Surface and Subsurface Behavior of a Natural Gas Storage Site over Time: The Case of the Cornegliano Gas Field (Po Plain, Northern Italy)

Abstract

Foredeep basins often host significant natural gas reservoirs within siliciclastic successions, as exemplified by the Po Plain (Northern Italy), one of Europe’s largest foredeep basins. Here, numerous depleted gas reservoirs have been successfully converted into underground gas storage (UGS) facilities. For safe and efficient storage operations, detailed reservoir characterization and continuous monitoring of surface and subsurface effects are crucial. This study investigates the Cornegliano Laudense reservoir during its first 5–7 years as a UGS facility, employing an integrated monitoring approach that combines traditional methods (InSAR for surface deformation, microseismic monitoring) with innovative techniques (Pulsed Neutron Log-PNL). The results clearly illustrate and quantify the significant increase in storage capacity over a relatively short operational period, primarily driven by the progressive displacement of formation water by injected gas. Despite increased stored gas volumes, monitoring revealed no adverse effects on surface stability or subsurface seismicity. This integrated methodology demonstrates substantial potential for refining predictive models, optimizing storage efficiency, and enhancing sustainable management practices for underground gas storage operations.

1. Introduction and Aims

Underground gas storage (UGS) plays a pivotal role in ensuring energy security and managing seasonal demand fluctuations. This is particularly important in the Po Plain (Northern Italy) region: this foredeep basin, one of Europe’s most industrialized and densely populated areas in Europe, has witnessed the transformation of several depleted gas fields into storage sites [1,2].

These sites, the effects of their activity, and the analysis of the surface/subsurface effects that can be induced are crucial for security reasons and to ensure an efficient and sustainable storage activity through time. The surface effects are mainly related to subsidence/uplift movements that could somehow affect the safety of urbanized areas and/or infrastructures. In the Po Plain, different research teams conducted subsidence analysis on the whole Po Plain and around storage sites, and multidisciplinary analysis of ground movements related to cyclical gas injection and withdrawal. Utilizing Interferometric Synthetic Aperture Radar (InSAR) and Global Navigation Satellite System (GNSS) data spanning nearly two decades, integrated geological, fluid-flow, and geomechanical models have been presented. These data and models provided insights into the poromechanical processes induced by storage operations, emphasizing the importance of both monitoring activities and such integrated approaches and underscoring the necessity of reliable ground movement predictions for the safety of storage systems and urban settlements [1,2,3,4,5]. In addition to having precise monitoring data and well-constrained geological models, it is essential to have a detailed understanding of the reservoir structure in order to monitor it over time by cross-referencing this information with data from storage activities. This combination of data allows a full and well-constrained understanding of the behavior of an active storage site. In the Cornegliano storage field, all these data are available for the last 5–7 years and have been analysed in an integrated way. The Cornegliano Gas Field in the Po Valley Petroleum Basin consists of a typical anticline, roughly E-W-oriented, bounded by a main thrust to the south and by its related backthrust to the north. Both structures belong to the outermost fronts of the Southern Alps that are buried under the Pleistocene deposits of the Central Po Plain (Northern Italy). This paper aims to investigate the following topics:

• Investigate the long-term effects of methane injection and withdrawal cycles on storage performance within a depleted siliciclastic reservoir in a foredeep basin.
• Determine whether continuous underground gas storage (UGS) operations influence surface deformation patterns around the reservoir site over extended operational periods.
• Assess whether prolonged storage operations contribute to induced seismicity and distinguish between operationally induced seismic events and natural tectonic activity.
• Explore how an integrated monitoring approach, combining traditional techniques (such as seismic and SAR) and innovative methods (such as Pulsed Neutron Logging), can enhance understanding of the reservoir’s geomechanical stability and environmental compatibility.
• Evaluate if insights gained from integrated monitoring and analyses can effectively refine predictive models and operational strategies, thereby reducing risks and uncertainties while ensuring operational safety and environmental compliance.

2. Background

2.1. Geological Settings-The Po Plain

The Po Plain Basin is the shared foreland of the Southern Alps and Northern Apennines belts (Northern Italy) and is the area where the outermost buried fronts of these two thrust-fold belts are closer. Presently, the Po Plain is an alluvial, E-W-elongated plain covering more than 40,000 square kilometers and limited by the Alps to the west and the north, by the Northern Apennines to the south, and by the Adriatic Sea to the east (Figure 1). From a structural point of view, the Po Plain is located over a portion of the Adria microplate between two opposing-verging fold and thrust belts [6,7,8,9]. The progressive tectonic evolution of the Southern Alps and Northern Apennines took place on a foreland affected by Mesozoic inherited structures, mainly NS-oriented normal faults that strongly affected the facies distribution and the successive advancement of the Southern Alps and Northern Apennines advancing tectonic fronts [10,11,12,13]. The main tectonic activity periods of the two thrust-fold belts are different: the Po Plain was in fact affected by south-vergent Southern Alps compressional tectonic events since the Middle Eocene and by the N-NE vergent Northern Apennines compression since the Lower Miocene [7,13]. In the northern and central parts of the Po Plain, the south-verging buried thrust fronts of the Southern Alps, partly outcropping along the Northern boundary of the alluvial plain, are present. To the south, the Pedeapenninic Thrust Front separates the outcropping Northern Apennines from the buried N-NE verging compressional structures. These buried thrusts are organized in three different structural arcs, whose geometry and rates of activity controlled the deposition and the architecture of the Plio-Pleistocene syntectonic successions of the Po Plain (onshore) and the Adriatic Sea (offshore) [13,14,15,16].

The Cornegliano Field is located in the Po Valley Petroleum Basin and consists of a typical anticline, bounded both to the North and to the south by reverse faults (Figure 2). The producing reservoir is composed of sands and shaly sands of the “Sabbie di Caviaga” formation, of Lower Pliocene age, at a depth range of 1350–1450 m below the surface. The reservoir is closed by marlstones at the base and a 400+ m-thick claystone cap rock. The Cornegliano Field was originally discovered by Eni S.p.A. in the 50 s and produced about 2.4 BSm³ (Billion standard cubic meters) of gas from 12 wells (12 of 20 wells drilled) until 1997. The Original Gas in Place was in the order of 3.2 BSm³, and therefore the field reached a recovery factor of 74%. A total of 20 wells were drilled between the 1950s and 1968, including 8 dry wells. All 20 wells were plugged and abandoned by Eni S.p.A. before the beginning of the conversion into Underground Gas Storage (UGS), which was operated by IGS in 2016, 33 years after the end of massive production in 1983.

2.2. Reservoir Structure

IGS began the conversion of the Cornegliano field into UGS in 2016 by drilling 14 new wells: 12 storage wells divided into 2 different areas called “cluster” and two additional wells used as microseismic monitoring wells. In 2017, after the first four wells had been drilled and the related evidence observed, a 3D seismic survey was performed to enhance the description of the reservoir structure; remaining drilling works greatly benefited from the data collected in the survey, and two further wells were completed down in the reservoir in 2019. Except for the two vertical wells, all the other wells were drilled as slanted wells, with inclinations of up to 90° at the target depth. The reservoir is divided into two main areas that are drained through two different clusters of wells (Cluster A and Cluster B). From the analysis of dynamic and pressure data, these two areas are apparently separated by a major fault (Figure 3), which turns out to produce partial compartmentalization between the two sectors and, thus, a partial flow barrier. Other minor faults are present in the two clusters, particularly in cluster A, but apparently, they do not affect or limit the fluid circulation inside the reservoir sectors.

The field is characterized by prevalent sand beds with intercalations of clayey facies. The reservoir is confined in a structural trap defined by a 179–430 m thick clayey-marly facies association on the top and by some folded and mineralized strata at the bottom, with a maximum thickness of 210 m. Existing wells produced gas from B, C1, C2, and C3 sandstone intervals (see

Table 1. Reservoir description.

Zones	Layers	Comments
B	B1—First “appearance” of sand below the cap rock	The B zone has overall poor reservoir quality with discontinuous sand bodies mostly developed in the Western part of the reservoir: area around wells A04 and C10.
	B2—Discontinuous sandstone bodies within the zone
C1	C1a—Upper sandstone layer within the zone	The C1 zone is characterized by medium to good reservoir quality with thickness of the sandstone layers up to 15 m. The sandstone bodies are continuous over the reservoir although not homogeneously distributed as those in the C2 zone.
	C1b—Lower sandstone layer within the zone
	C2a_1—Upper sandstone layer within the zone	The upper part of the C2 zone is characterized by good-quality sandstone with thicknesses of up to 20 m. In some wells, additional sandstone layers have been defined as C2a_1/L and C2a_2/L.
C2	C2a_2—Second sandstone layer within the upper part of the zone
	C2b—Lower sandstone layer within the zone
C3	C3a—Upper sandstone layer within the zone	The C3 zone is characterized by good-quality sandstone. As only a few wells entered this zone, the interpretation of the formation tops is uncertain and only realistic for the upper sandstone.

). Zones C1 and C2 are the main target zones for gas storage. Zone C2 has better reservoir properties and higher permeability and porosity than zone C1 and represents the main storage zone. The reservoir rocks are loose sands and shaly sands, characteristic of the Lower Pliocene “Sabbie di Caviaga” Formation, and are likely poorly cemented due to their relatively young geological age. This limited diagenetic maturation implies reduced compaction and cementation, making the formation mechanically weaker than older, more indurated reservoirs. As a result, these rocks exhibit elastic-brittle behavior under UGS operational conditions rather than ductile deformation. This is confirmed by the geomechanical model (not presented in this paper), which shows that stress paths during injection and withdrawal cycles remain below the Mohr–Coulomb failure envelope, and that plastic shear deformation is confined to discrete fault zones, not within the reservoir matrix itself. Initial reservoir pressure was 16.1 MPa (161 bar) at 1330 m. At the end of production (1997), reservoir pressure was 5 MPa (50 bar). Pressure at the drilling time of A01 and B01 wells was about 11.4 MPa (114 bar) at about 1400 m, with the gradient being 0.00815 MPa/m (0.081 bar/m) in the reservoir (from MDT pressure tests performed with the Modular Formation Dynamic Tester known as MDT). The stratigraphy and the log information regarding the Cornegliano reservoir are publicly available at ViDEPI https://www.videpi.com/videpi/cessati/pdf.asp (accessed on 30 July 2025) by looking at the Cornegliano 15 well located less than 2 km from the reservoir axial culmination.

The Cornegliano Laudense reservoir exhibits significant variability in petrophysical properties. Porosity values generally range between 15% and 29%, reflecting varying degrees of reservoir quality. Permeability shows considerable variation, extending from very low values around 0.01 millidarcies (mD), indicative of microporous to nanoporous textures, up to significantly higher permeability values reaching approximately 1700 mD, typical of macro-porous reservoir intervals. This broad range in both porosity and permeability highlights the reservoir’s heterogeneous nature and has implications for fluid flow dynamics and overall storage efficiency.

2.3. Storage Activities

The commercial activities of the Cornegliano UGS officially started on 13 December 2018, and after the so-called ramp-up phase, it became a fully operational underground gas storage facility. IGS operates the storage at a maximum static pressure of 16.1 MPa (161 bar), which corresponds to the original reservoir pressure at the time of its discovery. IGS operates differently from traditional seasonal gas storage models. Unlike conventional systems, which typically follow a seasonal pattern of injecting gas in summer and withdrawing in winter, IGS employs innovative strategies tailored for flexibility and market responsiveness. A first example of this innovation is the “5 + 2” product. On top of the standard seasonal product, the “5 + 2” operational model is proposed on a weekly basis during the withdrawal season, depending on the actual storage capacity. For this product, during the winter season, gas production occurs during the five working days (Monday to Friday) to meet industrial and commercial demand, while injection activities are concentrated over the weekend (Saturday and Sunday). This efficient scheduling allows IGS to optimize storage operations and adapt dynamically to market fluctuations, setting it apart in the energy storage landscape. Another example of IGS’s flexibility is the introduction of “intraday” products. These products enable shippers to switch between injection and production within a single gas day. This capability provides unparalleled responsiveness to real-time market needs.

In order to analyze and describe the evolution of the gas volume inventory from the beginning of the UGS conversion, Figure 4 illustrates the evolution of underground gas storage (UGS) volumes over time, providing a detailed breakdown of total gas storage and its distribution across two operational clusters, Cluster A and Cluster B. By examining the data from 1 December 2018 to March 2025, we aim to highlight seasonal trends, long-term variations, and the operational dynamics of the storage system. The gas volumes are represented proportionally, with “X” used as a reference value for scaling. The total gas volume (blue line) demonstrates a clear seasonal pattern, with injections typically occurring during the warmer months and withdrawal during colder months. Over the observed period, the total gas volumes have increased progressively, with seasonal peaks rising from an initial approximate value of 0.4X to nearly 1 (with 1 being the maximum volume of gas injected; for confidentiality and commercial reasons, all gas-volume values have been normalized to 1, keeping 1 as the present-day maximum value), reflecting substantial growth in storage capacity utilization (Figure 5). During withdrawal periods, the volume generally falls to around 0.4X–0.5X, with a noticeable dip in the winter of 2022–2023 (i.e., less than 0.3X), followed by a recovery in the following cycles. These variations align with the seasonal demand for gas, driven primarily by heating needs in winter and the gas market instability due to the war in Ukraine. Cluster A (green line) mirrors the seasonal trends seen in the total gas volume, with peak volumes initially around 0.3X, gradually increasing to approximately 0.6X on the normalized scale, indicating consistent performance and a critical role in gas storage operations. During withdrawal periods, the volumes typically drop to around 0.2X–0.3X. The regularity of these fluctuations reflects the cluster’s robust participation in both injection and withdrawal phases throughout the years. Cluster B (yellow line) represents the smallest contributor to total storage. Its peak volumes remain below 0.4X, and its seasonal fluctuations are less pronounced than those of Cluster A. The evolution of cluster B follows a different pattern compared to the one observed in Cluster A but nevertheless the contribution to the UGS is improving, and the gas-saturated area in the reservoir is getting better after every storage cycle. Overall, the data highlights the growing importance of underground gas storage in responding to seasonal demand variations. The steady increase in total storage peaks underscores the system’s expanding capacity and adaptability. The steady increase in total gas storage volumes, with the total volume of gas almost doubling over the observed period, is not the result of a single, isolated event. Instead, it reflects a combination of factors that have collectively enhanced the storage capacity of the system.

3. Methods

Here, we briefly illustrate the methods utilized to analyze the data recorded with different techniques and instruments. We also present the methodological approach to correlate injection/production data from the UGS and the main technologies used to obtain a solid dataset. This study utilized datasets collected during underground gas storage (UGS) operations from 2018 onwards. Data acquisition involved:

• Operational gas injection and withdrawal records analyzed with a hysteresis plot.
• Surface deformation data acquired using the Sentinel-1 satellite (SAR data analysis).
• Seismic monitoring data from surface and downhole systems.
• Subsurface measurements from Pulsed Neutron Logs (PNL).

3.1. Hysteresis Plot

A P/Z plot (or hysteresis plot) is a graphical tool commonly used in reservoir engineering to analyze gas reservoirs, especially in terms of storage management [19,20]. It is particularly useful for understanding the behavior of underground gas storage (UGS) facilities as they evolve over time. P/Z is the ratio between bottomhole pressure and the gas compressibility factor Z (as defined in Equation (1)). By plotting P/Z versus cumulative gas volumes, reservoir engineers can verify whether the total gas remains in place and assess how the reservoir pressure responds to gas injection and withdrawal cycles over time. The observed yearly trends exhibit typical seasonal cyclic patterns, corresponding to periods of gas injection (higher P/Z values, associated with higher pressure conditions) and withdrawal (lower P/Z values). Each operational cycle appears as a distinct trajectory on the graph, indicating variations in operational conditions or in the reservoir response, which may be influenced by operational management, geological heterogeneity, or evolving reservoir characteristics. Such monitoring is particularly important during the ramp-up phase, when a gradual deformation of the hysteresis cycles can be observed—the slope from the origin tends to decrease—as new reservoir areas, each with its own properties, become saturated with gas. The term Z of the ratio represents the gas compressibility factor, which is a crucial parameter in gas reservoir engineering that quantifies how much the behavior of real gas deviates from that of an ideal gas under given conditions of temperature and pressure. It is defined mathematically as:

$$\mathrm{z}={\displaystyle \frac{\mathrm{P}\mathrm{V}}{\mathrm{n}\mathrm{R}\mathrm{T}}}$$

(1)

where:

• P is the absolute pressure (Pa).
• V is the gas volume (m³).
• n is the number of moles of gas (mol).
• R is the universal gas constant (J/mol·K).
• T is the absolute temperature (K).

For an ideal gas, Z equals 1. However, real gases deviate from this ideal behavior due to molecular interactions and varying gas properties. Thus, for reservoir engineering purposes, the Z factor corrects the ideal gas law to accurately represent the physical properties of gases under reservoir conditions. It is a key element for calculating the true volume of gas under the pressure and temperature conditions encountered in gas storage operations. In reservoir management, the P/Z relationship is relevant because it helps in understanding reservoir depletion behavior, checking the total gas in place, and predicting future performance during injection and withdrawal phases. Lower Z factors at higher pressures reflect greater deviations from ideal gas behavior, typically due to higher compressibility and molecular interactions at reservoir conditions.

3.2. SAR (Synthetic Aperture Radar) Analysis

The analysis employed the Small Baseline Subset (SBAS) technique, processing datasets from ascending and descending satellite orbits. The investigation area extended approximately 30 km from the Cornegliano Laudense gas storage site, encompassing the core operational zones, as well as surrounding regions, to capture broader geological and anthropogenic influences. This area (the SAR dataset covers an area of about ~190 km × 70 km) aligns with the defined monitoring zones, including the Internal and Extended Domains, as stipulated in regulatory guidelines by the Italian Ministry of the Environment and Energy Security (formerly MiSE-UNMIG, 2014). High spatial resolution was achieved, with deformation maps produced at a ground pixel resolution of approximately 20m × 20 m providing detailed insights into surface displacement patterns. The data briefly summarized here are discussed in greater detail in [21]. These datasets provided complementary views of surface displacements. The SBAS analysis generated geocoded maps of deformation velocities, expressed in centimeters per year, which were overlaid on high-resolution optical imagery.

3.3. Surface Seismic Monitoring

The analysis of microseismicity around the Cornegliano Laudense gas storage site leverages a comprehensive monitoring system to detect and characterize small-scale seismic events. The monitoring and data analysis have been carried out by OGS on behalf of IGS and are illustrated in detail in [21]. The Cornegliano Laudense Monitoring Network (RMCL) comprises ten seismic stations, equipped with advanced borehole seismometers deployed at an average depth of 75 m. This configuration minimizes surface noise and anthropogenic disturbances, ensuring the capture of high-quality seismic data. The monitoring period extends from January 2017 to April 2024, encompassing both baseline conditions and various operational phases of the gas storage site. The monitoring area is divided into (i) the Internal Domain (DI) with a radius of 3 km around the reservoir (~28 km²), representing the zone of highest operational influence, (ii) the Extended Domain (DE) with a radius of 15 km (~706 km²), and (iii) the External Area (EA), defined as a 30 km radius (~2827 km²) from the site (Section 4.2.2). Data coming from DI, DE, and EA are expected to provide seismicity data coming from the reservoir directly, the operational area, and the natural tectonic-related seismicity, respectively. Seismic data has been processed using a dual approach, whereby the real-time system employs automated algorithms to detect, classify, and locate seismic events with minimal delay, triggering alerts for events exceeding a pre-defined magnitude threshold. For events of particular interest, data are manually reprocessed and refined. The monitoring and analytical activities are described in detail in [21].

3.4. Downhole Seismic Monitoring

The Cornegliano Laudense storage site is also equipped with a downhole seismic monitoring acquisition system installed in observation wells A04 and B03. The monitoring system is designed for real-time monitoring of low-magnitude seismic events associated with underground gas storage operations. Developed by Schlumberger Italiana S.p.A. at the request of Ital Gas Storage, this system represents an advanced infrastructure for geophysical surveillance of the subsurface. The system consists of an array of three-component geophones installed within the 9 5/8” casing in the two monitoring wells. Each array includes five sensors, spaced at 100 m intervals between 800 m and 1200 m in depth. Their arrangement has been optimized to detect microseismic events with a magnitude lower than 1.5 throughout the entire reservoir. The acquired data is transmitted in real-time to a dedicated IGS processing server, either at the Schlumberger offices or on the cloud, through a secure VPN connection. The data processing workflow is automated through Earthworm software (version 7.10), which performs microseismic event detection, automatic picking of P and S waves, and preliminary hypocenter determination. Subsequently, the events are manually verified and reprocessed using Schlumberger Petrel (version 2024) software, ensuring detailed and precise seismic characterization. The system’s output is visualized through the Seismic Focal Point (SFP), an interface that allows real-time consultation of detected events, with both automatic and manual updates. The SFP provides detailed information on the location, magnitude, and depth of events, along with their representation on georeferenced maps. The recorded seismicity is categorized, like the Cornegliano Laudense Monitoring Network (RMCL), into different domains based on distance and depth criteria, including an Internal Domain (Type A, 3 km radius), an Extended Domain (Type B, 15 km radius), and a broader Extended Area (Type C, 30 km radius). The microseismic monitoring reports confirm that all detected seismicity falls within natural background activity, regional tectonic movements, or anthropogenic noise unrelated to storage operations. This reinforces the reliability of the monitoring infrastructure in distinguishing between natural and induced seismicity. The use of downhole seismic arrays, combined with advanced waveform analyses, provides precise determinations of hypocenter depths, essential for correctly identifying fault systems involved in induced seismicity. Similar advanced waveform analyses applied to the Castor Underground Gas Storage (UGS) in Spain allowed accurate identification of fault reactivation occurring at depths significantly deeper than the injection zone, highlighting the effectiveness of waveform-based methodologies in improving event characterization [22]. Combined surface-downhole microseismic monitoring is a more effective approach in characterizing fracture geometry, dimension, and orientation, providing high-resolution insights compared to traditional single-level monitoring [23].

3.5. Pulsed Neutron Log

Pulsed Neutron Log (PNL) is an advanced technique widely utilized in the oil and gas industry to evaluate subsurface reservoir properties, particularly within cased wells [24]. The PNL tool is particularly valuable for monitoring underground gas storage reservoirs. It provides quantitative detection of gas presence and can also identify the mineralogy of the reservoir, as well as deposits within or around the well. Technology hinges on the principle of injecting high-energy neutrons into the surrounding formation and analyzing their interactions with formation nuclei in the matrix and fluids. These interactions yield critical insights into lithology, porosity, fluid volume, and saturation, making PNL a key tool for reservoir characterization and monitoring. PNL tools employ a pulsed neutron source, typically a deuterium-tritium (D-T) accelerator, to generate bursts of fast neutrons with an energy of approximately 14 MeV. Once emitted, these neutrons penetrate the formation and undergo a series of interactions. Initially, high-energy neutrons lose energy through elastic scattering, a process predominantly controlled by collisions with light nuclei, such as hydrogen. This energy loss ultimately reduces the neutrons to thermal energies. Subsequently, thermalized neutrons are captured by atomic nuclei, triggering the emission of gamma rays with characteristic energies specific to the capturing elements. Additionally, during the high-energy phase, inelastic scattering occurs, where neutrons excite atomic nuclei and induce the emission of gamma rays at unique energy levels associated with individual elements. Alongside the analysis of gamma rays’ elemental energy spectrum, the decay rate of thermal neutrons, measured as a function of time after the neutron pulse, provides additional insights into reservoir properties. The PNL tool is equipped with gamma-ray scintillator detectors to capture the energy spectra resulting from both inelastic scattering and neutron capture interactions. These measurements provide Log outputs sensitive to fluids and elemental concentration data for the surrounding formation. One of the primary parameters derived from this data is the macroscopic capture cross-section (Σ), which quantifies the efficiency of neutron absorption by the formation. This parameter is closely related to the formation’s fluid characteristics; among the nuclei existing in the formation, the chlorine is effective in capturing neutrons, hence Σ is related to the amount of water chlorine. Another critical measurement is the hydrogen index (HI), which is proportional to the amount of hydrogen atoms in the formation and serves as an indirect indicator of gas presence and changes. The presence of hydrogen in the fluid-filled pore spaces significantly affects neutron moderation, making PNL particularly effective in differentiating fluid types. The fast neutron cross-section measurement, as described later in the paragraph, is a rock property sensitive to the atomic density of the material, hence highly sensitive to gas. In combination with HI and the other PNL log outputs, it enables a quantitative assessment of gas volumes and saturation. The versatility of PNL lies in its ability to operate in cased wells, where conventional logging techniques are often ineffective. This capability is particularly valuable for monitoring reservoir dynamics over time, such as changes in fluid saturation due to water flooding, gas breakthrough, or other secondary recovery processes. Time-lapse logging with PNL enables operators to track these changes and optimize hydrocarbon recovery strategies (see Section 4.2.4). The use of time-lapse Pulsed Neutron Logging (PNL) for evaluating gas saturation is well established. In particular, the combined use of PNL neutron porosity and the Fast Neutron Cross-Section (FNXS) with the elemental concentrations for matrix correction has been successfully applied to assess residual gas saturation in cased wells [25]. Similarly, the GSEP (gas saturation evaluation parameter) parameter introduced by [26] offers a robust alternative by improving sensitivity to gas saturation variations even in complex lithologies. Moreover, the tool’s ability to identify elemental signatures facilitates lithological analysis and enables converting the ratio between carbon and oxygen into oil volume and saturation independently of salinity, when rock porosity is known. This aids in the differentiation of formations and the evaluation of their productive potential. Unlike thermal neutron capture cross-section, for which certain isotopes have extremely high values (such as Cl, B, and Gd), fast neutron cross-sections of all isotopes are similar. Thus, the Fast Neutron Cross-Section (FNXS) is approximately proportional to atom density. Therefore, while atom density is distinctly different from bulk density, this new nuclear property has similar utility in quantitative petrophysical interpretation as the bulk density (gamma–gamma density measurement). The mineral matrix and water have similar FNXS values. In contrast, due to its low atom density, gas has a significantly lower FNXS value. This is because, in terms of fast neutron attenuation, all the different elements have similar cross-sections, and hydrogen does not dominate as it would for capture cross-section, and the atomic density effect dominates. Thus, FNXS is not very sensitive to liquid-filled porosity variations but very sensitive to gas-filled porosity variations. It facilitates the computation of accurate formation porosity in gas-filled formations together with neutron porosity because its response contrasts with the traditional neutron porosity type response dominated by hydrogen [27]. The raw measurement provided by the tool is the net inelastic gamma-ray count rate (GRAT) measured in the farthest-spaced detector. The GRAT is characterized based on laboratory data and normalized to be transformed into formation FNXS. Gain and offset are applied to correct it for environmental effects; a manual adjustment may be required for the most accurate FNXS measurement, because the actual logging conditions are often different from the laboratory-based data. The alignment of temperature profiles and Joule–Thomson cooling effects with PNL measurements further validates the interpretation, ensuring a comprehensive understanding of reservoir dynamics. As an additional application, when used repeatedly, the PNL tool can indirectly detect changes in the quality of the gravel pack. The Fast Neutron Cross-Section (FNXS) is a formation nuclear property independent of thermal and capture cross-sections and highly sensitive to gas, regardless of the hydrogen index. As shown in Equation (2), FNXS is defined as the ability of the formation to interact with incident fast neutrons [28], and it is strongly inversely proportional to the inelastic gamma ray count rate.

$$\mathrm{F}\mathrm{N}\mathrm{X}\mathrm{S}={\mathsf{\Sigma }}_{\mathrm{e}\mathrm{l}\mathrm{a}\mathrm{s}\mathrm{t}\mathrm{i}\mathrm{c}}@14\mathrm{M}\mathrm{e}\mathrm{V}=\mathsf{\sigma }\ast \mathrm{A}\mathrm{t}\mathrm{o}\mathrm{m} \mathrm{D}\mathrm{e}\mathrm{n}\mathrm{s}\mathrm{i}\mathrm{t}\mathrm{y}$$

(2)

where Σ is the macroscopic elastic cross-section at 14MeV defined as the product of atom density and microscopic cross-section (σ) as shown in the equation (standard unit of cm⁻¹).

4. Results

4.1. Operational Gas Storage Analysis

The P/Z plots for Cluster A and Cluster B (Figure 6) illustrate the evolution of reservoir behavior over several years, spanning from 2018 to 2025. These plots capture the relationship between pressure, compressibility, and total gas in storage, offering insights into the operational performance and long-term dynamics of underground gas storage facilities. For Cluster A, the P/Z values exhibit a clear positive trend as the total gas in storage increases, reflecting a consistent evolution over the years. Over time, the data points progressively shift upward and to the right, as a demonstration of better operational performance. At low inventory levels, a minimum P/Z value is consistently observed across the analyzed operational cycles, remaining substantially stable over time and highlighting the critical role of cushion gas (CG) in preventing P/Z from dropping below an operational threshold in terms of minimum bottom hole pressure, typically in the range of 4–5 MPa (40–50 bar), which could otherwise lead to complications in gas production, thereby maintaining pressure stability and deliverability when lower inventories are reached. The yearly progression of the P/Z curves suggests a gradual improvement in reservoir behavior in terms of the total volume of gas stored, possibly driven by gas saturation evolution. The overall trend indicates a stable and increasingly efficient gas storage facility capable of adapting to operational demands. In Cluster B, a similar positive trend is evident, although the range of total gas in storage is narrower compared to Cluster A. Like in Cluster A, the P/Z values dip to a minimum at low inventory levels, underscoring the importance of cushion gas for maintaining stability. The yearly evolution of P/Z values in Cluster B reflects controlled and predictable operational performance, with the data points forming a more compact trend compared to Cluster A. This suggests that Cluster B operates within tighter constraints, due to its smaller storage. When comparing the two clusters, Cluster A shows a higher storage capacity. Despite this difference, both clusters show the critical role of cushion gas in stabilizing reservoir conditions at low inventories. Over the years, the P/Z trends in both clusters indicate a stabilization of reservoir behavior and optimization of performance. This evolution results from improved reservoir management, operational adjustments, and the natural stabilization of reservoir dynamics, as already discussed in the previous paragraph. Overall, the P/Z plots for both clusters reveal the adaptability of the UGS facilities to changing operational conditions. In 2019–2020, as illustrated in Figure 6b, Cluster B exhibits either no data or only sparse records. This is attributable to plant modifications carried out in that period, during which storage operations were suspended.

4.2. Monitoring Activities

All underground activities, particularly those related to gas storage, require careful monitoring and control to ensure operational safety and prevent undesirable impacts on natural site conditions, including alterations to the local stress field. While seismicity and ground deformation are traditionally critical factors monitored to manage emergencies or unforeseen events effectively, IGS had already proactively implemented comprehensive monitoring initiatives well before the official start of operations, going beyond the mandatory legislative requirements. On a voluntary basis, IGS initiated extensive, detailed monitoring of seismic activity and surface deformation during the pre-operational phase.

4.2.1. Ground Deformation Analysis

The monitoring activities presented in this section are based on the methodology and results described in [21], which provides the framework and supporting data for the analyses reported in Section 4.2.1 and Section 4.2.2. The monitoring of ground deformations associated with the Cornegliano gas storage site utilized Synthetic Aperture Radar (SAR) data from Sentinel-1 satellites, spanning the period from March 2015 to April 2024. The map (Figure 7) revealed consistent patterns of surface displacement in the vicinity of the Cornegliano Laudense reservoir, particularly around Cluster A, the primary operational site. Vertical displacement rates showed localized uplift of up to +2.4 mm/year and subsidence in adjacent areas, while horizontal components indicated shifts predominantly toward the northwest at rates reaching +2.0 mm/year in the ETRF14 reference frame. Temporal analysis identified progressive deformation trends well correlated with the operational cycles of gas injection and extraction, but the deformations are all within the expected ground deformation models carried out prior to the storage activities. In general, long-term data suggests a controlled relationship between reservoir operations and surface dynamics over the study area; nevertheless, localized anomalies were detected with deformation rates slightly deviating from the regional patterns, suggesting interactions with minor geological heterogeneities or localized operational influences. To ensure reliability, the InSAR-derived measurements were rigorously validated against data from the on-site Global Navigation Satellite System (GNSS) stations. For instance, the GNSS station at Lodi city recorded a vertical displacement rate of +2.3 mm/year and horizontal motion at +2.0 mm/year toward the northwest, closely matching the InSAR results. Across the dataset, discrepancies between InSAR and GNSS measurements were minimal, generally within ±0.2 mm/year for horizontal components and ±0.5 mm/year for vertical components, underscoring the robustness of the satellite-derived deformation maps.

Figure 7a represents a magnified section of the deformation velocity map of the area of interest. The velocities are expressed in centimeters per year and reflect displacements in the radar’s line-of-sight (LOS). The central feature of the map is the gas storage site, designated as Cluster A and marked by a red star with a white border. Around this site, distinct deformation signals are present, with patterns of uplift and subsidence that align with the operational activities of the reservoir. The figure also includes temporal graphs for four specific points of interest (P1, P2, P3, and P4) within the analyzed area (Figure 7b–e). The graphs illustrate the time evolution of surface displacement at each point, measured in LOS. By plotting the deformation trends over the observation period, the figure highlights how ground movement correlates with the operational cycles of gas injection and extraction. Moreover, a deformation effect is clearly identifiable in the area of Turano Lodigiano (Figure 7e, P4), located 15 km from the storage site and therefore independent of it. This observation confirms the results from previous analyses, showing a substantially linear deformation trend with a rate ranging from 0.5 cm/year to 1 cm/year. Overall, Figure 7 offers a detailed and localized view of ground deformation, leveraging high-resolution spatial and temporal data to reveal the dynamic responses of the area to storage activities. The image and associated graphs underscore the precision of the SBAS technique and its ability to monitor subtle surface changes over time.

4.2.2. Seismic Monitoring Results

During the monitoring period from November 2023 to April 2024, three seismic events were detected. These included one event in the Extended Domain and two in the External Area beyond 30 km (Figure 8). The most significant event, recorded on 25 February 2024, had a Local Magnitude (ML) of 2.6 and occurred at a depth of 28.5 km. The Local Magnitude (ML) is calculated using the attenuation formula of [29]. This event was attributed to tectonic activity unrelated to gas storage operations, given its depth and spatial location. The two remaining events, with magnitudes of ML 0.9 and ML 2.2 and depths of 1.9 km and 29 km, respectively, were consistent with shallow seismic activity frequently observed in the region. These findings align with historical seismicity patterns and appear not to be related to operationally induced seismicity. A statistical analysis of seismicity from 2017 to 2024 revealed no changes in frequency and magnitude associated with gas storage activities. The frequency, magnitude, and spatial distribution of detected events, in fact, remained within the expected natural variability. Furthermore, no discernible correlations were found between seismic event occurrence and the gas injection or withdrawal cycles, underscoring the controlled nature of reservoir operations.

4.2.3. Downhole Seismic Monitoring Results

The results presented in this section refer to the downhole seismic monitoring conducted at the Cornegliano Laudense storage site, aimed at detecting and classifying seismic events occurring at different depths and spatial domains in the vicinity of the reservoir. The monitored magnitudes (

Table 2. Summary of the recorded microseismic events.

Event Type	Recorded Magnitude	Recorded Depth	Events per Year (Range)
Type A	0.8–1.2 (2024)	4.4–4.6 km (2024)	0–2 events
Type B	−0.8 to 1.8	2.4–8.0 km	5–38 events
Type C	>1.6	4.1–44.7 km	88–148 events

) for Type A events in 2024 ranged from 0.8 to 1.2 at depths between 4.4 and 4.6 km, confirming minimal seismic activity in the vicinity of the reservoir. Type B events exhibited greater variability, with magnitudes ranging from −0.8 to 1.8 and depths between 2.4 km and 8.0 km, reflecting natural variability typical of regional tectonics. Type C events showed a broader magnitude spectrum (1.6 to 7.9) and deeper seismic activity (up to 44.7 km), correlating with regional geological dynamics rather than operational activities. A detailed statistical analysis indicated that increases in the number of Type B and C events observed during specific periods, particularly in 2024, remained within historical natural variations and did not correlate with operational phases such as gas injection or withdrawal. This demonstrates that the storage operations at Cornegliano Laudense have not induced significant seismicity nor altered the natural seismic background. Overall, the downhole seismic monitoring system effectively distinguishes between natural seismicity, anthropogenic noise, and potential operationally induced events. Its deployment has significantly improved the capability for real-time geophysical surveillance, contributing to the safe management of the storage site. In addition to automatic and manual data processing, the monitoring system includes a resolute team of seismic experts responsible for immediate notification to IGS in case of detection of anomalous seismic events. This ensures rapid evaluation and timely operational response to guarantee the continued safety and stability of the UGS operations.

4.2.4. Pulsed Neutron Log Results

In Cornegliano UGS, the PNL tool was first deployed during the early stages of the UGS conversion in 2019 in wells A02, A03, A05, and A06, after the first Log campaign performed just after the drilling phase. This initial run occurred under similar pressure conditions at the end of the injection campaign. During the 2021–2022 campaign, the PNL tool was used in wells A03, A05, A06, and B04. The choice to exclude A02 this time was due to its improved water behavior observed during the 2020–2021 campaign, while B04 was selected due to its atypical behavior, being completed only in the C1 layer yet showing some connections with other wells, likely from cluster A. It is noteworthy that the PNL tool is also run alongside Pressure and Temperature gauges (PT), which provide valuable data on gas/water interfaces. The most recent PNL campaign was conducted at the end of the injection cycle in October 2023, targeting wells A03, A06, B04, and B06. The possibilities for comparison are outlined in

Table 3. Pulsed Neutron Log measurement campaigns.

	A02	A03	A05	A06	B04	B06
2017	x	x	x	x	x	x
2019	x	x	x	x
2021		x	x	x	x
2023		x		x	x	x

, indicating that full comparisons can be made for A03, A06, and B04. For well B06, this was the first time the PNL tool was used, which was necessary to assess its improvement over time. One of the key insights from the 2023 survey is the detection of the gas/water interface, which was found to be 10 m lower compared to the 2021 survey. This observation confirms the consistent improvement of the well over the course of the campaign, as it is now impacted by water only at low inventory levels.

Figure 9 clearly illustrates this progressive saturation through the relationship between effective porosity, represented by the blue curve (the maximum theoretical porous volume available for gas), and actual gas saturation, depicted by the expanding red areas. The observed increase in gas saturation approaching the theoretical limit confirms that reservoir capacity has been effectively optimized. The data reveals the presence of gas across multiple layers, with distinct trends observed over time. Zone A exhibits consistent gas saturation, indicating stable storage conditions. Zone B demonstrates a steady increase in gas volume and saturation over the period analyzed. In Zone C, dynamic changes are evident, with gas saturation increasing in 2023. Zone D shows the lowest FNXS and neutron porosity values in 2023, reflecting an improvement in gas storage capacity.

5. Discussion

The integrated analysis performed at the Cornegliano Laudense natural gas storage facility between 2018 and 2025 has provided key insights into reservoir behavior, surface stability, and microseismicity. Data obtained from Pulsed Neutron Logging (PNL) confirmed a significant and progressive displacement of formation water by injected gas, particularly evident in deeper sandstone intervals (zones C1 and C2). The consistent lowering of the gas–water contact by approximately 10 m between 2021 and 2023 highlights improved gas saturation and efficient use of the available pore space without compromising geological stability. The robustness of the observed gas saturation trends is supported by the literature [26], which has demonstrated that such responses are consistent with fluid displacement under cyclic gas injection, even in low-porosity or high-clay content formations. This positive evolution is further supported by the trends in the P/Z plot, where the decreasing slope over time indicates enhanced storage efficiency. As injected gas progressively occupies more pore space, the compressibility-adjusted pressure (P/Z) decreases, reflecting an increased storage volume available at a given reservoir condition. Two primary mechanisms underpin this capacity expansion. First, ongoing improvements in reservoir management and gas injection techniques have allowed for greater utilization of pore space. Second, the progressive lowering of the aquifer beneath the reservoir has reduced its counteracting pressure, thereby facilitating greater accommodation of gas. These combined effects have led to approximately a 2.5X increase in total storage capacity observed in recent years. Surface deformation and microseismic monitoring data further validate the safety and efficiency of storage operations. Sentinel-1 SAR and GNSS analyses consistently indicated negligible ground deformation and subsidence, which remained well within natural geological limits. The absence of significant deformation confirms operational pressures are safely maintained within the elastic limits of the reservoir and surrounding geological structures. Microseismic monitoring throughout the operational period corroborates these findings, with seismic events consistently aligning with regional tectonic background levels and exhibiting no direct correlation with storage activities. These outcomes align closely with the results of [5], who studied similar anticline structural traps hosting underground gas storage (UGS) fields across the Po Plain basin. Using SAR cross-correlation analyses [5], they established threshold parameters (R and K indexes), effectively distinguishing UGS-related deformation from background noise and identifying limited areas of influence constrained strictly within storage boundaries. The similarities between their results and the Cornegliano Laudense observations strongly support the general validity of these criteria. Moreover, structural and operational parallels between Cornegliano Laudense and the anticline traps analyzed by [5] emphasize geological consistency within the Po Plain, facilitating the extension of operational insights to comparable storage sites. Reservoir geometry and bounding faults notably emerge as critical factors influencing deformation patterns and reservoir management strategies, underscoring the importance of detailed geological characterization. Recent studies on other gas storage projects (e.g., the Castor UGS, Spain) have shown that induced seismicity can result not only from direct pore pressure effects but also from complex stress-transfer mechanisms, including aseismic slip triggered by buoyancy forces and transient poroelastic effects [30]. Specifically, this paper demonstrated how seismicity was triggered in hydraulically disconnected, critically stressed basement faults, highlighting the relevance of stress-transfer phenomena in assessing seismic risk. This versatile approach facilitates accurate identification of reservoir sections suitable for further gas volume expansion, ensuring storage enhancements without inducing critical deformation or seismicity. Continued application of such comprehensive monitoring (Figure 10) and modelling strategies is recommended for future studies and operational practices in analogous geological contexts.

6. Conclusions

The methods applied at the Cornegliano Laudense site yielded crucial findings and practical implications, contributing substantially to enhancing future reservoir management and operational planning. The main outcomes of this study are summarized as follows:

• Gas storage capacity has effectively doubled since the beginning of operations, primarily through efficient displacement of formation water, without structural alterations to the geological framework.
• Integrated monitoring approaches combining satellite-based surface deformation tracking and microseismic monitoring have confirmed reservoir stability throughout operational phases.
• Surface deformation and seismic events have consistently remained within safe and predictable natural background limits.
• Observed geomechanical stability across multiple operational cycles demonstrates the intrinsic robustness of the reservoir, with no significant compaction or induced seismicity recorded.

Comparative studies with similar geological and operational contexts are recommended to validate and generalize the insights gained from the Cornegliano Laudense experience. Specifically, we recommend adopting multidisciplinary monitoring systems, performing detailed comparative analyses of reservoir dynamics and operational responses, and, where possible, establishing collaborative data-sharing platforms among UGS operators. In conclusion, the described approach serves as a robust benchmark for similar geological and operational scenarios, ensuring optimal reservoir performance and long-term environmental compatibility.