A natural hydrogen carbonate play: Sandino Forearc Basin, Pacific offshore Nicaragua Open Access

Offshore Nicaragua occupies the central segment of the Central American convergent margin where the Cocos oceanic plate is subducting beneath the leading edge of the Caribbean Plate (Mann, 2007) (Figure 1). The Nicaraguan sector of the Central American convergent margin extends from the subduction trench (Middle America Trench) to the Central American Volcanic Arc and includes the offshore and onshore Sandino forearc basin. The basin is separated from the active subduction trench by the Trench Slope and the so-called ‘Outer High’ (Figure 2). The Tempisque Basin forms the onshore southern projection of the forearc basin, and the exposed ophiolitic complexes of the Santa Elena and Nicoya Peninsulas of northern Costa Rica can be viewed as a continuation of the outer high system (Figure 2).

We propose that the generation of natural hydrogen will occur (or has occurred) along this convergent margin due to the action of subduction-related aqueous fluids (by serpentinization) on ultramafic rocks in the ‘upper plate’ (Hyndman and Peacock, 2003; Kawamoto et al., 2013; Vitale Brovarone et al., 2020). That this process has taken place along the Central American convergent margin is evidenced by serpentinized ophiolitic rocks encountered along the trench slope in DSDP drilling offshore Guatemala (see Leg 84, DSDP, 2007a).

In this paper, the natural hydrogen potential of the Sandino forearc basin is investigated through the hydrogen system generated in the forearc setting, and a shallow-water carbonate play is developed for exploration in the forearc basin.

The understanding of forearc architecture and evolution owes much to the work of the University of Texas Institute for Geophysics e.g., inter aliaWalther et al. (2000), Ranero et al. (2000), and Stephens (2014). Walther et al. (2000) show that the ‘upper plate’ underlying the outer high and trench slope is composed of high-velocity material and therefore is unlikely to be solely the product of accretionary transfer of unconsolidated abyssal plain and trench sediments. They called this the ‘margin wedge’ and interpreted it as an accreted or partially subducted oceanic plateau responsible for choking-off subduction in the early Tertiary.

The Caribbean Large Igneous Province (CLIP) that makes up most of the present-day Caribbean formed as an oceanic plateau between 139 and 90 Ma by passing over the Galapagos hot spot (Mann, 2007). The key to the tectonic history of the Central American convergent margin and the Sandino Basin lies in understanding the nature of the forearc basement terrains and their relationship to the CLIP.

The crust underlying most of Nicaragua was considered to be continental and was designated the Chortis Block (e.g., Pindell and Dewey, 1982). The separate Chorotega Block underlying Costa Rica and Panama (such as Nicoya and Santa Elena Peninsulas) was thought to have formed by obduction of an oceanic island arc complex onto the trailing edge of the CLIP during the Paleogene (Mann, 2007). In contrast, Baumgartner et al. (2008) and Baumgartner et al. (2012) propose the name Mesquito Composite Oceanic Terrane (MCOT) for the crust comprising the southern half of the Chortis Block underlying Nicaragua. In their model, the Santa Elena Peninsula, with its ultrabasic nappe emplaced on the Santa Rosa Accretionary Complex (Baumgartner and Denyer, 2006), forms the southern edge of the MCOT (Figure 2a).

The ophiolitic complexes of the Nicoya Peninsula were historically considered an obducted part of the CLIP, but subsequent work revised this view. Flores (2005) distinguishes three terranes on the Nicoya Peninsula: the Nicoya Complex sensu stricto, the Matambu, and Manzanilo Terranes. In their model, the Manzanilo Terrane was the original forearc, and the Nicoya Complex and Matambu Terranes collided with the forearc in the Late Cretaceous. In the Baumgartner et al. (2008) model, the Nicoya Complex sensu stricto first accreted to MCOT and then became reworked into the Nicoya Plateau in the Late Cretaceous. The Nicoya Plateau in this model therefore formed along the edge of the MCOT and was independent of the CLIP. In this way, Baumgartner et al. (2008) called all the elements of the Nicoya Peninsula the “Pre-CLIP Plateaus” (Figure 2a).

The Sandino Basin formed as a forearc basin (inter aliaRanero et al., 2000; Walther et al., 2000; Stephens, 2014) following the accretional events forming the Nicoya Complex in the Middle Campanian (Andjic et al., 2018). The margin was convergent from the Late Cretaceous through the early Tertiary, which corresponds to the main phase of forearc basin development. Uplift of the outer high, more prominent in the north in the form of the Corinto High, may have been the result of subduction alone (Ranero et al., 2000) or partly related to accretion or collision of an oceanic plateau some time in the early Tertiary (Walther et al., 2000; Stephens, 2014). Andjic et al. (2018) place the collision in the early to mid-Paleocene. Following this reconfiguration, subduction was renewed in the late Oligocene to early Miocene, and the convergent margin entered a phase of dominant extension during which some workers invoke Cocos slab roll-back (Mann, 2007; Stephens, 2014) and even slab detachment (Mann, 2007). Intermittent episodes of compression or strike-slip movement affecting the basin were induced by oblique convergence and rough crust subduction (Andjic et al., 2018).

The generic hydrogen system provides a framework for understanding the hydrogen systems generated by serpentinization in a variety of geologic settings (Figure 3). In the generic hydrogen system, two linked subsystems are described:

• Source-generation subsystem: The source rock (or protolith) is located in ultramafic basement rocks and therefore geologically separate from the sediments the hydrogen will migrate into. It requires a supply of water in the form of meteoric water, seawater, or subduction-derived fluids, depending on the geotectonic setting, and a plumbing system allowing the aqueous fluids to access the protolith (Jackson et al., 2024).

• Migration-retention subsystem: Hydrogen is generated geologically instantaneously by serpentinization and accesses the basin via an advective link either through faults or through fracture networks. Once inside the basin, hydrogen will migrate and become sealed and trapped in the same way as hydrocarbon gases (Hutchinson et al., 2024).

Here, we investigate the elements of a potential hydrogen system produced by serpentinization along the Nicaraguan sector of the forearc region of the Central American convergent margin.

The protolith comprises ultrabasic rocks within the forearc basement of the upper plate. Several previous geophysical investigations of the deep crustal structure of the forearc have been made (see inter aliaWalther et al., 2000; Cailleau and Oncken, 2008; Salares et al., 2013; Lücke and Arroyo, 2015). These studies have helped to elucidate the interplate boundary; the geometry of the subducting slab; and the nature of the basement of the trench slope, outer high, and Sandino Basin.

Walther et al. (2000) were the first to interpret a dense (high-velocity) basement underlying the outer high and trench slope. They called this the margin wedge, divided it into upper and lower parts, and postulated an origin as an accreted or partially subducted oceanic plateau. They also modeled a Mesozoic margin wedge under the main part of the Sandino Basin and a deeper elevated high-density mass underlying the flanks of the basin, which they interpreted as a mantle sliver. Salares et al. (2013) also obtain mantle-like seismic velocities, supporting the interpretation of a shallow mantle wedge beneath the Sandino Basin, at its shallowest only 10 km from the base of the basin. They also speculated that the mantle has undergone partial serpentinization on the basis of the velocities. In addition, the position of the downdip limit of the seismogenic zone below the Nicaraguan forearc was explained by Salares et al. (2013) as the hydration of the mantle wedge rocks by serpentinization. The low velocity of the mantle wedge was also attributed to serpentinization by DeShon et al. (2006).

For this study, we have investigated the nature of the forearc basement (beneath the trench slope, outer high, and Sandino Basin) by quantitative modeling of gravity along two regional seismic lines where the depth to basement could be constrained. Control on deeper structure and densities was provided by the previous studies of Walther et al. (2000), Cailleau and Oncken (2008), Salares et al. (2013), and Lücke and Arroyo (2015). The potential field data used in the interpretation were:

• Sandwell and Smith Gravity Model (Free Air) (map shown in Figure 4) (Sandwell et al., 2014)

• Gravity grids (free air, Bouguer, isostatic) provided by Getech

An acceptable fit between observed and calculated gravity was obtained along both profiles by invoking a crustal layer with densities varying between 2.7 and 2.9 gm/cm³ above a mantle wedge with a density of 3.3 gm/cm³ (Figure 4). The densities in the crust were varied laterally in discrete segments to finely tune the observed gravity response, but the model is more sensitive to the elevation of the mantle. A mantle body like Walther et al.’s mantle sliver is a requirement for the modeled profiles (Figure 4).

In the modeling, a distinction is made between the crust under the main basin and that under the outer high and trench slope corresponding to that made between the margin wedge and the Mesozoic margin wedge by Walther et al. (2000). The crust below the trench slope and outer high is attributed to the oceanic plateau (of Walther et al., 2000) that caused the most recent choking of subduction in the early Tertiary and may be represented by the ophiolitic rocks (now partially serpentinized) encountered in DSDP drilling off Guatemala (DSDP, 2007a). The nature and origin of the crustal type under the main basin are based on regional geology and the understanding of convergent margin evolution (described previously). This suggests that the basement extending under the Sandino Basin represents the outer (most recently accreted) parts of the MCOT of Baumgartner et al. (2008) and Baumgartner et al. (2012), including the equivalents sensu lato of the Santa Elena Complex (Ultrabasic Nappe and Santa Rosa Accretionary Complex) (Figure 2a). According to Baumgartner et al. (2008) the Pre-CLIP Plateau terranes of the Nicoya Peninsula do not extend northward under the Sandino Basin (Figure 2a).

This provides a range of potential protoliths in the upper plate across the convergent margin from mantle rocks of the sliver, or wedge, which reach to within 10 km of the base of the basin to the ophiolitic rocks of the MCOT, which may include paleo-mantle ultramafics such as the outcropping Santa Elena peridotite.

In the subduction-forearc setting, three serpentinization scenarios are recognized: (1) at the outer rise as a result of seawater percolating down faults formed by downbending, (2) self-generated by fluids within the subducting slab (slab serpentinization), and (3) in the forearc resulting from subduction-generated aqueous fluids percolating upward and serpentinizing the mantle sliver or accreted plateau above (Martin et al., 2020; Vitale Brovarone et al., 2020). Scenario (3) is the most directly responsible for the formation of a hydrogen system in the upper plate of the forearc and consequently the overlying basin sediments.

Studies show that large volumes of aqueous fluids are released upward by dehydration reactions in the subducting oceanic crust and sediments (Kawamoto et al., 2013; Wilson et al., 2014). Over several tens of millions of years, it is estimated that fluid fluxes from the subducting plate could be sufficient to serpentinize the entire forearc mantle wedge (Hyndman and Peacock, 2003). However, fluid infiltration is likely to be fracture-controlled and the serpentinization heterogeneous (op. cit.). Evidence for serpentinized forearc mantle has been reported from a number of subduction zones based on geophysical investigations and numerical simulations (Hyndman and Peacock, 2003; Abers et al., 2017). Hydrogen is a product of this serpentinization (Motti et al., 2003; Vitale Brovarone et al., 2020).

Fluids are released from the dehydration of subducted sediments, altered basalts, and gabbros (of oceanic crust) and from previously formed serpentinites. The processes and temperatures involved in progressive aqueous fluid release from a subducting slab are listed by Hyndman and Peacock (2003). In subducted sediments, we find porosity collapse (compaction), opal to quartz transition (approximately 80°C), and smectite to illite conversion (approximately 120°C). In basalts of oceanic crust, we find porosity collapse (approximately 200°C–400°C) and garnet-forming dehydration (approximately 500°C). In addition to these processes, the breakdown of serpentinite formed by previous phases of serpentinization also contributes to slab dehydration and fluid release (e.g., Gorczyk et al., 2007a, 2007b).

The depth of dehydration depends on the stability field of the various hydrous phases and on the local thermal structure (Deschamps et al., 2013). Dehydration is mostly in the first 100–170 km of subduction with the main dewatering taking place from 300°C to 600°C and at P < 1.5 GPa (op. cit.).

The main controls on the thermal structure are the age of the incoming slab and the descent rate (which is controlled by the convergence velocity at the trench and slab geometry) (Peacock et al., 2005; van Keken and Wilson, 2023). The older (colder and denser) the subducting crust, the higher the convergence rate and the higher the slab dip, the colder the subduction zone (Syracuse et al., 2010), and the deeper and closer to the trench the dehydration will occur.

Northwest of the Nicoya Peninsula, relatively smooth Cocos crust created at the rapid-spreading East Pacific Rise (EPR) is being subducted. The thickness of the Cocos Plate increases from 5–7 km off Nicaragua to 12 km off SE Costa Rica. The convergence rate increases south–eastward from 60 mm/yr off Guatemala to 90 mm/yr off southern Costa Rica. It has also been shown that convergence is oblique such that the forearc is translating to the north–west at 10–15 mm/yr (DeMets, 2001).

The Central American subduction zone is characterized by rapid convergence of 15–25 Ma oceanic lithosphere with steep slab dip (off Nicaragua, Peacock et al., 2005 report a dip of approximately 84°.) This combination of factors classifies the Costa Rica–Nicaraguan sector of the Central American convergent margin as a relatively cool subduction setting (Syracuse et al., 2010).

Peacock et al. (2005) carry out thermal modeling across the Costa Rica–Nicaragua subduction zone to predict P-T paths of subducting lithosphere and the loci or types of slab dehydration reactions. The Costa Rica–Nicaragua subduction zone is a nonaccreting margin characterized by a very small frontal prism and subduction erosion. Calculated P-T paths followed by subducting oceanic mantle are significantly cooler beneath Nicaragua than off Costa Rica because the age and dip are greater to the north (Peacock et al., 2005). For example, at 3 GPa, modeled temperatures at a point 12 km below the slab interface within the subducting mantle range from 370°C beneath Nicaragua to 460°C beneath SE Costa Rica (op. cit.).

Rapid subduction results in isotherms within and near the subducting slab being depressed hundreds of kilometers into the mantle. Rapid subduction chills the subduction zone forearc, but temperatures in the mantle wedge (where the serpentinization is taking place) remain high as a result of the corner flow (Peacock et al., 2005). In cold subduction zones, high-pressure serpentinization may extend to greater depths but may be less pervasive, at least in the mantle wedge (Abers et al., 2017).

The water storage capacity of the subducting Cocos Plate is demonstrated by the degree of serpentinization in the subducting plate at the trench (see Ivandic et al., 2010) and seismic velocity studies that demonstrate extensive hydration of the subducting slab in the forearc (Abers et al., 2003; van Avendonk et al., 2011; Thorwart et al., 2013). Evidence for fluid-filled fracture zones above the subduction zone is provided by seismic studies (van Avendonk et al., 2010; Thorwart et al., 2013; Salares et al., 2013).

Faulting across the trench slope has been linked to seepage of deep-sourced fluids arising from the dehydration of clays along the plate boundary and normal faults provide fluid-flow paths for the water to rise from the plate boundary to the seafloor (Salares et al., 2013).

A model is presented for a forearc hydrogen system along the Nicaraguan sector of the convergent margin (Figure 5).

Fluid release from the subducting slab would have been active from the onset of the current phase of subduction beginning in the late Oligocene. With rapid plate convergence, it can be expected that fluid release begins relatively quickly at short distances from the trench.

Using Peacock et al.’s (2005) model of present-day thermal structure, it is estimated that the temperature at the plate interface down to within the mantle wedge ranges from approximately 100°C to 500°C. Over this temperature range, fluid release is induced by mineral transformations in subducted sediments and by porosity collapse in both sediments and oceanic crust below the trench slope, outer high and flanks of the Sandino Basin (Figure 5). Aqueous fluids expelled from the subduction zone make their way upward by buoyancy assisted by fault and fracture permeability in the accretionary crust (Jackson et al., 2024).

The protolith is provided by margin wedge ophiolitic rocks and underlying mantle wedge rocks that come within 10 km of the base of the Sandino Basin. For serpentinization, hydrogen generation is greatest in a high-temperature window of 200°C–300°C (see Jackson et al., 2024 based on work by McCollom and Bach, 2009, Klein et al., 2013, and McCollom et al., 2022). Peacock et al.’s (2005) thermal structure model indicate that the high-temperature window in the forearc basement spans the crust to mantle wedge below the outer high and flanks of the Sandino Basin (Figure 5).

Evidence that serpentinization has taken place along this margin is provided by the drilling of serpentinized ophiolitic rocks along the trench slope off Guatemala (DSDP, 2007a), seismic velocity studies (van Avendonk et al., 2011), and geophysical models (Salares et al., 2013). The occurrence of methane and other hydrocarbons in surface seeps (Sahling et al., 2008), in DSDP boreholes, and in the form of gas hydrates along the trench slope could indicate the presence of hydrogen from serpentinization. Hydrocarbon gases were encountered at sites 566 and 570 of Leg 84 (DSDP, 2007a) and at site 496 of Leg 67 (DSDP, 2007b), which had to be abandoned due to encountering excess gas. At site 566, hydrocarbon gases were encountered in a fault zone in serpentinite in the trench slope basement. Isotopic analysis by Galimov and Shabayeva (1985) indicates that the hydrocarbon gases are of biogenic origin. We speculate that this could be due to the action of hydrogenotropic methanogens.

The hydrogen generated by serpentinization will either be in solution or gas phase and will make its way by advection into the basin by way of faulting affecting the basement and basin fill. Evidence for connecting faults or faulting and possible gas streaming from basement is presented on seismic data (Figure 6). When in the basin, under lower P-T conditions, hydrogen will migrate in the gas phase and become sealed and trapped in the same way as hydrocarbon gases.

Since the migration-retention subsystem of the hydrogen system works in the same way as the gaseous petroleum system (see Figure 3), hydrogen exploration can use the same techniques and concepts as petroleum exploration. This includes “play-based exploration” as described by Hutchinson et al. (2024) and Jackson et al. (2024). Petroleum exploration by seismic surveying and drilling (four offshore exploration wells) in the Sandino Basin has provided a good exploration database for hydrogen exploration. The database used for this study includes legacy seismic data (processed by DownUnder GeoSolutions for Geoex MCG), a regional 2D data set acquired in 2015 by BGP for Geoex MCG, and a large 3D data set acquired in 2018 under the auspices of Statoil (now Equinor).

Petroleum exploration first targeted Tertiary sandstone reservoirs in anticlines situated in the axial part of the basin. The possibility of a carbonate play along the flanks of the outer high first came to light with seismic refraction studies (e.g., Ranero et al., 2000; Walther et al., 2000) and the interpretation of modern reflection seismic data. The carbonate play was recognized by Struss et al. (2008) and is believed to have been the main target of the Equinor 3D survey.

Several shallow water/emergent carbonates are known from outcrop in the Tempisque basin, the southern onshore expression of the forearc basin (Bandini et al., 2008). Their regional development has been summarized by Baumgartner et al. (2015) as follows:

• Upper Campanian to Maastrichtian El Viejo Formation shallow-water clastic and related carbonate reef unit consisting of rudistid framestones, bioclastic grainstones, and sandstones. The shallow-water carbonate section is approximately 45 m thick on the southern Santa Elena Peninsula (Azema et al., 1985).

• Upper Paleocene to Early Eocene Barra Honda Formation. These are shallow-water limestones of variable thickness with maximum of 250 m consisting of algal micrites in the lower part and bioclastic limestones in the upper part (Azema et al., 1985).

• Upper Oligocene carbonate bank (calcareous sandstones and bioclastic limestones) of Punta Pelado and Punta Nosara in southern Nicoya (Baumgartner-Mora et al., 2008).

Carbonate bodies have been interpreted from geophysical data in the Sandino Basin. Walther et al. (2000) recognize a high-velocity zone below the TD of Corvina-2. The top of the zone is formed by the base Upper Eocene horizon in the Upper Brito Formation and the lithologies penetrated in the bottom 175 m are noticeably calcareous (as noted by Walther et al., 2000). High-amplitude/high velocity intervals on the flank of the outer high have been interpreted as carbonate ramps or reefs by Walther et al. (2000) and Ranero et al. (2000). Similar high-density bodies have been interpreted as carbonates in the gravity modeling conducted for this study (see Figure 4).

Carbonate developments have also been interpreted from 2D and 3D seismic data on the basis of seismic character, morphology, and stratigraphic relationships along the flanks of the outer high. The architecture of the basin and outer high has been determined from the mapping of two regional seismic horizons dated by a tie to the Corvina-2 well: the mid-Oligocene and the base Upper Eocene (Figure 7a and 7b, respectively). The carbonates are developed between and below these two seismic horizons and are interpreted as part of a major regional backstepping bank-ramp system developing from the Late Cretaceous to the mid-Tertiary (Figure 8).

These carbonates are ideally placed to receive charge from hydrogen migrating from the immediate basement below the outer high and flanks of the Sandino Basin. A qualitative assessment based on seismic character and well data indicates that they have a thick claystone-dominated overburden capable of retaining hydrogen in buried structural or stratigraphic traps. This forms the basis of a carbonate play with a play fairway that has been traced along the outer high by examination of 2D and 3D seismic data (Figure 9).

Within this play fairway, 3D seismic data have enabled the recognition of two prospect leads: the Jinotepe paleo-high updip from Corvina-2 and the Masachapa High. Here, we focus on the Jinotepe paleo-high, which is interpreted as a carbonate bank-ramp buildup with upper and lower tiers bounded by the mid-Oligocene and base Upper Eocene seismic horizons, respectively (Figure 10). This attributes Oligocene to mid-Eocene ages to carbonate buildup in this area. Possible gas effects have been observed in the shallow sedimentary section above the prospect (Figure 11)

To test the veracity of this conclusion, an absolute trough amplitude extraction was generated from 300–350 m and compared to a similar extraction from 500–550 m (Figure 11c and 11d, respectively). The extractions have generated a clear low-, soft-amplitude response in the area associated with the natural hydrogen play, which could indicate gas escape features in this location. At the lower interval, this response is not evident, although other potential zones are identified, which may also indicate gas escape features. Although there is a lack of empirical evidence to suggest that these features are related to hydrogen, gases seem to be emanating from the prospect lead through the sedimentary column above. This is perceived as a positive for the occurrence of a gas accumulation in the Jinotepe prospect lead.

There is a risk that hydrogen occupying the carbonate reservoirs has been diluted (or displaced) by gaseous hydrocarbons expelled from mature petroleum source rocks in the basin. The basin has been explored for hydrocarbons with little success, although gas was encountered in Argonauta-1, and gas effects (amplitude anomalies and gas chimneys) are observed on the 3D seismic data. Struss et al. (2008) carry out 1D and 2D basin modeling in the basin to mitigate the maturity risk in this low heat-flow regime. They identify potential source rocks in the Masachapa and Brito Formations, and their modeling indicates that they become oil mature from the late Miocene with a kitchen restricted to the deepest part of the basin in the east. Struss et al. (2008) conclude that hydrocarbons (including gas) from source rocks at deeper levels of the stratigraphy may charge updip carbonate traps to the west but acknowledge that such sources remain un-proven.

The Central American convergent margin is a geotectonic setting where aqueous fluids are being released by the dehydration of sediments and crust within the hydrous, downgoing Cocos slab. These fluids move upward by buoyancy, resulting in serpentinization of ultramafic rocks in the mantle wedge and accreted crustal components of the upper plate of the forearc. It is proposed that this serpentinization is responsible for the formation of a hydrogen system operating in the forearc offshore Nicaragua.

The configuration and thermal structure of the convergent margin is such that serpentinization is predicted to occur below the outer high and flanks of the Sandino forearc basin. Hydrogen released can access basin sediments via faults formed by accretionary tectonics within the obducted crustal rocks.

For hydrogen systems in continental settings, water is required to circulate deep into basement protoliths to reach optimum serpentinization temperatures (Hutchinson et al., 2024). In the convergent margin model presented here, the source of aqueous fluids is below and readily accessible to the protolith situated in the optimum temperature range. In the case of the Nicaraguan sector of the Central American convergent margin, the serpentinization and hydrogen generation sites lie below the outer high and Sandino Basin of the forearc and hydrogen can access sediments of the basin.

In the basin, shallow-water carbonates, interpreted from seismic data along the flanks of the outer high, form potential reservoirs and traps for hydrogen emanating from the underlying basement. Interpretation of 2D and 3D seismic data has allowed the identification of a hydrogen carbonate play fairway along the outer high system. The Jinotope paleo-high and Masachapa High are presented as prospect leads within this play.

We thank Geoex MCG for allowing the use of 2D and 3D seismic data from offshore Nicaragua and Getech for access to gravity data. We also acknowledge G. Firpo, who conducted the seismic-gravity modeling when he was at ERCL.

Data associated with this research are available and can be obtained by contacting the corresponding author.

Biographies and photographs of the authors are not available.