We revised the concept of convective, intrusive geothermal plays, considering that the tectonic setting is not, in our opinion, a discriminant parameter suitable for a classification. We analysed and compared four case studies: (i) Larderello (Italy), (ii) Mt Amiata (Italy), (iii) The Geysers (USA) and (iv) Kizildere (Turkey). The tectonic settings of these geothermal systems are different and a matter of debate, so it is hard to use this parameter, and the results of classification are ambiguous. We suggest a classification based on the age and nature of the heat source and the related hydrothermal circulation. Finally we propose to distinguish the convective geothermal plays as volcanic, young intrusive and amagmatic.
Geothermal energy is a renewable resource suitable for baseload power production. Various countries are working toward an increase of geothermal exploitation and research development. Although geothermal energy has been exploited for many decades in many countries, a clear and unique classification of geothermal systems has not been accepted worldwide, probably due to the strong variability of geological, geophysical and thermodynamic conditions. In the past, many authors proposed a classification of geothermal systems and resources, based mainly on temperature (e.g. Muffler, 1979; Sanyal et al., 2005). Recently Moeck et al. (2014) proposed an alternative scheme to classify geothermal systems, in the frame of “geothermal plays”, based on geological characteristics. The “play” concept hails from oil and gas exploration and corresponds to a “…model in the mind of the geologist of how a number of geological factors might combine to produce petroleum accumulation in a specific stratigraphic level of a basin” (Allen P. A. and Allen J. R., 2005). It is hard to import this concept to geothermal exploration due to the possible development of geothermal systems in many geodynamic settings with extremely various geological characteristics worldwide. On the other hand, we agree with Moeck (2014) on the need for a clear and widely accepted new catalogue of geothermal plays to support geothermal exploration activities at least in their very first activities. By merging different opinions and scientific discussions during a recent workshop held by the International Geothermal Association (IGA) in Essen, Germany (IGA, 2013), classification of the geothermal plays has been attempted as follows:
convective, volcanic field, divergent margins; convective, volcanic field, convergent margins; convective, intrusive, extensional; convective, intrusive, convergent; convective, extensional domains fault controlled; conductive, intracratonic basin; conductive, foreland basin/orogenic belt; conductive, basement (igneous and metamorphic).
In this paper we analyse and discuss the structural setting, the heat source and the reservoir characteristics of four important geothermal fields in exploitation over decades: (i) Larderello (Italy), (ii) Mt Amiata (Italy), (iii) The Geysers (USA) and (iv) Kizildere (Turkey). We classify them as convective and intrusive play types, and we stress the similar geological features that could depict this type of play.
The understanding of common features of convective and intrusive plays is important since they host some of the most productive geothermal fields in the world. The Larderello and Mt Amiata fields, both located in Tuscany (Italy), are two large convective geothermal systems with similarities but also many differences. Larderello is one of the few vapour-dominated systems worldwide, where the first geothermal power plant was installed in 1913. The Mt Amiata geothermal area is located close to the homonymous extinct volcano (0.3–0.2 Myr) and is characterized by a liquid-dominated system. The Geysers field is located in California (USA) close to the Clear Lake volcanic field and is the most productive vapour-dominated geothermal system in the world; it has been exploited since the 1960s. The Kizildere field, located in western Turkey, is a liquid-dominated system exploited for power production since 1984, the first field in Turkey. In our analysis we include also Kizildere, which has been differently classified (IGA and IFC, 2014, and references therein), since the aim of our paper is to open a discussion and comparison between the different proposed models.
We argue that the prospective resources are hardly classified on the basis of both tectonic setting and stratigraphic features, and we propose a new classification.
The geothermal fields of Larderello and Mt Amiata (southern Tuscany, Italy) are located in the inner part of the Northern Appennines, a sector of the Apennine orogenic belt developed as a consequence of the Cenozoic collision between the European (Corso–Sardinian block) and the Adria plates (Boccaletti et al., 2011). Southern Tuscany is characterized by a shallow Moho discontinuity (20–25 km depth), a reduced lithosphere thickness due to uprising asthenosphere and the delamination of crustal lithosphere (Gianelli, 2008). Many authors proposed a tectonic evolution of the Northern Apennines due to two main deformational processes: (i) a first one related to eastward-migrating compressional tectonics and (ii) a subsequent extensional tectonics migrating eastward which has been affecting the inner part of the orogenic belt since at least the early Miocene (Carmignani et al., 1994; Jolivet et al., 1998; Brogi, 2006, and reference therein). Alternative models have been proposed to describe the tectonic evolution of the inner Northern Apennines (Boccaletti et al., 1997; Bonini and Sani, 2002). These studies revealed a complex tectonic evolution during the Miocene–Pleistocene with alternating compressive and extensional tectonics events but suggest a prevalent contribution of compressive tectonics till the Pleistocene epoch, in contrast with an uninterrupted regional extensional tectonics active since at least the early Miocene as suggested by other authors. After the Pleistocene, southern Tuscany is characterized by active extensional tectonics as inferred from borehole breakout analysis (Montone et al., 2012). Considering Quaternary tectonics, recent studies have suggested an important role of strike-slip faults and step-over zones controlling the magma emplacement in the inner Northern Apennines (Acocella et al., 2006) and in the Mt Amiata area (Brogi and Fabbrini, 2009). Batini et al. (1985) presented a seismological study of the Larderello area, showing an intense seismic activity of low magnitude, partially induced, that could be correlated with seismically active structures.
The Geysers–Clear Lake geothermal field is located in northern California, between the San Andreas fault system and the Coast Range thrust (Stanley and Rodriguez al., 1995). This region belongs to the California Coastal Ranges, and its geological features are a consequence of the eastward subduction of the Farallon oceanic plate underneath the North America Plate since late Mesozoic times. The tectonic evolution of the region is quite complex. The late Mesozoic subduction system along western North America was replaced, in the Eocene period, by the Mendocino Triple Junction, which evolved in the San Andreas transform system (Stanley and Rodriguez, 1995).
The Geysers geothermal field is located between NW-trending right-lateral strike-slip faults that belong to the San Andreas fault system and exhibits normal and strike-slip faulting (Boyle et al., 2013). The analysis of seismicity (Oppenheimer, 1986; Boyle et al., 2013) indicates that most of the fault plane solutions show an extensional and strike-slip component. However, above 1 km depth, a reverse component is present.
The Kizildere geothermal field is located in the Denizli and Aydin provinces of western Turkey in the easternmost part of the Büyük Menderes Graben. The western Anatolian horst-and-graben system forms the eastern boundary of the Aegean extensional system, which is one of the most active extensional regions in the world and is undergoing a N–S extension (Gürer et al., 2009, and reference therein). The west Anatolian–Aegean area underwent continental collisions which started in the Mesozoic with a collisional zone migrating southward, down to the present position of the Cyprus–Hellenic subduction zone. The differential velocity fields of plates involved in the western Anatolia–Aegean region may explain the opening of the Aegean extensional system (Doglioni et al., 2002). The Büyük Menderes Graben is about 140 km long and up to 14 km wide, and approximately trends E–W in the Kizildere area. The main normal fault, bounding the northern margin of the graben, terminates eastward close to the Kizildere geothermal field where a horse-tailing termination could facilitate the hydrothermal circulation (Faulds et al., 2009). The Büyük Menderes Graben is characterized by intense seismicity, mainly concentrated in the Çameli–Denizli district close to the geothermal field and dominated by low-magnitude seismic swarms (Süer et al., 2010).
The geodynamic setting and the magmatic activity produce a huge geothermal
anomaly in southern Tuscany, with maximum peaks centred in the Larderello
and Mt Amiata areas with values of heat flow up to 1000 mW m
Mt Amiata is a young (0.3–0.2 Myr) extinct volcano belonging to the TMP made up of trachytes, trachylatites and olivine-latites (Gianelli, 2008). The volcanic edifice hosts an important reservoir of cold and drinkable water and overlies impermeable, clayey units. As for Larderello, the high-temperature hydrothermal circulation occurs in two deep-seated non-volcanic reservoir. Major bodies of intrusive rocks were never crossed by deep wells in Mt Amiata, but the heat source may be related to shallow intrusions inferred from geophysical data (Bernabini et al., 1995; Manzella et al., 1998; Finetti, 2006). This allows us to consider this geothermal play as intrusive.
The heat source of The Geysers geothermal field corresponds to a Quaternary
pluton complex (> 100 km
Both the huge vapour-dominated reservoir and the upper portion of the felsite are oriented NW–SE, sub-parallel to the right-lateral San Andreas Fault and related wrench faults. In fact, Hulen and Norton (2000) considered the emplacement of the felsite to be possibly related to pull-apart extension. The presence of a batholith or multiple silicic magma chambers at depth are supported by geophysical evidence, but a shallow intrusion cyclically replenished by new magma (at least 500 000 years each) is required to keep the present-day heat flow and thermal anomaly (Erkan et al., 2005).
Surface heat flow in western Turkey depicts wide thermal anomalies with
values up to 150 mW m
Faulds et al. (2009) compared the western Turkey region to the western Great Basin undergoing significant extension and relatively sparse volcanism and excluded a magmatic heat source at upper crustal levels for the geothermal systems in the area. The primary control of structural features accommodating deep hydrothermal circulation is therefore suggested.
On the other hand, geochemical and isotopic analyses of C, S and B (Özgür, 2002; Simsek, 2003; Özgür and Karamenderesi, 2015) support the hypothesis of relatively shallow and recent magmatic intrusion. On the basis of helium isotopic data, Güleç and Hilton (2006) suggest the occurrence of plutonic activity underneath the Büyük Menderes Graben. This is consistent with the fact that during the late Miocene to Quaternary an oceanic-island basalt (OIB)-type volcanism occurs during the recent extensional phase in the Anatolian–Aegean region (Agostini et al., 2007). The most recent Quaternary volcanic products are found in the Kula region located about 65 km NW of Kizildere field.
The C
Considering that young intrusion is inferred in Kizildere and remembering that intrusive rocks below Larderello and The Geysers were disputed and not clearly proved until these rocks were reached by deep drilling, we include Kizildere in the discussion of intrusive plays.
There are differences and similarities between the Larderello and Mt Amiata
geothermal reservoirs. Both areas host two reservoirs, the shallow being
hosted in sedimentary units and the deep in crystalline rocks. At Larderello
superheated steam is present at depths over 3.5 km and with temperatures
exceeding 350
Barelli et al. (2010) highlight that the shallow and deep reservoirs of the Mt Amiata system are in piezometric equilibrium as pointed out by the hydrostatic pressure distribution.
Thermal springs and diffuse gas discharge are abundant in both Larderello and Amiata geothermal fields and surrounding areas, with fierce manifestations in Larderello (Duchi et al., 1986; Minissale et al., 1991, 1997; Frondini et al., 2009).
At The Geysers the geothermal fluids are hosted principally by highly
deformed late-Mesozoic-age subduction-trench-related metasedimentary and
meta-igneous rocks of the Franciscan complex. The system is disrupted by
high-angle, generally northwest-trending faults related to the still-active
San Andreas Fault and low- to moderate-angle thrust faults. The Franciscan
rocks at The Geysers are intruded by a northwest-trending Plio–Pleistocene
multi-phase felsic pluton, which actually hosts a portion of the steam
reservoir and underwent further mineral recrystallization due to the
intrusion and related fluids. The configurations of the felsite and
reservoir coincide, strongly suggesting that the intrusion critically
influenced steam-field evolution (Hulen and Nielson, 1993). The two
reservoirs (shallow and deeper “high-temperature zone”) produce steam at
temperatures in the range 235–342
Geothermal surface manifestation are widely diffused counting several thermal springs in the surrounding area (Donnelly-Nolan et al., 1993).
The Kizildere field is characterized by three different liquid-dominated
reservoirs hosted in fractured sedimentary and metamorphic rocks. The first
well (KD1) was completed in 1968, reaching 198
In our opinion a worldwide-accepted temperature-based classification of geothermal resources is needed, because it provides a quantitative evaluation of power production. On the other hand, in agreement with the definition of geothermal play of Moeck (2014), in the first stage of exploration it is useful to take into account a catalogue of plays based on geological features. But we argue with the following question: are geological features clear enough for characterizing favourable conditions for geothermal resources?
The comparison of the main geological conditions among the Larderello and Mt Amiata (Italy), The Geysers (USA) and Kizildere (Turkey) geothermal fields led us to identify common features that may characterize the convective and intrusive play. The most important common feature of this type of play is the effectiveness of the heat source represented by shallow plutonic intrusions, although nowadays for Mt Amiata and Kizildere fields the magmatic contribution is only inferred, as it was inferred in Larderello and The Geysers at the beginning. This is a crucial point, and the term intrusive for a play is not so immediate to apply. For example, different models have been also proposed for Kizildere with the magmatic activity the main matter of debate, as in Larderello and The Geysers before drilling and coring granites and felsite. Another example is that recent acidic intrusions are considered to be the heat source of Larderello and Mt Amiata geothermal systems (Bertini et al., 2006; Gianelli, 2008), whereas in the conceptual model of the Larderello geothermal area of Brogi et al. (2003, and references therein) the geothermal area is located in a “basin and range”-like structure, the magmatic contribution as heat source being minimized and depicting an overall scenario of a fault-controlled system.
The efficacy of the heat source is a leading issue. In fact, with regard to the Italian and American fields the available information nowadays endorses the effectiveness of buried intrusion older than 1 Myr. Mathematical models exclude the possibility that intrusions of any reasonable, even large, size can supply enough heat and are able to feed large geothermal systems for more than 1 Myr (Norton and Knight, 1977; Calore et al. 1981; Cathless and Erendi, 1997). A continuous magma, and therefore heat, feeding is therefore necessary to maintain a geothermal system of the size of Larderello or The Geyser. In our opinion, the term “intrusive” should be accompanied by the term “Young”. We can define young as an intrusion at least coeval, or younger than the last tectonic phase affecting the geothermal area, and if isotopic dating is available it should be younger than approximately 1 Myr.
The age of the magmatism is used also for the catalogue of geothermal play proposed by IGA and IFC (2014), which, however, lacks a clear distinction of volcanic and intrusive plays. In our opinion a system fed by young intrusions with the geothermal reservoir hosted in the associated volcanites has different features with respect to a “convective, intrusive” play, with the reservoir hosted in sedimentary and crystalline units. The geothermal plays characterized by intrusions approximately older than 1 Myr and without evidence of melt or partial melting in the upper crust could be included into a “amagmatic play”, to be eventually sub-classified. Geochemical data on surface manifestations should be considered to support the cataloguing activities because they provide useful information about the hypothesis of magmatic contribution, helping in discriminating the intrusive and amagmatic systems.
Of course, large or composite batholiths are better heat sources than small dikes or laccoliths, which cannot induce thermal anomaly for a long period of time. The volume of the intrusion, however, is not a good discriminating parameter, because during the exploration it is difficult to define its size. Thus, further distinguishing the intrusive plays on the basis of the size of the intrusion is in our opinion not of practical use.
Another important issue is the convective heat transfer that implies the circulation of a thermovector fluid. This condition distinguishes the conventional system exploitable by current technologies from the conductive unconventional geothermal systems that require engineering stimulation. The four geothermal systems are classified as convective since they show a wide and effective hydrothermal circulation, even complex, considering the presence of more than one reservoir for each field. In our analysis we could count on geophysical data and well logs for fields in operation. Considering an initial stage of exploration for a play, without geophysical data, it is difficult to assess the regime of heat transfer at depth. A preliminary indication could be provided by the number and type of geothermal manifestations in the surrounding areas. Surface manifestations (e.g. hot springs and gas discharge) are common in the four fields of interest, disregarding the fluid phase in the reservoir (steam-dominated in Larderello and The Geysers, and liquid-dominated in Amiata and Kizildere). In the considered cases, the low-permeability layer acting as a cap rock is a low-permeability sedimentary or crystalline unit, and the abundance and distribution of natural manifestations, as well as the hydraulic head of steam or brine, are strictly related to the depth of the reservoir and the faults and fractures regime.
Attempts at evaluating the steam fraction in a geothermal reservoir have been proposed by D'Amore and Truesdell (1979), but its application before drilling and during the geochemical survey of natural manifestation is problematic. In any case the presence of steam phase is not an indication for intrusive or fault-controlled geothermal systems. For example Mt Amiata is an intrusive geothermal system with liquid-dominated reservoirs.
The comparison of these fields drove us to exclude lithological and stratigraphic conditions as key parameters to classify the plays. A geothermal reservoir can be hosted in various typologies of sedimentary and crystalline rocks. What really matters is the rock permeability.
What makes things even more complex is the geodynamic and structural setting, which may spatially vary in stress regime (from compressive to extensional or strike slip) and in time (polyphasic tectonic). In areas rich in data such as those we analysed the tectonic evolution is still under debate in the scientific community. We have shown that Larderello and Mt Amiata are located in the inner sector of an active orogenic belt that has undergone extensional tectonics since the Miocene or Pleistocene (non-univocal consensus about timing), and some authors have suggested the importance of recent strike-slip faults during the emplacement of plutons that represent the heat source.
The main elements in common in the four fields which we used for the classification are the hydrothermal circulation and the known or inferred plutonic heat source, respectively identifying the convective and intrusive terms. We observe that there are two other common parameters in the four areas: (i) relevant seismic activity and (ii) high heat flow. It is known that geothermal fields are common in tectonically active areas and earthquake swarms could be associated with areas of recent volcanic or geothermal activity (Sibson, 1996). The heat flow values depict huge thermal anomalies in the surrounding areas of the Larderello, Mt Amiata, The Geysers and Kizildere fields, with maximum values centred on the field in exploitation.
We compared the main geological features of the Larderello, Mt Amiata, The Geysers and Kizildere geothermal fields in order to describe the common elements that could be useful for the classification of geothermal plays based on the terminology proposed for the IGA workshop held in Essen, Germany (IGA, 2013). We classified these fields as convective and intrusive plays. The first term would indicate the presence of a reservoir suitable for economic exploitation with current technologies without engineering stimulation. The term intrusive is correlated with the plutonic heat source that feeds wide and highly productive geothermal systems. We do not adhere to the proposal to split this play into different kinds depending on the tectonic setting. Considering that a play should be defined in an unambiguous way, and should help in classifying resources and planning exploration decisions, we conclude that recognized resources, such as those we analysed, and even more so the prospective resources, can hardly be classified on the basis of tectonic setting. We explained that geodynamic and structural setting are still debated in such well-known fields, and a tectonics-based classification of geothermal plays could not simplify the exploration planning. The structural survey remains a milestone in a geothermal exploration project to understand tectonics evolution and to assess the faults and fractures systems that control hydrothermal circulation.
With regard to the classification of geothermal plays we suggest simplifying the classification of the convective plays, distinguishing volcanic, young intrusive and amagmatic. Our classification reduces the emphasis on the tectonic settings, which can be subjective and therefore lead to ambiguous conceptual models. Besides highlighting the importance of geochemical data for inferring magmatic heat source, we identify two more features that are common in the four fields: (i) they are seismically active, and (ii) they show high heat flow values and wide thermal anomalies. These features, more than structural and tectonic features, might be used for a sub-classification.
The authors would like to thank the anonymous reviewers for their valuable comments and suggestions, which greatly contributed to improve the final version of the paper. A special acknowledgement also goes to Samuele Agostini for the useful discussion on the subject of this study. Edited by: I. S. Moeck Reviewed by: two anonymous referees