Hydrogeological Processes in Terranes (Proceedings of the Antalya Symposium and Field Seminar, October 1990). _ IAHS Publ. no. 207, 1993. 107

FIRST RESULTS FROM THE MONITORING SYSTEM OF THE KARSTIC COMPLEX "GROTTE DI FRASASSI-GROTTA GRANDE DEL VENTO" (CENTRAL APENNINES, )

W. (V. U.) DRAGONI Earth Sciences Department, Perugia University, Piazza dell'Université, 06100 Perugia, Italy A. VERDACCHI Consorzio Frasassi, 60040 Genga (), Italy

ABSTRACT The karst complex of "Grotte di Frasassi-Grotta Grande del Vento" is located in the gorge cut by the River Sentino in the anticline of Mt. Valmontagnana, about 50 km from the town of Ancona (central Italy). In order to manage the in a rational way, and to get new information about the karstic processes at the "Grotta Grande del Vento", a computerized monitoring system was installed for temperature, humidity, rain, percolation and air velocity, inside and outside the cave complex. The first data collected suggest the following preliminary results: (a) Normally the flow of groundwater is towards the Sentino River. During floods this flow is reversed. The effect of the waters mixing can increase karstic dissolution. This hypothesis seems to be confirmed by the greater dimensions of the cavities close to the Sentino River, (b) As expected there is a close correlation between air flow through the cave and the temperature difference inside and outside the cave. However the data seem to show that in some zones of the caves the air flow is mainly controlled by the processes of condensation-evaporation, (c) The condensation phenomena probably play an important role in the karstic evolution of the system, (d) An initial estimation of the groundwater draining into the Sentino River along the Frasassi Gorge has been made (about 50 1/s); according to the Maillet equation the depletion constant of the river is 2.8 X 10"2 day"1, that of the aquifer is around 8.4 x 10"2 day"1.

INTRODUCTION In central Italy, about 50 km from Ancona, in the Frasassi Gorge, one of the most spectacular karst complexes in Europe has developed. The largest underground system in the area is that known as "Grotta Grande del Vento" (Great Cave of the Wind), the explored part of which extends for over 15 km.

The Grotta Grande del Vento system is economically quite important to the area: since being opened to tourists in 1974, approximately 400 000 have visited the cave each year. The part open to tourists is run by a public agency, the "Consorzio Frasassi", which has set up a Scientific Committee with the task of studying local karst processes and how they might be altered by the steady flow of tourists in the cave. For example the visitors alter the C02 content in the air of the cave and algae tend to grow in artificially lit areas, spoiling the best . The Committee is composed of small teams of scientists from various fields and is divided into three different sections: the first one deals with the chemical, physical and geological aspects of the problem (Physical Section), the second one is the Bio-Speleological Section; the third one is the Dating Section, with the task of dating the speleothems using isotopic methods.

In early November 1989 a continuous real-time monitoring system was installed for monitoring the physical data most relevant to the aims of the Scientific Committee. Up to now priority has been given to temperature, rainfall, percolation inside the caves, 108 W. (V. U.) Dragoni & A. Verdacchi atmospheric pressure, humidity, air velocity and carbon dioxide content in the air. We expect these data will make an important contribution towards the understanding of some of the phenomena taking place in a karst environment. In this paper the monitoring system will be briefly described, and some of the preliminary results inferred from the data obtained will be discussed.

GEOLOGICAL AND GEOMORPHIC FEATURES Figure 1 gives a simplified geological sketch of the area where the Frasassi Gorge is located. Here the geological situation has been synthesized. The reference list includes many entries which give information on the Gorge.

The Grotta Grande del Vento cave system has developed entirely within the Mt. Valmontagnana anticline core, which is situated on the western side of the Umbria- ridge. This anticline has an Apennine trend (northwest-southeast) with an Adriatic virgence (northwest). Towards the west it joins with the Camerino syncline, whereas towards the east it crosses the southern edge of the Pergola syncline.

, tssl 5 Do • 7 Ho • '9 HlO 11 12 —1— 13 0 14 O 15 • 16 Fig. 1 - Geological sketch of the Frasassi area. 1: Calcare Massiccio; 2: Bugarone and heteropic formations.; 3: Maiolica ; 4: Marne a Fucoidi; 5: Scaglia Bianca e Rosata; 6: Scaglia Cinerea e Variegata; 7: Bisciaro; 8: Schlier; 9: alluvial deposits; 10: debris and landslides; 11: fault; 12: uncertain fault; 13: anticline axis; 14: syncline axis; 15: major sulphatic spring; 16: beginning and end of the Frasassi Gorge. The monitoring system of the "Grotte di Frasassi-Grotta Grande del Vento" 109

The thickness of the sediments making up the Mt. Valmontagnana anticline is less than that of the adjacent synclines, due to settling in low depth conditions over a "structurally lifted block" existing during the entire Jurassic period. This brought about two different sequences in the sedimentation area: the Mt. Valmontagnana sequence (condensed sequence), and that of the adjacent synclines (complete sequence).

Triassic evaporitic formations ("Formazione di Burano"?), which do not outcrop in the Frasassi area, should be present below both sequences. Both sequences begin with the limestone formation of "Calcare Massiccio" (Hettangian-Sinemurian) which is made up of beds ranging in thickness from about 1 m to 10 m. This formation has a total thickness of between 500 and 700 m. The uppermost portion of the Calcare Massiccio has about 80 m of typical oolitic facies (Colacicchi & Pialli, 1974). On a large scale the Calcare Massiccio is the most permeable formation in the area. Below the degree of permeability of the geological formations is given with respect to the permeability of the Calcare Massiccio. In the condensed sequence, from Lias to Tithonian, the deposition of the "Bugarone" formation took place, after the Calcare Massiccio, in conditions of a "structurally lifted block". This formation is composed of thin beds of limestone nodules, reaches a maximum thickness of 50 m (Centamore et al., 1975), and, on the whole, can be considered as having medium permeability. The other sequence, deposited in pelagic basins, is heteropic with the Bugarone, about 450 m thick, and composed of limestone, marly-limestone, and cherty-limestone (from Sinemurian p.p. to Lower Tithonian), with an overall medium-low permeability.

From the Tithonian on, both sequences continue with the typical Umbria-Marche sequence: — "Maiolica" (Lower Tithonian p.p.-Aptian), made up of bedded limestone, with medium-high permeability; — "Marne a Fucoidi" (Aptian p.p.-Cenomanian), consisting of marls, with very low permeability; — "Scaglia bianca e rosata" (Middle Cenomanian-Middle Eocene p.p.), made up of limestone and marly-limestone in the upper portion, with medium-high permeability; — "Scaglia variegata e Cinerea" (Middle Eocene p.p.- Oligocène), made up of marls, marly-limestone, with low permeability; — "Bisciaro" (Lower Miocene), made up of marls and silty-clays, with very low permeability; — "Schlier" (Lower Miocene), made up of marls, silty-marls and clays, with very low permeability. Although there is a certain amount of tectonic activity in the Frasassi area at the present time (Forti & Postpischl, 1979; Centamore et al., 1978), the features of the zone are essentially due to past tectonic stages; in particular, conjugate strike-slip systems were most likely formed in the late Pliocene, the dextral faults striking NNW-SSE and the sinistral faults northeast-southwest. The latter are definitely prevalent (Menichetti, 1988; Cocchioni et al., 1988).

The north-south system of inverse faults existed before the strike-slip systems. The Sentino stream flows through the northern part of the Mt. Valmontagnana anticline, cutting deeply into the Calcare Massiccio core (Fig. 1). The latter is the formation which is most affected by karst phenomena, although they also occur in the Maiolica and in the Bugarone (Coltorti & Galdenzi, 1982). On the basis of the morphology of Frasassi Gorge, it appears that the deepening of the Sentino River into the Mesozoic formations underwent a progressive migration towards the north. 110 W. (V. U. ) Dragoni & A. Verdacchi

Fig. 2 - Map of the karstic system "Grotta Grande del Vento-Grotta del Fiume". The dimensions of the cavities increase close to the river; the main directions are parallel to the main fault systems.

The karst phenomena, enhanced by some primary porosity in the Calcare Massiccio, developed accordingly to pre-existing tectonic features, such as faults and joints (Figs 2 and 3). At the Frasassi Gorge at least four main karst plains can be recognized (Cattuto & Passeri, 1972; Cattuto, 1976); according to a more recent study there seem to be eight plains (Antinori, 1979). These plains are often matched by particular external morphologic features such as terraces, orographic benches, etc. This shows that the formation of the gorge was characterized by stasis periods.

Two types of groundwater are present in the area. One has a bicarbonate chemical composition, originating essentially from in loco infiltration; it floats in places on a water with sulphate-chloride composition. This second water, which acquires its chemical composition from contact with the Triassic evaporitic formations, belongs to a confined

Fig. 3 - Map of the karstic system "Grotta di Frasassi-Mezzogiorno". The cave is located in the Frasassi Gorge, at the same altitude the upper levels of Grotta Grande del Vento. Note the parallelism of the main directions of the cavities and how the dimensions of the cavities increase close to the paleo rivers. The monitoring system of the "Grotte di Frasassi-Grotta Grande del Vento" 111 aquifer. This aquifer, along the axis of the Mt. Valmontagnana anticline and in the presence of trans-tensive faults, becomes unconfined. Both waters are drained by the Sentino, and some sulphate-chloride springs are present along the Frasassi Gorge. The karst phenomena of the Frasassi Gorge are largely controlled by the interaction between the two waters and by the presence of H2S (Cucchi & Forti, 1988).

CLIMATIC AND HYDROLOGICAL CHARACTERISTICS The Sentino basin, at the end of the Frasassi Gorge (Fig. 1), has an area of 285 km2, with an average altitude of 495 m a.s.l. The average precipitation in the area is about 1200 mm/year, and the average temperature is approximately 13.5°C. Real évapotranspiration according to the Turc formula averages 650 mm/year. Precipitation reaches peaks in November and April, while July is the driest month. The greatest infiltration, strictly tied to water excess, takes place during November, December and January.

At the present time, continuous measurements are not being taken of the flow of the Sentino River. Daily measurements taken between 1926 and 1929 at the end of the gorge, give an average flow of approximately 7 m3/s, corresponding to about 775 mm/year over the entire basin, and to a runoff coefficient of 0.67. This latter value is considerably higher than that obtained using the Turc formula, on the hypothesis that the hydrogeological basin coincides with the catchments. It should be emphasized, however, that the average rainfall in the 1926-1929 period was about 300 mm higher than the average precipitation, therefore only further investigations may make clear how much this high runoff coefficient depends on underground contributions from neighbouring basins and not on abnormal pluviométrie conditions.

As regards the sulphate waters appearing in the gorge, Cigna & Giogelli (1989) proposed two possible underground permanence times, 5 or 35 years, based on isotopic methods.

Measurements taken using both a current meter and chemical methods show that the overall flow of the sulphate waters in the Frasassi Gorge is about 50 1/s, which is much higher than the 27 1/s estimated by Perrone (1911). Our measurements were taken during a dry period in the winter of 1990 when the average flow of the Sentino River at the end of Frasassi Gorge was approximately 1 m3/s. According to our measurements the depletion coefficient of the Sentino River, according to the Maillet equation, should be about 2.8 X 10 2 day"1, that of the sulphate aquifer 8.4 x 10"3 day"1.

THE MONITORING SYSTEM Attempts were made in the past to take continuous measurements of some parameters in the Frasassi caves, such as temperature, humidity, and C02 level in the air. However, until 1988 the instrumentation consisted of rather delicate mechanical systems (recorded on paper) that made it difficult to obtain reliable and continuous data. In 1989 a new, entirely computerized, system was set up for continuous, real-time monitoring of the main underground physical parameters, with magnetic data storage. Figures 4 and 5 show the planned system.

The completed system will consist of 33 sensors run by two TECA concentrator stations (small computers which organize and record sensor data a first time). In turn the TECA concentrators are connected to a computer which automatically processes and records the data. The use of the TECA stations was necessary by the distances, in some cases more 112 W. (V. U.) Dragoni & A. Verdacchi

Fig. 4 - Sketch of the monitoring system. 1: TECA concentrator; 2: hygrometer; 3: water thermometer; 4: pluviometer; 5: barometer; 6: psychrometer; 7: carbacidometer; 8: thermohygrometer; 9: tacheoanemometer; 10: air thermometer. than 1000 m, between the sensors and the computer. The voltage losses deriving from this would have compromised the reliability of the measurements.

All sensors were appropriately chosen for the environment in which they would be operating: underground climatic conditions are a considerable handicap for electronic systems, especially the humidity, which is almost always close to 100%.

The fact that each TECA concentrator can handle around 60 sensors makes expansion of the entire system easily possible. At the present time 23 of the 33 planned sensors are in operation. The sensors at present activated are:

six air temperature thermometers (sensitivity 0.01 degrees); two water temperature thermometers (sensitivity 0.01 degrees); six hygrometers (sensitivity 0.01%); two raingauges (sensitivity 0.25 mm); five tacheoanemometers (sensitivity 0.1 mm/s); two barometers (0.1 millibar).

Fig. 5 - Plan of the monitored portion of the Grotta Grande del Vento system (for explanation see Fig. 4). The monitoring system of the "Grotte di Frasassi-Grotta Grande del Vento" 113

In the next few months another 10 sensors should be put into operation for measuring C02 in the air of the cave. In the future it is hoped that gauges will be installed for measuring the water table level and the flow of the Sentino.

All the instruments have to be calibrated periodically to prevent problems deriving from instrumental inaccuracy. Up to now it was not possible to calibrate continuously all the sensors; however the data on which the following results are based have been recorded by calibrated instruments and are reliable.

INITIAL DATA AND IMPLICATIONS Internal meteorology and karst phenomena

The data on temperature, air velocity and direction have provided interesting indications as to the meteorology of the system and the development of karst phenomena. As usual in caves fairly near the surface, the average internal temperature is more or less constant and close to the average external temperature. For caves with openings at different heights there are two types of air circulation. Using T; for internal temperature and Te for external temperature, we have:

(a) if T; > Te, air enters from the lower openings and is warmed. This warming causes a decrease in density, therefore triggering a flow of air from the lower openings towards the upper openings. This situation is common in winter, and thus is conventionally called "winter circulation" (Fig. 6A).

(b) if T; < Te , air enters from the upper openings and is cooled. This cooling causes an increase in density, therefore triggering a flow of air from the upper openings towards the lower openings. This situation is common in summer, and thus is conventionally called "summer circulation". The two situations outlined above may take place within a 24-h span, when the day and night temperatures are alternately higher and lower than the internal temperature; this generally takes place in spring and autumn (Fig. 7).

Bi D! Ha ^4

Fig. 6 - Air circulation in the Grotta Grande del Vento cave. A: Winter- type circulation; B: summer-type circulation. 1: oversaturated area; 2: condensation zone; 3: underground water; 4: direction of air circulation; T;, Te: temperature measurement points (see Fig. 7); Pt, P2: air speed measurement points. 114 W. (V. U.) Dragoni & A. Verdacchi

Fig. 7 - Air flow and temperature of the Grotta Grande del Vento cave system. Data were recorded at minute intervals between 18 and 21 March 1990. T; = internal temperature (°C), Te = external temperature (°C), V^ = air speed (cm/s) at point Fl of Fig. 6, V2 = air speed at point P2 of Fig. 6. Up and down arrows show the air flow direction.

This outline does not take into consideration other factors influencing the flow of air in a cave, such as the velocity and direction of the external wind with respect to the openings, the variations in barometric pressure outside, and the effect of evaporation and condensation phenomena inside the cave. As regards the last two factors in particular, it should be kept in mind that, with other conditions being the same, humid air is less dense than dry air. This implies that, with other conditions being the same, a mass of humid air tends to rise when immersed in a mass of dry air. Thus, in a cave with two openings at different heights, a liquid water surface inside, the same temperature inside and outside and undersaturated air outside, a rising current is set off (Castellani & Dragoni, 1986).

Evaporation-condensation factors inside caves have been studied by various authors, who, although having little available data, have hypothesized as to the importance of such factors in the development of karst phenomena (for example, cf. Castellani & Dragoni, 1982, 1987; Cigna & Forti, 1986). The data and observations currently available at Frasassi seem to confirm these hypotheses, adding new elements.

Figure 6 is a conceptual sketch of the Frasassi system, topographically too complex to be shown here exactly. Figure 7 gives the internal and external temperatures and the air velocity registered in the period between 18 and 21 March 1990. Figure 7 gives the air temperatures; the air velocity through an upper opening, corresponding to passage P t in Fig. 6; and the air velocity recorded in an internal shaft (Falconara shaft) at about 450 m from the lower opening, corresponding to point P2 in Fig. 6, and also shown in Figs 2 and 4.

Figure 7 shows that the Grotta Grande del Vento cave system, the behaviour of the air flow in the external zones of the caves is different from that in the internal zones. As The monitoring system of the "Grotte di Frasassi-Grotta Grande del Vento" 115

expected in the external zones the air flow is controlled mainly by the value and sign of the difference (T; - Te). In the internal-lower zones the air flow is controlled mainly by the humidity-evaporation effect. The air flow is always ascending and the thermal effect only slows down the air velocity, without inversion. Probably only when the values (Te — T;) are bigger than a threshold value greater than 8°-9°C "summer circulation" reaches the sensor P2 at the Falconara shaft.

It is interesting to underline that in any of the cases there can be condensation on the cave walls. During "winter circulation" (Fig. 6A) the air rises, expands and cools. Condensation also takes place, obviously, in the upper openings near the outside, where the temperature is lower than the internal temperature. During "summer circulation" condensation could happen due to the internal ascending circulation and because the flow coming from outside could cool below dew point (Fig. 6B). The presence of condensation phenomena are suggested by the value of the relative humidity, which is always close to 100%, and by the frequent formation of fog in the upper part of the cave.

The evaporation of the internal lakes of the cave is indicated by the fact that the surface temperature of the internal lakes is generally lower than the air temperature by almost a tenth of a degree. Furthermore the air temperature in different points of the cave system shows that the innermost lower zone is constantly colder than the outer zones (about 1°C). This can be explained by admitting that the innermost lower zone is affected essentially by evaporation, which is an endothermic process; from this point of view, the peripheral upper-middle areas, where usually condensation should take place, are warmer due to the liberation of latent condensation heat.

The condensed water is aggressive against carbonates given the presence of C02 and H2S, so that it is difficult to deny the importance of condensation in post-phreatic karst processes.

Flood events and karst Data on the temperature of the aquifer have brought to light an aspect of the processes present in the Frasassi system, which until now has been overlooked. It was said previously that in ordinary conditions the Sentino drains the groundwater from the Mt. Valmontagnana limestone massif. During floods, however, water from the river temporarily feeds the aquifer. Figure 8 gives data on the temperature of the water inside the cave, at a linear distance of about 450 m from the Sentino River, during a flood which took place in November 1989; the same figure gives the recorded rainfall. During the flood the external water, colder than the internal water, was able to lower the temperature of the groundwater by up to 10°C in six days. As it is known, the mixture of two waters with different C02 contents creates an undersaturated mix, even if both are saturated in carbonates; the effect is increased when the two waters are at different temperatures (Bogli, 1960a,b).

It should be pointed out that the available data indicate a stratification of the C02 at Frasassi in the lowest levels of the cave, with concentrations in the air up to 1000 ppm (Castellani, 1989); a cooling of the water implies an increase in the C02 content of the water with a subsequent increase in the aggressiveness against carbonates. All of this suggests that flooding contributes considerably to the development of karst phenomena. This appears to be confirmed by the fact that, going towards the river, the dimensions of the caves get larger and larger (Figs 2 and 3). 116 W. (V. V.) Dragoni & A. Verdacchi

Fig. 8 - Temperature of groundwater in the cave at a point 450 m from the Sentino River (November 1989). 1: Rainfall in millimetres; 2: Temperature of the water.

FINAL REMARKS

The data collected so far have allowed us to confirm some hypotheses made in the past, and have also offered the possibility of pointing out some aspects of karst phenomena which would not have been observable with less sophisticated equipment. In the first place, we were able to better define the mechanisms and the characteristics of the circulation of air in the "Grotta grande del Vento" cave system. In this context, it was possible to indicate the areas where condensation may take place, according to the different types of circulation. The result of this is important as the condensation waters are aggressive. In this light, one could ascribe to their action the formation of particular karst features such as inverse gutters, mount milk, etc. often encountered at the Frasassi karst systems.

The data suggest the existence of a particular karst factor due to the action of the Sentino River flood waters. The development of the cave dimensions within the Frasassi karst complex, with respect to the distance from the river, seems to prove this hypothesis.

The results reached up to this point are qualitative and preliminary. The completion and final setting up of the monitoring system should permit more definitive conclusions and the making of quantitative models, possibly with results of general interest.

REFERENCES Antinori, A., 1979, Studio geologico-strutturale dell'area carsica di Frasassi. Tesi di Laurea, Univ.di Camerino. The monitoring system of the "Grotte di Frasassi-Grotta Grande del Vento" 117

Bogli, A., 1960a, Kalklosung und karrenbildung. Z. Geomorphol., suppl. 2, 4-21. Bogli, A., 1960b, Les phases de dissolution du calcaire et leur importance pour les problèmes karstiques. Rass. Speleol. It., 12, 167-180. Castellani, V. & Dragoni, W. (V. U), 1982, About the genesis of karstic cavities. 2° International Symposium on Utilization on Karstic Areas (Bari, May 1982). Castellani, V. & Dragoni, W. (V. U.), 1986, Evidence for karstic mechanism involved in the evolution of Moroccan hamadas. /. Speleol., 15, 57-61. Castellani, V. & Dragoni, W. (V. U.), 1987, Some consideration regarding karstic evolution of desert limestone plateaux. International Geomorphology 1986, pt. II, 1199-1205. Castellani, V., 1989, Frasassi e speleomonitoraggio. Speleologia, 18, 33-35. Cattuto, C, 1976, Correlazione tra piani carsici ipogei e terrazzi fluviali nella valle del fiume Esino (Marche). Boll. Soc. Geol. It., 95, 147-160. Cattuto, C. & Passeri, L., 1972, Relazioni tra idrologia carsica e litologia nell'area umbro-marchigiana. Acta Cong. Naz. Spel., XI, vol. 1, 227-238. Centamore, E., Catenacci, V., Chiocchini, M., Chicchini, V., Jacobacci, A., Martelli, G., Micarelli, A. & Valletta, M., 1975, Carta geologica d'ltalia: 1:50.000, Note illustrative del foglio 291 Pergola, 1-40. Centamore, E., Deiana, G., Dramis, F., Micarelli, A. Carloni, G. C, Francavilla, F., Nesci, O. & Moretti, A., 1978, Dati preliminari sulla neotettonica dei fogli 116 (Gubbio): 123 (Assisi), etc. Contributi preliminari alia realizzazione della carta neotettonica d' Italia, 113-148. CNR. Cigna, A. & Forti, P., 1986, The speleogenetic role of air flow caused by convection. Int. J. Speleol., 15, 41-52. Cigna, A. & Giorgelli, F., 1989, Underground water dating by tritium measurement. Ac. Int. Spel. Cong. (Budapest). Colacicchi, R. & Pialli, G., 1974, Significato paleogeografico di alcuni depositi nella parte sommitale del Calcare Massiccio (Nota preliminare). Boll. Soc. Geol. It., 92 (suppl), 173-187. Coltorti, M. & Galdenzi, S., 1982, Geomorfologia del complesso carsico: "Grotta del Mezzogiorno"(4 MA-AN) con riferimento ai motivi neotettonici, etc. Studi Geologici Camerti, VII, 123-132. Cocchioni, M., Coltorti, M., Dramis, F., Mariani, F. & Tazioli, G. F., 1988, Circolazione idrica e chimismo delle acque sotterranee dell'area carsica di Frasassi nelle Marche. Ac. Nat. Cong: "Carsismo nella gola di Frasassi", preprint. Cucchi, F. & Forti, P., 1988, Evoluzione speleogenetica del complesso carsico: "G. G. del Vento- G. del Fiume", S. Vittore, Marche. Ac. Nat. Cong. "Carsismo nella gola di Frasassi", preprint. Forti, P. & Postpischl, D., 1979, Determinazioni di dati neotettonici da analisi di concrezioni alabastrine. I e II contrib. alia Carta Neotettonica d'ltalia C.N.R., pp. 1399-1409, 634-644. CNR. Menichetti, M., 1988, Influenze strutturali nello sviluppo del carsismo della gola di Frasassi, (Appennino marchigiano). Ac. Nat. Cong. "Carsismo nella gola di Frasassi", preprint. Perrone, A., 1911, Memorie illustrative della carta idrografica d'ltalia: N.35, p. 203.