes the calculations very difficult~ because we are obliged to use pairs of liquids of very different composition and to introduce a large number of solid phases, instead of proceeding by progressive steps using few solid phases. The following transitions were considered:
- - lava 51 to the residual glass of the Avellino grey pumice
-- lava 30 to the Pompei grey pumice --Pompei grey to white pumice - - Avellino grey residual glass to white
pumice. The solid phases considered for each
transitions are those suggested by the previous qualitative reconstruction of the crystallization process. For the solid phases composition average values between the compositions found in the initial and final rocks were used. Results of computation, made using a XLFRAC program (STORMER and NICHOLLS, 1978) show that:
i) the Avellino white pumice can be obtained by removal of 72% of solid from the grey pumice residual liquid;
ii) the Pompei white pumice can derive from the grey pumice by fractionation of 22% solid;
iii) it is possible to obtain the residual liquid of the grey Avellino pumice by removal of 79% solid from lava 51 and the
THE 801VI~VIA-V~I~S MAGMA CHAMBER 313
TABLE 1 3 - Fractionation model (R 2 is the quadratic sum of residua).
In i t i a l Final liquid l iquid
PFSV 51 PFSV 118G
PFSV 118G PFSV 116
PFSV 30 PFSV 158
PPSV 158 PFSV 157
i Total wt%J so l i d phases (wt%) tO be removed
s o l i d f t . px pl
78.6 J 5.7 12.8 21.5 - -
71.0 - 30.7 - - 17.1
67.0 6.7 19.8 18.4 18.4 -
?1,7 - 13.9 - 6 . l -
mt [ Kfp
4.0 34.7
45.2
3.7
2 . t
R 2
0.4S I
Z 57
~..17
Pompei grey pumice by removal of 67% solid from lava 30.
These results are far from being sure (see Table 13) and have a purely indica- tive value. However it is to be noted that the results obtained for the transition from the grey to white pumice are consis- tent with the estimated volumes of erupt- ed products in the Pompei and Avellino pumice deposits (Table 14). With all the caution required by the case, we can assume that the liquid that produced the Plinian pumice of Pompei and Avellino corresponds to approx. 30 weight % of the initial liquid, whose volume can then be estimated using the measured volumes of the erupted pumlce. The estimates of volumes and densities of both the Pompei and Avellino pumice fall deposits (LIRER e t al. , 1973) are reported in Table 14 together with the equivalent phonolitic liquid volumes calculated by assuming for them a density of 2.2 g/cm 3. On the basis of these data the volumes of the primary liquids (assumed density = 2.6 g]cm 3)
entering the magma chamber are calculat- ed to be 2.31 km 3 for Pompei and 1.83 km 3 for Avellino. I t has to be noted that the volume estimates refer to Plinian pumice fall deposits only, and that pyro- clastic flow and surge deposits associated to them were not considered. The mainly phreatic nature of such deposits (SHERIDAN e t al., 1980) suggests however that their magmatic component was very scarce and should not modify appreciably the previous estimate. On the other hand the reduction of the magmatic fraction in the last eruptive phases and the parallel increase of xenoliths of intrusive and cumulitic rocks suggest that no magmatic liquid was left in the chamber after the Plinian eruptions, Therefore the volume that has been estimated for the initial liquid can be considered to represent also the volume of the Vesuvius shallow magma chamber.
The problem of the shape of the cham- ber is a difficult one. Following the sugges- tion of BARBERI and LEONI (1980), based
TABLE 14 - Data relevant to the calculation of the volume of the Pompei and Avellino p~rnary magmas. Pumice density and volumes from LINER et al., (1973).
~ange of pu- Pumice !quivalent Vol. prl- mice density
(g/cm-)
AVELLINO I grey PUMICE FALL facies DEPOSIT white
facies
grey POMPEI facies PUMICE FALL white DEPOSIT
facies
0.7-1.7
0.5-0.7
0.7-I.0
0.6-0.9
volu~s (Km-)
1.76 2.10
0.34
1.61 2.60
l .og
sol.phonolitic liquids 3 :d=2.2) (Kin)
0.66
0.82
mary liquid (d=2.6) as- suming 70wt% fractionation
on the comparison of metamorphic calcar- eous ejecta in the Plinian pumice with the Mesozoic series outcropping near Vesu- vius, we can accept a vertical dimension of at least 2 kin. The more logical resulting shape is cylindrical with a diameter of approx. 1.1 km for Avellino and slightly higher (1.2 kin) for Pompei.
CONCLUSIONS
As a whole the collected data appear consistent and the iniGal working hypoth- esis is fully confmned. A coherent body of volcanological and petrological data indi- cates that the phonolitic pumice erupted during the two most recent and important Plinian eruptions of Vesuvius originated mostly from the solid-liquid fractionation of a parent tephritic liquid in a shallow magma chamber. The chamber was locat- ed at a depth from approx~ 2 to 4 kin, within the Mesozoic limestone series that forms the sedimentary basement of the volcano. The volume of this chamber was of ca- 2 km 3 and its shape possibly cylind- rical with h ~ 2 km and diameter of 1 kin. The tephritic magma entered the cham- ber at a temperature of 1200°C. During a long period of rest, lasted for several centuries at least, it underwent a fractio- nation and residual phonolitic liquid was formed in the upper part of the chamber, after removal of about 70 weight % of solid phases that accumulated in the lower part of the chamber. Interaction with the calcareous country rocks also occurred, producing contact reaction skams whose solids were partly incorpo- rated into the magma (olivine, plagioclase) and inducing a late precipitation in the phonolitic liquid of anomalous solid phases (scapolite, cancrinite) with the consequent chemical modification of some magmatic phases (sanidine and perhaps clinopyroxene). Then the eruptions occurred when the phonolitic liquid had a temperature of 800°C (Avellino) and 850°C (Pompei), giving rise to a complex series of volcanic events, including an initial magmatic phase (pllrnice fall) followed by an increasing interaction of water with the magma and the hot solid
of the chamber (pyroclastic flows and surges).
The present study indicates the possi- bility of obtaining quantative data on shallow magma chambers feeding central volcanoes, also with a purely volcanologi- cal and petrological approacl~ The same approach can be utilized for a number of active or recent volcanoes where the erupted sequence can be sampled and volumetrically estimated. Other cases would probably result more simple than the Vesuvius example that is not an ideal one because it lacks of a continuous basic- salic series of products.
The results obtained on the Vesuvius magnm chamber can be used to test geophysical models and to plan an appro- priate strategy in organizing the geophysi- cal and geochemical surveillance of the volcano. Furthermore, they can serve for geothermal energy evaluations, consider- hag the magma chamber as a shallow heat source and computing models of thermal diffusion in the surrounding rocks. A study of this kind is in progress in cooperation with geophysicists of the University of Naples, using the results on the magnm chamber presented in this paper, the whole volcanic history of Somma-Vesuvius and its hydrogeological framework
ACKNOWLEDGEMENTS
This work has been financially support~ ed by CNR (Italian Geodynamics Project, publication n. 355) and by the CNR (Italy)- CNRS (France) cooperative pro- gram on exchanges of researchers visit.
The joint-venture AGIP-ENEL for geothermal energy exploration in Italy has sponsored, through Aquater S.p.A., the microprobe analyses made at the Labora- toire de Microanalyse Electronique of Orsay and the work of A. Sbrana. All these organizations are gratefully aknow- ledged.
REFERENCES
BARBERI, F. and LEONI, L., 1980, Metamorphic Carbonate Ejecta from Vesuvius Plinian Eruptions. Evidence of the Occurrence of
THE SOMMA-VESUVIUS MAGMA CHAMBER 315
Shallow Magma Chambers. Bull. Volcanol., 43-1, p. 107-120.
CARMXCttAEL, I. S. E., 1974, Igneous Petrology. McGraw Hill, 266 pp.
CLOCCHIATTI, R., 1975, Les inclusions vitreuses des cristaux de quartz. Mere. Soc. Geol. Fr., 54 (122), p. 1-96.
DELmRtaS, G., DI PAOLA, G. M., ROSI, M. and SANTACROCE, n., 1979, La storia eruttiva del complesso vulcanico Somma-Vesuvio rico- struita dalle successioni piroclastiche deI Monte Somma. Rend. Soc. It. Min. Petr., 35 (1), p. 411-438.
DOLF[, D. and TRIGILA, R., 1978, The Role of Water in the 1944 Vesuvius Eruption. Contr. Mineral. Petrol., 67, p. 297-304.
FERGUSON, A. g., 1973, On Hour Glass Sector Zoning in Clinopyroxene. Mineral. Mag., 39, p. 321-325.
FUDALI, R.F., 1963, Experimental Studies Bearing on the Origin of Pseudoleucite and Associated Problems of Alkalic Rocks Genesis. Geol. Soc. Amer. Bull., 74, p. 1101- 1126.
HAMmTON, D. L. and MACKENZIE, W. S., 1965, Phase~equilibrium Studies in the System NaAISiO. - KAlS i 05 - SiO 2 - H.zO. Mineral. ~lag., 34. p: 214-231.
HANN, H. H., VrETEN, K., 1971, Zur beberch- nung der kristaUchemischen formel und des Fe 3+ gehaltes yon klinopyroxen ans elek- troneustral-microanaIysen. Neues. Jahrb. Mineral. Monatsh., p. 310-314.
HERMES, O.D., 1977, Petrochemical Signific- ance of Xenolithic Nodules Associated with Potash-rich Lavas of Somma-Vesuvius Volcano. Unpub. NSF technical report, Rhode Is. University.
HOLLISTER, L.S. and GANCRAZ, A.J., 1971, Compositional Sector Zoning in Clinopyrox- ene in the Narce Area, italy. Amer. Mineral., 56, p. 959-979.
LEt~G, I. S., 1974, Sector Zoned Titanaugite: Morphology Cristal Chemistry. and Growth. Amer. Mineral., 59, p. 127-138.
Deposits from Somma-Vesuvius, Italy. Geol. Soc. Amer. Bull., 84, p. 759-772.
Lr_vrH, W.C., 1967, Studies in the System KAISiO~ - SiO 2 - H20 pt.1, Inferred Phase Relations and Petrologic Applications. Journ. Petrol. 8, p. 372-416.
NARI M., 1979, Le eruzioni di pomici (~Pliniane~O del Somma-Vesuvio. Unpubl. thesis, Pisa.
RAHMAN, S., 1975, Some AIuminous Clino- pyroxenes from Vesuvius and Mr. Somma, Italy. Mineral. Mag., 40, p. 43-53.
RITTMAN, A., 1933, Die geologisch bedingte evolution und differentiation des Somma- Vesuvmagmas. Zeitschr. f. Vul]~, XV, 1-2, Berli~
RosI, M., SANTACROCE, R., SHERIDAN, M., 1981, Volcanic Hazard of Vesuvius, Italy. Bull. B.R.G.M., (2) IV-2, p. 169-174.
SAVELL~, C., 1967, The Problem of rock Assimi- lation by Somma-Vesuvius: L Composition of Somma and Vesuvius Lavas. Contr. Min. Petro., 16, p. 328-353.
SHERIDAN, M., BAr~BERI, F., ROSl, M., SANTA- CROCE, R., 1981, A Model for Plinian Erup- tions of Vesuvius. Nature, 1,289, p. 282-285.
SCHERn~L0, A., 1935, La meionite del Somma- Vesuvio, Per. Mineral., 6, 3, p. 2.
STOaM~r~, J. C., NlCHOLLS, J., 1978, XLFRAC: a Program for the Interaction Testing of Magmatic Differentiation Models. Compu- ters in Geosciences, 4, 2, p. 143.
STRONG, D. F., 1969, Formation of the Hour Glass Structure in Augite. Mineral. Mag., 37, (288), p. 472-479.
THOMPSON, H., 1972, Oscillatory and Sector Zoning in Augite from a Vesuvian Lavas. Carnegie Inst. Wash. Year book 71, p. 463- 470.
WASS, S., 1973, The Origin and Petrogenetic Significance of Hour Glass Zoning in Titaniferous Clinopyroxenes. Mineral. Mag., 39 (302), p. 133-144.
Ms. received Nov. 1980; sent to review Nov. 1980. Revised ms. received and accepted July 1981.
