Coupled Microbiological–Isotopic Approach for Studying Hydrodynamics in Deep Reservoirs: The Case of the Val d’Agri Oilﬁeld (Southern Italy)

: The studies upstream of the petroleum industry include oil and gas geological exploration and are usually focused on geological, structural, geophysical, and modeling techniques. In this research, the application of a coupled microbiological–isotopic approach was explored to assess its potential as an adequate characterization and monitoring tool of geoﬂuids in oilﬁeld areas, in order to expand and reﬁne the information acquired through more consolidated practices. The test site was selected within the Val d’Agri oilﬁeld, where some natural hydrocarbon springs have been documented since the 19th century in the Tramutola area. Close to these springs, several tens of exploration and production wells were drilled in the ﬁrst half of the 20th century. The results demonstrated the e ﬀ ectiveness of the proposed approach for the analysis of ﬂuid dynamics in complex systems, such as oilﬁeld areas, and highlighted the capacity of microbial communities to “behave” as “bio-thermometers”, that is, as indicators of the di ﬀ erent temperatures in various subsurface compartments.


Introduction
Petroleum reservoirs are discovered in a wide range of geologic settings across the continents [1] and their monitoring is one of the key factors in the management of oil and gas resources [2]. Successful management requires an understanding of the structure of the reservoir, the distribution of fluids within the reservoir, drilling and maintaining wells which can produce fluids from the reservoir, transport and processing of produced fluids, refining and marketing the fluids, safely abandoning the reservoir when it can no longer produce, and mitigating the environmental impact of operations throughout the life cycle of the reservoir [3].
The studies on mineral oil and gas reservoirs are usually focused on geological, structural, and geophysical features, e.g., [4][5][6][7][8]. In this research, we explore, for the first time, the potential application of a coupled microbiological-isotopic approach as a useful tool for the characterization and monitoring of geofluids in oilfield areas, in order to expand and refine the information acquired through more consolidated practices.
The isotopes, and in particular the analysis of the stable isotopes 18 O and 2 H, allow identifying the origins of groundwater, and their use has long been consolidated in hydrogeological studies, e.g., [9][10][11].
On the other hand, prokaryotes (domains Bacteria and Archaea) have developed a high adaptive capacity in the most different habitats on the planet, and they can colonize even the harshest environments. They dominate global biogeochemical cycles, thus regulating ecosystem functions. Bacteria and archaea are ubiquitous in nature and have often been used to monitor water quality, e.g., [12,13], to increase knowledge of the hydrogeological characteristics of aquifer systems, e.g., [14,15], and to evaluate the potential of bioremediation of contaminated sites, e.g., [16][17][18][19]. Besides, over the last decades, broad phylogenetic and functional diverse microbial communities of several subsurface oil reservoirs have been described using the newly available molecular techniques, e.g., [20].
It is well known that only a small fraction of naturally occurring microorganisms can be cultivated in laboratory growth conditions. This has hindered, in the past, the full characterization of ecosystems and precluded the understanding of how they work and are regulated. Metagenomics and other "omics" are among the fastest advancing scientific tools at the basis of the recent and unprecedented access to genetic and functional information of entire microbial communities, contributing to knowledge about mechanisms and processes of essential ecosystem services, and to the emergence of innovative applications in many different areas. For instance, the next-generation sequencing (NGS) of 16S rRNA gene is now one of the most widely used applications for the taxonomic and phylogenetic evaluation of microbial community composition, e.g., [21][22][23], and it has opened the door to a deeper insight of complex environments.

Study Area
The Val d'Agri is a Quaternary NW-SE trending intramountain basin located within the southern Apennines thrust belt (southern Italy) (Figure 1), whose formation and evolution were controlled by brittle tectonics. The intense and recent deformation is testified by seismic activity in the last 40 ka, such as the M7 1857 Basilicata earthquake, e.g., [24][25][26]. The main Val d'Agri oilfield is hosted in a reservoir made of fractured, low-porosity carbonates belonging to the buried inner Apenninic Platform belt, e.g., [28][29][30]. Light to medium crude oil and  The main Val d'Agri oilfield is hosted in a reservoir made of fractured, low-porosity carbonates belonging to the buried inner Apenninic Platform belt, e.g., [28][29][30]. Light to medium crude oil and gas are stored in limestone and dolomite (Miocene to Cretaceous age) [31]. The carbonate reservoir lies below the Pliocene siliciclastic foredeep deposits and a thick mélange layer ( Figure 2). Hydrocarbons have been extracted since the mid-1900s through several wells at a depth ranging from 1.8 to 3.5 km below sea level.
The Apenninic Platform is about 7000 m thick and characterized by a bottom part made up of evaporites, sandstones, and conglomerates (Triassic age) lying above a crystalline basement. Well data were used to establish the progressive movement of the front of the chain towards the NE during the Pliocene-early Pleistocene.
Water 2019, 11, x FOR PEER REVIEW 3 of 18 gas are stored in limestone and dolomite (Miocene to Cretaceous age) [31]. The carbonate reservoir lies below the Pliocene siliciclastic foredeep deposits and a thick mélange layer ( Figure 2). Hydrocarbons have been extracted since the mid-1900s through several wells at a depth ranging from 1.8 to 3.5 km below sea level.
The Apenninic Platform is about 7000 m thick and characterized by a bottom part made up of evaporites, sandstones, and conglomerates (Triassic age) lying above a crystalline basement. Well data were used to establish the progressive movement of the front of the chain towards the NE during the Pliocene-early Pleistocene. The study area is in Tramutola village. It includes natural hydrocarbon springs whose presence was already known thanks to oral testimonies and several papers published at the end of the 19th century and early 20th century [32]. The hydrocarbon seepages pour in the stream Rio Cavolo, forming oil stains ( Figure 3).
From the late nineteenth century onward, research activities led to the discovery of the small and superficial Tramutola oil field exploited by Agip Mineraria in the 1930s and 1940s through 45 exploration and production wells. The study area is in Tramutola village. It includes natural hydrocarbon springs whose presence was already known thanks to oral testimonies and several papers published at the end of the 19th century and early 20th century [32]. The hydrocarbon seepages pour in the stream Rio Cavolo, forming oil stains (Figure 3).
From the late nineteenth century onward, research activities led to the discovery of the small and superficial Tramutola oil field exploited by Agip Mineraria in the 1930s and 1940s through 45 exploration and production wells.
These wells intercepted oil and/or gas from a few to several hundreds of meters below the ground (b.g.), e.g., [33]. From one of these wells (P art , artesian well) and the hydrocarbon springs, hydrogen sulfide (H 2 S) emissions occur and are immediately recognizable because of the typical smell of "rotten egg" (unpublished data).
The springs S1 and S2, as well as P art , are located along a W-E fault where the Apulian carbonate platform and Rio Cavolo Unit (Oligocene) [34] crop out ( Figure 4). The P art stratigraphy can be schematized as follows (from the top to the bottom): Rio Cavolo Unit from 0 to 44 m b.g., Apenninic Platform carbonates from 44 to 136 m b.g., tectonic mélange from 136 to 165 m b.g., Flysch Galestrino Formation from 165 to 352 m b.g., and Scisti Silicei Formation from 352 to 404 m b.g. Oil and gas were detected at a different depth within the Rio Cavolo Unit, Flysch Galestrino, and Scisti Silicei Formations [33]. These wells intercepted oil and/or gas from a few to several hundreds of meters below the ground (b.g.), e.g., [33]. From one of these wells (Part, artesian well) and the hydrocarbon springs, hydrogen sulfide (H2S) emissions occur and are immediately recognizable because of the typical smell of "rotten egg" (unpublished data).
The springs S1 and S2, as well as Part, are located along a W-E fault where the Apulian carbonate platform and Rio Cavolo Unit (Oligocene) [34] crop out ( Figure 4). The Part stratigraphy can be schematized as follows (from the top to the bottom): Rio Cavolo Unit from 0 to 44 m b.g., Apenninic Platform carbonates from 44 to 136 m b.g., tectonic mélange from 136 to 165 m b.g., Flysch Galestrino Formation from 165 to 352 m b.g., and Scisti Silicei Formation from 352 to 404 m b.g. Oil and gas were detected at a different depth within the Rio Cavolo Unit, Flysch Galestrino, and Scisti Silicei Formations [33].
The Rio Cavolo Unit is made up of clays, micaceous limestone, and rare marly layers. The tectonic mélange is between (overthrust) the Apennine Platform and the Lagonegrese Units [35]. The Flysch Galestrino Formation (Lower Cretaceous) is made up of clays, marls, and limestone, while the Scisti Silicei Formation (Upper Triassic-Jurassic) is composed of clays, marls, and chert [36,37].
From the hydrogeological point of view, the study site belongs to an area where different hydrogeological series complexes crop out. In more detail, the carbonate rocks belong to the so-called Mesozoic carbonate platform series complexes, whose permeability is very high, due to a welldeveloped fracture network and the presence of karst conduits [38]. The less permeable sedimentary successions belong to the syn-orogenic turbidite series complexes and both the outer and the inner basins series complexes. The rock masses belonging to these complexes are characterized by a permeability ranging from very low to low, due to a mixed pore-fracture network. However, their hydraulic conductivity can be locally enhanced due to well-developed damage zones associated to fault zones [38]. Along the Cavolo stream and the whole Agri Valley, alluvial sediments crop out. Due to the coexistence of fine and coarse sediments, the hydraulic conductivity of the alluvial complex range between less than 1 × 10 −8 and 2 × 10 −2 m/s [39].
No detailed studies have been published concerning the hydrogeological behavior of carbonate and siliciclastic media at the study area. However, the same hydrogeological units were deeply investigated and characterized in the wider context of the continental Italian southern Apennines (for siliciclastic low-permeability media, see, for example, Petrella and Celico, 2009 [40]; for highpermeability carbonate aquifers, see, for example, Petrella and Celico, 2013 [41], De Vita et al. 2012 [42], Allocca et al. 2015 [43], and Fiorillo et al. 2018 and 2019 [44,45]).
As per groundwater geochemistry in the study area, interesting results were obtained by Paternoster et al. 2005 [46]. The springs fed by carbonate aquifers have a Ca-HCO3 composition, while groundwaters flowing within the siliciclastic sediments have a high amount of As and Cu. However, the concentrations of As and first-row transition elements were usually below the maximum  permissible level for drinking water defined by Italian law. The authors link the availability of As, Pb, Cu, Zn, and Fe to the occurrence of iron oxi-hydroxides. Moreover, nitrate concentration seems to be influenced by the use of fertilizers.

Hydrogeological Investigations
The discharges of the hydrocarbon springs S1 and S2 were measured in low flow (July 2018), in early recharge (October 2018) and in late recharge (March 2019). The flow rate of the Part artesian well was not measurable, but based on some historical data, its order of magnitude is about 30 m 3 /h (unpublished data).

Water Sampling and Analyses
Three sampling campaigns have been carried out in July 2018, October 2018, and March 2019. Rainwater samples for isotopic analyses were collected monthly in two local rain samplers located at 1047 and 1290 m above sea level (a.s.l.).
The rainfall was collected using ten-liter polyethylene bottles containing about 300 mL of vaseline oil to prevent evaporation processes. Oil contamination was carefully avoided by syringing the water samples out of the bottle.
Groundwater samples for stable isotope (δ 18 O, δ 2 H), tritium, and microbiological (Next-Generation Sequencing of 16S rRNA gene) analyses were collected during the discharge measurements at springs S1 and S2, and at the Part well (screened at the well bottom).
The non-hydrocarbon spring S3, fed by the local carbonate aquifer and located at the contact between the high-permeability carbonate rocks and low-permeability siliciclastic successions, was analyzed for its isotopic content and used as a sort of endmember to compare hydrocarbon spring water with groundwater exclusively flowing within a relatively shallow aquifer system.
The localization of the rain samplers, springs, and the artesian well is reported in Table 1.
In addition, a water sample was also collected from the deep reservoir and analyzed to compare its isotopic signature and microbial community with those retrieved in the spring and Part groundwaters. . Geological map of the study area; the blue points and the triangle show the location of the investigated springs S1, S2, and S3 and the artesian well P art (the geological sketch is taken from Olita 2018, modified [33]).
The Rio Cavolo Unit is made up of clays, micaceous limestone, and rare marly layers. The tectonic mélange is between (overthrust) the Apennine Platform and the Lagonegrese Units [35]. The Flysch Galestrino Formation (Lower Cretaceous) is made up of clays, marls, and limestone, while the Scisti Silicei Formation (Upper Triassic-Jurassic) is composed of clays, marls, and chert [36,37].
From the hydrogeological point of view, the study site belongs to an area where different hydrogeological series complexes crop out. In more detail, the carbonate rocks belong to the so-called Mesozoic carbonate platform series complexes, whose permeability is very high, due to a well-developed fracture network and the presence of karst conduits [38]. The less permeable sedimentary successions belong to the syn-orogenic turbidite series complexes and both the outer and the inner basins series complexes. The rock masses belonging to these complexes are characterized by a permeability ranging from very low to low, due to a mixed pore-fracture network. However, their hydraulic conductivity can be locally enhanced due to well-developed damage zones associated to fault zones [38]. Along the Cavolo stream and the whole Agri Valley, alluvial sediments crop out. Due to the coexistence of fine and coarse sediments, the hydraulic conductivity of the alluvial complex range between less than 1 × 10 −8 and 2 × 10 −2 m/s [39].
No detailed studies have been published concerning the hydrogeological behavior of carbonate and siliciclastic media at the study area. However, the same hydrogeological units were deeply investigated and characterized in the wider context of the continental Italian southern Apennines (for siliciclastic low-permeability media, see, for example, Petrella and Celico, 2009 [40]; for high-permeability carbonate  [44,45]).
As per groundwater geochemistry in the study area, interesting results were obtained by Paternoster et al. 2005 [46]. The springs fed by carbonate aquifers have a Ca-HCO 3 composition, while groundwaters flowing within the siliciclastic sediments have a high amount of As and Cu. However, the concentrations of As and first-row transition elements were usually below the maximum permissible level for drinking water defined by Italian law. The authors link the availability of As, Pb, Cu, Zn, and Fe to the occurrence of iron oxi-hydroxides. Moreover, nitrate concentration seems to be influenced by the use of fertilizers.

Hydrogeological Investigations
The discharges of the hydrocarbon springs S1 and S2 were measured in low flow (July 2018), in early recharge (October 2018) and in late recharge (March 2019). The flow rate of the P art artesian well was not measurable, but based on some historical data, its order of magnitude is about 30 m 3 /h (unpublished data).

Water Sampling and Analyses
Three sampling campaigns have been carried out in July 2018, October 2018, and March 2019. Rainwater samples for isotopic analyses were collected monthly in two local rain samplers located at 1047 and 1290 m above sea level (a.s.l.).
The rainfall was collected using ten-liter polyethylene bottles containing about 300 mL of vaseline oil to prevent evaporation processes. Oil contamination was carefully avoided by syringing the water samples out of the bottle.
Groundwater samples for stable isotope (δ 18 O, δ 2 H), tritium, and microbiological (Next-Generation Sequencing of 16S rRNA gene) analyses were collected during the discharge measurements at springs S1 and S2, and at the P art well (screened at the well bottom).
The non-hydrocarbon spring S3, fed by the local carbonate aquifer and located at the contact between the high-permeability carbonate rocks and low-permeability siliciclastic successions, was analyzed for its isotopic content and used as a sort of endmember to compare hydrocarbon spring water with groundwater exclusively flowing within a relatively shallow aquifer system.
The localization of the rain samplers, springs, and the artesian well is reported in Table 1.
In addition, a water sample was also collected from the deep reservoir and analyzed to compare its isotopic signature and microbial community with those retrieved in the spring and P art groundwaters.
Electrical conductivity, temperature, and pH measurements were performed in situ with portable equipment (Hanna Instruments 9829).
All samples were stored in a refrigerated box and transported to the laboratory. Stable isotope analyses (δ 18 O, δ 2 H) were carried out at the Isotope Geochemistry Laboratory of the University of Parma (Italy), using a Delta Plus mass spectrometer (Thermo Fisher Scientific, Waltham, Massachusetts, USA) coupled to an automatic HDO device preparation system. The technique consists of bringing the liquid sample into an isotopic equilibrium, at a controlled temperature of 18 • C, with a pure gas (CO 2 in the case of oxygen and H 2 in the case of hydrogen). The isotopic equilibrium, in the case of hydrogen, would be reached very slowly, and platinum catalysts are used to accelerate the reaction. For the oxygen isotope determination, 5 cm 3 of water was equilibrated with pure CO 2 , while for hydrogen isotopes, 5 cm 3 of water was equilibrated with pure H 2 (platinum wire was used as a catalyzer of gas-liquid water equilibration). Equilibration times were 3 h for hydrogen and 8 h for oxygen. The isotope ratio is expressed as where "A" is 18 or 2, "B" is 16 or 1, "R" is the ratio of the isotopic abundances, "i" is the sample of interest, and % = 10 −3 . Analyses of 3 H were carried out at the Isotope Geochemistry Laboratory of Trieste University, Italy. To decrease measurement errors, the samples followed the procedure of the preventive electrolytic enrichment of tritium, where 250 g of the water sample was expected to be reduced, by electrolysis, to 20 g. The analyses for the determination of the tritium activity were carried out according to the procedures provided by Water and Environment News No. 3 (1998) [47]. The analytical prediction uncertainty was ± 0.1% for δ 18 O, ± 1% for δ 2 H, and ± 0.5 TU for 3 H.

Chemical Analyses
During the third sampling campaign in March 2019, water samples were collected from the hydrocarbon springs and the well, P art , to analyze Benzene, Toluene, Ethylbenzene, Xylene (BTEX) and Polycyclic aromatic hydrocarbons (PAHs) contents. Forty milliliter colorless glass vials were used for the BTEX analysis, while 1 L black glass bottles were used for PAH analysis. The analyses were performed at Biochemie Lab S.r.l. following the EPA 5030C 2003 + EPA 8015D 2003 protocol for BTEX and the EPA3510C 1996 + EPA 8270E 2018 protocol for PAH [48][49][50][51].

Microbiological Analyses: 16S Ribosomal RNA Gene Next Generation Sequencing (NGS)
For bacterial community analyses, water samples (4 L) were filtered through sterile mixed esters of cellulose filters (S-Pak TM Membrane Filters, 47 mm diameter, 0.22 µm pore size, Millipore Corporation, Billerica, MA, USA) within 24 h from the collection. Bacterial DNA extraction from filters was performed using the commercial kit FastDNA SPIN Kit for soil and FastPrep ® Instrument. After the extraction, DNA integrity and quantity were evaluated by electrophoresis in 0.8% agarose gel containing 1 µg/mL of Gel-RedTM. The bacterial community profiles in the samples were generated by NGS technologies at the Genprobio S.r.l. Laboratory. Partial 16S rRNA gene sequences were obtained from the extracted DNA by polymerase chain reaction (PCR), using the primer pair Probio_Uni and Probio_Rev, targeting the V3 region of the bacterial 16S rRNA gene sequence [52]. Amplifications were carried out using a Verity Thermocycler (Applied Biosystems) and PCR products were purified by the magnetic purification step involving the Agencourt AMPure XP DNA purification beads (Beckman Coulter Genomics GmbH, Bernried, Germany) in order to remove primer dimers. Amplicon checks were carried out as previously described [52]. Sequencing was performed using an Illumina MiSeq sequencer with MiSeq Reagent Kit v3 chemicals. The fastq files were processed using a custom script based on the QIIME software suite [53]. Paired-end read pairs were assembled to reconstruct the complete Probio_Uni/Probio_Rev amplicons. Quality control retained sequences with a length between 140 and 400 bp and mean sequence quality score > 20 while sequences with homopolymers > 7 bp and mismatched primers were omitted. To calculate downstream diversity measures, operational taxonomic units (OTUs) were defined at 100% sequence homology using DADA2 [54]; OTUs not encompassing at least two sequences of the same sample were removed. All reads were classified to the lowest possible taxonomic rank using QIIME2 [53,55] and a reference dataset from the SILVA database v132 [56]. The biodiversity of the samples (alpha-diversity) was calculated with the Shannon index.

Hydrogeological Settings
The discharge of the spring S1 varied slightly over time, showing a slight decrease (1.4 to 1.3 m 3 /h) from July to late October (dry period), and an increase (1.3 to 1.7 m 3 /h) in March (rainy period), in agreement with the distribution of precipitation in that area. The overall synchronicity between rainy periods and the increase in discharge at S1 clearly suggested active pathways within the feeding aquifer system. Unfortunately, the S2 discharge was not measurable after the winter period, therefore no speculation can be formulated concerning recharge processes. where "s(yx)" is the standard error of regression and "n" is the number of couples of data, "A" is the intercept of the regression line, and p = probability. For precipitation in Southern Italy [57], the resulting regression line is  (2) and H o : E (1) = E (2) cannot be rejected with high probability; thus we assume that the two regressions are not different.

Isotope Investigations
The isotopic data of spring and groundwaters collected in the Tramutola study area are located close to the local meteoric water line (2) suggesting a meteoric origin of the analyzed waters ( Figure 5). On the contrary, the samples taken from the deep reservoir are far from the line ( Figure 5); actually, they represent fossil waters that also interacted with the carbonate formation at an elevated temperature.
The isotopic composition of spring and groundwater samples did not vary widely over time; isotopic variations were lower than 2u, where "u" is prediction uncertainty for the δ 18 O and δ 2 H data.
As far as tritium is concerned, all samples showed a relatively high 3 H content (4.1 to 9.1 TU), if compared with tritium content in recent rainwaters analyzed in southern Italy (e.g., 4.6 TU [59]; 5.0 TU [10]; 6.2 to 10.8 [unpublished data]) and the wider Adriatic area [60]. The non-hydrocarbon spring, S3, was characterized by the highest tritium value (8.4 to 9.1 TU), in agreement with the rapid pathways within the carbonate aquifer. Moreover, the variation of TU values was lower than 2u, suggesting little interaction with waters having quite different TU values. The hydrocarbon springs S1 and S2 had similar tritium contents (5.6 to 6.9 TU and 6.8 to 7.5 TU, respectively), slightly lower than those characterizing the S3 spring water. As for S1 water, the variation over time was slightly higher than the 2σ error of the 3 H analyses, therefore suggesting the mixing of different endmembers, possibly related to longer (lower tritium content) and shorter (higher tritium content) pathways. Taking into consideration the homogeneous stable isotope content over time, both pathways are related to well-mixed groundwater. P art waters showed a more significant variation over time (4.1 to 6.2 TU), further confirming the existence of mixing between different endmembers: (i) one related to rainwater infiltrating relatively far from the observation well and (ii) a second one linked to closer pathways.
regressions are not different.
The isotopic data of spring and groundwaters collected in the Tramutola study area are located close to the local meteoric water line (2) suggesting a meteoric origin of the analyzed waters ( Figure  5). On the contrary, the samples taken from the deep reservoir are far from the line ( Figure 5); actually, they represent fossil waters that also interacted with the carbonate formation at an elevated temperature. The isotopic composition of spring and groundwater samples did not vary widely over time; isotopic variations were lower than 2u, where "u" is prediction uncertainty for the δ 18 O and δ 2 H data.

Chemical Analyses
Chemical analyses were performed on springs S1, S2, and well P art . Samples collected at the spring S3, whose waters are used for drinking purposes, were not considered for BTEX and PAHs determinations. The data revealed detectable PAH such as naphtalene (0.00231 µg/L) in the spring S1, and benzo

Next-Generation Sequencing Results
MiSeq runs produced an average of 61,682 sequences for the samples collected at the springs (S1 and S2) and from the artesian well, P art . An average of 80,061 reads was obtained from the analysis of the deep reservoir bacterial community ( Table 2). The 16S rRNA gene sequences generated in this study have been deposited in the NCBI Sequence Read Archive under the accession number PRJNA629324.
The rarefaction analysis (a measure used to estimate the alpha diversity in samples and gauge whether or not sequencing efforts captured the microbial diversity) highlighted a greater microbial diversity in the spring S2 compared to the spring S1, the artesian well, and the deep reservoir ( Figure S1). Proteobacteria, Chloroflexi, and Bacteroidetes were the three major phyla in waters from the spring S1 in July 2018 (85.55% of sequences). In October 2018 and March 2019, Epsilonbacteraeota were found with the highest percentages (72.25% and 52.48%, respectively) ahead of Proteobacteria (25.78% and 20.67%) and Bacteroidetes (0.86% and 23.63%).
Microbial communities in groundwater collected from the P art well were mainly characterized by Proteobacteria and Patescibacteria, with mean relative abundance values of 92.78% and 4.09%, respectively.
The phyla Proteobacteria, Synergistetes, and Firmicutes accounted for, on average, 94.82% of the sequences retrieved from the deep reservoir. Overall, Proteobacteria represented the dominant phylum in all the samples collected from July 2018 to March 2019, ranging from 42.65% to 50.88%.
The analysis of the microbial community composition at the family level ( Figure 6) revealed, in spring S1, a predominance of Helicobacteraceae (47.12%), Chlorobiaceae (7.63%), and unclassified microorganisms of the order Chloroflexales   The bacterial communities of the P art well were mainly characterized by the families Burkholderiaceae, Rhodocyclaceae, and Methylophilaceae, accounting for, on average, 29.16%, 23.49%, and 17.34% of sequences, respectively.
Desulfomicrobiaceae, Synergistaceae, and family III of the order Thermoanaerobacterales were found at the highest percentages in the deep reservoir, with mean values of 43.97%, 25.49%, and 20.19%, respectively.
Many of these genera encompass chemolithotrophic or phototrophic sulfur-oxidizing bacteria (SOB), which derive energy from the oxidation of reduced sulfur compounds, or use sulfide as electron donors for anoxygenic photosynthesis, like the green sulfur bacterium Chlorobium limicola [67], playing an important role in the element cycling in the environment. These results are not surprising, especially when considering the hydrogen sulfide emissions from the analyzed well and springs. In fact, the presence of this gas in waters, probably naturally generated in situ from reservoir biomass and sulfate-containing minerals through microbial sulfate reduction and/or thermochemical sulfate reduction, could have represented a driving factor shaping microbial community structure and function.

Taxonomy
Growth Temperature ( • C) Citations

Discussion and Conclusions
Both the hydrogeological behavior and the isotopic features of the studied hydrocarbon springs suggest that they are strictly related to active recharge in a local aquifer system, in agreement with findings related to the nearby non-hydrocarbon spring, whose waters are used for drinking purposes. At the same time, the hydrocarbon springs flow out along a fault zone, which enhances fluid flow, allowing the upflow of hydrocarbons and their mixing with the local groundwater, which is reasonably fed by the nearby carbonate aquifer. In detail, this is due to the fault crossing the sequence made of the Scisti Silicei Formation, Galestri Formation, Tectonic Mélange, Apenninic Carbonate platform, and Rio Cavolo Formation, characterized by oil and gas layers at different depths [33]. The hydraulic behavior of this fault zone is similar to that of faults present in other carbonate aquifers in southern Italy, where the existence of high-permeability damage zones, e.g., [83][84][85] and/or heterogeneous fault cores, allowing a significant fluid migration, has been revealed in previous research, e.g., [86][87][88][89][90][91][92][93]. In these contexts, bacterial cell filtration typical of low-permeability fine-grained media, e.g., [90,[94][95][96][97] is limited, and microorganisms can be used effectively as bio-tracers for specific hydrogeological and microbiological purposes, e.g., [97].
The groundwater intercepted by the P art well is also fed by a more prolonged pathway, as demonstrated by the tritium content lower than those detected in the hydrocarbon springs and S3. Taking into consideration the wider geological setting, the artesian well intercepted a relatively deep (compatible with mesophilic and psychrophilic bacteria), but active pathway within the Scisti Silicei aquifer (Figure 7). This aquifer is unconfined upgradient and downgradient with respect to the P art well, where the Scisti Silicei Formation crops out. Differently, it is confined (and locally artesian) where the Scisti Silicei is beneath the low-permeability flysch deposits. This deep groundwater naturally flows eastwards, towards the alluvial aquifer of the Agri Valley.
Water 2019, 11, x FOR PEER REVIEW 12 of 18 reasonably fed by the nearby carbonate aquifer. In detail, this is due to the fault crossing the sequence made of the Scisti Silicei Formation, Galestri Formation, Tectonic Mélange, Apenninic Carbonate platform, and Rio Cavolo Formation, characterized by oil and gas layers at different depths [33]. The hydraulic behavior of this fault zone is similar to that of faults present in other carbonate aquifers in southern Italy, where the existence of high-permeability damage zones, e.g., [83][84][85] and/or heterogeneous fault cores, allowing a significant fluid migration, has been revealed in previous research, e.g., [86][87][88][89][90][91][92][93]. In these contexts, bacterial cell filtration typical of low-permeability finegrained media, e.g., [90,[94][95][96][97] is limited, and microorganisms can be used effectively as bio-tracers for specific hydrogeological and microbiological purposes, e.g., [97]. The groundwater intercepted by the Part well is also fed by a more prolonged pathway, as demonstrated by the tritium content lower than those detected in the hydrocarbon springs and S3. Taking into consideration the wider geological setting, the artesian well intercepted a relatively deep (compatible with mesophilic and psychrophilic bacteria), but active pathway within the Scisti Silicei aquifer (Figure 7). This aquifer is unconfined upgradient and downgradient with respect to the Part well, where the Scisti Silicei Formation crops out. Differently, it is confined (and locally artesian) where the Scisti Silicei is beneath the low-permeability flysch deposits. This deep groundwater naturally flows eastwards, towards the alluvial aquifer of the Agri Valley. In the present study, the potential of a coupled microbiological-isotopic approach for monitoring geofluids in hydrocarbon reservoirs and, in detail, the capacity of microbial communities to "behave" as "bio-thermometers", has been assessed for the first time.
When analyzing the communities in the Part waters, collected at about 400 m b.g., and in spring waters, only mesophilic and psychrophilic microorganisms were detected. Differently, in the deep reservoir, thermophilic bacteria thriving at high temperatures, were found. These findings are consistent with geothermal curves and isotherms reported by Candela et al. [98] at the Val d'Agri oilfield (Figure 8), and demonstrate the usefulness of the proposed approach, at least at the study site.
The development and application of molecular biological methods to hydrogeological issues has led to increasing numbers of studies on the microbial communities of aquifer systems over the past few decades, e.g., [90,97]. For example, in the carbonate environments of southern Italy, the potential use of microorganisms as tracers has been examined with reference to the analysis of recharge and flow processes with excellent results, e.g., [90,97]. At the study site, the analysis of bacterial species in spring, groundwater, and deep reservoir samples and isotopic analyses proved to be an effective tool to obtain information on the subsurface dynamics and temperatures (Figure 8). In a broader perspective, the same approach could also be used for the comprehension of more complex phenomena in exploited oil fields. In the present study, the potential of a coupled microbiological-isotopic approach for monitoring geofluids in hydrocarbon reservoirs and, in detail, the capacity of microbial communities to "behave" as "bio-thermometers", has been assessed for the first time.
When analyzing the communities in the P art waters, collected at about 400 m b.g., and in spring waters, only mesophilic and psychrophilic microorganisms were detected. Differently, in the deep reservoir, thermophilic bacteria thriving at high temperatures, were found. These findings are consistent with geothermal curves and isotherms reported by Candela et al. [98] at the Val d'Agri oilfield (Figure 8), and demonstrate the usefulness of the proposed approach, at least at the study site.
The development and application of molecular biological methods to hydrogeological issues has led to increasing numbers of studies on the microbial communities of aquifer systems over the past few decades, e.g., [90,97]. For example, in the carbonate environments of southern Italy, the potential use of microorganisms as tracers has been examined with reference to the analysis of recharge and flow processes with excellent results, e.g., [90,97]. At the study site, the analysis of bacterial species in spring, groundwater, and deep reservoir samples and isotopic analyses proved to be an effective tool to obtain information on the subsurface dynamics and temperatures (Figure 8). In a broader perspective, the same approach could also be used for the comprehension of more complex phenomena in exploited oil fields.