Studio Geologico Dott. Giancarlo Carboni Via Nazionale 277 / 09039 (CA) P.lva 02671260921 Cod.Fisc. CRBGCR67M30L924H

Geothermics, VoI. 13, No.4, pp. 333 - 347,1984 . 0375 - 6505/84 $3.00 + 0.00 . Printed in Great Britain.

, \ , GRAVITY SURVEY AND INTERPRETATION OF BOUGUER h ANOMALIES IN THE CAMPIDANO GEOTHERMAL ARE .,: ~~. (, ITAL Y)* PUla -

R. BALIA,t M. CIMINALE,t M. LODDO,t G. PECORINI,§ G. RUINAt and R. TRUDut

I +Istuuto di Geofisica Mineraria, Università di , , +Dipartimento di Geologia e Geofisica, Università di Bari, Italy and §Dipartimento di Scienze della Terra, Università di Cagliari. Italy

(Received November 1983; accepted for publication February 1984) •

Abstract-A gravity survey of the Campidano geothermal fields and surrounding region was conducted in 1981. Il covered an area of 1900 km' and included 952 uniformly distributed stations. The Bouguer anornaly is generally negative within the Campidano grabén, reaching - IO mgal in the centrai zone, whereas a positive Bouguer anomaly prevails outside the graben, exceeding 20 mgal in several areas. The gravity data were interpreted using spectral analysis and two-dirnensional models, to determine the thickness of sediments and andesitic volcanics within the graben. The total thickness of these formations reaches 3000 m in the centre, bUI is reduced elsewhere, especially towards the sides of the graben. The thermal springs on both the eastern and western sides of the graben are associated with residual positive anomalies and are near very steep gradients in the Bouguer anomaly.

INTRODUCTION The geological and structural features, the presence of a large number of surface hydrogeothermal manifestations and their geochemical characteristics indicate that the island of Sardinia (Fig. 1) is one of the most promising regions in Italy for exploitation of geothermal resources. The increasing need to find new and renewable energy sources has stimulated interest in this island, especially in the Carnpidano graben (in the south) and Logudoro area (in the l north). Research is at an advanced stage and the first results can be synthesized as follows: (i) maximum heat flow in the Campidano graben is 4.0 H.F.U. (Loddo et al., 1982), (ii) the distribution of deep temperatures suggests values near !OO°C at depths between 1.0 and 1.5 km (Panichi and Squarci, 1982). Unfortunately there are not enough thermal and heat flow data at present to compile a reliable geotherrnal map for assessing the potenti al of the geothermal resources. A gravity survey was carri ed out in the Campidano graben prior to drilling a new series of geothermal boreholes. The gravity anomaly maps are shown in this paper (Figs 2 - 4). The main structural features are analysed qualitatively; the maps are also interpreted quantitatively by the spectral analysis technique and by means of two-dimensional models, along sections across the Campidano graben. This survey has irnproved the knowledge of the regional geologic framework and the underground structure and revealed a close relationship between the local gravity anomalies, structural trends and several geothermal areas.

GEOLOGIC SETTING Sardinia has a Paleozoic basement, consisting mainly of graniti c metamorphic rocks; its western sector is intersected bY""a roughly N - S trending Tertiary Rift (Sardinian Rift) I contaìning the Campidano graben (Fig. 1).

*Research carried out with financial contribution from the ltalian National Research Council, Finalized Energy Project, Geothermal Energy Sub-Project, contrae t No. 80.0924.

333 . .. .." l' -_._ _---_._ .._~~ i I

334 R. Balia, M. Cimina/e, M. Loddo, G. Pecorini, G. Ruina and R_ Trudu I lO' E N

41' 41' I I I I I !

F:""l1 ~ ~===~2 ~~~W []4

~5 Da

7~/

8* g@ 39' 39' 10 29 30km

IO' E Fig. I. Geological sketch map of Sardinia island. (I) Quaternary sedirnents, (2) Miocene deposits, (3) Miocene delta sand, (4) Tertiary basai t, andesite, rhyodacite, rhyolite, ignimbrite and tuFf, (5) Mesozoic limestone and dolornite, (6) Paleozoic granitic-schistose basement, (7) fault, (8) thermal spring, (9) borehole.

I The Rift is a deep, wide depression lying between the Paleozoic blocks of Nurra, Iglesiente- \ Sulcis and the eastern half of the Island; it is filled by volcanics (mainly andesite, ignimbrite and \ rhyodacite) and Oligo - Miocenic marine and continental deposits (Pala et al., 1982a). The voIcanic units of the Rift, reaching a maximum thickness of at least 1000 m, belong to the back-arc volcanism, that evolved during the drift and subsequent rotation (Middle - Upper \ I Gravity Survey oJ Campidano Area, Sardinia 335

9·00'

N t

39"

D Pno-Pìerstocene semmenrs (!)CI ac.enote ED easenrc and rtwcunc lava IIoW5 IPliocene) *HotSPring

DTefliarysediments • Gravity stauon

~ AndesitiC rcrmeucn '\ Gravity comours (Oligo-Miocene) \ tntervat 2 mGal

D pareozotc Basement _~UI1 ..•..•..""". /"

39°20' lOkm

8"40' E

Fig. 2. Bouguer anomaly map and generai geology of Campidano graben.

Oligocene - Burdigalian) of Sardinia from the Gulf of Leone up to its present position. The volcanic rocks generally overlie the Paleozoic basement and Mesozoic limestones (Nurra), or Eocenic sediments (Sulcis, Campidano, Trexenta). Upper Oligocene- Messinian marine and fluvial-lacustrine deposits (conglomerates, sandstones, tuffites, marls, limestones and clays), about 1000 m thick, are interbedded with, or overlie, the vo\canic rocks. The Campidano graben, whose limits are the Gulfs of Oristano and Cagliari, extends in the southern part of the island (Figs l and 2). The boundary faults in the graben have maximum throws up to 3000 - 4000 m. The depression is filled by a Pliocene - Quaternary sequence : delta (800 - 1000 m thick) mainly composed of continental sediments interbedded with basai tic lavas. te, (6) Ali these units lie in discordance on the Oligocene - Miocene volcanic - sedimentary deposits of the Rift. Four boreholes drilled by SAIS and AGIP within the Campidano area did not reach ente- the Paleozoic basement; three of them (OR l, 1800 m; OR 2, 1700 m; AGIP, 2400 m) ended in :and the Oligocene - Lower Miocene sediments or volcanics; borehole CI reached a depth of 1700 m, ending in Eocene continental deposits (clays and sandstones). The depth of the basement is 19 to consequently stili unknown and the occurrence of Mesozoic limestones above the Paleozoic pper crystalline rocks has still to be proved. From a structural point of view, the geophysical and r---·--·__··_-_·····- ._---~_ .. _._----_ .._._---_ .._---_ .. ----

I 336 R. Balia, M; Cimina/e, M. Loddo, G. Pecorini, G. Ruina and R. Trudu

S'40'E ,'00'

39-.0' 311'40'

l I I j I I ~

39"20' 39'2 I i 10km

S'40'e ,'00'

Fig. 3. Regional gravity anomaly (w<21t/15 krrr').

geological data lead us to distinguish three main basins in the Campidano graben; from north to south these are: Oristano, S. Gavino and Cagliari basins. They are bounded by sills or structural highs, whose overall features were identified in Trudu (1963). The main hydrothermal manifestations of Sardinia occur for the most part on the margins of the Tertiary Rift, along the major faults. They are prevalently chlorinated waters with temperatures of 45 -78°C and up to 50 I/s of discharge. Sodium bicarbonate (rich in CO,) springs occur along the post-Miocene boundary faults in the Campidano graben. CO, is connected with a limestone thermal metamorphism; on the contrary, the watershed area and the deep hydrological circulation seem to be represented by the Paleozoic basement (Bertorino et al., 1982; Caboi et al., 1982; Caboi and Noto, 1982; Dettori et al., 1982; Pala et al., 1982b). The hydrological role of the voIcano - sedimentary sequences of the rift and graben, up to now considered only as an impermeable cover, has stili to be c\arified. The main problems in the Campidano area are: (i) the depth of the Paleozoic basement, (ii) the structural relationship between each of the three above-rnentioned basins, (iii) the distribution and subsurface morphology of the andesitic voIcanics and (iv) the trend of the major fauIts that border and intersect the graben area, with special reference to hydrothermal circulation. .:

Gravity Survey of Campidano Area, Sardinia 337

S'40'E 9'00'

39'010'

/1

39'20' 39'2

fig. 4. Residual gravity anomaly (w>21t/15 krn ").

GRAVITY SURVEY Several geophysical surveys, including gravity and magnetic (Trudu, 1953, 1961, 1962, 1963), rorth to geoelectrical (Marchisio et al., 1982), magnetotelluric (Finzi Contini, 1982) and seismic sills or . prospectings (Martinis, 1969; Guerra, 1981), have been concerned directly or indirectly with the Campidano graben. There are not enough data, however, to make a quantitative analysis of the irgins of graben structure, although the information available on the southernmost part is sufficient. ers with The gravity survey recently carri ed out in the Campidano covers ali the area from Latitude in CO2) 39° 17' N to 39°45' N (Fig. 2). The measurements were made in 952 stations, with a density of 1 CO2 is station every 2 krn", using a LaCoste & Romberg model G gravimeter (No. 351) with a scale and the constant K == 1.06 mgal/div and instrumental drift of less than 0.05 mgal/day. The reference orirto et network consists of 12 stations of the gravity survey of Sardinia (Trudu, 1962). Elevation of our The n», stations was determined by trigonometric leveIling using a AGA model 14 geodimeter and a to now Wild model 12 theodolite. The Italian Military Geographic Institute bench-marks were also utilized. .sement, The mean estimated error is about 0.03 mgal for the gravity measurements and 0.2 m for the (iii) the elevation. j of the To obtain the Bouguer anornaly, ali the usual corrections were applied to the observed uhermaì gravity, In order to obtain more reliable results and highlight the pattern of the structure, the densities used for these corrections were obtained on the basis of a lithological evaluation and rr~'----'... . __._--_._ _ _ .. __ _ _ _-'-_ .._ ..-_.._-~.------_.. -' I 338 R. Balia, M. Ciminale, M. Loddo, G. Pecorini, G. Ruina and R. Trudu of measurements made on rock samples. The gravitational effect of the terrain was assessed up to Hayford zone L. This distance is sufficient for the limited area surveyed and for the moderate topographic relief. The mean estimated total error for the Bouguer anomaly is about 0.1 mga\. Absolute, residual and regional Bouguer anomaly maps were drawn by a computer program on the basis of the methods outlined by Hessing et al. (1972) and Bolondi et al. (1977). The resulting Bouguer anomaly map is shown in Fig. 2.

BOUGUER MAP: DESCRIPTION AND INTERPRETATION The Bouguer anomaly map (Fig. 2) is characterized by: (1) A large negative anomaly, which trends approximately NW - SE (similar to the graben axis and the main structural features of the island) and consists of a sequence of lows separated by small gravity highs. From the structural point of view, the gravity lows are related to the thickest sectors of the graben Tertiary fili, prevalently Miocene - Pliocene sediments. The lows coincide with the areas in which the top of the Paleozoic basement is deeper. The maximum depth of the top of the basement is probably near S. Gavino. On the other hand, near , in correspondence to a small gravity high, the basement seems to be uplifted with respect to the structural low of the Gulf of Cagliari. (2) The contour lines on the western border are oriented parallel to the graben axis for the enti re region. The horizontal gravity gradients show significant variations that can be related to the graduaI stepwise morphology dipping towards the centrai part of the depression. (3) The isoanomalies on the eastern border are, also, generally oriented parallel to the graben and are regularly spaced; in the Monastir and areas, on the contrary, the gravity gradients are unusually high and in the southwestern boundaries of the gradient reaches a maximum. (4) A gravity low marks the wide Miocene sedimentary basin of Trexenta. The southern boundary of this low shows a gentle slope and, consequently, small gradients only. A fault and/or a flexure directed SW - NE and perpendicular to the main structural framework may exist in this area, forming the lateral contact between the andesitic volcanics and the Miocene marly formation.

REGIONAL AND LOCAL GRA VITY ANOMALIES The distinction between regional anomalies, related to the deep events, and local anomalies of shallow origin, is a useful means of defining the structural features outlined in the Bouguer map. This procedure also facilitates the qualitative analysis of the depth range of the different anomalies and simplifies their quantitative interpretation. The distinction between local and regional anomalies is a very delicate procedure in gravity studies; there are several mathematical and empirical mcthods available for this purpose and the choice is essentially subjective. We chose a two-dimensional filtering method which, according to several authors (Fuller, 1967; Bozzi Zadro and Caputo, 1968; Bath, 1974) seems to offer a reasonable degree of objectivity and is also very effective for anomaly resolution. On the basis of the generai shape and wavelengths of the anomaly shown in the Bouguer map, several filtering tests were made using the following cut-off frequencies: ro = 21t/5; 21t/1O; 21t/ 15; 21t120 krn'. Analysis and comparison of the various residual and regional gravity contour maps led to a qualitative evaluation. In the 21t/5 km " gravity residual map, the anomalies have no great structural significance because they refer to shallow sources and/ or to the superficial part of the deeper geological structures. The 21t/1O, 211:/15 and 211:120 krrr ' residual maps are rather similar as regards shape and size of the anomalies. The 21t/15 krn" residual map was thus chosen for the qualitative interpretation in order to analyse the characteristics and structural lines of the graben. The contour lines of the regional map (Fig. 3) 'rudu Gravity Survey of Campidano Area, Sardinia 339 was assessed up are regular and smooth, yet the wide negative anomaly whose centre is localized near S. Gavino ed and for the was filtered in part only. This gravity low suggests that the area around S. Gavino must be made iornaly is about up of a great thickness of sediments and a deepening of the basement in the northern and by a computer southern area of the graben. The positive gravity trend from the graben centre towards the SE idi et al, (1977), and NW is probably due to very deep structures. The residual map (Fig. 4) al so shows the wide low which characterizes the originai Bouguer map, forming an elongated NW - SE structure in the centrai part of the graben with a c\ear N - S distortion of the axis approximately between Villacidro and Samassi. Campidano area can, therefore, be divided into two seemingly different parts, the southern region having a more r to the graben complex gravity pattern. There is a sequence of lows and highs in the southern region that lows separated cannot be observed in the northern region of the graben: this different pattern seems to be ~ related to the caused by tectonic discontinuities running north - south, dividing the depression into a series of ients. The lows blocks which lower or raise the basement, with consequent thickening or thinning of the The maximum Tertiary voIcanic - sedimentary sequence. This hypothesis is supported by the magnetic ier hand, near discontinuities shown in the aeromagnetic map (Cassano et ai., 1979) and, in part, by the le uplifted with surficial geological evidence, The low in the eastern region, in particular, coincides with the elongated basin of the - marshes, whose Miocene formations, mainly .en axis for the represented by recent stratigraphic sequences, have very sloping beds, forming a sync\ine, an be related to lowered by a series of faults. ession. The gravity high between , Serramanna and Samassi could, on the basis of the above- el to the graben mentioned block model, be interpreted as a structural high. There is little geologic evidence, ry, the gravity however, except for some wide terraces on the Samassi Formation (this unit, here in Pliocene a the gradient continental facies, has a maximum thickness of 200 m and overlies the Miocene sediments). We could ascribe this gravity high to the absence of the Pliocene - Miocene sequence in some , The southern layers. Borehole Cl, drilled on the southern border of the high, encountered no Lower Pliocene s only. A fault or Upper Miocene deposits within the stratigraphic sequence of Villasor. Assuming that the ramework may Miocene sequence north of Villasor was subjected to an intense erosion before the deposition of id the Miocene the Samassi Formation, then the high could indicate an uplift of Lower Miocene deposits only. The block sequence between S'Acquacotta and Villasor presents a large low corresponding to a wide alluvional plain. Steep westward gravity gradients clearly reveal the S'Acquacotta fault. Two gravity highs north and south of correspond to the Paleozoic schistose !ocal anomalies outcrops, associated subordinately to andesites. The highest gradient value corresponds to the in the Bouguer fault related to the hydrothermal manifestation of S'Acquacotta (48°C; rich in CO,). The of the different morphological features of the western border of the graben, near Vallermosa, are generally a good reflection of the trend of isoanomalies. As a consequence the S'Acquacotta fault, .dure in gravity trending in a northerly direction south of the thermal spring, must also assume a NW - SE .is purpose and direction south of the Vallermosa parallel. nethod which, East of the Campidano graben there is a gravity high connected with the andesitic reliefs that . 1974) seems to form a vo1canic ridge along the eastern border from - Serrenti to Villagreca. The positive fesolution. anomaly is probably related to the Paleozoic basement which outcrops near in the Bouguer Villagreca - Nuraminis and is partially covered by a thin layer of vo1canics in the vicinity of this es: ro = 2n/5; area. Il and regional The steep gravity gradient between the Serrenti - Nuraminis low and the above-described sidual map, the high c\early reveals the eastern main fault of the graben and the local structure. rrces and/ or to The main feature of northern Campidano graben is the large NW - SE gravity low of S. nd 2n120 km " Gavino (25 km in length), which can be correlated with the generai geclogic framework of the 'he 21t/15 krn " graben, Further north, the main structural lines of the gr aben are directed N - S, as in the to analyse the southernmost zone. ial map (Fig. 3) The gravity highs in the area (west of the graben) are c\early related to the granitic- 340 R. Balia, M. Cimina/e, M. Loddo, G. Pecorini, G. Ruina and R. Trudu schistose basement. In addition, the small gravity gradients indicate a graduallowering of the basement towards the centre of the graben. A steep horizontal gradient near Sardara corresponds to the eastern main fault and the gravity high coincides with the outcrops of the schistose basement. The perfect alignment of the isoanomalies and of the E - W faults limiting the basement north and southwards has already been mentioned. Near the intersection between the main fauIt and the W - E fault, located north of S. Maria de Is Acquas, are some hydrothermal manifestations (55°C; rich in CO,). Finally, the gravity low near Mogoro may provide new data on the thickness of the Oligocene - Miocene and Pliocene sediments (Samassi formation) and on basement depth beneath the thin basaltic cover.

QUANTIT ATIVE INTERPRETA TION The elongated structure of the graben and the regular and parallel trend of the contour lines has suggested an interpretation in terms of two-dimensional masses. Due to the paucity of geophysical and structural data, particular1y with regard to the depth of the basement (never met by drilling) it is impossible to construct models that totally reflect the deep geologica I situation. We have therefore tried to make a direct interpretation using the spectral analysis technique, before constructing a model for defining depth and thickness of the various units. The spectral technique is very effective for studying the potential fields and estimating theaverage depth of the deep structure and large-scale discontinuities, such as the Moho, or the Curie isotherm, and for solving complex structures such as vo\canoes (Spector and Grant, 1970; Curtis and Jain, 1975; Shuey et al., 1977; Mishra and Pedersen, 1982; Negi et al., 1983). Assuming that the deep discontinuity interface has a statistical behaviour, i.e. random geometrical parameters, then the amplitude spectrum of the magnetic or gravity fields along a profile wilI decrease linearly, in ' a well-defined longwave band, with respect to the corresponding wave number on a log +Iinear representation. The slope of this straight line provides an estimate of the average discontinuity depth. The basement can be referred to these parametri c or statistical models. The various problems arising from this method are related, in particular, to estimates of amplitude spectrum and to the method itself. Pedersen (1978) debated in detail the different problems associated with the reliability of the discontinuity - depth estimate; Mishra and Pedersen (1982), analysing the method, established the spectral bandwidth within which we can accept the linear trend of the fit between the amplitude spectra and the wave numbers, and the influence of high frequencies on the estirnate. After constructing several theoretical models, the authors suggest the precautions and the expedients which must be used in various situations. In eur case, the amplitude spectra were computed along 24 parallel profiles (at a distance of 1250 m), approximately normal to the graben axis and 32 km in length. The first profile (1) is shown in Fig. 5. The gravity values were deduced from the Bouguer map, preliminarily filtered by a high-pass filter to remove the regional long-wavelength field connected to the deep crustal structure and the mantle. To obtain stable spectra a running averages procedure on the arnplitudes was used for five frequencies (two for each side 01' the selected frequency). Figure 6 shows the arnplitude spectra versus the wave number for ali the selected profiles. The amplitudes at low frequencies are c\osely fitted by a straight line, but at high frequencies the large departure from linearity impedes estirnate of the depth of the shallow discontinuities. Spectral analysis has enabled us to define only one density discontinuity. The variation in the slope of the straight lines, which fits the values, indicates a change in the average depth of the discontinuity extending from profile l to 24, or from SE to NW along the Campidano axis. The -Ah .•• :lèi&ta>, ••

Gravity Survey of Campidano Area, Sardinia 341 .f the

j the \ CAMPIDANO axil if the 'eady = cated 0-:::;- :0,). , the lepth

Iines ty of leVer -gical ique, .ctral th of , and Jain, o 5 io km dom mga the line Fig. 5. Gravity profiles in the Campidano graben: profile I is the first of 24 selected for spectral analysis; profiles AA', these BB' , CC', DD' , EE' were used in the two-dimensional models. d, in 978) the shed trend of the depth of diseontinuity along a SE - NW direetion, reported in Fig. 7 (AA curve), I the shows some undulation. The maximum (about 3000 m) is rea ehed in the neighbourhood of S. nate. Gavino, while the minimum (1400 m) is on the above-mentioned struetural high, near l the Serramanna. The results of spectral analysis were interpreted by means of a two-dimensional model (Talwani et al., 1959) whose eonstraints are: ce of (1) discontinuity depth, as determined by the spectral analysis, is assumed as the interface :1) is between the Miocene andesites and the sehistose crystalline basement, terèd (2) the stratigraphic structure of the graben is assumed as follows: (i) sedimentary ustal Oligo - Mio - Pliocene sequenee formation, (ii) andesitic tufaceous Oligo - Miocene the formation and (iii) granitic - metamorphic basement, .ire 6 (3) the densities in the Bouguer anomalies eomputation are obtained from roek samples The coming from cores of sedimentary and andesitic formations (well Cl) and from samples of 5 the the basement outerops. ities. The average density values are: l the 2.65 g/em' for basement, fthe 2.45 g/cm' for andesites and tuffs, The 2.10 g/cm' for sediments. Il 342 R. Balia, M. Cim inale, M. Loddo, G. Pecorini, G. Ruina and R. Trudu l ;~ ~\;, ~

~ ~r "" 7 r "" · r ~ · r \ ,. r \ 11 r \ 1Z ~o:~ ~ ~ l'-''"~~ i\;

~ -. ~. ~\~,~. • 17 ~ • ••••• •••• •••• t , ~ ". -1 •• , ", ••••• ••••• • ", ;

:1 \ 00 " 1 \ 00 •• 1 \ .. 1 \ o •• 1 \ 00 •• 1 \ 000 ••

.,~~~.~~~ ....

o " e 12 o " 8 12 o • e 12 o " 8 12 o 4 8 12 o 4 8 12

F R E a u E N C Y , z n /32 .km·'

Fig. 6. Arnplitude spectra versus wavenumber for the 24 selected profiles in Campidano graben. Profile l is shown in Fig. 5.

~, tr o « !? o c ~ e~ :li o ai ai

o 5km TEAT'ARV s • o -1000 c _._._._.~._.-._._._ •.••••.•_ I '" f N T S

/ti o .- _.

-· ....·-· .•..•...._._._ ..• c I C Q) -2000 SE I o N NW Q) E

• A , • o z -3000 o I C

• A S

Fig. 7. Section along Carnpidano axis, obtained by spectral analysis (AA curve) and two-dimensional models (BB and CC curves).

i. b Gravity Survey oJ Campidano Area, Sardinia 343 As a consequence, the density differences (with respect to that of the basement) considered in the computation procedure are: -0.20 g/cm' for andesites, -0.55 g/crri' for sediments. The quantitative interpretation was carri ed out along five selected profiles (AA' , BB' , CC' , DD' , EE' of Fig. 5) across the graben axis, using the Bouguer anomaly values, filtered by the regional long-wavelength field. These results are shown in Figs 8 - 12. The best fit interpretation models of the gravity profiles show a change in size and morphology of the depression along the five graben sections, particularly as regards the thickness of the sediments and of the andesites in the underlying basement. There also appears to be a correlation between the change in gravity gradients and the rather steep dip of the boundary faults. In any case, the gravity gradient over the western margin of the depression is less than that over the eastern margin, indicating that the eastern main fault is more steeply sloping than the western one. Figure 7 shows the discontinuity trend along the graben axis, which was clearly revealed by spectral analysis (AA curve), and also the andesites - basement interface (BB curve), and the sediments - andesites interface (CC curve), as obtained from the two-dimensional models. The agreement between curves AA and BB is as good as the fit in the two-dimensional profiles. Thus, the results from the two different interpretation methods can be deemed reliable and, as a consequence, both the models of Figs 8 - 12 and the trends of Fig. 7 probably correspond roughly to the structure of the depression. The models used for the gravity interpretation are simplified and do not consider some details (i.e. the very thin andesitic volcanics present along or near the borders of the graben). A more detailed interpretation is presented in Fig. 13. own in

-S"'" ~ « 20 ~:z « '0 A A' .:>... :;: o CI' "..., -10 ~ a~ -20~ ~

ct sw NE

o

DISTANCE (km I

Fig. 8. Observed and computed Bouguer anomalies and section across the Campidano graben along profile AA' (see Fig. 5). The solid line in the upper part of the figure is computed gravity and dots are the observed gravity values. The BB and section shows the Paleozoic basernent density and the density anomalies of sediments and underlying volcanics with respect to the basement (in g/crrr'). 344 R. Balia, M. Cimina/e, M. Loddo, G. Pecorini, G. Ruina and R. Trudu

;; ~'" >- 20 ~ o B 1 z 10 8 -c >- s o ~ " -10

~-20~------~ ~ ~ sw NE

O

10 15 20 25 30

DISTANCE (km) Fig. 9. Observed and computed Bouguer anomalies and section across the Campidano graben, along profile BB' (for Fig. Il other details see Fig. 8).

'" ~'" ~ -c 10 ::!; o C ,.'".. O :;>- -10 ~ ~ -20 § Cl ~ -30 Cl: SW NE

o E~

~c, u.. Cl

Fig. I DISTANCE (km) Fig. lO. Observed and computed Bouguer anomalies and section across the Campidano graben, along pro file CC' (for other details see Fig. 8). kno," bord CONCLUSIONS Parti The gravity study of the Campidano graben has confirmed the presence of a wide negative previ anomaly, in the long-wavelength, near S. Gavino, where the Tertiary volcanic - sedimentary geotl filling, overlyingthe Paleozoie basement, is probably about 3000 m thiek. The study has also anon , , defined the main strueturallineaments of the graben more preeisely, eonfirming those already outci du Gravity Survey of Campidano Area, Sardinia 345

,5~ ~ 10r------, ~ ~ o -c D s -lO Ct:'< (!) -J -20 2 ili -30L------~ Ct: N[

30 o io 15 20 25 30 DISTANCE (km) profile BB' (for Fig. Il. Observed and computed Bouguer anomalies and section across the Campidano graben, along pro file DD' (for other details see Fig. 8).

~"'" ~ 20 o~ <: lO "....~ s o ~ (!) -.1-10 ~ 9 :2-20 Q:

SW NE N[ G

I 1 •...J:: e, 2 "'Cl

3 o lO 15 20 25 30 ] DISTANCE (km) 30 Fig. 12. Observed and computed Bouguer anomalies and section across the Campidano graben, along profile EE' (for other details see Fig. 8). pro file CC' (for known or supposed and clarified some other features. The study has also shown that the eastern border is steeper than the western one, in close relationship with the gravity gradients. Particularly interesting is the residual positive anomaly between Villasor and Samassi, vide negative previously unknown, which has been interpreted as a rise of the basement. The most important .sedimentary geothermal manifestations in the Campidano are associated with the residual positive udy has also anomalies, along the border of the graben, where the Paleozoic granitic metamorphic basement .hose already outcrops. This can be observed both at S'Acquacotta, where the steep gravity gradients mark 346 R. Balia, M. Cimina/e, M. Loddo, G. Pecorini, G. Ruina and R. Trudu

D Cassano, I CAMPIOANO aeroma m 1000 Villacidro Moulltains Geal. I, Curtis, C. maps ir Dettori, B Sardegi Finzi Cont in Sard Fuller, B. Expl, ( Guerra, l. In. Att Hessing, F Geoph; Loddo, M In ut« Marchisio, risultar Martinis, : pp. 56: c A M p D A N o Mishra, D Boreho!e Serramanna Villasor Campidano1 tion l~ Negi, J. ( State ( Paia, A., Geotei Pala, A.; Geate/ Panichi, ( Pedersen, 43,94 Shuey, R spectr NW SE Spector, ' Talwani, Fig. 13. Generalized transverse and longitudinal geologic sections of the Carnpidano graben: (l) Paleozoic basement, applic (2) Eocene sandstones and clays (Cixerri formation, (3) Oligocene- Lower Miocene andesitic-tufaceous formation, (4) Trudu, R Upper Oligocene - Miocene - Lower Pliocene sediments, (5) andesitic la vas and tuffs (Lower Miocene), (6) Pliocene Geal. (Samassi Formation) and Quaternary sediments, (7) Pliocene basalt. Trudu, R Trudu, R Trudu, F 5-23 the western main fault, and at S. Maria de Is Acquas, where the eastern fault crosses the fault associated with the hot spring of the same name. The lack of other surface geothermal manifestations along the southern part of the eastern main fault (where there are Paleozoic outcrops related to steep gradients) does not exclude the possibility that geothermal fluids are kept at depth by the impermeability of the cover.

REFERENCES Bath, M. (1974) Spectra/ Analysis in Geophysics. Eisevier, Amsterdam. Bertorino, G., Caboi, R., Caredda, A. M., Cidu, R., Fanfani, L., Panichi, C., Sitzia, R. and Zuddas, P. (1982) Alcune considerazioni sulla geochimica delle acque termali del Graben del Campidano. In Ricerche Geatermiche in Sardegna, C.N.R. P.F.E. RF lO, pp. 133-143, Roma. Bolondi, G., Rocca, F. and Zanoletti, S. (1977) Methods for contouring irregularly spaced data. Geophys. Prosp. 25,96-119. Bozzi Zadr o, M. and Caputo, M. (1968) Filtri multidimensionali e loro applicazioni geofisiche. Ann. Geof. 3, 287 - 304. Caboi, R., Cidu, R.; Pala, A. and Pecorini, G. (1982) Le acque fredde della Sardegna. Lineamenti idrogeologici ed idrogeochimici. In Ricerche Geatermiche in Sardegna, C.N.R. P .F.E. RF IO, pp. 25 - 55, Roma. Caboi, R. and Noto, P. (1982) Dati isotopici sulle acque termali e fredde dell'area campidanese. In Ricerche Geotermiche in Sardegna, C.N.R. P.F.E. RF lO, pp. 124-132, Roma. Gravity Survey OJ Campidano Area, Sardinia 347

Cassano, E., Marcello, A., Nannini, R., Pretti, S., Ranieri, G., Salvaderi, R. and Salvadori, l. (1979) Rilievo DI aeromagnetico della Sardegna e del mare circostante. Carte di anomalie magnetiche in scala 1/250000. 80/1. Servo Geol. Ital. C, 199 - 234. Curtis, C. E. and Jain, S. (1975) Determination of volcanic thickness and underlying structures from aeromagnetic maps in the silet area of Algeria. Geophysics 40, 79 - 90. Dettori, B., Zanzari, A. R. and Zuddas, P. (1982) Le acque termali della Sardegna. In Ricerche Geotermiche in Sardegna, C.N.R. P.F.E. RF lO, pp. 56-86, Roma. Finzi Contini, G. (1982) Prospezioni magnetotelluriche effettuate sul graben del Campidano. In Ricerche Geotermiche in Sardegna, C.N.R. P.F.E. RF lO, pp. 164-180, Roma. Fuller, B. D. (1967) Two-dimensional frequency analysis and design of grid operators. In Mining Geophysics, Soc. Expl. Geophys., VoI. II, 658 -708. Guerra, l. (198 I) Grandi profili sismici: struttura crostale della Sardegna sulla base di dati sismici e gravimetrici. In. Atti I Convegno G.N.G. T.S., C.N.R., pp. 639 - 645, Roma. Hessing, R. C., Lee, H. K., Pierce, A. and Powers, E. N. (1972) Autornatic contouring using bicubic functions. Geophysics 37, 669 - 674. Loddo, M., Mongelli, F., Pecorini, G. and Tramacere, A. (1982) Prime misure di flusso di-calore in Sardegna. In Ricerche Geotermiche in Sardegna, C.N.R., P.F.E. RF IO, pp. 181-209, Roma. Marchisio, M., Ranieri, G., Losito, G. and Censini, G. (1982) Prospezioni geoelettriche dipolari nel Campidano: risultati preliminari. In Ricerche Geotermiche in Sardegna, C.N.R. P.F.E. RF IO, pp. 144-163, Roma. Martinis, B. (1969) Sardegna (Ricerca petrolifera). In Enciclopedia del petrolio e del gas naturale, VoI. VI, E.N.I., pp. 565 - 571, Italia. Mishra, D. C. and Pedersen, L. B. (1982) Statistical analysis of potentials fields from subsurface reliefs. Geoexplora- tion 19, 247 - 265. Negi, J. G., Agrawal, P. K. and Rao, N. N. (1983) Three-dirnensional model of the Koina area of Maharashtra State (India) based on the spectral analysis of aeromagnetic data. Geophysics 48, 964 - 974. Pala, A., Pecorini, G., Porcu, A. and Serra, S. (1982a) Schema geologico strutturale della Sardegna. In Ricerche Geotermiche in Sardegna, C.N.R. P.F.E. RF IO, pp. 7 - 24, Roma. Pala, A., Pecorini, G., Porcu, A. and Serra, S. (1982b) Geologia e idrogeologia del Carnpidano, In Ricerche Geotermiche in Sardegna, C.N.R. P.F.E. RF IO, pp. 87 -103, Roma. . Panichi, C. and Squarci, P. (1982) Carta delle temperature sotterranee in Italia. C.N.R. P.F.E. RF lO, Roma. Pedersen, L. B. (1978) A statistical analysis of potential fields using a vertical circular cylinder and a dike. Geophysics 43, 943 - 963. Shuey, R. T., Schellinger, D. K., Truppi, A. and Alley, L. B. (1977) Curie depth determination from aeromagnetic spectra. Geophys. J. R. Astr. Soc. 50, 75 -101. SE Spector, A. and Grant, F. S. (1970) Statistical models for interpreting aeromagnetic data. Geophysics 35,293 - 302. ment, Talwani, M., Worzel, J. L. and Landisman, M. (1959) Rapid gravity computations for two-dirnensional bodies with n, (4) application to the Mendocino submarine fracture zone. J. geophys. Res. 64,49 - 59. Dcene Trudu, R. (1953) Prime note di gravimetria sarda. Sezione gravi metrica nella pianura del Carnpidano. 80/1. Servo Geol. /tal. 75, 929 - 940. Trudu, R. (1961) Ricognizione magnetometrica nel Campidano Meridionale. Res. Ass. Min. Sar. LXV, 33-44. Trudu, R. (1962) Rilevamento gravimetrico della Sardegna. 80/1. Geof. Teor. Appl. IV, 299-399. Trudu, R. (1963) Prime considerazioni sul rilevamento gravimetrico della Sardegna. Res. Ass. Min. Sar. LXVII, 5 -23. 'ault tern ~the

cune le in

'osp, r. 3,

.i ed rche